article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
algorithmic information theory ( ait ) is a the of individual objects , using , and concerns itself with the relationship between computation , information , and randomness .
the information content or complexity of an object can be measured by the length of its shortest description .
for instance the string `` 0101010101010101010101010101010101010101010101010101010101010101 '' has the short description `` 32 repetitions of 01'' , while `` 1100100001100001110111101110110011111010010000100101011110010110 '' presumable has no simple description other than writing down the string itself .
more formally , the ( ac ) of a string @xmath0 is defined as the length of the shortest program that computes or outputs @xmath0 , where the program is run on some fixed universal computer .
a closely related notion is the probability that a universal computer outputs some string @xmath0 when fed with a program chosen at random .
this ( ap ) is key in addressing the old philosophical problem of in a formal way .
the major drawback of ac and ap are their incomputability .
time - bounded `` levin '' complexity penalizes a slow program by adding the logarithm of its running time to its length .
this leads to computable variants of ac and ap , and ( us ) that solves all inversion problems in optimal ( apart from some huge multiplicative constant ) time . ac and ap also allow a formal and rigorous definition of randomness of individual strings that does not depend on physical or philosophical intuitions about nondeterminism or likelihood . roughly , a string is ( ar ) if it is incompressible in the sense that its algorithmic complexity is equal to its length .
ac , ap , us , and ar are the core subdisciplines of ait , but ait spans into many other areas .
it serves as the foundation of the ( mdl ) principle , can simplify proofs in computational complexity theory , has been used to define a universal similarity metric between objects , solves the maxwell demon problem , and many others .
formalizes the concept of simplicity and complexity .
intuitively , a string is simple if it can be described in a few words , like `` the string of one million ones '' , and is complex if there is no such short description , like for a random string whose shortest description is specifying it bit by bit .
typically one is only interested in descriptions or _ codes _ that are effective in the sense that decoders are on some computer .
the universal @xmath1 is the standard abstract model of a general - purpose computer in theoretical computer science .
we say that program @xmath2 is a description of string @xmath0 if @xmath2 run on @xmath1 outputs @xmath0 , and write @xmath3 .
the length of the shortest description is denoted by @xmath4 where @xmath5 is the length of @xmath2 measured in bits .
one can show that this definition is nearly independent of the choice of @xmath1 in the sense that @xmath6 changes by at most an additive constant independent of @xmath0 .
the statement and proof of this invariance theorem in @xcite is often regarded as the birth of algorithmic information theory .
this can be termed kolmogorov s thesis : the intuitive notion of ` shortest effective code ' in its widest sense is captured by the formal notion of kolmogorov complexity , and no formal mechanism can yield an essentially shorter code .
note that the shortest code is one for which there is a general decompressor : the kolmogorov complexity establishes the ultimate limits to how short a file can be compressed by a general purpose compressor .
there are many variants , mainly for technical reasons : the historically first `` plain '' complexity , the now more important `` prefix '' complexity , and many others .
most of them coincide within an additive term logarithmic in the length of the string . in this article
we use @xmath7 for the prefix complexity variant .
a has a separate input tape which it reads from left - to - right without backing up , a separate worktape on which the computation takes place , and a separate output tape on which the output is written .
we define a as the initial segment of the input that is scanned at the time when the machine halts , and the is the string that has been written to the separate output tape at that time . the conditional prefix complexity @xmath8 is the length of the shortest binary program @xmath9 on a universal prefix turing machine @xmath1 with output @xmath0 and input @xmath10 @xcite . for non - string objects
( like numbers @xmath11 , pairs of strings @xmath12 , or computable functions @xmath13 ) one can specify some default coding @xmath14 and define @xmath15 .
the most important properties are : * that @xmath7 is approximable from above in the limit but not computable , * the upper bounds @xmath16 and @xmath17 , * kraft s inequality implies @xmath18 , * the lower bound @xmath19 for `` most '' @xmath0 and @xmath20 for @xmath21 , * extra information bounds @xmath22 , * subadditivity @xmath23 , * symmetry of information @xmath24 , * information non - increase @xmath25 for computable functions @xmath13 , * and coding relative to a probability distribution ( mdl ) + @xmath26 for computable probability distributions @xmath27 , where all ( in)equalities hold within an additive constant .
furthermore , it shares many properties with shannon s entropy ( information measure ) , but @xmath7 has many advantages .
the properties above allow us to draw a schematic graph of @xmath7 as depicted in figure [ figk ] . _ schematic graph of prefix kolmogorov complexity @xmath6 with string @xmath0 interpreted as integer .
@xmath28 for ` most ' @xmath0 and @xmath29 for all @xmath0 for suitable constant @xmath30._,scaledwidth=55.0% ]
solomonoff ( 1964 ) considered the probability that a universal computer outputs some string when fed with a program chosen at random . this algorithmic `` solomonoff '' probability ( ap )
is key in addressing the old philosophical problem of induction in a formal way .
it is based on * ( choose the simplest model consistent with the data ) , * epicurus principle of multiple explanations ( keep all explanations consistent with the data ) , * bayes s rule ( transform the a priori distribution to a posterior distribution according to the evidence , experimentally obtained data ) , * ( universal ) turing machines ( to compute , quantify and assign codes to all quantities of interest ) , and * algorithmic complexity ( to define what simplicity / complexity means ) .
occam s razor ( appropriately interpreted and in compromise with epicurus principle of indifference ) tells us to assign high / low a priori plausibility to simple / complex strings @xmath0 . using @xmath7 as the complexity measure
, one could choose any monotone decreasing function of @xmath7 , e.g. @xmath31 .
the precise definition of ( ap ) , also called universal a priori probability , @xmath32 is the probability that the output of a ( so - called monotone ) universal turing machine @xmath1 starts with @xmath0 when provided with fair coin flips on the input tape .
formally , @xmath33 can be defined as @xmath34 where the sum is over all ( so - called minimal , not necessarily halting , denoted by * ) programs @xmath2 for which @xmath1 outputs a string starting with @xmath0 . since the shortest programs @xmath2 dominate the sum , @xmath32 is roughly @xmath31 .
@xmath33 has similar remarkable properties as @xmath7 .
additionally , the predictive distribution @xmath35 converges rapidly to 1 on ( hence predicts ) any computable sequence @xmath36 .
it can also be shown that @xmath33 leads to excellent predictions and decisions in general stochastic environments .
if married with sequential , it leads to an optimal reinforcement learning agent embedded in an arbitrary unknown environment @xcite , and a formal definition and test of intelligence .
a formally related quantity is the probability that @xmath1 halts when provided with fair coin flips on the input tape ( i.e. that a random computer program will eventually halt ) .
this halting probability , also known as chaitin s constant @xmath37 , or ` the number of wisdom ' has numerous remarkable mathematical properties , and can be used for instance to quantify goedel s incompleteness theorem .
consider a problem to solve for which we have two potential algorithms @xmath38 and @xmath39 , for instance breadth versus depth first search in a finite ( game ) tree .
much has been written about which algorithm is better under which circumstances .
consider the following alternative very simple solution to the problem : a meta - algorithm @xmath40 runs @xmath38 and @xmath39 in parallel and waits for the first algorithm to halt with the answer . since @xmath40 emulates @xmath38 and @xmath39 with half - speed ,
the running time of @xmath40 is the minimum of @xmath41time@xmath42 and @xmath41time@xmath43 , i.e. @xmath40 is as fast as the faster of the two , apart from a factor of 2 .
small factors like 2 are often minor compared to potentially much larger difference in running time of @xmath38 and @xmath39 .
( us ) extends this idea from two algorithms to _ all _ algorithms .
first , since there are infinitely many algorithms , computation time has to be assigned non - uniformly .
the optimal way is that @xmath40 devotes a time fraction of @xmath44 to each ( prefix ) program @xmath2 .
second , since not all programs solve the problem ( some never halt , some just print `` hello world '' , etc . ) @xmath40 has to verify whether the output is really a solution , and if not discard it and continue .
how does this fit into ait ?
a problem of ac @xmath7 is its incomputability .
time - bounded `` levin '' complexity penalizes a slow program by adding the logarithm of its running time to its length : @xmath45 it is easy to see that @xmath46 is just the logarithm of the running time ( without verification ) of @xmath40 , and is therefore computable .
while universal search is nice in theory , it is not applicable in this form due to huge hidden multiplicative constants in the running time .
another restriction is that verification needs to be fast .
hutter @xcite developed a more general asymptotically fastest algorithm , which removes the multiplicative constant and necessity of verification , unfortunately at the expense of an even larger additive constant .
schmidhuber @xcite developed the first practical variants of @xmath40 by carefully choosing the programming language ( @xmath1 ) , allocating time in @xmath40 adaptively , designing training sequences of increasing complexity , reusing subroutines from earlier simpler problems , and various other `` tricks '' .
he also defined the speed prior , which is to @xmath47 what ap is to ac .
the mathematical formalization of the concept of probability or chance has a long intertwined history .
the ( now ) standard axioms of probability , learned by all students , are due to kolmogorov ( 1933 ) . while mathematically convincing , the semantics is far from clear .
frequentists interpret probabilities as limits of observed relatives frequencies , objectivists think of them as real aspects of the world , subjectivists regard them as one s degree of belief ( often elicited from betting ratios ) , while cournot only assigns meaning to events of high probability , namely as happening for sure in our world .
none of these approaches answers the question of whether some _ specific individual _ object or observation , like the binary strings above , is random .
kolmogorov s axioms do not allow one to ask such questions .
von mises ( 1919 ) , with refinements to his approach by wald ( 1937 ) , and church ( 1940 ) attempted to formalize the intuitive notion of one string looking more random than another ( see the example in the introduction ) with partial success .
for instance , if the relative frequency of 1s in an infinite sequence does not converge to 1/2 it is clearly non - random , but the reverse is not true : for instance `` 0101010101 ... '' is not random , since the pair `` 01 '' occurs too often .
pseudo - random sequences , like the digits of @xmath48 , cause the most difficulties .
unfortunately no sequence can satisfy all randomness tests .
the mises - wald - church approach seemed satisfactory untill ville ( 1939 ) showed that some sequences are random according to their definition and yet lack certain properties that are universally agreed to be satisfied by random sequences .
for example , the relative frequency of 1 s in increasingly long initial segments should infinitely often switch from above 1/2 to below 1/2 and vice versa . martin - loef ( 1966 ) , rather than give a definition and check whether it satisfied all requirements , took the approach to formalize the notion of all effectively testable requirements in the form of tests for randomness .
the tests are constructive ( namely all and only lower semi - computable ) ones , which are typically all one ever cares about . since the tests are constructed from turing machines , they can be effectively enumerated according to the effective enumeration of the turing machines they derive from . since the set of sequences satisfying a test ( having the randomness property the test verifies ) has measure one , and there are only countably many tests , the set of sequences satisfying all such tests also has measure one .
these are the ones called algorithmic random|algorithmically `` martin - loef '' random ( ar ) .
the theory is developed for both finite strings and infinite sequences . in the latter case
the notion of test is more complicated and we speak of sequential tests . for infinite sequences one can show that these are exactly the sequences which are incompressible in the sense that the algorithmic prefix complexity of every initial segment is at least equal to their length
. more precisely , the infinite sequence @xmath49 an important result due to g.j .
chaitin and c. schnorr .
this notion makes intuitive sense : a string can be compressed
iff there are some regularities in the string
the string is non - random .
* ml - random sequences can not be effectively constructed .
yet we can give a natural example : the , @xmath37 is a real number between 0 and 1 , and the sequence of bits in its binary expansion is an infinite ml - random sequence . *
randomness of other objects than strings and sequences can also be defined .
* coupling the theory of ar with recursion theory ( downey and hirschfeldt 2007 ) , we find a hierarchy of notions of randomness , at least if we leave the realm of computability according to turing .
many variants can be obtained depending on the precise definition of `` constructive '' .
in particular `` relative randomness '' based on ( halting ) oracle machines leads to a rich field connected to recursion theory . * finally , the crude binary separation of random versus non - random strings can be refined , roughly by considering strings with @xmath50 for some @xmath51 .
if strings are interpreted as ( the expansion of ) real numbers , this leads to the notion of constructive or effective hausdorff ( fractal ) dimension .
despite the incomputability of its core concepts , ait has many , often unexpected , applications .
ait helps to tackle many philosophical problems in the sense that it allows one to formalize and quantify many intuitive but vague concepts of great importance as we have seen above , and hence allows one to talk about them in a meaningful and rigorous way , thus leading to a deeper understanding than without ait .
most importantly , ac formalizes and quantifies the concepts of simplicity and complexity in an essentially unique way .
a core scientific paradigm is occam s razor , usually interpreted as `` among two models that describe the data equally well , the simpler one should be preferred . ''
using ac to quantify `` simple '' allowed solomonoff and others to develop their universal theories of induction and action , in the field of .
ait is also useful in the foundations of thermodynamic and its second theorem about entropy increase , and in particular for solving the problem of . by ( often crudely ) approximating the `` ideal '' concepts ,
ait has been applied to various problems of practical interest , e.g. in linguistics and genetics .
the principle idea is to replace the universal turing machine @xmath1 by more limited `` turing '' machines , often adapted to the problem at hand .
the major problem is that the approximation accuracy is hard to assess and most theorems in ait break down .
the universal similarity metric by vitanyi and others is probably the greatest practical success of ait : a reasonable definition for the similarity between two objects is how difficult it is to transform them into each other .
more formally one could define the similarity between strings @xmath0 and @xmath10 as the length of the shortest program that computes @xmath0 from @xmath10 ( which is @xmath52 ) .
symmetrization and normalization leads to the universal similarity metric .
finally , approximating @xmath7 by standard compressors like lempel - ziv ( zip ) or bzip(2 ) leads to the normalized compression distance , which has been used to fully automatically reconstruct language and phytogenetic trees , and many other clustering problems . see for details and references . in science itself , ait can constructivize other fields : for instance , statements in shannon information theory and classical probability theory necessarily only hold in expectation or with high probability .
theorems are typically of the form `` there exists a set of measure x for which y holds '' , i.e. they are useful for ( large ) samples .
ar on the other hand can construct high - probability sets , and results hold for individual observations / strings . and
real numbers also have constructive counterparts .
naturally , ait concepts have also been exploited in theoretical computer science itself : ait , via the incompressibility method , has resolved many open problems in computational complexity theory and mathematics , simplified many proofs , and is important in understanding ( dissipationless ) reversible computing .
it has found applications in statistics , cognitive sciences , biology , physics , and economics .
ait can also serve as an umbrella theory for other more practical fields , e.g. , in machine learning , the minimum description length ( mdl ) principle can be regarded as a downscaled practical version of ac .
@xcite suggested to define the information content of an object as the length of the shortest program computing a representation of it .
@xcite invented the closely related universal a priori probability distribution and used it for time series forecasting . together with @xcite ,
this initiated the field of algorithmic information theory in the 1960s .
and others significantly contributed to the field in the 1970s ( see e.g. @xcite ) .
in particular the prefix complexity and time - bounded complexity are ( mainly ) due to him .
li and vitanyi @xcite is the standard ait textbook .
the book by calude @xcite focusses on ac and ar , hutter @xcite on ap and us , and downey and hirschfeldt @xcite on ar .
the at http://www.hutter1.net/ait.htm contains further references , a list of active researchers , a mailing list , a list of ait events , and more .
there is still no generally agreed upon notation and nomenclature in the field .
one reason is that researchers of different background ( mathematicians , logicians , and computer scientists ) moved into this field .
another is that many definitions are named after their inventors , but if there are many inventors or one definition is a minor variant of another , things become difficult .
this article uses descriptive naming with contributors in quotation marks .
not even the name of the whole field is generally agreed upon .
_ algorithmic information theory _ , coined by gregory chaitin , seems most appropriate , since it is descriptive and impersonal , but the field is also often referred to by the more narrow and personal term _ kolmogorov complexity_.
the ait field may be subdivided into about 4 separate subfields : ac , ap , us , and ar . the fifth item below refers to applications . *
( ac ) * * philosophical considerations * * properties of ac * * plain ( kolmogorov ) complexity * * prefix complexity * * resource bounded complexity * * other complexity variants * ( ap ) * * occam s razor and epicurus principle * * discrete algorithmic probability * * continuous algorithmic probability = a priori semimeasure * * universal sequence prediction * * the halting probability = chaitin s omega = the number of wisdom * ( us ) * * levin search * * levin complexity and speed prior * * adaptive levin search * * fastest algorithms for general problems * * optimal ordered problem solver * * goedel machines * ( ar ) / recursion theory * * recursion theory * * effective real numbers * * randomness of reals * * van mises - wald - church randomness * * martin - loef randomness * * more randomness concepts and relative randomness * * effective hausdorff dimension * * * minimum description / message length * * machine learning * * artificial intelligence * * computational complexity * * the incompressibility method * * ( shannon ) information theory * * reversible computing * * universal similarity metric * * thermodynamics * * entropy and maxwell demon * * compression in nature | this article is a brief guide to the field of algorithmic information theory ( ait ) , its underlying philosophy , and the most important concepts .
ait arises by mixing information theory and computation theory to obtain an objective and absolute notion of information in an individual object , and in so doing gives rise to an objective and robust notion of randomness of individual objects .
this is in contrast to classical information theory that is based on random variables and communication , and has no bearing on information and randomness of individual objects .
after a brief overview , the major subfields , applications , history , and a map of the field are presented . |
spacecraft exploration of the planets in our solar system and the discovery of exoplanets has attracted considerable attention in the atmospheric escape from planetary objects @xcite .
the cassini spacecraft currently improves our understanding of the atmospheric escape from titan @xcite .
the maven mission circuits around mars for studying its atmospheric composition @xcite . in july 2015 , the new horizons ( nh ) spacecraft completed its flyby of pluto and discovered flowing ice and an extended haze on the planet .
pluto already exhibits a planetary geology that comprises flowing ices , exotic surface chemistry , mountain ranges , and vast haze . analyzing pluto
s atmosphere reveals that pluto s surface has a reddish hue , a simple hydrocarbon in its atmosphere , and the temperature for hazes to form at altitudes higher than 30 kilometers above pluto s surface .
the hydrodynamic escape problem ( hep ) is crucial for investigating of the evolution of planetary atmospheres .
the hep for a single - constituent atmosphere is governed by the following euler equations with gravity and heat : @xmath0 where @xmath1 is the position vector from the center of the planet to the particle of the gas ; @xmath2 , and @xmath3 represent the density , velocity , pressure , and total energy of the gas respectively ; and @xmath4 are the gravitational constant , mass of the planet , and heating , respectively . in this paper , we are concerned with the three - dimensional inviscid hydrodynamic equations without thermal conduction in spherical symmetric space - time models , that is , we considered to be of the following form @xmath5 here , @xmath6 denotes the distance from the center of the planet , @xmath7 and @xmath8 are the altitudes of the inner and outer boundaries of the planetary atmosphere , respectively .
typically , @xmath7 and @xmath8 are the altitudes of the upper thermosphere and exobase .
the total energy @xmath3 is the sum of the kinetic energy and the internal energy of the gas flow , @xmath9 where @xmath10 is the adiabatic constant with @xmath11 .
the steady transonic solutions of are crucial because of an almost hydrodynamic equilibrium state near the bottom boundary .
the hybrid fluid / kinetic model @xcite seems to be realistic approach , which comprises the hydrodynamic escape result @xcite and a drifting maxwell - boltzmann distribution function that includes the bulk velocity @xmath12 in the velocity distribution @xcite .
tian and toon @xcite implemented a numerical simulation using a time - dependent hyperbolic system . a time - independent model experiences singularity at the sonic points @xcite .
for the relaxation methods in @xcite for free conduction , the achieved numerical solutions depend on close guess of initial data with positive velocity .
the first theoretical analysis for a steady hep was reported in @xcite . using the geometric singular perturbation method , the authors constructed smooth transonic stationary solutions issuing from subsonic states to supersonic states and various types of discontinuous stationary solutions for . for the time - evolutionary case ,
the global existence results are yet to be established . in this paper , the global existence of time - evolutionary transonic solutions to the hep in the hydrodynamic region @xmath13 is established .
the gravity and heat affecting intensity can be distinguished during the wave interaction , leading us to the effective development of the numerical simulation .
we define the notations as follows : @xmath14 using , we can rewrite in a compact form @xmath15 where @xmath16 and @xmath17 the complete model of the hep is given by the following free boundary value problem : @xmath18\times[0,\infty ) , \\ & u(x,0)=u_0(x)\in\omega,\ x\in[x_b , x_t ] , \\ & { \rho}(x_b , t)={\rho}_b(t),\ m(x_b , t)=m_b(t),\quad t>0 , \\ & { \rho}\big|_{{\sigma}},\ \frac{m}{{\rho}}\big|_{{\sigma}}>0,\ \mathfrak{kn}(u)\big|_{{\sigma}}\le 1 , \end{split}\right.\end{aligned}\ ] ] where the exobase of the atmosphere @xmath19 ( as well as @xmath20 ) must be determined and @xmath21 denotes the knudsen number of @xmath22 .
physically , the region @xmath20 is called the hydrodynamic region of .
the position of the inner boundary @xmath7 may be probed through astronomical observation . however , determining the outer boundary @xmath8 is usually difficult due to the transition of the kinetic and hydrodynamical regions .
determining the position of the outer boundary @xmath19 and solving in @xmath20 simultaneously is basically a free boundary problem , which makes it difficult to establish the global existence result . to overcome this difficulty
, we first propose the following associated initial - boundary value problem ( ibvp ) without vacuum in @xmath23 : @xmath24 where @xmath25 and @xmath26 is an open domain centered at some sonic state @xmath27 we call the set @xmath28 the transition surface or the sonic states .
the vacuum case is excluded from this formula because the atmospheric gas does not act as fluid when the density tends to zero . whether the glimm method can be applied to the vacuum case for the general system has remained unsolved for decades . in this paper , a new version of the glimm method
is used for establishing the existence of global entropy solutions of under the following conditions : 1 .
@xmath29 , @xmath30 @xmath31 , @xmath32 and @xmath33 are bounded positive functions with small total variations , and there exists @xmath34 sufficiently small such that @xmath35 and @xmath36 for @xmath37 ; 2 .
@xmath38 for @xmath39 and some positive constant @xmath40 ; 3 .
@xmath41 . under the condition @xmath42 ,
consists of global entropy solutions with positive velocity in @xmath43 ( main theorem i ) .
in addition , under a certain constraint of transonic initial data , in the complement of @xmath20 , denoted as @xmath44 , the wave speeds of glimm s approximate solutions @xmath45 to are positive , that is , the entropy flow @xmath22 to which @xmath45 converges is supersonic in @xmath44 so that the waves of @xmath22 in @xmath44 do not move into @xmath20 to interact with the waves in @xmath20 .
moreover , we prove that the knudsen number @xmath21 of @xmath22 in @xmath20 satisfies @xmath46 which implies that @xmath20 fulfills the physical meaning of the hydrodynamic region @xcite . using this strategy
, we can prove that the solution @xmath47 for is indeed the entropy solution of ( main theorem ii ) .
let us review some previous results related to this topic and clarify the motivation of the study .
when @xmath48 , the system is reduced to the strictly hyperbolic system , @xmath49 the entropy solution to the riemann problem was first constructed by lax @xcite . in particular , the solution is self - similar and consists of constant states separated by elementary waves : rarefaction waves , shocks , and contact discontinuities .
furthermore , the global existence of weak solutions to the cauchy problem was established by glimm @xcite , who considered lax s solutions as the build blocks of the scheme .
for the inhomogeneous hyperbolic systems , @xmath50 the cauchy problem was first studied by liu @xcite . for the cauchy problem of the general quasi - linear , strictly hyperbolic system @xmath51
the existence of entropy solutions was first established by dafermos and hsiao @xcite . in @xcite ,
system was studied under dissipative conditions by using the asymptotic expansion of the classical riemann solutions .
the aforementioned conditions contribute considerably in investigating the systems and . in the aforementioned studies ,
the source term was used for generating an extra stationary characteristic field in the riemann problem .
the time - independent wave curves generated by this filed are tangential to the classical 1-wave curves at sonic states , resulting in the nonuniqueness of solutions of a riemann problem .
the total variations of solutions may blow up in a finite time and the blow up phenomenon can be eliminated using a further dissipative assumption .
when @xmath52 and @xmath53 are independent of @xmath6 , bianchini - bressan @xcite studied the existence result by using regularization method .
luskin - temple @xcite and the authors in @xcite establish the existence result by combining glimm s scheme with the method of fractional steps .
base on these studies , we can consider the effect of the source in as the perturbations of the solutions to the homogeneous conservation laws .
the appearance of the source terms in breaks up the self - similarity of riemann waves .
but the effect is only up to @xmath54 , that is , the effect of the source terms on self - similar waves is of the order @xmath55 in each riemann cell .
more precisely , @xmath56 therefore , we can construct the approximate solution for our generalized riemann problem as @xmath57 in each riemann cell , where @xmath58 solves and @xmath59 solves the linearized system of around @xmath58 .
we have @xmath60 and @xmath61 for some contraction matrix @xmath62 depending on @xmath58 ( as well as @xmath63 ) .
this construction of the generalized riemann solver is in contrast to the fractional step scheme @xcite and other operator splitting methods @xcite .
for instance , in @xcite , the effect of the source on the solutions of the classical riemann problem is decoupled . in our case , the effect of the source on the solutions of the classical riemann problem is strongly coupled .
the estimates of wave interaction are more complicated than in @xcite . for the stability of the generalized glimm scheme , contrary to the methods used in @xcite ,
in which the positivity of the gas velocity is assumed , we can demonstrate , through the structure of the generalized glimm s approximate solution , that the escape velocity of the gas is globally positive , which matches the astronomical observation .
therefore , the uniform bounds of the total variations of the approximate solutions in can be achieved by showing that ( 1 ) the glimm functionals of @xmath58 are nonincreasing in time and ( 2 ) the perturbations have a uniform bound of the total variations in each time step .
in addition , we prove positivity of the gas velocity through a rigorous mathematical proof . based on the contraction matrix @xmath64 , we can achieve a more accurate formula of wave interaction estimates that lead to the decay result of the glimm functionals , and a new relation between the velocity and the glimm functionals in each time step .
consequently , the stability of the generalized glimm scheme and the global positivity of the gas velocity is obtained , as shown in section 3 .
based on @xmath42 and the estimation of interaction , the existence of global entropy solutions in transonic gas flow without any dissipative condition is established . we now introduce the definitions of weak solutions and entropy solutions for , and state the main theorems .
consider the initial - boundary value problem in .
a measurable function @xmath65 is a weak solution of if @xmath66 for any test function @xmath67 .
let @xmath26 be a convex subset of @xmath68 .
a pair @xmath69 is an entropy pair of if @xmath70 is convex on @xmath26 and @xmath71 furthermore , a measurable function @xmath22 is an entropy solution of if @xmath22 is a weak solution of and satisfies @xmath72 for every entropy pair @xmath69 and any positive test function @xmath73 . *
main theorem i. * _ consider the initial - boundary value problem with transonic initial data @xmath74 .
assume that the inner boundary data @xmath75 satisfies the condition @xmath42 and the heat @xmath76 satisfies the condition ( @xmath77 ) .
let @xmath78 be the sequence of approximate solutions of by using the generalized glimm scheme .
then , there exist a null set @xmath79 and a subsequence @xmath80 such that if @xmath81 , then @xmath82 is the positive entropy solution of .
in particular , the gas velocity is positive in @xmath43_. + * main theorem ii . * _ assume that the transonic initial data @xmath74 such that @xmath83 is decreasing and @xmath84 is increasing and @xmath85 , where @xmath86 as defined in .
there exists @xmath87 depending on the initial and boundary data such that @xmath88\times[0,\infty)$ ] is the hydrodynamic region of , which means @xmath89 is the global entropy solution of satisfying @xmath90 . _
+ this paper is organized as follows .
section 2 presents the generalized solvers for riemann and boundary - riemann problems based on the construction of the approximate solutions to these problems through the asymptotic expansion and operator splitting techniques .
the residuals of the solutions in each grid are calculated to preserve the consistency of the proposed scheme .
section 3 presents a generalized version of the glimm scheme .
moreover , the generalized wave interaction estimate , nonincreasing glimm functional , and estimate for the total variation of the perturbations in each time strip are obtained .
the global existence of the entropy solutions for is proved .
the hydrodynamic region is determined in the final section .
in this section , we introduce a new method of constructing the approximate solutions to the riemann and boundary - riemann problems of , which are the building blocks of the generalized glimm scheme of the hep .
the residuals of the approximate solutions will be estimated for maintaining the consistency of the generalized glimm scheme .
let us select the spatial resolution @xmath91 and the temporal step @xmath92 sufficiently small , which satisfies the courant - friedrichs - lewy ( cfl ) condition @xmath93 we define the inner region at the location @xmath94 and time @xmath95 @xmath96 and the boundary region at the lower boundary @xmath7 and time @xmath95 @xmath97 the riemann problem of in @xmath98 , denoted by @xmath99 , is given by @xmath100 and the boundary - riemann problem of in @xmath101 , denoted by @xmath102 can be expressed as @xmath103 where @xmath104 , and @xmath53 are defined in and ; @xmath105 and @xmath106 are the left and right constant states ; and @xmath107 and @xmath108 are the density and momentum at the boundary @xmath7 , respectively . by setting the source term @xmath109 in and , the corresponding classical riemann and boundary - riemann problems
are denoted by @xmath110 and @xmath111 .
the system is a strictly hyperbolic system whose jacobian matrix @xmath112 has three distinct real eigenvalues : @xmath113 where @xmath114 the corresponding right eigenvectors of @xmath112 are @xmath115 where the total specific enthalpy @xmath116 is @xmath117 and @xmath118 is the _ sound speed _ of the gas . here
, the gas is assumed to be ideal so that the pressure satisfies @xmath119 where @xmath120 is the molar gas constant and @xmath121 is the absolute temperature . according to and
, the sound speed can be expressed as @xmath122 since the laws of thermodynamics indicate that the absolute zero temperature can not be reached by only the thermodynamic process , it allows us to assume that @xmath123 furthermore , we have @xmath124 which implies that the first and third characteristic fields are genuinely nonlinear and the second characteristic field is linearly degenerate .
therefore , the entropy solutions for @xmath125 and @xmath126 consist of either shock waves , rarefaction waves first or third characteristic fields , or contact discontinuities from the second characteristic field . for each @xmath127 ,
the @xmath128-rarefaction wave is a self - similar function @xmath129 which satisfies @xmath130 where @xmath131 is the @xmath132 identity matrix and the admissible @xmath128-shock is a discontinuous function satisfying the rankine - hugoniot condition @xmath133=[f(u)],\ ] ] and lax s entropy condition @xmath134 where @xmath135 is the speed of the shock - front and @xmath136 $ ] denotes the difference of states across the shock .
according to lax s method @xcite , we can obtain the existence and uniqueness of the entropy solution for @xmath125 .
the solution consists of at most four constant states separated by shocks , rarefaction waves , or contact discontinuity .
however , for @xmath126 , even under the rankine - hugoniot and lax s entropy conditions , we may not obtain the uniqueness of the weak solutions when @xmath137 is near the transition surface @xmath28 , see figure 1 .
+ figure 1 .
two states @xmath138 and @xmath139 connect to @xmath140 by two different waves : + @xmath141 and @xmath142 .
both states satisfy the rankine - hugoniot and lax - entropy conditions . moreover , the total variation of these solutions can be large even @xmath143 , @xmath144 are small . to solve this problem , we impose an additional condition on the solutions : 1 . a weak solution @xmath145 is the entropy solution of @xmath126 if @xmath22 has the least total variation in @xmath146 within all weak solutions of @xmath126 . under the condition @xmath147 , we can select the unique entropy solution for @xmath126 . in addition , the entropy solution does not consist of the 0-speed shock from the first characteristic field attached on the boundary @xmath148 . the following theorem states the existence and uniqueness of entropy solutions for @xmath125 and @xmath126 .
@xcite suppose @xmath149 .
then , there is a neighborhood @xmath150 of @xmath151 such that if @xmath152 , @xmath125 has a unique solution consisting of at most four constant states separated by shocks , rarefaction waves , and contact discontinuity .
moreover , under the additional condition @xmath147 , there exist a neighborhood @xmath153 of @xmath137 and @xmath154 exist such that @xmath155 and @xmath126 admits a unique self - similar solution @xmath22 satisfying @xmath156 .
we can now construct the approximate solutions for @xmath157 and @xmath158 by using theorem 2.1 .
let @xmath159 be the entropy solution of @xmath125 .
then , for @xmath160 , the approximate solution @xmath22 of @xmath157 is given by @xmath161 where @xmath162 is constructed using the following steps : ( 1 ) linearizing of the system around the homogeneous solution @xmath58 .
( 2 ) averaging the coefficient of the linearized system .
( 3 ) applying the operator - splitting method to the modified system .
the detailed construction of @xmath162 is provided in appendix a. on the basis of these steps and and , the approximate solution @xmath65 can be expressed as @xmath163 where @xmath164,\ ] ] and @xmath165 is defined as @xmath166 and @xmath167 furthermore , through complex computation , we have @xmath168 consequently , @xmath169 as @xmath170 or @xmath171 , which is consistent with the entropy solution for homogeneous conservation laws
. moreover , the approximation in continues to be true when @xmath58 is a constant state .
the construction of the approximate solution for @xmath158 is similar to that for @xmath157 .
therefore , the approximate solution for @xmath158 is also given by .
the perturbation @xmath59 in @xmath158 may not satisfy @xmath172 because of , which means that the approximate solution @xmath22 may not match the boundary condition .
however by , the error between the approximation @xmath22 and the boundary data @xmath173 can be estimated by @xmath174 denotes the oscillation of a function @xmath175 in the set @xmath101 .
this indicates that such approximation does not affect the stability and consistency of the generalized glimm method , which will be discussed later . to demonstrate the consistency of the generalized glimm scheme , it is necessary to calculate the residuals of the approximate solutions for riemann and boundary - riemann problems .
given a measurable function @xmath22 , region @xmath176 , and test function @xmath177 , the residual of @xmath22 for in @xmath178 is defined as @xmath179 we have the following estimates .
let @xmath22 and @xmath180 be the approximate solutions of the riemann problem @xmath157 and the boundary - riemann problem @xmath158 respectively .
let @xmath181 be a test function .
suppose that @xmath57 and @xmath182 .
then the residuals of @xmath22 and @xmath180 can be estimated respectively by @xmath183 and @xmath184 where @xmath185 denotes the oscillation of a function @xmath186 in the set @xmath187 , and @xmath98 and @xmath101 are given by and , respectively .
we only demonstrate the calculation of @xmath188 ; the calculation of @xmath189 is similar . without loss of generality ,
let @xmath190 with @xmath191 , and @xmath58 consists of the 1-shock with speed @xmath192 , 2-contact discontinuity with speed @xmath193 , and 3-rarefaction wave with lower speed @xmath194 and upper speed @xmath195 . by and ,
@xmath196 according to the structure of @xmath58 , @xmath197 can be evaluated as @xmath198 + figure 2 .
typical solution to the classical riemann problem @xmath125 .
here @xmath58 consists of different constant states in the region @xmath199 , and self - similar in @xmath200 , where @xmath201 is as shown in figure 2 .
next , according to and and the complicate calculation , @xmath202 where @xmath203 since @xmath204 , @xmath205 , and @xmath206 when @xmath207 .
applying green s theorem and to @xmath208 , we have @xmath209 and @xmath210 next , @xmath211 is estimated as follows . in the region @xmath212\}$ ]
, @xmath58 is the 3-rarefaction wave ; therefore , by the approximate solution @xmath22 satisfies @xmath213 where @xmath214 .
it leads to @xmath215 where @xmath216 , and @xmath217 according to and further calculation , we have @xmath218\dot{\widetilde{u } } \nonumber \\ & & \hspace{2cm}+o(1){\delta}t,\end{aligned}\ ] ] where @xmath219 . applying the integration by parts to @xmath211 along with , we obtain @xmath220 according to , and - @xmath221 we estimate the second and third terms on the right - hand side of .
suppose that the state @xmath222 is connected to the state @xmath223 by 1-shock on the right and the @xmath224 by the 2-contact discontinuity on the left .
then , the rankine - hugoniot condition gives @xmath225 according to and , we obtain @xmath226 because @xmath227 the aforementioned final equality holds because of the complex computation ; therefore , we obtain : @xmath228 similarly , we have @xmath229 therefore , by and , can be rewritten as @xmath230 since @xmath58 is the entropy solution for @xmath231 , then according to the results of @xcite and , we have @xmath232 following , and - , we obtain .
the proof is complete .
in this section , we establish the stability of the generalized glimm scheme and consequently the compactness of the subsequences of approximate solutions @xmath233 to .
the stability , which is always the core of the glimm method , is obtained through the modified wave interaction estimates , nonincreasing property of the glimm functional , and uniform boundedness of the total variations of the perturbations in approximate solutions .
we prove the global existence of entropy solutions to by demonstrating the consistency of the scheme and entropy inequalities for weak solutions at the end of this section . in this subsection
, we introduce a nonstaggered generalized glimm scheme for the initial boundary value problem .
let us discretize the domain @xmath23 into @xmath234 where @xmath235 and @xmath236 are small positive constants satisfying the cfl condition .
the @xmath237th time strip @xmath238 is denoted by @xmath239 suppose that the approximate solution @xmath240 has been constructed in @xmath238 ; then , we choose a random number @xmath241 and define the initial data @xmath242 in @xmath238 by @xmath243 to initiate the scheme at @xmath244 , we set @xmath245 .
the points @xmath246 are called the @xmath247 of the scheme and the points @xmath248 are the mesh points on the boundary @xmath148 . then , @xmath249 in @xmath250 is constructed by solving the set of riemann problems @xmath251 with initial data @xmath252 and the boundary - riemann problem @xmath253 with initial - boundary data @xmath254 moreover , near the boundary @xmath148 , the approximate solution for satisfies the entropy condition @xmath147 in section 2 .
+ figure 3 .
approximate solution for ibvp with mesh curves .
according to , the approximate solution @xmath255 has an explicit representation @xmath256 where @xmath257 consists of the weak solutions to the corresponding classical riemann problems @xmath258 or boundary - riemann problems @xmath259 in @xmath260 .
the cfl condition ensures that the elementary waves in each @xmath260 do not interact with each other before time @xmath261 . repeating this process
, we construct the approximate solution @xmath262 of in @xmath43 by using the generalized glimm scheme with a random sequence @xmath263 , @xmath264 for all @xmath128 . to obtain the desired estimates
, it is favorable to consider curves comprising line segments joining mesh points rather then horizontal lines .
therefore , we define the mesh curves for the non - local glimm functionals introduced in @xcite .
a mesh curve @xmath265 for is a piecewise linear curve that connects the mesh point @xmath266 on the left with @xmath267 or @xmath268 on its right , @xmath269 together with the line segments joining the points @xmath270 and @xmath271 and some portion of the boundary ( see figure 3 and @xcite ) .
simultaneously , the mesh curve @xmath265 divides the domain @xmath272 into @xmath273 and @xmath274 regions such that @xmath274 contains the line @xmath275 . we can partially order two mesh curves by saying @xmath276 ( or @xmath277 is a _
successor _ of @xmath278 ) if every mesh point of @xmath277 is either on @xmath278 or contained in @xmath279 . in particular , @xmath277 is an _ immediate successor _ of @xmath278 if @xmath276 and all mesh points on @xmath277 except one are on @xmath278 .
a diamond region is a closed region enclosed by a mesh curve and its immediate successor . in this subsection
, several types of nonlinear wave interactions will be described and the classical wave strengths of @xmath280 in each time step will be estimated through wave interactions between the classical waves and the perturbations in the previous time step .
connecting all the mesh points through the mesh curves , the domain @xmath43 is decomposed as a union of the sets of diamond , triangular , and pentagonal regions .
the wave interactions can be divided into the following three types : 1 . in each diamond region , the incoming generalized waves from adjacent riemann problems interact with each other and emerge as the outgoing generalized waves of the riemann problem in the next time step ; 2 . in each triangular region , the incoming generalized waves from the riemann problem at the boundary interact with each other and emerge as the outgoing generalized waves of the boundary - riemann problem in the next time step ; 3 . in each pentagonal region ,
two families of incoming generalized waves , one from the boundary - riemann problem and the other from adjacent riemann problem , interact with each other and emerge as the outgoing generalized waves of the riemann problem in the next time step . in each diamond ( or triangular and pentagonal ) region , all the generalized waves comprise classical outgoing waves and perturbations .
therefore , the objective of wave interaction estimates is to estimate how the wave strengths of classical outgoing waves are influenced by the interaction or reflection of generalized incoming waves .
we start with the wave interaction estimates of type ( i ) .
suppose @xmath281 and let @xmath282 denote the generalized riemann solution of @xmath283 connecting the left constant state @xmath151 with the right constant state @xmath137 .
moreover let @xmath284 be the solution of the corresponding classical riemann problem @xmath285 .
then , the _ classical wave strength _ of @xmath282 is defined as the wave strength of @xmath284 , which is expressed as @xmath286 in other words , the jump discontinuity @xmath287 is resolved into @xmath288 , @xmath289 , and @xmath290 such that @xmath291 is connected to @xmath292 on the right by a @xmath293-wave of strength @xmath294 .
note that @xmath284 is independent of the choice of @xmath295 .
we say that an @xmath128-wave and a @xmath293-wave approach if either @xmath296 , or else @xmath297 and at least one wave is a shock . given another @xmath298 with classical wave strength @xmath299 , the _ wave interaction potential _ associated with @xmath300 , @xmath301 is defined as @xmath302 assume that @xmath303 is an immediate successor of @xmath265 .
let @xmath304 denote the diamond region centered at @xmath305 and enclosed by @xmath265 and @xmath303 .
four vertices of @xmath304 , see figure 4 , are denoted by @xmath306 or @xmath307 here @xmath308 are random numbers in @xmath309 .
define the matrix @xmath310 $ ] , where @xmath311 is the right eigenvector of @xmath112 associated with the eigenvalue @xmath312 , and @xmath313 is invertible in @xmath26 .
then , we have the following theorems on the wave interaction estimates .
[ thm3.1 ] let @xmath314 , and @xmath315 be constant states in some neighborhood contained in @xmath26 with @xmath316 and @xmath317 , where @xmath318 , and @xmath319 @xmath320 suppose that the classical wave strengths of the incoming generalized waves across the boundaries @xmath321 and @xmath322 of @xmath304 are @xmath323 + figure 4 .
classical wave strengths in the diamond region @xmath304 . respectively , and the classical wave strength of the outgoing generalized waves across the boundary @xmath324 ( see figure 4 ) is @xmath325 then there exist constants @xmath326 and @xmath327 such that @xmath328 and @xmath329 in particular , is reduced to the classical wave interaction estimate in @xcite when the system is without the source term .
according to the results of @xcite , we have @xmath330 where @xmath311 is the right eigenvector of @xmath112 associated with the eigenvalue @xmath312 , and @xmath331 is the kronecker delta .
in addition , the coefficients in and are all evaluated at @xmath315 .
similarly , @xmath332 where @xmath333 denotes the cubic terms of @xmath235 and @xmath334 . according to - , @xmath335 when @xmath336 and @xmath337 . by ,
@xmath338 according to the taylor expansion of @xmath339 around @xmath315 , @xmath340 according to and together with @xmath341 , the difference of @xmath137 and @xmath151 is expressed as @xmath342 according to and , @xmath343 and @xmath344 by comparing with , and , and using , we obtain @xmath345 where @xmath346 $ ] and @xmath347 .
to estimate , we need to evaluate all terms at the state @xmath315 . after a complex calculation and using the taylor expansion with respect to @xmath55 , the term @xmath348 is estimated as follows : @xmath349 where @xmath350 .
next , we estimate @xmath351 in .
define @xmath352 ; then , @xmath353 where @xmath354 with @xmath355 , and @xmath356 . according to and
, the difference between @xmath357 and @xmath358 is @xmath359 applying the taylor expansion of at @xmath315 along with , and using and , we have @xmath360 where @xmath361 define @xmath362 by a direct calculation of eigenvalues @xmath363 of @xmath364 and @xmath365 of @xmath366 evaluated at @xmath367 , and considering @xmath368 , we obtain @xmath369 @xmath370 we obtain that @xmath371 when @xmath92 is sufficiently small .
therefore , there exist non - singular matrices @xmath372 such that @xmath373\cdot q_{\delta t}^{-1 } , \\
\phi(u_m)=t_{\delta t}\cdot{\operatorname{diag}}[{\mu}_1,{\mu}_2,{\mu}_3]\cdot t_{\delta t}^{-1}.\end{aligned}\ ] ] since @xmath374 , if @xmath375 is selected , then , @xmath376||t_{\delta t}^{-1}|\le\max\limits_{1\le
i\le3}{\mu}_i\leq 1-{\zeta}{\delta}t .
\end{split}\ ] ] finally , by and , we obtain .
the proof is complete . for the cases ( ii ) and ( iii )
, we construct the approximate solution for boundary - riemann problem by . due to the construction of the approximate solutions to the boundary - riemann problems
, the approximate solutions do not match the boundary conditions .
we need to understand how the the errors on the boundary affect the interaction of waves .
therefore , the wave interaction near the boundary is more complicated than the wave interaction in case ( i ) . on the other hand , to estimate the wave interaction near the boundary , the exact direction of wave for each characteristic field must be known . because of the positivity of the initial and boundary velocities , we can prove that the velocity is globally positive , as shown in section 3.3 .
let us denote the boundary data at @xmath237th time strip by @xmath377 and let @xmath378 be the solution of the generalized boundary - riemann problem @xmath379 .
we define the strength of the 0-wave on the boundary as @xmath380 according to @xcite , the following theorem can be proved using the generalized version of goodman s wave interaction estimates near the boundary : ( boundary interaction estimate ) let @xmath381 and @xmath382 as defined previously , and @xmath383 , where @xmath384 are defined as in theorem 3.1 . , ( see figure 5 ) . + figure 5 .
wave strengths in the region near the boundary @xmath148 .
moreover , let @xmath385,\ ( \widehat{u}_b^{n+1},u_r^{n+1}):=[(\widehat{u}_b^{n+1},{u}_b^{n+1},u^{n+1}_z , u_r^{n+1})/(\varepsilon_0,\varepsilon_2,{\varepsilon}_3)]\ ] ] represent the solutions of @xmath386 and @xmath387 , respectively .
assume that @xmath388 $ ] is the 1-wave of @xmath389 right next to @xmath390 on the @xmath237th time strip ( see figure .
then there exists a constant @xmath391 such that @xmath392 where @xmath393 .
the boundary interaction estimate is calculated using the method reported in @xcite .
the estimation is constructed using the following steps .
( 1 ) decompose the wave interaction into two parts , the transmission part and the reflection part , and evaluate the interacted wave strength of these parts .
( 2 ) estimate the effect of the riemann solver on the wave strength . ( 3 ) combine the estimation of ( 1 ) and ( 2 ) to complete the proof .
step ( 1 ) : we divide the interaction of the waves into two parts , the transmission part and the reflection part .
the incoming wave @xmath394 interacts with @xmath395 , and @xmath396 in order ; it generates one penetrative wave @xmath397 through the boundary , see figure 6 , and two reflected waves @xmath398 and @xmath399 from the boundary , see figure 7 . for the transmission part , the wave strength @xmath397 and states @xmath400 and @xmath401 can be determined through the interaction estimates of waves @xmath402 , and @xmath403 . furthermore , according to lemma 4.2 ( a ) in @xcite and the triangle inequality , we obtain @xmath404 . + figure . 6 :
interaction of waves for the transmission part @xmath405 higher order term .
+ figure 7 : interaction of waves for the reflection part @xmath406 higher order term .
for the reflection part , the wave strengths @xmath398 and @xmath399 can be determined through the generalized boundary - riemann problem @xmath407 .
therefore , we estimate the interaction of @xmath398 and @xmath399 and the waves @xmath408 and @xmath403 . for any state @xmath409 ,
let us define @xmath410 according to , @xmath398 and @xmath399 satisfy @xmath411 is connected to @xmath412 on the right by waves @xmath413 , and @xmath414 .
wave strengths @xmath396 and @xmath408 are bounded ; using this and and , we have @xmath415 $ ] be the approximate solution of @xmath416 , for some @xmath417 . according to the result in section 2
, there exists a smooth function @xmath418 connecting two constant states such that @xmath419 and @xmath420 . by and @xmath421
, we obtain @xmath422 according to , @xmath423 step ( 2 ) : let us denote @xmath424 and @xmath425 , where @xmath64 is in . according to , , and , we obtain @xmath426 where @xmath427 , @xmath428 and @xmath429 as in . finally , by and , we have @xmath430 step ( 3 ) : finally , according to and , @xmath431 we complete the proof of the theorem . in this subsection
, we prove the nonincreasing of the glimm functional and provide the estimate of the total variations of the perturbations , which lead to the compactness of subsequences of the approximate solutions for .
let @xmath434 denote the approximate solution for by the generalized glimm scheme described in section 3.1 ; @xmath434 can be decomposed as @xmath435 where @xmath436 is the approximate solution obtained by solving homogeneous conservation laws in each time step and @xmath437 is the perturbation term ; @xmath436 and its total variation are uniformly bounded .
according to the results of @xcite , it can be accomplished that the glimm functional is nonincreasing in time .
let @xmath265 be a mesh curve , @xmath303 be the immediate successor of @xmath265 , and @xmath304 be the diamond region enclosed by @xmath265 and @xmath303 , centered at @xmath438 .
the glimm functional @xmath439 for @xmath440 over @xmath265 is defined as @xmath441 where @xmath442 is a sufficiently large constant , which will be determined later , and @xmath443 here , both constants @xmath444 and @xmath445 will be determined later , @xmath446 , @xmath447 is evaluated at the mesh point @xmath448 , and the presence of @xmath449 is because @xmath432 crosses @xmath265 and locates in some boundary triangle region ( see figure 5 ) .
+ we first consider the case that @xmath265 and @xmath303 differ in the diamond region away from the boundary . according to theorem 3.1 .
, let @xmath450 be the wave interaction potential associated with @xmath451 and @xmath452 and let @xmath453 by the condition @xmath454 , @xmath455 . according to
, we have the following inequalities @xmath456 where @xmath457 are in of theorem [ thm3.1 ] . by , , and , @xmath458 if @xmath442 satisfies @xmath459 for some @xmath39 , then @xmath460 coupling with , we obtain the estimate @xmath461 now , let @xmath462 , @xmath463 , denote the mesh curves that contain all mesh points @xmath464 at time @xmath465 ; therefore , @xmath466 is located on the time strip @xmath467 .
we select the positive number @xmath442 such that @xmath468 where @xmath40 will be determined later . then , @xmath469 . by and adding up recursive relation over all @xmath470 , and using @xmath471 and @xmath472
, we obtain @xmath473 where @xmath474 therefore , if @xmath475 sufficiently small , we have @xmath476 where @xmath477 and @xmath478 is given in . define @xmath479 . by the condition ( @xmath480 ) ,
@xmath481 this implies that @xmath482 .
moreover , @xmath483 and @xmath484 therefore , @xmath485 . according to and and the similarly argument in the previous step , we further obtain @xmath486 and @xmath487 which implies @xmath488 .
based on the selection of the constants @xmath442 and @xmath489 in , it can be verified that @xmath490 continue this process and by using induction , if holds true for all @xmath265 with @xmath491 , @xmath492 , it yields @xmath493 and therefore , @xmath494 in particular , @xmath495 .
therefore , the inequality leads to @xmath496 for @xmath491 , @xmath492 .
next , we consider the case that @xmath303 is an immediate successor of @xmath265 so that they only differ on boundary @xmath497 .
according to the conditions ( @xmath480 ) , , and , we obtain @xmath498 provided that constants @xmath499 , @xmath500 , and @xmath501 . now , let @xmath462 be the mesh curve located on the time strip @xmath467 and include the half - ray @xmath502 .
moreover , let @xmath503 .
if @xmath235 and @xmath504 are sufficiently small , then we have @xmath505 based on and the analysis similar to that in the interior wave interaction , we show that the glimm functional @xmath439 is nonincreasing in time .
therefore , @xmath506 is defined for @xmath507 and @xmath508 .
next , we verify that the total variation of the perturbation is bounded in any time step .
let us denote @xmath509 , where @xmath64 is given in .
then , @xmath510 according to and the definition of @xmath511 , we obtain @xmath512 since @xmath513 is finite , the total variation of @xmath514 is bounded .
because of the boundedness of the total variation of approximate solutions , the constant @xmath40 in can be easily determined by the initial and boundary data , heat profile , and gravity . by , , and the results in @xcite , the following theorem is achieved .
let @xmath515 be as chosen in , and let @xmath262 be an approximate solution of based on the generalized glimm scheme .
then , under the condition @xmath42 , for any given constant state @xmath516 , there exists a positive constant @xmath517 , depending on the radius @xmath518 of @xmath26 , such that if @xmath519 and the condition @xmath520 hold true for with the constant @xmath40 in , then @xmath521 is well - defined for @xmath522 and @xmath91 is sufficiently small .
furthermore , @xmath523 has a uniform bound on the total variation and satisfies the following properties : 1 .
2 . @xmath525 .
3 . @xmath526 .
4 . the velocity @xmath527 .
5 . the density @xmath528 , where @xmath529 is the constant in ( @xmath530 ) . we prove ( ii ) first .
choose a fixed @xmath517 such that @xmath531 . according to , if @xmath532 then , for a sufficiently small @xmath91 , @xmath533 where the constant @xmath40 is defined in .
next , for ( i ) , @xmath534 by and , @xmath535 by using the aforementioned constants @xmath442 and @xmath517 , we obtain that , for a sufficiently small @xmath235 , @xmath536 is defined on @xmath43 when the condition @xmath42 hold true .
in addition , @xmath536 and its total variation are uniformly bounded and independent of @xmath235 .
for ( iii ) , without loss of generality , let @xmath537 , @xmath538 , and let @xmath539 . according to , @xmath540 for some @xmath541 $ ] .
therefore , according to the corollary 19.8 in @xcite , the result ( iii ) is obtained immediately . according to theorem 3.4 . and oleinik s analysis in @xcite , the following theorem for the compactness of the subsequence of @xmath78 holds true .
assume that the condition @xmath42 hold true .
let @xmath542 be a family of approximate solutions obtained using the generalized glimm scheme .
then , there exist a subsequence @xmath543 of @xmath542 and measurable function @xmath22 such that 1 .
@xmath544 in @xmath545 as @xmath546 ; 2 . for any continuous function @xmath52
, we have @xmath547 in @xmath545 as @xmath548 . finally , the global existence of entropy solutions to is presented by demonstrating the consistency of the scheme and entropy inequalities for the weak solutions .
to achieve the consistency , the convergence of the residual to the approximate solutions @xmath542 of is given .
consider the problem with the condition @xmath42 .
assume that @xmath233 is a sequence of approximate solutions for , which are constructed using the generalized glimm scheme .
there exist a null set @xmath549 and subsequence @xmath550 such that the limit @xmath551 is an entropy solution of .
we first calculate the residual of @xmath542 . according to , @xmath552 . for convenience , we omit the symbol @xmath553 in @xmath554 , and @xmath59 in the rest of this subsection .
define @xmath555\times[t_n , t_{n+1}],\ n=0,1,2,\cdots;k=1,2,\cdots,\ ] ] and @xmath556 we concentrate on the case that @xmath557 is away from the boundary , that is , @xmath558 ; the case @xmath559 can be estimated similarly . for a test function @xmath560 , by summing over all @xmath557 , we have the following estimate @xmath561 : @xmath562 where @xmath563 according to glimm s argument in @xcite and , @xmath564 for almost random sequence @xmath565 , where @xmath566 is a probability space of the random sequence . therefore , by , and theorem 3.3 .
, there exist a null set @xmath549 and subsequence @xmath550 such that the limit @xmath567 is a weak solution of .
next , we show that the aforementioned weak solution @xmath22 is indeed an entropy solution satisfying the entropy inequality . given an entropy pair @xmath568 , define @xmath569 by using an argument similar to the proof of theorem 2.2 .
( see appendix b ) , we can estimate the residuals of @xmath570 in @xmath557 as @xmath571 summing over all @xmath557 , we have @xmath572 for a sufficiently small @xmath235 . based on glimm s argument , implies that @xmath573 for every entropy pair @xmath568 and positive test function @xmath73 .
the existence of the entropy solution of is established .
in section 3 , the global existence of the entropy solution to hep is established on the basis of the initial density with a positive lower bound . however , does not fulfill the physical meaning because of which the atmosphere density reaches vacuum as @xmath6 approaches infinity .
therefore , it is necessary to determine the hydrodynamic region , a subset of @xmath23 , in which our solution of is physically well - defined .
this section is devoted to establishing the main theorem ii . for a certain constraint on the transonic initial data @xmath574 , where @xmath83 is a decreasing function and @xmath84 is an increasing function in @xmath575 ,
there exists a region @xmath576\times[0,\infty)$ ] such that the wave speeds of the solutions to are positive in @xmath577 .
next , by adopting the knudsen number of the gas , we prove that there also exists a region @xmath578\times[0,\infty)$ ] such that the gas in @xmath579 no longer acts like fluid .
in other words , fails to model the hep outside @xmath580 , and the governed equations must be replaced by the kinetic equations . according to the assumed initial data described previously
, there exists a nonempty hydrodynamic region of hep such that @xmath581 .
let @xmath582 be the solution of constructed in theorem 3.5 .
, with initial data @xmath574 and boundary data @xmath583
. then , for any @xmath37 , we have @xmath584 for any @xmath37 , let @xmath585 , and let @xmath557 denote the riemann cell containing the point @xmath295 . by and the random choice process ,
the approximate solution in the @xmath237th time step satisfies @xmath586 for some @xmath587 such that @xmath588 .
more precisely , according to , @xmath589 where @xmath590 . since the density and momentum are globally positive according to theorem 3.3 .
, @xmath591{\delta}t,\end{aligned}\ ] ] where @xmath76 is the heat profile .
note that the grid @xmath557 is the domain of dependence of @xmath295 for one time step . by using backward induction on @xmath237 and passing to the limit as @xmath592 in theorem 3.5 .
, the result is obtained .
next , we show that there exist @xmath593 and @xmath576\times[0,\infty)$ ] such that the wave speeds of @xmath594 are positive .
recall that the mach number of @xmath22 is defined as @xmath595 assume that the transonic initial data @xmath74 satisfying the condition @xmath596 , where @xmath83 are decreasing , @xmath84 is increasing , and @xmath85 .
let @xmath409 be the solution of constructed using theorem 3.5 .
; then there exist @xmath597 and @xmath576\times[0,\infty)$ ] such that the characteristic speeds of the solution @xmath65 in @xmath577 are positive . according to
, the mach number @xmath598 satisfies @xmath599 let @xmath600 , where @xmath601 is defined in . according to lemma 4.1 .
, @xmath602 for all @xmath37 and therefore , @xmath603 next , we define @xmath604 according to lemma 4.1 . , @xmath605 the function @xmath606 is increasing based on the assumption of @xmath607 and @xmath84 . since the initial data is transonic with @xmath85 , @xmath608 . on the other hand , according to ( @xmath530 ) , the function @xmath606 is greater than 1 in the far field , when @xmath607 is near @xmath609 which is sufficiently small .
there exists @xmath597 such that @xmath610 , and @xmath611 for @xmath612 .
therefore , @xmath613 denote the region @xmath576\times[0,\infty)$ ] .
we have shown that @xmath614 , for all @xmath615 , that is , the characteristic speeds of the solution @xmath65 in @xmath577 are positive .
next , we want to determine the hydrodynamic region @xmath616 $ ] such that the constructed solution @xmath65 has mathematical and physical significance .
the knudsen number is defined as @xmath617 , the ratio of the mean free path of the molecules , @xmath618 , to the density scale height , @xmath619 , of the atmosphere .
the region of validity of the hydrodynamic equation is often classified using the knudsen number @xmath620 , which is useful for determining whether statistical mechanics or continuum mechanics formulation of fluid dynamics must be used . here
, @xmath621 is the collision cross section used in @xcite , @xmath622 is the number density , @xmath623 is the mass of a molecule , @xmath121 is the temperature of the gas , and @xmath624 is the boltzmann constant .
according to , @xmath625 the knudsen number can be computed as follows : @xmath626 the hydrodynamic equations are applied appropriately , where @xmath627 , so that many collisions occur over relevant length scales keeping the gas in thermal equilibrium . if @xmath628 , the continuum assumption of fluid mechanics maybe no longer be a good approximation because there are few collisions in this level to inhibit a molecule from escaping . for a transonic initial data
@xmath74 as in theorem 4.2 . and @xmath629 , @xmath630 let @xmath409 be the solution of constructed using theorem 3.5 . ; then , there exist @xmath631 and @xmath632\times[0,\infty)$ ] such that @xmath633 .
let us denote @xmath634 and @xmath635 ; then , @xmath636 and @xmath637 are of bounded variation functions based on the construction of solutions in theorem 3.5 .
, and the lower boundedness of the density in @xmath638 . moreover , according to and the assumption of the initial data , @xmath639 that is , @xmath640 is a decreasing function of @xmath6 .
define the function @xmath641 as @xmath642 according to , for any @xmath37 , the approximation of @xmath643 in the @xmath237th time step can be evaluated as @xmath644 where @xmath645 and @xmath646 , and @xmath647 are as defined in lemma 4.1 .
based on the similar argument in lemma 4.1 and , @xmath648 is a decreasing function and @xmath649 .
next , define @xmath650 since @xmath651 for all @xmath37 , @xmath652 according to and the assumption , @xmath653 . on the other hand , according to the decrease of the function @xmath648 and the assumption ( @xmath530 ) , @xmath654 in the far field whenever @xmath607 near @xmath609 .
there exists @xmath631 such that @xmath655 , and @xmath656 for @xmath657 $ ] .
therefore , @xmath658 denote the region @xmath632\times[0,\infty)$ ] .
therefore , the knudsen number @xmath659 for all @xmath660 . finally , according to theorem 4.2 . and theorem 4.3
, we define the hydrodynamic region of hep by using @xmath581 .
the wave speeds of the solution @xmath65 , constructed in theorem 3.5 . , are positive in the region @xmath661 .
moreover , the knudsen number @xmath662 in the region @xmath20 .
therefore , we obtain an entropy solution @xmath65 of that has both mathematical and physical significance in the hydrodynamic region @xmath20 .
here , we describe the construction of the riemann solver in . based on theorem 2.1 .
, let @xmath58 be the entropy solution for @xmath663 , we only need to construct the perturbation @xmath59 of @xmath22 from @xmath58 .
let us denote @xmath664 and consider the linearized system of around @xmath58 with initial data @xmath665 : @xmath666 where @xmath667 , \\ & & b(x , t)=h(x)g_u(x,\widetilde{u})=h(x)\left[\begin{array}{ccc } 0 & 1 & 0 \\ v^2-\tilde{u}^2 & 2\tilde{u } & 0 \\ \frac{{\gamma}-1}{2}\tilde{u}^3-\tilde{u}\widetilde{h } & \widetilde{h}-({\gamma}-1)\tilde{u}^2-v^2 & { \gamma}\tilde{u } \end{array}\right ] , \\ & & c(x , t)=h(x)g(x,\widetilde{u})=h(x)\left[\begin{array}{c } \tilde{m } \\
\tilde{{\rho}}(\tilde{u}^2+v^2 ) \\
\tilde{m}(\widetilde{h}+v^2)-\frac{xq}{2 } \end{array}\right],\end{aligned}\ ] ] and @xmath668 is given in .
due to the appearance of shocks or discontinuity in @xmath58 , the coefficients in may be discontinuous . to obtain the better regularity of approximate solutions for
, the averaging process of the coefficients in with respect to @xmath669 over @xmath670 $ ] is used . for a bounded variation function @xmath671 ,
the average of @xmath671 is defined as @xmath672 in addition , @xmath673 is continuous even across the shock and contact discontinuity .
we construct approximate solution for by solving @xmath674 based on the operator - splitting method , solutions for can be approximated through composing solutions for @xmath675 and @xmath676 since the system admits zero solution .
the solution for can be approximated as the solution of the follows : @xmath677 where @xmath678 and @xmath679 can be derived from @xmath680 .
the averaging process is reasonable because @xmath681 the matrix @xmath678 in has eigenvalues @xmath682 and the corresponding eigenvectors @xmath683 where @xmath684 the transformation matrix of @xmath685 is then given by @xmath686 $ ] so that @xmath685 can be diagonalized as @xmath687.\end{aligned}\ ] ] moreover , the fundamental matrix solved using @xmath688 is @xmath689.\end{aligned}\ ] ] using the transformation matrix @xmath690 , the state transition matrix of @xmath691 is given by @xmath692 where @xmath693 , \\
n_2(x ) & = \frac{1}{2v}\left[\begin{array}{ccc } 0 & 0 & 0 \\ 0 & 0 & 0 \\ ( \tilde{u}_{\star}-v){\delta}_1-(\tilde{u}_{\star}+v){\delta}_2 & { \delta}_2-{\delta}_1 & 2v \end{array}\right ] , \\
n_3(x ) & = \frac{1}{2v}\left[\begin{array}{ccc } \tilde{u}_{\star}+v & -1 & 0 \\ \tilde{u}_{\star}^2-v^2 & -(\tilde{u}_{\star}+v ) & 0 \\ ( \tilde{u}_{\star}+v){\delta}_2 & -{\delta}_2 & 0 \end{array}\right ] . \end{split}\ ] ] according to and the variation of constant formula , the solution for is given by @xmath694+\frac{\tilde{{\rho}}_{\star}}{2}(e^{h(\tilde{u}_{\star}-v)}-1)\left[\begin{array}{c } 1 \\ \tilde{u}_{\star}-v \\ { \delta}_2 \end{array}\right ] \nonumber \\ & & \hspace{2cm}+\frac{\tilde{{\rho}}_{\star}}{2{\gamma}\tilde{u}_{\star}}(e^{{\gamma}h\tilde{u}_*t}-1)\left[\begin{array}{c } 0 \\ 0 \\ -(\tilde{u}_{\star}+v){\delta}_1-(\tilde{u}_{\star}-v){\delta}_2 + 2\tilde{u}_{\star}(\widetilde{h}_{\star}+v^2)-\frac{xq}{\tilde{{\rho}}_{\star } } \end{array}\right ] \nonumber \\ \ ] ] where @xmath695 and @xmath696 $ ] . finally , replacing @xmath680 in by @xmath58 , the perturbation @xmath59 is obtained and is given in the form @xmath697 where @xmath698,\ ] ] with @xmath167
let @xmath568 be an entropy pair of . here , we describe the estimation of the residual @xmath699 @xmath700 where @xmath701 , and @xmath702 are defined as in theorem 3.5 . according to the taylor expansion of @xmath703 and , for any positive test function @xmath73 , @xmath704 in @xcite , if @xmath568 is an entropy pair , then @xmath705 in the sense of distribution on the rectangle @xmath557 .
multiplying the left hand side of by a positive test function @xmath706 , and integrating by parts over @xmath557 , we obtain @xmath707 the estimate of @xmath708 is similar to that of @xmath197 in theorem 2.2 .
let @xmath709 , @xmath710 , and let @xmath711 in @xmath712 consist of the 1-shock with speed @xmath192 , 2-contact discontinuity with speed @xmath193 , and 3-rarefaction wave with lower speed @xmath194 and upper speed @xmath195 .
then , @xmath713 according to and integration by parts , we have @xmath714 and @xmath715 next , we estimate @xmath716 .
according to the same argument as that for - and @xmath717 , we obtain @xmath718 by - , @xmath719 we estimate the second and third terms on the right - hand side of .
suppose that the state @xmath222 is connected to the state @xmath223 by 1-shock on the right , and to the state @xmath224 by 2-contact discontinuity on the left . by the result of @xcite , is replaced by @xmath720 therefore , according to and similar argument as , we obtain @xmath721 according to , , , , and , we obtain the result of : @xmath722 , _ hydrodynamic planetary thermosphere model 1 : response of the earth s thermosphere to extreme solar euv conditions and the significance of adiabatic cooling _ , j. geophys .
planets , 113 ( 2008 ) , e05008 . | the hydrodynamic escape problem ( hep ) , which is characterized by a free boundary value problem of euler equation with gravity and heat , is crucial for investigating the evolution of planetary atmospheres . in this paper , the global existence of transonic solutions to the hep is established using the generalized glimm method . the new version of riemann and boundary - riemann solvers ,
are provided as building blocks of the generalized glimm method by inventing the contraction matrices for the homogeneous riemann ( or boundary - riemann ) solutions .
the extended glimm - goodman wave interaction estimates are investigated for obtaining a stable scheme and positive gas velocity , which matches the physical observation .
the limit of approximation solutions serves as an entropy solution of bounded variations .
moreover , the range of the hydrodynamical region is also obtained .
+ + msc : 35l50 , 35l60 , 35l65 , 35l67 , 76n10 , 85a20 , 85a30 + + keywords : hydrodynamic escape problem ; nonlinear hyperbolic systems of balance laws ; generalized riemann and boundary - riemann problems ; generalized glimm scheme ; hydrodynamic region . |
in many scientific studies the knowledge of derivatives of a given quantity is of particular importance .
for example in theoretical physics , especially in thermodynamics , many quantities of interest require the calculation of derivatives of an underlying thermodynamic potential with respect to some external parameters such as temperature , volume , or chemical potentials .
in many cases the thermodynamic potentials can only be evaluated numerically and one is forced to employ numerical differentiation techniques which are error - prone as any numerical methods .
furthermore , the thermodynamic potential has to be evaluated at the physical point defined by minimizing the thermodynamic potential with respect to some condensates yielding the equations of motion ( eom ) .
generally , these equations can be solved only numerically and thus introduce additional implicit dependencies which makes the derivative calculations even more complicated . even in cases where
the thermodynamic potential and the implicit dependencies on the external parameters are known analytically , the evaluation of higher - order derivatives becomes very complex and tedious and in the end impedes their explicit calculation . in this work we present a novel numerical technique , based on algorithmic differentiation ( ad ) to evaluate derivatives of arbitrary order of a given quantity at machine precision .
compared to other differentiation techniques such as the standard divided differentiation ( dd ) method or symbolic differentiation , the ad produces truncation - error - free derivatives of a function which is coded in a computer program .
additionally , ad is fast and reduces the work required for analytical calculations and coding , especially for higher - order derivatives .
furthermore , the ad technique is applicable even if the implicit dependencies on the external parameters are known only numerically . in ref .
@xcite a comprehensive introduction to ad can be found .
first remarks about the computation of derivative of implicitly defined functions were already contained in @xcite .
however , a detailed description and analysis is not available yet .
additional information about tools and literature on ad are available on the web - page of the ad - community @xcite .
this work is organized in the following way : for illustrations we will introduce an effective model , the so - called linear sigma model with quark degrees of freedom in sec .
[ sec : model ] .
this model is widely used for the description of the low - energy sector of strongly interacting matter . as a starting point the basic thermodynamic grand potential and the eom of this model are calculated in a simple mean - field approximation in order to elucidate the technical problems common in such types of calculations . before we demonstrate the power of the ad method by calculating certain taylor expansion coefficients up to very high orders for the first time in sec .
[ sec : taylor ] , the ad method itself and some mathematical details are introduced in sec .
[ sec : ad ] .
details for the calculation of higher - order derivatives of implicit functions are given in the following sec .
[ sec : impfun ] . in sec .
[ sec : advsdd ] the results of the ad method are confronted with the ones of the standard divided differences ( dd ) method in order to estimate the truncation and round off errors
. finally , we end with a summary and conclusion in sec .
[ sec : summary ] .
in order to illustrate the key points of the ad method we employ a quantum field theoretical model @xcite . this model can be used to investigate the phase structure of strongly interacting matter described by the underlying theory of quantum chromodynamics ( qcd ) .
details concerning this effective linear sigma model ( l@xmath0 m ) in the qcd context can be found in reviews , see e.g. @xcite .
the quantity of interest for the exploration of the phase structure is the grand potential of the l@xmath0 m .
this thermodynamic potential depends on the temperature @xmath1 and quark chemical potential @xmath2 because the particle number can also vary .
it is calculated in mean - field approximation whose derivation for three quark flavors is shown explicitly in @xcite . for the l@xmath0mthe total grand potential @xmath3
consists of two contributions @xmath4 where the first part , @xmath5 , stands for the purely mesonic potential contribution and is a function of two condensates , @xmath6 and @xmath7 .
the second part , @xmath8 , is the quark contribution and depends on the two condensates as well as on the external parameters temperature @xmath1 and , for simplicity , only one quark chemical potential @xmath2 . since the quark contribution arises from a momentum - loop integration over the quark fields , it is given by an integral which can not be evaluated in closed form analytically .
readers who are unfamiliar with the physical details , may simply regard eq .
( [ eq : grand_pot ] ) as an only numerically known function and continue with the reading above eq .
( [ eq : eom ] ) , which introduces an implicit dependency on the parameters @xmath1 and @xmath2 whose treatment with the ad technique is the major focus of this work . explicitly , in mean - field approximation the quark contribution reads @xmath9 where a summation over three quark flavors @xmath10 is included .
the usual fermionic occupation numbers for the quarks are denoted by @xmath11 and for antiquarks by @xmath12 respectively . in this example
only two different single - particle energies , @xmath13 , emerge @xmath14 the first index @xmath15 denotes the combination of two mass - degenerate light - quark flavors ( @xmath16 ) and the other index @xmath17 labels the heavier strange quark flavor .
the expressions in parentheses in @xmath18 are the corresponding quark masses . in this way
, the dependency of the grand potential on the condensates , @xmath19 , @xmath20 enter through the quark masses , which has not been indicated explicitly in eq .
( [ eq : quark_pot ] ) .
the mesonic potential does not depend on the quark chemical potential nor on the temperature explicitly .
it is just a function of the two condensates and reads @xmath21 wherein all remaining quantities , e.g. @xmath22 are constant parameters . since the physical condensates , @xmath23 and @xmath24 ,
are determined by the extrema ( minima ) of the total grand potential with respect to the corresponding fields , they fulfill the equations of motion @xmath25 this in turn introduces an implicit @xmath1- and @xmath2-dependence of both condensates , @xmath26 these quantities represent the physical order parameters which , together with the grand potential , are the basis of the exploration of the phase structure of the model .
we denote the grand potential evaluated at @xmath27 @xmath20 , as @xmath28 in order to find the temperature and chemical potential behavior of the order parameters the integral in eq .
( [ eq : quark_pot ] ) and simultaneously the eom have to be solved numerically .
this already is an example suitable for an ad application , because a derivative of a only numerically solvable , implicit function is needed as input .
later we will be interested in higher - order derivatives of the grand potential with respect to , e.g. , the chemical potential .
for example , the quark number density at the physical point is defined by @xmath29 in cases without an implicit @xmath1- or @xmath2-dependence in the thermodynamic potential some progress can be made by calculating the corresponding derivatives explicitly and solving the corresponding equations numerically .
this might be feasible for lower - order derivatives , in particular , if parts of the derivative calculations can be performed by some computer algebra packages like mathematica or maple .
but for higher - order derivatives this procedure is error - prone and time - consuming and not applicable anymore . in the following sections the algorithmic differentiation technique for implicitly defined functions
is introduced on a general mathematical level .
suppose the function @xmath30 , @xmath31 describing an arbitrary algebraic mapping from @xmath32 to @xmath33 is defined by an evaluation procedure in a high - level computer language like fortran or c. the technique of algorithmic differentiation provides derivative information of arbitrary order for the code segment in the computer that evaluates @xmath34 within working accuracy .
for this purpose , the basic differentiation rules such as , e.g. , the product rule are applied to each statement of the given code segment .
this local derivative information is then combined by the chain rule to calculate the overall derivatives .
hence the code is decomposed into a long sequence of simple evaluations , e.g. , additions , multiplications , and calls to elementary functions such as @xmath35 or @xmath36 , the derivatives of which can be easily calculated .
exploiting the chain rule yields the derivatives of the whole sequence of statements with respect to the input variables . as an example , consider the function @xmath37 with @xmath38 that can be evaluated by the pseudo - code given on the left column of tab .
[ fig1 ] . on the right - hand side ,
the resulting statements for the derivative calculation @xmath39 are given .
@xmath40 for the vector @xmath41 one obtains the first column of the jacobian @xmath42 the vector @xmath43 , @xmath44 . correspondingly , the other unit vectors in @xmath45 yield the other two remaining columns of the jacobian @xmath46 .
table [ fig1 ] illustrates the so - called forward mode of ad , where the derivatives are propagated together with the function evaluation .
alternatively , one may propagate the derivative information from the dependents @xmath47 to the independents @xmath48 yielding the so - called reverse mode of ad . over the past decades , extensive research activities led to a thorough understanding and analysis of these two basic modes of ad , where the complexity results with respect to the required runtime are based on the operation count @xmath49 , i.e. , the number of floating point operations required to evaluate @xmath50 , and the degree @xmath51 of the computed derivatives . using the forward mode , one computes the required derivatives together with the function evaluation in one sweep as illustrated above .
the forward mode yields one _ column _ of the jacobian @xmath52 at no more than three times @xmath49 @xcite .
one _ row _ of @xmath52 , e.g. , the gradient of a scalar - valued component function of @xmath53 , is obtained using the reverse mode in its basic form also at no more than four times @xmath49 @xcite .
it is important to note that this bound for the reverse mode is completely independent of the number @xmath54 of input variables .
this observation is called _ cheap gradient result_. for the application discussed in the present work , the forward mode has been chosen for the efficient computation of higher - order derivatives which is illustrated in the following paragraphs . to this end
, we consider taylor polynomials of the form @xmath55 .
the expansion is truncated at the highest derivative degree @xmath51 which is chosen by the user .
the vector polynomial @xmath56 describes a path in @xmath57 which is parameterized by @xmath58 .
thus , the first two vectors @xmath59 and @xmath60 represent the tangent and the curvature at the base point @xmath61 . assuming that the function @xmath62 is sufficiently smooth , i.e. , @xmath51 times continuously differentiable
, one obtains a corresponding value path @xmath63 the coefficient functions @xmath64 are uniquely and smoothly determined by the coefficient vectors @xmath65 with @xmath66 . to compute this higher - order information ,
first we will examine for a given taylor polynomial @xmath67 the derivative computation based on `` symbolic '' differentiation .
let us generalize the previous relation @xmath68 given in eq .
( [ eq : ytay ] ) , and consider now a general smooth function @xmath69 as for example the evaluation of a @xmath70-function .
this function @xmath71 represents one of the intermediate values computed during the function evaluation as illustrated in table [ fig1 ] .
one obtains for the taylor coefficients @xmath72 the derivative expressions @xmath73 hence , the overall complexity grows rapidly with the degree @xmath51 of the taylor polynomial . to avoid these prohibitively expensive calculations the standard higher - order forward sweep of algorithmic differentiation is based on taylor arithmetic @xcite yielding an effort that grows like @xmath74 times the cost of evaluating @xmath75 .
this is quite obvious for arithmetic operations such as multiplications or additions , where one obtains the recursion shown in table [ tab : arith ] , .taylor coefficient propagation for arithmetic operations [ cols="^,<,^,^",options="header " , ] similar formulas can be found for all intrinsic functions .
this fact permits the computation of higher - order derivatives for the vector function @xmath34 as composition of elementary components .
the ad - tool adol - c @xcite uses the taylor arithmetic as described above to provide an efficient calculation of higher - order derivatives .
for the application considered here , higher - order derivatives of a variable @xmath76 are required , where @xmath47 is _ implicitly _ defined as a function of some variable @xmath77 by an algebraic system of equations @xmath78 naturally , the @xmath54 arguments of @xmath79 need not be partitioned in this regular fashion . to provide flexibility for a convenient selection of the @xmath80 _ truly _ independent variables @xmath81 , let @xmath82 be a projection matrix with only @xmath83 or @xmath84 entries that picks out these independent variables .
hence , @xmath85 is a column permutation of the matrix @xmath86 \in { { \mathbb r}}^{p\times n}$ ] .
then the nonlinear system @xmath87 has a regular jacobian , wherever the implicit function theorem yields @xmath47 as a function of @xmath48 .
therefore , we may also write with @xmath88 @xmath89 for the seed matrix @xmath90^\top \in { { \mathbb r}}^{n \times p}$ ] . now , we have rewritten the original implicit functional relation between @xmath48 and @xmath91 as an inverse relation @xmath92 .
assuming an @xmath93 that is locally invertible we can evaluate the required derivatives of the implicitly defined @xmath94 with respect to @xmath95 using the computation of higher - order derivatives described above in the following way . starting with a taylor expansion eq .
( [ eq : x ] ) of @xmath48 and a corresponding solution @xmath96 of eq .
( [ eq : f ] ) , one obtains for a sufficiently smooth @xmath97 the representation @xmath98 substituting the taylor expansion of @xmath48 into the previous equation yields @xmath99from the comparison of coefficients , it follows that @xmath100 as a next step , the structure of the taylor coefficients @xmath101 is analyzed .
for the first three coefficients , one has @xmath102 due to the definition of @xmath103 , where @xmath104 denotes the derivative of @xmath105 with respect to its argument . for the higher - order coefficients
, it is now shown that they have the structure @xmath106 where @xmath107 involves only derivatives of order @xmath108 with respect to @xmath58 and hence @xmath109 for @xmath110 , one obtains @xmath111 now , let the assumption hold for @xmath108 . then , one obtains for @xmath112 the equation @xmath113 \bigg|_{t=0 } \\ & = \frac{1}{j } \left({\bf h_z } ( { \bf z}({\bf x}(t ) ) ) \frac{\partial^j}{\partial t^j } { \bf z}({\bf x}(t))\right ) \bigg|_{t=0 } \\ & \phantom{=. } + \frac{1}{j } \left[\left ( \frac{\partial}{\partial t } { \bf h_z } ( { \bf z}({\bf x}(t ) ) ) \right ) \left ( \frac{\partial^{j-1}}{\partial t^{j-1 } } { \bf z}({\bf x}(t ) ) \right)\right ] \bigg|_{t=0 } \\ & \phantom{=.}+ \frac{1}{j}\left ( \frac{\partial } { \partial t } { \bf \tilde{h}}_{j-1 } ( { \bf z}({\bf x}(t)))\right ) \bigg|_{t=0}.\end{aligned}\ ] ] due to the assumptions , the function @xmath114 involves only derivatives of order @xmath108 with respect to @xmath58 since @xmath115 does only contain derivatives of order @xmath116 with respect to @xmath58 .
therefore , ( [ eq:14 ] ) is proven and it follows that @xmath117 due to the definition of @xmath103 . combining ( [ eq:15 ] ) with ( [ eq:2 ] ) ,
one obtains the equations @xmath118 and therefore @xmath119 where the jacobian @xmath120 and its factorization can be reused as long as the argument @xmath121 is the same . for this purpose , the jacobian @xmath120 can be evaluated exactly by using the forward mode of ad .
therefore , it remains to provide the missing contributions @xmath122 to compute the desired taylor coefficients @xmath103 .
one starts with the taylor expansion @xmath123 for @xmath124 , one performs the following steps 1 . a forward mode evaluation of degree @xmath112 . since @xmath125 this yields only the contribution @xmath122 .
2 . one system solve @xmath126 to compute @xmath103 .
this approach provides the complete set of taylor coefficients of the taylor polynomial @xmath127 that is defined by a given taylor polynomial for @xmath48 .
these taylor coefficients of @xmath127 are computed for a considerably small number of taylor polynomials @xmath56 to construct the desired full derivative tensor for the implicitly defined function @xmath91 according to the algorithm proposed in @xcite .
consider the following two nonlinear expressions @xmath128 describing the relation between the cartesian coordinates @xmath129 and the polar coordinates @xmath130 in the plane .
assume , one is interested in the derivatives of the second cartesian and the second polar coordinate with respect to the first cartesian and the first polar coordinate .
then one has @xmath131 , @xmath132 , @xmath133 , @xmath134 , and @xmath135 .
the corresponding projection and seed matrix are @xmath136 provided the argument @xmath91 is consistent in that its cartesian and polar components describe the same point in the plane , one has @xmath137 . in this simple case , one can derive for the implicitly defined functions @xmath138 and @xmath139 the desired derivatives explicitly by symbolic manipulation : @xmath140 the derivatives up to order 3 of @xmath141 will be used to verify the results from the differentiation of the implicitly defined functions .
these derivatives have the following representation : @xmath142 as shown in @xcite , these derivatives can be computed efficiently and exactly from a considerable small number of univariate taylor expansions like the one in eq . .
furthermore , this article also proposes a specific choice of the employed taylor polynomials .
for the example considered here , i.e. , @xmath132 and @xmath143 , one obtains the taylor expansions @xmath144 then , the procedure described in the previous section yields the following taylor expansions of @xmath145 with the base point @xmath146 @xmath147 { \bf z}^2 = { \bf z}_0 + \left ( \begin{array}{r } \;2 \\ -1 \\ \;1 \\ \!- \frac{2}{5 } \end{array } \right ) t + \left ( \begin{array}{c } \;\;0 \\ \!- \frac{2}{3 } \\ \;\;0 \\ \!- \frac{2}{75 } \end{array } \right ) t^2 + \left ( \begin{array}{c } \;\;0 \\ \!- \frac{2}{9 } \\ \;\;0 \\ \!- \frac{46}{1125 } \end{array } \right ) t^3 \end{array}\end{aligned}\ ] ] @xmath148 { \bf z}^4 = { \bf z}_0 + \left ( \begin{array}{r } 0\\ 5\\ 3\\ \frac{4}{5 } \end{array } \right ) t + \left ( \begin{array}{c } \;\;0 \\ \!- \frac{8}{3 } \\ \;\;0 \\ \!- \frac{68}{73 } \end{array } \right ) t^2 + \left ( \begin{array}{c } 0 \\ \frac{40}{9 } \\ 0 \\ \frac{1508}{1125 } \end{array } \right ) t^3 . \end{array}\end{aligned}\ ] ] from these numerical values , one can derive the desired derivatives in eq .
( [ eq:17 ] ) as given below @xmath149 { \displaystyle}\frac{\partial^2 y_1}{\partial z^2_1 } & = - { \displaystyle}\frac{25}{27 } = \frac{2}{9 } { \bf z}^1_{22 } = { \displaystyle}\frac{2}{9 } * \left(- \frac{25}{6}\right)\\{\displaystyle}\frac{\partial^2 y_1}{\partial z_1 \partial z_3 } & = { \displaystyle}\frac{20}{27 } = - \frac{5}{36 } { \bf z}^1_{22 } + \frac{1}{4 } { \bf z}^2_{22 } + \frac{1}{4 } { \bf z}^3_{22 } - \frac{5}{36 } { \bf z}^4_{22 } \\[.3cm]{\displaystyle}\frac{\partial^2 y_1}{\partial z^2_3 } & = - \frac{16}{27 } = \frac{2}{9 } { \bf z}^4_{22 } = \frac{2}{9 } * \left(- \frac{8}{3}\right ) \\[.3 cm ] { \displaystyle}\frac{\partial^3 y_1}{\partial z^3_1 } & = - { \displaystyle}\frac{100}{81 } = \frac{2}{9 } { \bf z}^1_{32 } = \frac{2}{9 } * - \frac{50}{9 } \\[.3 cm ] { \displaystyle}\frac{\partial^3 y_1}{\partial z^2_1 \partial z_3 } & = { \displaystyle}\frac{95}{81 } = - \frac{5}{27 } { \bf z}^1_{32 } + \frac{2}{3 } { \bf z}^2_{32 } - \frac{1}{3 } { \bf z}^3_{32 } + \frac{2}{27 } { \bf z}^4_{32 } \\[.3 cm ] { \displaystyle}\frac{\partial^3 y_1}{\partial z_1 \partial z^2_3 } & = - { \displaystyle}\frac{88}{81 } = \frac{2}{27 } { \bf z}^1_{32 } - \frac{1}{3 } { \bf z}^2_{32 } + \frac{2}{3 } { \bf z}^3_{32 } - \frac{5}{27 } { \bf z}^4_{32}\\ { \displaystyle}\frac{\partial^3 y_1}{\partial z^3_3 } & = \frac{80}{81 } = \frac{2}{9 } { \bf z}^4_{32 } = \frac{2}{9 } * \frac{40}{9}\end{aligned}\ ] ] where @xmath150 denotes the @xmath151th component of the @xmath152th taylor coefficient of the taylor expansion @xmath112 . as can be seen , the required 24 entries of the first three derivative tensors can be obtained from four univariate taylor expansions .
this computation of tensor entries from the taylor expansions , i.e. the exact coefficients for the taylor coefficients , is derived and analyzed in detail in @xcite . in the procedure described above
, the higher - order forward mode of ad is applied for each value of @xmath112 for @xmath124 . employing in addition to the higher - order forward mode the higher - order reverse mode ,
the number of forward and reverse sweeps can be reduced to @xmath153 . in this case , the values of the required @xmath154 is reconstructed from the information available due to the reverse mode differentiation .
the ad - tool adol - cprovides a corresponding efficient implementation of this algorithm and will be used in the numerical tests below , where the @xmath155- behavior of the higher - order derivative calculation can be observed in the measured runtimes .
in the following the previous general mathematical description of the ad technique is applied to the model example introduced in the beginning .
furthermore , the ad results are then compared to those obtained with the standard divided differentiation ( dd ) method .
the first step for the calculation of the @xmath152-th order derivatives of the grand potential with respect to @xmath2 , @xmath156 by means of the ad technique requires a suitable formulation of @xmath157 .
this can be accomplished by a taylor expansion of the condensates @xmath158 .
the required coefficients , i.e. , the derivatives @xmath159 can be calculated by applying the technique described in the previous section for implicit functions .
the next step consists in the calculation of the derivatives of @xmath157 w.r.t .
@xmath2 by using the taylor expansions of the condensates . in the following the procedure will be exemplified in detail . in this example only one , i.e.
@xmath160 , truly independent variable @xmath161 is considered and the temperature @xmath1 plays the role of a constant parameter .
the generalization to mixed derivatives with respect to @xmath1 and @xmath2 can also be realized but is omitted for simplicity .
firstly , the taylor coefficients for the condensates @xmath162 are needed .
this is done via the inverse taylor expansion capabilities of adol - c . for that purpose
the following function @xmath163 is introduced .
the @xmath164 dimensional argument @xmath165 splits into @xmath132 implicitly defined functions @xmath166 and @xmath167 truly independent variable @xmath168 , cf .
( [ eq : f ] ) .
furthermore , the projection matrix reads @xmath169 . in order to obtain the @xmath2-derivatives of the functions @xmath170 for fixed values of @xmath171
the following steps are required : 1 .
the numerical solution of the eom , see eq .
( [ eq : eom ] ) , yields the values of the condensates @xmath172 at the potential minimum . for these values
the condition @xmath173 with @xmath174 is obviously valid . 2 .
prepare the derivative calculation for @xmath175 usingadol - c .
3 . evaluate the taylor coefficients of @xmath176 at @xmath177 up to the highest derivative degree @xmath51 desired by the user . from now on ,
the taylor expansions of @xmath178 around @xmath179 are labeled as @xmath180 .
these taylor expansions are inserted in the grand potential which leads to the definition @xmath181 the function @xmath182 is exact in the explicit @xmath2-dependence but only exact up to order @xmath51 in the implicit dependence .
thus , the @xmath152-order derivatives of @xmath182 correspond to the derivatives of the original @xmath183 if the derivatives are evaluated at the expansion point @xmath179 and @xmath184 , i.e. , we have @xmath185 this equation is valid only at the point @xmath179 . in order to obtain the desired @xmath183 derivatives at another @xmath186 point , the expansion coefficients of @xmath187 have to be recalculated for each @xmath186 point .
however , this reduces the problem of calculating the derivatives of @xmath183 with only implicitly known functions @xmath162 to the calculation of the @xmath182 derivatives with explicitly known @xmath188 .
finally , the calculation of the derivatives of @xmath182 then requires two steps : 1 .
prepare the derivative calculation for @xmath189 with adol - cat @xmath179 which were chosen for the evaluation of the coefficients of @xmath188 .
2 . evaluate the taylor coefficients of @xmath189 . another method to approximate derivatives of a function
is based on divided differences which is explained in the following .
based on the definition of the derivative of a function @xmath10 at a point @xmath190 @xmath191 the simplest linear approximation for @xmath192 is obtained by calculating the right - hand side of eq .
( [ eq : def_deriv ] ) for a small but finite value of @xmath193 @xmath194 the problem with this approximation is that it involves two types of errors ( cf .
e.g. @xcite ) .
if @xmath193 is too large , the so - called truncation error induced by the used approximation or algorithm to calculate the derivative becomes significant . on the other side , when @xmath193 becomes too small another error , the rounding error yields cancellations in the enumerator of ( [ eq : dd ] ) and spoils the quality of the approximation .
since the two error sources compete with each other , one has to find an optimal value of @xmath193 for which the numerical error of the derivative evaluation is smallest . in general
, this optimal @xmath193 varies with @xmath190 , the point at which the derivative is calculated .
the truncation error is relatively easy to control . by comparing eq .
( [ eq : dd ] ) with a taylor expansion for @xmath195 around @xmath190 one sees that the truncation error of the linear approximation is of @xmath196 , i.e. , the error is a linear function of @xmath193 .
thus , decreasing @xmath193 will also decrease the truncation error .
in addition , by increasing the degree of the expansion the truncation error can also be further improved .
one such improved extrapolation , the richardson expansion , is based on the @xmath54-th order taylor expansion for @xmath195 and @xmath197 around @xmath190 for which a truncation error of the order @xmath198 can be derived . by repeating the algorithm for the determination of the truncation error
, a better approximation for the first derivative @xmath192 can be obtained .
similar improvements of the truncation error for higher - order derivatives are also known . as an example the corresponding approximations for the second derivative @xmath199 with three grid points @xmath200 @xmath201 + \mathcal{o}(h^{2})\ ] ] and with five grid points
@xmath202 + \mathcal{o}(h^{4 } ) \end{gathered}\ ] ] are itemized .
the grid points are given by @xmath203 for completeness the fourth - order derivative is quoted @xmath204 + \mathcal{o}(h^{4})\ .\end{gathered}\ ] ] where at least five grid points are needed for its calculation .
the disadvantage of such type of improvements is that the function has to be evaluated at several different grid points @xmath200 which are located in the vicinity of @xmath190 .
the other error source , the rounding error , depends on the used format of the floating point number representation in the computer .
a single precision ieee floating point number is stored in a 32-bit word , where 8 bits are used for the biased exponent and the fractional part of the normalized mantissa is a 23-bits binary number .
one bit in the ieee format is always reserved for the sign of the number .
a double precision number occupies 64 bits , with the biased exponent stored in 11 bits and the fractional part is stored on the remaining 52 bits .
thus , besides the fact that one can represent only a finite subset of all real numbers , all floating point calculations are furthermore rounded resulting in incorrect values .
the smallest positive number @xmath205 , where the floating point approximation for @xmath206 is indeed larger than one is called the machine precision .
when one rounds to the nearest representable number the machine precision is roughly @xmath207 where @xmath208 is the number of bits used to store the mantissa s fraction . for a single precision representation one finds @xmath209 and for a double precision number calculation @xmath210 .
this means that single precision numbers have at most about 7 accurate digits while double precision numbers have about 16 accurate digits .
but in general , due to the error propagation during the application of approximate algorithms the number of accurate digits for a numerical solution decreases .
therefore , the rounding error will be several orders of magnitude larger for a more complicated calculation such as the one for the thermodynamic potential . to minimize this source of error in the derivative calculation of the thermodynamic potential
, a larger value of @xmath193 is reasonable . in order to estimate these numerical errors and verify the quality of the adol - cevaluations the results of the derivative calculation obtained with ad
are confronted with the dd method . in figs .
[ fig : advsnum_t0 ] and [ fig : advsnum_cep ] the results of a dd evaluation as a function of @xmath193 in comparison with the ad calculation for the second - order and fourth - order derivative of the thermodynamic potential are shown .
[ fig : advsnum_t0 ] shows the @xmath2-derivatives of the potential evaluated at @xmath211 mev which is close to the crossover phase transition in the @xmath186 phase diagram .
one can clearly see the competition of the truncation and rounding errors . for the second - order derivative
the optimal value is around @xmath212 while for the fourth - order derivative a slightly larger value @xmath213 leads to more stable results . in fig .
[ fig : advsnum_cep ] the same derivatives are calculated at the point @xmath214 mev which is near the critical end point in the phase diagram . while in the previous fig .
[ fig : advsnum_t0 ] the rounding error dominates , the truncation error is now more important . for the second - order derivative almost no rounding error is visible in the resolution shown . since the truncation error for the five - point expression , eq .
( [ eq:2nd_five ] ) , is of the order @xmath215 and of the order @xmath216 for the corresponding three - point equation , eq .
( [ eq:2nd_three ] ) , the results of the five - point derivative is indistinguishable already for @xmath213 while for the three - point formula a smaller value of @xmath217 is required . for the fourth - order derivative the interval where the derivative does not vary with @xmath193
is very small . only for @xmath218
the dd result is close to the ad result . in summary
, one realizes that the dd derivatives require a very careful fine - tuning of the @xmath193 value .
the dd result coincides always with the dd results where the @xmath193 variation vanishes .
one finds that the ad technique is more efficient than the dd method .
the dd calculation always requires the evaluation of the function at several points , e.g. , for the fourth derivative five function evaluations are necessary . in our case
this involves the solution of the eom at these five nodes .
this is a time - consuming disadvantage of the dd method . with the ad
the eom need to be solved only once . despite the fact that it is required to generate an internal function representation of the evaluation of the eom solution and of the thermodynamic potential inside of adol - c ,
the ad implementation is much faster .
corresponding runtime measurements are illustrated in fig .
[ fig : mwad_runtime ] . the runtime of the dd approach can be described by the linear function @xmath219 where as the ad runtime performs like @xmath220 .
this result fits perfectly to the computational complexity of the ad approach described in sec .
[ sub:4.3 ] .
as previously illustrated , both error sources for a derivative calculation with the dd method are in general difficult to keep under control , in particular , if higher - order derivatives are involved . however , with the ad method it is possible to obtain higher - order derivatives with very high precision . in the following an explicit example
is given within the already introduced linear sigma model .
higher derivatives are required if one is interested , e.g. , in the extrapolation of monte carlo lattice simulations of strongly interacting matter ( lattice gauge theory ) to finite quark chemical potential . at finite quark chemical potential such types of monte carlo simulations
can not be directly performed @xcite .
one possible extrapolation to finite quark chemical potential is based on a taylor expansion around zero chemical potential @xcite . for this purpose
, we consider the same kind of expansion in the quark - meson system described by the l@xmath0 m .
an example is given by the coefficients in the expansion of the pressure @xmath221 which is related to the thermodynamic potential via @xmath222 at fixed temperature and small values of the quark chemical potential the pressure may be expanded in a taylor series around @xmath223 , @xmath224 where the expansion coefficients are given in terms of derivatives of the pressure @xmath225 the series is even in @xmath226 which reflects the invariance of the partition function under the exchange of particles and antiparticles . in fig .
[ fig : adtaylor ] the expansion coefficients @xmath227 to @xmath228 are shown as function of the scaled temperature @xmath229 . here , @xmath230 is the pseudocritical temperature at which the crossover transition occurs for vanishing chemical potential .
since the first three expansion coefficients @xmath231 and @xmath232 are already known and well - understood we do not show them again @xcite . in lattice gauge theory one
can currently calculate the first five coefficients , @xmath233 @xcite .
the higher coefficients @xmath234 with @xmath235 vanish for temperatures basically outside of a five percent window around @xmath230 .
thus , all coefficients are only shown in the range @xmath236 .
all curves are smooth oscillating functions around zero even up to the @xmath237 derivative order .
the amplitude of the oscillation and the number of roots around @xmath230 increases with the order @xmath54 .
thus , this oscillating behavior of the coefficients obviously requires a smaller @xmath193 in order to decrease the truncation error but then the rounding error increases . already in this example
the error sources are dramatic for such a high degree of derivatives .
therefore it is not reasonable and actually not possible to obtain the higher coefficients with standard techniques such as the dd method .
a novel numerical technique , which is based on algorithmic differentiation , for the calculation of arbitrarily high - order and high - precision derivatives has been presented .
the new feature of the technique is the additional treatment of implicitly defined functions .
in addition , the basic concepts of the algorithmic differentiation for explicit dependencies is discussed . as a demonstration of the successful extension to implicitly defined functions
the ad technique is applied to a quantum - field theoretical model for strongly - interacting matter . in this model
the implicitly defined functions are represented by the underlying equations of motion where the implicitly defined order parameter is known only numerically .
two important error sources namely , the rounding and truncation error , for a derivative calculation in general are discussed in detail . furthermore , the results with the improved ad method are confronted to those obtained by standard divided difference ( dd ) methods . in the comparison
the rounding and truncation errors can clearly be identified . while for a second - order derivative calculation the error sources are still controllable , they become intractable for higher orders . in the model example
higher - order derivative coefficients of a taylor expansion for the pressure are calculated up to @xmath237 order .
since these coefficients are calculated for the first time , no comparison with other results can be performed .
the obtained curves are very stable and smooth functions which demonstrates the power of the novel ad technique . in a forthcoming publication
@xcite this method will be applied to the more realistic polyakov - quark - meson model for three quark flavors @xcite .
the presented ad technique augmented by implicitly defined dependencies can be applied to a wide class of problems , where high - order derivatives are involved .
standard alternative methods for the derivative calculation such as the dd method fail due to uncontrollably increasing errors .
especially , in the case of only numerically known implicit dependencies , an analytic solution is actually not possible . here , the ad method is still applicable and displays its exceptional impact .
the work of mw was supported by the alliance program of the helmholtz association ( ha216/ emmi ) and bmbf grants 06da123 and 06da9047i .
we thank j. albersmeyer , f. karsch , a. krassnigg , r. roth and j. wambach for useful discussions and comments .
gell - mann , m. and levy , m. , nuovo cim . * 16 * ( 1960 ) 705 .
meyer - ortmanns , h. , rev .
* 68 * ( 1996 ) 473 .
schaefer , b .-
j . and wambach , j. , phys . part .
* 39 * ( 2008 ) 1025 .
schaefer , b .-
wagner , m. , phys . rev .
* d79 * ( 2009 ) 014018 . | scientific studies often require the precise calculation of derivatives . in many cases
an analytical calculation is not feasible and one resorts to evaluating derivatives numerically
. these are error - prone , especially for higher - order derivatives . a technique based on algorithmic differentiation
is presented which allows for a precise calculation of higher - order derivatives .
the method can be widely applied even for the case of only numerically solvable , implicit dependencies which totally hamper a semi - analytical calculation of the derivatives . as a demonstration
the method is applied to a quantum field theoretical physical model .
the results are compared with standard numerical derivative methods .
algorithmic differentiation , numerical differentiation , taylor expansion , quantum chromodynamics 02.60.gf , 02.70.bf , 12.38.aw , 11.30.rd |
the non - thermal nebulae around young energetic pulsars ( pulsar wind nebulae - pwne ) have been suspected to accelerate leptons to sufficiently large energies allowing production of tev @xmath1-rays ( e.g. gould 1965 , rieke & weekes 1969 , grindlay & hoffman 1971 , stepanian 1980 ) . in fact , the nebula around the crab pulsar was originally reported by the whipple group as a first tev @xmath1-ray source ( weekes et al .
later , detections of tev @xmath1-rays from nebulae around other pulsars have also been claimed ( e.g. around psr 1706 - 44 - kifune et al .
1995 , chadwick et al . 1998 , vela pulsar - yoshikoshi et al .
1997 , or psr 1509 - 58 ( msh 15 - 52 ) - sako et al .
2000 ) , but some of the reported emission features have not been confirmed by recent more sensitive observations ( see e.g. the results of hess collab . concerning psr 1706 - 44 - aharonian et al . 2005a , or vela pulsar - aharonian et al .
recent advances in the study of the pulsar wind nebulae obtained mainly by observations in the lower energy range ( from radio to x - ray emission ) are reviewed by gaensler & slane ( 2006 ) .
early detections of the tev @xmath1-ray emission from the pwne have been usually interpreted in the so called synchrotron self - compton model ( ssc model ) according to which relativistic leptons inside the nebula produce soft synchrotron photons .
these photons are next up - scattered by the same leptons to the @xmath1-ray energies as a result of inverse compton process ( ic ) .
also soft photons of other nature , microwave background radiation ( mbr ) , infrared or optical background present inside the pnw , are up - scattered by these leptons to @xmath1-ray energies .
the semi - phenomenological model of this type has been elaborated by de jager & harding ( 1992 ) and successfully applied to the best studied tev @xmath1-ray source around the crab pulsar .
these authors apply this model assuming the distribution of the magnetic field inside the pulsar wind nebula obtained by kennel & coroniti ( 1984 ) . using the known spatial distribution of the synchrotron emission inside the nebula
they get the spectra and distribution of relativistic leptons inside the nebula .
next , from the obtained distribution of leptons and soft radiation inside the nebula , they calculate the tev @xmath1-ray emission ( spectra and spatial distribution ) . based on such procedure , de jager &
harding predict that the higher energy tev @xmath1-ray emission should originate closer to the pulsar .
the model has been up - dated by atoyan & aharonian ( 1996 ) by including other possible soft radiation targets inside the nebula and more recently confronted with the observations of the crab nebula by hillas et al .
( 1998 ) . for the application of ssc model to nebulae around other pulsars
see e.g. du plessis et al .
( 1995 ) - psr 1509 - 58 , or de jager et al .
( 1996 ) - vela pulsar .
the classical model of de jager & harding ( 1992 ) applies the multiwavelength observations of the crab nebula predicting the tev @xmath1-ray spectra at the present time . it does not consider the evolution of the nebula in time
the first simple time dependent leptonic model for the pwne has been discussed by aharonian , atoyan & kifune ( 1997 ) and applied to the nebula around psr 1706 - 44 .
the authors analyze the evolution of the equilibrium energy spectrum of leptons injected at a constant rate , with a fixed spectrum ( independent of time ) into the region with a constant magnetic field .
they come to an interesting conclusion that pwne with relatively low magnetic field should become strong sources of @xmath1-ray emission .
therefore , the efficiency of energy conversion from the rotating pulsar to @xmath1-rays should increase with the age of the nebula .
observation of the tev @xmath1-ray emission up to @xmath080 tev by the hegra collab .
( aharonian et al .
2004 ) indicates the existence of particles with energies up to @xmath010@xmath2 ev inside the crab nebula .
acceleration of leptons to such energies requires special conditions in the acceleration region ( low synchrotron energy losses , very efficient acceleration mechanism ) .
note however , that recently the hess collab .
( aharonian et al .
2006b ) reported the @xmath1-ray spectrum from the crab nebula showing an exponential cut - off at energy @xmath014 tev . if real such discrepancy might suggest the presence of an additional component in the crab nebula spectrum above several tev , which could be interpreted as the contribution from relativistic hadrons inside the nebula ( see e.g. bednarek & bartosik 2003 , bb03 ) .
in fact , the existence of relativistic hadrons inside the pwne has been considered in the past by e.g. , cheng et al .
( 1990 ) , atoyan & aharonian ( 1996 ) , and bednarek & protheroe ( 1997 ) .
these hadrons interact with the matter of the supernova remnant and surrounding medium producing pions which decay to @xmath1-rays and neutrinos .
possible contribution of @xmath1-rays from hadronic processes to the observed flux from the crab nebula has been discussed in the above mentioned papers .
for example , atoyan & aharonian ( 1996 ) consider interaction of relativistic leptons and hadrons with the matter inside the crab nebula and conclude that bremsstrahlung and @xmath3 processes might give some interesting contribution to the observed spectrum provided that the effective density of matter inside the nebula is by one order of magnitude larger than the observed average density of matter . in this paper
we discuss a more complete hadronic - leptonic model which has been recently proposed independently by bednarek & bartosik ( 2003 ) and amato , guetta & blasi ( 2003 ) .
let us first review some new observational results in which context this model will be considered .
since the aim of this paper is to review the expected multi - messenger high energy signatures for acceleration of particles in the pwne , here we only mention the most interesting ( in our opinion ) new observational results obtained mainly by the hess collaboration in the tev @xmath1-ray energy range . for a more complete review of the hess results see e.g. , de jager ( 2006 ) , gallant ( 2006 ) , or aharonian ( 2006 ) .
the great advantage of the hess cherenkov telescope system is the ability of morphological studies of tev @xmath1-ray emission .
it has appeared that tev @xmath1-ray emission from the middle - aged pwne is extended and comes from the region up to several parsecs , e.g. msh 15 - 52 ( aharonian et al .
moreover , the emission is often asymmetric which is probably the result of interaction of the supernova with the surrounding medium , e.g. pwne associated with psr b1823 - 13 ( aharonian et al .
2005c ) , kookaburra complex ( aharonian et al .
2006c ) or around the vela pulsar ( aharonian et al .
such asymmetries are difficult to explain by the movement of the pulsars through the nebulae ( e.g. the observed offset of the vela pulsar by @xmath00.5 degree would require its velocity of @xmath0250 km s@xmath4 , while the measured value is only 65 km s@xmath4 , caraveo et al .
another interesting result of the hess collab . is the discovery of @xmath1-ray emission between 550 gev up to 65 tev from the region of the vela pulsar with unexpected very flat spectrum ( index @xmath01.45 ) and the cut - off at @xmath014 tev ( aharonian et al .
this extended tev emission offset from the location of the vela pulsar has been also confirmed the cangaroo collab .
( enomoto et al . 2006 ) .
in fact , all other pwne have the spectral indices in the range @xmath02.2 - 2.4 , i.e. more similar to that observed for the crab nebula .
the tev @xmath1-ray spectral features observed from the vela nebula seem to be more consistent with their hadronic origin ( e.g. bednarek & bartosik 2003 , horns et al .
2006 ) . since in the last few years
no essential development in the modeling of the high energy emission from the pwne in the framework of the pure leptonic models has been obtained , we concentrate here on the recently considered hadronic - leptonic ( hybrid ) models which try to include not only radiation processes due to the injected leptons but also processes due to injection of relativistic hadrons into the nebula .
these recent works were stimulated mainly by the studies of the mechanism of possible acceleration of leptons as a result of their interaction with the alfven waves generated by coherently gyrating heavy nuclei in the pwne ( for review see arons 1998 ) .
such hybrid leptonic - hadronic models also predict additional high energy messengers ( neutrinos , neutrons or charged particles contributing to the observed cosmic rays at the earth ) , whose observation will certainly put new constraints on the high energy processes occurring in the pwne .
possible acceleration of heavy nuclei , extracted from the surface of the neutron star , by the electric field induced in the outer gaps of the inner pulsar magnetosphere was suggested already by cheng et al .
however , these nuclei should partially photo - disintegrate when passing through the non - thermal radiation of the outer gaps .
neutrons from their disintegration are injected into expanding nebula surrounding the energetic pulsar .
the decay products of neutrons , i.e. protons , partially captured inside the nebula , interact with the matter and contribute to the @xmath1-ray emission . such model was suggested for the @xmath1-ray and neutrino production from the crab nebula ( bednarek & protheroe 1997 ) .
the nuclei can be additionally accelerated by the pulsar wind carrying significant amount of energy lost by the pulsar .
they can energetically dominate the relativistic pulsar wind as proposed by hoshino et al .
( 1992 ) and gallant & arons ( 1994 ) .
based on the observations of the wisps in the crab nebula , these authors conclude that the nuclei are accelerated somewhere inside the pulsar wind zone ( without specifying the details of the unknown acceleration process ) . after passing the pulsar wind shock ,
the nuclei generate alfven waves which energy is next absorbed by leptons . as a result , monoenergetic nuclei and leptons with a power law spectrum are injected into the nebula .
the radiation model for the pwne based on the above mentioned acceleration mechanism of leptons has been independently developed by amato , guetta & blasi ( 2003 ) and bednarek & bartosik ( 2003 ) .
these models try to include ( for the first time self - consistently ) relevant leptonic and hadronic radiation processes inside the nebula taking into account also the time evolution of the nebula .
we describe the main features of such a model and obtain results based on the approach presented in bednarek & bartosik .
note , that although both models differ in some details , they give not very different results in the case of the crab nebula the concerning contribution of hadronic processes to the highest energy part of the @xmath1-ray spectrum and possible detection of the pulsar wind nebulae by the future neutrino telescopes . such hybrid hadronic - leptonic model has been also considered as possible explanation of the @xmath1-ray emission from the recently discovered @xmath1-ray nebula associated with the vela pulsar ( horns et al .
2006 ) .
the time evolution of a supernova remnant under the influence of an energetic pulsar is described according to the general picture proposed by ostriker & gunn ( 1971 ) and rees & gunn ( 1974 ) .
let us denote the initial expansion velocity of the bulk matter in supernova envelope by @xmath5 and its initial mass by @xmath6 .
the expansion velocity can increase due to additional supply of energy to the nebula by the pulsar .
it can also decrease due to accumulation of the surrounding matter .
these processes are taken into account in order to determine the radius of the nebula at a specific time , t , by using the energy conservation , @xmath7 where @xmath8 is the pulsar energy loss , @xmath9 , and @xmath10 s changes with time according to @xmath11 , @xmath12 is the initial period of the pulsar , and @xmath13 g is the surface magnetic field of the pulsar
. the nebula increases its mass according to @xmath14 where @xmath15 is the density of surrounding medium and @xmath16 is the radius of the expanding envelope at the time _
, @xmath17 the average density of matter inside the nebula is @xmath18 the above set of equations is solved numerically in order to obtain the basic parameters of the expanding pwn .
also the location of the inner shock inside the pwn can be estimated by balancing the pressure of the expanding nebula with the pressure of the pulsar wind ( rees & gunn 1974 ) , @xmath19 where @xmath20 is the ratio of the magnetic energy flux to the total energy flux lost by the pulsar at the location of the pulsar wind shock at radius @xmath21 .
then , the magnetic field strength at the shock region can be estimated from @xmath22 where @xmath23 and @xmath24 are the radius and the surface magnetic field of the pulsar , and @xmath25 is the light cylinder radius .
the evolution of basic parameters characterizing the pulsar wind nebula , for some initial parameters of the supernova explosion and the pulsar , are shown on fig . 1 in bb03 .
as already noted above we are interested in the scenario in which rotating magnetospheres of neutron stars can accelerate not only leptons but also heavy nuclei , extracted from positively charged polar cap regions .
in fact , different aspects of the high energy phenomena around pulsars , such as the change in the drift direction of the radio sub - pulses ( gil et al .
2003 ) , the existence of morphological features inside the crab nebula ( so called radio wisps ) , and the appearance of extremely energetic leptons inside it ( gallant & arons 1994 ) , can be naturally explained by the presence of heavy nuclei .
arons and collaborators ( e.g. see arons 1998 ) postulate that the lorentz factors of iron nuclei accelerated somewhere in the inner magnetosphere and/or the crab pulsar wind zone should be , @xmath26 where @xmath27 and @xmath28 are the mass and charge of the iron nuclei , @xmath29 is the velocity of light , and @xmath30 is the total electric potential drop across the open magnetosphere , and @xmath31 is the acceleration factor determining the lorentz factor of nuclei in respect to the maximum one allowed by the pulsar electrodynamics .
following arons and collaborators , the authors assumed : ( 1 ) @xmath31 is not very far from unity , the value @xmath32 is adopted ; ( 2 ) iron nuclei take most of the spin down power of the pulsar , @xmath33 , where @xmath34 .
unfortunately , this values are not predicted at present by any model of the ion acceleration in the pulsar wind and can only be constrained by the high energy observations of the pwne .
the iron nuclei are extracted from the neutron star surface and accelerated during the pulsar radio phase when the efficient leptonic cascades heat the polar cup region .
they are farther accelerated in the pulsar wind zone due to the linear increase of the lorentz factor of the outflowing plasma ( as more recently discussed by contopoulos & kazanas 2002 ) or due to reconnection of oppositely directed magnetic fields in the wind ( michel 1982 , coroniti 1990 , lyubarsky & kirk 2001 ) .
the quasi - monoenergetic heavy nuclei , after crossing the pulsar wind shock , generate alfven waves in the down - stream region , which energy is resonantly transfered to leptons present in the wind ( hoshino et al .
1992 ) . as a result
, leptons are accelerated to high energies with a power law spectrum with the spectral index @xmath35 between @xmath36 and @xmath37 ( see gallant & arons 1994 ) , where @xmath38 and @xmath39 are the electron and proton mass , respectively .
the spectrum is normalized in such a way as to get the conversion efficiency of energy from the iron nuclei to the leptons equal to @xmath40 .
note that the radiation from leptons depends on the product of the energy conversion from the pulsar wind to nuclei , @xmath41 , and the acceleration efficiency of positrons by these ions , @xmath40 .
therefore , decreasing the first coefficient and increasing the second , one obtains the same level of radiation from positrons but a lower level of gamma - ray flux from hadronic interactions of ions with the matter inside the nebula . since the dependence of @xmath42 on time is not predicted by any theoretical model in the original paper by bednarek & bartosik ( bb03 ) , this value was kept constant during the evolution of the nebula .
relativistic particles accelerated by the mechanism discussed above are captured inside the pulsar wind nebula losing energy on different processes .
however , nuclei injected into the nebula at a specific time @xmath43 , escape from it at the time @xmath44 , if their diffusion distance in the turbulent magnetic field of the nebula , @xmath45 , is equal to the dimension of the nebula , @xmath16 , at the time @xmath44 .
the diffusion distance in the magnetic field of the nebula is obtained by integration ( bednarek & protheroe 2002 ) , @xmath46 where the diffusion coefficient is taken to be @xmath47 , and @xmath48 is the larmor radius of nuclei depending on the distance from the center of the nebula . @xmath45 and @xmath44 are calculated numerically .
the adiabatic losses of nuclei during their propagation inside the nebula are also included . due to these losses and interactions of nuclei with the matter
the energy of a nucleus at a specific time is @xmath49 where @xmath50 is its energy at the time @xmath43 , k is the inelasticity coefficient for collisions of nuclei with matter , and @xmath51 is the optical depth calculated from the known density of matter inside the nebula .
the details of these calculations are given in sections 4.2 and 4.3 in bednarek & protheroe ( 2002 ) .
all the above mentioned effects have been taken into account when considering leptons and hadrons inside the nebula .
having defined the model of the nebula expansion ( which gives the basic parameters of the nebula ) , and also the injection spectra of particles as a function of time after the supernova explosion ( the age of the nebula ) , the equilibrium spectra of leptons and hadrons inside the nebula can be obtained taking into account different energy loss processes .
leptons injected into the medium of the expanding supernova remnant suffer energy losses mainly on radiation processes , bremsstrahlung , synchrotron , and the inverse compton , and due to the expansion of the nebula .
the rate of their energy losses can be described by @xmath52 where @xmath53 and @xmath54 describe the bremsstrahlung and adiabatic losses ; @xmath55 and @xmath56 the synchrotron and ics energy losses .
the coefficients , @xmath57 , and @xmath56 , depend on time in a complicated way due to the changing conditions in the expanding nebula ( magnetic field , density of matter and radiation ) .
therefore , the above formula can not be integrated analytically at an arbitrary time after supernova explosion . in order to determine energies of leptons , @xmath58 , inside the nebula at a specific time @xmath59 , which have been injected with energies @xmath60 at an earlier time @xmath61
, we use the numerical approach .
however , knowing the parameters of the nebula determined at a given time @xmath61 , the evolution of the equilibrium spectrum of leptons during the time step @xmath62 is determined analytically .
next , the conditions inside the nebula are changed to values which are obtained from the expansion model of the nebula at time @xmath63 .
the equilibrium spectrum of leptons at , @xmath59 , is then obtained by summing over the spectra injected at specific time and over all time steps up to the present observed time @xmath59 , @xmath64 where @xmath65 is the injection spectrum of leptons at time t , @xmath66 , and the jacobian @xmath67 describes the change of energy of lepton during the period @xmath62 .
it is calculated analytically by solving eq .
the example equilibrium spectra of leptons inside the nebula at the specific time after explosion of supernova are given in fig . 3 in bb03 .
in order to calculate the equilibrium spectra of different types of nuclei inside the nebula at its given age , we take into account the energy losses on the interaction of nuclei with the matter of the nebula , the adiabatic energy losses , and the escape of nuclei from the nebula due to diffusion .
the example equilibrium spectra of nuclei inside the nebula at specific age are shown in fig . 2 in bednarek & bartosik ( 2003 ) .
having obtained the equilibrium spectra of relativistic leptons and hadrons as a function of time after supernova explosion we can calculate the energy spectra of photons produced inside the nebula by these particles in different radiation processes . in general , such spectra are characterized by three main components : two due to leptons ( synchrotron , ic ) and one due to hadrons ( @xmath3 decay ) . depending on the parameters of the pulsar , the supernova , the surrounding medium and on the age of the nebula , different components may dominate . for the pwne in a very early stage of development ( less than @xmath0100 yrs )
, @xmath1-rays from the interaction of hadrons with matter may dominate .
for young nebulae ( at the age of @xmath010@xmath68 yrs , the crab type ) , the power of the synchrotron spectrum dominates over that of the ic @xmath1-ray spectrum . in the case of the middle age nebulae ( at the age @xmath010@xmath69 yrs , the vela type ) ,
the power in synchrotron bump may become comparable to the power in the tev @xmath1-ray bump ( produced mainly due to the ic scattering of the mbr ) .
5.5truecm for reasonable initial parameters , the hybrid model predicts also the appearance of the small bump in the multiwavelength spectrum of the crab nebula at the highest energies , see bednarek & bartosik ( 2003 , 2005 ) or amato , guetta & blasi ( 2003 ) .
in fact , recent measurements of the cutoff in the crab nebula spectrum at @xmath014 tev ( aharonian et al .
2006a ) , combined with the previous measurements of the spectrum up to @xmath080 tev ( well described by a simple power law , aharonian et al . 2004 )
, might be interpreted as the evidence of such an additional component at @xmath0100 tev .
based on the hybrid model , bednarek & bartosik ( bb03 ) calculated expected emission from the nebula around the vela pulsar predicting significant contribution to the @xmath1-ray flux from the decay of @xmath3 .
however , in the original calculations it has been assumed that relativistic leptons are injected into the nebula as a result of resonant interactions with nuclei during the whole lifetime of the pulsar .
recent results obtained by the hess collab .
( aharonian et al .
2006a ) show that it may not be the case . as we noted above , the observed tev @xmath1-ray spectrum from the vela nebula
is flatter than @xmath70 , and very nicely resembles the pure hadronic component ( see fig . 7 in bb03 ) .
1 shows the @xmath1-ray spectra obtained in terms of the hybrid model which are re - calculated applying the modified assumption about the acceleration process of leptons by hadrons inside the nebula . in the present calculations ,
leptons are accelerated only during the first 2000 , 3000 , and 5000 yrs after the pulsar birth .
leptons accelerated in the past lose efficiently their energy on radiative and adiabatic processes .
therefore , the equilibrium spectrum of leptons inside the nebula shifts to lower energies with increasing age of the nebula due to the lack of freshly injected leptons at the present time . as a result ,
the ic spectrum obtained in such a modified model cuts - off at lower energies and the flat @xmath1-ray emission above @xmath01 tev due to the decay of @xmath3 can be clearly observed .
such model predicts the appearance of a steep component in the tev @xmath1-ray spectrum below @xmath01 tev which might be used to estimate the duration of the acceleration process of leptons . due to the proximity of the pulsar and diffusion of leptons from the acceleration site close to the pulsar wind shock
, it has been predicted that the tev @xmath1-ray emission from the vela pwn should be extended with characteristic dimension of the order of @xmath05 pc ( bednarek & bartosik 2005 ) .
this is consistent with the hess observations .
note , that the vela nebula is significantly displaced from the location of the vela pulsar which might be the result of unequal pressure on the nebula due to its interaction with the inhomogeneous medium as earlier suggested by blondin , chavalier & frierson ( 2001 ) .
note , however that the morphological studies of the tev @xmath1-ray emission features of the pulsar wind nebula g18.0 - 0.7 , close to the pulsar psr j1825 - 137 , show clear softening of the spectrum with increasing distance from the pulsar ( aharonian et al .
such feature is better explained by the leptonic model in which leptons lose radiatively and adiabatically energy when diffusing outside the pulsar .
provided that the hypothesis on hadron acceleration in some pwne is correct and assuming that significant amount of observed @xmath1-ray flux is due to hadrons , it is possible to estimate the expected neutrino flux and the event rates in the presently planned and constructed 1 km@xmath71 neutrino detectors ( icecube , km@xmath68net ) . such simple estimate , based on the observed tev @xmath1-ray flux , has been performed by amato & guetta ( 2003 ) for a few pwne .
for example these authors argue that @xmath012 neutrino events per km@xmath68 per year should be detected from the crab nebula .
more detailed calculations according to the hybrid model of amato , guetta & blasi ( 2003 ) predicts @xmath05 - 13 neutrino events per yr per km@xmath68 from the crab nebula , provided that hadron lorentz factors reach @xmath010@xmath69 - 10@xmath72 .
the event rates estimated for the crab and vela pulsars , based on a similar model by bednarek ( 2003 ) , are on the level of @xmath01 neutrino event within 1 km@xmath71 detector during 1 year , i.e at the level of the atmospheric neutrino background ( anb ) . however ,
when these nebulae were younger , then the event rates were significantly larger , decreasing with the age of the nebula , @xmath73 , proportionally to @xmath74 ( see fig . 1 in bednarek 2003 ) .
based on the observed fluxes of the tev @xmath1-rays from the pwne by the hess collab . ,
kappes et al .
( 2006 ) predicts on average @xmath75 neutrino events per year above @xmath765 tev from the direction of the vela nebula in the km@xmath68net detector , whereas the estimated anb is of the level of @xmath01 neutrino event .
the hybrid model for @xmath1-ray production in the pwne postulates also that from relatively close sources , the @xmath1-rays and neutrinos might be accompanied by neutrons .
in fact , @xmath1-ray emission up to @xmath065 tev from the direction of the vela nebula , requires the acceleration of hadrons to the lorentz factors of the order of @xmath010@xmath77 . in the past ,
when the vela pulsar was at the present age of the crab pulsar , nuclei might be accelerated to the lorentz factors of the order of magnitude larger , i.e. @xmath010@xmath72 .
the mean free path for neutrons extracted from nuclei with such lorentz factors ( or produced in hadronic interactions of protons ) is already @xmath0100 pc .
so then , significant number of neutrons from the vela pulsar might also reach the earth about @xmath78 yrs ago , from the distance of @xmath79 pc ( caraveo et al .
2001 ) . accepting the hybrid model and keeping in mind that pwne are at
present the best established tev @xmath1-ray sources , we should expect that particles escaping from the pwne might contribute to the cosmic rays ( cr ) in the galaxy .
in fact , pulsars have been suspected since their discovery as main sources of cosmic rays .
ostriker & gunn ( 1969 ) and karakua , osborne & wdowczyk ( 1974 ) postulated acceleration of crs above the knee region in the pulsar wind zones by the large amplitude electromagnetic waves generated by rotating neutron stars .
the contribution of particles accelerated by pulsars to the observed cosmic ray spectrum has was later discussed by cheng & chi ( 1996 ) , bednarek & protheroe ( 2002 ) , giller & lipski ( 2002 ) .
cheng & chi ( 1996 ) propose that nuclei injected from the pulsar magnetospheres during the pulsar glitches can contribute to the knee region in the cr spectrum .
giller & lipski ( 2002 ) derive the initial parameters of the pulsar population inside the galaxy required to explain the observed shape of the cr spectrum and its intensity up to the highest energies .
bednarek & protheroe ( 2002 ) estimate the contribution of heavy nuclei accelerated in the pulsar outer gaps ( cheng , ho & ruderman 1986 ) from the population of galactic radio pulsars to the crs above the knee region taking into account the propagation and escape conditions , adiabatic and collisional energy losses of the nuclei during their propagation in the pulsar wind nebulae .
in fact , a new component in the cr spectrum at energies above @xmath80 ev , supplied e.g. by pulsars , is required by the measurements of the mass composition , which suggests an increase of the average mass above the knee ( e.g. glasmacher et al . 1999 ,
ave et al . 2003 ) . based on the hybrid model , bednarek & bartosik ( 2004 ) calculated the spectra of nuclei which escape from the pwne around the pulsars with assumed initial parameters .
applying a few different models for the initial parameters of the pulsars formed inside the galaxy , concerning the distribution of the surface magnetic fields of the new born neutron stars ( derived by narayan 1987 ) and the initial pulsar periods ( derived by lorimer et al .
1993 ) , the contribution of the pulsar population to the crs in the galaxy has been estimated .
it is concluded that model b for the pulsar population presented in lorimer at al .
( 1993 ) is able to describe satisfactory the shape of the cr spectrum and the features of the mass composition in the energy range between the knee and the ankle ( see figs .
3b and 4 in bednarek & bartosik 2004 ) .
the fine features in the cr spectrum ( the knee region ) might also appear due to the presence of the nearby pulsars ( e.g. erlykin & wolfendale 2004 ) .
however , such pulsars would create an anisotropy above the present observational limit ( bhadra 2006 ) .
also extremely high energy crs , i.e. above the ankle at @xmath81 ev , might be accelerated in the wind regions of pulsars with super strong surface magnetic fields , so called magnetars .
blasi , epstein & olinto ( 2000 ) , propose that the highest energy crs are accelerated locally in our galaxy by the pulsars whose initial periods are shorter than @xmath010 ms and whose surface magnetic fields are @xmath82 g. if such pulsars inject heavy nuclei with energies @xmath83 ev , then their larmor radii are smaller than the dimensions of the galactic halo , allowing their efficient trapping inside the galaxy .
the authors argue that such nuclei can be accelerated to extremely high energies after the moment when supernova envelope becomes transparent .
however it is not clear if they are not captured inside the supernova envelope by the magnetic field of the pwn ?
for example , from the modeling of the pwne ( bb03 ) , it is possible to estimate the magnetic field inside the nebula , @xmath010 - 100 g , and its dimension , @xmath84 cm , at @xmath01 year after explosion . for such parameters ,
the larmor radius for iron nuclei are lower than the dimensions of the nebula allowing an efficient trapping .
then , the adiabatic energy losses of nuclei due to fast expansion of the nebula become important . in a similar model ,
arons ( 2003 ) suggests that pulsars with extreme parameters produced in the whole universe are responsible for the highest energy crs .
the winds of such pulsars disrupt the supernova envelopes allowing in this way acceleration of protons to extremely high energies ( of the order of @xmath85 ev ) , and their escape from the nebula .
the final shape of the cr spectrum at the highest energies in the universe depends on the rate of gravitational energy losses by the pulsar during its early stage after formation .
note that the pwne associated with such very young pulsars in the galaxy should also produce significant fluxes of neutrinos which might be detectable within a few years after their birth , see e.g. berezinsky & prilutsky ( 1978 ) , protheroe , bednarek & luo ( 1998 ) , beall & bednarek ( 2002 ) .
as reported in sect . 2 , the tev sources are often substantially displaced from the positions of the associated middle aged pulsars .
this is probably the effect of the interaction of a parent supernova shock wave with the surrounding inhomogeneous medium .
this expectation seems to be very natural since young supernovae are usually close to the star forming regions containing massive clouds .
particles escaping from the pwne can be partially captured by these high density clouds producing @xmath1-rays in collisions with the cloud matter .
in fact , the tev @xmath1-ray emission from one of the most massive compact stellar association , cyg ob2 , has been detected by hegra collab .
( aharonian et al . 2002 ) .
observation of other 14 young open clusters does not give positive detections due to the insufficient sensitivity ( aharonian et al .
recently , extended tev @xmath1-ray emission , coincident with the location of giant molecular clouds , has been also observed from the direction of the galactic center ( aharonian et al .
such extended @xmath1-ray sources can be also caused by the appearance of nearby energetic pulsars .
if such pulsars have been born typically @xmath010@xmath69 yrs ago with initial periods of the order of a few millisecond and the surface magnetic fields of the order of @xmath010@xmath86 g , then at present they should have periods close to @xmath0200 ms .
such radio pulsars are difficult to discover in the high density regions .
hadrons accelerated during the lifetime of a pulsar can be partially captured in dense ( @xmath87 @xmath88 ) , magnetized ( @xmath89 g ) molecular clouds , producing @xmath1-rays for a relatively long time .
note , that even hadrons with lorentz factors up to @xmath010@xmath90 might be captured in the molecular clouds with dimensions of several parsecs since their larmor radii are small enough .
possible radiation effects from the interactions of the pwne with high density regions have been considered in the context of the @xmath1-ray emission from the galactic center ( bednarek 2002 ) , and also proposed as a possible explanation of the unidentified tev @xmath1-ray source in the cyg ob2 region ( bednarek 2003 ) .
pulsars produce a class of the best established tev @xmath1-ray sources in which particles are accelerated above @xmath0100 tev . observations with
the hess telescopes allow for the first time to study morphology of these objects indicating that the interaction of parent supernova remnants with the surrounding medium strongly influences the evolution of the pulsar wind nebulae .
these interaction effects have not been considered in the past leptonic models for the production of high energy radiation inside pwne and strongly encourage the efforts to work on a much more advanced , time dependent , three dimensional model for the pwne . moreover
, new observations of the pwne at tev @xmath1-rays seem to suggest that acceleration of hadrons to high energies inside the pwne should be also seriously considered .
in fact , the unusual shape of the @xmath1-ray spectrum from the nebula associated with the vela pulsar ( relatively narrow peak at several tev ) and possibly weak evidence of a discrepancy between observations of the crab nebula at the highest energies by hegra and hess can be naturally explained by the existence of an additional component in the tev @xmath1-ray spectra from these two objects which is due to contribution from relativistic hadrons .
since hadrons are weakly coupled to matter , they can diffuse from the pwne , interact with the surrounding medium , and also contribute to the observed cosmic ray spectrum in the galaxy .
therefore , pwne should be considered as one of the serious candidates responsible for the bulk of the cosmic rays above the knee region , i.e. @xmath7610@xmath2 ev in the galaxy .
detection ( or non - detection ) of other neutral radiation from the pwne , i.e. neutrinos and neutrons , should give the answer to this question .
3 aharonian , f.a .
( hegracollab . ) a&a * 393 * l37 ( 2002 ) aharonian , f.a .
( hegracollab . )
apj * 614 * 897 ( 2004 ) aharonian , f.a .
( hess collab . ) a&a * 432 * l9 ( 2005a ) aharonian , f.a .
( hess collab . ) a&a * 435 * l17 ( 2005b ) aharonian , f.a .
( hess collab . ) a&a * 442 * l25 ( 2005c ) aharonian , f.a .
( hess collab . ) a&a * 438 * 827 ( 2006a ) aharonian , f.a .
( hess collab . )
a&a in press , astro - ph/0607333 ( 2006b ) aharonian , f.a .
( hess collab . )
a&a in press , astro - ph/0606311 ( 2006c ) aharonian , f.a .
( hess collab . )
a&a in press , astro - ph/06007548 ( 2006d ) aharonian , f.a .
( hegra collab . ) a&a * 454 * 775 ( 2006e ) aharonian , f.a .
( hess collab . )
* 439 * 695 ( 2006f ) aharonian , f.a .
this volume ( 2006 ) aharonian , f.a , atoyan , a.m. , kifune , t. mnras * 291 * 162 ( 1997 ) amato , e. , guetta , d. aph * 19 * 403 ( 2003 ) amato , e. , guetta , d. , blasi , p. a&a * 402 * 827 ( 2003 ) arons , j. in proc .
on pulsars and their nebulae , mem.soc.ast.ital .
, * 69 * 989 ( 1998 ) arons , j. apj * 589 * 871 ( 2003 ) atoyan , a.m. , aharonian f.a
. mnras * 278 * 525 ( 1996 ) ave m. et al .
aph * 19 * 61 ( 2003 ) beall , j.h .
, bednarek , w. apj * 569 * 343 ( 2002 ) bednarek , w. mnras * 331 * , 483 ( 2002 ) bednarek , w. mnras * 345 * , 847 ( 2003 ) bednarek , w. , bartosik , m. a&a * 405 * , 689 ( 2003 ) ( bb03 ) bednarek , w. , bartosik , m. a&a * 423 * , 405 ( 2004 ) bednarek , w. , bartosik , m. j.phys .
g * 31 * , 1465 ( 2005 ) bednarek , w. , protheroe , r.j .
, prl * 79 * , 2616 ( 1997 ) bednarek , w. , protheroe , r.j .
, aph * 16 * , 397 ( 2002 ) berezinsky , v.s . ,
prilutsky , o.f .
a&a , * 66 * 325 ( 1978 ) bhadra , a. aph * 25 * 226 ( 2006 ) blasi , p. , epstein , r.i . ,
olinto , a.v .
apj * 533 * l123 ( 2000 ) blondin , j.m . ,
chevalier , r.a . ,
frierson , d.m .
apj * 563 * 806 ( 2001 ) caraveo , p.a .
, de luca , a. , mignami , r.p . ,
bignami , g.f .
apj * 561 * 930 ( 2001 ) chadwick , p.m. et al .
aph * 9 * 131 ( 1998 ) cheng , k.s . ,
chi , x. a&a * 306 * 326 ( 1996 ) cheng , k.s .
, ho , c. , ruderman , m. apj * 300 * 500 ( 1986 ) cheng , k.s . , cheung , t. , lau , m.m .
, yu , k.n .
, kwok , w.m .
g * 16 * 1115 ( 1990 ) contopoulos , i. , kazanas , d. apj * 566 * 336 ( 2002 ) coroniti , f.v .
apj , * 349 * 538 ( 1990 ) de jager , o.c . in proc .
`` astrophysical sources of high energy particles and radiation '' , eds . t. bulik , b. rudak , g. madejski ( toru , poland ) , aip * 801 * 295 ( 2006 ) de jager , o.c . , harding a.k
. apj * * ( 1992 ) de jager , o.c .
apj * 457 * 253 ( 1996 ) du plessis et al .
apj * 453 * 746 ( 1995 ) enomoto , r. et al .
( cangaroo collab . )
apj * 638 * 397 ( 2006 ) .
erlykin , a.d . ,
wolfendale , a.w .
aph * 22 * 47 ( 2004 ) gaensler , b.m . ,
slane , p.o .
araa * 44 * 17 ( 2006 ) gallant , y. a. this volume ( 2006 ) gallant , y. a. , arons , j. apj * 435 * 230 ( 1994 ) gil , j. , melikidze , g.i . , geppert , u. a&a * 407 * 315 ( 2003 ) giller , m. , lipski , m. j.phys .
g * 28 * 1275 ( 2002 ) glasmacher , m.a.k .
aph * 10 * 291 ( 1999 ) gould , r.j .
prl * 15 * 577 ( 1965 ) grindlay , j.e . , hoffman , j.a . astrophys.lett . *
8 * 209 ( 1971 ) hillas , a.m. et al .
apj * 503 * 744 ( 1998 ) horns , d. , aharonian , f.a . ,
santangelo , a. this volume ( 2006 ) horns , d. , aharonian , f.a . , santangelo , a. , hoffman , a.i.d . ,
masterson , c. a&a * 451 * , l51 ( 2006 ) hoshino , m. , arons , j. , gallant , y. a. , langdon , b. apj * 390 * 855 ( 1992 ) kappes , a. , hinton , j. , stegmann , c. , aharonian , f.a .
apj , submitted , astro - ph/0607286 ( 2006 ) karakua , s. , osborne , j.l . , wdowczyk , j. j.phys.g * 7 * 437 ( 1974 ) kennel , c.f . ,
coroniti , f.v .
apj * 283 * 710 ( 1984 ) kifune , t. et al .
apj * 438 * l91 ( 1995 ) lorimer d.r . ,
bailes , m. , dewey , r.j . , harrison , p.a .
mnras * 263 * 403 ( 1993 ) lyubarsky , y. , kirk , j.g .
apj , * 547 * 437 ( 2001 ) michel , f.c .
rev.mod.phys . , * 54 * 1 ( 1982 ) narayan r. apj * 319 * 162 ( 1987 ) ostriker , j.p . ,
gunn , j.e .
apj * 157 * 1395 ( 1969 ) ostriker , j.p . , gunn , j.e .
apj * 164 * l95(1971 ) protheroe , r.j . , bednarek , w. , luo , q. aph * 9 * 1 ( 1998 ) rees , m.j . ,
gunn , j.e .
mnras * 167 * 1 ( 1974 ) rieke , g.h . , weekes , t.c .
apj * 155 * 429 ( 1969 ) sako , t. et al .
apj * 537 * 422 ( 2000 ) stepanian , a.a .
. obs . * 62 * 79 ( 1980 ) weekes , t.c .
apj * 342 * 379 ( 1989 ) yoshikoshi , t. et al .
apj * 487 * l65 ( 1997 ) | young pulsars produce relativistic winds which interact with matter ejected during the supernova explosion and the surrounding interstellar gas .
particles are accelerated to very high energies somewhere in the pulsar winds or at the shocks produced in collisions of the winds with the surrounding medium . as a result of interactions of relativistic leptons with the magnetic field and low energy radiation ( of synchrotron origin , thermal , or microwave background )
, the non - thermal radiation is produced with the lowest possible energies up to @xmath0100 tev .
the high energy ( tev ) @xmath1-ray emission has been originally observed from the crab nebula and recently from several other objects .
recent observations by the hess cherenkov telescopes allow to study for the first time morphology of the sources of high energy emission , showing unexpected spectral features .
they might be also interpreted as due to acceleration of hadrons .
however , theory of particle acceleration in the pwne and models for production of radiation are still at their early stage of development since it becomes clear that realistic modeling of these objects should include their time evolution and three - dimensional geometry . in this paper
we concentrate on the attempts to create a model for the high energy processes inside the pwne which includes existence not only relativistic leptons but also hadrons inside the nebula .
such model should also take into account evolution of the nebula in time .
possible high energy expectations based on such a model are discussed in the context of new observations .
example.eps gsave newpath 20 20 moveto 20 220 lineto 220 220 lineto 220 20 lineto closepath 2 setlinewidth gsave .4 setgray fill grestore stroke grestore |
scientists face a rapidly growing deluge of data , which requires highly sophisticated methodologies for analysis and compression . simultaneously , the complexity of the data is increasing , evidenced in particular by the observation that data becomes increasingly high - dimensional .
one of the most prominent features of data are singularities which is justified , for instance , by the observation from computer visionists that the human eye is most sensitive to smooth geometric areas divided by sharp edges .
intriguingly , already the step from univariate to multivariate data causes a significant change in the behavior of singularities . whereas one - dimensional ( 1d ) functions can only exhibit point singularities , singularities of two - dimensional ( 2d ) functions
can already be of both point as well as curvilinear type .
thus , in contrast to _ isotropic _ features point singularities , suddenly _ anisotropic _ features
curvilinear singularities are possible . and , in fact , multivariate functions are typically governed by _
anisotropic phenomena_. think , for instance , of edges in digital images or evolving shock fronts in solutions of transport - dominated equations .
these two exemplary situations also show that such phenomena occur even for both explicitly as well as implicitly given data .
one major goal both for the purpose of compression as well as for an efficient analysis is the introduction of representation systems for ` good ' approximation of anisotropic phenomena , more precisely , of multivariate functions governed by anisotropic features .
this raises the following fundamental questions : 1 . what is a suitable model for functions governed by anisotropic features ?
2 . how do we measure ` good ' approximation and what is a benchmark for optimality ?
3 . is the step from 1d to 2d already the crucial step or how does this framework scale with increasing dimension ?
which representation system behaves optimally ?
let us now first debate these questions on a higher and more intuitive level , and later on delve into the precise mathematical formalism .
each model design has to face the trade - off between closeness to the true situation versus sufficient simplicity to enable analysis of the model .
the suggestion of a suitable model for functions governed by anisotropic features in @xcite solved this problem in the following way .
as a model for an image , it first of all requires the @xmath2 functions serving as a model to be supported on the unit square @xmath3 ^ 2 $ ] .
these functions shall then consist of the minimal number of smooth parts , namely two . to avoid artificial problems with a discontinuity ending at the boundary of @xmath3 ^ 2 $ ] , the boundary curve of one of the smooth parts is entirely contained in @xmath4 .
it now remains to decide upon the regularity of the smooth parts of the model functions and of the boundary curve , which were chosen to both be @xmath5 .
thus , concluding , a possible suitable model for functions governed by anisotropic features are 2d functions which are supported on @xmath3 ^ 2 $ ] and @xmath5 apart from a closed @xmath5 discontinuity curve ; these are typically referred to as _ cartoon - like images _ ( cf .
chapter @xcite ) .
this provides an answer to ( p1 ) .
extensions of this 2d model to piecewise smooth curves were then suggested in @xcite , and extensions to 3d as well as to different types of regularity were introduced in @xcite .
the quality of the performance of a representation system with respect to cartoon - like images is typically measured by taking a non - linear approximation viewpoint .
more precisely , given a cartoon - like image and a representation system which forms an orthonormal basis , the chosen measure is the asymptotic behavior of the @xmath0 error of the best @xmath1-term ( non - linear ) approximation in the number of terms @xmath1 .
this intuitively measures how fast the @xmath6 norm of the tail of the expansion decays as more and more terms are used for the approximation .
a slight subtlety has to be observed if the representation system does not form an orthonormal basis , but a frame . in this case ,
the @xmath1-term approximation using the @xmath1 largest coefficients is considered which , in case of an orthonormal basis , is the same as the best @xmath1-term approximation , but not in general .
the term ` optimally sparse approximation ' is then awarded to those representation systems which deliver the fastest possible decay rate in @xmath1 for all cartoon - like images , where we consider log - factors as negligible , thereby providing an answer to ( p2 ) .
we already identified the step from 1d to 2d as crucial for the appearance of anisotropic features at all .
hence one might ask : is is sufficient to consider only the 2d situation , and higher dimensions can be treated similarly ? or : does each dimension causes its own problems ? to answer these questions ,
let us consider the step from 2d to 3d which shows a curious phenomenon .
a 3d function can exhibit point (= 0d ) , curvilinear (= 1d ) , and surface (= 2d ) singularities .
thus , suddenly anisotropic features appear in two different dimensions : as one - dimensional and as two - dimensional features .
hence , the 3d situation has to be analyzed with particular care .
it is not at all clear whether two different representation systems are required for optimally approximating both types of anisotropic features simultaneously , or whether one system will suffice .
this shows that the step from 2d to 3d can justifiably be also coined ` crucial ' .
once it is known how to handle anisotropic features of different dimensions , the step from 3d to 4d can be dealt with in a similar way as also the extension to even higher dimensions .
thus , answering ( p3 ) , we conclude that the two crucial dimensions are 2d and 3d with higher dimensional situations deriving from the analysis of those . within the framework we just briefly outlined
, it can be shown that wavelets do not provide optimally sparse approximations of cartoon - like images .
this initiated a flurry of activity within the applied harmonic analysis community with the aim to develop so - called _ directional _ representation systems which satisfy this benchmark , certainly besides other desirable properties depending in the application at hand . in 2004 ,
cands and donoho were the first to introduce with the tight curvelet frames a directional representation system which provides provably optimally sparse approximations of cartoon - like images in the sense we discussed .
one year later , contourlets were introduced by do and vetterli @xcite , which similarly derived an optimal approximation rate .
the first analysis of the performance of ( band - limited ) shearlet frames was undertaken by guo and labate in @xcite , who proved that these shearlets also do satisfy this benchmark . in the situation of (
band - limited ) shearlets the analysis was then driven even further , and very recently guo and labate proved a similar result for 3d cartoon - like images which in this case are defined as a function which is @xmath5 apart from a @xmath5 discontinuity surface , i.e. , focusing on only one of the types of anisotropic features we are facing in 3d .
the results mentioned in the previous subsection only concerned band - limited systems . even in the contourlet case , although compactly supported contourlets seem to be included , the proof for optimal sparsity only works for band - limited generators due to the requirement of infinite directional vanishing moments .
however , for various applications compactly supported generators are inevitable , wherefore already in the wavelet case the introduction of compactly supported wavelets was a major advance .
prominent examples of such applications are imaging sciences , when an image might need to be denoised while avoiding a smoothing of the edges , or in the theory of partial differential equations as a generating system for a trial space in order to ensure fast computational realizations .
so far , shearlets are the only system , for which a theory for compactly supported generators has been developed and compactly supported shearlet frames have been constructed @xcite , see also the survey paper @xcite .
it should though be mentioned that these frames are somehow close to being tight , but at this point it is not clear whether also compactly supported _ tight _ shearlet frames can be constructed .
interestingly , it was proved in @xcite that this class of shearlet frames also delivers optimally sparse approximations of the 2d cartoon - like image model class with a very different proof than @xcite now adapted to the particular nature of compactly supported generators .
and with @xcite the 3d situation is now also fully understood , even taking the two different types of anisotropic features curvilinear and surface singularities into account . in sect .
[ sec : cartoon - images ] , we introduce the 2d and 3d cartoon - like image model class .
optimality of sparse approximations of this class are then discussed in sect .
[ sec : linear - non - linear ] .
[ sec : pyram - adapt - shearl ] is concerned with the introduction of 3d shearlet systems with both band - limited and compactly supported generators , which are shown to provide optimally sparse approximations within this class in the final sect . [ sec : optim - sparse - appr ] .
we start by making the in the introduction of this chapter already intuitively derived definition of cartoon - like images mathematically precise .
we start with the most basic definition of this class which was also historically first stated in @xcite .
we allow ourselves to state this together with its 3d version from @xcite by remarking that @xmath7 could be either @xmath8 or @xmath9 .
for fixed @xmath10 , the _ class @xmath11 of cartoon - like image _ shall be the set of functions @xmath12 of the form @xmath13 where @xmath14}}^d$ ] and @xmath15 with @xmath16}}^d$ ] and @xmath17 for each @xmath18 . for dimension @xmath8
, we assume that @xmath19 is a closed @xmath5-curve with curvature bounded by @xmath20 , and , for @xmath9 , the discontinuity @xmath19 shall be a closed @xmath5-surface with principal curvatures bounded by @xmath20 .
an indiscriminately chosen cartoon - like function @xmath21 , where the discontinuity surface @xmath22 is a deformed sphere in @xmath23 , is depicted in fig .
[ fig : cartoon ] . with @xmath24 for dimension @xmath9 , where the discontinuity surface @xmath22 is a deformed sphere .
] since ` objects ' in images often have sharp corners , in @xcite for 2d and in @xcite for 3d also less regular images were allowed , where @xmath22 is only assumed to be _ piecewise _ @xmath5-smooth .
we note that this viewpoint is also essential for being able to analyze the behavior of a system with respect to the two different types of anisotropic features appearing in 3d ; see the discussion in subsection [ subsec:3d ] . letting @xmath25 denote the number of @xmath5 pieces , we speak of the extended class of _ cartoon - like images @xmath26 _ as consisting of cartoon - like images having @xmath5-smoothness apart from a piecewise @xmath5 discontinuity curve in the 2d setting and a piecewise @xmath5 discontinuity surface in the 3d setting .
indeed , in the 3d setting , besides the @xmath5 discontinuity surfaces , this model exhibits curvilinear @xmath5 singularities as well as point singularities , e.g.,the cartoon - like image @xmath21 in fig .
[ fig : cartoon - piecewise ] exhibits a discontinuity surface @xmath27 consisting of _ three _
@xmath5-smooth surfaces with point and curvilinear singularities where these surfaces meet . with @xmath28 , where the discontinuity surface @xmath22 is piecewise @xmath5 smooth . ]
the model in @xcite goes even one step further and considers a different regularity for the smooth parts , say being in @xmath29 , and for the smooth pieces of the discontinuity , say being in @xmath30 with @xmath31 .
this very general class of cartoon - like images is then denoted by @xmath32 , with the agreement that @xmath33 for @xmath34 . for the purpose of clarity , in the sequel we will focus on the first most basic cartoon - like model where @xmath34 , and add hints on generalizations when appropriate ( in particular , in sect .
[ sec : some - extensions ] ) .
after having clarified the model situation , we will now discuss which measure for the accuracy of approximation by representation systems we choose , and what optimality means in this case .
let @xmath35 denote a given class of elements in a separable hilbert space @xmath36 with norm @xmath37 and @xmath38 a dictionary for @xmath36 , i.e.,@xmath39 , with indexing set @xmath40 .
the dictionary @xmath41 plays the role of our representation system .
later @xmath35 will be chosen to be the class of cartoon - like images and @xmath41 a shearlet frame , but for now we will assume this more general setting .
we now seek to approximate each single element of @xmath35 with elements from @xmath41 by ` few ' terms of this system .
approximation theory provides us with the concept of best @xmath1-term approximation which we now introduce ; for a general introduction to approximation theory , we refer to @xcite . for this ,
let @xmath42 be arbitrarily chosen . since @xmath41 is a complete system , for any @xmath43 there exists a finite linear combination of elements from @xmath41 of the form @xmath44 such that @xmath45
. moreover , if @xmath41 is a frame with countable indexing set @xmath40 , there exists a sequence @xmath46 such that the representation @xmath47 holds with convergence in the hilbert space norm @xmath48 . the reader should notice that , if @xmath41 does not form a basis , this representation of @xmath49 is certainly not the only possible one .
letting now @xmath50 , we aim to approximate @xmath49 by only @xmath1 terms of @xmath41 , i.e. , by @xmath51 which is termed _ @xmath1-term approximation _ to @xmath49 .
this approximation is typically _ non - linear _ in the sense that if @xmath52 is an @xmath1-term approximation to @xmath49 with indices @xmath53 and @xmath54 is an @xmath1-term approximation to some @xmath55 with indices @xmath56 , then @xmath57 is only an @xmath1-term approximation to @xmath58 in case @xmath59 .
but certainly we would like to pick the ` best ' approximation with the accuracy of approximation measured in the hilbert space norm .
we define the _ best @xmath1-term approximation _ to @xmath49 by the @xmath1-term approximation @xmath60 which satisfies that , for all @xmath61 , @xmath62 , and for all scalars @xmath63 , @xmath64 let us next discuss the notion of best @xmath1-term approximation for the special cases of @xmath41 forming an orthornomal basis , a tight frame , and a general frame alongside an error estimate for the accuracy of this approximation .
let @xmath41 be an orthonormal basis for @xmath36 . in this case
, we can actually write down the best @xmath1-term approximation @xmath65 for @xmath49 . since in this case @xmath66 and this representation is unique , we obtain @xmath67 \phi_i + \sum_{i \in i \setminus { i}_n } { \langlef,\phi_i\rangle } \phi_i\big\|\\ & = \|({\langlef,\phi_i\rangle } -{c}_i)_{i \in { i}_n}\|_{\ell^2 } + \| ( { \langlef,\phi_i\rangle})_{i \in i \setminus { i}_n}\|_{\ell^2}.\end{aligned}\ ] ] the first term @xmath68 can be minimized by choosing @xmath69 for all @xmath70 . and the second term @xmath71 can be minimized by choosing @xmath72 to be the indices of the @xmath1 largest coefficients @xmath73 in magnitude . notice that this does not uniquely determine @xmath52 since some coefficients @xmath73 might have the same magnitude .
but it characterizes the set of best @xmath1-term approximations to some @xmath42 precisely .
even more , we have complete control of the error of best @xmath1-term approximation by @xmath74 assume now that @xmath41 constitutes a tight frame with bound @xmath75 for @xmath36 .
in this situation , we still have @xmath66 but this expansion is now not unique anymore . moreover , the frame elements are not orthogonal .
both conditions prohibit an analysis of the error of best @xmath1-term approximation as in the previously considered situation of an orthonormal basis . and
in fact , examples can be provided to show that selecting the @xmath1 largest coefficients @xmath73 in magnitude does not always lead to the _ best _ @xmath1-term approximation , but merely to _ an _ @xmath1-term approximation . to be able to still analyze the approximation error , one typically as will be also our choice in the sequel chooses the @xmath1-term approximation provided by the indices @xmath72 associated with the @xmath1 largest coefficients @xmath73 in magnitude with these coefficients , i.e. , @xmath76 this selection also allows for some control of the approximation in the hilbert space norm , which we will defer to the next subsection in which we consider the more general case of arbitrary frames .
let now @xmath41 form a frame for @xmath36 with frame bounds @xmath77 and @xmath78 , and let @xmath79 denote the canonical dual frame .
we then consider the expansion of @xmath49 in terms of this dual frame , i.e. , @xmath80 notice that we could also consider @xmath81 let us explain , why the first form is of more interest to us in this chapter . by definition
, we have @xmath82 as well as @xmath83 .
since we only consider expansions of functions @xmath49 belonging to a subset @xmath35 of @xmath36 , this can , at least , potentially improve the decay rate of the coefficients so that they belong to @xmath84 for some @xmath85 .
this is exactly what is understood by _
sparse approximation _ ( also called _ compressible approximations _ in the context of inverse problems ) .
we hence aim to analyze shearlets with respect to this behavior , i.e. , the decay rate of shearlet coefficients .
this then naturally leads to form .
we remark that in case of a tight frame , there is no distinction necessary , since then @xmath86 for all @xmath87 . as in the tight frame case , it is not possible to derive a usable , explicit form for the best @xmath1-term approximation .
we therefore again crudely approximate the best @xmath1-term approximation by choosing the @xmath1-term approximation provided by the indices @xmath72 associated with the @xmath1 largest coefficients @xmath73 in magnitude with these coefficients , i.e. , @xmath88 but , surprisingly , even with this rather crude greedy selection procedure , we obtain very strong results for the approximation rate of shearlets as we will see in sect . [
sec : optim - sparse - appr ] .
the following result shows how the @xmath1-term approximation error can be bounded by the tail of the square of the coefficients @xmath89 .
the reader might want to compare this result with the error in case of an orthonormal basis stated in .
[ lemma : n - term - frame - approx ] let @xmath90 be a frame for @xmath36 with frame bounds @xmath77 and @xmath78 , and let @xmath79 be the canonical dual frame .
let @xmath61 with @xmath62 , and let @xmath52 be the @xmath1-term approximation @xmath91 .
then @xmath92 recall that the canonical dual frame satisfies the frame inequality with bounds @xmath93 and @xmath94 . at first hand
, it therefore might look as if the estimate ( [ eq : n - term - frame - approx - bound ] ) should follow directly from the frame inequality for the canonical dual .
however , since the sum in ( [ eq : n - term - frame - approx - bound ] ) does not run over the entire index set @xmath95 , but only @xmath96 , this is not the case .
so , to prove the lemma , we first consider @xmath97 using cauchy - schwarz inequality , we then have that @xmath98 where we have used the upper frame inequality for the dual frame @xmath99 in the second step
. we can now continue ( [ eq : reisz - estimate ] ) and arrive at @xmath100 relating to the previous discussion about the decay of coefficients @xmath73 , let @xmath101 denote the non - increasing ( in modulus ) rearrangement of @xmath102 , e.g.,@xmath103 denotes the @xmath104th largest coefficient of @xmath105 in modulus .
this rearrangement corresponds to a bijection @xmath106 that satisfies @xmath107 strictly speaking , the rearrangement ( and hence the mapping @xmath108 ) might not be unique ; we will simply take @xmath101 to be one of these rearrangements .
since @xmath109 , also @xmath110 .
suppose further that @xmath111 even decays as @xmath112 for some @xmath113 , where the notation @xmath114 means that there exists a @xmath115 such that @xmath116 , i.e.,@xmath117 .
clearly , we then have @xmath118 for @xmath119 . by lemma [ lemma :
n - term - frame - approx ] , the @xmath1-term approximation error will therefore decay as @xmath120 where @xmath52 is the @xmath1-term approximation of @xmath49 by keeping the @xmath1 largest coefficients , that is , @xmath121 the notation @xmath122 , also written @xmath123 , used above means that @xmath124 is bounded both above and below by @xmath125 asymptotically as @xmath126 , that is , @xmath127 _ and _ @xmath128 .
we now return to the setting of functions spaces @xmath129 , where the subset @xmath35 will be the class of cartoon - like images , that is , @xmath130 .
we then aim for a benchmark , i.e. , an optimality statement , for sparse approximation of functions in @xmath131 . for this
, we will again only require that our representation system @xmath41 is a dictionary , that is , we assume only that @xmath132 is a complete family of functions in @xmath133 with @xmath40 not necessarily being countable . without loss of generality , we can assume that the elements @xmath134 are normalized , i.e.,@xmath135 for all @xmath95 . for @xmath136
we then consider expansions of the form @xmath137 where @xmath138 is a countable selection from @xmath40 that may depend on @xmath49 .
relating to the previous subsection , the first @xmath1 elements of @xmath139 could for instance be the @xmath1 terms from @xmath41 selected for the best @xmath1-term approximation of @xmath49 . since artificial cases shall be avoided , this selection procedure has the following natural restriction which is usually termed _
polynomial depth search _ : the @xmath104th term in @xmath140 is obtained by only searching through the first @xmath141 elements of the list @xmath140 , where @xmath142 is a polynomial . moreover ,
the selection rule may _ adaptively _ depend on @xmath49 , and the @xmath104th element may also be modified adaptively and depend on the first @xmath143th chosen elements .
we shall denote any sequence of coefficients @xmath89 chosen according to these restrictions by @xmath144 .
the role of the polynomial @xmath142 is to limit how deep or how far down in the listed dictionary @xmath140 we are allowed to search for the next element @xmath134 in the approximation .
without such a depth search limit , one could choose @xmath41 to be a countable , dense subset of @xmath133 which would yield arbitrarily good sparse approximations , but also infeasible approximations in practise .
using information theoretic arguments , it was then shown in @xcite , that almost no matter what selection procedure we use to find the coefficients @xmath145 , we can not have @xmath146 bounded for @xmath147 for @xmath148 .
[ theo : benchmark ] retaining the definitions and notations in this subsection and allowing only polynomial depth search , we obtain @xmath149 in case @xmath41 is an orthonormal basis for @xmath150 , the norm @xmath146 is trivially bounded for @xmath151 since we can take @xmath152 . although not explicitly stated , the proof can be straightforwardly extended from 3d to higher dimensions as also the definition of cartoon - like images can be similarly extended .
it is then intriguing to analyze the behavior of @xmath153 from thm .
[ theo : benchmark ] . in fact , as @xmath154 , we observe that @xmath155 .
thus , the decay of any @xmath156 for cartoon - like images becomes slower as @xmath7 grows and approaches @xmath6 , which as we just mentioned is actually the rate guaranteed for _ all _ @xmath157 .
[ theo : benchmark ] is truly a statement about the optimal achievable sparsity level : no representation system up to the restrictions described above can deliver approximations for @xmath158 with coefficients satisfying @xmath159 for @xmath160 .
this implies the following lower bound @xmath161 where @xmath162 is a decreasing ( in modulus ) arrangement of the coefficients @xmath145 . one might ask how this relates to the approximation error of ( best ) @xmath1-term approximation discussed before . for simplicity ,
suppose for a moment that @xmath41 is actually an orthonormal basis ( or more generally a riesz basis ) for @xmath133 with @xmath8 and @xmath9 .
then as discussed in sect .
[ subsubsec : onb ] the _ best _
@xmath1-term approximation to @xmath163 is obtained by keeping the @xmath1 largest coefficients . using the error estimate as well as ( [ eq : coeff - fastest - decay ] )
, we obtain @xmath164 i.e. , the best @xmath1-term approximation error @xmath165 behaves asymptotically as @xmath166 or worse . if , more generally , @xmath41 is a frame , and @xmath52 is chosen as in ( [ eq : frame - n - term - largest ] ) , we can similarly conclude that the asymptotic lower bound for @xmath165 is @xmath166 , that is , the optimally achievable rate is , at best , @xmath166 .
thus , this optimal rate can be used as a benchmark for measuring the sparse approximation ability of cartoon - like images of different representation systems .
let us phrase this formally .
[ def : optimal ] let @xmath132 be a frame for @xmath133 with @xmath8 or @xmath9 .
we say that _ @xmath41 provides optimally sparse approximations _ of cartoon - like images if , for each @xmath163 , the associated @xmath1-term approximation @xmath52 ( cf . ) by keeping the @xmath1 largest coefficients of @xmath167 satisfies @xmath168 where we ignore @xmath169-factors .
note that , for frames @xmath41 , the bound @xmath170 automatically implies that @xmath171 whenever @xmath52 is chosen as in eqn . .
this follows from lemma [ lemma : n - term - frame - approx ] and the estimate @xmath172 where we have used that @xmath173 .
hence , we are searching for a representation system @xmath41 which forms a frame and delivers decay of @xmath174 as ( up to log - factors ) @xmath175 as @xmath176 for any cartoon - like image .
we will next study two examples of more traditional representation systems the fourier basis and wavelets with respect to their ability to meet this benchmark .
for this , we choose the function @xmath21 , where @xmath78 is a ball contained in @xmath177}}^d$ ] , again @xmath8 or @xmath9 , as a simple cartoon - like image in @xmath26 with @xmath24 , analyze the error @xmath178 for @xmath52 being the @xmath1-term approximation by the @xmath1 largest coefficients and compare with the optimal decay rate stated in definition [ def : optimal ] .
it will however turn out that these systems are far from providing optimally sparse approximations of cartoon - like images , thus underlining the pressing need to introduce representation systems delivering this optimal rate ; and we already now refer to sect . [ sec : optim - sparse - appr ] in which shearlets will be proven to satisfy this property . since fourier series and wavelet systems are orthonormal bases ( or more generally , riesz bases ) the best @xmath1-term approximation is found by keeping the @xmath1 largest coefficients as discussed in sect .
[ subsubsec : onb ] .
the error of the best @xmath1-term fourier series approximation of a typical cartoon - like image decays asymptotically as @xmath179 .
the following proposition shows this behavior in the case of a very simple cartoon - like let @xmath180 , and let @xmath181 .
suppose @xmath182 , where @xmath78 is a ball contained in @xmath177}}^d$ ] .
then @xmath183 where @xmath52 is the best @xmath1-term approximation from @xmath41 .
we fix a new origin as the center of the ball @xmath78 .
then @xmath49 is a radial function @xmath184 for @xmath185 .
the fourier transform of @xmath49 is also a radial function and can expressed explicitly by bessel functions of first kind @xcite : @xmath186 where @xmath187 is the radius of the ball @xmath78 . since the bessel function @xmath188 decays like @xmath189 as @xmath190 , the fourier transform of @xmath49 decays like @xmath191 as @xmath192 .
letting @xmath193 and @xmath194 be the partial fourier sum with terms from @xmath53 , we obtain @xmath195 the conclusion now follows from the cardinality of @xmath196 as @xmath197 .
since wavelets are designed to deliver sparse representations of singularities see chapter @xcite we expect this system to outperform the fourier approach
. this will indeed be the case .
however , the optimal rate will still by far be missed .
the best @xmath1-term approximation of a typical cartoon - like image using a wavelet basis performs only slightly better than fourier series with asymptotic behavior as @xmath198 .
this is illustrated by the following result .
[ prop : wavelets - ball - decay ] let @xmath199 , and let @xmath41 be a wavelet basis for @xmath133 or @xmath200}}^d)$ ] .
suppose @xmath182 , where @xmath78 is a ball contained in @xmath177}}^d$ ]
. then @xmath201 where @xmath52 is the best @xmath1-term approximation from @xmath41 .
let us first consider wavelet approximation by the haar tensor wavelet basis for @xmath200}}^d)$ ] of the form @xmath202 where @xmath203 , @xmath204 , and @xmath205 for @xmath206 .
there are only a finite number of coefficients of the form @xmath207 , hence we do not need to consider these for our asymptotic estimate . for simplicity , we take @xmath208 . at scale
@xmath209 there exist @xmath210 non - zero wavelet coefficients , since the surface area of @xmath22 is finite and the wavelet elements are of size @xmath211 . to illustrate the calculations leading to the sought approximation error rate , we will first consider the case where @xmath78 is a cube in @xmath177}}^d$ ] . for this , we first consider the non - zero coefficients associated with the face of the cube containing the point @xmath212 . for scale @xmath213 ,
let @xmath214 be such that @xmath215 , where @xmath216 and @xmath124 and @xmath217 are the haar wavelet and scaling function , respectively .
assume that @xmath218 is located in the first half of the interval @xmath219}}$ ] ; the other case can be handled similarly .
then @xmath220 where we have used that @xmath221 will typically be of size @xmath222 .
note that for the chosen @xmath213 and @xmath214 above , we also have that @xmath223 for all @xmath224
. there will be @xmath225 nonzero coefficients of size @xmath226 associated with the wavelet @xmath227 at scale @xmath213 .
the same conclusion holds for the other wavelets @xmath228 , @xmath229 . to summarize , at scale @xmath213 there will be @xmath230 nonzero coefficients of size @xmath231 . on the first @xmath232 scales , that is @xmath233 , we therefore have @xmath234 nonzero coefficients .
the @xmath104th largest coefficient @xmath103 is of size @xmath235 since , for @xmath236 , we have @xmath237 therefore , @xmath238 hence , for the best @xmath1-term approximation @xmath52 of @xmath49 using a wavelet basis , we obtain the asymptotic estimates @xmath239 let us now consider the situation that @xmath78 is a ball .
in fact , in this case we can do similar ( but less transparent ) calculations leading to the same asymptotic estimates as above .
we will not repeat these calculations here , but simply remark that the upper asymptotic bound in @xmath240 can be seen by the following general argument : @xmath241 which holds for each @xmath242 . finally , we can conclude from our calculations that choosing another wavelet basis will not improve the approximation rate .
we end this subsection with a remark on _ linear _ approximations . for a linear wavelet approximation of @xmath49 one would use @xmath243 for some @xmath244 .
if restricting to linear approximations , the summation order is not allowed to be changed , and we therefore need to include all coefficients from the first @xmath232 scales . at scale @xmath209 ,
there exist a total of @xmath245 coefficients , which by our previous considerations can be bounded by @xmath246 .
hence , we include @xmath247 times as many coefficients as in the non - linear approximation on each scale .
this implies that the error rate of the linear @xmath1-term wavelet approximation is @xmath179 , which is the _ same _ rate as obtained by fourier approximations .
the key problem of the suboptimal behavior of fourier series and wavelet bases is the fact that these systems are not generated by anisotropic elements .
let us illustrate this for 2d in the case of wavelets .
wavelet elements are isotropic due to the scaling matrix @xmath248 .
however , already intuitively , approximating a curve with isotropic elements requires many more elements than if the analyzing elements would be anisotropic themselves , see fig .
[ fig : wavelet - curve ] and [ fig : shearlet - curve ] . considering wavelets with anisotropic scaling
will not remedy the situation , since within one fixed scale one can not control the direction of the ( now anisotropically shaped ) elements .
thus , to capture a discontinuity curve as in fig .
[ fig : shearlet - curve ] , one needs not only anisotropic elements , but also a location parameter to locate the elements on the curve and a rotation parameter to align the elongated elements in the direction of the curve .
let us finally remark why a parabolic scaling matrix @xmath249 will be natural to use as anisotropic scaling .
since the discontinuity curves of cartoon - like images are @xmath5-smooth with bounded curvature , we may write the curve locally by a taylor expansion .
let s assume it has the form @xmath250 with @xmath251 near @xmath252 for some @xmath253}}$ ] .
clearly , the translation parameter will be used to position the anisotropic element near @xmath254 , and the orientation parameter to align with @xmath255 .
if the length of the element is @xmath256 , then , due to the term @xmath257 , the most beneficial height would be @xmath258 . and , in fact , parabolic scaling yields precisely this relation , i.e. , @xmath259 hence , the main idea in the following will be to design a system which consists of anisotropically shaped elements together with a directional parameter to achieve the optimal approximation rate for cartoon - like images .
after we have set our benchmark for directional representation systems in the sense of stating an optimality criteria for sparse approximations of the cartoon - like image class @xmath26 , we next introduce classes of shearlet systems we claim behave optimally .
as already mentioned in the introduction of this chapter , optimally sparse approximations were proven for a class of band - limited as well as of compactly supported shearlet frames . for the definition of cone - adapted discrete shearlets and , in particular , classes of band - limited as well as of compactly supported shearlet frames leading to optimally sparse approximations , we refer to chapter @xcite . in this section
, we present the definition of discrete shearlets in 3d , from which the mentioned definitions in the 2d situation can also be directly concluded . as special cases ,
we then introduce particular classes of band - limited as well as of compactly supported shearlet frames , which will be shown to provide optimally approximations of @xmath260 and , with a slight modification which we will elaborate on in sect .
[ sec : some - extensions ] , also for @xmath261 with @xmath262 .
the first step in the definition of cone - adapted discrete 2d shearlets was a partitioning of 2d frequency domain into two pairs of high - frequency cones and one low - frequency rectangle .
we mimic this step by partitioning 3d frequency domain into the three pairs of _ pyramids _ given by @xmath263 and the centered cube @xmath264 this partition is illustrated in fig .
[ fig : pyramids ] which depicts the three pairs of pyramids and fig .
[ fig : partition ] depicting the centered cube surrounded by the three pairs of pyramids @xmath265 , @xmath266 , and @xmath267 . .
the arrangement of the six pyramids is indicated by the ` diagonal ' lines .
see fig .
[ fig : pyramids ] for a sketch of the pyramids . ]
the partitioning of frequency space into pyramids allows us to restrict the range of the shear parameters . without such a partitioning as , e.g.,in shearlet systems arising from the shearlet group
, one must allow arbitrarily large shear parameters , which leads to a treatment biased towards one axis .
the defined partition however enables restriction of the shear parameters to @xmath268 $ ] , similar to the definition of cone - adapted discrete shearlet systems .
we would like to emphasize that this approach is key to provide an almost uniform treatment of different directions in a sense of a ` good ' approximation to rotation .
pyramid - adapted discrete shearlets are scaled according to the _ paraboloidal scaling matrices _
@xmath269 , @xmath270 or @xmath271 , @xmath272 defined by @xmath273 and directionality is encoded by the _ shear matrices _
@xmath274 , @xmath275 , or @xmath276 , @xmath277 , given by @xmath278 respectively .
the reader should note that these definitions are ( discrete ) special cases of the general setup in @xcite .
the translation lattices will be defined through the following matrices : @xmath279 , @xmath280 , and @xmath281 , where @xmath282 and @xmath283 .
we are now ready to introduce 3d shearlet systems , for which we will make use of the vector notation @xmath284 for @xmath285 and @xmath286 to denote @xmath287 _ and _ @xmath288 .
[ def : discreteshearlets3d ] for @xmath289 , the _ pyramid - adapted discrete shearlet system _ @xmath290 generated by @xmath291 is defined by @xmath292 where @xmath293 where @xmath294 and @xmath295 . for the sake of brevity
, we will sometimes also use the notation @xmath296 with @xmath297 .
we now focus on two different special classes of pyramid - adapted discrete shearlets leading to the class of band - limited shearlets and the class of compactly supported shearlets for which optimality of their approximation properties with respect to cartoon - like images will be proven in sect .
[ sec : optim - sparse - appr ] .
let the shearlet generator @xmath298 be defined by @xmath299 where @xmath300 and @xmath301 satisfy the following assumptions : a. @xmath302 , @xmath303 , and @xmath304 b. @xmath305 , @xmath306}}$ ] , and @xmath307 thus , in frequency domain , the band - limited function @xmath298 is almost a tensor product of one wavelet with two ` bump ' functions , thereby a canonical generalization of the classical band - limited 2d shearlets , see also chapter @xcite . this implies the support in frequency domain to have a needle - like shape with the wavelet acting in radial direction ensuring high directional selectivity , see also fig .
[ fig : tiling-3d ] . in the frequency domain .
the two shearlet elements have the same scale parameter @xmath308 , but different shearing parameters @xmath309 . ]
the derivation from being a tensor product , i.e. , the substitution of @xmath310 and @xmath311 by the quotients @xmath312 and @xmath313 , respectively , in fact ensures a favorable behavior with respect to the shearing operator , and thus a tiling of frequency domain which leads to a tight frame for @xmath314 . a first step towards this result is the following observation .
[ thm : construction - bandlimited-3d ] let @xmath315 be a band - limited shearlet defined as in this subsection . then the family of functions @xmath316 forms a tight frame for @xmath317 . for each @xmath318 ,
equation implies that @xmath319 hence , using equation , we obtain @xmath320 for @xmath321 . using this equation together with the fact that @xmath322 is supported inside @xmath323}}^3 $ ] proves the theorem . by thm .
[ thm : construction - bandlimited-3d ] and a change of variables , we can construct shearlet frames for @xmath324 , @xmath325 , and @xmath326 , respectively .
furthermore , wavelet theory provides us with many choices of @xmath327 such that @xmath328 forms a frame for @xmath329 . since @xmath330 as a disjoint union , we can express any function @xmath331 as @xmath332 , where each component corresponds to the orthogonal projection of @xmath49 onto one of the three pairs of pyramids or the centered cube in the frequency space .
we then expand each of these components in terms of the corresponding tight frame .
finally , our representation of @xmath49 will then be the sum of these four expansions .
we remark that the projection of @xmath49 onto the four subspaces can lead to artificially slow decaying shearlet coefficients ; this will , e.g.,be the case if @xmath49 is in the schwartz class .
this problem does in fact not occur in the construction of compactly supported shearlets .
it is easy to see that the general form does never lead to a function which is compactly supported in spatial domain .
thus , we need to deviate this form by now taking indeed exact tensor products as our shearlet generators , which has the additional benefit of leading to fast algorithmic realizations .
this however causes the problem that the shearlets do not behave as favorable with respect to the shearing operator as in the previous subsection , and the question arises whether they actually do lead to at least a frame for @xmath314 .
the next results shows this to be true for an even much more general form of shearlet generators including compactly supported separable generators .
the attentive reader will notice that this theorem even covers the class of band - limited shearlets introduced in sect .
[ subsec : bandlimited ] .
[ thm : construction_suff ] let @xmath333 be functions such that @xmath334 and @xmath335 and @xmath336 .
define @xmath337 and @xmath338 for @xmath339 .
then there exists a constant @xmath340 such that the shearlet system @xmath341 forms a frame for @xmath342 for all @xmath343 with @xmath344 provided that there exists a positive constant @xmath345 such that @xmath346 for a.e @xmath347 , where @xmath348}}$ ] .
we next provide an example of a family of compactly supported shearlets satisfying the assumptions of thm .
[ thm : construction_suff ] .
however , for applications , one is typically not only interested in whether a system forms a frame , but in the ratio of the associated frame bounds . in this regard
, these shearlets also admit a theoretically derived estimate for this ratio which is reasonably close to @xmath349 , i.e. , to being tight .
the numerically derived ratio is even significantly closer as expected .
[ example : compact - for - pyramid-3d ] let @xmath350 be such that @xmath351 and @xmath352 , and define a shearlet @xmath298 by @xmath353 where the function @xmath354 is the low pass filter satisfying @xmath355 for @xmath356 the function @xmath357 is the associated bandpass filter defined by @xmath358 and @xmath359 the scaling function is given by @xmath360 in @xcite it is shown that @xmath359 and @xmath315 indeed are compactly supported .
moreover , we have the following result .
suppose @xmath298 is defined as in ( [ eq : def - psi - compact ] ) .
then there exists a sampling constant @xmath340 such that the shearlet system @xmath361 forms a frame for @xmath324 for any translation matrix @xmath362 with @xmath289 and @xmath363 . using upper and lower estimates of the absolute value of the trigonometric polynomial @xmath354 ( cf .
@xcite ) , one can show that @xmath315 satisfies the hypothesis of thm .
[ thm : construction_suff ] as well as @xmath364 where @xmath345 is a constant , for some sufficiently small @xmath340 .
we note that this inequality is an analog to ( [ eq : lower - linf - bound ] ) for the pyramid @xmath365 .
hence , by a result similar to thm .
[ thm : construction_suff ] , but for the case , where we restrict to the pyramid @xmath324 , it then follows that @xmath361 is a frame . to obtain a frame for all of @xmath314 we simply set @xmath337 and @xmath338 as in thm .
[ thm : construction_suff ] , and choose @xmath366 as scaling function for @xmath339 . then the corresponding shearlet system @xmath367 forms a frame for @xmath314 .
the proof basically follows from daubechies classical estimates for wavelet frames in @xcite and the fact that anisotropic and sheared windows obtained by applying the scaling matrix @xmath269 and the shear matrix @xmath368 to the effective support has _ effective _ support on @xmath78 if the ratio @xmath369 is `` close '' to @xmath349 . ] of @xmath370 cover the pyramid @xmath365 in the frequency domain .
the same arguments can be applied to each of shearlet generators @xmath315 , @xmath371 and @xmath372 as well as the scaling function @xmath359 to show a covering of the entire frequency domain and thereby the frame property of the pyramid - adapted shearlet system for @xmath314 .
we refer to @xcite for the detailed proof .
theoretical and numerical estimates of frame bounds for a particular parameter choice are shown in table [ tab : frame - bounds-3d ] .
we see that the theoretical estimates are overly pessimistic , since they are a factor @xmath373 larger than the numerical estimated frame bound ratios .
we mention that for 2d the estimated frame bound ratios are approximately @xmath374 of the ratios found in table [ tab : frame - bounds-3d ] .
p3.5cmp3cmp5.cm theoretical ( @xmath375 ) & numerical ( @xmath375 ) & translation constants ( @xmath376 ) + 345.7 & 13.42 & ( 0.9 , 0.25 ) + 226.6 & 13.17 & ( 0.9 , 0.20 ) + 226.4 & 13.16 & ( 0.9 , 0.15 ) + 226.4 & 13.16 & ( 0.9 , 0.10 ) + the compactly supported shearlets @xmath377 from example [ example : compact - for - pyramid-3d ] are , in spatial domain , of size @xmath378 times @xmath378 times @xmath379 due to the scaling matrix @xmath269 .
this reveals that the shearlet elements will become ` plate - like ' as @xmath380 . for an illustration
, we refer to fig .
[ fig : support - size-3d - shearlet ] .
band - limited shearlets , on the other hand , do not have compactly support , but their effective support ( the region where the energy of the function is concentrated ) in spatial domain will likewise be of size @xmath378 times @xmath378 times @xmath379 owing to their smoothness in frequency domain . from example [
example : compact - for - pyramid-3d].,width=264 ] contemplating about the fact that intuitively such shearlet elements should provide sparse approximations of surface singularities , one could also think of using the scaling matrix @xmath381 with similar changes for @xmath382 and @xmath383 to derive ` needle - like ' shearlet elements in space domain .
these would intuitively behave favorable with respect to the other type of anisotropic features occurring in 3d , that is curvilinear singularities .
surprisingly , we will show in sect .
[ sec : extensions-3d ] that for optimally sparse approximation plate - like shearlets , i.e. , shearlets associated with scaling matrix @xmath384 , and similarly @xmath385 and @xmath383 are sufficient .
let us also mention that , more generally , non - paraboloidal scaling matrices of the form @xmath386 for @xmath387 can be considered .
the parameters @xmath388 and @xmath389 allow precise control of the aspect ratio of the shearlet elements , ranging from very plate - like to very needle - like , according to the application at hand , i.e.,choosing the shearlet - shape that is the best matches the geometric characteristics of the considered data .
the case @xmath390 is covered by the setup of the multidimensional shearlet transform explained in chapter @xcite .
let us finish this section with a general thought on the construction of band - limited ( not separable ) tight shearlet frames versus compactly supported ( non - tight , but separable ) shearlet frames .
it seems that there is a trade - off between _ compact support _ of the shearlet generators , _ tightness _ of the associated frame , and _
separability _ of the shearlet generators .
in fact , even in 2d , all known constructions of tight shearlet frames do not use separable generators , and these constructions can be shown to _ not _ be applicable to compactly supported generators . presumably , tightness is difficult to obtain while allowing for compactly supported generators , but we can gain separability which leads to fast algorithmic realizations , see chapter @xcite . if we though allow non - compactly supported generators , tightness is possible as shown in sect .
[ subsec : bandlimited ] , but separability seems to be out of reach , which causes problems for fast algorithmic realizations .
in this section , we will show that shearlets both band - limited as well as compactly supported as defined in sect .
[ sec : pyram - adapt - shearl ] indeed provide the optimal sparse approximation rate for cartoon - like images from sect . [ sec : optimal - sparsity ] .
thus , letting @xmath391 denote the band - limited shearlet frame from sect .
[ subsec : bandlimited ] and the compactly supported shearlet frame from sect . [ subsec : compactsupport ] in both 2d and 3d ( see @xcite ) and @xmath392 , we aim to prove that @xmath393 where as debated in sect .
[ sec : non - line - appr ]
@xmath52 denotes the @xmath1-term approximation using the @xmath1 largest coefficients as in ( [ eq : frame - n - term - largest ] ) .
hence , in 2d we aim for the rate @xmath394 and in 3d we aim for the rate @xmath395 with ignoring @xmath169-factors . as mentioned in sect .
[ sec : optimal - sparsity ] , see ( [ eq : sought - sparsity ] ) , in order to prove these rate , it suffices to show that the @xmath104th largest shearlet coefficient @xmath396 decays as @xmath397 according to dfn . [ def : optimal ] this will show that among all adaptive and non - adaptive representation systems shearlet frames behave optimal with respect to sparse approximation of cartoon - like images . that one is able to obtain such an optimal approximation error rate
might seem surprising , since the shearlet system as well as the approximation procedure will be non - adaptive . to present the necessary hypotheses , illustrate the key ideas of the proofs , and debate the differences between the arguments for band - limited and compactly supported shearlets , we first focus on the situation of 2d shearlets .
we then discuss the 3d situation , with a sparsified proof , mainly discussing the essential differences to the proof for 2d shearlets and highlighting the crucial nature of this case ( cf .
[ subsec:3d ] ) .
as discussed in the previous section , in the case @xmath8 , we aim for the estimates @xmath398 and @xmath399 ( up to log - factors ) . in sect .
[ sec : heuristic - analysis ] we will first provide a heuristic analysis to argue that shearlet frames indeed can deliver these rates . in sect .
[ sec : required - hypotheses ] and [ sec : main - result ] we then discuss the required hypotheses and state the main optimality result .
the subsequent subsections are then devoted to proving the main result .
we start by giving a heuristic argument ( inspired by a similar argument for curvelets in @xcite ) on why the error @xmath165 satisfies the asymptotic rate @xmath394 .
we emphasize that this heuristic argument applies to both the band - limited and also the compactly supported case . for simplicity
we assume @xmath24 , and let @xmath400 be a 2d cartoon - like image .
the main concern is to derive the estimate ( [ eq:2d - sparsity - rate ] ) for the shearlet coefficients @xmath401 , where @xmath402 denotes either @xmath315 or @xmath403 .
we consider only the case @xmath404 , since the other case can be handled similarly . for compactly supported shearlet
, we can think of our generators having the form @xmath405 , @xmath406 , where @xmath407 is a wavelet and @xmath359 a bump ( or a scaling ) function . it will become important , that the wavelet ` points ' in the @xmath408-axis direction , which corresponds to the ` short ' direction of the shearlet . for band - limited generators
, we can think of our generators having the form @xmath409 for @xmath410 .
we , moreover , restrict our analysis to shearlets @xmath377 since the frame elements @xmath411 can be handled in a similar way .
a. shearlets @xmath377 whose support does not overlap with the boundary @xmath19 .
b. shearlets @xmath377 whose support overlaps with @xmath19 and is nearly tangent . c. shearlets @xmath377 whose support overlaps with @xmath19 , but not tangentially
. does not overlap with @xmath19 , ( b ) the support of @xmath377 does overlap with @xmath19 and is nearly tangent , ( c ) the support of @xmath377 does overlap with @xmath19 , but not tangentially .
note that only a section of the discontinuity curve @xmath19 is shown , and that for the case of band - limited shearlets only the effective support is shown . ]
it turns out that only coefficients from case ( b ) will be significant .
case ( b ) is , loosely speaking , the situation , where the wavelet @xmath407 crosses the discontinuity curve over the entire ` height ' of the shearlet , see fig .
[ fig : case - a - b - c ] . _ case ( a ) .
_ since @xmath49 is @xmath5-smooth away from @xmath19 , the coefficients @xmath413 will be sufficiently small owing to the approximation property of the wavelet @xmath407 .
the situation is sketched in fig . [
fig : case - a - b - c ] .
_ case ( b ) .
_ at scale @xmath414 , there are about @xmath415 coefficients , since the shearlet elements are of length @xmath378 ( and ` thickness ' @xmath379 ) and the length of @xmath22 is finite . by hlder s inequality , we immediately obtain @xmath416 for some constants @xmath417 . in other words
, we have @xmath415 coefficients bounded by @xmath418 . assuming the case ( a ) and ( c ) coefficients are negligible , the @xmath104th largest coefficient @xmath103
is then bounded by @xmath419 which was what we aimed to show ; compare to ( [ eq : coeff - fastest - decay - upper ] ) in dfn .
[ def : optimal ] .
this in turn implies ( cf .
estimate ( [ eq : sparsity - implies - approx ] ) ) that @xmath420 by lemma [ lemma : n - term - frame - approx ] , as desired it follows that @xmath421 where @xmath77 denotes the lower frame bound of the shearlet frame .
_ case ( c ) .
_ finally , when the shearlets are sheared away from the tangent position in case ( b ) , they will again be small .
this is due to the frequency support of @xmath49 and @xmath422 as well as to the directional vanishing moment conditions assumed in setup [ setup : assumptions - on - generators - bl-2d ] or [ setup : assumptions - on - generators - cs-2d ] , which will be formally introduced in the next subsection . after having build up some intuition on why the optimal sparse approximation rate is achievable using shearlets
, we will now go into more details and discuss the hypotheses required for the main result .
this will along the way already highlight some differences between the band - limited and compactly supported case . for this discussion
, assume that @xmath424 is piecewise @xmath425-smooth with a discontinuity on the line @xmath426 , @xmath427 , so that the function @xmath49 is well approximated by two 2d polynomials of degree @xmath428 , one polynomial on either side of @xmath429 , and denote this piecewise polynomial @xmath430 .
we denote the restriction of @xmath142 to lines @xmath431 , @xmath432 , by @xmath433 .
hence , @xmath434 is a 1d polynomial along lines parallel to @xmath429 going through @xmath435 ; these lines are marked by dashed lines in fig .
[ fig : directional - van - mom - spatial ] .
we now aim at estimating the absolute value of a shearlet coefficient @xmath412 by @xmath436 we first observe that @xmath413 will be small depending on the approximation quality of the ( piecewise ) polynomial @xmath142 and the decay of @xmath315 in the spatial domain . hence it suffices to focus on estimating @xmath437 . for this
, let us consider the line integration along the direction @xmath438 as follows : for @xmath439 fixed , define integration of @xmath440 along the lines @xmath431 , @xmath441 , as @xmath442 observe that @xmath443 is equivalent to @xmath444 . for simplicity ,
let us now assume @xmath445 .
then @xmath446 and , by the fourier slice theorem @xcite ( see also ) , it follows that @xmath447 note that @xmath448 if and only if @xmath449 therefore , to ensure @xmath450 for any 1d polynomial @xmath434 of degree @xmath428 , we require the following condition : @xmath451 these are the so - called _ directional vanishing moments _ ( cf .
@xcite ) in the direction @xmath452 .
we now consider the two cases , band - limited shearlets and compactly supported shearlets , separately .
if @xmath315 is a band - limited shearlet generator , we automatically have @xmath453 since @xmath454 , where @xmath455 is determined by the line @xmath456 .
hence , equation implies that , if the direction of @xmath457 , i.e. , of @xmath458 is _ not _ close to the direction of @xmath429 in the sense that @xmath459 , then @xmath460 however , if @xmath315 is a compactly supported shearlet generator , equation can never hold , since it requires that @xmath454 .
therefore , for compactly supported generators , we will assume that @xmath461 , @xmath462 , has sufficient decay in @xmath463 to force @xmath423 and hence @xmath464 to be sufficiently small .
it should be emphasized that the drawback that @xmath423 will only be ` small ' for compactly supported shearlets ( due to the lack of exact directional vanishing moments ) will be compensated by the perfect localization property which still enables optimal sparsity .
[ setup : assumptions - on - generators - bl-2d ] the generators @xmath465 are band - limited and @xmath466 in the frequency domain .
furthermore , the shearlet system @xmath467 forms a frame for @xmath2 ( cf . the construction in chapter @xcite or sect . [ subsec : bandlimited ] )
. [ setup : assumptions - on - generators - cs-2d ] the generators @xmath468 are compactly supported , and the shearlet system @xmath467 forms a frame for @xmath2 . furthermore , for all @xmath469 , the function @xmath315 satisfies where @xmath472 , @xmath473 , @xmath474 , and @xmath475 a constant , and @xmath476 satisfies analogous conditions with the obvious change of coordinates ( cf . the construction in sect .
[ subsec : compactsupport ] ) .
we are now ready to present the main result , which states that under setup [ setup : assumptions - on - generators - bl-2d ] or setup [ setup : assumptions - on - generators - cs-2d ] shearlets provide optimally sparse approximations for cartoon - like images .
[ thm : opt - sparse-2d ] assume setup [ setup : assumptions - on - generators - bl-2d ] or [ setup : assumptions - on - generators - cs-2d ] .
let @xmath478 .
for any @xmath479 and @xmath480 , the shearlet frame @xmath467 provides optimally sparse approximations of functions @xmath400 in the sense of dfn .
[ def : optimal ] , i.e. , @xmath481 where @xmath482 and @xmath483 is a decreasing ( in modulus ) rearrangement of @xmath105 .
before we delve into the proof of thm .
[ thm : opt - sparse-2d ] , we first carefully discuss the main differences between band - limited shearlets and compactly supported shearlets which requires adaptions of the proof . in the case of
compactly supported shearlets , we can consider the two cases @xmath484 and @xmath485 . in case the support of the shearlet intersects the discontinuity curve @xmath486 of the cartoon - like image @xmath49 , we will estimate each shearlet coefficient @xmath487 individually using the decay assumptions on @xmath322 in setup [ setup : assumptions - on - generators - cs-2d ] , and then apply a simple counting estimate to obtain the sought estimates ( [ eq:2d - approx - rate ] ) and ( [ eq:2d - sparsity - rate ] ) . in the other case ,
in which the shearlet does not interact with the discontinuity , we are simply estimating the decay of shearlet coefficients of a @xmath5 function .
the argument here is similar to the approximation of smooth functions using wavelet frames and rely on estimating coefficients at all scales using the frame property . in the case of band - limited shearlets
, it is not allowed to consider two cases @xmath488 and @xmath489 separately , since all shearlet elements @xmath296 intersect the boundary of the set @xmath78 .
in fact , one needs to first localize the cartoon - like image @xmath49 by compactly supported smooth window functions associated with dyadic squares using a partition of unity .
letting @xmath490 denote such a localized version , we then estimate @xmath491 instead of directly estimating the shearlet coefficients @xmath492 .
moreover , in the case of band - limited shearlets , one needs to estimate the sparsity of the sequence of the shearlet coefficients rather than analyzing the decay of individual coefficients . in the next subsections we present the proof
first for band - limited , then for compactly supported shearlets in the case @xmath24 , i.e. , when the discontinuity curve in the model of cartoon - like images is smooth .
finally , the extension to @xmath493 will be discussed for both cases simultaneously .
we will first , however , introduce some notation used in the proofs and prove a helpful lemma which will be used in both cases : band - limited and compactly supported shearlets . for a fixed @xmath213 ,
we let @xmath494 be a collection of dyadic squares defined by @xmath495 \times [ \tfrac{l_2}{2^{j/2 } } , \tfrac{l_2 + 1}{2^{j/2 } } ] : l_1,l_2 \in { { \ensuremath{\mathbb{z}}}}\}.\ ] ] we let @xmath496 denote the set of all indices @xmath497 in the shearlet system and define @xmath498 for @xmath499 , we define the set of ` relevant ' indices on scale @xmath213 as @xmath500 assume setup [ setup : assumptions - on - generators - bl-2d ] or [ setup : assumptions - on - generators - cs-2d ] .
let @xmath501 .
then the following assertions hold : _ ( i ) . _ since @xmath505 for both the band - limited and compactly supported setup , we have that @xmath506 as a consequence , there is a scale @xmath507 such that @xmath508 for each @xmath509 .
it therefore follows from that @xmath510 _ ( ii ) .
_ by assertion ( i ) and estimate , we have that @xmath511 from this , the value @xmath512 can be written as a function of the total number of coefficients @xmath513 .
we obtain @xmath514 this implies that @xmath515 and @xmath516 where @xmath103 as usual denotes the @xmath104th largest shearlet coefficient in modulus .
since we assume @xmath24 , we have that @xmath517 .
as mentioned in the previous section , we will now measure the sparsity of the shearlet coefficients @xmath518 .
for this , we will use the weak @xmath519 quasi norm @xmath520 defined as follows . for a sequence @xmath521 ,
we let , as usual , @xmath522 be the @xmath104th largest coefficient in @xmath523 in modulus .
we then define : @xmath524 one can show @xcite that this definition is equivalent to @xmath525 we will only consider the case @xmath526 since the case @xmath527 can be handled similarly . to analyze the decay properties of the shearlet coefficients @xmath528 at a given scale parameter @xmath529 , we smoothly localize the function @xmath49 near dyadic squares .
fix the scale parameter @xmath318 . for a non - negative @xmath530 function @xmath531 with support in @xmath3 ^ 2 $
] , we then define a smooth partition of unity @xmath532 where , for each dyadic square @xmath533 , @xmath534 .
we will then examine the shearlet coefficients of the localized function @xmath535 . with this smooth localization of the function @xmath49 , we can now consider the two separate cases , @xmath536 and @xmath537 .
let @xmath538 where the union is disjoint and @xmath539 is the collection of those dyadic squares @xmath533 such that the edge curve @xmath19 intersects the support of @xmath540 .
since each @xmath541 has side length @xmath378 and the edge curve @xmath19 has finite length , it follows that @xmath542 similarly , since @xmath49 is compactly supported in @xmath3 ^ 2 $ ] , we see that @xmath543 the following theorems analyzes the sparsity of the shearlets coefficients for each dyadic square @xmath533 . to derive the sought estimates ( [ eq:2d - approx - rate ] ) and ( [ eq:2d - sparsity - rate ] ) for dimension @xmath8 , we will study two separate cases : those shearlet elements @xmath296 which do not interact with the discontinuity curve , and those elements which do . _
case 1_. : : the compact support of the shearlet @xmath296 does not intersect the boundary of the set @xmath78 , i.e.,@xmath555 . _ case 2_. : : the compact support of the shearlet @xmath296 does intersect the boundary of the set @xmath78 , i.e. , @xmath556 . for _ case 1 _
we will not be concerned with decay estimates of single coefficients @xmath557 , but with the decay of sums of coefficients over several scales and all shears and translations .
the frame property of the shearlet system , the @xmath5-smoothness of @xmath49 , and a crude counting argument of the cardinal of the essential indices @xmath558 will be enough to provide the needed approximation rate .
the proof of this is similar to estimates of the decay of wavelet coefficients for @xmath5 smooth functions .
in fact , shearlet and wavelet frames gives the same approximation decay rates in this case .
due to space limitation of this exposition , we will not go into the details of this estimate , but rather focus on the main part of the proof , _
case 2_. for _ case 2 _ we need to estimate each coefficient @xmath557 individually and , in particular , how @xmath559 decays with scale @xmath213 and shearing @xmath214 . without loss of generality we can assume that @xmath560 with @xmath561 .
we let then @xmath562 denote the area of integration in @xmath557 , that is , @xmath563 further , let @xmath429 be an affine hyperplane ( in other and simpler words , a line in @xmath564 ) that intersects @xmath562 and thereby divides @xmath562 into two sets @xmath565 and @xmath566 , see the sketch in fig . [
fig : trunc - and - lin - part - w - m - h ] .
we thereby have that @xmath567 the hyperplane will be chosen in such way that the area of @xmath568 is sufficiently small .
in particular , @xmath569 should be small enough so that the following estimate @xmath570 do not violate ( [ eq:2d - sparsity - rate ] ) . if the hyperplane @xmath429 is positioned as indicated in fig .
[ fig : trunc - and - lin - part - w - m - h ] , it can indeed be shown by crudely estimating @xmath571 that ( [ eq : truncated - estimate-2d - cs ] ) does not violate estimate ( [ eq:2d - sparsity - rate ] ) .
we call estimates of this form , where we have restricted the integration to a small part @xmath565 of @xmath562 , _ truncated _ estimates . hence
, in the following we assume that reduces to @xmath572 . for the term
@xmath573 we will have to integrate over a possibly much large part @xmath566 of @xmath562 . to handle this
, we will use that @xmath422 only interacts with the discontinuity of @xmath574 along a line inside @xmath562 .
this part of the estimate is called the _
linearized _ estimate , since the discontinuity curve in @xmath573 has been reduced to a line . in @xmath573 we are , of course , integrating over two variables , and we will as the inner integration always choose to integrate along lines parallel to the ` singularity ' line @xmath429 , see fig . [
fig : trunc - and - lin - part - w - m - h ] .
the important point here is that along these lines , the function @xmath49 is @xmath5-smooth without discontinuities on the entire interval of integration .
this is exactly the reason for removing the @xmath565-part from @xmath562 . using the fourier slice theorem we will then turn the line integrations along @xmath429 in the spatial domain into line integrations in the frequency domain .
the argumentation is as follows : consider @xmath575 compactly supported and continuous , and let @xmath576 be a projection of @xmath125 onto , say , the @xmath577 axis , i.e.,@xmath578 .
this immediately implies that @xmath579 which is a simplified version of the fourier slice theorem . by an inverse fourier transform
, we then have @xmath580 and hence @xmath581 the left - hand side of ( [ eq : fourier - slice - thm-2 ] ) corresponds to line integrations of @xmath125 along vertical lines @xmath582 . by applying shearing to the coordinates @xmath583
, we can transform @xmath429 into a line of the form @xmath584 , whereby we can apply ( [ eq : fourier - slice - thm-2 ] ) directly .
we will make this idea more concrete in the proof of the following key estimate for linearized terms of the form @xmath585 .
since we assume the truncated estimate as negligible , this will in fact allow us to estimate @xmath557 .
[ thm : decay - hyperplane ] let @xmath586 be compactly supported , and assume that @xmath315 satisfies the conditions in setup [ setup : assumptions - on - generators - cs-2d ] .
further , let @xmath558 be such that @xmath587 .
suppose that @xmath588 and that @xmath22 is linear on the support of @xmath422 in the sense @xmath589 for some affine hyperplane @xmath590 of @xmath564 .
then , a. if @xmath429 has normal vector @xmath591 with @xmath592 , @xmath593 b. if @xmath429 has normal vector @xmath591 with @xmath594 , @xmath595 c. if @xmath429 has normal vector @xmath596 with @xmath427 , then @xmath597 _ cases ( i ) and ( ii)_. we first consider the cases ( i ) and ( ii ) . in these cases ,
the hyperplane can be written as @xmath598 for some @xmath599 .
we shear the hyperplane by @xmath600 for @xmath427 and obtain @xmath601 which is a line parallel to the @xmath577-axis . here
the power of shearlets comes into play , since it will allow us to only consider line singularities parallel to the @xmath577-axis .
of course , this requires that we also modify the shear parameter of the shearlet , that is , we will consider the right hand side of @xmath602 with the new shear parameter @xmath603 .
the integrand in @xmath604 has the singularity plane exactly located on the line @xmath605 , i.e.,on @xmath606 . to simplify the expression for the integration bounds
, we will fix a new origin on @xmath606 , that is , on @xmath607 ; the @xmath577 coordinate of the new origin will be fixed in the next paragraph .
since @xmath49 is only nonzero of @xmath78 , the function @xmath49 will be equal to zero on one side of @xmath606 , say , @xmath608 .
it therefore suffices to estimate @xmath609 for @xmath610 and @xmath611 .
let us assume that @xmath612 .
the other case can be handled similarly . since @xmath315 is compactly
supported , there exists some @xmath613 such that @xmath614}}^2 $ ] . by a rescaling argument
, we can assume @xmath615 .
let @xmath616 with this notation we have @xmath617 .
we say that the shearlet normal direction of the shearlet box @xmath618 is @xmath619 , thus the shearlet normal of a sheared element @xmath377 associated with @xmath620 is @xmath621 .
now , we fix our origin so that , relative to this new origin , it holds that @xmath622 then one face of @xmath623 intersects the origin . next , observe that the parallelogram @xmath624 has sides @xmath625 , @xmath626 as it is only a matter of scaling , we replace the right hand side of the last equation with @xmath349 for simplicity . solving the two last equalities for @xmath577
gives the following lines : @xmath627 we shows that @xmath628 where the upper integration bound for @xmath408 is @xmath629 ; this follows from solving @xmath630 for @xmath408 and using that @xmath631 .
we remark that the inner integration over @xmath577 is along lines parallel to the singularity line @xmath632 ; as mentioned , this allows us to better handle the singularity and will be used several times throughout this section .
we consider the one - dimensional taylor expansion for @xmath633 at each point @xmath634 in the @xmath577-direction : @xmath635 where @xmath636 and @xmath637 are all bounded in absolute value by @xmath638 . using this taylor expansion in ( [ eq : hp - claim2-int - domain ] ) yields @xmath639 where @xmath640 and @xmath641 we next estimate each integral @xmath642 @xmath643 separately .
_ integral @xmath642_. we first estimate @xmath644 . the fourier slice theorem ,
see also ( [ eq : fourier - slice - thm ] ) , yields directly that @xmath645 by the assumptions from setup [ setup : assumptions - on - generators - cs-2d ] we have , for all @xmath646 , @xmath647 for some @xmath648 .
hence , we can continue our estimate of @xmath642 by @xmath649 and further , by a change of variables , @xmath650 since @xmath651 .
_ integral @xmath652_. we start estimating @xmath653 by @xmath654 applying the fourier slice theorem again and then utilizing the decay assumptions on @xmath322 yields @xmath655 since @xmath656 , we have @xmath657 .
the following estimate of @xmath658 then follows directly from the estimate of @xmath642 : @xmath659 from the two last estimate , we conclude that @xmath660 .
_ case ( iii)_. finally , we need to consider the case ( iii ) , in which the normal vector of the hyperplane @xmath429 is of the form @xmath596 for @xmath672 . for this , let @xmath673 . as in the first part of the proof
, it suffices to consider coefficients of the form @xmath674 , where @xmath675 with respect to some new origin .
as before , the boundary of @xmath676 intersects the origin . by the assumptions in setup [ setup
: assumptions - on - generators - cs-2d ] , we have that @xmath677 which implies that @xmath678 therefore , we have @xmath679 since a shearing operation @xmath274 preserves vanishing moments along the @xmath408 axis .
now , we employ taylor expansion of @xmath680 in the @xmath408-direction ( that is , again along the singularity line @xmath681 ) . by ( [ eq : hp - shear - preserves - vm ] ) everything but the last term in the taylor expansion disappears , and we obtain @xmath682 which proves claim ( iii ) . to see this claim , note that for a fixed @xmath213 and @xmath214 we need to count the number of translates @xmath687 for which the support of @xmath377 intersects the discontinuity line @xmath688 , @xmath689 , inside @xmath541 . without loss of generality
, we can assume that @xmath690}}^2 $ ] , @xmath691 , and @xmath692 , where @xmath620 is defined as in ( [ eq : cpjk ] ) .
the shearlet @xmath377 will therefore be concentrated around the line @xmath693 , see also fig .
[ fig : directional - van - mom - spatial ]
. we will count the number of @xmath694 for which these two lines intersect inside @xmath541 since this number , up to multiplication with a constant independent of the scale @xmath213 , will be equal to @xmath695 .
first note that since the size of @xmath541 is @xmath696 , only a finite number of @xmath697 translates can make @xmath698 whenever @xmath699 is fixed . for a fixed @xmath700
, we then estimate the number of relevant @xmath357 translates .
equating the @xmath408 coordinates in @xmath429 and @xmath701 yields @xmath702 without loss of generality , we take @xmath703 which then leads to @xmath704 hence @xmath705 .
this completes the proof of the claim . for @xmath499
, we will consider the shearlet coefficients larger than @xmath512 in absolute value .
thus , we define : @xmath706 where @xmath707 .
since the discontinuity line @xmath429 has finite length in @xmath3 ^ 2 $ ] , we have the estimate @xmath708 .
assume @xmath429 has normal vector @xmath591 with @xmath709 .
then , by thm . [ thm : decay - hyperplane](i )
, @xmath710 implies that @xmath711(i ) and the estimates and , we have that @xmath712 where , as usual , @xmath713 . by lem . [
lem : suff - lambda - eps - cond](ii ) , this leads to the sought estimates . on the other hand ,
if @xmath429 has normal vector @xmath714 or @xmath591 with @xmath715 , then @xmath716 implies that @xmath717 which follows by assertions ( ii ) and ( iii ) in thm .
[ thm : decay - hyperplane ] .
hence , we have @xmath718 note that @xmath719 , since @xmath720 for each @xmath533 , and that the number of shear parameters @xmath214 for each scale parameter @xmath318 is bounded by @xmath721 .
therefore , @xmath722 this implies our sought estimate which , together with the estimate for @xmath723 , completes the proof of thm .
[ thm : opt - sparse-2d ] for @xmath24 under setup [ setup : assumptions - on - generators - cs-2d ] .
we now turn to the extended class of cartoon - lime images @xmath725 with @xmath493 , i.e. , in which the singularity curve is only required to be piecewise @xmath5 .
we say that @xmath726 is a corner point if @xmath22 is not @xmath5 smooth in @xmath217 .
the main focus here will be to investigate shearlets that interact with one of the @xmath727 corner points .
we will argue that thm .
[ thm : opt - sparse-2d ] also holds in this extended setting .
the rest of the proof , that is , for shearlets _ not _ interacting with corner points , is of course identical to that presented in sect .
[ sec : proof - band - limited-2d ] and [ sec : proof - comp - supp-2d ] . in the compactly supported case one can simply count the number of shearlets interacting with a corner point at a given scale . using lem .
[ lem : suff - lambda - eps - cond](i ) , one then arrives at the sought estimate . on the other hand , for the band - limited case
one needs to measure the sparsity of the shearlet coefficients for @xmath49 localized to each dyadic square .
we present the details in the remainder of this section . in this case
, it is sufficient to consider a dyadic square @xmath728 with @xmath318 such that @xmath541 contains a singular point of edge curve .
especially , we may assume that @xmath213 is sufficiently large so that the dyadic square @xmath545 contains a single corner point of @xmath19 .
the following theorem analyzes the sparsity of the shearlet coefficients for such a dyadic square @xmath729 .
the proof of thm .
[ thm : piece_nonsmooth ] is based on a proof of an analog result for curvelets @xcite .
although the proof in @xcite considers only curvelet coefficients , essentially the same arguments , with modifications to the shearlet setting , can be applied to show thm .
[ thm : piece_nonsmooth ] .
finally , we note that the number of dyadic squares @xmath728 containing a singular point of @xmath19 is bounded by a constant not depending on @xmath213 ; one could , e.g.,take @xmath727 as this constant . therefore , applying thm .
[ thm : piece_nonsmooth ] and repeating the arguments in sect .
[ sec : proof - band - limited-2d ] completes the proof of thm .
[ thm : opt - sparse-2d ] for @xmath493 for setup [ setup : assumptions - on - generators - bl-2d ] .
_ case 1_. : : the shearlet @xmath296 intersects a corner point , in which two @xmath5 curves @xmath733 and @xmath734 , say , meet ( see fig .
[ fig : shear - a ] ) . _ case 2_. : : the shearlet @xmath296 intersects two edge curves @xmath733 and @xmath734 , say , simultaneously , but it does not intersect a corner point ( see fig . [
fig : shear - b ] ) .
intersecting two edge curves @xmath733 and @xmath734 which are part of the boundary of sets @xmath735 and @xmath736 .
@xmath737 and @xmath738 are tangents to the edge curves @xmath733 and @xmath734 in points contained in the support of @xmath296.,title="fig:",height=134 ] ( -60,44 ) ( -110,70 ) ( -33,28 ) ( -30,55 ) ( -19,13 ) ( -12,74 ) intersecting two edge curves @xmath733 and @xmath734 which are part of the boundary of sets @xmath735 and @xmath736 .
@xmath737 and @xmath738 are tangents to the edge curves @xmath733 and @xmath734 in points contained in the support of @xmath296.,title="fig:",height=134 ] ( -80,44 ) ( -110,70 ) ( -26,32 ) ( -27,58 ) ( -19,13 ) ( 1,72 ) since there exist only finitely many corner points with total number not depending on scale @xmath318 and the number of shearlets @xmath296 intersecting each of corner points is bounded by @xmath721 , we have @xmath740 as illustrated in fig . [
fig : shear - b ] , we can write the function @xmath49 as @xmath741 where @xmath742 ^ 2)$ ] and @xmath743 are two disjoint subsets of @xmath3 ^ 2 $ ] .
as we indicated before , the rate for optimal sparse approximation is achieved for the smooth function @xmath744 .
thus , it is sufficient to consider @xmath745 with @xmath746 ^ 2)$ ] . by a _
truncated _ estimate , we can replace two boundary curves @xmath733 and @xmath734 by hyperplanes of the form @xmath747 in the sequel , we assume @xmath748 and mention that the other cases can be handled similarly . next define @xmath749 for each @xmath750 , where @xmath751 denotes the dyadic squares containing the two distinct boundary curves . by an estimate similar to , we obtain @xmath752 applying thm .
[ thm : decay - hyperplane](i ) to each of the hyperplanes @xmath737 and @xmath738 , we also have @xmath753 for @xmath18 . without loss of generality , we may assume that @xmath754 .
then , and imply that @xmath755 and @xmath756 as follows : @xmath757 note that @xmath758 since the number of @xmath533 containing two distinct boundary curves @xmath733 and @xmath734 is bounded by a constant independent of @xmath213 .
the result is proved .
when passing from 2d to 3d , the complexity of anisotropic structures changes significantly .
in particular , as opposed to the two dimensional setting , geometric structures of discontinuities for piecewise smooth 3d functions consist of two morphologically different types of structure , namely surfaces and curves . moreover , as we saw in sect .
[ sec : optim - sparse - appr-2d ] , the analysis of sparse approximations in 2d heavily depends on reducing the analysis to affine subspaces of @xmath564 . clearly , these subspaces always have dimension one in 2d . in dimension three
, however , we have subspaces of dimension one and two , and therefore the analysis needs to performed on subspaces of the ` correct ' dimension .
this issue manifests itself when performing the analysis for band - limited shearlets , since one needs to replace the radon transform used in 2d with a so - called x - ray transform . for compactly supported shearlets
, one needs to perform the analysis on carefully chosen hyperplanes of dimension two .
this will allow for using estimates from the two dimensional setting in a slice by slice manner .
as in the two dimensional setting , analyzing the decay of individual shearlet coefficients @xmath492 can be used to show optimal sparsity for compactly supported shearlets while the sparsity of the sequence of shearlet coefficients with respect to the weak @xmath519 quasi norm should be analyzed for band - limited shearlets . for case ( b ) there are at most @xmath759 coefficients at scale @xmath414 , since the plate - like elements are of size @xmath378 times @xmath378 ( and ` thickness ' @xmath379 ) .
by hlder s inequality , we see that @xmath760 for some constants @xmath417 .
hence , we have @xmath759 coefficients bounded by @xmath761 . assuming the coefficients in case ( a ) and ( c ) to be negligible , the @xmath104th largest shearlet coefficient @xmath103 is therefore bounded by @xmath762 which in turn implies @xmath763 hence
, we meet the optimal rates ( [ eq : approx - fastest - decay - upper ] ) and ( [ eq : coeff - fastest - decay - upper ] ) from dfn .
[ def : optimal ] .
this , at least heuristically , shows that shearlets provide optimally sparse approximations of 3d cartoon - like images .
the hypotheses needed for the band - limited case , stated in setup [ setup : assumptions - on - generators - bl-3d ] , are a straightforward generalization of setup [ setup : assumptions - on - generators - bl-2d ] in the two - dimensional setting . [ setup : assumptions - on - generators - bl-3d ] the generators @xmath764 are band - limited and @xmath466 in the frequency domain .
furthermore , the shearlet system @xmath341 forms a frame for @xmath314 ( cf . the construction in sect . [ subsec : bandlimited ] ) . for the compactly supported generators we will also use hypotheses in the spirit of setup [ setup : assumptions - on - generators - cs-2d ] , but with slightly stronger and more sophisticated assumption on vanishing moment property of the generators i.e.,@xmath765 and @xmath666 .
[ setup : assumptions - on - generators - cs-3d ] the generators @xmath766 are compactly supported , and the shearlet system @xmath341 forms a frame for @xmath314 .
furthermore , the function @xmath315 satisfies , for all @xmath767 , for @xmath770 , where @xmath765 , @xmath666 , @xmath474 , and @xmath475 a constant , and @xmath476 and @xmath372 satisfy analogous conditions with the obvious change of coordinates ( cf . the construction in sect .
[ subsec : compactsupport ] ) .
[ thm : opt - sparse-3d ] assume setup [ setup : assumptions - on - generators - bl-3d ] or [ setup : assumptions - on - generators - cs-3d ] .
let @xmath24 .
for any @xmath479 and @xmath480 , the shearlet frame @xmath341 provides optimally sparse approximations of functions @xmath771 in the sense of dfn .
[ def : optimal ] , i.e. , @xmath772 where @xmath773 and @xmath483 is a decreasing ( in modulus ) rearrangement of @xmath105 .
the proof of thm .
[ thm : opt - sparse-3d ] for band - limited shearlets follows the same steps as discussed in sect .
[ sec : proof - band - limited-2d ] for the 2d case . to indicate the main steps
, we will use the same notation as for the 2d proof with the straightforward extension to 3d .
the proofs of thm .
[ thm : nonsmooth-3d ] and [ thm : smooth-3d ] follow the same principles as the proofs of the analog results in 2d , thm .
[ thm : nonsmooth ] and [ thm : smooth ] , with one important difference : in the proof of thm .
[ thm : nonsmooth ] and [ thm : smooth ] the radon transform ( cf .
( [ eq : fourier - slice - thm ] ) ) is used to deduce estimates for the integral of edge - curve fragments . in 3d
one needs to use a different transform , namely the so - called x - ray transform , which maps a function on @xmath23 into the sets of its line integrals .
the x - ray transform is then used to deduce estimates for the integral of the _ surface _ fragments .
we refer to @xcite for a detailed exposition . by thm .
[ thm : both-3d ] , we can now prove thm .
[ thm : opt - sparse-3d ] for the band - limited setup and for @xmath778 with @xmath24 .
the proof is very similar to the proof of thm .
[ thm : opt - sparse-2d ] in sect .
[ sec : proof - band - limited-2d ] , wherefore we will not repeat it . in this section we will consider the key estimates for the linearized term for compactly supported shearlets in 3d . this is an extension of thm .
[ thm : decay - hyperplane ] to the three - dimensional setting .
hence , we will assume that the discontinuity surface is a plane , and consider the decay of the shearlet coefficients of shearlets interacting with such a discontinuity .
[ thm : decay - hyperplane-3d ] let @xmath298 be compactly supported , and assume that @xmath315 satisfies the conditions in setup [ setup : assumptions - on - generators - cs-3d ] .
further , let @xmath558 be such that @xmath587 .
suppose that @xmath774 and that @xmath22 is linear on the support of @xmath422 in the sense that @xmath779 for some affine hyperplane @xmath36 of @xmath23 .
then , a. if @xmath36 has normal vector @xmath780 with @xmath781 and @xmath782 , @xmath783 b. if @xmath36 has normal vector @xmath780 with @xmath784 or @xmath785 , @xmath786 c. if @xmath36 has normal vector @xmath787 with @xmath788 , then @xmath789 fix @xmath558 , and let @xmath774 .
we first consider the case ( ii ) and assume @xmath784 .
the hyperplane can be written as @xmath790 for some @xmath791 . for @xmath792 ,
we consider the restriction of @xmath36 to the slice @xmath793 .
this is clearly a line of the form @xmath794 for some @xmath795 , hence we have reduced the singularity to a line singularity , which was already considered in thm .
[ thm : decay - hyperplane ] .
we apply now thm .
[ thm : decay - hyperplane ] to each on slice , and we obtain @xmath796 the first term @xmath797 in the estimate above is due to the different normalization factor used for shearlets in 2d and 3d , the second term is the conclusion from thm .
[ thm : decay - hyperplane ] , and the third is the length of the support of @xmath422 in the direction of @xmath798 .
the case @xmath785 can be handled similarly with restrictions to slices @xmath799 for @xmath800 .
this completes the proof of case ( ii ) .
neglecting truncated estimates , thm .
[ thm : decay - hyperplane-3d ] can be used to prove the optimal sparsity result in thm .
[ thm : opt - sparse-3d ] .
the argument is similar to the one in sect .
[ sec : proof - comp - supp-2d ] and will not be repeated here .
let us simply argue that the decay rate @xmath801 from thm .
[ thm : decay - hyperplane-3d](ii ) is what is needed in the case @xmath802 .
it is easy to see that in 3d an estimate of the form @xmath803 will guarantee optimal sparsity .
since we in the estimate @xmath804 have no control of the shearing parameter @xmath309 , we have to use a crude counting estimate , where we include all shears at a given scale @xmath213 , namely @xmath805 .
since the number of dyadic boxes @xmath541 where @xmath19 intersects the support of @xmath49 is of order @xmath806 , we arrive at @xmath807 paralleling the two - dimensional setting ( see sect .
[ sec : case - l - neq-1 ] ) , we can extend the optimality result in thm .
[ thm : opt - sparse-3d ] to the cartoon - like image class @xmath808 for @xmath478 , in which the discontinuity surface @xmath19 is allowed to be _ piecewise _ @xmath5 smooth .
moreover , the requirement that the ` edge ' @xmath19 is piecewise @xmath5 might be too restrictive in some applications . therefore , in @xcite , the cartoon - like image model class was enlarged to allow less regular images , where @xmath19 is piecewise @xmath809 smooth for @xmath810 , and not necessarily a @xmath5 .
this class @xmath811 was introduced in sect .
[ sec : cartoon - images ] consisting of _ generalized _ cartoon - like images having @xmath29 smoothness apart from a piecewise @xmath30 discontinuity curve .
the sparsity results presented above in thm .
[ thm : opt - sparse-3d ] can be extended to this generalized model class for compactly supported shearlets with a scaling matrix dependent on @xmath812 .
the optimal approximation error rate , as usual measured in @xmath813 , for this generalized model is @xmath814 ; compare this to @xmath395 for the case @xmath815 considered throughout this chapter .
for brevity we will not go into details of this , but mention the approximation error rate obtained by shearlet frames is slightly worse than in the @xmath34 case , since the error rate is not only a poly - log factor away from the optimal rate , but a small polynomial factor ; and we refer to @xcite the precise statement and proof .
capturing anisotropic phenomenon in 3d is somewhat different from capturing anisotropic features in 2d as discussed in sect .
[ subsec:3d ] .
while in 2d we ` only ' have to handle curves , in 3d a more complex situation can occur since we find two geometrically very different anisotropic structures : curves and surfaces .
curves are clearly one - dimensional anisotropic features and surfaces two - dimensional features .
since our 3d shearlet elements are plate - like in spatial domain by construction , one could think that these 3d shearlet systems would _ only _ be able to efficiently capture two - dimensional anisotropic structures , and _ not _ one - dimensional structures . nonetheless , surprisingly , as we have discussed in sect .
[ sec : some - extensions ] , these 3d shearlet systems still perform optimally when representing and analyzing 3d data @xmath808 that contain _ both _ curve and surface singularities ( see e.g.,fig . [
fig : cartoon - piecewise ] ) .
p. kittipoom , g. kutyniok , and w .- q lim , _ construction of compactly supported shearlet frames _ , preprint .
s. kuratsubo , _ on pointwise convergence of fourier series of the indicator function of @xmath817 dimensional ball _
, j. fourier anal .
* 16 * ( 2010 ) , 5259 . | multivariate functions are typically governed by anisotropic features such as edges in images or shock fronts in solutions of transport - dominated equations .
one major goal both for the purpose of compression as well as for an efficient analysis is the provision of optimally sparse approximations of such functions .
recently , cartoon - like images were introduced in 2d and 3d as a suitable model class , and approximation properties were measured by considering the decay rate of the @xmath0 error of the best @xmath1-term approximation .
shearlet systems are to date the only representation system , which provide optimally sparse approximations of this model class in 2d as well as 3d .
even more , in contrast to all other directional representation systems , a theory for compactly supported shearlet frames was derived which moreover also satisfy this optimality benchmark .
this chapter shall serve as an introduction to and a survey about sparse approximations of cartoon - like images by band - limited and also compactly supported shearlet frames as well as a reference for the state - of - the - art of this research field . |
because they do not interact with each other and can be guided via optical fibers over long distances with relatively low losses , photons appear as ideal information carriers and are therefore put forward as the flying qubits in most of quantum communication protocols .
the design of memories able to reliably store and retrieve photonic states is , however , still an open problem . the most commonly studied protocol , considered to implement such a quantum memory , is electromagnetically induced transparency ( eit ) @xcite .
this protocol was implemented in various systems such as cold atoms , gas cells , or doped crystals @xcite .
although the doppler broadening might seem to lead to strong limitations , eit - based light storage in warm alkali vapors gives good results and is still a subject of active investigation @xcite . in the last years
, some experiments were also performed in a raman configuration , using pulses which are highly detuned from the optical resonances in gas cells @xcite .
the eit - based storage protocol in a @xmath0 atomic system relies on the long - lived raman coherence between the two ground states which are optically coupled to the excited level . when a strong coupling beam is applied along one of the two transitions , a narrow transparency window limited by the raman coherence decay rate
is opened along the other leg of the system .
because of the slow - light effect associated with such a dramatic change of the medium absorption properties , a weak probe pulse on the second transition is compressed while propagating through the medium .
when this pulse has fully entered the atomic medium , it can be mapped onto the raman coherences which are excited by the two - photon process by suddenly switching off the coupling beam .
it can be safely stored during times smaller than the lifetime of raman coherence .
finally , the signal pulse can be simply retrieved by switching on the coupling beam again . in the raman configuration ,
the coupling and probe pulses are optically far off - resonance but still fulfill the two - photon transition condition .
the advantage is a large bandwidth , that allows to work with data rates higher than in the usual eit regime @xcite .
atoms at room temperature in a gas cell are particularly attractive for light storage because of the simplicity of their implementation .
the effects of the significant doppler broadening can be minimized using co - propagating coupling and probe beams , so that the two - photon resonance condition can be verified for all velocity classes : all the atoms can thus participate to the eit phenomenon as soon as they are pumped in the probed level . as a consequence , handy simple gas cells have turned out to be attractive for slow or even stopped light experiments @xcite . in a previous work @xcite
, we have reported on an added phase shift recorded for eit - based light storage experiments carried out in a helium gas at room temperature when the coupling beam is detuned from the center of the doppler line .
the simple model that we have derived could not satisfactorily account for our observations that were recorded for intermediate detunings , e.g. close to the doppler broadening of the transition . in the present paper , we come back to this problem and provide new experimental results , _ i.e. _ time - dependent measurements of the retrieved signal phase shift , as well as numerical results obtained through the simulation of the full system of maxwell - bloch equations .
the behaviour of these phase shifts with the coupling detuning seems satisfactorily accounted for by our simulations .
we also perform numerical calculations in the raman regime .
the paper is organized as follows . in section [ secii ]
we present the system and setup and describe how to measure the time - dependent phase shift of the retrieved pulse with respect to the coupling beam .
we also briefly recall the system of maxwell - bloch equations which governs our system and describe their numerical integration . in section [ seciii ] ,
we provide our experimental and numerical results and show that they qualitatively agree .
we also apply our simulations to the far off - resonant raman case .
finally , we conclude in section [ seciv ] and give possible perspectives of our work .
the atoms preferably used for eit storage experiments are alkali atoms , mainly rubidium and sometimes sodium or caesium .
we choose here to work with metastable @xmath1he atoms , which have the advantage of a very simple structure without hyperfine levels : transitions are thus far enough one from another to investigate the effect of detunings of the coupling and probe beams on light storage and retrieval . in our setup represented in fig .
[ experimental scheme ] , a @xmath2-cm - long cell is filled up with @xmath3 torr of helium atoms which are continuously excited to their metastable state @xmath4 by a radio - frequency ( rf ) discharge at 27 mhz .
each of the metastable ground states @xmath5 is hence fed with the same rate , denoted by @xmath6 .
the cell is isolated from magnetic field gradients by a three - layer @xmath7-metal shield to avoid spurious dephasing effects on the different zeeman components . a strong circularly - polarized field , called the control beam , propagates along the quantization axis @xmath8 .
its power is set at @xmath9 mw for a beam diameter of @xmath10 mm . as shown in fig .
[ atomic struture ] , the coupling field drives the transitions @xmath11 and @xmath12 . owing to the spontaneous transitions @xmath13 and @xmath14 , the atoms end up in the state @xmath15 within a few pumping cycles after the coupling beam has been switched on . as the atoms are at room temperature , the doppler broadening in the cell is @xmath16 .
we denote by @xmath17 the detuning of the coupling frequency @xmath18 with respect to the natural frequency @xmath19 of the transition @xmath20 , at the center of the doppler line .
once optical pumping is achieved , a weak signal pulse is sent through the atomic medium along the @xmath8 axis .
its polarization is circular and orthogonal to that of the coupling beam : the signal therefore couples the state @xmath21 to @xmath22 and we denote by @xmath23 the detuning of the signal frequency @xmath24 from the center of the doppler profile .
both signal and coupling beams are derived from the same laser diode , and their frequencies and amplitudes are controlled by two acousto - optic modulators . due to the efficiency of optical pumping through the coupling beam
, we assume that the state @xmath25 remains essentially unpopulated during the whole process and we accordingly neglect the driving of the transition @xmath26 by the signal field .
submitted to the coupling and signal fields , the atoms therefore essentially evolve in the three - level @xmath0 system @xmath27 ( see fig . [ atomic struture ] ) as long as the detunings @xmath28 , of the coupling and signal fields respectively , are small enough to avoid exciting neighbouring transitions .
thanks to the absence of hyperfine structure , the range of allowed values for @xmath29 is , however , much larger than in alkali vapor experiments : indeed , on the positive detuning side the nearest state ( @xmath30 ) is @xmath31 ghz away from optical resonance @xmath32 , while , on the negative detuning side , the nearest state ( @xmath33 ) is @xmath34 ghz away .
under eit conditions , the coupling beam opens a transparency window for the weak signal beam , which can therefore propagate without absorption through the medium if its spectrum is not too wide @xcite . in the experimental results we present
hereafter , we used a signal pulse , which consists of a smoothly increasing exponential followed by an abruptly decaying exponential of respective characteristic times @xmath35 and @xmath36 ns .
its maximum power is about @xmath37 and the beam diameter is about @xmath10 mm .
different dissipative mechanisms influence the width of the eit window besides spontaneous emission , such as collisions and transit of the atoms in and out of the beams .
these phenomena result in the decay of atomic coherences at the rates @xmath38khz for the raman coherence @xmath39 and @xmath40 for the optical coherence @xmath41 .
we have shown previously that velocity changing collisions redistribute the pumping of the atoms over an effective width sligthly smaller than the doppler linewidth @xcite . in our conditions , with a coupling power of 18 mw , this effective width is experimentally estimated to be @xmath42ghz .
due to power broadening , the width of the transparency window is then of the order of @xmath43 khz .
the highly dispersive character of the medium under eit conditions can be used furthermore to store and retrieve a weak signal pulse : due to eit dispersion , the signal pulse indeed travels with a reduced group velocity and is temporally contracted .
once the pulse has entered the cell , the coupling beam can be switched off : information about the signal pulse is then stored into the raman coherences . after a storage time @xmath44 , the control beam is switched on again , which releases the signal from the atomic coherence and ensures eit absorption - free propagation .
note that in our experimental setup , the optical depth is only about 3.5 and the pulse can thus not be fully compressed in the cell .
due to the finite width of the transparency window and the finite length of the cell , a part , typically @xmath45 % of the incoming signal energy , leaks out before the coupling beam is turned off and the storage period begins @xcite .
a typical experimental record is given in fig .
[ data processing]b .
the first part of the detected signal is the leak and after a storage time @xmath46s , once the coupling beam is turned on again , the retrieved signal is clearly visible .
the same vertical scale is used in graphs a , b and c. a ) accumulated plot of the interference signal @xmath47 between the probe and coupling beams ( blue ) .
upper and lower envelopes @xmath48 are shown in orange and red , respectively .
b ) signal intensity @xmath49 at the exit of the cell , deduced from the previous accumulated plot a ) and from a measurement of the coupling intensity @xmath50 .
c ) interference signal @xmath47 between the probe and coupling beams ( blue ) .
it is contained between the upper and lower envelopes @xmath48 , shown in orange and red , respectively .
one can note the presence of , width=272 ] in @xcite , we investigated the relative phase @xmath51 of the signal with respect to the coupling beam , and showed the existence of an optical detuning dependent extra phase shift @xmath52 between the incident and retrieved pulse .
this quantity can be measured through mixing the signal emerging from the cell with a small fraction of the control beam , via polarization optics .
the resulting intensity thus takes the form @xmath53.\label{intensity}\ ] ] in eq .
( [ intensity ] ) , the contrast factor @xmath54 , which ideally equals @xmath55 , accounts for non - perfect alignment of the beams and is measured for each set of data .
@xmath50 denotes the intensity of the small fraction of the coupling field which is mixed and interferes with the signal field .
it takes the same constant value @xmath50 during the writing and retrieval periods , while it vanishes during the storage time .
the value @xmath50 is measured in the absence of the signal ( one assumes that the introduction of the signal pulse does not substantially affect the measurement of @xmath50 ) .
the phase of the coupling beam is varied via a piezoelectric actuator from one experimental run to another : the scan is slow enough so that it is assumed to be constant during both the writing or retrieval steps .
@xmath49 is the time - dependent intensity of the signal beam emerging from the cell .
we denote by @xmath56 and @xmath57 the intensities of the leak and retrieved pulses , respectively . accordingly
, we introduce @xmath58 and @xmath59 , the relative phases of the leak and retrieved pulses , with respect to the coupling beam . to obtain the extra phase shift @xmath52 between the incident and retrieved pulses
, we measure the relative phases @xmath60 , @xmath61 by homodyne detection . repeating the same writing - storage - retrieval sequence for many different postions of the piezoelectric actuator ,
we obtain an accumulated plot whose upper / lower envelopes correspond respectively to @xmath62 ( see fig . [ data processing]a ) . given the previously measured value of @xmath50 , one can infer @xmath49 from @xmath63 and @xmath54 from @xmath64 ( see fig .
[ data processing]b ) . for a given position of the piezo - actuator , one
can then obtain @xmath58 and @xmath59 through fitting the experimental record with eq .
( [ intensity ] ) at each time @xmath65 ( see figs . [ data processing]c , d ) .
it was verified both experimentally and numerically that the phase of the leak is constant ( @xmath66 ) and can therefore be taken as a reference for the time - dependent relative phase of the retrieved pulse @xmath59 .
this time independence of the leak phase is ensured by the fact that its spectral content is much narrower than the eit bandwidth .
this is obtained thanks to the shape of the signal pulse : a slow exponential increase , followed by a sharp decrease . the part of the pulse which contains only low frequencies enters first and gives form to a leak , whose phase is constant .
the `` extra phase shift '' is then measured as @xmath67\equiv\varphi_{eit}(t)$ ] .
let us stress that in @xcite , we assumed that the relative phases @xmath68 of the leak and retrieved pulses were time - independent : therefore , we directly extracted effective `` averaged '' values for @xmath54 and @xmath52 by performing a two - parameter fit of the data with eq .
( [ intensity ] ) . here ,
by contrast , we measure the time - dependence of the phases and provide experimental plots for @xmath69 , without any assumption on its behaviour . for numerical simulations
, we described the system in the one - dimensional approximation . on the dimensions of the atomic sample ,
the coupling and probe transverse profiles are assumed to remain constant .
these fields can therefore be cast under the form @xmath70\mbox { , } \end{aligned}\ ] ] where @xmath71 denote the respective slowly - varying amplitudes of the control and signal fields , @xmath72 and @xmath73 stand for their respective frequencies and wavenumbers , while @xmath74 ( @xmath75 and @xmath76 define an arbitrary basis in the plane perpendicular to the propagation direction @xmath77 ) .
following e.g. @xcite , we model the atomic sample as a continuous medium of uniform linear density @xmath78 , and define the average density matrix of the slice @xmath79 $ ] by @xmath80 we moreover define the density matrix elements @xmath81 where @xmath82 refer to the atomic levels and can take the values @xmath83 or @xmath84 ( see fig . [ atomic struture ] ) .
we introduce the slowly - varying coherences @xmath85 and @xmath86 defined by : @xmath87 and write the bloch equations in the rotating wave approximation , for the class of velocity which is at the center of the doppler profile : @xmath88,\label{bloch5}\\ \partial_{t}\tilde{\sigma}_{1e } & = & -\left(\gamma_{1e}+\mbox{i}\delta_{s}\right)\tilde{\sigma}_{1e}+\mbox{i}\left\ { \omega_{s}^{*}\left(\sigma_{ee}-\sigma_{11}\right)-\omega_{c}^{*}\tilde{\sigma}_{1 - 1}\right\ } .\label{bloch6}\end{aligned}\ ] ] here , @xmath89 is the population decay rate of the state @xmath90 , and the rabi frequencies @xmath91 are defined by * @xmath92 * where @xmath93 are the relevant matrix elements of the dipole operator @xmath94 . to take into account all the atoms that are distributed in different velocity classes over the doppler linewidth , we developed a simple model , in which the optical coherence decay rates @xmath95 are replaced by the effective doppler width @xmath96 .
this gives satisfactory results thanks to the redistribution of the pumping by velocity changing collisions @xcite .
all our simulations were performed using this purely homogeneous broadening model .
consequently , we call @xmath29 the optical detuning , implicitly defined with respect to the center of the doppler profile . to ensure that the full population remains constant , the discharge - assisted ground - state feeding rate @xmath0 has been set to @xmath97 , the transit rate of the atoms through the laser beam
. moreover , while the state @xmath98 is fed with the rate @xmath99 , the state @xmath21 is effectively fed with the rate @xmath100 . as can be checked by considering the full 6-level system including not only the @xmath0 system of interest but also the states @xmath101
, the state @xmath21 is indeed directly fed by the rf discharge with the rate @xmath6 , but also indirectly via the state @xmath25 whose population is ( almost ) immediately transferred to @xmath21 through optical pumping . finally , in the medium , the fields propagate according to the helmholtz equation , written in the slowly - varying envelope approximation @xmath102 where @xmath103 . the set of maxwell - bloch equations eqs ( [ bloch1]-[propagation ] )
was numerically solved in matlab using the lax discretization method @xcite .
the medium was split into @xmath104 spatial steps of @xmath105 while the whole storage / retrieval sequence was split into @xmath106 timesteps of @xmath55 ps .
we present and discuss our numerical results in the following section .
all the experimental results and simulations are performed at two - photon resonance , which means that the coupling and signal optical detunings are equal . fig .
[ results ] shows experimental records for the time - dependent extra phase shift @xmath69 , achieved with different values of the optical detunings @xmath107 , between @xmath108 and @xmath55 ghz .
the detunings are set here on the positive side , where the nearest state ( @xmath30 ) is nearly @xmath31 ghz away .
each curve is obtained after averaging over 15 sets of data , recorded at different times , for different positions of the homodyne detection piezo - actuator .
the traces recorded on the oscilloscope present some spurious oscillations at a period of about @xmath109 ns . this noise is generated by the acousto - optic modulators and could be removed by numerically filtering the spurious frequencies during the data processing . in fig .
[ results ] , the time origin corresponds to the begining of the retrieval , when the coupling beam is turned on again . at that time , the probe intensity starts increasing to form the retrieved pulse .
we only plot the evolution of the extra phase shift @xmath69 when the signal intensity is high enough , typically from roughly @xmath104 ns to @xmath110 after the start of the retrieval .
one can see that @xmath69 is not constant over the retrieval , and its magnitude increases with the optical detuning @xmath111 .
the experimental plots are compared with numerical simulations of the full maxwell - bloch equations derived as explained in the previous section .
these simulations are in good agreement with experimental results : both present the same general shape for @xmath112 , and the same qualitative behaviour with the optical detuning @xmath111 .
one possible source for the observed discrepancies is our oversimplified treatment of the velocity distribution in the doppler profile . here
, we indeed assumed that velocity changing collisions are efficient enough to instantaneously and perfectly redistribute atoms pumped in the probed level @xmath21 over the effective doppler profile so that all these atoms contribute coherently to the storage process as if the broadening were homogeneous .
although this approximation is commonly used ( see @xcite ) , it might be severely questioned here , especially in optically detuned conditions that change the thermal equilibrium . in particular ,
the absorption of the coupling beam measured experimentally could not be well reproduced by the simulations .
this should also have an effect on the storage efficiency and on the temporal shape of the phase .
note that in the simulation program , in order to `` minimize '' this problem , we use an averaged coupling intensity over the length of the cell as the input parameter , instead of the real coupling intensity measured at the entrance of the cell . note also that we have checked our numerical results agree with the analytic approximate solutions presented in @xcite . to understand better the physical origin of the extra phase - shift @xmath69
, we compared the time - dependent relative phase @xmath113 between the signal and coupling beams at the exit of the cell , obtained : _ i _ ) during the storage and retrieval of a weak signal pulse ( @xmath114 before @xmath115 and @xmath116 after the storage time @xmath117 , and _ ii _ ) during the direct eit - propagation of the same weak pulse in the medium , while the coupling amplitude remains constant .
to compute @xmath118 in case _ i _ ) , we used the full simulation of maxwell - bloch set of equations , whereas in case _ ii _ ) we simply propagated each spectral component @xmath119 of the incoming pulse with the corresponding susceptibility @xmath120 .
[ susceptibility ] simultaneously displays the results we obtained in both cases , for two different values of the optical detuning @xmath121 .
the shape and order of magnitude for @xmath113 are clearly the same in cases _ i _ , _ ii _ ) : the main effect of the storage is to introduce a delay corresponding to the storage time @xmath44 .
here we chose @xmath122 but we checked both experimentally and theoretically that this phase shift does not depend on this storage time .
this suggests that the observed extra phase shift @xmath52 is essentially due to the propagation under eit conditions .
indeed , the stored part presents a sharp decrease associated with many frequency components , and is thus very sensitive to dispersive effects .
this problem should be taken into account for high speed information applications , like experiments performed in the raman configuration @xcite .
we have thus performed simulations in the far detuned regime as presented on fig .
[ raman ] .
the simulation results shown here are plotted for three different optical detunings @xmath123 ghz , @xmath124 ghz , @xmath125 ghz , much higher than the @xmath126ghz doppler broadening .
they were obtained for a number of atoms @xmath78 which is @xmath45 times higher than in our experimental case , and for a coupling power of @xmath127 mw .
these simulation results demonstrate a similar effect on the retrieved signal pulse phase , even slightly stronger than in our experimental conditions .
we have also checked that our numerical results in the raman configuration agree with the analytic approximate solutions presented in @xcite .
in this paper , we have experimentally investigated a time - dependent extra phase shift that appears in a storage - retrieval experiment , performed in a room temperature atomic cell , in optically detuned conditions .
this phase shift varies with time and does not depend on the storage time .
we have provided numerical simulations which qualitatively agree with the experimental results : in particular , it appears that the magnitude of the relative phase depends on the optical detuning , while its temporal shape is mainly given by the spectrum of the incoming pulse .
we explain the existence of this extra - phase by propagation effects that can be understood by a simple propagation model under eit conditions with optically detuned beams .
discrepancies may be due to an approximate treatment of doppler broadening in the cell .
the results presented here might be of importance , particularly for light storage experiments performed in the far - detuned raman regime , as reported in @xcite .
how these results translate into the regime of quantum light is an intriguing feature that we intend to address in a future work . the work of m .- a.m .
is supported by the dlgation generale larmement ( dga ) , france and the work of j. lugani by an indo - french cefipra funding .
we also thank the labex palm and the rgion ile de france dim nanok for funding . | we report measurements of the time - dependent phases of the leak and retrieved pulses obtained in eit storage experiments with metastable helium vapor at room temperature .
in particular , we investigate the influence of the optical detuning at two - photon resonance , and provide numerical simulations of the full dynamical maxwell - bloch equations , which allow us to account for the experimental results . |
cyclic codes have a lot of applications in communication system , storage system and computers . the decoding error probability and the undetected error probability
are closely related with the weight distributions .
for example , permutation decoding , majority decoding , locator decoding , decoding from the covering polynomials and so on @xcite . in general
the weight distributions are complicated @xcite and difficult to be determined .
in fact , as shown in @xcite and @xcite , the problem of computing weight distribution of a cyclic code is connected to the evaluation of certain exponential sums , which are generally hard to be determined explicitly . for more researches ,
refer to @xcite for the irreducible case , @xcite for the reducible case , and @xcite for recent studies .
especially , for related problems in the binary case with two nonzeros , refer to @xcite and @xcite . in this paper , we focus on the application of higher moments of the exponential sum to determine the weight distribution of a class of ternary cyclic codes with three nonzeros , combining with not only quadratic form but also macwilliams identities , with the help of the computer algebra system magma .
let @xmath0 be a prime .
a linear @xmath1 $ ] code is a @xmath2-dimensional subspace of @xmath3 with minimum ( hamming ) distance @xmath4 .
an @xmath5 $ ] linear code @xmath6 over @xmath7 is called cyclic if @xmath8 implies that @xmath9 where @xmath10 . by identifying the vector @xmath11 with @xmath12/(x^n-1),\ ] ] any linear code @xmath6 of length @xmath13 over @xmath7
represents a subset of @xmath14/(x^n-1)$ ] which is a principle ideal domain . the fact that the code is cyclic is equivalent to that the subset is an ideal .
the unique monic polynomial @xmath15 of minimum degree in this subset is the generating polynomial of @xmath6 , and it is a factor of @xmath16 .
when the ideal does not contain any smaller nonzero ideal , the corresponding cyclic code @xmath6 is called a minimal or an irreducible code . for any @xmath17 ,
the weight of @xmath18 is @xmath19 .
the weight enumerator of a code @xmath6 is defined by @xmath20 where @xmath21 denotes the number of codewords with hamming weight @xmath22 .
the sequence @xmath23 is called the weight distribution of the code , which is an important parameter of a linear block code .
assume that @xmath24 and @xmath25 for an even integer @xmath26 .
let @xmath27 be a primitive element of @xmath28
. in this paper , section [ secii ] presents the basic notations and preliminaries about cyclic codes .
section [ seciii ] determines the weight distributions of a class of cyclic codes over @xmath29 with nonzeros @xmath30 , and they are verified by using matlab .
note that the length of the cyclic code is @xmath31 .
final conclusion is in section [ seciv ] .
this paper is the counterpart of our another result in @xcite .
in this section , relevant knowledge from finite fields @xcite is presented for our study of cyclic codes .
it is about the calculations of exponential sums , the sizes of cyclotomic cosets and the ranks of certain quadratic forms .
first , some known properties about the codeword weight are listed .
then lemma [ fl002 ] , lemma [ es02 ] and lemma [ r001 ] are about the calculations of exponential sums
. finally , lemma [ gcd002 ] , lemma [ rq002 ] and corollary [ rq02 ] are about the ranks of relevant quadratic forms .
let @xmath0 be an odd prime , @xmath26 be a positive integer and @xmath27 is a primitive element of @xmath28 .
assume the cyclic code @xmath6 over @xmath7 has length @xmath32 and non - conjugate nonzeros @xmath33 , where @xmath34 .
then the codewords in @xmath6 can be expressed by @xmath35 where @xmath36 and @xmath37 is the trace mapping from @xmath28 to @xmath7 .
therefore the hamming weight of the codeword @xmath38 is : @xmath39 where @xmath40 ( @xmath22 is imaginary unit ) , @xmath41 , \mathbb{f}_q^*=\mathbb{f}_q\backslash \{0\}$ ] , @xmath42 and @xmath43 . for general functions of the form @xmath44
where @xmath45 , there are quadratic forms @xmath46 and corresponding symmetric matrices @xmath47 satisfying that @xmath48 ( lemma 1 , @xcite)[fl002 ] 1 . for the quadratic form @xmath49 , @xmath50 2 . for @xmath51 ,
if @xmath52 has solution @xmath53 , then @xmath54 where @xmath55 .
otherwise @xmath56 .
here @xmath57 denotes the legendre symbol .
lemma [ es02 ] is from @xcite , lemma [ r001 ] is from @xcite , also refer to @xcite for the calculations of exponential sums that will be needed in the sequel .
[ es02 ] for the quadratic form @xmath58 corresponding to @xmath59 , see ( [ qf01 ] ) 1 .
if the rank @xmath60 of the symmetric matrix @xmath61 is even , which means that @xmath62 , then @xmath63 2 . if the rank @xmath60 of the symmetric matrix @xmath61 is odd , which means that @xmath64 , then @xmath65 where @xmath66 and @xmath67 .
[ r001 ] let @xmath58 be the quadratic form corresponding to @xmath59 , see ( [ qf01 ] ) .
if the rank @xmath60 of the symmetric matrix @xmath61 is odd , then the number of quadratic forms with exponential sum @xmath68 equals the number of quadratic forms with exponential sum @xmath69 where @xmath67 .
the cyclotomic coset containing @xmath70 is defined to be @xmath71 where @xmath72 is the smallest positive integer such that @xmath73 . in the following ,
lemma [ gcd002 ] and lemma [ rq002 ] are from @xcite , also refer to @xcite for the binary case of lemma [ gcd002 ] .
[ gcd002 ] if @xmath74 is odd , then for @xmath75 , the cyclotomic coset @xmath76 has size @xmath77 if @xmath78 is even , then for @xmath75 , the cyclotomic coset @xmath76 has size @xmath79 for @xmath80 with corresponding quadratic form @xmath81 where @xmath82 , the following result is about its rank .
[ rq002 ] let @xmath26 be a positive integer , @xmath83 .
the rank @xmath84 of the symmetric matrix @xmath85 satisfies @xmath86 .
the following corollary is a special case of lemma [ rq002 ] .
[ rq02 ] the rank @xmath87 of the symmetric matrix @xmath88 corresponding to @xmath89 has five possible values : @xmath90 * note that in section [ seciii ] , a nonzero solution of a equation system means that all the variable values are nonzero .
in this section , the main results of this paper are obtained , that is the weight distribution of the cyclic code @xmath91 with nonzeros @xmath92 and @xmath93 for the case when @xmath26 is even , here @xmath24 . for this , the first five moments of exponential sum @xmath94 are computed in subsections [ sec3.1 ] , [ sec3.2 ] , [ sec3.3 ] , and the macwlliams identities are calculated in subsection [ sec3.4 ] . for an odd prime @xmath0 and even integer @xmath26 ,
this subsection calculates the first three moments of the exponential sum @xmath94 ( equation ( [ es00001 ] ) ) , see lemma [ esl001 ] and its another form lemma [ es002 ] , where the analysis of the third moment bases on the property of lemma [ esl002 ] .
( theorem 6.26 . ,
@xcite ) [ se01 ] let @xmath95 be a nondegenerate quadratic form over @xmath28 , @xmath96 odd , in an even number @xmath13 of indeterminates . then for @xmath97 the number of solutions of the equation @xmath98 in @xmath99 is @xmath100 where @xmath101 for @xmath102 , @xmath103 , @xmath104 is the quadratic character of @xmath28 and @xmath105 .
[ esl002 ] let @xmath0 be an odd prime , @xmath25 and @xmath106 .
then the solutions of the following equation in @xmath107 @xmath108 have the form @xmath109 where @xmath110 and @xmath111 .
first , it can be checked that the pairs @xmath112 given by the lemma satisfy equation ( [ se001 ] ) .
second , according to lemma [ se01 ] , the number of solutions of equation ( [ se001 ] ) in @xmath113 is @xmath114 since @xmath115 is a quadratic residue of @xmath116 .
furthermore , when @xmath117 varies through the nonzero elements of @xmath118 , @xmath119 in ( [ se002 ] ) gives all the solutions of ( [ se001 ] ) in @xmath113 including those solutions in the subfield @xmath120 .
in fact , @xmath121 and @xmath122 imply @xmath123 .
so for @xmath124 , @xmath125 . [ esl001 ]
let @xmath0 be an odd prime satisfying @xmath126 and @xmath25 .
then there are the following results about the exponential sum @xmath94 ( equation ( [ es00001 ] ) ) corresponding to @xmath127 1 .
@xmath128 2 .
@xmath129 3 .
@xmath130 . from definition , changing the order of summations , ( i )
can be calculated as follows @xmath131 equation ( ii ) can also be calculated in this way @xmath132 where @xmath133 is the number of solutions to the equation system @xmath134 if @xmath135 satisfy the above system , then @xmath136 and @xmath137 . since @xmath126 , that is @xmath138 , we have @xmath139 , a contradiction .
so , the only solution to above system is @xmath140 and @xmath141 . as to ( iii )
, we have @xmath142 where @xmath143 and @xmath144 is the number of solutions of @xmath145 to study equation system ( [ es01 ] ) , consider the last two equations .
canceling @xmath146 there is @xmath147 after simplification , it becomes @xmath148 from ( [ es0002 ] ) it can be checked that @xmath149 . in the same way , it implies that @xmath150 . in this case , @xmath151 , so only the first two equations of ( [ es01 ] ) are necessary to be considered .
the following paragraph of proof is similar to what has been done in @xcite .
for the first one of equation ( [ es01 ] ) , by lemma [ esl002 ] there exist @xmath152 such that @xmath153 and @xmath154 where @xmath155 .
substituting to the second equation , @xmath156 since @xmath138 , @xmath157 .
after simplification , the above equation becomes @xmath158 .
set @xmath159 , we have @xmath160 , that is @xmath161 now , since @xmath149 , @xmath162 that is @xmath163 where @xmath164 . also , because @xmath150 , we can obtain @xmath165 .
combining the two results , we find that @xmath166 if @xmath167 and @xmath168 satisfy ( [ es000201 ] ) and ( [ es000202 ] ) , we have @xmath169 . since @xmath170 , there exist integers @xmath171 such that @xmath172 that is equivalent to @xmath173 .
it is easy to check that @xmath174 is a factor of @xmath175 , so if there exist solutions , the number of solutions is @xmath174 .
it can be checked that there exist solutions in @xmath176 of equation system ( [ es01 ] ) .
finally , @xmath177 and @xmath178 .
using the following notations , lemma [ esl001 ] can be restated in lemma [ es002 ] when @xmath26 is even .
corresponding to lemma [ fl002 ] and corollary [ rq02 ] , we introduce the following notations for convenience .
let @xmath179 where @xmath180 and @xmath181 . also , denote @xmath182 for @xmath181 . and @xmath183 for @xmath184 , where @xmath22 is the imaginary unit .
by lemma [ r001 ] , set @xmath185 for @xmath184 , since @xmath186 is odd .
[ es002 ] let @xmath0 be an odd prime satisfying @xmath126 , and @xmath25 where @xmath26 is an even integer .
then the notations defined in ( [ 0001 ] ) and ( [ 0002 ] ) satisfy the following equations @xmath187 substituting the symbols of ( [ 0001 ] ) and ( [ 0002 ] ) to lemma [ esl001 ] , we have the following four equations @xmath188 @xmath189 where the first one comes from the fact that there are @xmath190 elements in the set @xmath191 . also , note that @xmath192 when @xmath193 .
using @xmath194 for @xmath184 , the result is obtained by simplification . for the fourth moment of @xmath94 in the particular case of @xmath24 , there is the following result about the number of solutions of the equation system @xmath195 in lemma [ cm001 ] , which is denoted by @xmath196 .
( @xcite)[cm001 ] let @xmath24 and @xmath25 , then @xmath197 using lemma [ cm001 ] and @xmath198 in lemma [ esl001 ] , @xmath199 is calculated in lemma [ cm01 ] where @xmath200
. [ cm01 ] let @xmath24 and @xmath25 .
the number of solutions of the following equation system @xmath201 is @xmath202 .
corresponding to lemma [ esl001 ] , the result of the fourth moment is provided in lemma [ s04 ] by applying lemma [ cm01 ] .
[ s04 ] let @xmath24 and @xmath25 .
then @xmath203 corresponding to lemma [ es002 ] , lemma [ s04 ] can be rewritten as the following corollary using the symbols of ( [ 0001 ] ) and ( [ 0002 ] ) .
[ c02 ] let @xmath24 and @xmath25 where @xmath26 is an even integer .
then @xmath204 for the fifth moment of @xmath94 , we need magma @xcite to find the number of solutions of the following equation system @xmath205 which is denoted by @xmath206 . as in @xcite ,
the irreducible components corresponding to the projective variety defined by ( [ e03 ] ) are listed in table [ tab02 ] using magma .
it is easy to be verified that every block of table [ tab02 ] contains a system of three equations ( note that ` @xmath207 ' is omitted ) , the solutions of which satisfy ( [ e03 ] ) .
furthermore , the union of all the solutions in each block presents the solutions of ( [ e03 ] ) exactly .
those equation systems are circulant symmetric about the variables . in general , few works were provided to deal with the moments using five variables .
but in this paper , magma helps us on the reduction of such systems in lemma [ cm02 ] , lemma [ s004 ] and corollary [ c03 ] .
for relevant knowledge of algebraic geometry , the reader is referred to @xcite . [ cols= " < , < , < , < , < , < , < , < " , ] + by which only the cases @xmath208 and @xmath209 are the possible solutions which can correspond to codewords since coordinates should be different . and
the number of such solutions is @xmath210 which corresponds to @xmath211 codewords , since every four - tuple @xmath212 corresponds to @xmath213 solutions of ( [ e003 ] ) .
in fact , if @xmath214 is a weight - four codeword with nonzero positions and values @xmath215 , then @xmath216 , @xmath217 and @xmath218 can all represent weight - four codewords . combing the above three cases , @xmath219 the result of the lemma
is obtained .
[ sm02 ] let @xmath220 where @xmath221 is an even integer .
the notations defined in equations ( [ 0001 ] ) and ( [ 0002 ] ) satisfy the following equations @xmath222 where @xmath223 @xmath224 and @xmath225 are defined in lemma [ w001 ] .
define the following notations for simplification @xmath226 in addition , the usage of macwilliams identities in the following paragraph , implies the condition @xmath221 , refer to lemma [ gcd002 ] . by equation ( [ c01 ] ) and lemma [ es02 ] , @xmath6 has seven possible nonzero weights @xmath227 with the above notations , the first four moments of codeword weight can be computed @xmath228 @xmath229 @xmath230 @xmath231 from equations ( [ sm001 ] ) , ( [ fmi001 ] ) and ( [ smi001 ] ) we have @xmath232 and the first one of equation ( [ sm004 ] ) is obtained after simplification . according to the fourth moment of macwilliams moment identity ( [ sm003 ] ) and equations ( [ fmi001 ] ) , ( [ smi001 ] ) , ( [ tmi001 ] ) and
( [ fmi002 ] ) , @xmath233 and the second one of equation ( [ sm004 ] ) is obtained after simplification . in this subsection ,
the parameters defined in equations ( [ 0001 ] ) and ( [ 0002 ] ) are calculated in lemma [ sm03 ] , and the weight distribution of the cyclic code @xmath6 is determined in theorem [ w02 ] .
[ sm03 ] let @xmath220 where @xmath221 is an even integer .
the notations defined in equations ( [ 0001 ] ) and ( [ 0002 ] ) satisfy the following equations @xmath234 where @xmath235 and @xmath236 are defined in lemma [ sm02 ] which needs @xmath221 . from lemma [ es002 ] , corollary [ c02 ] , corollary [ c03 ] and lemma [ sm02 ] , the notations satisfy the following equations @xmath237 solving the above equation system , the result is obtained .
let @xmath220 where @xmath221 is an even integer .
the number of solutions of the equation system @xmath238 is @xmath239 the following moment of exponential sum @xmath94 satisfies @xmath240 where @xmath241 is the number of solutions of ( [ m05 ] ) . solving the above equation for @xmath241 , the result is obtained . equation ( [ m05 ] ) considers the case for @xmath242 variables . using the seventh moment of @xmath94
, the number of solutions can be calculated when there are @xmath243 variables , etc .
[ w02 ] let @xmath220 where @xmath221 is an even integer .
the cyclic code @xmath6 with nonzeros @xmath244 and @xmath245 has seven nonzero weights , @xmath246 where @xmath27 is a primitive element of the finite field @xmath28 .
it is interesting to note about the weights of the cyclic code @xmath6 .
if there is a weight of the form @xmath247 , then there is a weight of the form @xmath248 .
so , it seemed as if the weights are symmetric about the value @xmath249 which is also a weight of @xmath6 .
as the following example illustrates , in general the higher the value @xmath22 the less number of corresponding weights .
this phenomenon may be explained to the fact that the linear part in the exponential position of the parameter @xmath250 acts as a center role in the formation of the weights .
let @xmath220 where @xmath251 .
the cyclic code @xmath6 has nonzeros @xmath244 and @xmath245 where @xmath27 is a primitive element of the finite field @xmath28 .
using matlab , it can be found that it has seven nonzero weights @xmath252 which is verified by using matlab .
since the weight distributions played an important role in the application of cyclic codes , this paper focuses on the determination of a class of cyclic codes with three nonzeros .
relevant results received a lot of attention by using methods with lower moments of exponential sum . here
we try to apply higher moments to deal with the problem .
a. johansen and t. helleseth , `` a family of @xmath26-sequences with five valued cross correlation , '' _ ieee trans .
inf . theory _ ,
2 , pp . 880887 , feb .
a. johansen , t. helleseth , and a. kholosha , `` further results on m - sequences with five - valued cross correlation , '' _ ieee trans .
theory , _ vol .
57925802 , dec . 2009 .
t. kasami , `` a decoding procedure for multiple - error - correcting cyclic codes , '' _ ieee trans .
inf . theory _ , vol .
2 , pp . 134138 , apri .
1964 .
x. liu and y. luo , `` the weight distributions of some cyclic codes with three or four nonzeros over @xmath29 , '' arxiv:1302.0394v1 . j. luo , y. tang , and h. wang , `` cyclic codes and sequences : the generalized kasami case , '' _ ieee trans .
inf . theory _ ,
21302142 , may . 2010 .
r. j. mceliece , `` irreducible cyclic codes and gauss sums , '' in:_combinatorics , part i : theory of designs , finite geometry and coding theory_. math .
centre tracts , amsterdam : math .
centrum , no .
55 , pp . 179196 , 1974 .
s. g. s. shiva , k. c. fung , and h. s. y. tan , `` on permutation decoding of binary cyclic double - error - correcting codes of certain lengths ( corresp . ) , '' _ ieee trans .
inf . theory _ ,
5 , pp . 641643 , 1970 . | cyclic codes have efficient encoding and decoding algorithms .
the decoding error probability and the undetected error probability are usually bounded by or given from the weight distributions of the codes .
most researches are about the determination of the weight distributions of cyclic codes with few nonzeros , by using quadratic form and exponential sum but limited to low moments . in this paper , we focus on the application of higher moments of the exponential sum to determine the weight distributions of a class of ternary cyclic codes with three nonzeros , combining with not only quadratic form but also macwilliams identities .
another application of this paper is to emphasize the computer algebra system magma for the investigation of the higher moments . in the end ,
the result is verified by one example using matlab .
submitted paper cyclic code , exponential sum , macwilliams identities , quadratic form , weight distribution |
the kerr spacetime is one of the most important exact solutions to the einstein vacuum equations .
its relevance stems from the uniqueness theorems for black holes which state that under rather general conditions the kerr spacetime is the only asymptotically flat , stationary , vacuum black hole solution see e.g. the introduction of @xcite for a critical review of the issue of black hole uniqueness and the involved assumptions .
although , the kerr spacetime is very well understood from a spacetime perspective , the same can not be said if one adopts a 3 + 1 point of view which would be the case if one tries to numerically calculate the spacetime from some cauchy initial data .
as soon as one moves away from a 3 + 1 gauge which is adapted to the stationary and axial symmetries which can occur in some applications , in particular in numerical ones an analysis of the kerr spacetime and initial data sets thereof becomes very complicated .
the explicit nature of the kerr solution makes it tempting to perform detailed calculations in order to , say , verify a particular property of the spacetime .
this approach usually leads to very lengthy expressions which can be very hard to analyse .
in a sense one could say that exact solutions contain too much information .
in this case it can be more convenient to adopt more abstract approaches and , accordingly , it may prove useful to have at hand a characterization of kerr initial data .
the question of providing an invariant characterization of initial data sets for the schwarzschild spacetime has been addressed in @xcite . in particular , the analysis of @xcite provides an algorithmic characterization of schwarzschild data .
that is , a procedure is provided to verify whether a given initial data set for the einstein field equations will render a development which is isometric to a portion of the schwarzschild spacetime .
one of the most important algebraic properties of the kerr spacetime is that its weyl tensor is of petrov type d. the close relation between vacuum spacetimes with a weyl tensor of petrov type d and killing spinors has been exploited in @xcite to provide a characterization of initial data sets whose developments will be of petrov type d. this characterization relies on one being able to decide whether a set of overdetermined partial differential equations has solutions for a given initial data set .
accordingly , such a characterization is not algorithmic .
although not explicitly stated in @xcite , from that analysis it should be possible to obtain a characterization of kerr initial data by adding some global conditions .
the characterization of initial data sets discussed in @xcite has followed the general strategy of starting from a given tensorial ( respectively , spinorial ) _ spacetime _ characterization of the class of spacetimes under consideration .
necessary conditions on the initial data set are obtained by performing a 3 + 1 decomposition of the spacetime characterization .
given a set of necessary conditions on the initial data , it is then natural to address the question of sufficiency
. this is , usually , the most challenging part of the procedure as one has to discuss the evolution of complicated tensorial objects .
the idea behind this is to show that if the necessary conditions are satisfied on some open subset of the initial hypersurface , then one can possibly , under some additional assumptions recover the spacetime characterization on the development of the open subset on the initial hypersurface from which one started . in this article a particular characterization of kerr initial data is addressed .
our starting point is a certain spacetime characterization of the kerr spacetime given in @xcite .
this characterization was developed with the aim of providing an alternative way of proving the uniqueness of kerr spacetime among the class of stationary , asymptotically flat black holes .
this expectation has been recently fulfilled in @xcite , where a proof of the uniqueness of kerr which does not make assumptions on the analyticity of the metric has been given . at the heart of the spacetime characterization given in @xcite cfr .
theorem [ kerr - characterization] is a certain tensor , _ the mars - simon tensor _ , whose construction requires the existence of a timelike killing vector .
the mars - simon tensor is a spacetime version of the _ simon tensor _ , a tensor , defined in the quotient manifold of a stationary spacetime , which characterizes the kerr spacetime see @xcite .
following the general strategy for the construction of characterizations of initial data sets out of spacetime characterizations , necessary conditions for kerr initial data are deduced from a 3 + 1 splitting of the mars - simon tensor .
accordingly , one assumes that the spacetime one is working with has a timelike killing vector . this requirement can be encoded in the initial data by requiring that the data has a suitable _ killing initial data _ ( kid ) .
the mars - simon tensor has the same symmetries as the weyl tensor , and hence its 3 + 1 decomposition can be given in terms of its _ electric _ and _ magnetic _ parts . in order to discuss the propagation of the mars - simon tensor we make use of a framework for discussing the propagation tensorial fields using _ superenergy - type _
tensors this framework has been discussed in e.g. @xcite .
it should be pointed out that the characterization discussed in this article is not algorithmic .
that is , like the one for type d initial data discussed in @xcite it depends on being able to decide whether a certain overdetermined system of partial differential equations admits a solution .
the article is structured as follows : in section [ preliminaries ] our main conventions are fixed and relevant aspects of the 3 + 1 formalism are discussed .
section [ weyl : c ] discusses the properties and causal propagation of weyl candidates i.e .
tensors with the same symmetries of the weyl tensor .
section [ section : mstensor ] is concerned with the properties of the mars - simon tensor .
section [ cp - ms ] discusses the causal propagation of the mars - simon tensor .
section [ section : application ] applies the previous discussion to the construction of a characterization of kerr initial data .
our main result is provided in theorem [ theorem : kerr_data ] . in section [ schwarzschild ]
we particularize to the case of the schwarzschild spacetime , where , with the aim of the results of @xcite it is possible to obtain an initial data characterization which is algorithmic .
some concluding remarks are provided in section [ conclusions ] .
finally , some technical details , too lengthy to be presented in the main text are presented in the appendix a. all the tensor computations of this paper have been performed with the software _
_ xact _ is a suite of mathematica packages which has among its many features the capability to efficiently canonicalize tensor expressions by the use of powerful algorithms based on permutation group theory @xcite .
currently no other software , either free or commercial , is capable to handle the tensor computations needed in this paper .
let @xmath0 denote a smooth orientable spacetime . the following conventions will be used : plain greek letters @xmath1 denote abstract indices and boldface latin characters @xmath2 will be used for component indices .
the signature of the metric tensor @xmath3 will be taken to be @xmath4 , while @xmath5 , @xmath6 and @xmath7 denote , respectively the riemann , ricci and weyl tensors of @xmath3 .
the tensor @xmath8 is the volume element which is used to define the hodge dual of any antisymmetric tensor denoted by attaching a star @xmath9 to the tensor symbol .
sometimes we will need to work with complex tensors in which case the complex conjugation of a tensor is denoted by an overbar . the operator @xmath10 symbolizes the lie derivative with respect to the vector field @xmath11 .
let @xmath12 be a unit timelike vector , @xmath13 defined on @xmath14 .
then any tensor or tensorial expression can be decomposed with respect to @xmath12 and the way to achieve it is the essence of the orthogonal splitting ( also known as 3 + 1 formalism ) which is described in many places of the literature see e.g. @xcite ) .
we review the parts of this formalism needed in this work .
the spatial metric is defined by @xmath15 and it has the algebraic properties @xmath16 , @xmath17 .
we shall call a covariant tensor @xmath18 _ spatial _ with respect to @xmath19 if it is invariant under @xmath20 i.e. if @xmath21 with the obvious generalization for any mixed tensor .
this property implies that the inner contraction of @xmath12 with @xmath22 ( taken on any index ) vanishes .
the orthogonal splitting of a tensor expression consists in writing it as a sum of terms which are tensor products of the unit normal and spatial tensors of lesser degree or the same degree in which case the unit normal is absent . in order to find the orthogonal splitting of expressions containing covariant derivatives we need to introduce
the _ spatial derivative _ @xmath23 which is an operator whose action on any tensor field @xmath24 , @xmath25 is given by @xmath26 from equation ( [ spatialcd ] ) it is clear that @xmath27 is spatial .
the results just described hold for an arbitrary unit timelike vector @xmath12 but in our framework we only need to consider integrable timelike vectors which are characterized by the condition @xmath28}=0 $ ] ( frobenius condition ) . in this case
there exists a foliation of @xmath14 such that the vector field @xmath12 is orthogonal to the leaves of the foliation .
we shall denote by @xmath29 , @xmath30 , the family of leaves of this foliation and @xmath31 is called the _ initial data hypersurface _ it is assumed that @xmath32 .
the tensor @xmath19 plays the role of the _ first fundamental form _ for any of the leaves while the symmetric tensor @xmath33 defined by @xmath34 can be identified with the _ second fundamental form_. combining the previous definition with the frobenius condition we easily derive @xmath35 where @xmath36 is the acceleration of @xmath12 . by using these quantities
one is in principle able to work out the orthogonal splitting of any tensorial expression .
we supply below the explicit results of calculations which will be needed repeatedly in the sequel .
* orthogonal splitting of the volume element : @xmath37 here @xmath38 is the _ spatial volume element _ which is a fully antisymmetric spatial tensor .
* orthogonal splitting of an antisymmetric tensor .
if @xmath39 is an arbitrary antisymmetric tensor then its orthogonal splitting takes the form @xmath40}-q_\rho\varepsilon_{\mu\nu}^{\phantom{\mu\nu}\rho},\ ] ] where @xmath41 , @xmath42 are clearly spatial tensors .
it is straightforward from the previous equation to find an analogous formula for the orthogonal splitting of @xmath43 .
if we work with the _ self - dual _ part of @xmath39 , denoted by @xmath44 and defined by @xmath45 then its orthogonal splitting takes the form
@xmath46},\ \mathcal{p}_\mu\equiv p_\mu+\mbox{i}\ q_\mu . \label{decompose_h}\ ] ] * orthogonal splitting of the weyl tensor : @xmath47\nu}- l_{\nu[\lambda}e_{\rho]\mu}- n_{[\lambda}b_{\rho]\tau}\epsilon^\tau_{\phantom{\tau}\mu\nu } -n_{[\mu}b_{\nu]\tau}\epsilon^\tau_{\phantom{\tau}\lambda\rho } \right),\label{weyl-1}\ ] ] where @xmath48 denote the weyl tensor _ electric _ and _ magnetic _ parts respectively , and @xmath49 .
the tensors @xmath50 and @xmath51 are symmetric , traceless , and spatial .
equation ( [ weyl-1 ] ) also holds for any rank-4 tensor possessing the same algebraic properties as the weyl tensor .
such tensors will be designated collectively as _ weyl candidates _
see section [ weyl : c ]
. we can also choose to work with the self - dual weyl tensor @xmath52 which is given by @xmath53 using ( [ weyl-1 ] ) and the analogous formula for the orthogonal splitting of @xmath54 one can obtain an expression for the orthogonal splitting of @xmath52 @xmath55\mu}+ l_{\mu[\rho}\mathcal{e}_{\lambda]\nu}+\mbox{i}\
\varepsilon_{\rho\lambda\alpha}n_{[\mu}\mathcal{e}_{\nu]}^{\phantom{\nu]}\alpha}+ \mbox{i}\ \varepsilon_{\mu\nu\alpha}n_{[\rho}\mathcal{e}_{\lambda]}^{\phantom{\nu]}\alpha } ) , \label{decompose - weyl}\ ] ] where @xmath56 again a similar formula to ( [ decompose - weyl ] ) holds for any weyl candidate .
* let @xmath57 be any spatial vector .
then the orthogonal splitting of @xmath58 is @xmath59 note that since @xmath60 is spatial then @xmath61 is also spatial .
we briefly review the standard formulation of the cauchy problem for the vacuum einstein equations . in this formulation one
considers a 3-dimensional connected riemannian manifold @xmath62 we use small plain latin letters @xmath63 for the abstract indices of tensors on this manifold and an isometric embedding @xmath64 .
the map @xmath65 is an isometric embedding if @xmath66 where @xmath67 realises the pullback , @xmath68 , of tensor fields from @xmath14 to @xmath69 .
the metric @xmath70 defines a unique affine connection @xmath71 ( levi - civita connection ) by means of the standard condition @xmath72 the riemann tensor of @xmath71 is @xmath73 and from it we define its ricci tensor by @xmath74 and its scalar curvature @xmath75 in @xmath69 indices are raised and lowered with @xmath70 and its inverse @xmath76 .
let @xmath62 be a riemannian manifold and suppose that there exists a symmetric tensor field @xmath77 on it which verifies the conditions ( vacuum constraints ) @xmath78 where @xmath79 .
provided that @xmath70 and @xmath77 are suitably smooth there exists an isometric embedding @xmath65 of @xmath69 into a globally hyperbolic , vacuum solution @xmath80 of the einstein field equations .
the set @xmath81 is then called a vacuum initial data set and the spacetime @xmath80 is the data development .
furthermore the spacelike hypersurface @xmath82 is a cauchy hypersurface in @xmath83 .
[ vacuum - data ] a statement of this theorem containing precise regularity conditions on @xmath70 , @xmath77 is formulated in reference @xcite . under the conditions of the theorem [ vacuum - data ] , we may construct a foliation of @xmath83 with @xmath12 as the timelike unit vector which is orthogonal to the leaves and , in the above notation , set @xmath84 .
it is then clear that @xmath85 .
other key properties are @xmath86 using these properties and the ricci identity for @xmath12 one can show that @xmath87 another property which is needed later on is the following one : for any spatial tensor @xmath88 , @xmath89 with respect to @xmath12 we have the property @xmath90
fundamental for our discussion will be tensors of rank 4 having the same algebraic properties as the weyl tensor .
more precisely , one defines a _ weyl candidate _ as any rank 4 tensor @xmath91 fulfilling the properties @xmath92\gamma\delta}=c_{\alpha\beta\gamma\delta}= c_{\gamma\delta\alpha\beta } , \quad c^{\alpha}_{\phantom{\alpha}\alpha\gamma\delta}=0 , \quad c_{\alpha[\beta\gamma\delta]}=0 . \label{weyl_candidatep}\ ] ] from ( [ weyl_candidatep ] ) we easily deduce that @xmath93 and indeed @xmath94 is also a weyl candidate . given a weyl candidate @xmath91 , its _ weyl current _
@xmath95 is defined by @xmath96 the tensor @xmath97 is trace - free , antisymmetric in the last pair of indices and it has the property @xmath98}=0 $ ] . from ( [ weyl_current ] )
we deduce the following identities @xmath99\mu\nu}= -\frac{1}{3}\eta_{\alpha\beta\gamma\sigma}j^{*\sigma}_{\phantom{*\sigma}\mu\nu},\quad \nabla_{\beta}j^{\beta}_{\phantom{\beta}\gamma\delta}= r_{[\gamma}^{\phantom{[\gamma}\sigma\rho\mu}c_{\delta]\sigma\rho\mu } , \label{weyl_bianchi}\ ] ] where @xmath100 . equations ( [ weyl_current])-([weyl_bianchi ] ) have counterparts involving @xmath101 which we will not discuss . also , if we combine ( [ weyl_current ] ) and ( [ weyl_bianchi ] ) we get a wave equation for the weyl candidate : @xmath102\delta\beta\alpha}-2r_{\chi\phantom{\alpha}\rho}^{\phantom{\chi}\alpha \phantom{\rho}\delta}c_{\beta\gamma\alpha\delta}+ 2r_{[\rho}^{\phantom{[\rho}\alpha}c_{\chi]\alpha\beta\gamma}+ 2r_{\beta\phantom{\alpha}[\rho}^{\phantom{\beta}\alpha\phantom{[\rho}\delta } c_{\chi]\delta\gamma\alpha } \nonumber\\ \fl \hspace{3cm}+\eta_{\chi\rho\lambda\pi}\nabla^{\pi } j^{*\lambda}_{\phantom{*\lambda}\beta\gamma}+2\nabla_{[\chi}j_{\rho]\beta\gamma}. \label{box_weyl}\end{aligned}\ ] ] the previous considerations hold regardless of whether the weyl candidate is real or _ complex_.
indeed , from any real weyl candidate @xmath103 we may construct a complex weyl candidate @xmath104 by means of @xmath105 a weyl candidate constructed in such a way is _ self - dual _ is given the name of `` self - dual '' . ]
@xmath106 reciprocally , if a complex weyl candidate @xmath104 fulfils ( [ self - dual ] ) then its real part @xmath107 and its imaginary part @xmath108 are also weyl candidates related via @xmath109 besides the basic algebraic properties of a weyl candidate just explained we need to introduce a further concept which will play a crucial role in the sequel .
[ causal - prop ] let @xmath0 be a spacetime and consider a weyl candidate @xmath110 defined in the whole of @xmath14 .
we say that the weyl candidate @xmath110 propagates causally on @xmath14 if for any embedded spacelike hypersurface @xmath111 the condition @xmath112 implies @xmath113 on @xmath114 .
given a set @xmath115 , a point @xmath116 belongs to the future cauchy development @xmath117 if any past - inextendible causal curve containing @xmath118 intersects @xmath115 .
there is a similar notion of @xmath119 and the set @xmath120 is called the _ cauchy development _ of @xmath115 .
it is well - known that the interior of @xmath114 is globally hyperbolic and since @xmath115 is an embedded spacelike hypersurface then it will be a cauchy hypersurface for @xmath114 see e.g. @xcite for an elementary introduction to these concepts of causality theory . the simplest example of a weyl candidate which propagates causally is the weyl tensor @xmath7 of a vacuum spacetime see e.g. @xcite .
it is possible to address the causal propagation of a weyl candidate by following the general techniques given in @xcite see also @xcite in which the causal propagation of any tensor is analysed .
these ideas are reviewed next .
an essential object to study the causal propagation of a weyl candidate is its _ bel - robinson tensor_. given a real weyl candidate @xmath91 , we define its bel - robinson tensor by @xmath121 the previous definition has been described in a more general framework in @xcite as the _ superenergy tensor _ of the weyl candidate @xmath91 . in our particular context we prefer to call this tensor the bel - robinson tensor of @xmath91 in analogy with the bel - robinson tensor constructed outof the weyl tensor @xcite .
the bel - robinson tensor of a weyl candidate has the following properties see @xcite for detailed proofs of points ( [ first ] ) , ( [ second ] ) and ( [ third ] ) ; point ( [ fourth ] ) follows by taking the covariant divergence of ( [ b - r ] ) and then using repeatedly ( [ weyl_current ] ) and ( [ weyl_bianchi ] ) . [ sproperties ] if @xmath122 is the bel - robinson tensor of the weyl candidate @xmath110 then 1 .
[ first ] @xmath123 , @xmath124 .
[ second]generalized dominant property : if @xmath125 , @xmath126 , @xmath127 , @xmath128 are causal future - directed vectors then @xmath129 .
this property admits an alternative formulation : let @xmath130 be any orthonormal frame with @xmath131 the timelike vector .
then the following inequality holds @xmath132 where the component indices refer to the frame @xmath133 .
[ third ] @xmath134 a timelike vector @xmath135 such that @xmath136 .
[ fourth]the covariant divergence of @xmath122 is given by @xmath137 next , we explain how the properties summarized in theorem [ sproperties ] are useful in the study of the causal propagation of @xmath138 .
we follow closely @xcite in our exposition .
let @xmath117 be the future cauchy development of a closed achronal hypersurface @xmath115 and pick up any event @xmath139 .
furthermore , let @xmath140 denote the causal past of the event @xmath141 .
the set @xmath142 is compact and it contains the set @xmath143 .
the boundary of @xmath144 is @xmath145 where @xmath146 is the future cauchy horizon of @xmath147 .
also , since the interior of @xmath117 is globally hyperbolic we deduce that it can be foliated by a family of spacelike hypersurfaces @xmath29 , @xmath148 , where @xmath149 , @xmath150 , @xmath151 and @xmath152 .
the figure [ figure1 ] shows a schematic view of this geometric construction .
[ figure1 ] , @xmath147 , @xmath117 and @xmath146 see the main text for further details . ]
let @xmath153 be the timelike normal 1-form to the leaves of @xmath154 and define the quantity @xmath155 here , @xmath156 is the positive measure on the spacetime @xmath14 constructed from the volume form @xmath157 ( any measure defined from a volume form will be understood as positive ) and @xmath158 is the measure on the hypersurface @xmath159 obtained from its volume form @xmath160 . the scalar @xmath161 is called the _ superenergy density _ and is everywhere non - negative see point ( [ second ] ) of theorem [ sproperties] which entails @xmath162 .
also , from the definition of @xmath163 we get @xmath164 where @xmath165 is defined by @xmath166 again point ( [ second ] ) of theorem [ sproperties ] implies that @xmath167 is also a non - negative quantity and hence @xmath168 .
now , according to the gauss theorem we have @xmath169 where @xmath170 and @xmath171 denote the measures on the hypersurfaces @xmath147 and @xmath146 respectively induced by the volume forms @xmath172 , and @xmath173 . in this last expression @xmath174
is a causal vector which is orthogonal to the null hypersurface @xmath146 , points outward to @xmath144 ( therefore it is future - directed ) and is such that @xmath175 .
note that these three conditions do not fix the vector @xmath174 univocally . by combining ( [ dwdt ] ) and ( [ gauss ] ) we deduce @xmath176 now , we need to estimate the right hand side of ( [ estimate1 ] ) . according to point ( [ second ] ) of theorem [ sproperties ] , the quantity @xmath177 is positive and therefore @xmath178 on the other hand , if @xmath110 vanishes on @xmath115 then @xmath179 and @xmath180 next , we set up an orthonormal frame @xmath130 with @xmath181 and we consider the set of components @xmath182 in such a frame . since these are continuous functions
we deduce that in the compact set @xmath144 the estimate @xmath183 , @xmath184 holds for some constant @xmath185 .
hence @xmath186 where in the last step , point ( [ second ] ) of theorem [ sproperties ] was used . from here , we deduce @xmath187 using the above estimates in ( [ estimate1 ] ) we get @xmath188 now it only remains to estimate @xmath189 and to that end ( [ covdivbr ] ) is used .
the actual estimate will depend on the particular expression of the weyl current .
for example , the simplest case arises when @xmath190 where no further estimates are needed and grnwall s lemma entails @xmath191 which in turn implies that the scalar @xmath192 vanishes in @xmath144 . from the arbitrariness of @xmath141 we deduce that the superenergy density is zero in @xmath117 and by a similar argument as the above one can show that this is also the case for @xmath119 .
hence point 3 of theorem [ sproperties ] implies that @xmath113 on @xmath193 .
consequently the tensor @xmath110 propagates causally whenever the weyl current is zero .
_ point ( [ third ] ) of theorem [ sproperties ] entails @xmath112 if and only if @xmath194 .
accordingly , the causal propagation of a weyl candidate can also be formulated in terms of the vanishing of the superenergy density at the hypersurface @xmath115 .
[ remark : se_density ] _
we start this section by recalling some basic background about vacuum spacetimes possessing a killing vector see e.g. @xcite for a complete account .
let @xmath0 be a vacuum solution of einstein field equations and let @xmath195 be a killing vector , @xmath196 .
this condition enables us to define a 2-form , the _ killing form _ , by @xmath197 } = \nabla_\mu\xi_\nu . \label{killing-2form}\ ] ] elementary manipulations lead to @xmath198 the _ self - dual killing form _
@xmath199 is then @xmath200 with @xmath201 the hodge dual of @xmath202 . from the previous definition it is straightforward to obtain the property ( self - duality ) @xmath203 we shall write @xmath204 .
the killing form is said to be degenerate if @xmath205 .
other basic algebraic properties of @xmath199 are @xmath206\rho}=0,\quad \mathcal{f}_{\mu\nu}\overline{\mathcal{f}}^{\mu\nu}=0 . \label{f2}\ ] ] from a killing vector @xmath195 we may define its associated _ ernst 1-form _ by @xmath207 where @xmath208 is the norm of the killing vector and @xmath209 denotes its _ twist form_. it is well - known that in a vacuum spacetime @xmath210 is closed i.e . @xmath211}=0 $ ] .
thus locally if we are in a simply connected region there exists a scalar field @xmath212 , the _ ernst potential _ , such that @xmath213 .
the self - dual weyl tensor and the self - dual killing form are related by @xmath214 a property which is easily obtained from ( [ covdivf ] ) .
finally , we introduce the tensor @xmath215 defined by @xmath216 the tensor @xmath215 has the obvious symmetries @xmath217\lambda\rho}= \mathcal{i}_{\lambda\rho\mu\nu}$ ] .
this tensor can be regarded as a metric in the space of self - dual 2-forms .
we have now all the ingredients to define the _ mars - simon tensor_. this is a four - rank tensor @xmath218 whose expression is given by @xmath219 from the previous expression we deduce that in a vacuum spacetime admitting a killing vector , the mars - simon tensor is determined up to a choice for the ernst potential .
we make a choice such that @xmath220 on @xmath14 .
our interest in the mars - simon tensor lies in the following result see @xcite for a proof .
let @xmath0 be a smooth , vacuum spacetime with a killing vector @xmath195 .
let @xmath199 denote the associated self - dual killing form . if there is a non - vanishing real constant @xmath221 such that the conditions @xmath222 hold on a non - empty @xmath223 then @xmath80 is locally isometric to the kerr spacetime .
[ kerr - characterization ] condition ( [ kerr_condition1 ] ) of theorem [ kerr - characterization ] can be reformulated in the form @xmath224 where the scalar @xmath225 is defined for any killing vector @xmath226 and any real constant @xmath221 by the expression @xmath227 a very important property of the scalar @xmath225 is that it fulfils the differential equation @xmath228 to see it , differentiate both sides of ( [ mars - simon : modified ] ) and use ( [ covdivfs ] ) to replace the covariant derivatives of @xmath199 .
the result is @xmath229 next use ( [ mars - simon : tensor ] ) to write @xmath230 in terms of @xmath231 and replace all the occurrences of @xmath232 by means of ( [ 2formelement ] ) to obtain @xmath233 we use here ( [ self - duality ] ) and ( [ ernst1 ] ) with the result @xmath234 from which ( [ cdxi ] ) is deduced .
in @xcite it is proved that the mars - simon tensor @xmath235 is a weyl candidate .
furthermore , it was shown that its weyl current is a linear function of @xmath235 in which case the wave equation ( [ box_weyl ] ) is homogeneous in @xmath235 .
the derivation of these results are discussed as some of the intermediate calculations will be needed in section [ cp - ms ] .
the mars - simon tensor is a weyl candidate .
[ mars - weyl ] _ proof . _ from ( [ mars - simon : tensor ] ) it is clear that @xmath236\alpha\lambda}=\mathcal{s}_{\mu\nu[\alpha\lambda]}= \mathcal{s}_{\alpha\lambda\mu\nu}$ ] . using ( [ f2 ] ) one can show that @xmath237 and therefore the mars - simon tensor is traceless .
to finish our proof we need to show that the mars - simon tensor fulfils the cyclic identity . to that end
we compute @xmath238}$ ] getting @xmath239}=\frac{6}{1-\sigma } \mathcal{f}_{\mu[\nu}\mathcal{f}_{\rho\lambda]}-\frac{\mbox{i}\ \mathcal{f}^2}{2(1-\sigma ) } \eta_{\mu\nu\rho\lambda}.\ ] ] the right hand side of this expression vanishes by virtue of the identity @xmath240}=\frac{1}{12}\mbox{i}\
\mathcal{f}^2 \eta_{\mu\nu\rho\lambda},\ ] ] which in turn is a consequence of the algebraic property @xmath241}= \mathcal{f}_{[\mu\nu}\mathcal{f}_{\rho\lambda]}$ ] and the fact that in four dimensions @xmath242}$ ] is proportional to @xmath243 .
a direct computation shows that the mars - simon tensor is self - dual @xmath244 therefore @xmath245 and @xmath246 are then related by duality as explained in section [ weyl : c ] : @xmath247 any property of @xmath248 will admit an equivalent counterpart formulated in terms of either its real or imaginary part .
in particular , this shows that both @xmath245 and @xmath246 are weyl candidates and any of them contains the same information as the original mars - simon tensor .
[ covdivms ] the covariant divergence of the mars - simon tensor is given by @xmath249\beta\alpha\rho}+\mathcal{f}_{\beta}^{\phantom{\beta}\rho } \mathcal{s}_{\gamma\delta\alpha\rho})+2g_{\beta[\gamma}\mathcal{s}_{\delta]\alpha\lambda\rho}\mathcal{f}^{\lambda\rho}\right)\xi^\alpha . \label{covdivmarssimon}\ ] ] [ cdms ] _ proof .
_ to calculate @xmath250 we start from ( [ mars - simon : tensor ] ) and take the covariant divergence of @xmath218 .
next we apply the relations ( [ covdivfs ] ) , @xmath251 , @xmath252 and @xmath253 .
finally we expand all the occurrences of @xmath254 using ( [ 2formelement ] ) with the result @xmath255 } ) \mathcal{f}^2 ) \nonumber \\ & & + \frac{1}{1-\sigma}((2\mbox{i}\ \eta_{\beta\gamma\delta\alpha } \overline{\mathcal{w}}_{\rho\phantom{\alpha}\pi\kappa}^{\phantom{\rho}\alpha } \xi^\rho+4g_{\beta[\delta}\overline{\mathcal{w}}_{\gamma]\alpha\pi\kappa}\xi^\alpha)\mathcal{f}^{\pi\kappa}-12\mathcal{f}_{\beta}^{\phantom{\beta}\rho}\overline{\mathcal{w}}_{\gamma\delta\alpha\rho}\xi^\alpha)\end{aligned}\ ] ] the dependence on the ernst 1-form can be removed from this expression if we write @xmath210 in terms of @xmath199 and @xmath195 by means of ( [ ernst1 ] ) .
the final step is to rearrange the terms in the resulting expression in order to render it in the form of ( [ covdivmarssimon ] ) .
this is accomplished by using ( [ mars - simon : tensor ] ) to write @xmath256 in terms of the mars - simon tensor , expanding again all the occurrences of @xmath254 by means of ( [ 2formelement ] ) and using ( [ f2 ] ) , ( [ self - duality ] ) where necessary .
equation ( [ covdivmarssimon ] ) enables us to define the weyl current of the mars - simon tensor as @xmath257\beta\alpha\rho}+\mathcal{f}_{\beta}^{\phantom{\beta}\rho } \mathcal{s}_{\gamma\delta\alpha\rho})+2g_{\beta[\gamma}\mathcal{s}_{\delta]\alpha\sigma\rho}\mathcal{f}^{\sigma\rho}\right)\xi^\alpha .
\label{ms - weylcurrent}\ ] ] since @xmath258 is linear in the mars - simon tensor , we conclude in view of ( [ box_weyl]) that @xmath259 fulfils a homogeneous wave equation .
this interesting property was used in @xcite to show that the _ domain of outer communication _ of an asymptotically flat smooth stationary spacetime is , under certain additional conditions , isometric to the kerr domain of outer communication i.e .
the _ uniqueness of kerr_. using @xmath260 together with the self - duality of the mars - simon tensor one deduces @xmath261 that is to say , @xmath262 is also self - dual with respect to the block of indices @xmath263 . in our forthcoming calculations
it is necessary to use the orthogonal splitting of @xmath248 with respect to a unit timelike vector field @xmath12 . since @xmath248 is a weyl candidate we can use ( [ decompose - weyl ] ) to find such a orthogonal splitting .
the resulting split is given by @xmath264\mu}+ l_{\mu[\rho}\mathcal{t}_{\lambda]\nu}+\mbox{i}\ \varepsilon_{\rho\lambda\alpha}n_{[\mu}\mathcal{t}_{\nu]}^{\phantom{\nu]}\alpha}+ \mbox{i}\ \varepsilon_{\mu\nu\alpha}n_{[\rho}\mathcal{t}_{\lambda]}^{\phantom{\nu]}\alpha } ) , \label{decompose - ms}\ ] ] where @xmath265 the tensor @xmath266 is symmetric , spatial and traceless .
alternatively , the orthogonal splitting of the mars - simon tensor can be calculated directly from ( [ mars - simon : tensor ] ) . to that end
one needs to find the orthogonal splitting of the different quantities which appear in ( [ mars - simon : tensor ] ) .
the orthogonal splitting of @xmath267 is given in ( [ decompose - weyl ] ) and the orthogonal splitting of @xmath199 is calculated by means of ( [ decompose_h ] ) and is given by @xmath268 } , \label{decompose - f}\ ] ] where @xmath269 one also needs the orthogonal splitting of @xmath270 which is easily calculated from ( [ 2formelement ] ) and ( [ decompose_eta ] ) so that @xmath271\nu}+2n_{[\nu}h_{\mu][\lambda}n_{\rho]}+\mbox{i}\ ( n_{[\mu}\varepsilon_{\nu]\rho\lambda}+n_{[\rho}\varepsilon_{\lambda]\mu\nu } ) .
\label{decompose_i}\ ] ] we insert relations ( [ decompose - weyl ] ) , ( [ decompose - f ] ) and ( [ decompose_i ] ) into the formula ( [ mars - simon : tensor ] ) and then equal the resulting expression to ( [ decompose - ms ] ) . the result of that is the relation @xmath272 which shall be used later on . also needed later on is the relation @xmath273 which is derived from equation ( [ decompose - f ] ) .
the killing vector @xmath195 can be decomposed in the form @xmath274 where @xmath275 , @xmath276 are called the _ killing lapse _ and the _ killing shift _ respectively .
we need to find a formula relating @xmath277 to @xmath275 , @xmath278 . to that end , we start by taking the covariant derivative of both sides of @xmath279 and replace @xmath280 by means of ( [ gradnormaltoextrinsic ] ) . the result is @xmath281 next we use in this expression the property @xmath282 and equation ( [ decompose_cdt ] ) to replace @xmath283 and @xmath284 respectively yielding @xmath285 the antisymmetric part of this expression gives the orthogonal splitting of the 2-form @xmath202 which is @xmath286}+n_{[\mu}d_{\nu]}y+d_{[\mu}y_{\nu]}- n_{[\mu}\mathcal{l}_{\vec{n}}y_{\nu]}. \label{ot - f}\ ] ] also the symmetric part of ( [ decompose_cdxi ] ) is the killing condition @xmath287 .
explicitly one has @xmath288 from which we deduce @xmath289 which renders the relation ( [ ot - f ] ) in the form @xmath290\rho}y^\rho+d_{[\mu}y_{\nu]}+2n_{[\mu}d_{\nu]}y . \label{decompose_f}\ ] ] from this expression we may compute the orthogonal splitting of @xmath291 which is @xmath292\alpha\beta}d^\alpha y^\beta , \label{decompose_fdual}\ ] ] where equation ( [ decompose_eta ] ) was used along the way . inserting equations ( [ decompose_f ] ) and ( [ decompose_fdual ] ) in ( [ mcalf ] ) and applying ( [ define - e ] ) on the resulting expression we deduce @xmath293 which is the required relation .
in this section a proof of the following result is provided . the mars - simon tensor @xmath248 propagates causally .
[ ms - propagation ] _ proof .
_ according to ( [ re - im ] ) it is enough to show that either @xmath245 or @xmath246 propagates causally .
we choose to work with the former , so let us set @xmath294 and define @xmath295 .
equation ( [ covdivmarssimon ] ) entails @xmath296 hence , @xmath297 is the weyl current of the weyl candidate @xmath298 . to study the causal propagation of @xmath298 we follow the general procedure explained in section [ causal - prop - weyl ] .
we denote by @xmath299 the bel - robinson tensor constructed from the weyl candidate @xmath298 . to prove our result
, we need to find a good estimate for the quantity @xmath300 see equation ( @xmath301 ) .
one can write @xmath302 where ( [ covdivbr ] ) was used . in order to work out the right hand side of equation ( [ q - est ] ) we calculate the orthogonal splitting of @xmath298 and @xmath303 with respect to @xmath12 .
the tensors @xmath298 and @xmath304 are related to @xmath248 and @xmath262 by means of the relations @xmath305 and therefore their orthogonal splittings can be calculated once those of @xmath259 and @xmath262 are known .
the orthogonal splitting of @xmath259 is given in equation ( [ decompose - ms ] ) and the orthogonal splitting of @xmath262 can be calculated from equation ( [ ms - weylcurrent ] ) given that we know the orthogonal splitting of all the quantities which appear in the definition of @xmath262 see ( [ ot - mcalj ] ) and ( [ ot - mcalz])-([ot - mcalp ] ) in appendix [ ot - j ] for the precise formulae .
inserting these orthogonal splittings in ( [ q - est ] ) we get @xmath306\nonumber\\ & & \hspace{5cm}-6\mbox{i}\ \varepsilon_{\mu\nu\rho}\mathcal{t}^{\lambda\nu } \overline{\mathcal{t}}_{\lambda}^{\phantom{\sigma}\rho}y\left(\frac{\mathcal{e}^\mu}{\sigma-1}-\frac{\overline{\mathcal{e}}^\mu } { \overline{\sigma}-1}\right ) .
\label{step1}\end{aligned}\ ] ] now , using the equations ( [ ots1])-([ots2 ] ) of appendix [ ot - br ] we can deduce the property @xmath307 which when combined with ( [ step1 ] ) yields @xmath308 . \label{step2}\ ] ] equation ( [ step2 ] ) is the key to get the required estimate . to obtain it ,
let us introduce an orthonormal frame @xmath130 with @xmath309 .
since @xmath310 , @xmath276 are spatial vectors they can be written as a linear combination of the vectors @xmath311 .
therefore , point ( [ second ] ) of theorem [ sproperties ] entails the estimates @xmath312\leq m_1 b_{\mu\nu\alpha\beta}n^\mu n^\nu n^\alpha n^\beta,\\ -\mbox{re}\left[\frac{12}{1-\sigma}b_{\mu\nu\alpha\beta } \mathcal{e}^\mu n^\alpha n^\beta y^\nu\right]\leq m_2 b_{\mu\nu\alpha\beta}n^\mu n^\nu n^\alpha n^\beta,\end{aligned}\ ] ] for some scalar continuous functions @xmath313 , @xmath314 defined on @xmath144 .
using these estimates in equation ( [ step2 ] ) we obtain our final estimate @xmath315 where @xmath316 is the upper bound of @xmath317 in the compact set @xmath144 and thus ( [ gronwall ] ) can be written in the form @xmath318 since @xmath179 , then an application of grnwall s lemma enables us to conclude that @xmath319 , @xmath320 which , according to the considerations made at the end of section [ causal - prop - weyl ] proves the causal propagation .
_ we stress that the fact that @xmath248 satisfies a wave equation and the property ( [ covdivmarssimon ] ) do not , by themselves , suffice to show the causal propagation of this tensor .
if we impose that @xmath248 and some directional derivative of this quantity vanish on a spacelike hypersurface @xmath115 then using the wave equation and standard results on hyperbolic partial differential equations allow us to conclude that @xmath248 vanishes on a neighbourhood of @xmath115 . in doing so , extra conditions needed to be imposed so that @xmath248 is indeed zero in @xmath114 .
this approach leads , in principle , to a weaker result than the one given in theorem [ ms - propagation ] .
_ consider now any vacuum spacetime admitting a killing vector @xmath195 and use @xmath195 to construct the mars - simon tensor @xmath248 and as above , define its real part @xmath298 and the bel - robinson tensor @xmath299 of @xmath298 .
for any unit timelike vector @xmath12 , define the positive scalar @xmath321 one can find an explicit expression for @xmath322 if we use the property see equation ( [ ots0 ] ) in [ ot - br] @xmath323 - 12\left|\frac{\mathcal{e}^\mu\mathcal{e}_\mu}{\sigma-1}\right|^2 + 36\left|\frac{\mathcal{e}^\mu\overline{\mathcal{e}}_\mu}{\sigma-1}\right|^2 , \label{se - density}\ ] ] which shows that @xmath322 can be expressed in terms of @xmath324 , @xmath277 and @xmath212 exclusively .
when @xmath322 vanishes on a cauchy hypersurface , the causal propagation of @xmath248 entails the following result .
let @xmath0 be a globally hyperbolic vacuum spacetime and suppose that @xmath14 admits a killing vector @xmath195 .
if there are a cauchy hypersurface @xmath111 and a real constant @xmath221 such that @xmath325 and @xmath326 where @xmath327 is the unit timelike normal to @xmath115 , then @xmath14 is locally isometric to the kerr spacetime .
[ positive - f ] _ proof . _ if @xmath328 is zero on @xmath115 then by ( [ phi ] ) @xmath329 .
therefore , remark [ remark : se_density ] leads us to @xmath330 using that the mars - simon tensor propagates causally , one deduces that @xmath331 on @xmath14 .
in particular , this implies that ( [ cdxi ] ) is rendered in the form @xmath332 which can be integrated to yield @xmath333 for some complex constant @xmath334 .
the condition @xmath326 implies that @xmath335 and hence @xmath225 vanishes on @xmath14 .
theorem [ kerr - characterization ] can now be applied and we conclude that @xmath14 is locally isometric to the kerr spacetime .
we explain in this section how to use previous results to find conditions which ensure that the data development of a vacuum initial data set is isometric to a subset of the kerr spacetime .
the basic idea behind is to express the result of theorem [ positive - f ] in terms of conditions on a vacuum initial data set @xmath336 .
let @xmath336 be a vacuum initial data set and assume that there exist two scalar fields @xmath337 , @xmath338 , a vector field @xmath339 and a real constant @xmath221 , all defined on @xmath69 , fulfilling the following conditions @xmath340-\frac{3}{4}m^4|1-\tilde{\sigma}|^6 + 36\left|\frac{\mathcal{e}^i\overline{\mathcal{e}}_i}{\tilde{\sigma}-1}\right|^2=0 , \label{scalar0}\\ \fl d_j\tilde{\sigma}=2\tilde{y}k_{jl}\tilde{y}^l-2\ \mbox{i}\ \varepsilon_{jml } ( k_{k}^{\phantom{k}l}\tilde{y}^k\tilde{y}^m+\tilde{y}^m d^l\tilde{y})+2\tilde{y}d_j\tilde{y}+2\tilde{y}^l d_{[l}\tilde{y}_{j]},\label{dsigma}\\ \fl d_{(i}\tilde{y}_{j)}+\tilde{y } k_{ij}=0 , \label{kid1 } \\ \fl d_i d_j\tilde{y}+\mathcal{l}_{\tilde{y}^l } k_{ij } = \tilde{y}(r_{ij } + k k_{ij } - 2 k_{il } { k^{l}_{\phantom{l}j } } ) , \label{kid2}\end{aligned}\ ] ] where @xmath341 then the data development @xmath0 of @xmath336 is locally isometric to an open subset of the kerr spacetime . [ theorem : kerr_data ] [ killing_existence]_the conditions ( [ kid1 ] ) and ( [ kid2 ] ) are the _ killing initial data _ ( kid ) equations .
it is well - known that if both a solution @xmath342 and the initial data set @xmath336 are _ suitably _ smooth , then the development of the initial data @xmath336 possesses a killing vector such that the pull - back of its restriction to @xmath69 coincides with the given killing data see for example , @xcite , and in particular @xcite for an explicit set of smoothness conditions on the killing data . alternatively ,
if the killing data is _ transversal _ , that is @xmath343 , then one can make use of the concept of killing development to ensure the existence of a spacetime with a killing vector associated to the killing data under question see @xcite .
again , under suitable regularity conditions it can be shown that the killing development always includes the cauchy development of @xmath336 see again @xcite : the killing development is unique and maximal among the class of spacetimes containing the relevant cauchy data and a killing vector with non - closed orbits ; hence , if the cauchy development has a timelike killing vector , then it must be a subset of the killing development . _ _ proof .
_ let @xmath14 be the data development of the vacuum initial data set @xmath81 and let @xmath195 be the killing vector mentioned in remark [ killing_existence ] .
as usual consider a foliation @xmath29 , @xmath148 , of @xmath14 with @xmath84 and let @xmath12 be the unit timelike normal to the leaves of this foliation .
let us adopt the formalism and notation explained in section [ section : mstensor ] with @xmath195 being the killing vector .
then the orthogonal splitting of @xmath195 is @xmath344 .
as explained in remark [ killing_existence ] we have the property @xmath345 which combined with equations ( [ ei ] ) and ( [ def - mce ] ) entails @xmath346 .
the orthogonal splitting of @xmath210 is calculated on one hand from equation ( [ ernst1 ] ) using the expressions ( [ decompose - f ] ) , ( [ def - mce ] ) and the property @xmath344 .
the result is @xmath347 where @xmath348}.\label{def - smu}\end{aligned}\ ] ] on the other hand , the orthogonal splitting of @xmath210 can be calculated from the relation @xmath349 yielding @xmath350 combining the previous equation with the pull - back of equation ( [ def - smu ] ) under @xmath65 and ( [ dsigma ] ) we deduce @xmath351 upon a suitable choice of the ernst potential @xmath212 , this equation can be integrated on @xmath69 to give @xmath352 .
next consider the mars - simon tensor @xmath248 constructed with the killing vector @xmath195 and calculate its orthogonal splitting with respect to @xmath12 equation ( [ decompose - ms ] ) . using ( [ tgoestoe ] ) and our previous results , we get @xmath353 expanding this product and using conditions ( [ e2])-([scalar0 ] ) we conclude @xmath354 moreover , equation ( [ e2 ] ) implies @xmath355 see ( [ ot - f2 ] ) .
thus , theorem [ positive - f ] applies and therefore we conclude that @xmath0 must be locally isometric to an open subset of the kerr spacetime .
theorem [ theorem : kerr_data ] asserts that under certain conditions the development of a vacuum initial data set is isometric to an open subset of the kerr spacetime but nothing is said as to the resulting development being one of the specializations of the kerr spacetime : the schwarzschild solution . in this section
we show the form which theorem [ theorem : kerr_data ] takes in that case .
this complements previous work of the present authors about the same issue @xcite . to state our result
we need a preliminary lemma whose proof can be found in @xcite .
let @xmath336 be a vacuum initial data set and assume that there exist a scalar @xmath337 and a vector field @xmath339 , all defined on @xmath69 , fulfilling the following conditions @xmath356}+ 2 \tilde{y}_{[i}d_{j]}\tilde{y } + 2 \tilde{y}_{[i}k_{j]l}\tilde{y}^l = 0,\label{kids-1}\\ & & \tilde{y}_{[i}d_{j}\tilde{y}_{k]}=0\label{kids-2}.\end{aligned}\ ] ] then there exists an integrable killing vector @xmath195 in the data development of @xmath336 .
[ lemma : static_killing ] with the aid of this lemma we can now derive conditions guaranteeing that the development of a vacuum initial data set is isometric to an open subset of the schwarzschild spacetime let @xmath336 be a vacuum initial data set and assume that there exist two scalar fields @xmath337 , @xmath338 and a vector field @xmath339 , all defined on @xmath69 , fulfilling the conditions of theorem [ theorem : kerr_data ] with ( [ kid1])-([kid2 ] ) replaced by ( [ kids-1])-([kids-2 ] ) .
then the data development @xmath14 is locally isometric to an open subset of the schwarzschild spacetime . [
theorem : schwarzschild_data ] _ proof .
_ the first step in the proof of this theorem is to use lemma [ lemma : static_killing ] to show that the data development @xmath14 admits an integrable killing vector @xmath195 .
the remaining part of the proof follows the same pattern as that of theorem [ theorem : kerr_data ] and therefore we reach the conclusion that @xmath14 is isometric to an open subset of a specialization of the kerr spacetime possessing an integrable killing vector which , as is well known , is the schwarzschild spacetime . _
it is worth recalling that the analysis of @xcite has provided an explicit formula for a _ timelike kid candidate _ in terms of concomitants of the initial data @xmath357 .
hence , a combination of theorem [ theorem : schwarzschild_data ] with the such timelike kid candidate renders an algorithmic characterization of schwarzschild initial data alternative to the one given in @xcite . _
the main result obtained in this article , theorem [ theorem : kerr_data ] provides a characterization of initial data for the kerr spacetime by first requiring that a set of overdetermined elliptic differential equations the kid equations admit a solution ; and then in turn requiring that certain objects constructed out of the solutions to the kid equation satisfy very particular relations .
this characterization fails to be algorithmic in two ways .
first , it is in practise a non - trivial problem to verify whether a certain initial data set admits a kid or not .
a pointwise procedure for doing this has been discussed in @xcite .
alternatively , one could resort to a non - time symmetric generalization of the geometric invariant characterizing static time symmetric initial data sets constructed in @xcite which requires solving a fourth - order elliptic partial differential equation . in any case
and this leads to the second way the result fails to be algorithmic even if one could ascertain the existence of a kid on the initial data , one requires to know the kid in an explicit manner in order to be able to verify the conditions ( [ e2 ] ) , ( [ scalar0 ] ) and ( [ dsigma ] ) .
it is plausible that from an hypothetical characterization of the kerr spacetime in terms of , say , concomitants of the weyl tensor analogous to that for the schwarzschild spacetime given in @xcite one would be able to construct , following the ideas of @xcite , kid candidates for the initial data set @xmath336 .
if such a kid candidate were available , then theorem [ theorem : kerr_data ] would constitute a powerful tool to analyse the behaviour of dynamical black hole spacetimes , both from an analytic and a numerical point of view .
we would like to thank jmm senovilla , m mars , w simon and cm losert - vk for useful remarks on the manuscript .
javk is supported by an epsrc advanced research fellowship .
agp gratefully acknowledges financial support from the spanish `` ministerio de ciencia e innovacin '' under the postdoctoral grant ex-2006 - 0092 .
given the symmetries of the weyl current @xmath262 it is immediate to realise that its orthogonal splitting must take the form @xmath358\mu}+n_\mu
n_{[\nu}\mathcal{v}_{\rho]}+ \mathcal{p}_{\mu\nu\rho}. \label{ot - mcalj}\ ] ] it takes a greater effort , however , to find the explicit expression of the spatial tensors @xmath359 , @xmath360 , @xmath361 , @xmath362 . these expressions are found by inserting into equation ( [ ms - weylcurrent ] ) the orthogonal splitting of @xmath248 shown in ( [ decompose - ms ] ) , the orthogonal splitting of @xmath199 shown in ( [ decompose - f ] ) and the relation @xmath344 .
a computer calculation yields the results @xmath363\alpha\rho}+3\mathcal{e}_{[\mu}\mathcal{t}_{\nu]\rho}y^\rho+y_{[\nu}\mathcal{t}_{\mu]\alpha}\mathcal{e}^\alpha-\mbox{i}\ \varepsilon_{\mu\nu\rho}\mathcal{e}^\alpha\mathcal{t}_\alpha ^{\phantom{\alpha}\rho}y),\label{ot - mcalz}\\ & & \hspace{-2cm}\mathcal{x}_{\mu\nu}\equiv \frac{1}{1-\sigma}\bigg(\mathcal{e}^\alpha\big(4\mbox{i } y(\mathcal{t}_{(\alpha}^{\phantom{(\alpha}\rho}\varepsilon_{\nu)\mu\rho}-\varepsilon_{\nu\alpha\rho}\mathcal{t}_{\mu}^{\phantom{\mu}\rho})+6\mathcal{t}_{\nu\mu}y_\alpha-2\mathcal{t}_{\mu\alpha}y_{\nu}-4\mathcal{t}_{\nu\alpha}y_\mu+4h_{\nu\mu}\mathcal{t}_{\alpha\rho}y^\rho \big ) \nonumber\\ & & \hspace{1cm}+6\mathcal{e}_\nu\mathcal{t}_{\mu\alpha}y^\alpha\big ) \bigg),\label{ot - mcalx}\\ & & \hspace{-2cm}\mathcal{v}_\mu\equiv\frac{\mathcal{e}^\alpha}{1-\sigma}\left(4\mbox{i}\ y^\rho(\varepsilon_{\mu\lambda(\rho}\mathcal{t}_{\alpha)}^{\phantom{\alpha)}\lambda}+\varepsilon_{\alpha\rho\lambda}\mathcal{t}_{\mu}^{\phantom{\mu}\lambda})-2y\mathcal{t}_{\mu\alpha}\right),\label{ot - mcalv}\\ & & \hspace{-2cm}\mathcal{p}_{\mu\nu\rho}\equiv\frac{2}{1-\sigma}\bigg(2\ \mbox{i}\ \varepsilon{}_{\nu}{}_{\rho}{}_{\lambda } \
\mathcal{e}{}_{\mu } \mathcal{t}{}_{\alpha}{}^{\lambda } y{}^{\alpha } + 2\mathcal{e}{}_{[\rho } ( 3y\mathcal{t}{}_{\nu]\mu } + \mbox{i } \
\varepsilon_{\nu]\mu\lambda}\mathcal{t}{}_{\alpha}{}^{\lambda } \
y{}^{\alpha } ) \nonumber\\ & & \hspace{1cm}+\mathcal{e}{}^{\alpha}\bigg(-2\mbox{i } \big(\varepsilon{}_{\rho}{}_{\alpha}{}_{\lambda}\mathcal{t}{}_{[\nu}{}^{\lambda } \
y{}_{\mu ] } + \varepsilon{}_{\nu}{}_{\alpha}{}_{\lambda } \mathcal{t}{}_{[\mu}{}^{\lambda } y{}_{\rho ] } + 2\varepsilon{}_{\mu}{}_{\alpha}{}_{\lambda}\mathcal{t}{}_{[\nu}{}^{\lambda } y{}_{\rho]}\nonumber \\ & & \hspace{1 cm } + y{}^{\lambda } ( \varepsilon{}_{\nu}{}_{\rho}{}_{\alpha } \mathcal{t}{}_{\mu}{}_{\lambda}+\varepsilon{}_{\mu}{}_{\alpha}{}_{[\nu } \mathcal{t}{}_{\rho]}{}_{\lambda})\big)\nonumber\\ & & \hspace{1cm}+2h{}_{\mu}{}_{[\nu } \big(\mathcal{t}{}_{\rho]}{}_{\alpha } y + \mbox{i}\ y{}^{\lambda } ( \varepsilon{}_{\rho]}{}_{\lambda}{}_{\delta } \mathcal{t}{}_{\alpha}{}^{\delta } + \ \mathcal{t}{}_{\rho]}{}^{\delta}\varepsilon{}_{\alpha}{}_{\lambda}{}_{\delta}+ \ \varepsilon{}_{\rho]}{}_{\alpha}{}_{\delta } \mathcal{t}{}_{\lambda}{}^{\delta})\big)\bigg)\bigg).\label{ot - mcalp}\end{aligned}\ ] ] in @xcite the orthogonal splitting of the bel - robinson tensor was calculated and the different parts resulting from such splitting were studied in detail .
the calculations are similar for the bel - robinson tensor of a weyl candidate and therefore we only need to adapt these results for the particular case of @xmath298 . in this case
the orthogonal splitting of @xmath299 reads see section ( 5.1 ) of @xcite @xmath365 where the different parts of this splitting are given by @xmath366 where @xmath367 and @xmath368 are respectively the _ electric _ and _ magnetic _ parts of the weyl candidate @xmath369 . from ( [ br - decomposition ] ) and ( [ br - parts ] )
we get @xmath370 where the relations @xmath371 need to be used . | exploiting a 3 + 1 analysis of the mars - simon tensor , conditions on a vacuum initial data set ensuring that its development is isometric to a subset of the kerr spacetime are found .
these conditions are expressed in terms of the vanishing of a positive scalar function defined on the initial data hypersurface .
applications of this result are discussed . |
the low - energy properties of graphene have been widely investigated theoretically and experimentally in recent years @xcite .
it is well - known that the low - lying elementary excitations of graphene are massless dirac fermions , which have linear dispersion and display quite different behaviors from ordinary electrons with parabolic dispersion . at half - filling state ,
the density of states of massless dirac fermions vanishes linearly with energy @xmath1 near the fermi level . due to this fact
, there is essentially no screening on the coulomb interaction between dirac fermions .
the unscreened , long - range coulomb interaction was argued @xcite to be responsible for a plenty of unusual physical properties , including the logarithmic velocity renormalization @xcite , the logarithmic specific heat correction @xcite , the presence of quantum critical point @xcite , and the marginal fermi liquid quasiparticle lifetime @xcite .
when the unscreened coulomb interaction is sufficiently strong , the semimetal ground state of graphene may no longer be stable .
there exists an interesting possibility that the massless dirac quasiparticles and quasiholes are bound into pairs through the attractive coulomb interaction between them . as a consequence ,
the massless dirac fermions acquire a finite mass and the ground state of graphene becomes insulating .
this semimetal - insulator transition is usually called excitonic instability in the literature @xcite .
it can be identified as the non - perturbative phenomenon of dynamical chiral symmetry breaking conventionally studied in the context of particle physics @xcite .
both dyson - schwinger equation @xcite and lattice simulation approaches @xcite found that such excitonic instability occurs only when the fermion flavor is less than a critical value @xmath2 and the coulomb strength is larger than a critical value @xmath3 .
if we fix the physical fermion flavor @xmath4 , then the semimetal - insulator transition happens at a single critical point @xmath3 .
the effective coupling parameter of coulomb interaction can be defined as @xmath5 with @xmath6 being the dielectric constant and @xmath7 being the effective velocity . in the clean limit , the physical magnitude of this parameter is around 3 or 4 for graphene in vacuum . in the same limit , we found by solving gap equation to the leading order of @xmath8 expansion that the critical strength @xmath9 @xcite . using analogous gap equation approach ,
the critical coupling is found to be @xmath10 and @xmath11 respectively in ref .
@xcite and ref .
in addition , the monte carlo study @xcite performed in lattice field theories found that the critical strength @xmath12 at @xmath13 .
our critical coupling is much more closer in magnitude to that of monte carlo study .
once a fermion mass gap is generated , the low - energy properties of graphene fundamentally change . below the energy scale set by the fermion gap
, the density of states of fermions is substantially suppressed , which would produce important consequences .
it is interesting to study some observable physical quantities those can serve as signatures for the existence of excitonic instability .
the effects of dynamical fermion gap have been discussed by several authors @xcite . in this paper , we calculate the specific heat and susceptibility of dirac fermions and other low - energy excitations in both semimetal and excitonic insulator phases .
these quantities can be compared with experimental results and hence may help to understand the physical consequence of excitonic instability . in the semimetal phase ,
the coulomb interaction is not strong enough to trigger excitonic pairing instability , but it is strong enough to produce unusual properties .
as found by vafek @xcite , the long - range coulomb interaction gives rise to logarithmic @xmath14-dependence of fermion specific heat , which is clearly not behavior of normal fermi liquid . in this paper , we re - derive the same qualitative result by a different method .
we also calculate the susceptibility of massless dirac fermions and show that it also exhibits logarithmic @xmath14-dependence due to long - range coulomb interaction . in the insulator phase ,
the fermion density of states is suppressed by the excitonic gap .
intuitively , the specific heat and susceptibility of dirac fermions should drop significantly from their corresponding magnitudes in the semimetal phase .
our explicit computations will show that this is true .
however , the massive dirac fermions are not the true low - lying elementary excitations in the insulating state . at the low energy regime ,
the only degree of freedom is the massless goldstone boson which originates from the dynamical breaking of continuous chiral symmetry .
the goldstone bosons make dominant contribution to the total specific heat at low temperature , but make no contribution to the total susceptibility . in section [ sec :
model ] , we set up the hamiltonian of the system and define the physical quantities in which we are interested .
we calculate the free energy , specific heat and susceptibility in section [ sec : clean ] .
we finally summarize the results and discuss some relevant problems in section [ sec : conclusion ] .
the hamiltonian of massless dirac fermions in single layer graphene is given by @xmath15 where the coulomb interaction potential @xcite is @xmath16 where @xmath17 .
as mentioned in introduction , it is convenient to define a dimensionless coulomb coupling as @xmath18 .
usually , dirac fermion in two spatial dimensions is described by two - component spinor field whose @xmath19 representation can be formulated by pauli matrices @xmath20 .
however , it is not possible to define a @xmath19 matrix that anticommutes with all these matrices .
therefore , there is no chiral symmetry in this representation . here , we adopt four - component spinor field @xmath21 to describe the massless dirac fermion @xcite .
the conjugate spinor field is defined as @xmath22 .
the @xmath23 @xmath24-matrices can be defined as @xmath25 , which satisfy the standard clifford algebra @xmath26 with metric @xmath27 .
obviously , there are two @xmath23 matrices @xmath28 which anticommute with all @xmath29 .
the total hamiltonian preserves a continuous u(2n ) chiral symmetry @xmath30 .
the mass term generated by excitonic pairing will break this global chiral symmetry dynamically to subgroup @xmath31 .
meanwhile , according to the goldstone theorem , there appear massless goldstone bosons due to the breaking of continuous chiral symmetry .
these bosons are the only gapless excitations in the symmetry broken phase and hence play an important role in determining the low - energy behaviors of the system .
although the physical fermion flavor is actually @xmath4 , in the following we consider a general @xmath32 in order to perform @xmath8 expansion . for convenience , we work in units where @xmath33 throughout the paper .
the electronic structure of graphene is very special in that the @xmath34-conduction bands and @xmath35-valence bands touch at two inequivalent @xmath36 points .
this is the reason why the low - energy fermionic excitations have a linear dispersion .
when the strong , long - range coulomb interaction opens an excitonic gap at the dirac point , the chiral symmetry of total hamiltonian is broken , resembling the non - perturbative phenomenon of dynamical chiral symmetry breaking in qed@xmath37 @xcite .
this mechanism was first proposed in graphene by khveshchenko @xcite and has been extensively studied @xcite in the following years .
gusynin _ et _ _ al .
_ discussed the influence of excitonic fermion gap on various transport quantities , including electrical and hall conductivity @xcite .
the results were compared directly with the experiments in graphene .
recently , kotov _ et _ _ al .
_ studied the effect of fermion gap on the interacting potential @xcite and found an effective weak confinement of fermions .
they also argued that the massive phase exhibits much more interesting behavior than the massless one .
the effect of fermion gap on quasiparticle lifetime and spectral function was discussed in @xcite .
this kind of excitonic instability may also exist in other correlated electron systems than graphene .
for instance , it was suggested by one of the authors that such instability can provide a qualitative understanding on the field - induced thermal metal - insulator transition observed in the vortex state of high temperature cuprate superconductor @xcite . in this paper , we calculate the specific heat and susceptibility by including the effect of coulomb interaction in both the semimetal and insulator phases .
these are physical quantities those can be measured by experiments and hence can help us to build interesting connections between theoretical predictions and experimental facts .
technically , we will follow the procedures utilized in the paper of kaul and sachdev @xcite . in this framework , all propagators and correlation functions are written in the matsubara imaginary time formalism . at finite temperature ,
the fermion propagator is @xmath38 where @xmath39 is the fermion frequency .
although generally the excitonic fermion gap should depend on momentum , energy , and temperature , we assume a constant mass gap @xmath40 throughout the paper to simplify calculations . the bare coulomb interaction function is simply @xmath41 in an interacting electron gas , the collective excitations screen the bare coulomb interaction and convert @xmath42 to @xmath43 where the polarization function @xmath44 is defined as @xmath45 } { k^2(q+k)^{2}},\ ] ] with @xmath46 and @xmath47 .
the first two orders in @xmath8 expansion of free energy @xmath48 are @xmath49 where @xmath50 is the leading , noninteracting term and @xmath51 the corrections from the coulomb interaction .
the leading term of fermion free energy is defined as @xmath52 $ ] . to calculate the free energy , we will first sum over the matsubara frequencies @xmath53 and then perform the integration over the intermediate variables and momentum @xmath54 , dropping all terms those are independent of temperature and volume @xcite .
the volume factor is neglected throughout this paper and we only consider free energy in unit volume . the interaction correction to free energy is given by @xmath55.\ ] ] we choose the zero temperature free energy @xmath56 as the reference free energy @xcite , and then define the following regularized free energy @xmath57.\end{aligned}\ ] ] here , we follow the strategy of ref . @xcite and introduce a magnetic field @xmath58 . for fermions , the field shifts frequency as @xmath59 , where @xmath60 .
the specific heat @xmath61 and susceptibility @xmath62 can be defined as @xmath63 which are divided to free and interaction terms , respectively . for a normal fermi liquid ,
the specific heat and susceptibility should behave as @xmath64 and @xmath65 according to the analysis in ref .
if we write the specific heat as @xmath66 , then @xmath67 should be @xmath68 similarly , the susceptibility can also be written as @xmath69 with @xmath70 being @xmath71 in the presence of fermion mass @xmath40 , the specific heat ( susceptibility ) no longer behaves as @xmath72 ( @xmath73 ) .
however , in order to make direct comparison , we still express specific heat ( susceptibility ) in terms of @xmath67 ( @xmath70 ) , which will depend on temperature @xmath14 .
the definitions presented in this section will be used to calculate the free energy , specific heat , and susceptibility in the next section .
in the presence of a constant fermion mass @xmath40 , the noninteracting free energy is @xmath74 \nonumber \\ & = & -t\int \frac{d^{2}\mathbf{k}}{2\pi^{2 } } \ln \left[1 + e^{-\frac{\sqrt{\mathbf{k}^{2}+m^{2}}}{t}\pm i\theta\frac{h}{t}}\right ] \nonumber \\&= & - \frac{1}{\pi}\bigg\{\frac{1}{3}\left(\frac{m}{t}\right)^{3 } + \frac{1}{4}\frac{m}{t}\mathrm{li}_{2}\left[-e^{\frac{m}{t}\pm i\theta\frac{h}{t}}\right ] \nonumber \\ & & - \frac{1}{4}\mathrm{li}_{3}\left[-e^{\frac{m}{t}\pm i\theta\frac{h}{t}}\right]\bigg\}t^{3}.\end{aligned}\ ] ] here , @xmath75 and @xmath76 are polylogarithmic functions .
it is easy to get the following noninteracting term for fermion specific heat @xmath77 \nonumber \\ & & -6\left(\frac{m}{t}\right)\mathrm{li}_{2}\left[-e^{\frac{m}{t}}\right ] + 6 \mathrm{li}_{3}\left[-e^{\frac{m}{t}}\right]\bigg\}.\end{aligned}\ ] ] this function is plotted in fig .
[ fig : spheat_0 ] . taking the @xmath78 limit of eq .
( [ eq : spheat_0 ] ) , the specific heat in the semimetal phase is @xmath79 with @xmath80 .
similarly , the noninteracting susceptibility is given as @xmath81\big\},\ ] ] which is plotted in fig .
[ fig : sus_0 ] . taking the @xmath78 limit of eq .
( [ eq : sus_0 ] ) , the susceptibility in the semimetal phase is @xmath82 with @xmath83 . from eq .
( [ eq : spheat_0 ] ) , eq .
( [ eq : sus_0 ] ) , and fig .
[ fig : free_0 ] , it is easy to see that the @xmath14-dependence of specific heat and susceptibility in the insulating phase differs significantly from the corresponding @xmath72 and @xmath73 behaviors in the semimetal phase .
this is not unexpected because the excitonic gap strongly suppresses the fermionic excitations at low temperature .
however , although the low - energy fermion excitations are strongly suppressed in the insulating phase , there exists another kind of gapless excitation : goldstone boson .
the presence of gapless goldstone bosons is the characteristic property of excitonic instability .
they are composed of dirac fermions ( quasiparticles ) and anti - fermions ( quasiholes ) , but carry no electric charge themselves .
the goldstone bosons do not contribute to the susceptibility because they do not couple to external magnetic field @xmath58 , but they do contribute to the total specific heat of the system .
in particular , the free energy of goldstone bosons is @xmath84 = -\frac{2\zeta(3)}{\pi}t^3,\ ] ] while the corresponding specific heat is @xmath85 which has been obtained in ref .
. apparently , a @xmath72 term of specific heat appears in both the semimetal phase and the insulating phase .
it seems that these two phases have similar specific heat , albeit contributed from different elementary excitations .
however , such similarity actually does not exist because it disappears once the interaction correction to free energy is incorporated .
we now include the interaction correction to the free energy .
note the goldstone bosons are neutral , so the coulomb interaction only affects the free energy of dirac fermions . to calculate the free energy @xmath51
, we should first know the polarization function . in the presence of finite fermion mass @xmath40 and external magnetic field @xmath58 ,
the polarization function @xmath86 can be calculated by the methods presented in @xcite .
here we only write down the final expression : @xmath87 \nonumber \\&&\,\ , -\frac{n}{4\pi}\int_{0}^{1}dx \frac{(1 - 2x)\omega_{m } \sin(\frac{1}{t}x \omega_{m } + \frac{1}{t}\theta h)}{d_{m } } \nonumber \\&&\,\ , -\frac{n}{2\pi}\int_{0}^{1}dx \frac{m^{2}+x(1-x)\omega_{m}^{2}}{e_{f } } \frac{\sinh(\frac{1}{t}e_f)}{d_{m}},\end{aligned}\ ] ] at finite temperature . here
we introduced the following abbreviated notations @xmath88 the zero - temperature limit of polarization function ( eq .
( [ eq : pifull ] ) ) is @xmath89 using eq .
( [ eq : pifull ] ) and eq .
( [ eq : pi_0 ] ) , the free energy @xmath51 ( eq . ( [ eq : delta_f ] ) ) can be directly computed . in the semimetal phase with @xmath78 ,
the polarization function is @xmath90,\end{aligned}\ ] ] at finite temperature with @xmath91 and @xmath92 at zero temperature .
using these expressions , the free energy of dirac fermion is written as @xmath93 where @xmath94 } { \pi(1+y^{2})\left[\frac{1}{\lambda}+\frac{1}{\sqrt{1+y^{2}}}\right ] } , \\
j(x , y,\mathbf{q } ) & = & \frac{2\exp\left[-\mathbf{q}\sqrt{x(1-x)(1+y^{2})}\right ] } { \pi(1+y^{2})\left[\frac{1}{\lambda}+\frac{1}{\sqrt{1+y^{2}}}\right ] } \nonumber \\ & & \times\cos(xy\mathbf{q } + \theta\frac{h}{t}).\end{aligned}\ ] ] here , a variable @xmath95 is introduced , with @xmath1 being the continuous form of @xmath96 when @xmath97 . for finite @xmath98
, @xmath99 damps rapidly with growing @xmath98 , so @xmath100 makes the dominant contribution to the free energy
. we can expand the function @xmath99 near this point and obtain @xmath101}{(1+y^{2 } ) \left[\frac{1}{\lambda}+\frac{1}{\sqrt{1+y^{2}}}\right ] } \nonumber\\&\approx & \frac{8t^3}{\pi^3}\int_{\delta}^{\frac{\lambda}{t}}\mathbf{q}\,d\mathbf{q } \int_{0}^{\infty}dy \frac{1}{(1+y^{2})\left[\frac{1}{\lambda } + \frac{1}{\sqrt{1+y^{2}}}\right ] } \nonumber \\&&\times \int_{0}^{\frac{1}{2}}\frac{dx}{\exp\left[2\mathbf{q}\sqrt{x}\sqrt{(1+y^{2})}\right ] } \nonumber \\&= & \frac{4}{\pi^3}\eta(\lambda)\,t^{3}\ln\frac{\lambda}{t},\end{aligned}\ ] ] where @xmath102}{\lambda \sqrt{1-\lambda^2}}-\frac{\pi}{2\lambda } .\ ] ] comparing with @xmath99 , the form of @xmath103 is more complicated owing to the cosine term @xmath104 .
the computation becomes difficult if we make taylor expansion of the cosine function . by plotting the dependence of function @xmath103 on its variables
, we found that the dominant regime is @xmath105 .
hence , we simply take @xmath103 as @xmath106 } { ( 1+y^{2})\left[\frac{1}{\lambda}+\frac{1}{\sqrt{1+y^{2}}}\right]}\cos(\theta \frac{h}{t}),\ ] ] which then leads to @xmath107 taking @xmath108 , the total free energy now has the form @xmath109 it is easy to get the following specific heat and susceptibility @xmath110 here , the ultraviolet cutoff @xmath111 can be taken to be of order @xmath112ev , which is determined by @xmath113 with lattice constant @xmath114 . from these results , we know that both specific heat and susceptibility of massless dirac fermions exhibit logarithmic @xmath14-dependence due to long - range coulomb interaction .
these are non - fermi liquid behaviors .
the appearance of such singular fermion specific heat was first pointed out by vafek @xcite . here
, we obtained the same qualitative @xmath14-dependence by a different method . in ref .
@xcite , the calculation of free energy was performed on the basis of the retarded vacuum polarization functions and retarded fermion propagator @xmath115 , while in our case the polarization functions and fermion propagator are expressed in the matsubara formalism .
strictly speaking , these two polarization functions are equivalent and should lead to the same results .
we numerically compute the free energy using both the polarization functions obtained in the present paper and that in ref .
@xcite , and found that the results are very close to each other ( the maximum proportional error of the coefficient @xmath116 is @xmath117 ) . in order to get
an analytic expression for free energy , some approximations to the polarization functions is unavoidable . in ref .
@xcite , the dominant contribution of polarization function comes from @xmath118 at both @xmath119 and @xmath120 regions ( after analytic continuation the momentum becomes @xmath121 ) , while in our calculation the dominant momentum region is @xmath122 ( @xmath123 in the continuous form ) .
for this reason , our analytic expression for the free energy differs from that of ref .
@xcite ( the approximation of @xmath124 might partly explain the difference ) .
after comparing the analytical results with numerical results , we found that our analytical result is slightly lower than the numerical result while the analytical result in ref .
@xcite is slightly greater than the numerical result . for @xmath125 ,
the analytical and numerical results for the coefficient are @xmath126 and @xmath127 respectively in our work and @xmath128 and @xmath129 respectively in ref .
@xcite .
the contribution of coulomb interaction to the free energy in the semimetal phase with @xmath78 is shown in fig .
[ fig : deltaf ] .
the free energy behaves as @xmath130 ( logarithmic correction ) for several different values of @xmath131 .
we now turn to the insulator phase where @xmath132 .
the free energy can be obtained by substituting eq .
( [ eq : pifull ] ) and eq .
( [ eq : pi_0 ] ) into eq .
( [ eq : delta_f ] ) .
the dependencies of specific heat and susceptibility on different fermion mass @xmath40 for @xmath133 are shown in fig .
[ fig : spheat_m4 ] and [ fig : sus_m4 ] , respectively .
the results for other choices of @xmath131 are similar and thus not shown . here
, we use the absolute values @xmath134 and @xmath135 , instead of @xmath136 and @xmath137 which are negative . from fig .
[ fig : spheat_m4 ] and [ fig : sus_m4 ] , we see that the fermion gap leads to remarkable suppression of the interaction correction to fermion specific heat and susceptibility . in summary , in the semimetal phase
the long - range coulomb interaction gives rise to non - fermi liquid behavior of specific heat and susceptibility . in the insulator phase , the fermion specific heat and susceptibility
are both significantly suppressed by the excitonic gap , but the total specific heat has a finite value due to the massless goldstone bosons . for different interaction strength @xmath131 .
the red line is the reference free energy @xmath138 .
it appears that free energy displays the same logarithmic behavior for different @xmath131.,width=259 ]
in this paper , we calculated the specific heat and susceptibility in graphene .
the ground state of graphene is semimetal when the coulomb interaction strength @xmath139 , but becomes insulator when @xmath140 .
the most prominent feature of semimetal phase is the appearance of logarithmic @xmath14-dependence of specific heat and susceptibility due to long - range coulomb interaction .
these are non - fermi liquid behaviors . in the insulating phase , because the interaction correction to fermion excitations is strongly suppressed by the excitonic gap , the total specific heat is solely determined by the contribution from goldstone bosons , while the susceptibility drops significantly . apparently , both specific heat and susceptibility manifest quite different behaviors in the two sides of the critical point @xmath0 . note that the semimetal and insulator phases both contain massless excitations : massless dirac fermion in the former and massless goldstone boson in the latter .
they have different statistics and exhibit completely different behaviors .
for example , the massless dirac fermions can transfer heat current and produce a universal thermal conductivity @xcite at @xmath141 , while the goldstone bosons only contribute a @xmath142 term , which vanishes rapidly as @xmath143 .
the massless dirac fermions also gives rise to a universal electric conductivity @xcite , although the predicted electronic conductivity is at invariance with experimental result ( the famous missing @xmath34 ) .
the goldstone bosons do not contribute to electric conductivity since they are neutral .
we should point out that the goldstone bosons are exactly massless only when the lagrangian respects a continuous chiral symmetry .
if the continuous chiral symmetry is explicitly broken by some contact four - fermion interaction , then the goldstone bosons are no longer strictly massless .
instead , they have a small mass as the result of dynamical breaking of appropriate continuous chiral symmetry @xcite . in this case , our discussion and calculation about the free energy contribution from goldstone bosons should be modified and a small mass should be included . in reality , there are various four - fermion interactions in the graphene @xcite . if the contact four - fermion interaction has the form @xmath144 , then the continuous chiral symmetry is not explicitly broken and the goldstone bosons are still massless . if the four - fermion interaction term is @xmath145 , then the system has only discrete chiral symmetry and there are no massless goldstone bosons @xcite . therefore , the specific heat of goldstone bosons presented in sec.[sec : clean ] is valid only when the continuous chiral symmetry is not explicitly broken by any four - fermion interaction term .
we finally comment on the validity of @xmath8 expansion .
the excitonic insulating transition requires the coulomb interaction between dirac fermions be sufficiently strong . in this strong coupling regime
, @xmath8 seems to be the only available expansion parameter , even if it is not small ( @xmath146 for graphene ) . in our specific case
, the fermion mass plays the dominant role in the insulator phase .
it suppresses significantly the coulomb interaction contribution to fermion specific heat .
this implies that , within the @xmath8 expansion , the next - to - leading order contribution could be neglected since it is much less than the leading order contribution .
it is reasonable to speculate that higher order corrections in @xmath8 expansion are also suppressed by the dynamical fermion mass . in the semimetal phase
, there is no such suppressing effect , so higher order corrections might be important . as shown in the context , the analytical calculation of next - to - leading order correction is already very complicated , including higher order corrections will make analytical calculation intractable .
the specific heat of massless dirac fermions may be analyzed by renormalization group approach @xcite , which found power - law @xmath147 behavior after summing up all orders of logarithmic corrections @xcite .
however , the exponent @xmath148 can only be calculated by performing @xmath8 expansion .
therefore , the validity of @xmath8 expansion also needs to be studied in this approach .
we thank g. cheng and j .- r .
wang for discussions .
w. l. is grateful to r. asgari for very helpful correspondence .
this work is supported by national science foundation of china under grant no .
10674122 . v. p.
gusynin , s. g. sharapov , and j. p. carbotte , phys .
96 ( 2006 ) 256802 ; v. p. gusynin and s. g. sharapov , phys .
b. 73 ( 2006 ) 254511 ; v. p. gusynin , v. a. miransky , s. g. sharapov , and i. a. shovkovy , phys .
b. 74 ( 2006 ) 195429 . | the strong coulomb interaction between massless dirac fermions can drive a semimetal - insulator transition in single - layer graphene by dynamically generating an excitonic fermion gap .
there is a critical interaction strength @xmath0 that separates the semimetal phase from the insulator phase .
we calculate the specific heat and susceptibility of the system and show that they exhibit distinct behaviors in the semimetal and insulator phases .
massless dirac fermion , semimetal - insulator transition , non - fermi liquid behavior 73.43.nq , 71.10.hf , 71.30.+h |
let us consider a portfolio of @xmath0 loans .
let the notional of loan @xmath1 be equal to the fraction @xmath2 of the notional of the whole portfolio .
this means that if loan @xmath1 defaults and the entire notional of the loan is lost the portfolio loses fraction @xmath2 or @xmath3 of its value . in practice
when a loan @xmath1 defaults a fraction @xmath4 of its notional will be recovered by the creditors . thus the actual loss given default ( lgd ) of loan @xmath1 is @xmath5 fraction or @xmath6 of the notional of the entire portfolio .
we now describe the gaussian m - factor model of portfolio losses from default .
the model requires a number of input parameters . for each loan
@xmath1 we are give a probability @xmath7 of its default .
also for each @xmath1 and each @xmath8 we are given a number @xmath9 such that @xmath10 .
the number @xmath9 is the loading factor of the loan @xmath1 with respect to factor @xmath11 .
let @xmath12 and @xmath13 be independent standard normal random variables .
let @xmath14 be the cdf of the standard normal distribution . in our model loan @xmath1 defaults if @xmath15 this indeed happens with probability @xmath7 .
the factors @xmath16 are usually interpreted as the state of the global economy , the state of the regional economy , the state of a particular industry and so on .
thus they are the factors that affect the default behavior of all or at least a large group of loans in the portfolio .
the factors @xmath17 are interpreted as the idiosyncratic risks of the loans in the portfolio .
let @xmath18 be defined by @xmath19 we define the random loss caused by the default of loan @xmath1 as @xmath20 where @xmath4 is the recovery rate of loan @xmath1 .
the total loss of the portfolio is @xmath21 an important property of the gaussian factor model is that the @xmath22 s are not independent of each other .
their mutual dependence is induced by the dependence of each @xmath22 on the common factors @xmath16 .
historical data supports the conclusion that losses due to defaults on different loans are correlated with each other .
historical data can also be used to calibrate the loadings @xmath9.the @xmath22 s are not independent of each other .
their mutual dependence is induced by the dependence of each @xmath22 on the common factors @xmath16 .
historical data supports the conclusion that losses due to defaults on different loans are correlated with each other .
historical data can also be used to calibrate the loadings @xmath9 .
when the values of the factors @xmath16 are fixed , the probability of the default of loan @xmath1 becomes @xmath24 the random losses @xmath22 become conditionally independent bernoulli variables with the mean given by @xmath25 and the variance given by @xmath26 by the central limit theorem the conditional distribution of the portfolio loss @xmath23 given the values of the factors @xmath16 can be approximated by the normal distribution with the mean @xmath27 and the variance @xmath28 then the joint distribution of the factors @xmath16 and the portfolio loss @xmath23 can be approximated by a distribution with density @xmath29 where @xmath30 stands for the gaussian density with mean @xmath31 and variance @xmath32 .
let @xmath33 . we define a tranche loss profile @xmath34 by @xmath35 number @xmath36
is called the attachment point of a tranche , while @xmath37 is called the detachment point of a tranche .
the expected loss of a tranche is then @xmath38 this can be rewritten as a double integral @xmath39 the inside integral with respect to @xmath23 is easily done analytically because @xmath23 has a simple normal distribution for fixed values of the factors @xmath16 .
the outside integral has to be computed numerically .
however , since it is an integral of a bounded smooth function with respect to m - dimensional gaussian density , it is one of the simpler integrals to compute numerically .
the expected loss of a tranche can be written as a double integral @xmath65 the inside integral with respect to @xmath23 can be done analytically for fixed values of the factors @xmath16 .
the outside integral has to be computed numerically .
however , since it is an integral of a bounded smooth function with respect to m - dimensional gaussian density , it is one of the simpler integrals to compute numerically .
in this section we apply the proposed algorithm to the single factor gaussian model of a portfolio with 125 names . we choose a 125 name portfolio because it is the size of the standard djcdx.na.ig portfolio .
we choose a single factor model because it is the one most frequently used in practice .
we evaluate the expected loss for four different tranches .
all tranches have attachment point @xmath40 or @xmath41 .
the detachment points are 3% , 7% , 10% and 15% .
we take these detachment points because they are the ones most frequently used in practice in order to evaluate the base correlation .
the parameters of the porfolio are @xmath42 in figure [ resultsfig ] we compare the expected loss computed using @xmath43 monte carlo samples with the expected loss computed using formula ( [ theformula ] ) .
the agreement between the two is good .
[ resultsfig ] to obtain the results in figure [ resultsfig ] we only needed to perform a single one dimensional numerical integration for each tranche .
this is an improvement over the moody s ft method @xcite which requires computing a large number of fourier transforms for each tranche .
each individual fourier transform is as computationally expensive as ( [ theformula ] ) .
when the approximation to the conditional distribution of @xmath23 given by the central limit theorem is deemed insufficiently accurate , an arbitrarily accurate representation of the conditional distribution of the portfolio loss @xmath23 can be obtained from its hermite series expansion . for historical reasons this expansion is also known as the charlier series expansion @xcite , @xcite .
let @xmath44 be the c.d.f . of the conditional distribution of the portfolio loss @xmath23 .
so that @xmath45 for each fixed value of the factors @xmath16 we define the normalized conditional loss @xmath46 by @xmath47 let @xmath48 be the c.d.f . of the distribution of the normalized conditional portfolio loss @xmath46 . so that @xmath49 we define the hermite polynomial @xmath50 of degree @xmath51 by @xmath52 let @xmath53 be defined by @xmath54 then we have @xmath55 the series above converges in the sense of distributions ( generalized functions ) @xcite .
a good reference on the theory of distributions ( generalized functions ) is @xcite .
let us pick a finite @xmath0 .
then we have @xmath56 as before the approximation is in the sense of generalized functions .
equation ( [ expn ] ) implies that the distribution of the normalized conditional portfolio loss @xmath46 can be approximated by a distribution with the density @xmath57 the function @xmath58 is not necessarily nonnegative and therefore may not be a probability density in the strict sense .
however , as is explained in @xcite , this does not affect the validity of our final result ( [ theformula ] ) .
therefore we may treat @xmath58 as a real probability density .
the distribution of the unnormalized loss @xmath23 can be approximated by a distribution with density @xmath59 the joint distribution of the factors @xmath16 and the portfolio loss @xmath23 can be approximated by a distribution with density @xmath60 where @xmath61 stands for the gaussian density with mean @xmath62 and variance @xmath63 .
observe that the coefficient @xmath53 depends only on the moments of the distribution @xmath48 .
since @xmath22 s are independent bernoulli random variables these moments are known analytically .
thus in the case under consideration all the @xmath53 s are known analytically .
if in equation ( [ den1 ] ) we set @xmath64 we obtain the normal density approximation proposed in the previous section .
we show later that it gives good numerical results even when the portfolio size is too small for the normal approximation to be accurate .
in this section we test the proposed algorithm on several portfolios of smaller size . for these portfolios
the approximation to the conditional distribution of the portfolio loss @xmath23 given by the central limit theorem is not very accurate , because of their small size .
however , the hermite expansion produces very good results .
we apply the proposed algorithm to the single factor gaussian model of a portfolio with @xmath51 names .
we take @xmath51 to be 25 ( size of dj itraxx australia ) , 30 ( size of dj itraxx ex japan ) , 50 ( size of dj itraxx cj ) and 100 ( size of djcdx.na.hy ) .
we choose a single factor model because it is the one most frequently used in practice . for each @xmath51
we compute the loss of the equity tranche with the attachment point @xmath40 or @xmath41 and the detachment point 3% .
the parameters of the porfolio are @xmath66 where @xmath67 .
finally , we choose @xmath68 in ( [ expn ] ) . in figure [ resultsfig2 ] we compare the expected loss computed using @xmath43 monte carlo samples with the expected loss computed using formula ( [ theformula2 ] ) .
the agreement between the two is good .
[ resultsfig2 ] to obtain the results in figure [ resultsfig2 ] we only needed to perform a single one dimensional numerical integration for each tranche .
this is an improvement over the fourier transform based methods @xcite which require computing a large number of fourier transforms for each tranche .
each individual fourier transform is as computationally expensive as ( [ theformula2 ] ) .
the hermite expansion ( [ expn ] ) can be used to achieve arbitrary accuracy when the normal approximation is insufficiently accurate .
the proposed algorithm is fast because the inside integral in ( [ theformula2 ] ) can be done analytically .
we also comment that the algorithm can be extended trivially to the case of non - constant recovery rates and recovery rates correlated with the state of the factor variables .
i thank my adviser a. chorin for his help and guidance during my time in uc berkeley .
i thank mathilda regan and valerie heatlie for their help in preparing this article .
i am also grateful to ting lei , sunita ganapati and george wick for encouraging my interest in finance .
last , but not least , i thank my family for their constant support . | we propose a fast algorithm for computing the expected tranche loss in the gaussian factor model with arbitrary accuracy using hermite expansions .
no assumptions about homogeneity of the portfolio are made .
it is intended as an alternative to the much slower fourier transform based methods @xcite . |
a key development in the study of globular clusters ( gcs ) in external galaxies has been the discovery that most large galaxies have bimodal gc color distributions ( e.g. , zepf & ashman 1993 ; forbes , brodie & grillmair 1997 ; gebhardt & kissler - patig 1999 ; kundu & whitmore 2001a ; larsen 2001 ) .
these are usually described as blue ( metal - poor ) and red ( metal - rich ) gc subpopulations , although additional substructure may be present .
the red gc system properties appear to be intimately tied to those of their parent galaxy , suggesting that the red gcs formed along with the bulk of the galaxy field stars ( forbes 1997 ; forbes & forte 2001 ; larsen 2001 ) . in both spirals and ellipticals
, they are thought to be associated with the bulge / spheroid component ( forbes , brodie & larsen 2001 ) .
the blue gcs are among the oldest and most metal - poor stellar systems observable .
therefore , they provide a probe of very early epochs of star formation in the universe . whether or not the properties of blue gcs correlate with
the mass of their parent galaxy has been controversial ( forbes 1997 , c^ ot ' e 2000 ; burgarella , kissler - patig & veronique 2001 ; forbes & forte 2001 ; larsen 2001 , lotz 2003 ) , and no clear demonstration of such a relation exists in the literature
. however , the issue is an important one in the context of gc and galaxy formation .
if a correlation exists , it implies that the blue gcs , or at least a significant proportion of them , `` knew '' about the galaxy to which they would ultimately belong .
this indicates that their formation was affected by the local conditions and that they may have already been contained within the dark matter halo of their final galaxy .
the detailed chemical and age structure within the blue gc systems of galaxies of various types and environments would then offer one of the few observational constraints on the properties of the protogalactic clouds which combined to build the galaxies that we observe today .
such a correlation would also rule out any formation mechanism whereby all metal - poor gcs form _
completely _ independently of a host galaxy ( e.g. , peebles & dicke 1968 ) .
our aim here is to consider galaxies over a large luminosity range , use only high - quality data , and analyze the photometry in a uniform manner . in this way we will reduce the random and systematic errors that could disguise the existence of a blue gc
host galaxy correlation .
in particular , we have added new data on the gc systems of dwarf and low - luminosity elliptical galaxies , and utilized a bayesian statistical method to find the peak of the blue gc color distribution .
we chose to explore a possible correlation between the mean @xmath4 color ( i.e. , the mode / peak of the gaussian distribution ) of metal - poor gc systems and the @xmath5 of their host galaxies , since most gc system studies have been carried out in the @xmath6 and @xmath7 bands . when using only optical colors the age - metallicity degeneracy is always a concern , but recent spectroscopic results suggest that , despite the presence of a small fraction of intermediate - age gcs in some early - type galaxies ( goudfrooij 2001 ; larsen 2003 ; strader 2003a ) , both metal - poor and metal - rich gc subpopulations appear to be very old ( @xmath8 gyr ) within model uncertainties ( e.g. , larsen 2002 ; larsen & brodie 2002 ) .
our sources for massive early - type galaxies were larsen ( 2001 ) and kundu & whitmore ( 2001a , 2001b ) , who determined the peaks of the blue and red gc subpopulations by fitting gaussians as part of the kmm routine ( ashman , bird , & zepf 1994 ) .
in addition , several nearby luminous spiral galaxies have enough blue gcs to have their mean colors accurately determined . these include the milky way and m31 ( harris 1996 , barmby 2000 ) as well as several sculptor group galaxies ( olsen 2004 ) .
our method ( see below ) utilizes photometry for individual gcs , and we derive the peaks and errors ourselves rather than just using those reported in the literature .
therefore only galaxies with high quality hubble space telescope data ( which has minimal contamination ) and for which we had access to the photometry are included . to probe the metal - poor gc systems of low - luminosity galaxies
, we also included local group dwarf galaxies in our sample .
these were primarily taken from the compilation of forbes ( 2000 ) , though we have used new spectroscopic metallicities for old lmc gcs ( beasley , hoyle , & sharples 2002 ) whenever possible .
the metallicities of fornax gcs were taken from the study in strader ( 2003b ) , and we have added ngc 4147 to the metal - poor sagittarius dwarf gcs ( bellazzini 2003 ) .
the local group sample was supplemented with the m81 dwarf ddo78 , which has one gc ( sharina , silchenko , & burenkov 2003 ) .
most of the dwarf gcs have spectroscopic and/or color - magnitude diagram ( cmd ) based metallicities ( presumably superior to those obtained from their @xmath4 colors ) , and these were converted into @xmath4 colors using the galactic relation of barmby ( 2000 ) .
we included only genuinely _ old _ gcs , excluding , for example , intermediate - age gcs in the magellanic clouds . while further detections of gcs in dwarf galaxies outside the local group have been claimed ( e.g. , in the m81 group ; karachentsev 2000 ) , we included only those whose identities
have been confirmed by spectroscopy .
finally , we note that since the majority of our sample galaxies are in groups or clusters , at present we can only claim to be exploring the existence of a correlation in these environments . for all galaxies with four or more gcs , we used bayesian markov chain monte carlo ( mcmc ) methods , implemented in the package winbugs ( spiegelhalter 2003 ) , to find the mean color of the blue gcs .
see gilks , richardon , & spiegelhalter ( 1996 ) and strader & brodie ( 2004 ) for more details on mcmc methods and for comparisons to classical methods , e.g. , maximum likelihood . for the luminous early - type galaxies , as well as the milky way and m31 , we fit a homoscedastic ( equal variance ) gaussian mixture model to the @xmath4 colors to estimate the values of the blue and red peaks , as well as credible posterior intervals ( the bayesian equivalent of confidence intervals ) around those values .
indeed , the accurate estimation of these posterior intervals is one of the main advantages of using bayesian method over the standard least - squares framework . in the remainder of the galaxies , which had at most eighteen blue gcs , a single gaussian was fit only to the metal - poor gcs ( taken to be those with [ fe / h ] @xmath9 ) . for galaxies with fewer than four gcs ,
the mean of the individual @xmath4 colors was used , with a fixed 0.05 mag ( equivalent to @xmath10 dex ) 1@xmath11 error as a conservative estimate of the actual error . using winbugs
, we fit a hierarchical linear model to the galaxy absolute magnitudes and mean gc colors .
we made the _ a priori _ choice of a linear model because such models in ( @xmath12 , @xmath4 ) observational space reflect power - law models in luminosity / metallicity ( @xmath13 , @xmath14 ) theoretical space .
our model has the following form : @xmath15 diffuse ( noninformative ) priors were placed upon @xmath16 , @xmath17 , and @xmath18 .
@xmath19 is the initial estimated blue peak for the @xmath20 galaxy given the priors on @xmath16 and @xmath17 .
cosmic variance and measurement errors are then hierarchically propagated to model the observed blue peak @xmath21 for each galaxy .
the resulting best fit weighted model is @xmath22 , which is significant at the @xmath23 level .
an unweighted model @xmath24 has a nearly identical slope at an even higher level of significance ( @xmath25 ) .
this suggests no significant bias is introduced by our calculated errors , and that the slope of the weighted model is not being defined by a small number of luminous galaxies with very accurately measured peaks .
the nonparametric spearman rank correlation test confirms the correlation between @xmath5 and mean @xmath4 for blue gcs without assuming a specific model , finding a probability @xmath26 that there is no monotonic relationship between the two variables .
we also separately fit weighted and unweighted models to only the spheroidal galaxies in our sample ( 42 of the total 53 galaxies ) , since the younger stellar populations in disk galaxies may complicate our use of @xmath5 as a tracer of galaxy mass .
the resulting models are @xmath27 and @xmath28 , respectively , and very similar to those for the full sample .
the `` universal '' slope of @xmath29 found above is consistent at the @xmath30 level with the results of larsen ( 2001 ) , @xmath31 , and of burgarella ( 2001 ) , @xmath32 .
the bottom part of figure 1 shows mean @xmath4 color for metal - poor gcs vs. parent galaxy @xmath5 .
while there are not enough disk galaxies to meaningfully derive a slope for these galaxies alone , an examination of figure 1 suggests that they may have systematically bluer metal - poor gc systems than spheroid - dominated galaxies : all eight disk galaxies with @xmath33 fall below the best - fit model line for the full sample .
the apparent offset at @xmath34 is @xmath35 mag , equivalent to just over @xmath36 dex in metallicity .
this difference is likely partially due to our use of @xmath5 as a tracer for galaxy mass , but an interpretation with a simple closed - box chemical evolution model also suggests that , at fixed luminosity , disk galaxies had a larger gas fraction than spheroidal galaxies at the time their metal - poor gcs formed .
this , in turn , implies that either ( i ) metal - poor gcs in disk galaxies are older than their spheroidal counterparts , or ( ii ) an initial phase of chemical enrichment ( to @xmath37 ) proceeded more quickly in ellipticals than in spirals . in a recent simulation of gc formation in a hierarchical cosmology ,
kravtsov & gnedin ( 2004 ) predict a relation close to @xmath38 for gcs in giant galaxies .
though they do not distinguish between gc subpopulations , at the redshift that they stop their simulation ( @xmath39 ) all of their gcs are metal - poor ( with [ fe / h ] @xmath40 ) , so a comparison to our relation is justified . for the purposes of this comparison , we convert the observed @xmath4 vs. @xmath5 relation into @xmath41 form using the galactic color - metallicity relation @xmath42 } + ( 1.15\pm0.02)$ ] ( barmby 2000 ) .
we find @xmath1 , a relation that is substantially shallower than the kravtsov & gnedin prediction .
could systematic statistical biases be responsible for our relation ?
first , we note that many of the possible sources of error apply _ only _ to the low - luminosity galaxies in our sample , e.g. , the conversion from spectroscopic and cmd - based [ fe / h ] to @xmath4 color .
these possible biases can be addressed by using a homogeneous , unaffected subset of our full sample . taking only those galaxies that have @xmath4 hst data and that show visual bimodality
, we find a best fit weighted model with a slope of 0.0098 , nearly identical to that obtained for the full sample .
this shows that the leverage of the dwarf galaxies in our sample is not responsible for the correlation , in contrast to the conclusions drawn by burgarella ( 2001 ) . nor can _ any _ systematic issue with the dwarf galaxies in our sample produce the observed relation .
one can then ask what systematic biases might affect our homogeneous subsample of luminous galaxies . such biases could occur in two ways : ( i ) the detected correlation is a product of selection effects of either the gcs in a given galaxy or the galaxies themselves , or ( ii ) the metal - poor gc peaks are being mismeasured in such a way as to produce our observed relation
. consider point ( i ) first : the galaxy subsample , while by no means carefully chosen as a true volume or magnitude - limited sample , comprises a good fraction of nearby luminous early - type galaxies , and it is reasonable to expect it to be generally representative of early - type galaxies in similar environments .
regarding the selection of gcs in a given galaxy , our relation might arise if both galaxy luminosity and metal - poor gc metallicity were related through a third confounding variable .
possibilities here include gc luminosity or galactocentric distance . despite many studies which have searched for a gc metallicity - luminosity relation ( e.g. , van den bergh 1996 )
, no such correlation has been observed in any galaxy to date .
furthermore , while gc systems as a whole have negative color gradients ( i.e. , become bluer at large radii ) , this appears to be solely due to the differing radial distributions of the red and blue subpopulations , with little or no radial variation in the colors of the subpopulations themselves ( e.g. , lee , kim , & geisler 1998 ; harris , harris , & mclaughlin 1998 ) .
thus , it does not appear likely that selection effects could have produced our relation . point ( ii ) above posits that our observed correlation is an artifact of the assumptions used to estimate the color distributions , namely , that the observed distributions can be well - described as a superposition of two gaussians with equal variance .
criticisms could be made of the assumptions of ( a ) bimodality , ( b ) gaussian distributions , or ( c ) homoscedasticity . since galaxies in this subsample
were selected to be _ visually _ bimodal , as well as generally exhibiting strong statistical evidence against unimodality ( see larsen 2001 ) , arguing against the assumption of bimodality requires instead that three or more subpopulations of gcs are present in the majority of the galaxies .
no convincing demonstrations of such multimodal systems have been made , even though some galaxies now have very deep hst data ( e.g. , m87 ; jord ' an 2002 ) .
assumption ( b ) can be motived _ a priori _ by the lognormal metallicity distributions predicted by a simple closed - box chemical evolution model , appealing to the central limit theorem , or to the dominance of gaussian photometric errors .
it could be argued that gc color distributions could have heavier tails than gaussians , and be better fit by , e.g. , student s @xmath43 distributions ( as the galactic gc luminosity function appears to be ; secker 1992 ) . as a test of this hypothesis , we fit the gc colors of m87 with a model very similar to that in sec . 2 , except that we used @xmath43 distributions ( instead of gaussians ) with the degrees of freedom as a free parameter .
the posterior distributions of this parameter had most of their mass at large values , suggesting that gaussian fits are indeed appropriate ( since a @xmath43 tends to a gaussian as the degrees of freedom tend to infinity ) . regarding assumption ( c ) of homoscedasticity , due to the degeneracy between variance and other parameters in mixture model fits , it is generally desirable assume a homoscedastic model unless there is _ a priori _ evidence against it ( e.g. , ashman 1994 ) . for our sample ,
the blue and red gcs have similar luminosity functions ( e.g. , larsen 2001 ) , so photometric errors could not be the source of unequal variances .
thus , any differences would have to be intrinsic .
the galactic gc system , the only one which has metallicities measured for nearly all clusters , appears to be homoscedastic within measurement errors ( c^ ot ' e 1999 ) .
in addition , any proposed heteroscedasticity would need to be finely tuned , with an ever larger fraction of metal - rich gcs `` wrongly '' assigned to the metal - poor peak with increasing galaxy luminosity , to have any prospect of producing our relation . as a simple numerical example , assume that all galaxies have a `` base ''
mean blue gc color of @xmath4 = 0.90 .
our relation indicates that the most luminous galaxies in our sample have blue gc colors @xmath44 and red gc colors @xmath45 . since the final peak is essentially a number - weighted average of the individual gc colors , and such galaxies have similar numbers of blue and red clusters , an unreasonable @xmath46 of the red gcs would have to be regularly misassigned to the blue peak to produce the colors we see for such galaxies ( this is in addition to the _ ad hoc _ nature of the tuning noted above ) . in sum
, we find no evidence that systematic statistical biases could have led to the observed correlation between mean metal - poor gc color and parent galaxy luminosity .
the classic fall & rees ( 1988 ) classification of gc formation scenarios into primary , secondary , and tertiary has in the last decade given way to a new triad : accretion ( e.g. , c^ ot ' e , marzke , & west 1998 ) , major merger ( e.g. , ashman & zepf 1992 ) , and _ in situ _ ( e.g. , forbes 1997 ) .
the primary focus of each of these formation scenarios is the gc systems of giant ellipticals ( ges ) .
however , much of the following discussion applies also to the formation of less - luminous ellipticals ( es ) and disk galaxies .
c^ ot ' e ( 1998 , 2000 , 2002 ) were able to produce bimodal metallicity distributions for the gc systems of luminous galaxies under the fundamental assumption that all galaxies have one intrinsic population of gcs , which depends on galaxy luminosity .
they used observations of a sample of dwarf and giant galaxies to find a `` primordial '' relation between gc metallicity and parent galaxy luminosity . by varying an assumed mass spectrum of protogalactic fragments
, they were able to produce a wide variety of gc metallicity distributions , some of which matched the bimodal form seen in most giant galaxies .
it should be noted that they assume dwarf galaxies in the local universe ( and especially in the local group ) are the surviving counterparts of these protogalactic fragments .
if instead these gaseous fragments merged themselves out of existence at @xmath47 the epoch of blue gc formation , then the model is essentially equivalent to an _ in situ _ scenario ( see below ) .
the c^ ot ' e ( 2002 ) primordial mean metallicity-@xmath12 relation ( see their figure 1 ) is @xmath48 .
this relation predicts that galaxies with @xmath49 will have one intrinsic population of gcs with @xmath50 . by contrast , the blue gc relation we derived above and the red gc relation from larsen ( 2001 ) predict peaks at @xmath51 and @xmath52 , respectively . in figure 1
we have overplotted red gc peaks for fifteen galaxies from larsen ( 2001 ) . to these
we have added four less - luminous galaxies ( which are in our blue sample as well ) for which we have remeasured the red peaks ourselves , using original data from larsen and kundu & whitmore ( 2001a , 2001b ) .
the dotted line is a new linear fit to the red peaks , weighted by the number of gcs in each peak .
the dashed line is the ct ' e relation . while the observations are marginally consistent with the primordial relation at the blue and red ends of the luminosity range , in the intermediate - metallicity region between @xmath53 and @xmath54 there is a clear , @xmath36 mag difference between the observed peaks and the primordial prediction .
thus , the c^ ot ' e primordial relation for a single intrinsic population of gcs appears to be less consistent with current observations than the existence of two separate ( blue and red ) gc metallicity luminosity relations .
c^ ot ' e ( 2002 ) showed that by using their primordial relation , a wide variety of gc color distributions could be produced .
an examination of their figure 1 suggests that , given appropriate input assumptions , the accretion model could reproduce the blue gc relation .
however , the ability of the accretion model to successfully reproduce our observed relation is critically dependent on the validity of the primordial relation . by contrast , a metal - poor gc galaxy luminosity correlation arises naturally under _ in situ _
scenarios for gc formation , e.g. , harris & pudritz ( 1994 ) and forbes ( 1997 ) . in such scenarios ,
enrichment of the interstellar medium is linked to the depth of the galactic potential , _ independent _ of the details of gc formation in a particular scenario . a blue gc
galaxy relation places limits on the degree to which an intrinsic population of metal - poor gcs in a luminous galaxy has been affected by accretion and/or mergers , with substantial implications for galaxy formation models ( see below ) .
a caveat is that the cosmic variance in our model ( 1@xmath11 scatter of @xmath35 mag ) does not rule out the formation of the metal - poor gc systems of ges like m49 and m87 from the accretion of several slightly less luminous galaxies with redder - than - average blue gcs .
however , our correlation implies that _ in the mean _ the blue gc systems of ges could not have formed primarily from the accretion of less luminous galaxies , especially @xmath55 dwarfs , whose blue gcs have mean @xmath56 .
the paucity of metal - poor field stars in the halos of luminous e galaxies ( e.g. , ngc 5128 ; harris & harris 2000 ) is another argument against significant accretion , since such stars would presumably accompany the accreted metal - poor gcs ( harris 2003 ) .
we are not arguing that accretion does not take place ; witness the existence of the sagittarius dwarf ( ibata , gilmore , & irwin 1994 ) , the recently discovered canis major galaxy ( martin 2004 ; forbes , strader & brodie 2004 ) , relic stellar streams in the halo of m31 ( ibata 2001 ) , and more distant systems ( e.g. , forbes 2003 )
. however , we do suggest that it is not the primary mechanism by which the gc systems of most luminous galaxies are built .
we note that a major merger scenario for the formation of ges is also difficult to reconcile with our results , since the mean colors of metal - poor gcs in typical sb / sc spirals like the milky way and m31 are bluer than those of ges ( see figure 1 ) .
the favored site for the formation of blue gcs in the major merger scenario is in pregalactic fragments ( ashman & zepf 1992 ) .
ashman ( 2003 ) states `` some variation in mean metallicity [ of metal - poor gcs ] would not rule out this option provided it did not correlate with properties of the current parent galaxy '' . as noted above for the accretion model , while cosmic variance would allow the occasional merger of two @xmath57 spirals with redder - than - average blue gcs , the existence of a global correlation suggests that major mergers are unlikely to be the dominant mechanism for the creation of the gc systems of ges . independent arguments in favor of this point have been made based upon the relative _ numbers _ of metal - poor gcs in spirals and ges ( see harris 2003 for a summary ) , though some recent work ( e.g. , rhode & zepf 2004 ) suggests that these `` specific frequency '' arguments may only pose a problem for the blue gc systems of ges .
we note that proponents of the merger scenario have themselves argued ( see , e.g. , the discussion in ashman & zepf 1998 ) that major mergers can not be the _ exclusive _ mechanism for for forming ges . _ in situ _ scenarios have been criticized with straw man comparisons to monolithic collapse models ( e.g. , eggen , lynden - bell , & sandage 1962 ) , though , for example , forbes ( 1997 ) argued that the initial collapse probably involved `` some chaotic merging of many small subunits '' as in the galactic formation model of searle & zinn ( 1978 ) .
this idea was reflected in simulations of gc formation by beasley ( 2002 ) .
one of the most significant problems with the _ in situ _ scenario is the lack of a definitive physical mechanism for the truncation of metal - poor gc formation .
an initial starburst , accompanied by the formation of metal - poor gcs , could conceivably ionize and/or remove a substantial mass of gas from inner regions of the protogalaxy , with the subsequent infall of enriched gas forming a population of metal - rich gcs and the bulk of the galaxy starlight ( forbes 1997 ) . however , this feedback mechanism would be increasingly ineffective in more massive halos , with little suppression of star formation expected for halos with virial velocities @xmath58 km / s ( dekel & woo 2003 ) .
thus , feedback is unlikely to be the source of the near - universal metallicity gap observed in the gc systems of massive galaxies with @xmath59 several hundred km / s .
cosmic reionization might also result in a rapid truncation of blue gc formation ( e.g. , santos 2003 ) , though some instead argue for reionization as the trigger for metal - poor gc formation ( cen 2001 ) . among the advantages of the reionization hypothesis
are its consistency with early structure formation theories , as well as the prediction of the observed differences in metal - poor gc specific frequency between ellipticals and spirals .
since ellipticals ( located in high - density peaks ) collapse earlier than spirals of similar mass , they have a larger fraction of their final mass assembled and available for gc formation before reionization . by a similar argument
, one might expect the metal - poor gc metallicity galaxy luminosity relation to have a steeper slope in denser environments .
a study comparing the blue gc systems of field and cluster galaxies might then offer the intriguing possibility of elucidating the relative roles of galaxy mass and environment in shaping the initial phase of local metal enrichment .
our findings also have implications for popular hierarchical structure formation models ( e.g. , kauffmann , white , & guiderdoni 1993 ; cole 2000 ; somerville , primack , & faber 2001 ) .
for example , semi - analytic models generically predict a substantial fraction of mass assembly at @xmath60 .
however , as already noted , relatively luminous es can not have accreted very many dwarf galaxies and their populations of gcs , since this would destroy our observed correlation .
the ages of metal - poor gcs in dwarf galaxies place lower limits on the possible epochs of significant accretion .
we note that our results do not constrain ( i ) the accretion of galaxies before the formation of their gc systems , ( ii ) the accretion of galaxies without gcs ( which typically have @xmath61 ) , or ( iii ) the formation of field galaxies .
however , if the ages of old metal - poor gcs in local group dwarfs ( @xmath2 gyr ; johnson 1999 ) are typical of those elsewhere ( an untested assumption ) , then the bulk of galaxy assembly in groups and clusters must have occurred at @xmath62 . for more likely mean gc ages of @xmath63 gyr , the primary epoch of mass assembly
is pushed to @xmath64 or higher .
winbugs code is available upon request from js .
we acknowledge support by the national science foundation through grant ast-0206139 and a graduate research fellowship ( js ) .
graeme smith , soeren larsen , michael beasley , michael pierce , and michael west provided valuable comments on the manuscript .
suggestions from the referee keith ashman significantly improved the paper .
we thank soeren larsen , arunuv kundu , and knut olsen for providing us with their data .
ashman , k. m. & zepf , s. e. 1992 , , 384 , 50 ashman , k. m. & zepf , s. e. 1998 , globular cluster systems ( cambridge : cambridge university press ) ashman , k. m. , bird , c. m. , & zepf , s. e. 1994 , , 108 , 2348 ashman , k. m. 2003 , extragalactic globular cluster systems , 3 barmby , p. , huchra , j. p. , brodie , j. p. , forbes , d. a. , schroder , l. l. , & grillmair , c. j. 2000 , , 119 , 727 beasley , m. a. , baugh , c. m. , forbes , d. a. , sharples , r. m. , & frenk , c. s. 2002 , , 333 , 383 beasley , m. a. , hoyle , f. , & sharples , r. m. 2002 , , 336 , 168 bellazzini , m. , ibata , r. , ferraro , f. r. , & testa , v. 2003 , , 405 , 577 burgarella , d. , kissler - patig , m. , & buat , v. 2001 , , 121 , 2647 cen , r. 2001 , , 560 , 592 cole , s. , lacey , c. g. , baugh , c. m. , & frenk , c. s. 2000 , , 319 , 168 c^ ot ' e , p. 1999
, , 118 , 406 c^ ot ' e , p. , marzke , r. o. , & west , m. j. 1998 , , 501 , 554 c^ ot ' e , p. , marzke , r. o. , west , m. j. , & minniti , d. 2000 , , 533 , 869 c^ ot ' e , p. , west , m. j. , & marzke , r. o. 2002 , , 567 , 853 de vaucouleurs , g. , de vaucouleurs , a. , corwin , h. g. , jr . , buta , r. j. , paturel , g. , & fouque , p. 1991
, third reference catalogue of bright galaxies ( new york : springer ) dekel , a. & silk , j. 1986 , , 303 , 39 dekel , a. & woo , j. 2003 , , 344 , 1131 eggen , o. j. , lynden - bell , d. , & sandage , a. r. 1962 , , 136 , 748 fall , s. m. & rees , m. j. 1988 , iau symp .
126 : the harlow - shapley symposium on globular cluster systems in galaxies , 126 , 323 fitzgibbons , g. l. 1990 , ph.d . thesis , univ .
florida , gainesville forbes , d. a. , brodie , j. p. , & grillmair , c. j. 1997 , , 113 , 1652 forbes , d. a. , masters , k. l. , minniti , d. , & barmby , p. 2000 , , 358 , 471 forbes , d. a. , brodie , j. p. , & larsen , s. s. 2001 , , 556 , l83 forbes , d. a. & forte , j. c. 2001 , , 322 , 257 forbes , d. a. , beasley , m. a. , bekki , k. , brodie , j. p. , & strader , j. 2003 , science , 301 , 1217 forbes , d. a. , strader , j. , & brodie , j. p. 2004 , aj , in press gebhardt , k. & kissler - patig , m. 1999 , , 118 , 1526 gilks , w. , richardson , s. , & spiegelhalter , d. 1996 , markov chain monte carlo in practice ( london : chapman and hall ) goudfrooij , p. , mack , j. , kissler - patig , m. , meylan , g. , & minniti , d. 2001 , , 322 , 643 harris , w. e. 1996 , , 112 , 1487 harris , w. e. 2003 , extragalactic globular cluster systems , 317 harris , w. e. & pudritz , r. e. 1994 , , 429 , 177 harris , g. l. h. & harris , w. e. 2000 , , 120 , 2423 harris , w. e. , harris , g. l. h. , & mclaughlin , d. e. 1998 , , 115 , 1801 ibata , r. a. , gilmore , g. , & irwin , m. j. 1994 , , 370 , 194 ibata , r. , irwin , m. , lewis , g. , ferguson , a. m. n. , & tanvir , n. 2001 , , 412 , 49 johnson , j. a. , bolte , m. , stetson , p. b. , hesser , j. e. , & somerville , r. s. 1999 , , 527 , 199 jord ' an , a. , c^ ot ' e , p. , west , m. j. , & marzke , r. o. 2002 , , 576 , l113 karachentsev , i. d. et al .
2000 , , 363 , 117 karachentsev , i. d. et al .
2003 , , 404 , 93 kauffmann , g. , white , s. d. m. , & guiderdoni , b. 1993 , , 264 , 201 kravtsov , a. & gnedin , o. 2004 , , submitted kundu , a. & whitmore , b. c. 2001 , , 121 , 2950 kundu , a. & whitmore , b. c. 2001 , , 122 , 1251 larsen , s. s. , brodie , j. p. , huchra , j. p. , forbes , d. a. , & grillmair , c. j. 2001 , , 121 , 2974 larsen , s. s. , brodie , j. p. , beasley , m. a. , & forbes , d. a. 2002 , , 124 , 828 larsen , s. s. & brodie , j. p. 2002
, , 123 , 1488 larsen , s. s. , brodie , j. p. , beasley , m. a. , forbes , d. a. , kissler - patig , m. , kuntschner , h. , & puzia , t. h. 2003 , , 585 , 767 lee , m. g. , kim , e. , & geisler , d. 1998 , , 115 , 947 lotz , j. , miller , b. , ferguson , h. , & stiavelli , m. 2003 , extragalactic globular cluster systems , 211 martin , n. f. , ibata , r. a. , bellazzini , m. , irwin , m. j. , lewis , g. f. , & dehnen , w. 2004 , , 348 , 12 olsen , k. , miller , b. , suntzeff , n. , schommer , r. , & bright , j. 2004 , aj , in press peebles , p. j. e. & dicke , r. h. 1968 , , 154 , 891 pierce , m. j. & tully , r. b. 1992 , , 387 , 47 pritchet , c. j. & van den bergh , s. 1999 , , 118 , 883 prugniel , p. & simien , f. 1996 , , 309 , 749 rhode , k. l. & zepf , s. e. 2004 , , 127 , 302 santos , m. r. 2003 , extragalactic globular cluster systems , 348 schlegel , d. j. , finkbeiner , d. p. , & davis , m. 1998 , , 500 , 525 searle , l. & zinn , r. 1978 , , 225 , 357 secker , j. 1992 , , 104 , 1472 sharina , m. e. , silchenko , o. k. , & burenkov , a. n. 2003 , , 397 , 831 somerville , r. s. , primack , j. r. , & faber , s. m. 2001 , , 320 , 504 spiegelhalter , d. , thomas , a. , best , n. , & lunn , d. , 2003 , _ winbugs user manual _ strader , j. , brodie , j. p. , schweizer , f. , larsen , s. s. , & seitzer , p. 2003
, , 125 , 626 strader , j. , brodie , j. p. , forbes , d. a. , beasley , m. a. , & huchra , j. p. 2003
, , 125 , 1291 strader , j. , & brodie , j. p. 2004 , in preparation van den bergh , s. 1996 , , 112 , 2634 zepf , s. e. & ashman , k. m. 1993 , , 264 , 611 | we demonstrate a significant ( @xmath0 ) correlation between the mean color of _ metal - poor _ globular cluster ( gc ) systems and parent galaxy luminosity .
a bayesian markov chain monte carlo method is introduced to find the mean color , and is easily generalizable to quantify multimodality in other astronomical datasets .
we derive a gc color galaxy luminosity relation of the form @xmath1 . when combined with evidence against a _ single _ primordial gc metallicity
galaxy luminosity relation for protogalactic fragments , the existence of such a correlation is evidence against both accretion and major merger scenarios as an explanation of the entire metal - poor gc systems of luminous galaxies .
however , our relation arises naturally in an _ in situ _ picture of gc formation , and is consistent with the truncation of metal - poor gc formation by reionization .
a further implication is that the ages of metal - poor gcs in dwarf galaxies constrain the main epoch of galaxy formation in hierarchical models .
if the ages of old metal - poor gcs in local group dwarfs ( @xmath2 gyr ) are typical of those in dwarfs elsewhere , then the bulk of galaxy assembly ( at least in clusters and groups ) must have occurred at @xmath3 , contrary to the predictions of some structure formation models . |
the electron / positron linac at kek injects 8-gev electron and 3.5-gev positron beams into kekb rings , where the cp - violation study is carried out . since the efficiency of the experiment
can be increased by shortening the injection time , several mechanisms have been introduced to accomplish this@xcite .
especially , much effort has been made to improve the positron injection time , since it is longer compared with that of electrons@xcite .
one of such effort is a two - bunch acceleration plan , which has been studied and applied@xcite . in this scheme
two bunches of positrons are accelerated in one rf pulse , which is 50 hz ; they may double the injection rate .
the time space between two bunches , however , is restricted by the rf frequencies of the linac and the rings , and the smallest space is 96.29 ns , since the common frequency is 10.38 mhz .
thus , a precise beam control and diagnosis are necessary .
the beam diagnosis used so far has been made by employing strip - line - type beam - position monitors ( bpm ) , wire scanners for transverse profiles and streak cameras for longitudinal profiles . in order to maintain stable beams , it is essential to have the beam instrumentations work for both of the two bunches .
the two - bunch read - out of bpms is especially important , because it is used in a number of orbit and energy feedback loops to stabilize the beams .
along the 600-m linac , 90 bpms are installed and their signals are transferred to one of 18 measurement stations .
signals are delayed and combined so as not to overlap each other , and are fed into a 5-gs / s waveform digitizer ( sony - tektronix tds-680b / c)@xcite , as in fig .
although the bpm signal is a fast bipolar , the readout precision is optimized using the interpolation function of the digitizer .
all 18 digitizers are triggered by a single distributed signal , which is synchronized with beam repetition and rf frequencies .
the waveform is read through the gpib , and a signal from each electrode is analyzed with a predetermined response function once per second by a vme computer ( force 68060 ) .
the response functions include 3rd - order position - mapping functions , attenuation factors of various components and position offsets against the center of the corresponding quadrupole magnet derived from a beam - based alignment .
since the timing and amplitude ranges of bpm signals are different depending on the beam modes and locations , the process is driven by a control database system@xcite .
the acquired beam positions at 18 stations are sent to central computers once per second and are served for various beam - energy and orbit feedback systems to maintain stable beam operation .
the bpm system was improved for two - bunch operation . as written above , it is important to acquire the beam positions of two bunches along the linac simultaneously to study the beams . in our instrumentation ,
signals from those two bunches appear as two signals separated by 96.29 ns on the waveform . although it was sometimes necessary to add more delay lines so as to avoid waveform overlapping , there was no need to add any specific hardware to handle such signals with small separations .
the calibration factors were re - examined since delay lines were added , and the beam - timing database for the signal analysis was extended to accommodate two - bunch information .
processing functions / commands for bpms on the central computers are also extended or added for two bunches , while keeping the old functions as before for single - bunch operations . with these modifications ,
the bpm processing system was extended for two - bunch operation without any performance loss in either precision and speed .
it has been used in beam operation since march , 2001 .
most of the operation software which utilize the bpm information was extended to meet both single- and two - bunch operations .
one of such examples is fig .
[ fig2 ] , which measures the beam energies of two bunches by correlation between a steering - magnet field and the beam - position response at the bunching section .
in order to accelerate the beams properly , the beam characteristics of two bunches need to be adjusted so as to be the same .
for example , in order to adjust the beam - energy differences , we change the beam timing and rf pulse timing .
the beam timing can be changed by 10-ps steps@xcite and the rf pulse timing can be changed by 1.75-ns steps at each sector independently .
most of other parameters in the linac are not sensitive against time separation of 96.29 ns . with such adjustments ,
the 10-nc primary electron bunches are accelerated up to 3.7 gev and positrons are generated as shown in fig .
[ fig3 ] .
the beam feedback loops in the linac for energy and orbit stabilization@xcite were also extended to control two - bunch beams .
since we do nt have many mechanisms to control two bunches independently , most feedback loops were modified to use positions derived from the charge - weighted averages of two bunches . with these changes
, those loops can maintain the average orbit and energy . in software ,
only the monitoring function was extended to read the average positions if two bunches are accelerated .
for positron injection , about 20 beam feedback loops are used , and they are all extended for two bunches . while normal energy and orbit feedback loops use charge - weighted average positions , feedback loops to minimize the energy differences use the position difference between two bunches , as shown in fig .
[ fig4 ] .
although the energy difference does not change frequently , such loops stabilize the beam over the long term .
the data - acquisition system for the linac bpms was upgraded to provide beam positions in two - bunch operation without losing any original features . along with improvements of the streak camera and wire scanner systems
, it has still been indispensable to study and operate on linac beams .
the system is also used by many operation software programs , including beam - energy and orbit feedback systems . | in order to double the positron injection rate into the kekb ring , a two - bunch acceleration scheme has been studied at the linac , in which bunches separated by 96 ns are accelerated in 50 hz . in this scheme stabilization of the energy and orbit of each bunch
is indispensable .
thus , the beam energy and orbit feedback systems have been upgraded .
since beam characteristics are acquired through beam - position monitors ( bpm ) , their read - out system was improved to meet two - bunch requirements . combined waveforms from bpm s
were adjusted with delay cables avoiding overlaps so as to enable the simultaneous measurements of the beam positions of two bunches .
the beam energies of two bunches were balanced by tuning the rf pulse timings , and the average energy was stabilized by adjusting the accelerating rf phases .
the average beam orbits were also stabilized .
slow feedback systems at the injector section for charge and bunching stabilities are being planned as well .
these systems were successfully used in the test beams and will be employed during routine operation . |
the jack symmetric polynomials @xcite are a system of orthogonal polynomials expressing the excited states of an integrable one - dimensional quantum many - body system with the trigonometric type potential called the calogero - sutherland model @xcite .
to be precise , the jack polynomials are eigenfunctions of a hamiltonian @xmath5 which is obtained by a certain transformation of the calogero - sutherland hamiltonian . here
@xmath6 is a parameter appearing in the calogero - sutherland model .
the excited states can be constructed from the jack polynomials . ]
these are one - parameter deformations of the schur symmetric polynomials . in general , being integrable means that the model has sufficiently many conserved quantities , and that system can be analytically solved . like the calogero - sutherland model , many of the integrable systems are not physical models of particles existing in the real world . however , the mathematical structure of the integrable models , e.g. , excellent solvability , can be used to advantage in many fields of mathematics . let us consider symmetric functions which are defined as a projective limit of symmetric polynomials with finite variables ( * ? ? ?
* chap . 1 ) .
in the case of the jack polynomials , the infinite - variable limit exists and is called the jack symmetric functions .
the jack functions are parametrized by partitions or young diagrams , and has the complex parameter @xmath6 ( see also footnote [ footnote : hbeta ] ) .
actually we can consider the parameter @xmath6 as an indeterminate , and then the jack functions are defined over the field @xmath7 .
the surprising result due to mimachi and yamada is that the jack functions associated to rectangular young diagrams have a one - to - one correspondence with singular vectors of the virasoro algebra @xcite .
the virasoro algebra is constructed by the infinitesimal conformal transformations in two dimensions , and is the lie algebra generated by @xmath8@xmath9 and the central element @xmath10 satisfying the relations @xmath11=(n - m)l_{n+m}+ c \ : \frac{n(n^2 - 1)}{12}\delta_{n+m,0 } , \quad n , m \in \mathbb{z},\ ] ] @xmath12=0,\quad n \in \mathbb{z}.\ ] ] this is an essential algebra to two - dimensional conformal field theories required for string theory and statistical mechanics . to obtain the irreducible representations of the virasoro algebra is important not only in representation theory but also in the conformal field theories .
the irreducibility of highest weight representations can be determined by special vectors called singular vectors in the highest weight representation .
although the singular vectors have an integral representation , the expression formula of the jack functions by the dunkl operator @xcite is more useful .
further , various properties of jack functions are known .
thus , the expression of the singular vectors by the jack functions is very convenient and beneficial . as a @xmath13-difference deformation of the jack polynomials
, there is a system of orthogonal polynomials with rich theory called the macdonald polynomials @xcite . for later use
let us introduce the notation for macdonald symmetric function , which is the infinite - variable version of the macdonald polynomial .
we denote by @xmath14 the macdonald symmetric function associated to the partition @xmath15 . here
@xmath13 and @xmath16 are free parameters , and they can be considered as complex numbers or indeterminates . in this paper , we regard the power sum symmetric functions @xmath17 as variables of the macdonald functions ( for more detail , see appendix [ sec : macdonald and hl ] ) .
the macdonald polynomials are also simultaneous eigen - functions of commuting @xmath13-difference operators , now called macdonald difference operators .
let us also mention that they are related to the ruijsenaars model @xcite which is a relativistic extension of the calogero - sutherland model .
the @xmath13-deformation like the macdonald functions makes theory clearer and often mathematically easier to handle .
for example , the jack functions can be characterized as the hamiltonian @xmath5 ( see footnote [ footnote : hbeta ] ) , but they have degenerate eigenvalues , and difficulties arise when we prove their orthogonality and coincidence with the singular vectors .
in the theory of the macdonald functions , this degeneracy problem can be eliminated and the discussion is clearer .
also , the hamiltonian @xmath5 has an infinite number of commuting operators .
however , it is difficult to write down these operators explicitly @xcite , and in the macdonald theory we have an explicit formula for the commuting family of difference operators having @xmath14 as simultaneous eigenfunctions . for the above reason
, it can be said that macdonald s theory is more beautiful . in the @xmath18 ( @xmath19 )
limit with @xmath6 fixed , the macdonald functions are reduced to the jack functions . on the other hand , in @xmath3 limit with
@xmath16 fixed , they are reduced to the symmetric functions called the hall - littlewood functions .
the hall - littlewood functions have a close connection to the character of the general linear group over finite fields , and they are also a generalization of the schur functions ( * ? ? ? * chap .
it is one of the advantages that it is possible to unify and generalize the two generalizations of the schur functions .
some applications in knot invariants @xcite and stochastic processes @xcite are also known .
the macdonald functions are one of the important symmetric functions for modern mathematics .
awata , kubo , odake and shiraishi introduced in @xcite a @xmath13-deformation of the virasoro algebra , which is named the deformed virasoro algebra .
this deformed algebra is designed so that singular vectors of verma modules correspond to macdonald symmetric functions @xmath14 .
the deformed virasoro algebra is an associative algebra defined over the base field @xmath20 , where @xmath13 and @xmath16 are the same parameters as in @xmath14 .
the generators are denoted by @xmath21 ( @xmath22 ) , and the defining relation is @xmath23 = -\sum_{l=1}^{\infty}f_l(t_{n - l}t_{m+l}-t_{m - l}t_{n+l } ) -\frac{(1-q)(1-t^{-1})}{1-p}(p^n - p^{-n } ) \delta_{n+m,0},\ ] ] where @xmath24 and @xmath25 are the structure constants defined by @xmath26 it is shown that the singular vectors of the deformed virasoro algebra coincide with the macdonald functions associated with rectangular young diagrams .
it is also possible to obtain the jack and macodnald functions associated with general partitions from the singular vectors of the @xmath27-algebra and the deformed @xmath27-algebra ( which is the ( deformed ) virasoro algebra when @xmath28 ) @xcite . to be exact , singular vectors of the ( deformed ) @xmath27-algebra can be realized by @xmath29 families of bosons under the free field representation . by a certain projection to one of these bosons
, we can obtain the jack ( or macdonald ) functions associated with young diagrams with @xmath29 edges ( see figure [ fig : youngdiag_n-1edges ] ) .
edges , width=226 ] * 1.2 . *
the representation theory of the virasoro algebra plays an essential role in the two - dimensional conformal field theories . in 2009 , while studying the low energy effective theory of m5-branes , alday , gaiotto and tachikawa discovered the correspondence between the correlation functions of two - dimensional conformal field theories and the partition functions of four - dimensional supersymmetric gauge theories ( agt conjecture ) @xcite .
gauge theory has a long history and is an attractive theory studied by a lot of mathematicians and physicists .
although it is difficult to calculate the partition functions of gauge theories in general , nekrasov gave an explicit formula ( nekrasov formula ) for the instanton partition function of four - dimensional @xmath30 supersymmetric gauge theory in 2002 @xcite .
the nekrasov formula @xmath31 is written by the summation of the terms @xmath32 parametrized by tuples of young diagrams : @xmath33 these terms are given in a factorized form , and as @xmath34 increases , the amount of calculation becomes enormous .
however , it can be calculated by a simple combinatoric method .
the discovery of @xcite is the following relation between two - dimensional and four - dimensional field theories . the nekrasov formula for the four - dimensional @xmath35 gauge theory with four matters in ( anti-)fundamental representation ( actually , it is the nekrasov formula of the @xmath36 gauge theory divided by the @xmath2 factor @xmath37 ) coincides with the four - point conformal block of the two - dimensional conformal field theory .
basics of the conformal field theories were established by belavin , polyakov and zamolodchikov ( bpz ) in 1984 @xcite .
they described the critical phenomenon of the two - dimensional ising model which is a model of the ferromagnet , and so on .
the primary fields @xmath38 are operators on the representation space of the virasoro algebra such that @xmath39=z^n\left ( z\frac{\partial}{\partial z } + h(n+1 ) \right ) v(z ) , \quad z , h\in \mathbb{c}.\ ] ] the primary fields are the main research object in the conformal field theories . here
@xmath40 is called the conformal dimension of the primary field . furthermore , in the conformal field theories , it is a fundamental problem to calculate the correlation functions of the primary fields . generally , in the quantum field theories ,
the calculations of correlation functions are difficult , and usually it is often solved by approximation .
bpz succeeded in determining the exact forms of correlation functions in the conformal field theories .
in particular , they derived differential equations with regular singularities for the correlation functions .
however , the research by bpz was performed mainly for primary fields with the special conformal dimension , i.e. the minimal models , and they did not investigate the correlation functions in general forms . even if we derive the differential equations of the correlation functions , it is difficult to find their solutions . from the standpoint of conformal field theories , the agt conjecture that states the agreement between the nekrasov formulas and the conformal blocks ( originally in the liouville theory , that is the theory having the primary field with generic conformal dimensions ) is studied under the expectation that general formulas for the correlation functions can be obtained .
various extensions were made immediately after the agt conjecture was discovered .
first of all , the original agt conjecture deals with the four - dimensional gauge theory in the case that the number of ( anti-)fundamental matters @xmath41 is @xmath42 .
immediately after this original conjecture @xcite , the cases with @xmath43 were studied in @xcite .
these cases can be obtained from the case of @xmath44 by applying the same degenerate limits to the nekrasov formula and the conformal block .
especially when @xmath45 , the conformal block degenerates to the inner product of the vector @xmath46 called the whittaker vector of the virasoro algebra .
case , the degenerate conformal blocks can be realized by the inner product of certain vectors that are the general form of the vector @xmath46 .
] moreover , it is also expected that the four - dimensional gauge theories with the higher gauge group @xmath47 correspond with the @xmath27-algebra @xcite . the jack functions and the macdonald functions also play an important role in the agt conjecture .
for example , the expansion coefficients of the whittaker vector @xmath46 by the jack functions are clarified @xcite . in addition , it is known that a good basis called aflt basis @xcite can be regarded as a sort of generalization of the jack functions .
the aflt basis is a basis in the representation space of the algebra @xmath48 , which is first introduced by alba , fateev , litvinov and tarnopolskiy , and the conformal block can be combinatorially expanded by this basis .
the aflt basis is an orthogonal basis which parametrized by pairs of young diagrams @xmath49 . in the @xmath27 algebra case ,
it is parametrized by @xmath0-tuples of young diagrams and exists in the representation space of the algebra @xmath50 . by inserting the identity @xmath51 with respect to the aflt basis @xmath52
, the calculation of correlation functions @xmath53 is attributed to that of the matrix element @xmath54 , where @xmath38 is a sort of the primary field defined by some relations with generators of the virasoro algebra and the heisenberg algebra .
then the three - point functions @xmath55 are factorized and coincide with the significant factors called the nekrasov factors , which compose the nekrasov formula .
namely , if we expand the correlation functions by using the aflt basis , then the form of its expansion is quite the same as that of the nekrasov formula ( [ eq : nek formula ] ) .
further , the conformal block of the algebra @xmath56 coincides with the partition function @xmath31 of @xmath36 gauge theory .
actually , such a good basis does not exist in the representation space of the virasoro algebra .
since the @xmath2 factor contributes and complicates the agt conjecture , we need the adjustment by the heisenberg algebra .
since the aflt basis correspond to the torus fixed points in the instanton moduli space , it is also called the fixed point basis . in @xcite ,
the original agt conjecture is `` proved '' with the help of the aflt basis ( the generalized jack functions ) and the free field representation . in @xcite by using zamolodchikov reccursion relation . some proofs from geometric representation theory
are also given in @xcite .
] however , this `` proof '' is based on another conjecture .
to explain it in more detail , recall that the free field representation of the conformal blocks can be written by the dotsenko - fateev integral @xmath57 , where @xmath58 means some integrals of the integrand @xmath59
. then @xmath59 can be expanded by a sum of the products of the generalized jack functions @xmath60 and their dual functions @xmath61 , which are parametrized by tuples of young diagrams .
this expansion formula is called the cauchy formula . at that time , it was conjectured that the integral value of each term @xmath62 directly corresponds to @xmath32 in the nekrasov formula .
this is the scenario of the `` proof . ''
although this proof is straightforward without using recurrence formulas etc , since the integral value of the generalized jack functions is still a conjecture , it is necessary to prove it in order to complete this proof .
for that , we need to investigate more properties of the generalized jack functions
. @xmath13-deformed version of the agt conjecture is also provided .
that is , the deformed virasoro/@xmath1-algebra is related to five - dimensional gauge theories ( 5d agt conjecture ) @xcite . in the simplest case
, it is shown that the inner product of the whittaker vector of the deformed virasoro algebra coincides with the instanton partition function ( k - theoretical partition function ) of the five - dimensional @xmath63 pure @xmath36 gauge theory .
also the same approach as @xcite is taken in the @xmath13-deformed case .
in other words , it is conjectured that the @xmath13-deformed dotsenko - fateev integral corresponds to the partition function with @xmath44 matters , and this conjecture is checked by using the generalized macdonald functions @xcite .
the @xmath13-deformed version of the aflt basis @xcite ( that is , the generalized macdonald functions ) exists in the representation space of the level @xmath0 representation of the ding - iohara - miki algebra ( dim algebra ) . the dim algebra ( explained in appendix [ sec : def of dim ] ) has the face of a @xmath13-deformation of the @xmath64 algebra as introduced by miki in @xcite , and the deformed virasoro/@xmath1-algebra appear in its representation @xcite .
since the dim algebra has a lot of background , there are a lot of other names such as quantum toroidal @xmath65 algebra @xcite , quantum @xmath64 algebra @xcite , elliptic hall algebra @xcite and so on .
the dim algebra has a hopf algebra structure which does not exist in the deformed virasoro/@xmath1-algebra , and the dim algebra is associated with the macdonald functions having rich theory .
that is free field representation of the macdonald s difference operator , it is discovered that @xmath66 form a part of representation of the dim algebra @xcite .
see also fact [ fact : lv
. 1 rep of dim ] .
] unlike the case of the generalized jack functions , the generalized macdonald functions can be constructed by the coproduct of the dim algebra @xcite .
it is a surprising phenomenon that the structure of the coproduct of the dim algebra has information on the partition functions of the five - dimensional gauge theories .
furthermore , in the @xmath13-deformed case , awata - kanno s and iqbal - kozkaz - vafa s refined topological vertices @xcite are also reproduced by the matrix elements of some intertwining operator of the dim algebra , and the coincidence between the correlation function of the dim algebra and the 5d nekrasov formula is proved @xcite .
the agt conjecture with respect to the @xmath13-deformed aflt basis @xcite ( recalled in section [ sec : reargument of di alg and agt ] ) is almost parallel to the undeformed case , and it suffices to consider the algebra @xmath67 , denoted by @xmath68 , which is generated by certain operators @xmath69 ( @xmath70 , @xmath9 ) obtained by the level @xmath0 representation of the dim algebra .
the level @xmath0 representation is that on a fock module @xmath71 with the highest weight @xmath72 . the vertex operator @xmath73 on this fock module is defined by the relation ( definition [ df : dim vertex op ] ) @xmath74 where @xmath75 .
@xmath76 can be regarded as an analog of the virasoro primary field .
the generalized macdonald functions are defined to be the eigenfunctions of the generator @xmath77 constructed by the copoduct of the dim algebra .
then , it is conjectured that the matrix elements of @xmath76 with respect to the generalized macdonald functions reproduce the five - dimensional nekrasov factors . under this conjecture
, the four - point conformal block of @xmath76 corresponds to the 5d @xmath36 nekrasov formula with @xmath78 matters . *
* the first main theorem in this thesis is the formula for the kac determinant of the algebra @xmath68 ( theorem [ thm : kacdet ] ) : @xmath79 where @xmath80 , @xmath81 , and @xmath82 denotes the number of the @xmath0-tuples of young diagrams of the size @xmath34 . for the definition of @xmath83 , see notations in the latter part of this section .
this determinant can be proved by using the fact that the generators @xmath69 can be decomposed into the deformed @xmath1-algebra part and the @xmath2 part by a linear transformation of the bosons , and using the screening currents of the deformed @xmath1-algebra . by this formula
, we can solve the conjecture ( * ? ? ? * conjecture 3.4 ) that the following pbw type vectors of the algebra @xmath68 ( definition [ df : ordinary pbw vct ] ) are a basis : @xmath84 we also discover that singular vectors of the algebra @xmath68 correspond to some generalized macdonald functions as the second main theorem . by this result
, we can get singular vectors from generalized macdonald functions
. can express all singular vectors of the algebra @xmath68 is incompletely understood .
however , the kac determinant can be proved by the only vanishing points given by the singular vectors @xmath85 ( see ( [ eq : sing vct chi(i)rs ] ) ) corresponding to the simple roots , because the determinant has @xmath86 weyl group invariance .
] the singular vectors are intrinsically the same as those of the deformed @xmath1-algebra . however , as the projection of the bosons is necessary for the correspondence with the ordinary macdonald functions , the result of this thesis that does not need projections can be thought to be a generalization of @xcite . as a corollary of this fact
, we can find a new relation of the ordinary macdonald functions and the generalized macdonald functions by the projection of the bosons .
furthermore , since screening operators are written by integrals , we can also get an integral representation of generalized macdonald functions . concretely
, the vector @xmath87 defined to be @xmath88 is a singular vector . here @xmath89 denotes the screening operator , the @xmath0-tuple of parameters @xmath90 is a function of @xmath91 , and for non - negative integers @xmath92 , @xmath93 , @xmath94 ( for more details , see section [ sec : sing vct and gn mac ] ) .
the singular vector @xmath87 coincides with the generalized macdonald function with the @xmath0-tuple of young diagrams in figure [ fig : youngdiag_onlyrightside ] ( theorem [ thm : sing vct and gn mac 2 ] .
( main theorem ) ) .
in fact , figure [ fig : youngdiag_n-1edges ] means the same young diagram being on the rightmost side in figure [ fig : youngdiag_onlyrightside ] .
hence the projection of this generalized macdonald function corresponds to the ordinary macdonald functions associated with the rightmost young diagram with @xmath29 edges in figure [ fig : youngdiag_onlyrightside ] ( corollary [ cor : projection of gn mac ] ) .
when the condition @xmath95 for the number of screening currents and parameter @xmath91 in @xmath87 is removed , the above figure is not a young diagram .
however it turns out that the vector @xmath87 coincides with the generalized macdonald function obtained by cutting off the protruding part and moving boxes to the young diagrams on the left side .
for example , if @xmath96 for all @xmath97 , the corresponding @xmath0-tuple of young diagrams of the generalized macdonald function is figure [ fig : youngdiag_somerectangle ] ( theorem [ thm : sing vct and gn mac 2 ] .
( main theorem ) ) . * 1.4 . *
furthermore , we investigate behavior in the limit to the hall - littlewood functions , @xmath3 , of the deformed virasoro algebra and the algebra @xmath68 .
also 5d agt conjecture is studied in this limit .
the reason of considering such a limit is that the situation becomes simple and some problems are solved .
in particular , the simplest 5d agt conjecture can be proved , and pbw type vectors can be expressed in terms of hall - littlewood functions . by virtue of the theory of hall - littlewood functions ,
we can obtain and prove an explicit formula ( theorem [ thm : main theorem ] ) for the four - point correlation function of a certain operator @xmath98 , which is the limit @xmath3 of the vertex operator @xmath76 associated with @xmath99 . here , @xmath71 is the fock module with the highest weight @xmath100 .
@xmath101 here for a partition @xmath15 , @xmath102 , and @xmath103 is the same one in ( [ eq : kac det in intro ] ) .
the function @xmath104 can be calculated by the generalized hall - littlewood functions in the same way as @xcite .
however , we can obtain this formula by inserting the identity with respect to the pbw type vectors .
we call this hall - littlewood limit @xmath3 ` crystallization ' after the use of the quantum groups @xcite , where the parameter @xmath13 represents the temperature in the rsos model @xcite which has symmetry of the deformed virasoro algebra , and the limit @xmath105 can be considered as the zero temperature limit . although our studies are mathematically different from the notion of the original crystal base of quantum groups , the physical meaning and the motivation to simplify phenomena are the same . to investigate their mathematical relationship
is an interesting open problem . on the other hand , little
is known about the physical meaning of the hall - littlewood limit in the gauge thoery at present .
* in this thesis , the r - matrix of the dim algebra is also investigated . the result with respect to the r - matrix
is based on the collaborative researches @xcite , and only works of the author is described . in general , a r - matrix is defined as a solution of the yang - baxter equation , and is closely related to the solvable lattice models , the knot invariants and so on .
further , it is well - known that r - matrices can be constructed by hopf algebras such as the quantum groups . in general , a hopf algebra @xmath106 with the coproduct @xmath107 is called quasi - cocommutative if there exists an invertible element @xmath108 in the algebra @xmath109 such that @xmath110 this @xmath108 is called the universal r - matrix .
if @xmath108 also satisfies the relations @xmath111 ( see definition of @xmath112 in section [ sec : r - matrix ] ) then @xmath106 is called quasi - triangular and @xmath108 satisfies the yang - baxter equation @xmath113 .
the dim algebra is known to be quasi - triangular @xcite . in this thesis ,
the representation matrix of the universal r - matrix @xmath108 is explicitly calculated . in the tensor product of the level 1 representation of the dim algebra ( we denote it by @xmath114 ) , it is block - diagonalized at each level of the free boson fock space .
also , it can be seen that the action of @xmath108 on the generalized macdonald functions corresponds to the exchange of spectral parameters , partitions , and variables in the generalized macdonald functions . moreover , by using the renormalized generalized macdonald functions ( the integral form @xmath115 ) , it can be conjectured that @xmath116 where @xmath117 is the vector obtained by exchanging partitions , variables and spectral parameters in @xmath118 ( see definitions in section [ sec : explicit cal of r ] ) . as a consequence
, we have conjecture ( conjecture [ conj : r - matrix by int form ] ) of the explicit formula for the representation matrix @xmath119 of the universal r - matrix in the basis of @xmath118 : @xmath120 in @xcite , the rtt relation of the dim algebra is also studied using this r - matirx .
this thesis is organized as follows . in section
[ sec:5d agt ] , two examples of the 5d agt conjecture are reviewed .
one is the correspondence between the whittaker vector of the deformed virasoro algebra and the partition function of the 5d pure gauge theory .
the other is the conjecture on the aflt basis using the level @xmath0 representation of the dim algebra . in section [ sec : kac det and sing vct ] , we give a factorized formula for the kac determinant of the algebra @xmath68 .
its proof depends on some results of the deformed @xmath1-algebra .
the relationship between the singular vectors and the generalized macdonald functions is also revealed . in section [ sec : crystallization ] , we investigated the @xmath105 limit of the deformed virasoro algebra , the algebra @xmath68 and the 5d agt conjecture .
in particular , the simplest 5d agt conjecture is proved in this limit . in section [ sec : r - matrix ] , the explicit form of the representation of the universal r - matrix of the dim algebra is calculated .
its general form is also conjectured in terms of the generalized macdonald functions . in section [ sec : properties of gn mac ] , properties of the generalized macdonald functions are studied .
first , to state the existence theorem of the generalized macdonald functions , we need partial orderings among @xmath0-tuples of partitions . in this thesis , by using the partial orderings @xmath121 ( see definition [ def : ordering1 ] ) and @xmath122 ( see definition [ df : ordering elaborated version ] ) , the existence theorem is proved . however , in @xcite , another ordering @xmath123 is used and the proof of existence theorem ( * ? ? ?
* proposition 3.8 ) is omitted .
we justify the theorem ( * ? ? ?
* proposition 3.8 ) by comparing @xmath122 and @xmath123 in subsection [ sec : partial orderings ] . in subsection
[ sec : realization of rank n rep ] , we also investigate the action of the generators @xmath124 and higher rank hamiltonians on the generalized macdonald functions .
their actions are based on a conversion rule called spectral duality that exchanges the level @xmath0 representation and the rank @xmath0 representation of the dim algebra .
furthermore , in subsection [ sec : limit to beta ] , the @xmath125 limit is also studied . since
the generalized jack functions have degenerate eigenvalues , their cauchy formula used in the senario of proof of the agt conjecture @xcite is non - trivial . by taking the limit from the macdonald functions
, we can justify the orthogonality of the generalized jack functions and show the cauchy formula . in appendix
[ sec : macdonald and hl ] , the definition and basic facts of the ordinary macdonald functions and the hall - littlwood functions are briefly reviewed following @xcite . in appendix
[ sec : def of dim ] , the definition of the dim algebra and the level @xmath0 representation are explained following mainly @xcite .
moreover we also describe the definition of another representation of the dim algebra called level @xmath126 representation or the rank @xmath0 representation . in appendix
[ sec : proofs and chekc of crystal ] , we present some proofs and checks of conjectures in section [ sec : crystallization ] . at last in appendix [ sec : ex of r - matrix ] , we give explicit examples of r - matrix at level 2 .
* @xmath127 denote the set of positive integers , integers , rational numbers , real numbers , complex numbers , respectively .
* @xmath128 denotes the set of non - negative integers . *
@xmath129 denotes the set of integers except @xmath130 . *
@xmath131 denotes the kronecker delta . *
@xmath132 $ ] denotes the ring of polynomials in @xmath133 over a field @xmath134 . *
@xmath135 denotes the cardinality of set . * functions
@xmath136 depending on multiple variables @xmath137 ( @xmath138 ) are occasionally written as @xmath139 or @xmath140 for abbreviation . * for a partition @xmath15 , @xmath141 and @xmath142 denote the power sum symmetric function and the monomial symmetric function , respectively . * for @xmath143
, @xmath144 denotes the elementary symmetric function .
let us explain the notation of partitions and young diagrams .
a partition @xmath145 is a non - increasing sequence of integers @xmath146 .
we write @xmath147 .
the length of @xmath15 , denoted by @xmath148 , is the number of elements @xmath149 with @xmath150 .
partitions are identified if all elements except @xmath130 are the same .
for example , @xmath151 .
@xmath83 denotes the number of elements that are equal to @xmath152 in @xmath15 , and we occasionally write partitions as @xmath153 .
for example , @xmath154 .
the partitions are identified with the young diagrams , which are the figures written by putting @xmath149 boxes on the @xmath152-th row and aligning the left side .
for example , if @xmath155 , its young diagram is .
the conjugate of a partition @xmath15 , denoted by @xmath156 , is the partition whose young diagram is the transpose of the diagram @xmath15 .
for example , the conjugate of @xmath155 is @xmath157 . for a partition @xmath15 and a coordinate @xmath158 , define @xmath159 @xmath160
is called arm length and @xmath161 is called leg length . in the diagram , they mean the numbers of boxes in right side from or below the box being in the @xmath152-th row and @xmath162-th column .
for example , if @xmath163 , then @xmath164 , @xmath165 .
note that they can take negative values as @xmath166 , @xmath167 . for a partition @xmath15
, we define @xmath168 .
this means the sum of the numbers obtained by attaching a zero to box in the top row of the young diagram of @xmath15 , a @xmath169 to each box in the second row , and so on . for @xmath0-tuple of partitions @xmath170 ,
define @xmath171 .
if @xmath172 , we occasionally use the symbol `` @xmath173 '' as @xmath174 .
the author would like to express his deepest gratitude to his supervisor hidetoshi awata for a great deal of advice . without his guidance and persistent help , this thesis would not have been possible .
the author shows his greatest appreciation to hiroaki kanno for his insightful comments and suggestions , and h. fujino , t. matsumoto , a. mironov , al .
morozov , and .
morozov and y. zenkevich for the collaborative researches .
some of the results in this thesis are based on the collaborations with them .
the author also would like to thank m. hamanaka , k. iwaki , t. shiromizu , s. yanagida and friends for valuable discussions and supports .
the author is supported in part by grant - in - aid for jsps fellow 26 - 10187 .
we start with recapitulating the result of the whittaker vector of the deformed virasoro algebra and the simplest five - dimensional agt correspondence .
let @xmath13 and @xmath16 be independent parameters and @xmath175 .
the deformed virasoro algebra is the associative algebra over @xmath20 generated by @xmath21 ( @xmath22 ) with the commutation relation @xmath176 = -\sum_{l=1}^{\infty}f_l(t_{n - l}t_{m+l}-t_{m - l}t_{n+l } ) -\frac{(1-q)(1-t^{-1})}{1-p}(p^n - p^{-n } ) \delta_{n+m,0},\ ] ] where the structure constant @xmath177 is defined by @xmath26 the relation ( [ comm.rel.of qvir ] ) can be written in terms of the generating function @xmath178 as @xmath179,\ ] ] where @xmath180 .
the deformed virasoro algebra is introduced in @xcite .
let @xmath181 be the highest weight vector such that @xmath182 , @xmath183 ( @xmath184 ) , and @xmath185 be the verma module generated by @xmath181 .
similarly , @xmath186 is the vector satisfying the condition that @xmath187 , @xmath188 ( @xmath189 ) .
@xmath190 is the dual module generated by @xmath186 .
the pbw type vectors @xmath191 for partitions @xmath15 form a basis over @xmath185 .
also , @xmath192 form a basis over @xmath190 . here
@xmath193 is a partition or a young diagram .
the bilinear form @xmath194 is uniquely defined by @xmath195 .
this bilinear form is called the shapovalov form .
the whittaker vector @xmath196 is defined as follows . for a generic parameter @xmath197 ,
define the whittaker vector is also called the gaiotto state or the irregular vector . ]
@xmath196 by the relations @xmath198 similarly , the dual whittaker vector @xmath199 is defined by the condition that @xmath200 this vector is in the form @xmath201 and its norm is calculated as @xmath202 , where @xmath203 denotes the inverse matrix element of the shapovalov matrix @xmath204 .
it is useful to consider the free field representation of the deformed virasoro algebra . by the heisenberg algebra generated by @xmath137 ( @xmath9 ) and @xmath205 with the relations @xmath206=n\frac{1-q^{|n|}}{1-t^{|n| } } \delta_{n+m,0 } ,
\qquad [ a_n,{q}]=\delta_{n,0},\ ] ] the generating function @xmath207 can be represented as @xmath208 here @xmath209 .
let @xmath210 be the highest weight vector in the fock module of the heisenberg algebra such that @xmath211 ( @xmath212 ) , and @xmath213 . then @xmath214 .
furthermore , @xmath215 can be regarded as the highest weight vector @xmath181 of the deformed virasoro algebra with highest weight @xmath216 . in @xcite ,
awata and yamada conjectured an explicit formula for @xmath196 in terms of macdonald functions under the free field representation , and yanagida proved it in @xcite .
the simplest five - dimensional agt conjecture is that the inner product @xmath217 coincides with the five - dimensional ( k - theoretic ) nekrasov formula for pure @xmath35 gauge theory @xcite : @xmath218 where @xmath160 and @xmath161 are the arm length and the leg length defined in introduction , and @xmath156 is the conjugate of @xmath15 .
[ fact : simpleagt ] for @xmath219 , @xmath220 this fact is conjectured in @xcite and proved in @xcite when the parameter @xmath13 is generic .
we now turn to the dim algebra @xcite .
let us recall the aflt basis in the 5d agt correspondence of the @xmath47 gauge theory along @xcite .
in this section , we use @xmath0 kinds of bosons @xmath221 ( @xmath222 , @xmath223 ) and @xmath224 with the relations @xmath225= n \frac{1-q^{|n|}}{1-t^{|n| } } \ , \delta_{i , j } \ , \delta_{n+m , 0},\ ] ] @xmath226=0 , \quad
[ u_i , u_j]=0 , \qquad ( \forall i , j , n).\ ] ] here @xmath224 is the substitution of zero mode @xmath227 , which is realized in two different ways in sections [ sec : kac det and sing vct ] and [ sec : crystallization ] , respectively . let us define the vertex operators @xmath228 and @xmath229
. set @xmath230 define generators @xmath231 by @xmath232 where @xmath233 denotes the usual normal ordered product , and @xmath234 the generator @xmath235 arises from the level @xmath0 representation of ding - iohara - miki algebra @xcite , and is obtained by acting the coproduct of the dim algebra to the vertex operator @xmath66 @xmath0 times ( see appendix [ sec : def of dim ] ) .
the other generators @xmath69 appear in the commutation relations of generators @xmath236 ( @xmath237 ) .
when we just consider the agt conjecture , it suffices to deal with the subalgebra @xmath67 in some completion of the endomorphism of the algebra of @xmath0-tensored fock modules for our heisenberg algebra .
we denote the algebra @xmath67 by @xmath68 .
if @xmath238 , the commutation relations of the generators are @xmath239 where @xmath240 is the multiplicative delta function and the structure constant @xmath241 is defined by @xmath242 @xmath243 these relations are equivalent to @xmath244= -\sum_{l = 1}^{\infty } f^{(1)}_l ( x^{(1)}_{n - l } x^{(1)}_{m+l } - x^{(1)}_{m - l } x^{(1)}_{n+l } ) + \frac{(1-q)(1-t^{-1})}{1-p}(p^m - p^n ) x^{(2)}_{n+m } , \\ & [ x^{(2)}_n , x^{(2)}_m]= -\sum_{l = 1}^{\infty } f^{(2)}_l ( x^{(2)}_{n - l } x^{(2)}_{m+l } - x^{(2)}_{m - l } x^{(2)}_{n+l } ) , \\ & [ x^{(1)}_n , x^{(2)}_m]= -\sum_{l = 1}^{\infty } f^{(1)}_l ( p^{l } x^{(1)}_{n - l } x^{(2)}_{m+l } - x^{(2)}_{m - l } x^{(1)}_{n+l } ) .\end{aligned}\ ] ] the proof is similar to the calculation of the deformed virasoro algebra or the deformed @xmath1-algebra . in the formula ( [ eq : rel .
of generator x^1 ] ) , we use @xmath245 for an @xmath0-tuple of parameters @xmath246 , define @xmath247 and @xmath248 to be the highest weight vectors such that @xmath249 ( @xmath250 , @xmath251 ) , @xmath252 and @xmath253 .
@xmath254 is the highest weight module generated by @xmath255 , and @xmath256 is the dual module generated by @xmath257 .
the bilinear form ( shapovalov form ) @xmath258 is uniquely determined by the condition @xmath259 .
[ df : ordinary pbw vct ] for an @xmath0-tuple of partitions @xmath260 , set @xmath261 the pbw theorem can not be used because the algebra @xmath68 is not a lie algebra , but in @xcite it was conjectured that the pbw type vectors @xmath262 and @xmath263 are a basis over @xmath71 and @xmath256 , respectively .
this conjecture can be solved by the kac determinant of the algebra @xmath68 , which is proved in section [ sec : kac det and sing vct ] . in this section ,
we consider another type of the pbw basis , since it has good expression in @xmath105 limit in terms of the hall - littlewood functions ( see section [ sec : crystallization of n=2 ] ) .
[ df : another pbw vct ] for @xmath260 , set @xmath264 let us review the aflt basis in @xmath71 , which is also called generalized macdonald functions . in order to state its existence theorem , let us prepare the following ordering .
[ def : ordering1 ] for @xmath0-tuple of partitions @xmath265 and @xmath266 , @xmath267 here @xmath268 . note that the second condition can be replaced with @xmath269 .
we can state the existence theorem of generalized macdonald functions in the basis of products of macdonald functions @xmath270 , where @xmath271 are macdonald symmetric functions defined in appendix [ sec : macdonald and hl ] with substituting the bosons @xmath272 for the power sum symmetric functions @xmath17 . for each @xmath0-tuple of partitions @xmath265
, there exists a unique vector @xmath273 such that @xmath274 where @xmath275 is a constant , @xmath276 is the eigenvalue of @xmath277 .
similarly , there exists a unique vector @xmath278 such that @xmath279 then the eigenvalues are @xmath280 although the ordering of definition [ def : ordering1 ] is different from the one in @xcite , the eigenfunctions @xmath281 are quite the same .
the proof is similar to the one in section [ sec : partial orderings ] , which follows from triangulation of @xmath77 . by this proposition
, it can be seen that @xmath281 is a basis over @xmath71 , and the eigenvalues of @xmath277 are non - degenerate . in section [
sec : partial orderings ] , a more elaborated ordering is introduced and a relationship between these orderings is explained . in section [ sec : limit to beta ] , it is shown that these vectors @xmath281 correspond to the generalized jack functions defined in @xcite in the @xmath18 limit . to use generalized macdonald functions in the agt correspondence , we need to consider its integral form . in this paper
, we adopt the following renormalization , which is slightly different from that of @xcite .
[ df : integral form of gn macdonald ] define the vectors @xmath282 and @xmath283 , called the integral forms , by the condition that @xmath284 the coefficients @xmath285 and @xmath286 are polynomials in @xmath287 , @xmath288 and @xmath289 with integer coefficients . if @xmath238 , the transition matrix @xmath290 is as follows : @xmath291 @xmath292 @xmath293 by using these integral forms , the five dimensional agt conjecture can be stated in the following form .
* conjecture 3.11 and conjecture 3.13 ) ) the norm of @xmath118 reproduces the nekrasov factor : @xmath294 where @xmath295 .
[ df : dim vertex op ] call the linear operator @xmath296 the vertex operator if it satisfies @xmath297 and @xmath298
. then the relations for the fourier components are @xmath299 for @xmath223 .
if @xmath300 , it is known that @xmath76 exists and is given by @xmath301 where @xmath302 is the operator from @xmath303 to @xmath304 satisfying the relation @xmath305 .
the matrix elements of @xmath306 with respect to generalized macdonald functions are @xmath307 under these conjectures , we can obtain a formula for multi - point correlation functions of @xmath76 by inserting the identity @xmath308 .
in particular , the formula for the four - point functions agrees with the 5d @xmath309 nekrsov formula with @xmath78 matters .
an m - theoretic derivation of this formula is also given by @xcite .
in this section , we give the formula for the kac determinant of the algebra @xmath68 and prove it . moreover , it is shown that singular vectors correspond to the generalized macdonald functions . in order to prove the kac determinant
, we need screening currents of the algebra @xmath68 . to construct them ,
it is necessary to realize the operator @xmath224 and the highest wight vector @xmath247 in terms of the charge operator @xmath310 and @xmath227 ( @xmath311 ) .
let @xmath310 be the operator satisfying the relation @xmath312=\delta_{n,0}\delta_{i , j},\ ] ] @xmath210 be the highest weight vector in the fock module of the heisenberg algebra such that @xmath313 for @xmath314 .
for an @xmath0-tuple of complex parameters @xmath315 with @xmath316 , we realize the highest wight vector @xmath247 and @xmath224 as @xmath317 where @xmath6 is defined by @xmath19 .
then they satisfy the required relation @xmath318 .
similarly , let @xmath319 be the dual highest weight vector , and @xmath320 .
these highest wight vectors are normalized by @xmath321 , and satisfy the condition of the shapovalov form @xmath322 . and @xmath16 are assumed to be generic in this section . ]
we obtain the formula for the kac determinant with respect to the pbw type vectors @xmath323.-algebra are proved in @xcite . ]
[ thm : kacdet ] let @xmath324 .
then @xmath325 where @xmath80 , @xmath81 , and @xmath326 is the number of entries in @xmath15 equal to @xmath152 .
@xmath82 denotes the number of @xmath0-tuples of young diagrams of size @xmath34 , i.e. , @xmath327 .
in particular , if @xmath300 , @xmath328 if @xmath329 and @xmath330 for any numbers @xmath152 , @xmath162 and integers @xmath331 , @xmath332 , then the pbw type vectors @xmath323 ( resp .
@xmath333 ) are a basis over @xmath71 ( resp .
@xmath334 ) .
it can be seen that the representation of the algebra @xmath68 on the fock module @xmath71 is irreducible if and only if the parameters @xmath335 satisfy the condition that @xmath329 and @xmath330 .
the proof of theorem [ thm : kacdet ] is given in the next section .
it is known that the level @xmath0 representation of the dim algebra introduced in the last section can be regarded as the tensor product of the deformed @xmath27-algebra and the heisenberg algebra associated with the @xmath2 factor @xcite .
this fact is obtained by a linear transformation of bosons .
the point of proof of theorem [ thm : kacdet ] is to construct singular vectors by using screening currents of the deformed @xmath27-algebra under the decomposition of the generators @xmath336 into the deformed @xmath27-algebra part and the @xmath2 part . in general , a vector @xmath337 in the fock module @xmath71 is called the singular vector of the algebra @xmath68 if it satisfies @xmath338 for all @xmath152 and @xmath339 .
the singular vectors obtained by the screening currents are intrinsically the same one of the deformed @xmath1-algebra . from this singular vector
, we can get the vanishing line of the kac determinant in the similar way of the deformed @xmath27-algebra . at first , in the @xmath340 case , we introduce the following bosons . + * u(1 ) part boson * @xmath341 @xmath342 + * orthogonal component of @xmath343 for @xmath344 * @xmath345 + * fundamental boson of the deformed @xmath27-algebra part * @xmath346 @xmath347 then they satisfy the following relations @xmath348=-\frac{(1-q^{|n|})(1-t^{-|n|})(1-p^{|n|})}{n(1-p^{n|n| } ) } \delta_{n+m,0 } , \qquad [ { b^{(i)}}_n , b'_m]=[{h^{(i)}}_n , b'_m]=[{q^{(i)}}_h , b'_m]=0,\ ] ] @xmath349=\frac{(1-q^{|n|})(1-t^{-|n|})}{n}\delta_{i , j}\delta_{n+m,0 } -p^{(n-\frac{i+j}{2})|n|}\frac{(1-q^{|n|})(1-t^{-|n|})(1-p^{|n|})}{n(1-p^{n|n|})}\delta_{n+m,0},\ ] ] @xmath350= -\frac{(1-q^n)(1-t^{-n})(1-p^{(\delta_{i , j}n-1)n})}{n(1-p^{nn})}p^{nn \,\theta(i > j ) } \delta_{n+m,0},\ ] ] @xmath351=\delta_{i , j}-\frac{1}{n } , \qquad \sum_{i=1}^n p^{-in } { h^{(i)}}_n = 0 , \qquad { q^{(n)}}_{\lambda}=\sum_{i=1}^n q^{(i)}_h=0 .\ ] ] where @xmath352 is @xmath169 or @xmath130 if the proposition @xmath353 is true or false , respectively .
correspond to the fundamental bosons @xmath354 in @xcite ] using these bosons , we can decompose the generator @xmath336 into the u(1 ) part and the deformed @xmath1-algebra part .
that is to say , @xmath355 @xmath356 and @xmath357 @xmath358 @xmath359 is the generator of the deformed @xmath27-algebra .
let us introduce the new parameters @xmath360 and @xmath361 defined by @xmath362 then the inner product of pbw type vectors can be written as @xmath363 hence , its determinant is also in the form @xmath364 where @xmath365 is a polynomial in @xmath360 ( @xmath366 ) which is independent of @xmath361 . now in @xcite , the screening currents of the deformed @xmath27-algebra are introduced : @xmath367 where @xmath368 is the root boson defined by @xmath369 and @xmath370 .
the bosons @xmath368 and @xmath371 commute with @xmath372 , and it is known that the screening charge @xmath373 commutes with the generators @xmath374 . therefore , @xmath373 commutes with any generator @xmath375 , and it can be considered as the screening charges of the algebra @xmath68 . define parameters @xmath376 , @xmath377 by @xmath378 and @xmath379 , and set @xmath380 . then @xmath381 . for any number @xmath382 , the vector arising from the screening current @xmath89
, @xmath383 is a singular vector .
@xmath85 is in the fock module @xmath384 with the parameter @xmath335 satisfying @xmath385 for @xmath386 and @xmath387 , @xmath388 .
the @xmath389 vectors obtained by this singular vector @xmath390 contribute the vanishing point @xmath391 in the polynomial @xmath59 .
similarly to the case of the deformed @xmath27-algebra ( see @xcite ) , by the @xmath86 weyl group invariance of the eigenvalues of @xmath392 , the polynomial @xmath59 has the factor @xmath393 ( @xmath394 ) . considering the degree of polynomials @xmath395 , we can see that when @xmath396 , the kac determinant is @xmath397 where @xmath398 is a rational function in parameters @xmath13 and @xmath16 and independent of the parameters @xmath289 . if @xmath300 , @xmath399 is clearly in the form @xmath400 next , the prefactor @xmath398 can be computed in general @xmath0 case by introducing another boson @xmath401 the commutation relation of the boson @xmath402 is @xmath403 = -\frac{(1-t^{-n})(1-q^n)}{n}\delta_{i , j } , \quad n>0.\ ] ] define the determinants @xmath404 and @xmath405 with @xmath406 given by the expansions @xmath407 where @xmath408 and @xmath409 are @xmath410 by using these determinants , the kac determinant can be written as @xmath411 here @xmath412 is the determinant of the diagonal matrix @xmath413 .
this factor is independent of the parameters @xmath289 , and we have @xmath414 . in ( [ eq : kacdet1 ] ) , the factor depending on @xmath289 in @xmath399 was already clarified .
hence , we can determine the prefactor @xmath398 by computing the leading term in @xmath415 .
that is , the prefactor @xmath398 can be written as @xmath416 where we introduce the function @xmath417 which gives the leading term of @xmath418 as the polynomial in @xmath419 , and @xmath420 . to calculate this leading term ,
define the operators @xmath421 , @xmath422 by @xmath423 @xmath424 is arising from only the operator @xmath425 in @xmath426 .
then @xmath427 let the matrices @xmath428 and @xmath429 be given by @xmath430 where @xmath431 and @xmath432 are defined in the usual way : @xmath433 then @xmath424 is expressed as @xmath434 since the matrix @xmath428 is lower triangular with respect to the partial ordering @xmath435 and @xmath436 are defined as follows : @xmath437 @xmath438 then we have @xmath439 unless @xmath440 . ] and its diagonal elements are @xmath441 we have @xmath442 the transition matrix @xmath443 is upper triangular with respect to the partial ordering @xmath436 , and all diagonal elements are @xmath169 . thus @xmath444 . similarly by considering the base transformation to @xmath445 , it can be seen that @xmath446 therefore the prefactor @xmath398 is @xmath447 hence theorem [ thm : kacdet ] is proved . in this subsection ,
the singular vectors of the algebra @xmath68 are discussed .
trivially , when @xmath448 , the kac determinant ( [ eq : kacdet for dim ] ) degenerates , and it can be easily seen that the vectors @xmath449 are singular vectors . since the screening operator @xmath89 is the same one of the deformed @xmath27-algebra , the situation of the singular vectors of @xmath68 except contribution arising when @xmath448 is the same as the deformed @xmath27-algebra . we discover that singular vectors obtained by the screening currents @xmath89 correspond to generalized macdonald functions .
algebra and the aflt basis is investigated in @xcite . ] first , we have the following simple theorem .
[ thm : sing vct and gn mac 1 ] for a number @xmath450 , if @xmath451 and the other @xmath452 are generic , there exists a unique singular vector @xmath85 in @xmath71 , and it corresponds to the generalized macdonald function @xmath281 with @xmath453 that is , @xmath454 existence and uniqueness are understood by the formula for the kac determinant ( [ eq : kacdet for dim ] ) in the usual way .
actually , the unique singular vector @xmath85 is the one of ( [ eq : sing vct chi(i)rs ] ) .
since the screening charges commute with @xmath277 , the singular vector is an eigenfunction of @xmath77 of the eigenvalue @xmath455 . using the relations @xmath385 for @xmath386 and @xmath387 , @xmath388
, we have @xmath456 where @xmath457 is the eigenvalue of the generalized macdonald functions introduced in ( [ eq : eigenvalue of gn mac ] ) .
thus , the singular vector @xmath85 and the generalized macdonald function @xmath458 are in the same eigenspace of @xmath459 . moreover , by comparing the eigenvalues @xmath460 , it can be shown that the dimension of the eigenspace of the eigenvalue @xmath461 is @xmath169 even when @xmath451 .
therefore , this theorem follows .
let us consider more complicated cases
. for variables @xmath462 ( @xmath463 ) , define the function @xmath464 by @xmath465 then it satisfies @xmath466 .
we focus on the following singular vectors @xmath467 where the parameter @xmath468 is @xmath469 , @xmath470 , and for non - negative integers @xmath92 and @xmath471 ( @xmath463 ) , @xmath472 then the singular vector @xmath473 is in the fock module @xmath71 of the highest weight @xmath474 defined by @xmath475 , @xmath476 , @xmath477 and @xmath478 now we obtain the following main theorem with respect to the generalized macdonald functions and the singular vectors of the dim algebra . this theorem can be regarded as a generalization of the result in @xcite .
[ thm : sing vct and gn
mac 2 ] let parameters @xmath289 satisfy @xmath479 for all @xmath152 .
* ( a ) . * if @xmath93 for all @xmath97 , then the singular vector @xmath473 coincides with the generalized macdonald function @xmath480 with @xmath481 : @xmath482 see figure [ fig : youngdiag_onlyrightside ] in introduction . *
if @xmath483 for all @xmath97 , the singular vector @xmath473 coincides with the generalized macdonald function associated with the tuple of young diagrams @xmath484 : @xmath485 see figure [ fig : youngdiag_somerectangle ] in introduction .
the proof is quite similar to that of theorem [ thm : sing vct and gn mac 1 ] .
the eigenvalue of this singular vector is @xmath486 on the other hand , the eigenvalue of the generalized macdonald function in the case * ( a ) * is calcurated as follows .
firstly , @xmath487 and @xmath488 hence , by using the equation @xmath489 , we can see that @xmath490 this is equal to the eigenvalue of the singular vector .
also , it can be seen that the dimension of the eigenspace of the eigenvalue @xmath491 is @xmath169 even when @xmath479 .
if the condition @xmath95 does not hold , figure [ fig : youngdiag_onlyrightside ] is not a young diagram . in this case , the singular vector @xmath492 corresponds to the generalized macdonald function with the @xmath0-tuple of young diagram obtained by cutting off the protruding parts and moving the boxes to the young diagram in the left side .
that is , the case * ( b)*. the proof in the case * ( b ) * is exactly the same as the case * ( a ) * , so it is omitted .
it is known that projections of the singular vectors @xmath492 in the case * ( a ) * onto the diagonal components of the boson @xmath493 correspond to ordinary macdonald functions ( * ? ? ? * ( 35 ) ) .
hence , ordinary macdonald functions are obtained by the projection of generalized macdonald functions .
[ cor : projection of gn mac ] when @xmath479 for all @xmath152 , @xmath494 here , @xmath17 denotes the ordinary power sum symmetric functions .
next , we consider a crystallization of the results of subsection [ sec : review of simplest 5d agt ] , namely the behavior in the @xmath3 limit of the deformed virasoro algebra and the simplest 5d agt correspondence . in this limit ,
the scaled generators @xmath495 satisfy the commutation relation @xmath496 = & -(1-t^{-1})\sum_{\ell = 1}^{n - m}{\tilde t } _ { n-\ell}{\tilde t } _ { m+\ell } \quad ( n > m > 0
\quad \mbox{or}\quad 0>n > m ) , \\
[ { \tilde t } _
n,{\tilde t } _ 0 ] = & -(1-t^{-1})\sum_{\ell = 1}^{n}{\tilde t } _ { n-\ell}{\tilde t } _ { \ell } -(t - t^{-1})\sum_{\ell = 1}^{\infty } t^{-\ell } { \tilde t } _ { -\ell}{\tilde t } _ { n+\ell } \quad ( n > 0 ) , \\
[ { \tilde t } _
0,{\tilde t } _ m ] = & -(1-t^{-1})\sum_{\ell = 1}^{-m}{\tilde t } _ { -\ell}{\tilde t } _ { m+\ell } -(t - t^{-1})\sum_{\ell = 1}^{\infty } t^{-\ell } { \tilde t } _ { m-\ell}{\tilde t } _ { \ell } \quad ( 0 > m ) , \allowdisplaybreaks[4 ] \\
[ { \tilde t } _
n,{\tilde t } _ m ] = & -(1-t^{-1}){\tilde t } _ { m}{\tilde t } _ { n } -(t - t^{-1})\sum_{\ell = 1}^{\infty } t^{-\ell } { \tilde t } _ { m-\ell}{\tilde t } _ { n+\ell } \nonumber \\ & + ( 1-t^{-1})\delta_{n+m,0 } \quad ( n > 0 > m).\end{aligned}\ ] ] in @xcite , the above algebra is introduced and its free field representation is given .
let the bosons @xmath497 ( @xmath22 ) satisfy the relations @xmath498=n \frac{1}{1-t^{|n|}}\delta_{n+m,0}$ ] , @xmath499 = \delta_{n,0}$ ] .
these bosons can be regarded as the @xmath105 limit of the bosons @xmath137 and @xmath500 in ( [ eq : comm rel of qt - boson ] ) , i.e. , @xmath501 , @xmath502 .
then @xmath503 is represented as @xmath504 \tilde{\lambda}^+(z ) + \theta[\,n\geq0\ , ] \tilde{\lambda}^-(z ) \right ) z^n,\ ] ] where @xmath505 and @xmath506 $ ] is 1 or 0 if the proposition @xmath353 is true or false , respectively . by this free field representation ,
we can write the pbw type vectors in terms of hall - littlewood functions @xmath507 defined in appendix [ sec : macdonald and hl ] : @xmath508 here @xmath509 is an abbreviation for @xmath510 , and @xmath215 and @xmath511 are the same highest weight vectors in section [ sec : review of simplest 5d agt ] such that @xmath512 and @xmath513 .
these expressions are the consequences of jing s operators ( fact [ fact : jing s operator ] ) . because of ( [ eq : inner prod of hl poly ] ) ,
they are diagonalized as @xmath514 where @xmath515 is defined in appendix [ sec : macdonald and hl ] .
since @xmath516 is non - degenerate , there is no singular vector in the limit @xmath3 .
the disappearance of singular vectors can be understood by the fact that the highest weight which has singular vectors diverges at @xmath517 .
the whittaker vector of this algebra is similarly defined .
define the whittaker vector @xmath518 by the relation @xmath519 similarly , the dual whittaker vector @xmath520 is defined by @xmath521 then the crystallized whittaker vector is in the simple form @xmath522 and its inner product is @xmath523 on the other hand , recalling the nekrasov formula @xmath524 given in ( [ eq : nek formula for pure ] ) of subsection [ sec : review of simplest 5d agt ] , we can take the crystal limit with the following trick . the renormalization @xmath525 controls divergence in the @xmath3 limit ( @xmath526 , @xmath527 : fixed ) : @xmath528 @xmath529 removing parts
which have singularity in the nekrasov factor , we have @xmath530 hence , @xmath531 @xmath532 if @xmath533 or @xmath534 for any integer @xmath34 , @xmath535 , then @xmath536 at @xmath3 .
therefore , the sum with respect to partitions @xmath15 , @xmath537 can be rewritten as the sum with respect to integers @xmath34 , @xmath535 , i.e. , @xmath538 after some simple calculation , we get ( [ eq : tildez ] ) . using these calculations , we can get the following theorem which is an analog of the simplest 5d agt relation ( fact [ fact : simpleagt ] ) , and prove it more easily than the generic case .
@xmath539 note that the left hand side is independent of @xmath97 .
@xmath540 can be rewritten as @xmath541 which has simple poles at @xmath542 with @xmath543 , @xmath544 and @xmath545 .
then @xmath546 note that @xmath547 thus @xmath548 is an odd function in @xmath331 .
therefore @xmath549 residues at all singularities in @xmath500 of @xmath550 vanish , but @xmath551 .
hence @xmath550 is independent of @xmath500 .
therefore , @xmath552 in this paper , we discuss the crystallization only of the deformed virasoro algebra .
it is expected that the limit can be taken for the general deformed @xmath27-algebra .
however in the case of @xmath553 , an essential singularity seems to appear , and at present we do not know how to take an appropriate limit . to find an appropriate limit procedure and apply the agt conjecture for the deformed @xmath27-algebra
@xcite we need further studies . in the crystallized case ,
the screening current diverges , which is one of the reasons why in this limit singular vectors disappear .
hence it may be difficult to apply the agt correspondence studied by @xcite .
next , we discuss a crystallization of the results of subsection [ sec : reargument of di alg and agt ] . in this subsection and the next subsection , unlike section [ sec : kac det and sing vct ] ,
the operators @xmath224 are assumed to be independent of the parameter @xmath13 in order to avoid difficulty in taking the @xmath105 limit .
let us realize the operators @xmath224 and the vector @xmath247 as @xmath554 then they also satisfy relation @xmath252 .
similarly @xmath555 .
moreover , we consider the case that the parameters @xmath289 are independent of @xmath13 .
the case where the parameters @xmath289 depend on @xmath13 is briefly described in section [ sec : another type of limit ] . at first , let us demonstrate the @xmath105 limit in the @xmath300 case . in this subsection , we use the same bosons @xmath497 and @xmath205 as subsection [ sec : crystal of qvir ] . since singularity in @xmath76 can be removed by normalization @xmath556 , define the vertex operator @xmath557 by @xmath558 if @xmath300 , @xmath559 are ordinary macdonald functions , and their integral forms @xmath560 have , at @xmath517 , the relation @xmath561 hence , the matrix elements @xmath562 can be written in terms of integrals by virtue of jing s operators @xmath563 and @xmath564 defined in ( [ eq : jing s op ] ) and ( [ eq : dual jing s op ] ) . using the usual normal ordered product @xmath233 with respect to the bosons @xmath497 , be the heisenberg algebra generated by the bosons @xmath497 ( @xmath9 ) , @xmath205 and @xmath169 .
@xmath565 is the algebra obtained by making @xmath566 commutative .
the normal ordered product @xmath233 is defined to be the linear map from @xmath565 to @xmath566 such that for @xmath567 , @xmath568 and @xmath569 . in the next subsection
, the same symbol @xmath233 denotes the normal ordered product with respect to the bosons @xmath570 which is defined similarly .
] we have @xmath571 thus @xmath572 where @xmath573 , @xmath574 , @xmath575 , and the integration contour is @xmath576 .
this integral reproduces the @xmath3 limit of the nekrasov factor .
[ def : crystal nek factor ] set @xmath577 where @xmath578 is the set of boxes in @xmath537 whose arm length @xmath579 is not zero .
for example , if @xmath580 , @xmath581 .
this nekrasov factor has the property @xmath582 for any @xmath15 .
therefore , the conjecture in the crystallized case of @xmath300 is @xmath583 the case of some particular partitions can be checked by calculating the contour integral ( appendix [ sec : check of n=1 conjecture ] ) .
next , let us consider the @xmath105 limit in the case of @xmath238 . in this case
, the generator @xmath584 of the heisenberg algebra is renormalized as @xmath585 and the generator @xmath586 is used as it is . by this normalization , it is possible to take the limit . also , the algebraic structure of @xmath587 and @xmath588 does not change .
then @xmath587 and @xmath588 have the form @xmath589 moreover , let us use the bosons @xmath570 ( @xmath590 ) and @xmath591 with the relation @xmath592=n\frac{1}{1-t^{|n| } } \delta_{i , j } \
, \delta_{n+m , 0 } , \quad [ b_n^{(i ) } , { q}^{(j ) } ] = 0,\ ] ] and regard @xmath593 , @xmath594 .
let us define the generator at @xmath3 .
[ def : crystal first generator ] set @xmath595 [ prop : free field rep of x^1 ] definition [ def : crystal first generator ] is well - defined , i.e. , @xmath596 has no singularity at @xmath517 , and its free field representation is @xmath597 \tilde{\lambda}^{1}(z ) + \theta [ n\leq 0 ] \tilde{\lambda}^{2}(z ) \right\ } z^n,\ ] ] where @xmath598 is defined in section [ sec : crystal of qvir ] and @xmath599 define @xmath600 and @xmath601 by @xmath602 we can see @xmath603 is well - behaved in the limit @xmath3 by the form of @xmath604 . if @xmath339 , @xmath605 if @xmath606 , @xmath607 and if @xmath608 , @xmath609 thus @xmath610 is well - defined and ( [ eq : free field rep of txo ] ) is the natural free field representation . for the second generator ,
the following rescale is suitable .
set @xmath611 the free field representation of @xmath612 is given by @xmath613 where @xmath614 this proposition is easily obtained by calculating @xmath615 .
we can calculate the commutation relation of these generators as follows .
the generators @xmath610 and @xmath612 satisfy the relations @xmath616 & = -(1-t^{-1 } ) \sum_{l=1}^{n - m } \tilde{x}^{(1)}_{n - l } \tilde{x}^{(1)}_{m+l } \qquad ( n > m>0 \;\ ; \mathrm{or } \;\ ; 0>n > m ) , \\ [ \tilde{x}^{(1)}_n , \tilde{x}^{(1)}_0 ] & = -(1-t^{-1 } ) \sum_{l=1}^{n-1 }
\tilde{x}^{(1)}_{n - l } \tilde{x}^{(1)}_{l } -(1-t^{-1 } ) \sum_{l=1}^{\infty } \tilde{x}^{(1)}_{-l } \tilde{x}^{(1)}_{n+l } + ( 1-t^{-1 } ) \tilde{x}^{(2)}_{n } \quad ( n>0 ) , \\ [ \tilde{x}^{(1)}_n , \tilde{x}^{(1)}_m ] & = -(1-t^{-1})\sum_{l=0}^{\infty } \tilde{x}^{(1)}_{m - l } \tilde{x}^{(1)}_{n+l } + ( 1-t^{-1 } ) \tilde{x}^{(2)}_{n+m } \quad ( n>0>m ) , \\ [ \tilde{x}^{(1)}_0 , \tilde{x}^{(1)}_m ] & = -(1-t^{-1 } ) \sum_{l=1}^{-m-1 } \tilde{x}^{(1)}_{-l } \tilde{x}^{(1)}_{m+l } -(1-t^{-1 } ) \sum_{l=1}^{\infty }
\tilde{x}^{(1)}_{m - l } \tilde{x}^{(1)}_{l } + ( 1-t^{-1 } ) \tilde{x}^{(2)}_{m } \quad ( 0>m ) , \end{aligned}\ ] ] @xmath617 & = ( 1-t^{-1 } ) \sum_{l=1}^{\infty } \tilde{x}^{(2)}_{m - l } \tilde{x}^{(1)}_{n+l } \qquad ( n>0,\ ; \forall m ) , \\ [ \tilde{x}^{(1)}_0 , \tilde{x}^{(2)}_m ] & = -(1-t^{-1 } ) \sum_{l=1}^{\infty } ( \tilde{x}^{(1)}_{-l } \tilde{x}^{(2)}_{m+l}- \tilde{x}^{(2)}_{m - l } \tilde{x}^{(1)}_{l } ) \quad ( \forall m ) , \\ [ \tilde{x}^{(1)}_n , \tilde{x}^{(2)}_m ] & = -(1-t^{-1 } ) \sum_{l=1}^{\infty } \tilde{x}^{(1)}_{n - l } \tilde{x}^{(2)}_{m+l } \qquad ( n<0,\ ; \forall m ) , \end{aligned}\ ] ] @xmath618 = -(1-t^{-1 } ) \sum_{l=1}^{\infty } ( \tilde{x}^{(2)}_{n - l } \tilde{x}^{(2)}_{m+l}- \tilde{x}^{(2)}_{m - l } \tilde{x}^{(2)}_{n+l } ) \qquad ( \forall n , m).\ ] ] these are obtained by the following relation of generating functions : @xmath619 where @xmath620 and for ( [ relation of tlo tlt ] ) we used the formula @xmath621 the algebra generated by @xmath622 and @xmath623 is closely related to the hall - littlewood functions . in particular ,
the pbw type vectors can be written as the product of two hall - littlewood functions . for a pair of partitions @xmath624 , set @xmath625 we have the expression of these vectors in terms of the hall - littlewood functions .
[ prop : pbw vct rep by hall ] @xmath626 where @xmath627 the vectors
@xmath628 do not have such a good expression .
this proposition is proved by the theory of jing s operator .
then the vectors @xmath629 are partially diagonalized as the following proposition . furthermore , with the help of hall - littlewood functions
, we can calculate the shapovalov matrix @xmath630 and its inverse @xmath631 .
[ prop : shapovalov in terms of hl poly ] we can express @xmath632 by the inner product @xmath633 of hall - littlewood functions defined in appendix [ sec : macdonald and hl ] : @xmath634 ( [ eq : shapovalov ] ) follows from proposition [ prop : pbw vct rep by hall ] .
( [ eq : inverse shapovalov ] ) can be obtained by the equation @xmath635 which is shown by inserting the complete system with respect to @xmath636 into the equation @xmath637 .
existence of the inverse matrix @xmath638 leads linear independence of @xmath639 .
since there are the same number of linear independent vectors as the dimension of each level of @xmath71 , we can see that @xmath639 forms a basis over @xmath71 . if @xmath16 is not a root of unity and @xmath640 , @xmath639 ( resp .
@xmath641 ) is a basis of @xmath71 ( resp .
@xmath642 ) . next , let us introduce generalized hall - littlewood functions which are specialization of generalized macdonald functions and give some crystallized versions of the agt conjecture .
[ def : gn hall - littlewood ] define the vectors @xmath643 and @xmath644 as the @xmath3 limit of generalized macdonald functions , i.e. , @xmath645 we call the vectors @xmath643 generalized hall - littlewood functions .
these are the eigenvectors of @xmath646 : @xmath647 moreover the eigenvalues are @xmath648 however there are too many degenerate eigenvalues to ensure the existence of generalized hall - littlewood functions .
it is difficult to characterize @xmath649 as the eigenfunction of only @xmath650 .
for example , @xmath651 and @xmath652 have the relation @xmath653 , but @xmath654 .
is given under the hypothesis that the vector @xmath559 has no singulality in the limit @xmath3 .
if we can show the existence therem of both generalized macdonald and generalized hall - littlewood functions by using the same partial ordering and the same basis , this hypothesis is guaranteed . ]
let us define the transition matrices @xmath655 and @xmath656 by the expansions @xmath657 then up to the degree 2 the matrix elements @xmath658 are given by @xmath659 @xmath660 up to the degree 2 the matrix elements @xmath661 are given by @xmath662 @xmath663 similarly to the case of generic @xmath13 , we define the integral forms of generalized hall - littlewood functions and give a conjecture of their norms .
the integral forms @xmath664 and @xmath665 are defined by @xmath666 note that the coefficients @xmath667 and @xmath668 can be zero at @xmath517 .
[ conj : inner prod of gn hl ] @xmath669 next , let us define the vertex operator at crystal limit .
the vertex operator @xmath670 is the linear operator satisfying the relations @xmath671 @xmath672
the existence of such an operator is shown by the renormalization @xmath673 . in the relation of @xmath76 and @xmath69
, it is understood that this renormalization is appropriate by considering the shift of @xmath674 such that @xmath557 and not all @xmath675 are commutative and the relation does not diverge .
we give some simple properties of the vertex operator @xmath557 .
@xmath676 for any @xmath250 , @xmath677 these follow from the commutation relations of @xmath557 .
especially note that the three - point function which has generators @xmath675 on the left side , i.e. , @xmath678 , remains only in the case of special young diagrams @xmath679 with only one vertical column .
[ conj : mat element of phi wrt gn hl ] the matrix elements of @xmath557 with respect to the integral form @xmath680 are @xmath681 under these conjectures [ conj : inner prod of gn hl ] and [ conj : mat element of phi wrt gn hl ] , we obtain the formula for correlation functions of the vertex operator @xmath557 . for example , the function corresponding to the four - point conformal block is @xmath682 ( [ eq : expansion by aflt ] ) is the agt conjecture in the limit @xmath3 with help of the aflt basis .
however , in the crystallized case , we can prove another formula for this four - point correlation function by using the pbw type basis . at first , let us show the following two lemmas . [ lem : matrix el . of tphi wrt pbw ] the matrix elements with respect to pbw type vector @xmath683 and @xmath684 are @xmath685 for @xmath686 , by ( [ eq : comm rel of tphi and txt ] ) and the relation @xmath687 , @xmath688
repeating this calculation , we get @xmath689 where @xmath690 .
when @xmath691 , by similar calculation @xmath692 by using above two formulas ( [ eq : preparation for i>1 ] ) and ( [ eq : preparation for i=1 ] ) , if we write @xmath693 , @xmath694 we have explicit form of formulas for some parts of the inverse shapovalov matrix .
[ lem : exlicit form of inv .
@xmath695 in this proof , we put @xmath696 .
hall - littlewood function @xmath697 is the elementary symmetric function @xmath698 times @xmath699 .
elementary symmetric functions have the generating function @xmath700 hence by the @xmath701 case of fact [ fact : spetialization of hl poly ] , @xmath702 therefore , the lemma follows from proposition [ prop : shapovalov in terms of hl poly ] .
we give other proofs of this lemma in appendix [ seq : another proof of lemma ] , and the form of @xmath703 can be found in appendix [ sec : explicit form of < q , q > ] . by the property ( [ eq : simple property1 ] ) , proposition [ prop : shapovalov in terms of hl poly ] and lemmas [ lem : matrix el .
of tphi wrt pbw ] and [ lem : exlicit form of inv .
shapovalov ] , we can show the following theorem .
[ thm : main theorem ] @xmath704 in this way , the explicit formula for the correlation function can be obtained , where we do nt use any conjecture . the formulas ( [ eq : expansion by aflt ] ) and
( [ eq : expansion formula by pbw ] ) are compared in appendix [ comparison of two formula ] .
we expect that these works will be generalized to @xmath705 case . finally , we present other types of the crystal limit . in this paper , we investigated the crystal limit while the parameters @xmath289 , @xmath706 and @xmath707 are fixed .
however , it is also important to study the cases when these parameters depend on @xmath13 .
for example , let us consider the case that @xmath708 , @xmath709 , @xmath710 ( @xmath711 ) and @xmath712 , @xmath713 , @xmath714 are independent of @xmath13 or fixed in the limit @xmath3 .
let @xmath715 for all @xmath716 and @xmath717 .
then the nekrasov formula for generic @xmath13 case ( @xmath238 ) @xmath718 depends only on the partitions of the shape @xmath719 in the limit @xmath3 , where @xmath720 is fixed , and coincides with the partition function of the pure gauge theory ( [ eq : tildez ] ) : @xmath721 where @xmath722 .
hence , we expect that the vector @xmath723 corresponds to the whittaker vector in the section [ sec : crystal of qvir ] in this limit , though we were not able to properly explain it . in this way , by considering the various other values of @xmath724 and @xmath725 , we can find special behavior of @xmath726 and the conformal block @xmath727 and may prove the relation .
these are our future studies .
in general , a bialgebra @xmath106 is called quasi - cocommutative if there exists an invertible element @xmath728 such that for all @xmath729 , @xmath730 this @xmath108 is called the universal r - matirx .
the dim algebra is quasi - cocommutative @xcite . in this section ,
we explicitly calculate the representation of @xmath108 .
moreover , the expression of its representation matrix is generally conjectured .
the point of calculation is to make use of the condition that the generalized macdonald functions are the eigenfunctions of @xmath77 , to reduce the degree of freedom of the matrix in advance . in this section
, we formally write the universal @xmath108-matrix as @xmath731 and set @xmath732 , @xmath733 , @xmath734 .
occasionally , we explicitly write the variable of the generalized macdonald functions like @xmath735 and the pbw basis of the bosons is written as @xmath736 firstly , by definition of @xmath108 , we have @xmath737 here @xmath738 .
the formula for the coproduct of the dim algebra and the definition of the representation @xmath739 are given in appendix [ sec : def of dim ] .
hence , @xmath740 are eigenfunctions of @xmath741 . by comparing the forms of @xmath741 and @xmath742 ,
the eigenfunctions of @xmath741 are obtained by replacing @xmath743 with @xmath744 and @xmath745 with @xmath746 . moreover , by checking their eigenvalues , it can be seen that @xmath740 are proportional to @xmath747 : @xmath748 where @xmath749 are proportionality constants . by this property
, the representation matrix of @xmath750 is block - diagonalized at each level of the boson @xmath343 .
the proportionality constants @xmath751 can be calculated by using the generalized macdonald functions in the @xmath752 case .
similarly to the @xmath238 case , from the relation @xmath753 we have @xmath754 where @xmath755 and @xmath756 are constants .
since the generalized macdonald functions satisfy @xmath757 the proportionality constants have the relation @xmath758 . for example
let us describe the calculation of the representation matrix at level 1 .
the following are examples of @xmath759 at level 1 : @xmath760 } \rangle } \\ { | p_{\emptyset,[1],\emptyset } \rangle } \\
{ | p_{[1],\emptyset,\emptyset } \rangle } \end{array } \right ) = a(u_1,u_2,u_3 ) \left ( \begin{array}{c } { | a_{\emptyset,\emptyset,[1 ] } \rangle } \\ { | a_{\emptyset,[1],\emptyset } \rangle } \\ { | a_{[1],\emptyset,\emptyset } \rangle } \end{array } \right),\ ] ] @xmath761 by the above discussion , if we set the matrix @xmath762 } & 0 & 0 \\ 0 & k^{(12)}_{\emptyset,[1],\emptyset } & 0 \\ 0 & 0 & k^{(12)}_{[1],\emptyset,\emptyset } \\
\end{array}\right ) a^{(12)}(u_1,u_2,u_3 ) a^{-1}(u_1,u_2,u_3),\ ] ] @xmath763 then the representation matrix of @xmath764 in the basis of the generalized macdonald functions is the transposed matrix of @xmath765 : @xmath766 } \right\rangle } & { \left| p_{\emptyset,[1],\emptyset } \right\rangle } & { \left| p_{[1]\emptyset,\emptyset } \right\rangle } \\
\end{array}\right ) = \left(\begin{array}{ccc } { \left| p_{\emptyset,\emptyset,[1 ] } \right\rangle } & { \left| p_{\emptyset,[1],\emptyset } \right\rangle } & { \left| p_{[1]\emptyset,\emptyset } \right\rangle } \\
\end{array}\right ) { } ^t b^{(12)}.\ ] ] the constants @xmath767 are determined as follows . at first , since scalar multiples of r - matrices are also r - matrices , we can normalize as @xmath768 .
this means that @xmath769 .
next , we consider the base change from @xmath770 to the bosons @xmath771 : @xmath772 then @xmath773 is in the form @xmath774 where @xmath775 is a function of @xmath776 . since for the action of @xmath777 to @xmath778 , variables @xmath779 and @xmath780 should not appear , we get equations @xmath781 . by solving these equations
, we can see that @xmath782 substituting this value into the matrix ( [ eq : b12 containing k ] ) , we have @xmath783 in this way , we obtain the explicit expression @xmath773 of representation matrix of universal @xmath108 at level 1 .
of course , it is possible to calculate the representation matrix of @xmath784 in the same way , but by using symmetry with respect to @xmath785 at different @xmath152 , we can easily understand the forms of @xmath784 and @xmath786 : @xmath787 indeed , we can check that they satisfy the yang - baxter equation @xmath788 incidentally , in the basis of the generalized macdonald functions , @xmath789 @xmath790 the representation matrix of @xmath791 is the @xmath792 matrix block at the lower right corner of @xmath765 or @xmath773 .
for example , in the basis of generalized macdonald functions , its representation matrix is @xmath793 next , let us explain the case at level 2 .
the generalized macdonald functions at level 2 in the @xmath752 case are expressed as @xmath794 } \rangle } & { | p_{\emptyset,\emptyset,[1,1 ] } \rangle } & { | p_{\emptyset,[1],[1 ] } \rangle } & { | p_{[1],\emptyset,[1 ] } \rangle } & { | p_{\emptyset,[2],\emptyset } \rangle } & { | p_{\emptyset,[1,1],\emptyset } \rangle } & { | p_{[1],[1],\emptyset } \rangle } & { | p_{[2],\emptyset,\emptyset } \rangle } & { | p_{[1,1],\emptyset,\emptyset } \rangle } \end{array } \right ) \nonumber \\ & = { } ^t\mathcal{a } \left ( \begin{array}{ccccccccc } { | p'_{\emptyset,\emptyset,[2 ] } \rangle } & { | p'_{\emptyset,\emptyset,[1,1 ] } \rangle } & { | p'_{\emptyset,[1],[1 ] } \rangle } & { | p'_{[1],\emptyset,[1 ] } \rangle } & { | p'_{\emptyset,[2],\emptyset } \rangle } & { | p'_{\emptyset,[1,1],\emptyset } \rangle } & { | p'_{[1],[1],\emptyset } \rangle } & { | p'_{[2],\emptyset,\emptyset } \rangle } & { | p'_{[1,1],\emptyset,\emptyset } \rangle } \end{array } \right),\end{aligned}\]]where @xmath795 denotes the product of ordinary macdonald functions @xmath796 , and the matrix @xmath797 is given in appendix [ sec : ex of r - matrix ] . in the same manner , we can get the representation matrix of @xmath108 . at first , @xmath765 at level 2 is in the form @xmath798 then we can find the proportionality constants such that all @xmath799 are zero just by solving equations @xmath800 ( @xmath801 ) .
we have also checked that the representation matrix @xmath802 obtained in this way satisfies the yang - baxter equation up to level 3 .
the explicit expressions of @xmath108 at level 2 are written in appendix [ sec : ex of r - matrix ] .
the proportionality constants are in the form @xmath803 these proportionality constants can be simplified by using the integral forms of the generalized macdonald functions @xmath804 , which are defined in section [ sec : reargument of di alg and agt ] .
define its opposite version by @xmath805 and the constants @xmath806 by @xmath807 by these renormalized functions , the relation ( [ eq : propo rel . ] ) can be written as @xmath808 then it is conjectured that @xmath809 this equation has been checked at @xmath810 .
therefore , the representation matrix @xmath119 of @xmath108 in the basis of the integral forms can be expressed as the following conjecture .
[ conj : r - matrix by int form ] @xmath811 since the formula ( [ eq : formula for r matrix ] ) means the expansion coefficients of @xmath117 in front of @xmath118 , by using the transition matrix defined by @xmath812 and its opposite version @xmath813 , the r - matrix can be calculated by the matrix operation @xmath814 this formula gives a much simpler way to get explicit expressions as compared with deducing them from the universal r - matrix @xcite .
incidentally , the proportionality constants are conjectured to be @xmath815 where @xmath816 is the nekrasov factor defined in ( [ eq : def of nek factor ] ) , section [ sec : review of simplest 5d agt ] .
the second equality follows from the formula ( eq.(2.34 ) in @xcite , eq.(102 ) in @xcite ) @xmath817 where @xmath818 is the framing factor @xcite .
the equation ( [ eq : conj of prop const ] ) has been checked up to level 3 .
the existence theorem of generalized macdonald functions can be stated by the ordering @xmath121 in definition [ def : ordering1 ] . in this subsection
, we introduce a more elaborated ordering . using this ordering
, we can find more elements which is @xmath130 in the transition matrix @xmath819 , where @xmath820 , and get more strict condition to existence theorem .
[ df : ordering elaborated version ] for @xmath0-tuples of partitions @xmath265 and @xmath266 , @xmath821 for all @xmath822 . here
@xmath823 denote that @xmath824 for all @xmath152 . if @xmath752 and the number of boxes is @xmath825 , then @xmath826 & & ( \emptyset , \emptyset , ( 2,1 ) ) \ar[ld]\ar[rd ] & & ( \emptyset , \emptyset , ( 1,1,1))\ar[ld ] & \\ & & ( \emptyset , ( 1 ) , ( 2))\ar[lld]\ar[d]\ar[rrrrd ] & & ( \emptyset , ( 1 ) , ( 1,1))\ar[lllld]\ar[d]\ar[rrd ] & & \\ ( \emptyset , ( 2 ) , ( 1))\ar[d]\ar[rrd]\ar[rrrrd ] & & ( ( 1 ) , \emptyset , ( 2))\ar[d ] & & ( ( 1 ) , \emptyset , ( 1,1))\ar[lld ] & & ( \emptyset , ( 1,1 ) , ( 1))\ar[d]\ar[lld]\ar[lllld ] \\ ( \emptyset , ( 3 ) , \emptyset)\ar[d ] & & ( ( 1 ) , ( 1 ) , ( 1))\ar[lld]\ar[d]\ar[rrd]\ar[rrrrd ] & & ( \emptyset , ( 2,1 ) , \emptyset)\ar[lllld]\ar[rrd ] & & ( \emptyset , ( 1,1,1 ) , \emptyset)\ar[d ] \\ ( ( 1 ) , ( 2 ) , \emptyset)\ar[rrd]\ar[rrrrd ] & & ( ( 2 ) , \emptyset , ( 1))\ar[d ] & & ( ( 1,1 ) , \emptyset , ( 1))\ar[d ] & & ( ( 1 ) , ( 1,1 ) , \emptyset)\ar[lllld]\ar[lld ] \\ & & ( ( 2 ) , ( 1 ) , \emptyset)\ar[ld]\ar[rd ] & & ( ( 1,1 ) , ( 1 ) , \emptyset)\ar[ld]\ar[rd ] & & \\ & ( ( 3 ) , \emptyset , \emptyset ) & & ( ( 2,1 ) , \emptyset , \emptyset ) & & ( ( 1,1,1 ) , \emptyset , \emptyset ) .
& } \ ] ] here @xmath827 stands for @xmath828 . by using the following conjecture , we can state the existence theorem .
[ conj : action of eta_n ] let @xmath829 . in the action of @xmath830 ( @xmath250 ) on macdonald functions @xmath831
, there only appear partitions @xmath537 contained in @xmath15 , i.e. , @xmath832 [ thm : another existence thm ] under the conjecture [ conj : action of eta_n ] , for an @xmath0-tuple of partitions @xmath265 , there exists an unique vector @xmath273 such that @xmath833 at first , @xmath830 satisfies @xmath834 if we act @xmath835 on the product of the macdonald functions , then @xmath836 where @xmath837 is a polynomial of degree @xmath838 of @xmath839 .
such that @xmath840= n a^{(j)}_{-n}$ ] , the polynomial @xmath841 satisfies @xmath842 = ( |\lambda^{(i)}| - |\mu| ) c'_{\lambda^{(i ) } , \mu}$ ] .
] hence @xmath843 therefore one can easily diagonalize it and we have this theorem . in the basis of monomial symmetric functions @xmath844
, we have @xmath845 where @xmath846 ( @xmath847 ) .
thus the partial ordering @xmath848 defined as follows also triangulates @xmath77 .
@xmath849 it can be shown that the partial ordering @xmath848 is equivalent to the ordering @xmath850 introduced in @xcite .
therefore theorem [ thm : another existence thm ] supports the existence theorem in ( * ? ? ?
* proposition3.8 ) .
a representation of the dim algebra called rank @xmath0 representation is provided in @xcite in terms of a basis @xmath851 called aflt basis .
this rank @xmath0 representation corresponds to the @xmath0-fold tensor product of the level ( 0,1 ) representation described in appendix [ sec : def of dim ] .
the level ( 0,1 ) representation can be considered as the spectral dual to the level ( 1,0 ) representation which is realized by the heisenberg algebra . in this subsection ,
based on this spectral duality we present conjectures for explicit expressions of the action of @xmath852 on the generalized macdonald functions , which are defined to be eigenfunctions of the hamiltonian @xmath77 .
we can also conjecture the eigenvalues of higher rank hamiltonians on the generalized macdonald functions from those of the spectral dual generators provided in @xcite .
our conjectures mean that the generalized macdonald functions concretely realize the spectral dual basis to @xmath851 in @xcite .
although we already define the integral forms @xmath118 of the generalized macdonald functions , let us use another renormalization @xmath854 of them , which is defined by @xmath855 where @xmath856 is the nekrasov factor .
this renormalization is the almost same as @xmath118 .
their difference is conjectured to be the scalar multiplication of only monomials in parameter @xmath13 , @xmath16 and @xmath289 .
it is expected that the basis @xmath857 corresponds to the aflt basis were already conjectured to be the aflt basis in this original sense . ] in @xcite and realizes the rank @xmath0 representation through the spectral duality @xmath858 .
that is to say , for any generator @xmath859 in the dim algebra , the action of @xmath860 on the integral forms @xmath854 is in the same form as one of @xmath861 on the basis @xmath862 @xcite , where @xmath863 .
indeed , we can check that the action of @xmath864 on the generalized macdonald functions is as the following conjecture .
let us denote adding a box to or removing it from the young diagram @xmath265 through @xmath865 and @xmath866 respectively .
we also use the notation @xmath867 for the triple @xmath868 , where @xmath869 is the coordinate of the box of the young diagram @xmath870 .
[ conj : action of x ] @xmath871 where @xmath872 and for the triple @xmath873 , we put @xmath874 these actions of @xmath124 in this conjecture come from the corresponding actions of the generators @xmath875 and @xmath876 in @xcite respectively , i.e. , @xmath877 and @xmath878 in our notation , which are the spectral duals of @xmath879 and @xmath880 . incidentally , introducing the coefficients @xmath881 by @xmath882 i.e. , @xmath883 , we can further conjecture that @xmath884 conjecture [ conj : action of x ] with respect to @xmath885 and the formula ( [ eq : rel between cplus and cminus ] ) are checked on a computer for @xmath886 for @xmath300 , for @xmath887 for @xmath888 and for @xmath889 for @xmath890 .
conjecture [ conj : action of x ] with respect to @xmath891 is also checked for the same size of @xmath266 . for each integer @xmath892 ,
the spectral dual of @xmath893 is @xmath894 defined by @xmath895 and @xmath896\cdots ] ] , \qquad k \geq 2.\ ] ] according to @xcite , @xmath894 are spectral dual to @xmath897 and consequently mutually commuting ; @xmath898=0 $ ] .
thus the generalized macdonald functions @xmath281 are automatically eigenfunctions of all @xmath894 , i.e. , @xmath899 , and @xmath894 can be regarded as higher hamiltonians for the generalized macdonald functions . since @xmath894 are the spectral duals to @xmath897 ; @xmath900 , their eigenvalues are expected to be @xmath901 where @xmath902 is defined in ( [ eq : b+ ] ) .
the eigenvalues @xmath903 correspond to those of the rank @xmath0 representation of the generators @xmath897 in @xcite . in the @xmath904 case , the conjecture ( [ eq : higher eigenvlue ] ) can be proved .
we have checked it for @xmath886 for @xmath300 , for @xmath887 for @xmath238 , for @xmath889 for @xmath752 and for @xmath905 for @xmath890 in the @xmath906 case . in this subsection , we show that the generalized macdonald functions are reduced to the generalized jack functions introduced by morozov and smirnov in @xcite in the @xmath18 limit ( see also the sub - thesis @xcite ) . although the scenario of proof of agt correspondence is given in @xcite , the orthogonality of the generalized jack functions are non - trivial since there are degenerate eigenvalues . the cauchy formula used in the scenario of proof
can not be proved without the orthogonality .
however , the eigenvalues of generalized macdonald functions are non - degenerate .
hence , we can prove the orthogonality of the generalized jack functions by using the limit in this section .
when taking this limit , we set @xmath907 @xmath908 , @xmath19 , @xmath909 and take the limit @xmath910 with @xmath6 fixed . to expose the @xmath911 dependence of the generators @xmath343 in the heisenberg algebra ,
they are realized in terms of @xmath0 kinds of power sum symmetric functions @xmath912 with @xmath913 in the ring of symmetric functions @xmath914 in the variables @xmath915 ( @xmath70 , @xmath916 ) : @xmath917 since the operators @xmath224 become parameter @xmath289 in the representation space @xmath71 , they are transformed as @xmath918 from the beginning in this section . with the above transformation , the generator @xmath77 can be regarded as the operator over the ring of symmetric functions @xmath914 .
define the isomorphism @xmath919 by @xmath920 where @xmath921 is defined in section [ sec : explicit cal of r ] .
the inner product @xmath922 over @xmath914 is defined by @xmath923 this inner product naturally realize the one over the fock module @xmath71 , i.e. , @xmath924 let @xmath925 be the adjoint operator of @xmath77 with respect to the inner product . since the generalized macdonald functions @xmath926 have non - degenerate eigenvalues the orthogonality clearly follows : @xmath927 where @xmath928 is defined to be the eigenfunctions of the adjoint operator @xmath925 of the eigenvalue @xmath460 .
now let us take the @xmath125 limit .
at first , consider the @xmath911 expansion of @xmath929 .
by @xmath930 and @xmath931 , @xmath932 , we have @xmath933 hence the @xmath911 expansion is @xmath934 thus we get @xmath935 where @xmath936 @xmath937 for @xmath938 , we define operators @xmath894 by @xmath939 with respect to @xmath940 , @xmath941 and @xmath942 , all homogeneous symmetric functions belong to the same eigenspace .
hence , the eigenfunctions of @xmath943 are the same to those of @xmath944 . in addition
, we have @xmath945 @xmath946 consequently the limit @xmath18 of the generalized macdonald functions are eigenfunctions of the differential operator @xmath947 . as a matter of fact , @xmath947 plus the momentum @xmath948 corresponds to the differential operator of @xcite , the eigenfunctions of which are called generalized jack symmetric functions . , @xmath949 .
] as in ( * ? ?
* proposition 3.7 ) , we can triangulate @xmath947 similarly . moreover if @xmath950 ( resp .
@xmath951 ) if and only if @xmath952 and @xmath953 @xmath954 for all @xmath955 and @xmath956 . ] and @xmath6 is generic , then @xmath957 .
( @xmath958 , @xmath959 are eigenvalues of @xmath947 . )
therefore we get the existence theorem of the generalized jack symmetric functions .
there exists a unique symmetric function @xmath960 satisfying the following two conditions : @xmath961 where @xmath962 denotes the product of monomial symmetric functions @xmath963 .
( @xmath964 is the usual monomial symmetric function of variables @xmath965 . ) from the above argument and the uniqueness in this proposition we get the following important result .
[ prop : limit of gnmac ] the limit of the generalized macdonald symmetric functions @xmath966 to @xmath6-deformation coincide with the generalized jack symmetric functions @xmath960 .
that is @xmath967 for the dual functions @xmath968 and @xmath969 , a similar proposition holds . by the orthogonality ( [ eq : orthogonality of gn mac ] ) , proposition [ prop : limit of gnmac ] and
the fact that the scalar product @xmath922 reduces to the scalar product @xmath970 which is defined by @xmath971 we obtain the orthogonality of the generalized jack symmetric functions .
if @xmath972 , then @xmath973 by this proposition , we can prove the cauchy formula for generalized jack symmetric functions in the usual way .
for example , in the @xmath238 case , we have @xmath974 where @xmath975 .
this is the necessary formula in the scenario of proof of the agt conjecture @xcite .
we give examples of proposition [ prop : limit of gnmac ] in the case @xmath238 .
the generalized macdonald symmetric functions of level 1 and 2 have the forms : @xmath976 @xmath977 @xmath978 also the generalized jack symmetric functions have the forms : @xmath979 @xmath980 @xmath981 if we take the limit @xmath18 of @xmath982 , then @xmath983 appears .
in this subsection , we briefly review some properties of hall - littlewood functions and macdonald functions following ( * ? ? ? * chap .
iii , vi ) .
let @xmath984^{s_n}$ ] be the ring of symmetric polynomials of @xmath0 variables and @xmath985 be the ring of symmetric functions .
the inner product @xmath986 over @xmath197 is defined such that for power sum symmetric functions @xmath987
( @xmath988 ) , @xmath989 where @xmath990 is the number of entries in @xmath15 equal to @xmath152 . for a partition @xmath15 ,
macdonald functions @xmath991 are uniquely determined by the following two conditions @xcite : @xmath992 here @xmath142 is the monomial symmetric function and @xmath993 is the ordinary dominance partial ordering , which is defined as follows : @xmath994 in this paper , we regard power sum symmetric functions @xmath17 ( @xmath143 ) as the variables of macdonald functions , i.e. , @xmath995 . here
@xmath996 is an abbreviation for @xmath997 . in this paper
, we often use the symbol @xmath998 , which is the polynomial of bosons @xmath999 obtained by replacing @xmath17 in macdonald functions with @xmath999 .
next , let the hall - littlewood function @xmath1000 be given by @xmath1001 .
if @xmath1002 , then for a partition @xmath15 of length @xmath1003 , the hall - littlewood polynomial @xmath1000 with @xmath1004 is expressed by @xmath1005 where @xmath1006 .
note that @xmath1007 .
the action of the symmetric group @xmath1008 of degree @xmath0 is defined by @xmath1009 for @xmath1010 .
it is convenient to introduce functions @xmath1011 , which are defined by scalar multiples of @xmath1012 as follows : @xmath1013 where @xmath1014 .
they are diagonalized as @xmath1015 these functions @xmath1011 can be constructed by using jing s operators @xmath1016 and @xmath1017 @xcite , which is defined by @xmath1018 where @xmath497 is the bosons realized by @xmath1019 [ fact : jing s operator ] let @xmath210 be the vector such that @xmath1020 ( @xmath339 ) .
then for a partition @xmath15 , we have @xmath1021 furthermore , the following specialization formula is known .
[ fact : spetialization of hl poly ] let @xmath331 be indeterminate . under the specialization @xmath1022 the hall - littlewood function @xmath1023 is specialized as @xmath1024
in this section , we recall the definition of the dim algebra and the level @xmath0 representation . for the notations ,
we follow @xcite .
the dim algebra has two parameters @xmath13 and @xmath16 .
let @xmath1025 be the formal series @xmath1026 then this series satisfies @xmath1027 .
define the algebra @xmath1028 to be the unital associative algebra over @xmath20 generated by the currents @xmath1029 , @xmath1030 and the central element @xmath1031 satisfying the defining relations @xmath1032
= \dfrac{(1-q)(1 - 1/t)}{1-q / t } \big ( \delta(\gamma^{-1}z / w ) \psi^+(\gamma^{1/2}w)-
\delta(\gamma z / w ) \psi^-(\gamma^{-1/2}w ) \big ) , \\ & g^{\mp}(z / w)x^\pm(z)x^\pm(w)=g^{\pm}(z / w)x^\pm(w)x^\pm(z ) .
\end{aligned}\ ] ] note that @xmath1033 are central elements in @xmath1028 .
let us contain the invertible elements @xmath1034 and @xmath1035 in the definition of @xmath1028 .
further , set @xmath1036 .
this algebra @xmath1028 is an example of the family topological hopf algebras introduced by ding and iohara @xcite .
this family is a sort of generalization of the drinfeld realization of the quantum affine algebras .
however , miki introduce a deformation of the @xmath64 algebra in @xcite , which is the quotient of the algebra @xmath1028 by the serre - type relation .
hence we call the algebra @xmath1028 the ding - iohara - miki algebra ( dim algebra ) .
since the algebra @xmath1028 has a lot of background , there are a lot of other names such as quantum toroidal @xmath65 algebra @xcite , quantum @xmath64 algebra @xcite , elliptic hall algebra @xcite and so on .
this algebra has a hopf algebra structure .
the formulas for its coproduct are @xmath1037 and @xmath1038 , where @xmath1039 and @xmath1040 . since we do not use the antipode and the counit in this thesis , we omit them .
the dim algebra @xmath1028 can be realized by the heisenberg algebra defined in section [ sec : review of simplest 5d agt ] .
[ fact : lv . 1 rep of dim ] the morphism @xmath1041 defined as follows is a representation of the dim algebra : @xmath1042 where @xmath1043 not that the zero mode @xmath1044 of @xmath1045 can be essentially identified with the macdonald difference operator @xcite . by using the coproduct of @xmath1028 , we can consider its tensor representations . for an @xmath0-tuple of parameters @xmath1046 , define the morphism @xmath1047 by @xmath1048 where @xmath1049 is inductively defined by @xmath1050 , @xmath1051 and @xmath1052 .
the representation @xmath1047 is called the level @xmath0 representations . in section [ sec :
reargument of di alg and agt ] , for simplicity , we write the @xmath152-th bosons as @xmath1053 the generator @xmath235 is defined by @xmath1054 the @xmath739 is also called the level @xmath1055 representation or the horizontal representation in @xcite in order to distinguish another representation of the dim algebra , which is called the level @xmath126 representation or the vertical representation . to define the level @xmath126 representation , we introduce some notations . for a partition @xmath193 and a number @xmath1056
, we set @xmath1057 here , @xmath1058 and @xmath1059 are considered as elements in @xmath1060 $ ] .
let @xmath419 be an indeterminate parameter and @xmath1061 be the fock module generated by the highest weight vector @xmath210 .
the morphism @xmath1062 defined as follows gives a representation of the dim algebra : @xmath1063 and @xmath1064 , where @xmath1065 denotes the ordinary macdonald function @xmath1066 associated with the partition @xmath15 . in this thesis , the vectors @xmath1065
are realized by the macdonald functions @xmath1067 along @xmath1068 .
however , they can also be regarded as the abstract vectors labeled by young diagrams .
note that the factors ( [ eq : a+])-([eq : b- ] ) can be written by the contribution from the edges of the young diagram @xmath15 , i.e. , the positions which we can add a box in or remove it from .
the tensor representation of @xmath1069 is also called the rank @xmath0 representation , which is given in @xcite in terms of the edge contribution .
this representation is connected to the level @xmath1055 representation under the change of basis and the automorphism of the dim algebra , that is defined as follows .
this connection is called the spectral duality ( * ? ? ?
* section 5 ) .
there exists an automorphism @xmath1070 such that @xmath1071 @xmath1072 and @xmath1073 .
although the algebra in @xcite slightly differs from the algebra @xmath1028 in the serre - type relation , this fact holds .
this automorphism is of order four . by using this automorphism
, we can check the correspondence between two representations of the dim algebra . in section [ sec : realization of rank n rep ] , we briefly explain the spectral duality and check it with respect to the generators @xmath852 and @xmath893 .
in this subsection , let us explain other proofs of lemma [ lem : exlicit form of inv .
shapovalov ] by the method of contour integrals .
the generating function of elementary symmetric functions and jing s operator makes the equation @xmath1074 where we put @xmath696 , @xmath1075 , and @xmath1076 .
it suffices to show that @xmath1077 , which is proved by a recursive relation of @xmath59 as follows .
the contour integral @xmath1078 surrounding origin is represented as that surrounding @xmath1079 . since @xmath1080 ,
the residue of @xmath1081 at @xmath1082 is @xmath130 .
hence , the only residue at @xmath1083 is left , and it is @xmath1084 by change of variable @xmath1085 , @xmath1086 therefore @xmath1087 thus the lemma [ lem : exlicit form of inv .
shapovalov ] is proved .
although it is slightly hard , one can also prove this lemma by reversing the order of integration , i.e. , first perform over a variable @xmath1088 surrounding origin .
indeed @xmath1088 has the pole only at @xmath1089 , and its residue satisfies @xmath1090 where for @xmath1091 , @xmath1092 and @xmath1093 . by the assumption that @xmath1094 for @xmath1095 with @xmath1096
, we inductively get @xmath1097 by virtue of the equation @xmath1098 which is also proved by induction with respect to @xmath1099 , it can be seen that @xmath1100 .
the formula for @xmath1102 is given in the lemma [ lem : exlicit form of inv .
shapovalov ] and the last subsection .
we also have an explicit form of @xmath1103 and @xmath1104 . by the proposition [ prop : shapovalov in terms of hl poly ]
, it suffices to give the explicit form of @xmath1101 .
@xmath1105 the proof is similar to the previous subsection .
set @xmath1106 then @xmath1107 by jing s operator .
integration of @xmath1088 around @xmath1079 makes recursive relation @xmath1108 and leads this proposition . for general partitions @xmath15 and @xmath537
, we can get the integral representation of @xmath1109 .
however , it is very hard to give their explicit formula . the integral formula ( [ eq : n=1 conjecture ] )
can be checked by the similar way to subsections [ seq : another proof of lemma ] and [ sec : explicit form of < q , q > ] .
let us set @xmath1110 then @xmath1111 , @xmath1112 .
the integration of @xmath1081 around @xmath130 give the relation @xmath1113 on the other hand , the integration of @xmath1114 around @xmath1079 makes @xmath1115 thus @xmath1116 these agree with the right hand side of ( [ eq : n=1 conjecture ] ) . in this subsection
, we compare two formulas ( [ eq : expansion by aflt ] ) and ( [ eq : expansion formula by pbw ] ) which are obtained by the other basis . comparing the coefficients of @xmath1117
, we have the equation @xmath1118 note that the left hand side is the summation with respect to pairs of partitions @xmath1119 and the right hand side is the summation with respect to single partitions @xmath15 .
the right hand side depends only on the ratio @xmath1120 though the left hand side does nt look that way . for a single partition @xmath15 ,
let us define @xmath1121 to be the set of all pairs of partitions @xmath1122 such that a permutation of the sequence @xmath1123 coincides with @xmath15 .
for example , if @xmath1124 , @xmath1125 then we obtain a strange factorization formula with respect to the partial summation of left hand side in ( [ eq : comparison of two formula ] ) @xmath1126 where @xmath1127 and @xmath1128 is introduced in definition [ def : crystal nek factor ] .
if we prove that left hand side of ( [ eq : strange facorization formula ] ) depends only on @xmath1129 , ( [ eq : strange facorization formula ] ) is easily seen by checking the case of @xmath1130 .
this equation almost reproduces each term of the right hand side in ( [ eq : comparison of two formula ] ) . hence ( [ eq : comparison of two formula ] ) may be proved by this equation . if ( [ eq : comparison of two formula ] ) holds , the agt conjecture at @xmath3 with the help of the aflt basis ( [ eq : expansion by aflt ] ) is completely proved .
examples of the generalized macdonald functions at level 3 in the @xmath752 case : @xmath1131 @xmath1132 @xmath1133 b. sutherland , `` exact results for a quantum many body problem in one - dimension , '' http://dx.doi.org/10.1103/physreva.4.2019[_phys .
rev . _ * a4 * ( 1971 ) 20192021 ] .
b. sutherland , `` exact results for a quantum many body problem in one - dimension .
2 . , '' http://dx.doi.org/10.1103/physreva.5.1372[_phys . rev .
_ * a5 * ( 1972 ) 13721376 ]
. i. g. macdonald , _ symmetric functions and hall polynomials_. oxford classic texts in the physical sciences . the clarendon press , oxford university press , new york , second ed . , 2015 .
m. aganagic and s. shakirov , `` knot homology and refined chern - simons index , '' http://dx.doi.org/10.1007/s00220-014-2197-4[_commun .
phys . _ * 333 * no . 1 , ( 2015 )
187228 ] , http://arxiv.org/abs/1105.5117[arxiv:1105.5117 [ hep - th ] ] .
i. cherednik , `` jones polynomials of torus knots via daha , '' _ int .
res . not .
imrn _ no .
23 , ( 2013 ) 53665425 , http://arxiv.org/abs/1111.6195[arxiv:1111.6195 [ math.qa ] ] .
e. gorsky and a. negut , `` refined knot invariants and hilbert schemes , '' _ j. math .
* 104 * ( 2015 ) 403435 , http://arxiv.org/abs/1304.3328[arxiv:1304.3328 [ math.rt ] ] .
a. borodin and i. corwin , `` macdonald processes , '' http://dx.doi.org/10.1007/s00440-013-0482-3[_probab .
theory related fields _ * 158 * no . 1 - 2 , ( 2014 ) 225400 ] , http://arxiv.org/abs/1111.4408[arxiv:1111.4408 [ math.pr ] ] .
j. shiraishi , h. kubo , h. awata , and s. odake , `` a quantum deformation of the virasoro algebra and the macdonald symmetric functions , '' http://dx.doi.org/10.1007/bf00398297[_lett . math . phys . _
* 38 * ( 1996 ) 3351 ] , http://arxiv.org/abs/q-alg/9507034[arxiv:q-alg/9507034 [ q - alg ] ] .
k. mimachi and y. yamada , `` singular vectors of virasoro algebra in terms of jack symmetric polynomials , '' _ rims kokyuroku _ * 919 * ( 1995 ) 6878 .
h. awata , y. matsuo , s. odake , and j. shiraishi , `` excited states of the calogero - sutherland model and singular vectors of the @xmath27 algebra , '' http://dx.doi.org/10.1016/0550-3213(95)00286-2[_nuclear phys .
b _ * 449 * no . 1 - 2 , ( 1995 )
347374 ] , http://arxiv.org/abs/hep-th/9503043[arxiv:hep-th/9503043 [ hep - th ] ] .
h. awata , h. kubo , s. odake , and j. shiraishi , `` quantum @xmath1139 algebras and macdonald polynomials , '' http://dx.doi.org/10.1007/bf02102595[_commun . math .
* 179 * ( 1996 ) 401416 ] , http://arxiv.org/abs/q-alg/9508011[arxiv:q-alg/9508011 [ q - alg ] ] .
l. f. alday , d. gaiotto , and y. tachikawa , `` liouville correlation functions from four - dimensional gauge theories , '' http://dx.doi.org/10.1007/s11005-010-0369-5[_lett .
* 91 * ( 2010 ) 167197 ] , http://arxiv.org/abs/0906.3219[arxiv:0906.3219 [ hep - th ] ] .
n. a. nekrasov , `` seiberg - witten prepotential from instanton counting , '' http://dx.doi.org/10.4310/atmp.2003.v7.n5.a4[_adv .
_ * 7 * no . 5 , ( 2003 ) 831864 ] , http://arxiv.org/abs/hep-th/0206161[arxiv:hep-th/0206161 [ hep - th ] ] .
a. a. belavin , a. m. polyakov , and a. b. zamolodchikov , `` infinite conformal symmetry in two - dimensional quantum field theory , '' http://dx.doi.org/10.1016/0550-3213(84)90052-x[_nucl .
* b241 * ( 1984 ) 333380 ] .
m. bershtein and o. foda , `` agt , burge pairs and minimal models , '' http://dx.doi.org/10.1007/jhep06(2014)177[_jhep_ * 06 * ( 2014 ) 177 ] , http://arxiv.org/abs/1404.7075[arxiv:1404.7075 [ hep - th ] ] .
d. gaiotto , `` asymptotically free @xmath1140 theories and irregular conformal blocks , '' http://dx.doi.org/10.1088/1742-6596/462/1/012014[_j .
* 462 * no . 1 , ( 2013 ) 012014 ] , http://arxiv.org/abs/0908.0307[arxiv:0908.0307 [ hep - th ] ] .
n. wyllard , `` a(n-1 ) conformal toda field theory correlation functions from conformal
n = 2 su(n ) quiver gauge theories , '' http://dx.doi.org/10.1088/1126-6708/2009/11/002[_jhep_ * 11 * ( 2009 ) 002 ] , http://arxiv.org/abs/0907.2189[arxiv:0907.2189 [ hep - th ] ] .
s. yanagida , `` whittaker vectors of the virasoro algebra in terms of jack
symmetric polynomial , '' http://dx.doi.org/10.1016/j.jalgebra.2011.02.039[_j
. algebra _
* 333 * ( 2011 ) 278294 ] , http://arxiv.org/abs/1003.1049[arxiv:1003.1049 [ math.qa ] ] .
v. a. alba , v. a. fateev , a. v. litvinov , and g. m. tarnopolskiy , `` on combinatorial expansion of the conformal blocks arising from agt conjecture , '' http://dx.doi.org/10.1007/s11005-011-0503-z[_lett .
* 98 * ( 2011 ) 3364 ] , http://arxiv.org/abs/1012.1312[arxiv:1012.1312 [ hep - th ] ] .
v. a. fateev and a. v. litvinov , `` integrable structure , w - symmetry and agt relation , '' http://dx.doi.org/10.1007/jhep01(2012)051[_jhep_ * 01 * ( 2012 ) 051 ] , http://arxiv.org/abs/1109.4042[arxiv:1109.4042 [ hep - th ] ] .
a. belavin and v. belavin , `` agt conjecture and integrable structure of conformal field theory for c=1 , '' http://dx.doi.org/10.1016/j.nuclphysb.2011.04.014[_nucl .
phys . _ * b850 * ( 2011 ) 199213 ] , http://arxiv.org/abs/1102.0343[arxiv:1102.0343 [ hep - th ] ] .
a. morozov and a. smirnov , `` towards the proof of agt relations with the help of the generalized jack polynomials , '' http://dx.doi.org/10.1007/s11005-014-0681-6[_lett .
phys . _ * 104 * no . 5 , ( 2014 )
585612 ] , http://arxiv.org/abs/1307.2576[arxiv:1307.2576 [ hep - th ] ] .
l. hadasz , z. jaskolski , and p. suchanek , `` proving the agt relation for @xmath1141 antifundamentals , '' http://dx.doi.org/10.1007/jhep06(2010)046[_jhep_ * 06 * ( 2010 ) 046 ] , http://arxiv.org/abs/1004.1841[arxiv:1004.1841 [ hep - th ] ] .
s. yanagida , `` norms of logarithmic primaries of virasoro algebra , '' http://dx.doi.org/10.1007/s11005-011-0502-0[_lett .
* 98 * ( 2011 ) 133156 ] , http://arxiv.org/abs/1010.0528[arxiv:1010.0528 [ math.qa ] ] .
d. maulik and a. okounkov , `` quantum groups and quantum cohomology , '' http://arxiv.org/abs/1211.1287[arxiv:1211.1287 [ math.ag ] ] .
o. schiffmann and e. vasserot , `` cherednik algebras , w - algebras and the equivariant cohomology of the moduli space of instantons on @xmath1142 , '' http://dx.doi.org/10.1007/s10240-013-0052-3[_publ .
hautes tudes sci . _
* 118 * ( 2013 ) 213342 ] , http://arxiv.org/abs/1202.2756[arxiv:1202.2756 [ math.qa ] ] .
a. braverman , m. finkelberg , and h. nakajima , `` instanton moduli spaces and @xmath1143-algebras , '' http://arxiv.org/abs/1406.2381[arxiv:1406.2381 [ math.qa ] ] .
f. nieri , `` an elliptic virasoro symmetry in 6d , '' http://arxiv.org/abs/1511.00574[arxiv:1511.00574 [ hep - th ] ] .
a. iqbal , c. kozcaz , and s .- t .
yau , `` elliptic virasoro conformal blocks , '' http://arxiv.org/abs/1511.00458[arxiv:1511.00458 [ hep - th ] ] .
h. awata and y. yamada , `` five - dimensional agt conjecture and the deformed virasoro algebra , '' http://dx.doi.org/10.1007/jhep01(2010)125 [ _ jhep _ * 01 * ( 2010 ) 125 ] , http://arxiv.org/abs/0910.4431[arxiv:0910.4431 [ hep - th ] ] .
h. awata and y. yamada , `` five - dimensional agt relation and the deformed beta - ensemble , '' http://dx.doi.org/10.1143/ptp.124.227[_prog .
_ * 124 * ( 2010 ) 227262 ] , http://arxiv.org/abs/1004.5122[arxiv:1004.5122 [ hep - th ] ] .
y. zenkevich , `` generalized macdonald polynomials , spectral duality for conformal blocks and agt correspondence in five dimensions , '' http://dx.doi.org/10.1007/jhep05(2015)131[_jhep_ * 05 * ( 2015 ) 131 ] , http://arxiv.org/abs/1412.8592[arxiv:1412.8592 [ hep - th ] ] .
h. awata , b. feigin , a. hoshino , m. kanai , j. shiraishi , and s. yanagida , `` notes on ding - iohara algebra and agt conjecture , '' _ rims kokyuroku _ * 1765 * ( 2011 ) 1232 , http://arxiv.org/abs/1106.4088[arxiv:1106.4088 [ math - ph ] ] . k. miki , `` a @xmath1144 analog of the @xmath64 algebra , '' http://dx.doi.org/10.1063/1.2823979[_j . math
_ * 48 * no . 12 , ( 2007 ) 123520 ,
b. feigin , a. hoshino , j. shibahara , j. shiraishi , and s. yanagida , `` kernel function and quantum algebra , '' _ rims kokyuroku _ * 1689 * ( 2010 ) 133152 , http://arxiv.org/abs/1002.2485[arxiv:1002.2485 [ math.qa ] ] .
b. feigin , m. jimbo , t. miwa , and e. mukhin , `` quantum toroidal @xmath65 and bethe ansatz , '' http://dx.doi.org/10.1088/1751-8113/48/24/244001[_j .
a _ * 48 * no . 24 , ( 2015 ) 244001 , 27 ] , http://arxiv.org/abs/1502.07194[arxiv:1502.07194 [ math.qa ] ] .
bourgine , m. fukuda , y. matsuo , h. zhang , and r .- d .
zhu , `` coherent states in quantum @xmath1145 algebra and qq - character for 5d super yang - mills , '' http://arxiv.org/abs/1606.08020[arxiv:1606.08020 [ hep - th ] ] .
i. burban and o. schiffmann , `` on the hall algebra of an elliptic curve , i , '' http://dx.doi.org/10.1215/00127094-1593263[_duke math .
j. _ * 161 * no . 7 , ( 2012 ) 11711231 ] , http://arxiv.org/abs/0505148[arxiv:0505148 [ math.ag ] ] .
b. feigin , k. hashizume , a. hoshino , j. shiraishi , and s. yanagida , `` a commutative algebra on degenerate @xmath1146 and macdonald polynomials , '' _ j. math .
_ * 50 * no .
9 , ( 2009 ) 095215 , 42 , http://arxiv.org/abs/0904.2291[arxiv:0904.2291 [ math.co ] ] .
h. awata and h. kanno , `` refined bps state counting from nekrasov s formula and macdonald functions , '' http://dx.doi.org/10.1142/s0217751x09043006[_int .
phys . _ * a24 * ( 2009 ) 22532306 ] , http://arxiv.org/abs/0805.0191[arxiv:0805.0191 [ hep - th ] ] .
a. iqbal , c. kozcaz , and c. vafa , `` the refined topological vertex , '' http://dx.doi.org/10.1088/1126-6708/2009/10/069[_jhep_ * 10 * ( 2009 ) 069 ] , http://arxiv.org/abs/hep-th/0701156[arxiv:hep-th/0701156 [ hep - th ] ] .
h. awata , b. feigin , and j. shiraishi , `` quantum algebraic approach to refined topological vertex , '' http://dx.doi.org/10.1007/jhep03(2012)041 [ _ jhep _ * 03 * ( 2012 ) 041 ] , http://arxiv.org/abs/1112.6074[arxiv:1112.6074 [ hep - th ] ] .
m. kashiwara , `` crystalizing the @xmath500 analog of universal enveloping algebras , '' http://dx.doi.org/10.1007/bf02097367[_commun . math .
* 133 * ( 1990 ) 249260 ] .
s. l. lukyanov and y. pugai , `` multipoint local height probabilities in the integrable rsos model , '' http://dx.doi.org/10.1016/0550-3213(96)00221-0[_nucl .
phys . _ * b473 * ( 1996 ) 631658 ] , http://arxiv.org/abs/hep-th/9602074[arxiv:hep-th/9602074 [ hep - th ] ] .
h. awata , h. kanno , a. mironov , a. morozov , a. morozov , y. ohkubo , and y. zenkevich , `` toric calabi - yau threefolds as quantum integrable systems .
@xmath1147 -matrix and @xmath1148 relations , '' http://dx.doi.org/10.1007/jhep10(2016)047[_jhep_ * 10 * ( 2016 ) 047 ] , http://arxiv.org/abs/1608.05351[arxiv:1608.05351 [ hep - th ] ] .
h. awata , h. kanno , a. mironov , a. morozov , a. morozov , y. ohkubo , and y. zenkevich , `` anomaly in rtt relation for dim algebra and network matrix models , '' http://arxiv.org/abs/1611.07304[arxiv:1611.07304 [ hep - th ] ] .
b. feigin , e. feigin , m. jimbo , t. miwa , and e. mukhin , `` quantum continuous @xmath1149 : semiinfinite construction of representations , '' http://dx.doi.org/10.1215/21562261-1214375[_kyoto j. math .
_ * 51 * no . 2 , ( 2011 )
337364 ] , http://arxiv.org/abs/1002.3100[arxiv:1002.3100 [ math.qa ] ] . b. l. feigin and a. i. tsymbaliuk , `` equivariant @xmath1150-theory of hilbert schemes via shuffle algebra , '' http://dx.doi.org/10.1215/21562261-1424875[_kyoto j. math . _ * 51 * no . 4 , ( 2011 ) 831854 ] , http://arxiv.org/abs/0904.1679[arxiv:0904.1679 [ math.rt ] ] . s. yanagida , `` whittaker vector of deformed virasoro algebra and macdonald symmetric functions , '' http://dx.doi.org/10.1007/s11005-016-0821-2[_lett
phys . _ * 106 * no . 3 , ( 2016 ) 395431 ] , http://arxiv.org/abs/1402.2946[arxiv:1402.2946 [ math.qa ] ] . h. nakajima and k. yoshioka , `` instanton counting on blowup . i. 4-dimensional pure gauge theory , '' http://dx.doi.org/10.1007/s00222-005-0444-1 [ _ invent .
_ * 162 * ( 2005 ) 313355 ] , http://arxiv.org/abs/math/0306198[arxiv:math/0306198 [ math.ag ] ] . h. nakajima and k. yoshioka , `` instanton counting on blowup .
@xmath1150-theoretic partition function , '' http://dx.doi.org/10.1007/s00031-005-0406-0[_transform .
groups _ * 10 * no
. 3 - 4 , ( 2005 ) 489519 ] , http://arxiv.org/abs/math/0505553[arxiv:math/0505553 [ math.ag ] ] .
s. yanagida , `` five - dimensional su(2 ) agt conjecture and recursive formula of deformed gaiotto state , '' http://dx.doi.org/10.1063/1.3505826[_j .
* 51 * ( 2010 ) 123506 ] , http://arxiv.org/abs/1005.0216[arxiv:1005.0216 [ math.qa ] ] .
s. yanagida , `` norm of the whittaker vector of the deformed virasoro algebra , '' http://arxiv.org/abs/1411.0462[arxiv:1411.0462 [ math.qa ] ] .
j. ding and k. iohara , `` generalization and deformation of drinfeld quantum affine algebras , '' http://dx.doi.org/10.1023/a:1007341410987[_lett .
phys . _ * 41 * ( 1997 ) 181193 ] .
c . tan , `` an m - theoretic derivation of a 5d and 6d agt correspondence , and relativistic and elliptized integrable systems , '' http://dx.doi.org/10.1007/jhep12(2013)031[_jhep_ * 12 * ( 2013 ) 031 ] , http://arxiv.org/abs/1309.4775[arxiv:1309.4775 [ hep - th ] ] .
bouwknegt and k. pilch , `` the deformed virasoro algebra at roots of unity , '' _ comm .
_ * 196 * no . 2 , ( 1998 ) 249288 , http://arxiv.org/abs/9710026[arxiv:9710026 [ math.qa ] ] .
p. bouwknegt and k. pilch , `` on deformed w algebras and quantum affine algebras , '' _ adv .
* 2 * ( 1998 ) 357397 , http://arxiv.org/abs/math/9801112[arxiv:math/9801112 [ math.qa ] ] .
m. fukuda , s. nakamura , y. matsuo , and r .- d .
zhu , `` sh@xmath1151 realization of minimal model cft : triality , poset and burge condition , '' http://dx.doi.org/10.1007/jhep11(2015)168[_jhep_ * 11 * ( 2015 ) 168 ] , http://arxiv.org/abs/1509.01000[arxiv:1509.01000 [ hep - th ] ] .
y. ohkubo , h. awata , and h. fujino , `` crystallization of deformed virasoro algebra , ding - iohara - miki algebra and 5d agt correspondence , '' http://arxiv.org/abs/1512.08016[arxiv:1512.08016 [ math - ph ] ] .
h. awata , h. kubo , s. odake , and j. shiraishi , `` virasoro type symmetries in solvable models , '' in _ extended and quantum algebras and their applications to physics tianjin , china , august 19 - 24 , 1996_. 1996 .
http://arxiv.org/abs/hep-th/9612233[arxiv:hep-th/9612233 [ hep - th ] ] .
m. taki , `` on agt - w conjecture and q - deformed w - algebra , '' http://arxiv.org/abs/1403.7016[arxiv:1403.7016 [ hep - th ] ] .
a. morozov and y. zenkevich , `` decomposing nekrasov decomposition , '' http://dx.doi.org/10.1007/jhep02(2016)098[_jhep_ * 02 * ( 2016 ) 098 ] , http://arxiv.org/abs/1510.01896[arxiv:1510.01896 [ hep - th ] ] .
m. taki , `` refined topological vertex and instanton counting , '' http://dx.doi.org/10.1088/1126-6708/2008/03/048[_jhep_ * 03 * ( 2008 ) 048 ] , http://arxiv.org/abs/0710.1776[arxiv:0710.1776 [ hep - th ] ] .
y. ohkubo , `` existence and orthogonality of generalized jack polynomials and its @xmath13-deformation , '' http://arxiv.org/abs/1404.5401[arxiv:1404.5401 [ math - ph ] ] .
s. mironov , a. morozov , and y. zenkevich , `` generalized jack polynomials and the agt relations for the @xmath1152 group , '' http://dx.doi.org/10.1134/s0021364014020076[_jetp lett . _
* 99 * ( 2014 ) 109113 ] , http://arxiv.org/abs/1312.5732[arxiv:1312.5732 [ hep - th ] ] .
n. h. jing , `` vertex operators and hall - littlewood symmetric functions , '' http://dx.doi.org/10.1016/0001-8708(91)90072-f[_adv .
_ * 87 * no . 2 , ( 1991 )
226248 ] .
h. awata , h. kanno , t. matsumoto , a. mironov , a. morozov , a. morozov , y. ohkubo , and y. zenkevich , `` explicit examples of dim constraints for network matrix models , '' http://dx.doi.org/10.1007/jhep07(2016)103[_jhep_ * 07 * ( 2016 ) 103 ] , http://arxiv.org/abs/1604.08366[arxiv:1604.08366 [ hep - th ] ] . | in this thesis , we obtain the formula for the kac determinant of the algebra arising from the level @xmath0 representation of the ding - iohara - miki algebra .
this formula can be proved by decomposing the level @xmath0 representation into the deformed @xmath1-algebra part and the @xmath2 boson part , and using the screening currents of the deformed @xmath1-algebra .
it is also discovered that singular vectors obtained by its screening currents correspond to the generalized macdonald functions .
moreover , we investigate the @xmath3 limit of five - dimensional agt correspondence . in this limit ,
the simplest 5d agt conjecture is proved , that is , the inner product of the whittaker vector of the deformed virasoro algebra coincides with the partition function of the 5d pure gauge theory .
furthermore , the r - matrix of the ding - iohara - miki algebra is explicitly calculated , and its general expression in terms of the generalized macdonald functions is conjectured .
@xmath4 + phd thesis submitted to nagoya university + graduate school of mathematics + march 2017 |
the observed evolution of the space density of clusters of galaxies provides a powerful constraint on the underlying cosmological model .
many authors have demonstrated both analytically and numerically that the expected abundance of clusters , as a function of cosmic epoch , is a sensitive test of the mean mass density of the universe ( @xmath14 ) and the type of dark matter ( press & schechter 1974 ; lacey & cole 1993 , 1994 ; oukbir & blanchard 1992 , 1997 ; henry 1997 ; eke et al .
1996 , 1998 ; viana & liddle 1996 , 1999 ; bryan & norman 1998 ; reichart et al . 1999 ; borgani et al .
1999 ) .
measurements of the evolution of the cluster abundance have made significant progress over the past decade .
for example , in their seminal work , gioia et al .
( 1990 ) and henry et al .
( 1992 ) computed the luminosity function of x ray clusters extracted from the _ einstein _ extended medium sensitivity survey ( emss ) and concluded that the x - ray cluster luminosity function ( xclf ) evolved rapidly over the redshift range of @xmath15 .
the launch of the rosat satellite heralded a new era of investigation into the xclf .
the rosat all sky survey ( rass ) has provided new determinations of the local xclf and has demonstrated that there is little observed evolution in the xclf out to @xmath16 ( ebeling et al .
1997 ; de grandi et al .
1999 ) in agreement with the earlier work of kowalski et al .
in addition , the rosat satellite has supported several investigations of the distant x ray cluster population ( rixos , castander et al .
1995 ; sharc , burke et al .
1997 , romer et al . 1999 ; rdcs , rosati et al .
1998 ; warps , jones et al . 1998 ; vikhlinin et al .
1998a ; nep , henry et al .
initially , such investigations reported a deficit of high redshift , low luminosity clusters consistent with the original emss result ( castander et al .
however , over the last few years , there has been a growing consensus for a non evolving xclf .
first , nichol et al . (
1997 ) re examined the emss cluster sample and determined that the statistical evidence for evolution of the emss xclf had decreased in light of new rosat data .
second , several authors have now conclusively shown that the xclf does not evolve out to @xmath17 for cluster luminosities of @xmath18 ( collins et al . 1997 ; burke et al . 1997 ; rosati et al . 1998 ; jones et al . 1998 ) .
above @xmath19 , recent work has indicated that the xclf may evolve rapidly in agreement with the original claim of gioia et al .
reichart et al .
( 1999 ) highlighted a deficit of luminous ( @xmath20 ) emss clusters at @xmath21 _
i.e. _ the emss survey has both the sensitivity and area to find such clusters but does not detect them .
moreover , vikhlinin et al .
( 1998b ) has recently reported evidence for a deficit of luminous clusters at @xmath22 based on the @xmath23 rosat survey ( vikhlinin et al .
1998a ) . in this paper ,
we report on the first determination of the bright end of the xclf that is independent of the emss . in sections [ sample ] & [ sf ] , we outline the bright sharc sample of clusters used herein and its selection function . in sections
[ lfs ] & [ discuss ] , we present the derivation of the xclf and discuss its implications . throughout this paper , we use @xmath24 and @xmath25 to be consistent with other work in this field .
all quoted luminosities are in the hard rosat passband [ @xmath26 kev ] and are aperture and k corrected ( see romer et al .
1999 for details ) .
the details of the construction of the bright sharc survey are presented in romer et al .
the bright sharc was constructed from 460 deep ( @xmath27 ksecs ) , high galactic latitude ( @xmath28 ) , rosat pspc pointings which cover a unique area of @xmath29 . using a wavelet based detection algorithm , @xmath30 x
ray sources were detected in these pointings of which @xmath31 were measured to be significantly extended ( @xmath32 ; see nichol et al .
1997 ) relative to the rosat pspc point spread function .
the bright sharc represents the brightest 94 of these 374 extended cluster candidates above a rosat count rate of 0.0116 @xmath33 .
this corresponds to a flux limit of @xmath34 [ 0.52.0 kev ] for the average neutral hydrogen column density of the bright sharc and a cluster temperature of @xmath35 kev . over the past two years
, we have optically identified the most likely x ray emitter for @xmath36 of these @xmath37 bright sharc cluster candidates and have discovered @xmath38 clusters , @xmath39 groups of galaxies and @xmath40 nearby galaxies ( the remainder are blends of x ray sources _
e.g. _ agns & stars ; see romer et al .
we find @xmath0 clusters in the range @xmath41 ( median redshift of @xmath2 ) and have independently detected cluster rxj0152 - 7363 ( @xmath42 based on 3 galaxy redshifts obtained at the cfht ) which is one of the most luminous , high redshift x
ray clusters ever detected ( see romer et al . 1999 ) .
this cluster has also been detected by the warps and rdcs surveys ( see ebeling et al . 1999 ; rosati , private communication ) .
an important part of any survey is a solid understanding of the selection function _
i.e. _ the efficiency of finding objects as a function of both cosmological and operational parameters . in the case of the emss cluster sample ,
the selection function is somewhat straightforward since the emss optically identified all sources regardless of their observed x ray extent .
this is not the case for the bright sharc and therefore , the most direct way of modelling the selection function is through monte carlo simulations .
the details of such simulations are given in adami et al .
( 1999 ) but we present here some initial results . the bright sharc selection function is obtained by adding artificial clusters to pspc pointings and determining if these clusters would have satisfied the bright sharc selection criteria . in this way
, we can accurately model the effects of the signal to noise cut , the extent criteria and blending of sources . to date
, we have only simulated artificial clusters at four different intrinsic luminosities , @xmath43 ( @xmath44 _ i.e. _ equally spaced in log space ) , and covering a redshift range of @xmath45 in steps of @xmath46 .
these artificial clusters were constructed assuming an isothermal king profile with @xmath47 kpc and @xmath48 .
the effects of changing the cluster profile and its parameters are explored in detail in adami et al .
( 1999 ) . for each combination of @xmath49 and @xmath50 ,
we added , one at a time , 10 artificial clusters to each of 10 pspc pointings ( randomly chosen from all pointings available in the sharc survey ) .
the positions of the clusters in these pointings were chosen at random but recorded for later use when computing the area surveyed ( see below ) . for each artificial cluster , we computed the expected number of rosat pspc counts for its given luminosity , redshift and the exposure time of the pointing .
we then took a poisson deviate about the expected number of counts and distributed the counts at random assuming a redshifted king profile .
the pointing was then processed in the same fashion as the real data ( see romer et al . 1999 ; nichol et al .
1997 ) thus allowing us to determine if the cluster would have been selected for the bright sharc . in figure [ area
] , we show the results of these initial simulations . the effective area of the bright sharc was computed by splitting the pspc field of view into 4 annulii ( @xmath51 , @xmath52 , @xmath53 and @xmath54 ; the central part of the pspc was excluded ) and multiplying the area in each of these annulii by the measured success rate of detecting clusters in these same annulii ( as a function of the input cluster redshift and luminosity ) .
we then summed the effective area in these annulii over all pointings used again as a function of luminosity and redshift to provide the total effective area sampled by the simulations .
finally , we re scaled the results to obtain the expected area for all 460 pspc pointings used in the bright sharc .
the luminosity function of the bright sharc was determined using the @xmath55 methodology outlined in avni & bahcall ( 1980 ) , henry et al .
( 1992 ) and nichol et al .
( 1997 ) , where @xmath56 is the available sample volume for any given cluster in the survey . using the selection function presented in figure [ area ] ,
@xmath57 can be computed for a cluster of luminosity @xmath58 using @xmath59 where @xmath60 and @xmath61 are the lower and upper bounds of the redshift shell of interest , @xmath62 is the volume per unit solid angle for that redshift shell and @xmath63 is the effective area of the bright sharc from figure [ area ] . in practice ,
the integral in eqn .
[ va ] is replaced by a sum over the discrete values of @xmath63 obtained from the simulations .
linear interpolation was used where necessary to obtain finer resolution in both luminosity and redshift space .
the luminosity function was derived by summing the @xmath56 values for all clusters in the bright sharc as a function of luminosity _
_ @xmath64 where @xmath65 is the width of the luminosity bins and @xmath66 is the number of clusters in that luminosity bin ( table [ datat ] ) . for the results presented here , we have restricted ourselves to the luminosity range @xmath67 . in figure
[ lf ] , we present the bright sharc xclf and compare it to the southern sharc xclf ( burke et al . 1997 ) and measurements of the local xclf ( ebeling et al .
1997 ; de grandi et al . 1999 ) .
we provide , in table [ datat ] , the data points displayed in figure [ lf ] together with the redshift and luminosity ranges studied .
we also provide the number of clusters in each bin .
we have not performed a parametric fit to the data because of the limited dynamic range in luminosity available from our present simulations . [ cols="^,^,^,^,^ " , ]
figure [ lf ] demonstrates that the high redshift xclf does not evolve below @xmath68 , or below @xmath69 in the xclf ( @xmath70 from ebeling et al .
1997 ) . using a kolmogorov
smirnov ( ks ) test similar to that discussed by de grandi et al .
( 1999 ) , we find that the bright sharc unbinned data is fully consistent with the low redshift xclf over the luminosity range @xmath71 ( we find a ks probability of 0.32 ) .
this result is consistent with previous work ( nichol et al .
1997 ; burke et al . 1997 ; rosati et al . 1998 ; jones et al . 1998 ) but pushes the evidence for a non evolving xclf to the highest luminosities presently reached by rosat data .
this is illustrated by the fact that we find 6 distant clusters in the luminosity range @xmath72 ( table [ datat ] ) which is more than any other rosat archival survey .
the issue therefore becomes the degree of observed evolution in the xclf above @xmath69 .
we have investigated evolution in the xclf above @xmath69 in two separate ways .
first , we have one very luminous high redshift cluster in the bright sharc rxj0152.7 - 1357 ( @xmath42 ; @xmath73 ) which probes the xclf above @xmath69 . the implied position of this cluster in the xclf is given in figure [ lf ] and table [ datat ] ( we used a redshift range of @xmath74 and luminosity range of @xmath75 when computing the volume sampled by this cluster ) .
as can be seen in figure [ lf ] , the implied space density of rxj0152 - 1357 agrees with the local xclf and may be evidence for a non - evolving xclf above @xmath76 .
however , we must remain cautious since rxj0152.7 - 1357 has a complex x ray morphology indicative of an on going merger which may have enhanced its luminosity ( see ebeling et al .
such disturbed or non spherical morphologies ( both in the x rays and optical ) appear to be common at these high redshifts _
e.g. _ ms1054.4 - 0321 at @xmath77 ( donahue et al .
1999 ) and rxj1716.6 + 6708 at @xmath78 ( henry et al . 1998 ; gioia et al . 1999 ) and may indicate that we are witnessing the epoch of massive cluster formation .
the second path of investigation is to repeat the analysis of collins et al .
( 1997 ) and vikhlinin et al .
( 1998b ) and compute the number of expected bright sharc clusters at these bright luminosities assuming a non evolving xclf . in the luminosity range @xmath79 and redshift range @xmath80
, we would predict @xmath8 clusters based on the de grandi et al .
( 1999 ) xclf ( or @xmath81 clusters using the ebeling et al 1997 xclf ) . at present
, the bright sharc contains only one confirmed cluster in this range , rxj1120.1 + 4318 at @xmath82 and @xmath83 .
the poisson statistical significance ( gehrels 1986 ) of this observed deficit is @xmath9 ( or @xmath84 for ebeling et al .
1997 ) . one way to increase the statistical significance of any possible deficit of high redshift x ray luminous clusters is to combine the bright sharc and @xmath85 survey of vikhlinin et al .
( 1998a ) as both surveys should have similar selection functions .
we have determined that @xmath86 rosat pspc pointings are in common between the two surveys ( see romer et al .
1999 ) , or @xmath87 of the area of the bright sharc .
vikhlinin et al .
( 1998b ) have noted that their survey probably contains only 2 candidate clusters above @xmath88 and @xmath22 ( based on photometric redshift estimates ) , while they would have expected 9 clusters above this redshift and luminosity limit .
in fact , bright sharc spectroscopy of one of these two candidate clusters rxj1641 + 8232 shows that it is at @xmath89 ( based on 3 galaxy redshifts ) . the other candidate cluster rxj1641 + 4001 is not in the bright sharc .
therefore , the bright sharc plus the @xmath90 survey cover a total area of @xmath91 and scaling the aforementioned numbers appropriately , we would expect @xmath13 ( using the de grandi et al .
1999 xclf ) or @xmath92 ( using the ebeling et al .
1997 xclf ) clusters in the luminosity range @xmath93 and redshift range @xmath12 . by comparison , this joint survey contains only 1 confirmed@xmath94 x ray luminous , high redshift cluster . from gehrels ( 1986 ) , the statistical significance of this deficit is now 99.5% or 97.5% respectively .
we note however that there are still candidates in both surveys that require further optical follow up _
i.e. _ one of the three unidentified bright sharc candidates mentioned in section [ sample ] could be a high redshift cluster .
the above analyses suffer from small number statistics and incomplete optical follow up .
moveover , the local xclfs are also affected by small number statistics at these high x
ray luminosities .
therefore , the issue of evolution above @xmath95 in the xclf remains unclear .
however , we note that four independent surveys of distant clusters of galaxies have now seen a potential deficit of x - ray luminous high redshift clusters ( sharc , warps , emss , @xmath96 survey ) .
it is worth stressing that some level of xclf evolution is expected above @xmath95 at high redshift and the degree of such evolution is a strong indicator of @xmath14 ( see blanchard & oukdir 1992 , 1997 ; reichart et al .
1999 ) .
the way to resolve these problems is to construct larger samples of clusters over bigger areas of the sky .
this is certainly possible for the local xclf _
e.g. _ using the on going reflex survey of clusters being constructed from the rass ( see bhringer et al .
however , for the distant xclf , it is unlikely that a significant amount of further area ( @xmath97 ) can be added to the present surveys using the existing rosat pointing archive ; one may have to wait for the xmm satellites ( see romer 1998 ) .
the next major improvements to the results presented in this paper will be to obtain a more detailed view of the bright sharc selection function as well as to complete the optical follow up of the remaining candidates .
another improvement would be the possible combination of all the rosat distant surveys ( sharc , @xmath98 , rdcs , nep & warps ) thus maximising the amount of volume sampled at high redshift .
the authors would like to thank alain blanchard , jim bartlett , francisco castander , harald ebeling , pat henry , andrew liddle , piero rosati & alex vikhlinin for helpful discussions over the course of this work .
this research was supported through nasa adp grant nag5 - 2432 ( at nwu ) and nasa ltsa grant nag5 - 6548 ( at cmu ) .
am thanks the carnegie mellon undergraduate research initiative for financial assistance .
we thank alain mazure and the igrap
las ( marseille , france ) for their support .
adami , c. , ulmer , m. p. , pildis , r. a. , romer , a. k. , holden , b. p. , & nichol , r. c. , 1999 , apj , in preparation + avni , y. , bahcall , j. n. , 1980 , apj , 235 , 694 + bhringer , h. , guzzo , l. , collins , c.a . , neumann , d.m .
, schindler , s. , schuecker , p. , cruddace , r. , chincarini , g. , de grandi , s. , edge , a.c . ,
macgillivray , h.t . , shaver , p. , vettolani , g. , voges , w. 1998 , in wide field surveys in cosmology , editions frontieres .
isbn 2 - 8 6332 - 241 - 9 , 261 .
+ borgani , s. , rosati , p. , tozzi , p. , norman , c. 1999 , in press ( astro - ph/9901017 ) .
+ bryan , g. l. , norman , m. l. , 1998 , apj , 495 , 80 + burke , d.j . ,
collins , c.a . ,
romer , a.k . ,
holden , b.p . ,
nichol , r.c .
1997 , apj , 488 , l83 .
+ castander , f.j . ,
bower , r.g . ,
ellis , r.s .
, aragon - salamanca , a. , mason , k.o . ,
hasinger , g. , mcmahon , r.g . ,
carrera , f.j . ,
mittaz , j.p.d .
, perez - fournon , i. , lehto , h.j .
1995 , nature , 281 , 59 .
+ collins , c.a . , burke , d.j . ,
romer , a.k . , sharples , r.m . , nichol , r.c .
1997 , apj , 479 , l117 .
+ de grandi , s. , bhringer , h. , guzzo , l. , molendi , s. , chincarini , g. , collins , c.a . , cruddace , r. , neumann , d.m . ,
schindler , s. , schuecker , p. , & g. , voges , w. , 1999 , apj , 514 , in press ( astro - ph/9902067 ) + donahue , m. , voit , g.m . ,
gioia , i. , luppino , g. , hughes , j.p . ,
stocke , j.t .
1998 , apj , 502 , 550 .
+ ebeling , h. , edge , a. c. , fabian , a. c. , allen , s. , crawford , c. s. , bhringer , h. , 1997 , apj , 479 , l101 + ebeling , h. , edge , a.c . , bhringer , h. , allen , s.w . ,
crawford , c.s . ,
fabian , a.c . , voges , w. , huchra , j.p .
1998 , mnras , 301 , 881 .
+ ebeling , h. , jones , l.r . , perlman , e. , scharf , c. , horner , d. , wegner , g. , malkan , m. , fairley , b. , mullis , c. r. , 1999 , , submitted ( astro - ph/9905321 ) eke , v. r. , cole , s. , frenk , c. , 1996 , mnras , 282 , 263 + eke , v. r. , cole , s. , frenk , c. , & henry , j. p. , 1998 ,
mnras , 298 , 1145 + gehrels , n. 1986 , apj , 303 , 336 + gioia , i. m. , henry , j.p . , maccacaro , t. , morris , s.l . , stocke , j.t . ,
wolter , a. 1990 , , 356 , l35 .
+ gioia , i. m. , henry , j.p . , mullis , c. r. , ebeling , h. , wolter , a. 1999 , , accepted ( astro - ph/9902277 )
. + henry , j.p . ,
gioia , i.m . ,
maccacaro , t. , morris , s.l . ,
stocke , j.t . , wolter , a. 1992 , apj , 386 , 408 .
+ henry , j.p .
1997 , apj , 489 , l1 .
+ henry , j.p . ,
gioia , i.m . , mullis , c. r. , clowes , d. l. , luppino , g. a. , bhringer , h. , briel , u. g. , voges , w. , huchra , j. p. , 1997
, aj , 114 , 1293 + jones , l.r . , scharf , c. , ebeling , h. , perlman , e. , wegner , g. , malkan , m. , horner , d. 1998 , apj , 495 , 100 .
+ kowalski , m. p. , ulmer , m. p. , cruddace , r. g. , wood , k. s. , 1984 , apj supp .
, 56 , 403 .
+ lacey , c. , cole , s. 1993 , mnras , 262 , 627 + lacey , c. , cole , s. 1994 , mnras , 271 , 676 + nichol , r.c . , holden , b.p . ,
romer , a.k . ,
ulmer , m.p . ,
burke , d.j . , collins , c.a . ,
apj 1997 , 481 , 644 .
+ oukbir , j. , blanchard , a. , 1992 , a&a , 262 , 21 + oukbir , j. , blanchard , a. , 1997 , a&a , 317 , 365 + press , w. h. , schechter , p. 1974
, , 187 , 425 + reichart , d.e . , nichol , r.c . , castander , f.j . , burke , d.j . , romer , a.k . , holden , b.p . , collins , c.a . , ulmer , m.p . ,
apj 1999 , accepted ( astro - ph/9802153 ) .
+ romer , a. k. , 1998 , in wide field surveys in cosmology , editions frontieres .
isbn 2 - 8 6332 - 241 - 9 , 261 .
( astro - ph/9809198 ) + romer , a.k . ,
nichol , r.c . ,
holden , b.p . , ulmer , m.p . ,
pildis , r.a . ,
merrelli , a. j. , adami , c. a. , burke , d.j . ,
collins , c.a . ,
metevier , a. , kron , r. commons , k. , 1999 , apjs , submitted .
+ rosati , p. , della ceca , r. , norman , c. , giacconi , r. , 1998 , apj , 492 , 21 .
+ viana , p.t.p . ,
liddle , a.r .
1996 , mnras , 281 , 323 + viana , p.t.p . ,
liddle , a.r .
1999 , in press ( astro - ph/9803244 ) + vikhlinin , a. , mcnamara , b.r . ,
forman , w. , jones , c. , quintana , h. , hornstrup , a. 1998a , apj 502 , 558 .
+ vikhlinin , a. , mcnamara , b.r . ,
forman , w. , jones , c. , quintana , h. , hornstrup , a. 1998b , apj 498 , l21 . | we present here initial results on the x ray cluster luminosity function ( xclf ) from the bright sharc sample of distant x ray clusters of galaxies .
this sample is 97% complete in its optical identifications and contains @xmath0 x ray luminous clusters in the redshift range @xmath1 ( median @xmath2 ) and @xmath3 [ 0.52.0 kev ] .
we present a preliminary selection function for the bright sharc based on monte carlo simulations . using this selection function ,
we have computed the bright sharc xclf and find it to be fully consistent with a non evolving xclf to @xmath4 and @xmath5 . at @xmath6
, we find evidence for a deficit of clusters compared to that expected from a non evolving xclf .
we detect only one such cluster in the redshift range @xmath7 when we would expect @xmath8 clusters based on the local xclf of de grandi et al .
( 1999 ) .
the statistical significance of this deficit is @xmath9 . to increase the statistical significance of this possible deficit ,
we have combined the bright sharc and the @xmath10 survey of vikhlinin et al ( 1998a ) .
this joint survey covers @xmath11 and contains only one confirmed @xmath12 , @xmath6 cluster , while we would expect @xmath13 such clusters based on the local xclf ( de grandi et al .
1999 ) .
the statistical significance of the deficit in this joint survey increases to 99.5% .
these results remain preliminary because of incompletenesses in the optical follow up and uncertainties in the local xclf . |
this study finds the leaky pipeline phenomenon exists for women past the postdoctoral level at a level of around 15% .
some may wonder if this is a big enough leak to be a problem . in human terms , however , a 15% leak means that we are missing one out of six women who , in an equitable society , would have been physics faculty members .
there are so few women at the faculty level in physics , that losing one out of every six is in fact a serious concern .
especially if we think about what it must take to convince someone to leave a field when , by that point in their careers , they have committed their working lives to physics , and have gone through at least a decade of higher education to get there .
not all women who become physics faculty members have experienced gender discrimination during their careers . however , many do , and it is unfortunate that the combination of gender discrimination and a ` glass ceiling ' phenomenon in the field is preventing more women from becoming physics faculty members .
the author is an experimental particle physicist , and has observed over the years the serious obstacles that her female colleagues have had to face as they try to advance in the field .
there is indeed widespread discrimination against many women in physics , and women with children seem to be particularly vulnerable ; for instance , the author is personally acquainted with three female physicists who , after having children , had to work for free or a substantially reduced rate compared to their peers , simply to remain in the field .
the only other choice available to them was to simply drop out of the academic pipeline all together .
conversely , the author knows literally hundreds of male physicists past the doctoral level , but is not aware of a single male who has had their pay cut off or substantially reduced for any reason .
given some of the chilling incidents of discrimination against females that the author has personally observed to transpire within physics academia , it is somewhat surprising that the relative leak of females in the academic pipeline past the postdoctoral level is * only * 15% .
a male colleague once mentioned to the author that he felt very sorry for many of his female colleagues in physics because he felt the message that they were persistently given was that ` yes !
the good news is that you * can * succeed in physics as a female !
( you just need to be prepared to chew your own leg off to do so ) ' .
tragically , some of the women who have ultimately made it to the faculty level likely did have to ` chew their own leg off ' to get there .
even more tragically , their stories are almost never told because they ( quite rightly ) fear repercussions to their career if they speak out .
the chilly climate that removes one out of every six female potential physics faculty members needs to be changed if particpation of women in physics is to be increased at all levels .
probability that a female who graduated in a particular year will be a physics professor in 2005 at one of the ` top 50 ' american physics universities .
points are the actual distribution , and the histogram indicates the predicted distribution , obtained assuming that females and males from the same graduating class have the same relative probability of being a professor in 2005 . ]
year - of - phd of female professors in 2005 ( points ) , for assistant , associate and full professors . the histograms indicate the predicted year - of - phd distributions , obtained assuming that females and males from the same graduating class have the same relative probability of being a professor in 2005 . ]
year - of - phd of all female professors in 2005 ( points ) .
the histogram indicates the predicted year - of - phd distribution , obtained assuming that females and males from the same graduating class have the same relative probability of being a professor in 2005 . ]
american institute of physics 2005 report _ women in physics and astronomy _
+ http://www.aip.org/statistics/trends/reports/women05.pdf national research council 1995 report _ research - doctorate programs in the united states : continuity and change _ | the author has recently examined the departmental web pages of the ` top 50 ' physics research universities , as ranked by the national research council ( nrc ) @xcite .
most of the departmental web pages contained biographical data ( ie ; year and institute of phd , etc ) of their faculty members .
of the approximately 1750 faculty members at the ` top 50 ' universities that were examined , approximately 100 were female , and around 1425 had available biographical data .
based on this data , the * predicted * fractions of female faculty members at the ` top 50 ' universities are @xmath0,@xmath1 , and @xmath2 at the assistant , associate , and full faculty levels , respectively . the * observed * fractions are @xmath3 , @xmath4 , @xmath5 , respectively . the overall observed number of women faculty is about 15% less than expected , and the depletion is statistically significant .
unfortunately , the study finds that the `` leaky pipeline '' is found to be alive and well for women in academic physics above the postdoctoral level , at all stages of the faculty career ladder .
this result is stark contrast with the conclusion of the american institute of physics ( aip ) 2005 report on women in physics and astronomy ; the aip report concludes that women are actually * more * likely to be hired at the faculty level than their male peers . in this paper
, we will discuss the two key flaws in the aip analysis that led to their faulty conclusion , then describe in detail the analysis performed by the author that corrects these flaws to get an accurate estimate of the ` leakiness ' of the academic pipeline for women physicists past the postdoctoral level .
_ to be submitted to the american journal of physics _
many studies have documented the `` leaky pipeline '' phenomenon for women in the academic hard sciences ( see , for instance , references @xcite and @xcite ) . a recent report that caught the author s interest was the american institute of physics 2005 report on women in physics and astronomy @xcite ; this report is rather unique in that it concludes that the leaky pipeline phenomenon does not exist past the doctoral level .
in fact , the report concludes that the observed fraction of female faculty members is actually * higher * than expected .
however , careful examination of the report reveals that the analysis that led to this conclusion was flawed ; first , the report lumps faculty members at phd granting universities together with faculty members at teaching colleges . at teaching colleges ,
the faculty are more likely to be female , yet much less likely to teach physics to physicists .
thus the aip report gives little indication of the fraction of women at the faculty level at the universities in america that produce the majority of physics doctoral degrees .
second , the analysis performed for the report did not properly take into account the differing age distributions of male and female faculty members ( the report used instead the combined age distribution of both males and females , which of course is completely dominated by the males , since they constitute over 90% of the sample ) .
this flaw has a significant effect on the predicted fraction of female faculty ( the predicted fraction goes up when the analysis is performed taking the differing age distributions properly into account ) . to perform a detailed analysis that corrects both of these problems , the author began with an examination of the departmental web pages of the ` top 50 ' physics universities , as ranked by the national research council ( nrc ) @xcite .
these universities produce the majority of bsc s and phd s in america .
most of the departmental web pages contained biographical data ( ie ; year and institute of phd , etc ) of their faculty members . in this study astronomers
were excluded because they have a different fraction of women participating in the field than other areas of physics .
adjunct , visiting , research , and affiliated professors were also excluded .
the study also excluded faculty members who had received their degree from a non - american institution . of 1743 faculty members at
the ` top 50 ' universities that were ultimately examined , 101 were female , and a total of 1425 had available biographical data . to obtain the predicted fraction of female professors from this data , we begin with the number of phd s granted each year to both males and females in america @xcite ( see figure [ fig : phd ] ) .
we then work out the probability that a male in a particular phd graduating class will be a professor at one of the ` top 50 ' universities in 2005 ; we do this by dividing the year - of - phd distribution of male professors with the distribution of the number of male phd s graduating each year .
if the leaky pipeline does not exist , the female ` be - a - professor - in-2005 ' probability will be exactly the same as for the males in each graduating class . in this manner , we obtain a prediction of the number of female professors we expect to see at the ` top 50 ' universities in 2005 .
figure [ fig : prob ] shows the actual probability versus year - of - phd that a female physicist will be a faculty member at one of the ` top 50 ' universities in 2005 . the histogram indicates the predicted distribution ,
obtained assuming that females have the same relative probability of being a professor in 2005 as males from the same graduating class .
the actual distribution is systematically lower than the predicted distribution .
figure [ fig : all ] shows the year - of - phd distributions of female assistant , associate , and full professors .
the histograms again indicate the predicted distribution , obtained assuming that females and males from the same graduating class have the same relative probability of being a professor in 2005 .
figure [ fig : summary ] shows the year - of - phd distributions of all female professors .
every point in the actual distribution is lower than the predicted .
based on this data , the * predicted * fractions of female faculty members at the ` top 50 ' universities are @xmath0,@xmath1 , and @xmath2 at the assistant , associate , and full faculty levels , respectively .
the * observed * fractions are @xmath3 , @xmath4 , @xmath5 , respectively .
it is interesting to note that the fraction of female associate professors is actually higher than predicted .
however , figure [ fig : summary ] shows an overall depletion of professors for years - of - phd 1984 and onwards .
it thus appears that the excess may well be due to women languishing longer at the associate professorship level than their male peers ( ie ; the excess probably reflects a ` clog ' in the academic pipeline at the associate professor level ) . the overall observed number of women faculty is about 15% less than expected , and the depletion is statistically significant at a level of @xmath6 standard deviations . |
in biological systems , a large number of elements , such as cells and organs , interact with each other and adapt system - wide behavior spontaneously in response to changes of environment and of physical constraints .
such spontaneous and adaptive dynamics have been studied in the framework of collective dynamics @xcite . to clarify the mechanism of adaptability , however , we need to investigate how the dynamics of interactions between elements should depend on the global behavior to establish adaptability .
collective dynamics has been studied as a distributed system to investigate how global structures emerge from the interaction of elements .
many types of distributed systems have been modeled by using differential equations , and the system behaviors have been described by appropriate basin switching between attractors @xcite .
for instance , previous studies investigated retrieval of a specific pattern by neural oscillator networks @xcite . in these studies , the stable pattern was predicted by lyapunov function analysis , and the basin switching was accomplished by controlling the potential profiles .
complex dynamics observed in brain systems have been studied in terms of chaotic itinerancy .
transient behavior between ordered and disordered states has been shown in some coupled oscillator systems . @xcite .
the aim of this study is to propose a distributed system showing appropriate basin switching by using the pathfinding problem .
pathfinding strategies have been proposed in both theoretical and experimental contexts . in laboratory experiments , pathfinding algorithms based on self - organization processes have been studied @xcite . in @xcite , a continuous model for the experimental results of @xcite has been proposed . in their model
, a conservation mass has been assumed , which allows the system to find the shortest path @xcite . in the present study
, we modify the loop searching system proposed in @xcite to apply it to the pathfinding problem in the framework of a distributed system ; that is , no conservation mass is assumed .
the path connecting the start and goal points is represented by a loop path containing both the start point and the goal point .
in such a network structure , a pathfinding problem can be regarded as a loop finding problem .
therefore , we propose a two - layer network system to construct such loop networks . in addition , in order to improve the finding time of the desired path , the effects of interactions between the dynamics of the two layers are considered .
the presented system shows the following three properties .
( i ) the system spontaneously finds a path connecting the start and goal nodes if such a path exists .
( ii ) the system shows the self - recovery property , that is , the system finds another possible path when the existing path is broken due to the removal of paths .
( iii ) in a tree structure network , the finding time increases in proportion to the depth of the tree .
the graphs we consider in this study have unidirectional edges between vertices and are acyclic .
the graph has two special vertices referred to as the start vertex and the goal vertex . in this section ,
we propose a construction procedure for a system that shows properties ( i ) to ( iii ) .
the path is defined by a stable solution of the system . to avoid confusion
, we use the terms , _ vertex _ and _ edge _ for the given graph and _ node _ and _ link _ for the same properties of the system .
the assumptions for the system are given as follows .
* we assume two layers exist , and a pair of nodes is placed on the layers at every vertex position ( fig.[fig : bilayer ] ) . for notational convenience , we refer to one of the layers as the p - layer ( positive ) and to the other one as the n - layer ( negative ) .
the node numbers corresponding to the @xmath0th vertex in the p - layer and in the n - layer are given by @xmath1 and @xmath2 , respectively .
* the excitatory links in the p - layer and the n - layer take place along the forward and backward direction of the edges , respectively .
that is , if the network of the p - layer has an excitatory link from node @xmath3 to node @xmath1 , then the network of the n - layer has an excitatory link from node @xmath2 to node @xmath4 .
bidirectional inhibitory links take place at every branching point of excitatory links .
additional assumptions for the inhibitory interactions are given in ( s7 ) and ( s7 ) .
* each node contains a large number of oscillators , and the state of each node is determined by the dynamics of the group oscillators corresponding to the node . * each p - layer and n - layer has two special nodes , corresponding to the start vertex and the goal vertex , referred to as the start node and the goal node , respectively .
the start nodes of the p - layer and the n - layer are labeled @xmath5 and @xmath6 , respectively .
similarly , the goal nodes of the p - layer and the n - layer are labeled @xmath7 and @xmath8 , respectively .
the network has an excitatory link from @xmath6 to @xmath5 and from @xmath7 to @xmath8 . * the network in the p - layer and the n - layer without the two links placed at the start and goal nodes is acyclic .
if a connecting path from the start vertex to the goal vertex exists , then a connecting path from the start node to the goal node exists in the p - layer and a connecting path from the goal node to the start node exists in the n - layer from ( s2 ) .
therefore , from ( s5 ) , loops of excitatory links are formed ; these loops contain nodes @xmath9 and @xmath10 .
the governing equations are as follows : @xmath11/\tau_j \\ & \hspace{0cm}+\mu_1 f(s_{k^\pm};\bar{s})+\mu_2\sum_{l^\pm\in \lambda^\pm } a_{l^\pm , k^\pm}f(s_{l^\pm};\bar{s } ) + w(t),\\ & \dot{v}^j_{k^\pm } = \varepsilon(u^j_{k^\pm}- q v^j_{k^\pm } + r)/\tau_j\\ & \hspace{0 cm } + \mu_3\sum_{m^{\pm}=1^\pm}^{k^\pm}\sum_{l^\pm\in\lambda^\pm}a_{m^\pm , l^\pm}a_{m^\pm , k^\pm } \\ & \hspace{2 cm } \times f(s_{m^{\pm},k^{\pm},l^\pm};\bar{s})f(s_{l^\pm}+\sigma s_{l^\mp};\gamma \bar{s})\\ & s_{k^\pm}=\frac{1}{j}\sum_{j=1}^{j}u^{j}_{k^\pm},\quad k^\pm=1^\pm,2^\pm,\dots , k^\pm,\ j=1,2,\dots , j , \end{aligned}\ ] ] where @xmath12 , @xmath13 is dimensionless time , a dot above a variable indicates the derivative of that variable with respect to @xmath13 , and @xmath14 ( @xmath15 ) and @xmath16 are positive constants .
the dynamics of each oscillator are described by the fitzhugh - nagumo equation . @xmath17 ( or @xmath18 ) is the total number of vertices . from ( s4 ) , each node consists of @xmath19 oscillators and the variables @xmath20 and @xmath21 correspond to the activator and inhibitor , respectively , of the @xmath22th oscillator in node @xmath0 .
the state of each node @xmath23 is determined by @xmath24 , the average of @xmath25 ( @xmath26 ) in node @xmath23 .
the strength of the self - feedback @xmath27 controls the degree of phase synchronization of the node , where large and small values of @xmath27 induce synchronization or incoherent oscillation , respectively . parameters @xmath28 and @xmath29 correspond to the strengths of excitatory and inhibitory connections , respectively [ fig.[fig : bilayer](b ) ] .
we assume that the interaction function @xmath30 has a threshold for activation ; thus , the state of the connection switches dynamically between the on and off states depending on the state of the node . the regulation of the on - off switching of the connecting nodes depends on @xmath31 and is defined by the heaviside function with a threshold @xmath32 , where @xmath33 for @xmath34 and @xmath35 for @xmath36 .
the interactions affect all elements uniformly ; that is , they are independent of @xmath22 .
the implementation of the heaviside function is needed to realize condition c3 , given later .
excitatory interaction directed from node @xmath37 to node @xmath0 is expressed as @xmath38 , where @xmath39 and @xmath40 indicate the presence and absence of such interactions , respectively . since the start and goal nodes
have a connecting path from the n - layer to the p - layer and a path from the p - layer to the n - layer from ( s5 ) , @xmath41 .
it is noted that @xmath42 except for @xmath43 and @xmath44 and that @xmath45 for all @xmath23 . from ( s2 ) , the system has inhibitory interactions at every branching point of the network ( see condition c2 ) .
the activation of the inhibitory interaction between nodes @xmath0 and @xmath37 is also regulated by @xmath46 [ fig.[fig : bilayer](b ) ] .
we use the notation @xmath47 to indicate that node @xmath48 regulates activation of inhibitory interactions between nodes @xmath0 and @xmath37 .
assumptions ( s7 ) and ( s7 ) correspond to the inhibitory functions of the presented system and the previous system , respectively .
* for the presented system , we take @xmath49 . except for the links at the start and goal nodes ,
the dynamics of the p - layer and the n - layer are coupled through this inhibitory interaction . from ( s7 ) , the inhibitory interactions are activated when @xmath50 is larger than @xmath51 and @xmath52 is larger than @xmath32 . to compare the finding time of the presented system with that of the previous system
, we use the previous system in section [ sec : results ] . since the previous system is fundamentally the same as ( [ eqn : model ] ) with @xmath53 and @xmath54 ,
we assume ( s7 ) when we examine numerical experiments of the previous system .
* for the previous system , we take @xmath53 and @xmath54 .
we assume ( s7 ) unless otherwise mentioned . for simplicity ,
we assume that the parameters @xmath55 , @xmath56 , @xmath57 , and @xmath58 are independent of @xmath23 and we set @xmath59 , @xmath60 , @xmath61 , and @xmath62 ; @xmath63 is a small amount of random noise in the interval @xmath64 $ ] ; the time constants @xmath65 take random values from the interval between @xmath66 and @xmath67 , where the values of @xmath66 and @xmath67 are set to @xmath68 .
the distribution of @xmath65 is the same for all nodes ; that is , it is independent of @xmath23 .
the regulatory rules proposed in @xcite are employed for the presented model .
each node can have three qualitatively different states : synchronized oscillation with a large amplitude , synchronized oscillation with a small amplitude , and incoherent oscillation .
we refer to these states as large - amplitude synchronized oscillation ( lso ) , small - amplitude synchronized oscillation ( sso ) and incoherent oscillation ( inc ) , respectively .
the state lso means that @xmath31 exceeds @xmath32 for every oscillation .
the use of the fitzhugh - nagumo equation admits the presence of sso when inhibitory interactions are exerted . by regarding lso as the on state and both sso and inc as the off state
, the proposed system can be regarded as a boolean network .
the parameter values @xmath14 are set so that the regulatory rules and the conditions ( c1,c2 , and c3 in fig .
[ fig : condition ] ) are satisfied .
the parameter values are set to @xmath69 and @xmath70 . as discussed later
, the state of a node shows inc when no interactions from other nodes are exerted .
the value of @xmath16 is taken from an appropriate regime satisfying the conditions .
here we discuss the regulatory rules for nodes in the p - layer .
the same argument can be applied , _ mutatis mutandis _
, to the regulatory rules for the nodes in the n - layer .
since the interactions exert uniformly on the oscillators belonging to the same node , phase resetting of oscillators is observed due to the excitatory interactions . for the given parameter values ,
the lso state is propagated along the nodes connected by excitatory links . as shown in fig.[fig : condition ] , the state of node @xmath1 becomes lso when that of node @xmath3 is lso [ fig.[fig : condition ] ( c1 ) ] .
the node @xmath1 becomes inc when the node @xmath3 is in one of the off states ( sso or inc ) since no interactions are exerted on node @xmath1 . at the branching point of excitatory links
, we can see the competition between the excitatory and inhibitory interactions . for the inhibitory interactions ,
the regulatory rules are determined by the states of the node in both layers .
the inhibitory interactions from node @xmath71 to node @xmath72 become active when both node @xmath1 and node @xmath2 are in the on state , or @xmath73 .
thus , the inhibitory interactions from node @xmath71 to node @xmath72 become inactive when either the state of node @xmath1 or that of node @xmath2 is one of the off states ( sso or inc ) .
when the state of node @xmath74 is lso , the excitatory interactions are exerted on node @xmath1 and node @xmath3 from nodes @xmath74 . in the case that the states of node @xmath4 and @xmath3 are lso as shown in fig.[fig : condition ] ( c2)(i ) , the states of node @xmath1 and @xmath2 become sso due to the inhibitory interactions .
the role of @xmath1 and @xmath3 is interchanged when the states of node @xmath1 and @xmath2 are lso . in the case that node @xmath4 and @xmath2
are both in one of the off states , as shown in fig.[fig : condition ] ( c2)(ii ) and ( iii ) , the state of nodes @xmath1 and @xmath3 becomes lso .
when the state of node @xmath74 is one of the off states , since @xmath75 and the inhibitory interactions are inactive , the nodes @xmath1 and @xmath3 become inc , independent of the state of the nodes in the n - layer [ fig.[fig : condition ] ( c2 ) ( iv ) and ( v ) ] .
parameter @xmath27 is set so that the single node ( @xmath76 ) shows transient behavior sso @xmath77 lso @xmath77 inc when sso is taken as an initial data point ( c3 ) .
figure [ fig : bif ] shows the bifurcation diagram of the single node ( node @xmath78 ) .
it is observed that such a state transition from lso to inc occurs at a limit point , @xmath79 . by moving @xmath27 off the limit point of lso , transient behavior from lso to inc
is observed , and the transition time increases as the distance between @xmath27 and the limit point becomes smaller .
note that the transient time of lso becomes longer as the distance from the limiting point becomes smaller .
the threshold @xmath32 is taken to be sufficiently large to ensure that the values of @xmath24 are smaller than @xmath32 , and the state of the node converges to inc when no input is received ( see the horizontal dotted line in fig.[fig : bif ] ) . in section [ sec : results ] , we can observe that a loop chain of lso states can be stable due to the regulatory rule ( c1 ) .
the self - recovery process is established owing to the regulatory rule ( c2 ) with the transient behavior ( c3 ) . and
@xmath2 are inhibited when nodes @xmath3 ,
@xmath4 , and @xmath74 are in an lso state .
( ii)(iii ) the state of node @xmath1 and @xmath3 becomes lso when the state of node @xmath74 is the on state and nodes @xmath2 and @xmath4 are in off states ( sso or inc ) .
( iv)(v ) the states of nodes @xmath1 and @xmath3 are inc when node @xmath74 is in the off state , independent of the state of nodes in the n - layer .
for all patterns , the roles of @xmath1 and @xmath3 can be interchanged .
the possible patterns for the n - layer are given similarly .
( c3 ) transient behavior when a node receives no input .
the solution converges to an inc for any initial condition .
node starting with an sso approaches a quasi - stable lso during a finite time interval and finally converges to inc.,scaledwidth=50.0% ] ) .
the vertical axis indicates the maximum value of @xmath80 .
the horizontal dotted line indicates the threshold @xmath32 .
the limit point is @xmath79 ( vertical dashed line).,scaledwidth=40.0% ]
in this section , we confirm that the presented system shows the properties ( i)-(iii ) listed in section [ sec : intro ] .
a path is defined as a chain of lso states belonging to the corresponding nodes in both the p - layer and the n - layer .
in addition , we demonstrate that the presented system tends to choose shorter paths when multiple paths exist .
we test the presented system in the network shown in fig.[fig : perturb ] . from ( s2 ) , bidirectional inhibitory links are placed between nodes @xmath81 , @xmath82 , @xmath83 , and @xmath84 .
initially , we take the start and goal vertices as vertices 1 and 6 , respectively ; that is , @xmath85 and @xmath86 [ fig.[fig : perturb](a ) ] . at @xmath87 ,
the network has three possible paths connecting the start and goal vertices : path @xmath88 , path @xmath89 , and path @xmath90 . in numerical results shown in [ fig.[fig : perturb](a ) ] , path @xmath90 is selected from the possible paths . to confirm the self - recovery property , edge 8@xmath779 of the existing edge
is removed at @xmath91 [ fig.[fig : perturb](b ) ] ; the parameter values of @xmath92 and @xmath93 are changed from @xmath94 to @xmath95 at @xmath91 .
then , it is observed that the system can find a new path @xmath96 .
the system can find a new path when the positions of the start and goal vertices are changed .
for example , in fig.[fig : perturb](c)(d ) , it is observed that the system finds a path when the position of the goal vertex is changed from 6 to 11 at @xmath97 ( @xmath98 ) and from @xmath99 to @xmath100 at @xmath101 ( @xmath102 ) . due to the addition of the excitatory links at @xmath101
, inhibitory links are added between @xmath103 and @xmath104 from ( s2 ) and ( s7 ) .
when the network has no connecting paths , all of the node states become inc from the regulatory rules . in order to estimate the increase in the rate of finding as the node number increases , we use a hierarchical tree network , shown in fig.[fig : tree ] : @xmath105 ( @xmath106 , @xmath107 ) , where @xmath108 is the depth of the hierarchy .
the start position is fixed at vertex 1 , and the goal vertex is changed between the leftmost vertex ( vertex @xmath109 ) and the rightmost vertex ( vertex @xmath110 ) at the bottom of the tree when the system has succeeded in finding a desired path .
for example , in the case of @xmath111 ( fig.[fig : tree ] ) , the path contains vertices 1 , 2 , and 4 when the goal vertex is 4 and vertices 1 , 3 , and 7 when the goal vertex is 7 . the average finding time is calculated across 50 trials
now we compare the finding time for the presented system [ ( s7 ) ] with @xmath112 and the previous system [ ( s7 ) ] and @xmath108 are taken as control parameters . in the case of the previous system ,
the finding time increases substantially as @xmath108 increases [ fig.[fig : tree ] ( b ) ] .
in contrast , the finding time increases at an almost constant rate for the presented system .
next , we investigate the dependence of @xmath16 on the finding time .
it is observed that the finding time is improved and the distributions tend to be narrow when @xmath16 is increased for @xmath113 and @xmath114 .
figure [ fig : tree ] ( d ) shows the distributions of the finding time when @xmath16 is taken as a control parameter with @xmath115 .
we note that the distributions tend to be wide again for @xmath116 .
this implies that the large @xmath16 basically decreases the finding time ; however , a large @xmath16 also decreases the reliability when the vertex number increases .
the difference of the finding time between the two systems comes from the timing of the activation of the inhibitory interaction .
the inhibitory interactions play a role in path selection at the branching point of excitatory links . in the presented system ,
the activation of the inhibitory interaction at node @xmath23 is regulated by both @xmath117 and @xmath118 .
this means that the inhibitory interaction for a large @xmath16 becomes active ( i.e. , path selection initiates ) after an lso wave propagates to , at least , either the start node or the goal node . in the example network in fig.[fig : hikaku ] ( a ) , the path selection initiates after the system succeeds in finding a desired path .
this prevents the system from selecting wrong paths , and thus results in a decrease in finding time .
in contrast , for the previous system , since the inhibitory interactions activate when either @xmath117 or @xmath118 is larger than @xmath32 , the selection of a path at branching points occurs before the system has found both the start and goal vertices , as shown in the example network in fig.[fig : hikaku ] ( b ) . and of the previous system ( dashed line ) .
( c ) magnified figure of ( b ) .
( d ) the distribution of finding time of the presented system with @xmath115 and @xmath119 . , scaledwidth=50.0% ] as discussed in section [ sec : time ] , for the presented system , the inhibitory interactions become active when an lso wave , which has already passed through either the start node or the goal node , first reaches the branching point .
therefore , it is expected that the system chooses a path for which the distance between the branching point and the start or goal node is the shortest ; this is , the node distance between the branching point and the start ( or goal ) node is smallest .
the ratio of the chosen path to the shortest path is calculated for a network shown in fig.[fig : shortest ] , where the distance of the path is defined by the number of nodes in the path .
the difference of the node number @xmath120 is taken as a control parameter .
as initial data , the state of the start node is set to lso and the other nodes are set to inc .
the number of trials is 100 with different random seeds .
it is confirmed that the presented system with @xmath112 and @xmath121 selects shorter paths more frequently as @xmath120 increases .
, respectively .
( b ) network structures .
we fix @xmath121 .
, scaledwidth=40.0% ]
in this study , by building on a previous system @xcite , we have presented a pathfinding system .
the finding process uses only local interaction between nodes ; that is , the presented system is in a class of distributed systems .
the system inherits the self - recovery property from the previous system .
in fact , we have demonstrated that the system can find a new path when perturbation of the existing paths occurs .
we further modify the inhibitory interactions to improve the finding time of the desired path .
it has been numerically confirmed that the finding time increases linearly as the depth of the hierarchy of the tree - like network increases .
the self - recovery process is explained by using a network shown in fig.[fig : recover ] .
for example , if the goal vertex is changed from @xmath122 to @xmath100 [ fig.[fig : recover](a ) ] , the state of nodes @xmath123 , @xmath124 , and @xmath125 successively becomes inc due to ( c1 ) , and the state of node @xmath126 is changed from sso to lso due to ( c3 ) [ fig.[fig : recover](b ) ] . due to ( c1 ) , the state of nodes @xmath127 , @xmath128 , @xmath125 , and @xmath129 successively becomes lso , and the desired path is found [ fig.[fig : recover](c)(d ) ] .
it should be noted that the lso state of node @xmath130 needs to be maintained during the process ( c ) to ( d ) .
this means that the transient time from lso to inc should be chosen to be large enough compared to the lso wave speed and network size by taking @xmath27 near the limit point in fig.[fig : bif ] .
if the graph contains a cyclic loop that does not contain the start vertex or the goal vertex , the system may fail to find the connecting path between the start and goal nodes , since an lso loop along the cyclic loop is stable , and the system may choose it .
thus , we assumed ( s6 ) .
this work was supported by a grant - in - aid for scientific research on innovative areas `` neural creativity for communication '' ( no .
4103 ) ( 21120003 and 21120005 ) of mext , japan and a grant - in - aid for scientific research ( c ) ( 25400199 ) of jsps , japan . | this study propose a continuous pathfinding system based on coupled oscillator systems .
we consider acyclic graphs whose vertices are connected by unidirectional edges .
the proposed model autonomously finds a path connecting two specified vertices , and the path is represented by phase - synchronized oscillatory solutions . to develop a system capable of self - recovery ,
that is , a system with the ability to find a path when one of the connections in the existing path is suddenly removed , we implemented three - state boolean - like regulatory rules for interaction functions . we also demonstrate that appropriate installation of inhibitory interaction improves the finding time . |
the kuiper belt is currently the observational frontier of our solar system .
presumably the best kept remnants of the icy planetesimals that formed the outer planets , kuiper belt objects ( kbos ) have been the subjects of intense study in the past @xmath315 years .
one intriguing kbo is 136108 haumea ( formerly 2003 el@xmath4 ) .
first famous for its super - fast rotation and elongated shape , haumea went on to surprise us with a host of interesting properties .
haumea s spin frequency of one rotation every @xmath5 hr is unparalleled for an object this large @xcite .
its shape is rotationally deformed into a @xmath0 km triaxial ellipsoid @xcite to balance gravitational and centripetal accelerations .
to attain such a fast rotation , haumea might have suffered a giant impact at the time when the kuiper belt was massive enough to render such events likely .
infrared spectroscopy has revealed a surface covered in almost pure h@xmath1o ice @xcite which gives haumea an optically blue colour @xcite .
the surfaces of the remaining pluto - sized kbos ( eris , pluto and makemake ) are covered in ch@xmath6 ice instead , granting them the tag ` methanoids ' .
two satellites were discovered in orbit around haumea @xcite , the largest of which is also coated in even purer h@xmath1o ice @xcite .
the two satellites have nearly coplanar orbits with fast - evolving , complex dynamics due mainly to tidal effects from the primary @xcite .
haumea s bulk density , derived assuming it is near hydrostatic equilibrium , is @xmath7 g @xmath2 @xcite .
the surface material has density @xmath8 in the same units implying that the interior must be differentiated and haumea must have more rock - rich core .
a number of kbos showing signs of h@xmath1o ice in their surface spectra all lie close to haumea in orbital space @xcite ; this , plus the unusually fast spin , the differentiated inner structure and the two small satellites also covered in h@xmath1o ice , all have been taken as evidence that haumea is the largest remnant of a massive collision that occured @xmath9 gyr ago @xcite . however , several potential members of the collisional family have been eliminated based on infrared photometry ( snodgrass et al .
, poster at this meeting ) . and
@xmath10 data were taken over 3 nights to ensure repeatability .
the effect of the dark red spot is apparent at rotational phases @xmath11 : the maximum and minimum that bracket that region appear darker and the @xmath12-band flux is consistently lower than the @xmath10-band flux indicating the spot is redder than elsewhere .
we measure a lightcurve period @xmath13 hours and a photometric range @xmath14 mag .
the rotationally averaged colour is @xmath15 mag .
best fit jacobi ellipsoid models are overplotted : a thick solid grey line shows how the uniform surface model fails to fit the dark red spot region , and thinner lines show that a small ( @xmath16 ) and very dark ( @xmath17 ) spot or a large ( @xmath18 ) and not very dark ( @xmath19 ) spot fit the data equally well . * b ) * cartoon representation of the three spot models considered in a ) showing the location of the spot on the surface of haumea . ]
we observed haumea in mid-2007 using the university of hawaii 2.2 m telescope with the goal of measuring its lightcurve in two bands , @xmath12 and @xmath10 ( fig . [ fig1]a ) .
our high - quality photometry @xcite shows two important features : 1 .
the lightcurve is not symmetric as would be expected from a uniform ellipsoidal body .
there is a clear asymmetry between the two sets of minima and maxima indicating the presence of a dark region on the surface ( fig .
[ fig1]a ) . a model lightcurve generated by placing a dark spot on the equator of haumea , visible at both minimum and maximum cross - section ( fig .
[ fig1]b ) , successfully fits the data .
2 . upon aligning the @xmath12 and @xmath10 lightcurve data
we verify that the @xmath12 points lie consistently below the @xmath10 points precisely at the location of the dark spot .
in other words , the dark spot is also redder than the average surface . in the rest of the paper we use drs to refer to the dark red spot . in our model ( fig .
[ fig1 ] ) the size and relative darkness of the drs are degenerate : the spot may be as small as a few percent of the projected cross - section of haumea and be about 20% as reflective as the rest of the surface , or it may be as large as to take a full hemisphere of haumea being then only 5% less reflective than elsewhere .
the same degeneracy applies to colour vs. spot size .
however , assuming the drs colour is within the range of values typically found in the solar system , @xmath20 , then when directly facing the observer the spot must take between 20% and 60% of the projected cross - section of haumea , and have an albedo between 55% and 65% .
this combination of colour and albedo is consistent with , e.g. eris , makemake and the bright regions on pluto and on saturn s satellite iapetus ; it is inconsistent with pluto s darker regions , with pluto s satellite charon , with saturn s irregular satellite phoebe and with centaurs chiron and pholus .
the orange dashed line shows a spectrum of crystalline ice ( indicated by the 1.65 @xmath21 m feature ) while the solid black line corresponds to ice with a lower degree of crystallinity .
* b ) * time - resolved 4-band spectrum of haumea [ adapted from @xcite ] . each line is a spectrum at a given rotational phase . at rotational phases when the drs faces the observer ( @xmath11 ) the @xmath12 band is depressed and the @xmath22 band
is enhanced .
spectra at each rotational phase are plotted relative to @xmath10 band and all rotational phases have been normalised by @xmath23 .
inset shows spectra before normalisation at @xmath23 . ]
prompted by the fact that haumea is covered in h@xmath1o ice , we set out to investigate how the properties of the ice changed close the drs region by monitoring the infrared h@xmath1o - ice absorption band at 1.5 @xmath21 m ( fig .
[ fig2 ] ) .
we collected two sets of data : time - resolved , quasi - simultaneous broadband @xmath24 ( to probe the continuum ) and medium band ch@xmath6s ( to probe the 1.5 @xmath21 m ) data using ukirt , and a similar dataset at the subaru telescope this time using broadband filters @xmath24 and @xmath22 .
both telescopes are located atop mauna kea .
neither of the filters ch@xmath6s or @xmath22 fully probe the 1.5 @xmath21 m band , and both are affected by a narrower band at 1.65 @xmath21 m which , if present , indicates that the ice is mostly crystalline .
our ukirt measurements indicate that the ch@xmath6s/@xmath24 flux ratio decreases by a few percent close to the drs , while our subaru measurement show that the @xmath22/@xmath24 flux ratio increases also by a few percent . this apparent contradiction can be explained if the drs is richer in crystalline h@xmath1o - ice than the rest of the surface .
that would change the shape of the 1.5 @xmath21 m and 1.65 @xmath21 m bands and cause exactly what is observed see fig .
@xcite report _
_ nicmos observations of haumea using the broadband filters f110w and f160w centered respectively at 1.1 @xmath21 m and 1.6 @xmath21 m . they find that the f160w / f110w flux ratio decreases at the drs , consistent with our ukirt observations . infrared spectra obtained at 4-m - class telescopes show no rotational variations , likely due to lack of sensitivity .
using the time - resolved @xmath25 data from the uh 2.2 m and @xmath26 data from subaru we constructed 4-band spectra of haumea as it rotates .
we did this by interpolating each of the four lightcurves and measuring the relative differences between the bands .
the result is shown in fig .
[ fig2]b where we plot 3 spectra away from the drs ( @xmath27 ) and 3 spectra close to the drs ( @xmath28 ) .
the figure shows that the drs material is a more efficient @xmath12 absorber than the rest of haumea , hence the redder @xmath12@xmath29@xmath10 colour , and that it is redder in visible - to - infrared colour . in
3 we saw that the drs displays bluer or redder behaviour in the @xmath26 wavelength range depending on exactly which filter bandpasses are used .
it is very unlikely that the dark spot is a topographical feature such as a mountain or valley as that would produce an achromatic change in brightness .
instead , the spot region exhibits slight but persistent visible and infrared colour properties that distinguish it from the rest of haumea s surface .
no atmosphere has been detected on haumea rendering an explanation based on irregular condensation of gases on the surface unlikely .
the fact that the drs absorbs @xmath12-band light more efficiently could indicate the presence of hydrated minerals @xcite .
alternatively the redder tint could be due to the presence of irradiated organic materials which would also explain the blue behaviour observed in some infrared bandpasses @xcite .
if confirmed , the higher degree of ice crystallinity at the drs could signal a recent temperature rise .
the drs could be a region where material from haumea s interior is trickling out .
the high bulk density of haumea indicates the presence of a more mineral - rich core .
if warmer , deep - lying material would find its way to the surface it would appear darker and presumably redder than h@xmath1o ice .
the slight increase in temperature and the presence of h@xmath1o are useful ingredients for mineral hydration and h@xmath1o - ice crystallinization .
the drs could also be the site of a recent impact of a @xmath30 km kbo onto haumea .
small kbos are dark and most are believed to be covered in red , irradiated organic mantles .
the collision would locally raise the temperature , thereby accelerating the transition from amorphous - to - crystalline h@xmath1o ice , and the impactor material would probably leave a visible trace on the surface
. the drs could also be due to something completely different .
time - resolved spectroscopy of haumea using @xmath31 m telescopes should help determine the composition of the drs and help solve the mystery of its origin .
i am grateful to david jewitt for comments on the manuscript and to the royal society for the support of a newton fellowship . | kuiper belt object 136108 haumea is one of the most fascinating bodies in our solar system .
approximately @xmath0 km in size , it is one of the largest kuiper belt objects ( kbos ) and an unusually elongated one for its size .
the shape of haumea is the result of rotational deformation due to its extremely short 3.9-hour rotation period . unlike other 1000 km - scale kbos which are coated in methane ice the surface of haumea
is covered in almost pure h@xmath1o - ice .
the bulk density of haumea , estimated around 2.6 g @xmath2 , suggests a more rocky interior composition , different from the h@xmath1o - ice surface .
recently , haumea has become the second kbo after pluto to show observable signs of surface features .
a region darker and redder than the average surface of haumea has been identified , the composition and origin of which remain unknown .
i discuss this recent finding and what it may tell us about haumea . |
for decades , galaxy merging has been a popular explanation for the observed evolution in galaxy properties .
galaxy mergers were first invoked to explain the morphological transformation of galaxies @xcite .
merging remains the backbone in cosmological simulations in building up large galaxies ( e.g. * ? ? ?
* ; * ? ? ?
gas - rich major mergers at high redshifts ( @xmath10 ) are thought to trigger starburst and active galactic nuclei ( agn ) episodes , quench star formation , and lead to bulge formation , thereby building the massive ellipticals in the local universe @xcite .
an alternative scenario has been proposed more recently , in which massive galaxies at high redshift are clumpy disks which are very efficient in turning incoming cold gas into stars @xcite .
the most luminous agns and ultra - luminous infrared galaxies ( ulirgs ) take place in major galaxy mergers @xcite .
merging galaxies have enhanced star formation activity compared to isolated ones ( @xcite ; but also see @xcite ) . as galaxy merging may have profound influence on how the galaxy population evolved to this day ,
quantifying its rate of occurrence is essential to judge whether it explains any of the observed evolutionary trends . as the timescale for galaxy mergers is on the order of a gyr ( e.g. * ? ? ?
* ) , the conventional way to measure galaxy merger rate is to divide the observed fraction of galaxies undergoing mergers by a typical merging ( observability ) timescale at different redshift bins .
merging galaxies can be identified as close galaxy pairs or galaxies displaying disturbed morphologies , and the timescale required to convert the merger fraction to merger rate depends on the specific selection technique . in this work
we use the pair selection method , as the merger fraction measured from morphological selection ( e.g. * ? ? ?
* ; * ? ? ?
* ) and the merging observability timescale are dependent on the imaging depth and resolution . the advent of multi - wavelength blank field observations in the past decade have enabled many improvements in the measurement of merger fractions , including the following : ( 1 ) the merger fraction of massive galaxies can be measured beyond @xmath12 ; ( 2 ) the photometric redshifts allow more accurate removal of the pairs projected along the line - of - sight ; ( 3 ) the stellar masses derived from the spectral energy distribution ( sed ) fitting provide the stellar mass ratio of galaxy pairs , which is a more physically meaningful proxy for the dynamical interaction than a single - band flux ratio ; ( 4 ) deeper and wider area surveys provide larger samples , which in turn allow the dependence of merger fractions on different parameters to be explored .
multiple authors have measured the merger fraction at @xmath13 , presenting somewhat conflicting results : does the merger fraction increase with redshift @xcite , remain constant , or even diminish @xcite ? as shown in @xcite , the variation of the parent galaxy selection and and mass ratio limits can contribute to some of the discrepancies across studies .
on the other hand , the average merging observability timescale is hard to estimate due to the large possible variety of orbital parameters and viewing angles , as well as the lack of observed dynamical information on a galaxy - to - galaxy basis .
the uncertainties in the implied merger rates are discussed thoroughly in @xcite . in this work , we present the largest sample of photometrically selected mergers at @xmath14=0.1 - 3 to date from stellar mass complete catalogs
. the _ ks_-band selected catalog from the ultravista / cosmos survey @xcite covers a large area , allowing us to expand our merger sample to more than five times times larger than previous studies .
we complement the ground - based ultravista catalog with the space - based 3dhst+candels @xcite catalog , which is deeper and has higher spatial resolution , to study possible systematic effects in measuring merger fractions .
the remainder of the paper is structured as follows : section [ sec : data ] describes the ultravista and the 3dhst+candels catalogs used in our study .
we present the criteria for selecting massive galaxies and mergers , as well as the completeness of the catalogs . in section [
sec : method ] we present the method of measuring the merger fractions as a function of redshift .
we compare the merger fractions measured using the two catalogs , as well as the selection using the stellar mass ratio and @xmath1-band flux ratio .
we examine the stellar mass ratio distribution of the selected mergers .
we discuss the two main sources of uncertainties in the merger fraction measurements .
we show that we are complete to detecting minor mergers up to @xmath15 .
finally we convert the merger fractions to merger rates , and infer the merger contribution in the stellar mass , size , velocity dispersion and number density evolution of massive galaxies . based on our findings ,
we address some broader questions in the context of galaxy evolution in section [ sec : discussion ] : what do the merger rates imply for the evolution of massive quiescent galaxies ? is merging an influential process in the cosmic star formation history or not ? we also discuss the future prospects of merger fraction studies .
the conclusions of this work is summarised in section [ sec : conclusions ] . in appendix
[ sec : missing_mergers ] we present the simulations we perform to test for the completeness limits of the faintest possible satellites . appendix [ sec : pf_lit ] provides an in - depth comparison to similar merger fraction measurements in the literature .
all magnitudes are quoted in the ab system .
a cosmology of @xmath16 = 70 km s@xmath17 mpc@xmath17 , @xmath18 = 0.3 and @xmath19 = 0.7 is adopted throughout this work .
we use the _ ks_-band selected catalog for the ultravista survey compiled by @xcite .
the ultravista survey targets the cosmos field @xcite with the eso vista survey telescope .
the effective survey area of ultravista is 1.62 deg@xmath20 .
the catalog contains psf - matched photometry in 30 photometric bands covering the wavelength range 0.15 - 24@xmath21 and includes the _ galex _
@xcite , cfht / subaru @xcite , ultravista @xcite , s - cosmos @xcite , and zcosmos @xcite datasets .
the ultravista source detection is performed on the _ ks_-band image with a @xmath22 aperture , which has a limiting magnitude of @xmath23 ( @xmath24-aperture ) . in total
there are 154 803 detected sources with reliable photometry having _
_ ks__@xmath25 , which is the 90% completeness limit and the adopted luminosity limit in this work .
the stellar masses quoted in this paper are derived assuming a chabrier imf .
further details regarding the photometric redshifts ( photo-@xmath14 s ) and sed fitting can be found in @xcite . to complement the
ground - based _
yjhk@xmath26 _ imaging from vista , we use the 3dhst catalog presented in @xcite and @xcite , which includes _ hst _ imaging from the candels survey @xcite over five fields : cosmos , goods - north and south , aegis , and uds with a combined usable area of @xmath27 deg@xmath20 .
@xcite performed photometry ( aperture of @xmath28 ) on the psf matched images and compiled a photometric catalog with photo-@xmath14 s and sed best fits .
we only use the objects marked with good photometry to ensure reliable photo-@xmath14 s and stellar masses .
we use close galaxy pairs as a probe for galaxy mergers following similar criteria used in the literature @xcite . in the ultravista catalog
, there are 9829 massive ( log@xmath4 ) galaxies in the redshift range of @xmath29 , and 380 ( @xmath30 ) of them are covered by the _
h_-band imaging from the candels and 3dhst cosmos surveys . around these massive galaxies ,
we search for galaxy satellites fulfiling the following criteria : 1 . within a projected separation of @xmath31 kpc @xmath32 .
stellar mass ratio @xmath33 of 1:1 - 4:1 as major merger , 4:1 - 10:1 as minor merger .
3 . the 1@xmath34 confidence intervals of the photo-@xmath14 s of the pair overlap .
we calculate @xmath35 using the angular scale based on the photo-@xmath14 s of the more massive galaxy . as
the fwhm of the ground - based ultravista @xmath36-band image is @xmath37 , corresponding to a maximum of 9.7 kpc @xmath32 at @xmath38 , we use 10 kpc @xmath32 as the lower limit of @xmath35 to ensure that no close pairs are missed due to blending . in section
[ sec : compare_merger_rates ] we explore the use of different @xmath35 bins up to 100 kpc @xmath32 .
we explore the use of the @xmath7-band flux ratio as a probe for the stellar mass ratio in section [ sec : pf_ratios ] , which we demonstrate to have a profound impact on the merger fraction evolution at @xmath39 .
the redshift distribution of massive galaxies and pairs are listed on table [ table : pf_massratio ] .
we assess the completeness limit of the massive galaxies and their 4:1 and 10:1 satellites in two aspects : the stellar mass completeness and the surface brightness limits .
we detail our analysis in appendix [ sec : missing_mergers ] and give the summary as follows .
we find that the surface brightness limit is the constraining factor for detecting the satellites of massive galaxies .
if completeness is only estimated by comparing the magnitude - redshift distribution to deeper catalogs , the completeness limits may be overstated .
we find that ultravista ( 3dhst+candels ) is complete to @xmath40 and @xmath41 ( @xmath42 and @xmath15 ) for major and minor mergers respectively . in this work ,
the data points at redshift bins which are mass incomplete are either omitted or plotted as lower limits , to ensure that incompleteness does not affect our conclusions .
despite the fact that 3dhst+candels is deeper than ultravista and can probe the merger fractions to higher redshifts , we demonstrate in section [ sec : pf_uvista ] that we do not get a higher merger fraction , both major and minor , with 3dhst+candels compared to ultravista , suggesting that there is not a significant population of mergers that have faint quiescent satellites only detectable in the 3dhst+candels catalog .
[ fig : pf_massratio ] [ fig : pf_fluxratio ]
the relation between the number of observed galaxy pairs ( @xmath43 ) and the number of ongoing physical galaxy mergers ( @xmath44 ) can be described as @xmath45 .
the quantity @xmath43 is defined as the number of galaxy pairs observed that satisfy a projected separation and mass ( or flux ) ratio criteria , e.g. pairs fulfiling the first two criteria listed in [ sec : select_pairs ] . among the observed pairs , some are galaxy pairs of physical proximity , while some pairs are galaxies projected along a similar line - of - sight .
the line - of - sight projected galaxy pairs can be corrected for using redshift measurements ( photometric or spectroscopic ) or statistical arguments based on the galaxy mass or luminosity function . in this work
we apply a photo-@xmath14 criterion as listed in section [ sec : select_pairs ] to correct for projected pairs ( @xmath46 ) .
we have demonstrated in @xcite that using the photo-@xmath14 s to correct for chance alignments yield results consistent with statistical corrections at @xmath47 . in this work
we do not correct for physical galaxy pairs at matching redshifts that are not energetically bound to merge , i.e. we assume @xmath48 = 0 .
cosmological simulations can provide a statistical estimate of @xmath48 to account for the unbound galaxy pairs in cluster environments with high relative velocities .
however , the interpretation may be complicated by the presence of a third neighbor which is not uncommon @xcite , or these pairs simply require more time before the eventual coalescence @xcite . galaxy fly - bys may be frequent @xcite but it remains unexplored how high - speed encounters may impact the mass distribution and light profiles of galaxies .
even if the cores do not coalesce , mass from the satellite may still be deposited onto the host galaxy , and the energy exchange can lead to size growth akin to a `` real '' merger @xcite .
it is not well understood how @xmath48 evolves with the environment and redshift . at higher redshift
, massive galaxies are expected to be less clustered than at the present day , so the effect is likely more dominant at low redshift .
future studies of the dynamical properties of galaxy pairs at different redshifts and environments may provide new insights into this effect , but for now we do not have enough information to correct for it .
we note that by including non - energetically bound pairs in our selection , the merger fractions derived in this paper are formally upper limits .
hereafter we refer to @xmath44 as @xmath49 for simplicity .
we define the merger fraction as the fraction of massive galaxies that are merging with a less massive companion , i.e. @xmath50 . the major and minor merger fractions ( @xmath51 and @xmath52 ) in redshift bins are listed on table [ table : pf_massratio ] and plotted on figure [ fig : pf_massratio ] ( left ) .
we parameterise the merger fractions within the completeness limits by a power law using least squares fitting . in the case of @xmath51 declining beyond @xmath38 in ultravista , the reduced @xmath53 value for the power law fit exceeds 10 indicating a bad fit so we fit the data points with a quadratic function instead .
we list the best fitting parameters in table [ table : bestfit ] . using the stellar mass ratio selection
, we find that @xmath51 ( @xmath52 ) increases from @xmath54 to reach a peak at @xmath55 , remains relatively constant to @xmath56 ( @xmath57 ) and then diminishes towards higher redshift .
a comparison between the merger fractions derived from the ground - based ultravista and the deeper , higher resolution 3dhst+candels reveals very similar @xmath51 and @xmath52 in both samples .
in fact , @xmath51 is slightly lower in 3dhst+candels than in ultravista at @xmath58 .
if we include the pairs without photo-@xmath14 information ( columns 3 and 7 on table [ table : pf_massratio ] ) in our merger sample , the @xmath51 of 3dhst+candels at this redshift bin becomes consistent with the one from ultravista .
this illustrates that space - based data is not required for measuring the galaxy merger fraction .
in fact , ground - based data with a large survey volume such as ultravista provides the optimal dataset , as the sample is adequately large to measure the redshift dependence of the merger fractions in finer redshift bins .
we elaborate on the uncertainties of merger fraction measurements in section [ sec : cv ] .
cccc + ultravista & stellar mass ratio & @xmath59 & @xmath60 + ultravista & @xmath1 flux ratio & @xmath61 & @xmath62 + 3dhst & stellar mass ratio & @xmath63 & @xmath64 + 3dhst & @xmath1 flux ratio & @xmath65 & @xmath66 + + ultravista & stellar mass ratio & @xmath67 & @xmath68 + ultravista & @xmath1 flux ratio & @xmath69 & @xmath70 + 3dhst & stellar mass ratio & @xmath71 & @xmath72 + 3dhst & @xmath1 flux ratio & @xmath73 & @xmath74 + [ table : bestfit ] ccccccccccc + @xmath75 & 628 & 23 & 12 & 201 & 3.66@xmath76 & & 29 & 6 & 170 & 4.62@xmath77 + @xmath78 & 772 & 40 & 4 & 99 & 5.18@xmath79 & & 52 & 3 & 117 & 6.74@xmath80 + @xmath81 & 1618 & 158 & 2 & 170 & 9.77@xmath82 & & 146 & 6 & 179 & 9.02@xmath83 + @xmath84 & 1692 & 184 & 6 & 169 & 10.87@xmath85 & & 140 & 9 & 140 & 8.27@xmath86 + @xmath87 & 1426 & 142 & 5 & 133 & 9.96@xmath88 & & 134 & 10 & 143 & 9.4@xmath89 + @xmath90 & 1163 & 133 & 8 & 99 & 11.44@xmath91 & & 81 & 13 & 102 & 6.96@xmath92 + @xmath93 & 1087 & 99 & 9 & 97 & 9.11@xmath94 & & 42 & 20 & 125 & 3.86@xmath95 + @xmath96 & 560 & 40 & 9 & 63 & 7.14@xmath97 & & 15 & 4 & 63 & 2.68@xmath98 + @xmath99 & 536 & 28 & 13 & 56 & 5.22@xmath100 & & 18 & 9 & 73 & 3.36@xmath101 + @xmath102 & 347 & 20 & 3 & 31 & 5.76@xmath103 & & 10 & 4 & 46 & 2.88@xmath104 + + @xmath105 & 123 & 10 & 1 & 43 & 8.13@xmath106 & & 7 & 1 & 44 & 5.69@xmath107 + @xmath108 & 61 & 3 & 0 & 9 & 4.92@xmath109 & & 5 & 0 & 14 & 8.2@xmath110 + @xmath111 & 88 & 5 & 0 & 14 & 5.68@xmath112 & & 9 & 1 & 16 & 10.23@xmath113 + @xmath114 & 63 & 6 & 0 & 20 & 9.52@xmath115 & & 5 & 1 & 18 & 7.94@xmath116 + @xmath117 & 45 & 0 & 0 & 7 & 0.0@xmath118 & & 3 & 1 & 7 & 6.67@xmath119 + + @xmath105 & 84 & 3 & 2 & 22 & 3.57@xmath120 & & 6 & 5 & 35 & 7.14@xmath121 + @xmath108 & 72 & 6 & 2 & 13 & 8.33@xmath122 & & 7 & 2 & 13 & 9.72@xmath123 + @xmath111 & 57 & 1 & 2 & 7 & 1.75@xmath124 & & 3 & 2 & 4 & 5.26@xmath125 + @xmath114 & 65 & 1 & 0 & 6 & 1.54@xmath126 & & 1 & 3 & 13 & 1.54@xmath127 + @xmath117 & 37 & 0 & 0 & 6 & 0.0@xmath118 & & 1 & 0 & 9 & 2.7@xmath128 + + @xmath105 & 66 & 4 & 0 & 11 & 6.06@xmath129 & & 6 & 2 & 25 & 9.09@xmath130 + @xmath108 & 77 & 6 & 1 & 9 & 7.79@xmath131 & & 4 & 1 & 11 & 5.19@xmath132 + @xmath111 & 74 & 3 & 0 & 9 & 4.05@xmath133 & & 6 & 5 & 12 & 8.11@xmath134 + @xmath114 & 47 & 2 & 0 & 7 & 4.26@xmath135 & & 2 & 0 & 8 & 4.26@xmath136 + @xmath117 & 39 & 4 & 0 & 6 & 10.26@xmath137 & & 2 & 1 & 7 & 5.13@xmath138 + + @xmath105 & 102 & 8 & 3 & 54 & 7.84@xmath139 & & 3 & 5 & 76 & 2.94@xmath140 + @xmath108 & 124 & 7 & 4 & 23 & 5.65@xmath141 & & 11 & 10 & 38 & 8.87@xmath142 + @xmath111 & 141 & 19 & 1 & 31 & 13.48@xmath143 & & 9 & 5 & 22 & 6.38@xmath144 + @xmath114 & 86 & 5 & 3 & 16 & 5.81@xmath145 & & 3 & 1 & 21 & 3.49@xmath146 + @xmath117 & 54 & 4 & 1 & 8 & 7.41@xmath147 & & 1 & 3 & 14 & 1.85@xmath148 + + @xmath105 & 87 & 11 & 6 & 29 & 12.64@xmath149 & & 9 & 6 & 19 & 10.34@xmath150 + @xmath108 & 103 & 5 & 2 & 13 & 4.85@xmath151 & & 8 & 4 & 17 & 7.77@xmath152 + @xmath111 & 162 & 11 & 2 & 15 & 6.79@xmath153 & & 11 & 2 & 38 & 6.79@xmath154 + @xmath114 & 98 & 3 & 0 & 12 & 3.06@xmath155 & & 2 & 2 & 10 & 2.04@xmath156 + @xmath117 & 65 & 4 & 0 & 11 & 6.15@xmath157 & & 0 & 2 & 16 & 0.0@xmath158 + [ table : pf_massratio ] [ fig : mtol_uvista ] [ fig : mtol_3dhst ] selected by stellar mass ratio ( blue ) or @xmath1 flux ratio ( red ) from the 3dhst+candels catalog .
the ratios are defined such that the mass- ( flux-)ratio selected mergers will have ratios of 1 - 10 . on the top panel
we show the histogram of the stellar mass ratios , and at bottom the histogram of the @xmath1 flux ratios .
the solid , dashed and dotted gray lines represent the 1:1 , 1:4 & 4:1 , 1:10 & 10:1 ratios respectively . from the top panel
, we observe that a large excess of flux ratio selected mergers in 3dhst+candels have @xmath1 flux ratios between 1 and 10 , but have stellar mass ratios between 10 and 100 .
this explains the rising merger fractions observed in figure [ fig : pf_fluxratio ] ( right ) due to bright satellites with log@xmath159 being included in the flux ratio selected sample .
[ fig : mr_fr_histogram ] , scaledwidth=50.0% ] merger fraction measurements have led to conflicting conclusions regarding whether it increases with redshift at @xmath39 @xcite or not @xcite .
the former studies use the single band flux ratio from _ hst _ _ h_-band imaging to estimate the mass ratio , rather than full the stellar mass ratio from sed fits used in the latter studies .
we explore the possibility of a systematic effect regarding the ratio used in the merger selection .
we repeat the selection of mergers with the @xmath1-band flux ratio instead of using the stellar mass ratio on the same dataset presented in section [ sec : data ] , namely the ultravista and 3dhst+candels catalogs .
the results are presented in figure [ fig : pf_fluxratio ] ( right ) .
it is apparent that the combination of using the flux ratio to select mergers and the 3dhst+candels catalog leads to an increasing redshift trend of @xmath51 ( @xmath52 ) up to @xmath160 ( @xmath15 ) where the catalog is complete for major ( minor ) satellites .
this is in contrast to the flat or even diminishing evolution found when mergers are selected by the stellar mass ratio ( figure [ fig : pf_massratio ] , left ) , as well as using flux ratio to select mergers from ultravista ( filled circles in figure [ fig : pf_fluxratio ] , right ) .
our results are in good agreement with the trends found in literature ( see appendix [ sec : pf_lit ] for details of the comparison ) meaning that we are able to reproduce the increasing redshift trend of the merger fraction if mergers are selected by flux ratio . by comparing the mergers selected in the overlapping area of the ultravista and candels - cosmos surveys
, we find that the flux - ratio selected satellites at @xmath10 are close to the survey depth of ultravista dr1 ( @xmath161 , @xcite ) , and therefore fainter satellites are missed due to low surface brightness .
we interpret the difference between the flux ratio selected merger fraction between the ultravista and the 3dhst - candels samples as being due to the observation limit of the ultravista dr1 data .
this is expected to improve for the forthcoming data release of ultravista in which the survey depth of the four ultra - deep stripes will be @xmath162 mag deeper . in order to explain the difference between the flux and stellar mass ratio selections using the 3dhst+candels catalog , we compare the stellar mass ratio and flux ratio distribution of the mergers using both selection techniques in figure [ fig : mr_fr_histogram ] .
we display the results for the redshift bin @xmath163 where the discrepancy in the merger fraction is most significant between the two selection techniques .
we find that almost all of the stellar mass ratio ( 1:1 - 10:1 ) selected mergers have @xmath7-band flux ratio in the same range . on the other hand ,
flux ratio selected mergers ( 1:1 - 10:1 ) include mergers with stellar mass ratios in the same range , as well as mergers with more extreme stellar mass ratios ( @xmath16410:1 ) . among the major _ flux ratio _ pairs at @xmath163 in 3dhst+candels , only @xmath165 have major _ stellar mass _ ratios .
the remaining pairs consist of minor _ stellar mass _ ratio ( @xmath166 ) and mostly very minor _ stellar mass _ ratio ( @xmath167 ) with @xmath16810:1 .
this demonstrates that the observed @xmath7-band flux is a biased tracer of the stellar mass at @xmath10 . using the @xmath7-band flux ratio as a probe for the stellar mass ratio leads to the inclusion of bluer , less massive galaxies as satellites . in another words , at @xmath10 most of the satellites are star - forming blue galaxies that are bright in the rest - frame optical b- or v - bands .
we conclude that the flux ratio selection yields a higher merger fraction than mass ratio selection at all redshifts for two reasons : ( 1 ) the observed @xmath7-band probes bluer rest - frame bands at higher @xmath14 ; ( 2 ) lower @xmath169 satellites enter the sample @xcite , where @xmath169 is the ratio of the stellar mass to the rest - frame @xmath170-band luminosity .
we illustrate the redshift dependence of @xmath169 in figure [ fig : mtol_z ] .
there is overall @xmath169 redshift evolution in both the massive galaxies and their satellites , in which the ratio increases over cosmic time .
both catalogs show a similar @xmath169 evolution except for the @xmath7-band flux ratio selected pairs in the candels+3dhst sample , where the evolution is steeper implying the inclusion of lower @xmath169 at @xmath10 than for the stellar mass ratio selection .
at @xmath171 the observed @xmath1-band roughly corresponds to the rest - frame @xmath172 and @xmath170 bands .
our simulations in appendix [ sec : sb_limit ] indicate that we are complete to @xmath173 for major ( minor ) mergers in 3dhst+candels , therefore the @xmath169 evolution can not be explained by observational effects and is intrinsic .
the @xmath169 evolution reflects the higher star formation activity at @xmath174 compared to that of the present day ( e.g. * ? ? ?
having shown that the use of the flux and stellar mass ratio can reproduce the discrepancy in merger fraction in literature , we proceed to find the ratio that best describes the dynamics and future evolution of the merging galaxies .
although using the @xmath7-band flux ratio selection is biased towards star - forming but low stellar mass satellites , the use of the stellar mass ratio may be biased against gas - rich satellites at @xmath13 .
galaxies appear to be more gas - rich at higher redshift and at lower masses @xcite .
such a dependence implies that the baryon mass ratio is closer to unity than the stellar mass ratio , since cold gas mass is included into the baryon mass calculation .
the baryon mass of a galaxy is a better probe of its total mass ( which also includes dark matter ) than the stellar mass alone , as shown in cosmological simulations @xcite .
a merger can be _ major _ or _ minor _ depending on whether the stellar mass , baryon mass or total mass is considered for the mass ratio @xcite .
intermediate mass galaxies of log@xmath175 are the satellites to the massive galaxies studied here , and their molecular gas mass may not be negligible in the total mass budget that governs the dynamics of the galaxies , especially at @xmath174 . if the cold gas fraction increases with redshift and decreases with stellar mass as previously claimed @xcite , there is a redshift - dependent underestimation if we use the stellar mass to trace the baryon mass .
the correction is likely larger at higher redshift due to the higher gas fraction . therefore merging with these gas - rich satellites with stellar mass ratios more extreme than 10:1
may contribute to the star formation budget of the massive galaxies @xcite , in the form of gas accretion or very minor mergers if characterised by the stellar mass ratio .
we note that gas - rich satellites are not equivalent to gas - rich mergers ( e.g. * ? ? ?
* ) , which is usually defined as the average gas fraction of both galaxies . despite the importance of the gas content in the merger definition as well as its contribution to star formation activity , direct measurements of the molecular gas mass are so far only available for limited samples of galaxies @xcite , mostly starbursting sub - millimeter galaxies .
alma surveys of large samples of normal " star - forming galaxies will shed light on this topic in the future @xcite .
ccccccccc @xmath176 & 7.8 & 0.18 & 0.39 & 0.43 & 6.7 & 0.20 & 0.39 & 0.44 + @xmath177 & 6.2 & 0.21 & 0.16 & 0.27 & 8.0 & 0.18 & 0.11 & 0.21 + @xmath178 & 7.5 & 0.17 & 0.59 & 0.61 & 7.3 & 0.18 & 0.19 & 0.26 + @xmath179 & 4.7 & 0.27 & 0.58 & 0.64 & 3.6 & 0.32 & 0.63 & 0.70 + @xmath180 & 5.0 & 0.33 & 0.86 & 0.92 & 2.9 & 0.45 & 0.80 & 0.92 + [ table:3dhst_cv ] it is apparent from figure [ fig : pf_3dhst ] that a considerable scatter exists for the merger fractions measured in the individual fields of the 3dhst+candels . the small survey area ( @xmath20.05deg@xmath20 for each of the five fields )
could lead to systematic uncertainties comparable to or larger than the poisson uncertainties .
we list the fractional errors ( @xmath181 ) of the merger fraction measurements of the candels+3dhst sample in table [ table:3dhst_cv ] .
the poisson uncertainties of the merger fractions are calculated as @xmath182 .
we compute the standard deviation of the merger fraction in each field from the combined mean as @xmath183 , where @xmath184 represents the measurement of each of the five fields .
the cosmic variance is simply the observed variance in excess of the poisson random noise , given by @xmath185 .
the cosmic variance is a comparable or sometimes larger contributor to the total error budget of the merger fraction measurements than the poisson uncertainty , as visualised in figure [ fig : pf_3dhst ] .
more specifically , in the redshift range of @xmath186 the @xmath51 measured from aegis is @xmath187 , whereas the same quantity is measured to be @xmath188 in goods - n .
while each of these quantities are @xmath189 from the @xmath190 averaged over the five candels fields , if the individual measurements are taken at face value without including the cosmic variance in the error budget , the results can differ by a maximum of @xmath191 depending on the field used .
combining the measurements from the five candels fields is crucial to mitigate cosmic variance , also known as the field - to - field variance @xcite .
the cosmic variance affecting the merger fraction measurements depends primarily on the number densities of the massive galaxies and their satellites , as well as the cosmic volume probed , as shown by @xcite .
here we use their parametrisation to estimate the relative cosmic variance for the ultravista and 3dhst+candels samples .
if we assume that the number densities of the massive galaxies and their satellites are not different in ultravista than in the combined five fields of 3dhst+candels , the cosmic variance has a dependence on the comoving volume as @xmath192 . since the comoving volume is proportional to the survey area , and ultravista covers @xmath193 larger area than the fields of 3dhst+candels combined , we expect the @xmath194 of ultravista to be @xmath195 that of 3dhst+candels .
another prominent error of the merger fraction is the poisson number count of pairs . as @xmath196 is proportional to @xmath197 , and again assuming similar number densities of satellites in both fields , we expect @xmath198 and therefore the poisson errors should be @xmath199 smaller in ultravista than that in 3dhst+candels .
this implies that the total fractional error of merger fraction measured from ultravista to be @xmath200 that of 3dhst+candels .
to summarise , we caution against drawing conclusions from merger fraction measurements based on individual candels - sized fields .
the merger fraction measurements from the five 3dhst+candels fields combined are comparable to those from ultravista which covers @xmath201 larger area , albeit with larger poisson uncertainties and in coarser redshift bins .
we call for including cosmic variance as a systematic uncertainty for pencil beam surveys such as 3dhst+candels for merger fraction measurements @xcite .
minor dry mergers are often invoked as the primary driver of the observed size evolution of quiescent massive galaxies from @xmath174 to 0 .
predictions from numerical simulations and virial arguments @xcite suggest that they are more efficient than major dry mergers in puffing up the sizes of quiescent galaxies per unit mass added . from previous minor merger fraction measurements @xcite and this work ( see section [ sec : merger_rates ] ) it is inferred that massive galaxies undergo less than one minor merger since @xmath174 . however ,
if the sole explanation of the observed size evolution is minor merging , multiple minor mergers are required ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
here we investigate the possibilities of missing faint satellites to massive galaxies at @xmath39 . as discussed in section [ sec : pf_ratios ] , we find that neither the major nor minor merger fractions in the candels deep fields are higher than those in the candels wide fields , although measurements from individual fields are subject to high cosmic variance ( see section [ sec : cv ] ) .
additionally , the merger fractions from stellar mass ratio selected mergers of ultravista and 3dhst+candels are remarkably consistent ( figure [ fig : pf_massratio ] , left ) , even in the redshift bins where ultravista is incomplete for low surface brightness galaxies .
even though the candels _ h_-band imaging is @xmath1643 magnitudes deeper and has @xmath202 smaller psf compared to ultravista , ultravista has the advantage that it probes a redder band ( _ ks _ ) where high redshift galaxies are brighter . to make a robust claim that we do not miss minor mergers lying just below the surface brightness limits ( sb ) of our surveys , we refer to the simulation performed for the completeness limits as introduced in appendix [ sec : sb_limit ] . in short
, we confirm that we do not miss minor mergers up to @xmath15 in 3dhst+candels .
we arrive at this conclusion by making the most conservative assumption that the faintest possible satellite is a maximally old , dust - free galaxy of log@xmath203 for a range of light profiles .
the completeness limits hold except for the extreme cases not simulated : ( 1 ) they have very compact sizes ( @xmath204 kpc ) and sersic index @xmath205 so that they have insufficient contiguous pixels above the detection threshold ; ( 2 ) they have very large sizes ( @xmath206 kpc ) and low @xmath207 so they have low sb ; ( 3 ) their dust extinction causes them to be fainter than a dust - free maximally old galaxy .
these size limits are motivated by the scaling relations for quiescent or early - type galaxies @xcite and simulation assumptions regarding the size of the stellar halo @xcite . unless these intermediate mass galaxies have light profiles very different from the more massive galaxies at similar redshift and similar mass galaxies at lower redshifts , ( 1 ) and ( 2 ) are not likely explanations .
the rest - frame optical faintest galaxies at @xmath10 should be quiescent and therefore should be dust - free , therefore ( 3 ) is not a likely explanation either . from binary merger simulations @xcite , the observability timescales of major and minor mergers are very short at @xmath208 kpc @xmath32 ( @xmath209 gyr ) and therefore we do not expect many close pairs blended by the psf .
as long as the lower @xmath35 limit for the close pair search is set according to the seeing and sb limit of the data , the resolution is not expected to cause a bias in the merger fraction .
as lower mass galaxies are more abundant than massive galaxies , one may expect that minor mergers are more frequent than major mergers from a statistical argument .
minor mergers are expected to be visible as pairs for longer than major mergers , according to dynamical friction timescales arguments and binary simulations @xcite .
therefore one intuitively expects the minor merger fraction and rate to be higher than the major ones .
however , cosmological simulations indicate that the major and minor merger rates are comparable in the stellar mass range probed in this work @xcite due to the stellar mass dependence on the @xmath210 relation . with our large complete sample of mergers , we can study the relative fractions of mergers of different stellar mass ratios ( @xmath211 ) .
we present our merger fractions in various @xmath211 bins in figure [ fig : pf_mu ] .
the merger fraction decreases as the @xmath211 gets more extreme .
the minor ( @xmath212 ) merger fractions are comparable to the major merger ( @xmath213 ) fractions at all redshifts .
this is in qualitative agreement with previous observations @xcite . for our sample of stellar mass ratio selected mergers from both datasets ,
the geometric number - weighted mean stellar mass ratio is @xmath214 4:1 - 5:1 and the mass - weighted mean stellar mass ratio is @xmath215 3:1 - 4:1 .
this is in consistency with various model predictions @xcite except @xcite , who find @xmath215 5:1 but @xmath216 16:1 .
their simulation is able to resolve down to 100:1 mergers , whereas we impose a cut at 10:1 mergers .
we attribute the discrepancy to a higher minor merger rate of their simulated massive galaxies , as well as our imposed cutoff at @xmath211=10:1
. -band flux ratio ) , except for a declining tail towards lower stellar mass ratios for the @xmath7-band flux ratio selection as discussed in section [ sec : pf_ratios ] . on this plot
we display the stellar mass selected ratio mergers from 3dhst+candels for illustration .
only the data points in which they are complete in stellar mass and surface brightness are shown ( see table [ table : completeness_all ] ) .
the major merger fractions appear to be comparable to the minor merger fractions at all redshifts .
, scaledwidth=50.0% ] the goal of measuring the galaxy merger fraction is to determine the time integral of the merger rate , defined as the number of mergers ( @xmath217 ) that a massive galaxy experiences on average over a time span .
the merger rate can be compared to the observed evolution of the galaxy population , such as in numbers , mass , size , etc .
, so that we can infer if galaxy merging is likely a driver .
merger rates scale as the number of mergers ( @xmath218 ) occurred during the time span ( @xmath219 ) defined by the redshift bin , divided by the time span , ( rate @xmath220 ) .
we measure @xmath221 as the number of observed merging galaxies ( @xmath222 ) divided by the observability timescale of mergers ( @xmath223 ) , i.e. rate @xmath224 .
the two common definitions of merger rates can be generalized as follows ( * ? ? ?
* and references therein ) : \(1 ) the number of merger events per unit time per unit volume ( @xmath221 ) : @xmath225 = \frac{n_{merge , obs}(z ) / \tau_\mathrm{obs}}{v_\mathrm{comoving}(z ) } = \frac{n_\mathrm{merge}}{\tau_\mathrm{obs } } \ ] $ $ \ ] ] where @xmath226 refers to the number of major ( or minor ) satellites around massive galaxies in that redshift , @xmath227 is the average observable timescale for the mergers of the mass ratio range observed to be within @xmath35 , and @xmath228 is the comoving volume projected by the survey area within the concerned redshift interval .
\(2 ) the number of merger events per galaxy per unit time ( @xmath229 ) is defined as : @xmath230 = \frac{\gamma ( z)}{n_\mathrm{massive}(z ) } = \frac{n_\mathrm{merge } } { n_\mathrm{massive } \tau_\mathrm{obs } } = \frac{f_\mathrm{merge}}{\tau_\mathrm{obs } } \ ] $ $ \ ] ] where @xmath231 is the number density of massive galaxies per unit volume .
the number of mergers a massive galaxy undergoes on average ( @xmath232 ) is simply the time integral of the merger rate per galaxy : @xmath233 $ $ \ ] ] where @xmath234 is the hubble time , and @xmath235^{1/2}$ ] @xcite with the @xmath236 s denoting the density parameters .
merger rates can be inferred by observing the merger fraction as a function of redshift , and then a merging timescale is assumed to convert the fraction to a rate .
the assumed merging timescale either comes from binary merger simulations @xcite , cosmological simulations @xcite , or approximation using the dynamical friction timescale . here
we briefly discuss the various options and justify the merger timescales used in this work .
the dynamical friction timescale @xcite is a suitable approximation for dark matter halo mergers of large mass ratios ( i.e. minor mergers ) .
however , it remains uncertain whether it can describe mergers with baryons or major mergers in which violent relaxation is the dominant mechanism determining the duration of the merger .
the timescales from binary simulations and cosmological simulations are conceptually distinct . in binary merger simulations ( e.g. * ? ? ?
* ) , two galaxies are set on approaching orbits , and the observability timescale ( @xmath223 ) samples the distribution of pre - coalescence pairs as a function of @xmath35 .
the timescale @xmath223 is a well - defined quantity which is directly applicable to the merger fraction to rate conversion .
this direct simulation method provides an accurate and comprehensible description of merging for the assumed conditions of relative velocity , gas fraction , morphology , etc . on the other hand ,
merging timescale ( @xmath237 ) defined in cosmological simulations @xcite depends on how the start and end of merging are defined , for example whether the end is the final coalescence of the two galaxy cores or when most of the mass of the satellite galaxy is deposited onto the massive one .
another complication is that there are different treatments of mapping stellar masses to the dm halos in cosmological simulations ( e.g. * ? ? ?
* ; * ? ? ?
we note that merging timescales for major mergers derived using cosmological simulations are shown to be @xmath238gyr longer compared with simulations that include baryons @xcite .
most importantly , @xmath223 instead of @xmath237 should be used to convert the observed fractions into rates . therefore in this work we use the @xmath223 from @xcite .
the cosmological simulations are useful to weigh the timescales of mergers from binary simulations with different assumptions , such as gas fraction , orbital parameters , as discussed in details in @xcite . due to the systematic uncertainties in these assumptions , as well as random uncertainties due to viewing angles of pairs
projected in 2d , the merging ( observability ) timescale can only be determined at best to @xmath239 accuracy ( * ? ? ?
* and references therein ) .
the merger rates derived using equations [ eqt : mergerrate_vol ] and [ eqt : nmerger ] normalised to timescales of 1 gyr are shown in table [ table : merger_rates_normalized ] .
we plot the inferred merger rates on figure [ fig : merger_rates ] . as expected from the merger fractions , we find the merger rates from ultravista are consistent with those from 3dhst+candels within the completeness range , and that the flux ratio selection method gives an increasing trend while the stellar mass ratio selection method gives a flat or diminishing trend for the 3dhst+candels catalog .
we list the best fitting parameters for the observed merger rates to a power law in table [ table : bestfit ] for easy comparison to literature .
as the merger rate uncertainties are considerably larger than the measured merger fractions due to the @xmath239 uncertainty in @xmath223 , the redshift dependence is weaker and we therefore deem a quadratic fit which has one more degree of freedom than the power law unnecessary .
we show the integrated number of major and minor mergers in table [ table : nmerger ] for the two catalogs and selection methods .
we find that at @xmath10 the observed merger rates using the stellar mass ratio selection are lower than predicted from the semi - analytical models ( sams ) of @xcite as shown in figure [ fig : merger_rates ] , but are consistent with the gas - poor merger rate ( @xmath240 , where the gas fraction @xmath241 is defined as the ratio of the total gas mass to the total baryon mass of the merging galaxies ) .
in general the sams predict that the galaxy merger rates increase monotonically with redshift .
our measurements using the @xmath7-band flux ratio selection show an increasing trend similar to the gas - rich merger rate of @xcite ( @xmath242 ) , even though the @xmath7-band flux is not a direct tracer of cold gas mass or star formation rate .
this lends support to our claim in section [ sec : pf_ratios ] that using the stellar mass ratio as a probe for the baryon mass ratio may be subject to a bias against gas - rich mergers at @xmath10 , an epoch at which cold gas fraction is non - negligible especially for intermediate mass galaxies @xcite .
we also compare the merger rates inferred from the merger fractions of various @xmath35 bins in figure [ fig : merger_rates_difft ] .
we only show results for the stellar mass ratio selection , but the following conclusions also hold for the @xmath7-band flux ratio selection .
we find that the merger rates are consistent for different @xmath35 bins once the suitable observability timescales from @xcite are applied . on average , the merger rates derived from mergers with @xmath35 = 10 - 100
kpc @xmath32 kpc @xmath32 is still small compared to the typical photo-@xmath14 uncertainty .
the typical photo-@xmath14 error at @xmath243 is @xmath244 @xcite , corresponding to 84 mpc / h at @xmath41 .
therefore we do not expect the photo-@xmath14 uncertainty to constrain widely separated pairs . ]
are up to @xmath245 higher than for smaller @xmath35 bins , although still consistent within the large uncertainties due to the 50% uncertainty in the merger observability timescale .
this implies that there are more widely separated mergers ( @xmath246 kpc @xmath32 ) than expected from the timescales of binary merger simulations .
possible explanations could be : ( 1 ) the large scale environment of galaxies are probed at separations of @xmath247 kpc @xmath32 , therefore we may include galaxies in the same over - densities that are not bound to merge ; ( 2 ) the merging observability timescales for wide pairs may be systematically longer than the assumed tilted polar orbit for close pairs , e.g. relative velocities of merging pairs are higher than assumed in the binary simulations ( typically @xmath248 km s@xmath17 ) which may be true in over - densities , or if the merger orbit is more like a circular orbit the merging timescale can be up to @xmath249 longer @xcite .
we note that the discrepancy is larger at lower redshift , hinting that the effect could be related to large - scale structure formation .
cosmological simulations may provide estimates of these effects . although we do not use the timescale of @xcite for galaxy merger fraction measurements for the reasons explained in section [ sec : compare_timescales ] , for comparison we note that using it leads to lower merger rates than those derived using the shorter timescales of @xcite as expected from the inverse scaling between timescale and rate .
cccccc + @xmath75 & 0.111@xmath250 & 0.109@xmath251 & & 0.14@xmath252 & 0.138@xmath253 + @xmath78 & 0.076@xmath254 & 0.09@xmath255 & & 0.098@xmath255 & 0.117@xmath256 + @xmath81 & 0.182@xmath255 & 0.125@xmath257 & & 0.169@xmath258 & 0.116@xmath257 + @xmath84 & 0.131@xmath257 & 0.123@xmath259 & & 0.1@xmath260 & 0.094@xmath260 + @xmath87 & 0.084@xmath261 & 0.083@xmath261 & & 0.08@xmath261 & 0.079@xmath261 + @xmath90 & 0.072@xmath262 & 0.072@xmath262 & & 0.044@xmath263 & 0.044@xmath263 + @xmath93 & 0.051@xmath263 & 0.045@xmath263 & & 0.021@xmath264 & 0.019@xmath265 + @xmath96 & 0.02@xmath265 & 0.028@xmath263 & & 0.008@xmath266 & 0.01@xmath265 + @xmath99 & 0.014@xmath265 & 0.016@xmath265 & & 0.009@xmath266 & 0.011@xmath265 + @xmath102 & 0.01@xmath265 & 0.015@xmath264 & & 0.005@xmath266 & 0.007@xmath265 + + + @xmath105 & 0.115@xmath267 & 0.5@xmath268 & & 0.099@xmath269 & 0.43@xmath270 + @xmath108 & 0.066@xmath258 & 0.096@xmath271 & & 0.086@xmath272 & 0.124@xmath273 + @xmath111 & 0.08@xmath258 & 0.073@xmath254 & & 0.078@xmath258 & 0.071@xmath274 + @xmath114 & 0.033@xmath259 & 0.031@xmath260 & & 0.025@xmath260 & 0.024@xmath260 + @xmath117 & 0.024@xmath260 & 0.023@xmath260 & & 0.014@xmath262 & 0.013@xmath262 + [ table : merger_rates_normalized ] cccccc + 10 - 30 kpc @xmath32 & @xmath275 & @xmath276 & & @xmath277 & @xmath275 + 10 - 100 kpc @xmath32 & @xmath278 & @xmath279 & & @xmath280 & @xmath281 + + + 10 - 30 kpc @xmath32 & @xmath282 & @xmath283 & & @xmath284 & @xmath285 + 10 - 100 kpc @xmath32 & @xmath281 & @xmath286 & & @xmath287 & @xmath288 + [ table : nmerger ] we compute the merger - driven stellar mass accretion rate as @xmath289 [ @xmath290 / gyr / galaxy ] = @xmath291 , where @xmath292 is the median stellar mass of the massive galaxies , @xmath229 is the major ( minor ) merger rate , and @xmath293 is the median stellar mass ratio of the major ( minor ) mergers .
all these quantities are redshift dependent so we are able to calculate the merger - driven stellar mass growth as a function of time .
there is controversy regarding whether merging triggers significant star formation episodes compared to isolated galaxies ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
@xcite study the evolution of the age- , mass - metallicity relation of massive galaxies since @xmath294 to @xmath295 , and report that neither new star formation nor chemical enrichment is needed for the evolution of massive quiescent galaxies .
additionally , we do not have measurements of the gas fraction of our merger sample .
therefore we note that our analysis only accounts for the accretion of existing stars and ignores stars formed during mergers , setting the lower limit on the merger contribution to the stellar mass growth .
we show the stellar mass accretion rate as a function of redshift in figure [ fig : mass_growth ] . for the average massive galaxy of log@xmath4
, we find that major ( minor ) merging leads to an average stellar mass growth of and @xmath296 during @xmath5 .
this amounts to a total of @xmath297 being accreted via 1:1 - 10:1 mergers , implying that the average @xmath298 galaxies increase their stellar masses by at least @xmath299 through accreting existing stars from satellite galaxies from @xmath15 to 0.1 .
our results are in agreement with similar observations for bright central galaxies in galaxy clusters @xcite and field galaxies @xcite up to @xmath12 , showing that major merging plays a significant role in the mass assembly of massive galaxies ( and therefore its number density evolution ) independent of the environment .
our stellar mass accretion rates are also consistent with simulation predictions @xcite with the exception of @xcite .
@xcite follow the history of simulated massive galaxies and find that by @xmath300 , 80% of the stars in massive galaxies are formed at @xmath301 ex - situ of the original halo at @xmath302 , and are accreted at @xmath303 with an average rate of @xmath304/yr .
their average mass accretion rate stays relatively flat at @xmath10 and decreases at lower redshift , which is qualitatively similar to our observed trends but on average @xmath305 higher , as seen in fig .
[ fig : mass_growth ] . as we discussed in section
[ sec : expect_minor_merger ] , this is explained by the higher minor merger rates in their simulations compared to the observations of this works and others .
we emphasise that the stellar mass accretion rate presented here does not include new stars formed due to merger - triggered star formation episodes , and therefore represents a lower limit of the true merger - driven stellar mass growth rate ( see also the discussion in section [ sec : pf_ratios ] ) .
dry merging provides a channel to increase the sizes of compact ( @xmath306 kpc ) massive quiescent galaxies ( qgs ) at @xmath10 by a few factors to @xmath295 ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , as discussed in section [ sec : few_minor ] .
we use our measured stellar mass accretion rate to infer an upper limit on the size evolution due to `` dry '' dissipationless merging .
since qgs are expected to remain quiescent for the build - up of the red sequence , and the dissipation from gas in merging galaxies can reduce the efficiency of puffing up sizes of galaxies , for this exercise we make the simplistic assumption that all observed mergers are dissipationless .
the aim of the test is to investigate to what extend the observed frequency of galaxy merging can explain the size evolution of qgs .
we find that the merger fractions of massive galaxies and the quiescent subset are consistent within their uncertainties , therefore we simply use the merger fractions of the overall massive galaxy population in the following analysis .
the virial theorem and more sophisticated merger simulations have been used to predict the size evolution due to dry merging .
the size evolution can be parameterised as @xmath307 , where @xmath308 for major merging and @xmath309 for minor merging predicted using the virial theorem @xcite , or alternatively @xmath310 for major merging and @xmath311 for minor merging according to the simulations of @xcite .
the high value of @xmath312 for minor merging in @xcite implies that it is very efficient in increasing the sizes of galaxies , and likely represents an upper limit due to the high dark matter content and extended stellar haloes of the satellites assumed in their simulation . for each redshift bin
, we multiply the average stellar mass accretion rate ( see section [ sec : mass_growth ] ) with the time elapsed in the redshift bin to get the stellar mass accreted , and scale the predicted size growth to the stellar mass accretion using the @xmath312 values as discussed above .
the maximum merger - driven size growth using both catalogs are plotted in figure [ fig : size_evolution ] .
we observe that the total amount of merging can only increase the size of massive qgs by a factor of two , from 1.5 kpc at @xmath15 to @xmath313 kpc at @xmath295 .
this result is insensitive to the size growth model used , meaning that the virial theorem provides a good approximation of the size evolution due to dissipationless merging .
the observed size evolution of massive qgs ( or early - type galaxies ) has been presented in various works . on figure
[ fig : size_evolution ] we compare our predicted merger - driven size evolution to two recent measurements using candels .
@xcite report an average size growth of @xmath314 from @xmath15 to 0 , with a redshift dependence of @xmath315 , consistent with previous works including @xcite and @xcite . on the other hand ,
@xcite report a consistent but slightly stronger size growth of @xmath316 times in the same redshift range , with a redshift dependence of @xmath317 , similar to the finding of @xcite .
both works report the scatter of the stellar - mass size relation to be consistent with being constant .
the difference of the observed size evolution may be due to the stellar mass threshold , as well as the size measurement technique . as the primary focus of this paper is not the observed size evolution
, we can only conclude that merging increases the sizes of a @xmath298 qg by a factor of two at most from @xmath318 to 0 . while this is insufficient to explain the observed average size growth of a factor of 3 - 5
, it is enough to bring the average sizes of massive qgs to @xmath11 below the local mean stellar mass - size relation if the redshift dependence is on the milder end of the observations ( @xmath315 ) like in @xcite . if the sizes follow a normal distribution , the massive qgs already formed and quenched since @xmath318 evolve through merging to form the smallest 16% ( 2% ) of local massive qgs since they lie at @xmath11 ( @xmath319 ) below the mean .
if the sizes follow a skewed distribution instead , as shown by @xcite , the fraction can be even higher ( e.g. up to the smallest 12.5% for @xmath319 below mean following chebyshev s inequality ) .
this may be a more relevant representation if these compact qgs end up to lie below the local mass - size relation , while the majority of later quenched qgs occupy the upper part of the relation .
recent measurements of compact massive qgs reveal that their number densities peak at @xmath320 , and decrease at lower redshifts @xcite , therefore they must undergo structural changes .
incidentally this is the same redshift range in which our merger rate peaks ( major : @xmath321 , minor : @xmath322 , see fig .
[ fig : merger_rates ] ) .
we will further the discussion on the observed size evolution in section [ sec : etg_evolution ] . even though there may be a significant number of minor mergers rejected by the stellar mass ratio criterion ( flux ratio between 1:1 and 10:1 , but stellar mass ratio more extreme than 10:1 ) , these mergers are more likely to have non - negligible gas mass and more dissipation so it does not help to solve the problem of the observed size evolution .
the gas content of merging galaxies may explain the scatter of the redshift - size evolution @xcite . however without gas measurements we are not able to test this hypothesis at this point .
the virial theorem predicts that equal - mass mergers do not change the stellar velocity dispersion @xmath323 , and minor mergers reduces the @xmath323 by @xmath324 if the satellite has a @xmath323 much lower than the massive galaxy it is merging with @xcite . using the stellar mass accretion rate we estimate that 4:1 - 10:1 minor mergers can only reduce the @xmath323 of massive galaxies by 6% from @xmath15 to 0.1 .
if we relax the assumption and allow 1:1 - 4:1 mergers to be equally efficient in reducing @xmath323 , the total stellar mass accreted implies that the @xmath323 decreases by maximum 25% from @xmath15 to 0.1 . from this
we conclude that merging is insufficient to reduce the high @xmath323 ( @xmath325 km s@xmath17 ) observed in @xmath174 qgs @xcite by @xmath326 to match the average of the local population .
this is consistent with claims that the addition of lower @xmath323 galaxies to the quiescent population at later times contribute to the decreasing average @xmath323 of the overall massive qg population @xcite .
we note that if a significant amount of dark matter is accreted by these massive qgs , the total mass increases and therefore the velocity dispersion and the sizes may change without any observable stellar mass growth . ) above two stellar mass thresholds .
the blue ( red ) filled circles represent the observed @xmath327 of massive galaxies of log@xmath8 ( 11.4 ) , and the error bars represent the poisson error of the number counts .
the triangles represent the major - merger driven @xmath327 growth using two merger observability timescales ( @xmath3280.5 gyr : downward triangles , dashed lines and yellow shades ; 1.0 gyr : upward triangles , solid lines , purple shades ) .
the colored shades show the uncertainty on @xmath327 propagated from the poisson errors of the number of mergers . the predicted major - merger driven @xmath327 growth accounts for the formation of `` new '' massive galaxies above the threshold due to major merging , as well as the reduction in numbers of massive galaxies that merge with each other ( a minor effect as observed ) .
the predicted growth is normalised to the observed @xmath327 of massive galaxies @xmath329 to which we are complete for major mergers .
we only perform this exercise on the ultravista catalog , because the 3dhst+candels contain too few galaxies above these stellar mass thresholds for meaningful @xmath327 constraints .
we find that the slope of the observed @xmath327 evolution of the most massive galaxies follows the predicted slope due to major merging , if the @xmath330 gyr @xcite for major merging which is roughly the average of the two timescales shown . to keep the slope consistent with the observed number densities
, a maximum of @xmath331 stellar mass can be added in addition to major merging , implying @xmath332 for mechanisms other than major and minor merging .
, scaledwidth=50.0% ] to understand what the merger rates from section [ sec : compare_merger_rates ] imply for the overall galaxy evolution , in this section we aim to quantify the contribution of merging to the observed increase in the number density ( @xmath327 ) of massive galaxies in the redshift range @xmath333 . as shown in section [ sec : mass_growth ] , most of the stellar mass accreted is through major merging , so in this section we only consider major merging for which our samples are complete to higher redshifts .
merging can affect the number counts of massive galaxies in two counteracting ways .
on one hand , merging among lower mass galaxies can increase the number of massive galaxies above a stellar mass threshold . on the other hand , merging among massive galaxies already above the mass threshold will lead to a decreased number count .
we denote @xmath334 as the number of mergers with individual stellar masses lower than a given threshold , but with the sum of their stellar masses above the threshold @xcite , and @xmath335 as the number of mergers with the individual stellar masses of both galaxies above the threshold .
the net change of @xmath327 due to major merging is @xmath336 , where @xmath337 and @xmath338 are the comoving volume and the elapsed time of the redshift range , and @xmath223 is the merger observability timescale given the projected separation ( @xmath35 ) range .
the @xmath223 for major mergers with @xmath35 = 10 - 30kpc @xmath32 is about 0.6 - 0.7 gyr @xcite with an error of @xmath339 gyr . in this exercise
we show the results of two values of @xmath223 ( 0.5 and 1.0 gyr ) . since we assume that no new stars are formed during mergers for the reasons discussed in section [ sec : mass_growth ] , the presented quantities mark the minimum merger contribution to the formation of new massive galaxies .
we present the results in figure [ fig : numdens_growth ] .
we find that major merging alone can explain the @xmath327 evolution of galaxies more massive than @xmath340 if @xmath223 lies between 0.5 - 1 gyr .
if @xmath223 was systematically much longer than 1 gyr , then additional mechanisms may be required to explain the @xmath327 evolution of these very massive galaxies .
we note that 3dhst+candels is inadequate for tracing the @xmath327 growth of the most massive galaxies .
the volume probed is too small leading to large cosmic variance on the observed number density and therefore is not shown .
taking our results further , we use the observed @xmath327 evolution of the most massive galaxies to constrain the upper limit of the stellar masses that can be added in addition to major merging .
we increase the stellar masses of all the galaxies by an arbitrary factor , and count the number of galaxies @xmath341 that cross the given mass thresholds .
its contribution to the @xmath327 evolution is @xmath342 .
we find that the observed @xmath327 evolution is marginally consistent with a maximum 15% of stellar mass growth of the overall massive galaxy population in addition to major merging since @xmath318 .
any non - major merging stellar mass growth beyond @xmath331 would overproduce the number of the most massive galaxies . as shown in section [ sec : mass_growth ] , minor merging accounts for @xmath343 of the stellar mass accreted .
therefore we conclude that there remains little room ( @xmath332 ) for the most massive galaxies to increase their stellar masses by mechanisms other than major and minor merging , such as star formation or very minor mergers ( @xmath34410:1 ) .
there are comparative studies of the possible mechanisms that can explain the size evolution @xcite .
merging , in particular dry minor merging , appears to be a viable means to explain the observed size and velocity dispersion evolution .
however , even when we assume that all mergers were dry ( dissipationless ) , the size evolution inferred from our merger fraction can only account for a factor of two of size increase from @xmath3452.5 to 0.1 .
this is marginally consistent with being @xmath11 below the mean stellar - mass size relation of the measurement of @xcite , but @xmath346 compared to that of @xcite .
this necessitates additional mechanisms to explain the observed size increase for the bulk of the population
. the apparent strong size evolution may be in part due to observational effects .
our observations indicate that massive galaxies tend to merge with galaxies with lower stellar mass - to - light ratios ( see figure [ fig : mtol_z ] and section [ sec : pf_ratios ] ) . if the younger , bluer stars of the companion are added to the outskirts of massive galaxies consisting of older stellar populations @xcite , then the half - light radius ( @xmath347 ) measured in rest - frame optical bands increases .
this scenario is supported by the observed negative colour gradients @xcite , and is consistent with the observation of @xcite that the @xmath347 of massive galaxies are smaller when measured at longer wavelengths .
@xcite show that the half - mass radii of massive qgs are on average @xmath348 smaller than the half - light radii measured from the rest - frame @xmath349-band .
therefore the observed size evolution is perhaps in part due to the radial dependence of the @xmath350 .
since the number- and mass - weighted average stellar mass ratio is @xmath2 4:1 for the mergers in this work , the satellites may strip off their stars at the outskirts like the 5:1 intermediate mass ratio merger simulated by @xcite , lending support to merging as a viable explanation for the observed size evolution and color gradients .
it is important to distinguish between the growth of individual galaxies and the evolution of the overall population .
the number density of the massive qgs evolves with redshift , for instance massive ( @xmath298 ) galaxies are 30 times more abundant at @xmath351 than @xmath352 ( e.g. @xcite and references therein , also see section [ sec : numdens_evolution ] ) .
therefore if larger , later quenched galaxies are continuously added to the qg population , it may be sufficient to increase the average sizes of qgs ( more details about the so - called `` progenitor bias '' in @xcite ) .
this assumes that the sizes of qgs are correlated with their age or time since being quenched , a trend which is observed in some works @xcite but not in others @xcite .
another implication is that the scatter of the size evolution is expected to increase if the progenitor bias is the sole explanation for the observed size evolution , which contradicts the constant scatter observed @xcite .
additionally , the progenitor bias alone does not explain the disappearance of compact qgs observed at @xmath10 @xcite .
the number density of compact qgs peaks at @xmath353 and decreases towards lower and higher redshifts .
our merger fractions ( stellar mass ratio selected ) peak at @xmath354 , and one may speculate on a causal relation between the two observations . a fixed number density selection may provide a more direct comparison between massive qgs at @xmath174 and their descendants at lower redshifts ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
if the descendants of compact massive qgs at @xmath355 are the most compact qgs in clusters in the local universe , the sizes of individual qgs will only need to increase by a factor of @xmath356 @xcite , which is in good agreement with the size evolution inferred from our merger rates .
apart from the observational effects and the progenitor bias discussed above , alternative means to increase the sizes of individual qgs have been proposed .
some examples include agn and/or supernova feedback @xcite , adiabatic cooling via the mass loss of old stars @xcite , and halo size evolution @xcite .
it is beyond the scope of this work to draw conclusions on the relative contributions of the possible options in explaining the size evolution .
we emphasise that our results provide a strong constraint : whichever mechanisms are responsible for the observed size evolution , there is little room for further stellar mass to be created or added ( @xmath9 at most for @xmath357 ) for the most massive galaxies ( @xmath358 ) in order not to over - produce the observed numbers at different redshifts . the open question of whether merging is a major contributor to the cosmic star formation history ( sfh ) has been tackled in different ways : do merging galaxies have higher star formation rates compared to isolated ones @xcite ? at each epoch , are star - forming galaxies primarily mergers or isolated disks @xcite ? these different perspectives can lead to seemingly contradictory conclusions . despite the apparent offset of visually identified mergers from the sfr-@xmath359 relation ( dubbed main - sequence , @xcite ) , merging galaxies only show disturbed morphologies for a limited time ( @xmath360 gyr , e.g. @xcite ) .
if the duty cycle of mergers is interpreted as the cause for the scatter of the sfr-@xmath359 relation , major mergers account for a majority of the total sf at @xmath361 @xcite .
@xcite have shown that mergers can enhance sfr to @xmath362 kpc , and such widely separated merging galaxies are likely not identified in morphological selected samples which probe later - stage mergers . on the other hand , the existence of isolated star - forming disks has been used as evidence against mergers being a contributor of cosmic sf budget based on the assumption that mergers destroy disks ( e.g. * ? ? ?
while mergers _ can _ destroy disks and remain a popular explanation for bulge formation @xcite , various works have shown that disks can reform after gas - rich mergers @xcite . even though galaxy merging may not increase the total amount of stars formed from the available cold gas reservoir , it can trigger starburst episodes by temporarily enhancing the star formation efficiency , leading to faster cold gas depletion @xcite .
detailed studies of the sfh of individual galaxies can provide an answer to whether most stars in galaxies are formed during merging or isolated phases ( continuous vs bursty sfh ) .
in section [ sec : pf_ratios ] we have shown that using the @xmath7-band flux ratio to select mergers leads to an increasing merger fraction evolution , as opposed to the flat or diminishing trend seen using stellar mass ratio selected pairs .
the former merger fraction share a similar redshift evolution as the cosmic star formation rate density ( e.g. @xcite and references therein ) albeit with considerable uncertainties : both rise from @xmath295 to @xmath12 and reach a plateau or increase mildly from @xmath12 to @xmath318 .
this may be a hint that at @xmath363 , massive galaxies are primarily merging with low stellar mass ( @xmath364 10:1 ) but gas - rich satellites .
these mergers are classified as major or minor depending on whether the baryon mass or stellar mass ratio is used .
when inferring the merger contribution to the cosmic star formation budget , we need to account for these `` missing '' mergers @xcite that did not enter the stellar mass ratio selection .
future surveys of the molecular gas mass of high-@xmath14 galaxies are needed to make progress on this issue .
the merger fraction of massive galaxies is @xmath365 , resulting in low number densities of mergers ( @xmath366 mpc@xmath367 ) at @xmath10 . as we show in section [ sec : cv ] , cosmic variance is the dominant source of uncertainty in merger fraction measurement with candels - sized surveys , due to the small survey area and low source number density .
we note that the merger fractions measured from ultravista and 3dhst+candels yield very consistent results ( see figure [ fig : pf_3dhst ] ) , even at the redshifts where ultravista is expected to be incomplete for low surface brightness satellites .
this is due to the fact that most satellites have lower @xmath350 ratios ( see section [ sec : pf_ratios ] and figure [ fig : mtol_z ] ) .
as long as the lower limit of @xmath35 is set so that no close pairs are missed due to blending , and the relevant observability timescales are applied for the @xmath35 range @xcite , deep ground - based nir surveys like ultravista and uds provide as accurate results as _ hst _ surveys .
ground - based surveys have the additional advantage of larger sample sizes , so that the evolution can be probed in finer redshift bins with small poisson uncertainties .
put another way , large area surveys are crucial to mitigate cosmic variance and poisson uncertainties in galaxy merger fraction measurements .
a limitation of the pair selection is that a minimum @xmath35 must be imposed to match the resolution of the imaging data , for example 10 kpc @xmath32 in this work .
if the scientific interest is on the incidence of late stage mergers of @xmath368kpc @xmath32 among agns or ulirgs ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , alternative merger identifications may be a more appropriate choice ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
photometric redshifts ( photo-@xmath14 s ) are essential in removing line - of - sight projected pairs from merger samples .
the projected pair fraction is redshift dependent and can reach @xmath369 at @xmath370 ( see table [ table : pf_massratio ] ) .
statistical simulations can provide an estimate for the number of projected pairs , however photo-@xmath14 s are crucial for selecting real mergers for spectroscopic follow - up .
one may expect photometric samples of mergers to include more mergers due to the larger uncertainties of photo-@xmath14 s than spec-@xmath14 s , however the merger fractions presented in this work using photometrically selected mergers are in agreement or even lower than those using spectroscopic selected mergers @xcite . aside from the variations of the parent sample
as discussed in @xcite , this may be an indication that the selection effects associated with the spectroscopic merger samples outweigh the uncertainties of photo-@xmath14 s in photometric merger samples , e.g. mass - incompleteness ( due to flux - limited selection ) , slit / fiber placement incompleteness , limited sample sizes and so on .
therefore we argue that large - area ( @xmath371 deg@xmath20 ) surveys with accurate photo-@xmath14 s currently provide the most time - efficient datasets for measuring galaxy merger fractions . on the theoretical front ,
the merging probability of galaxy pairs in close physical separations need to be quantified as a function of redshift and environment , as discussed in section [ sec : method ] .
it is also important to understand how galaxy fly - bys can impact the structure and dynamics of massive galaxies .
these are subtle yet crucial quantities that fold into the interpretation of the inferred galaxy merger rates , which are paramount in determining whether galaxy merging is a significant driver of its evolution .
we present the largest sample of photometrically selected mergers at @xmath333 from mass - complete catalogs , using complementary datasets of a large area ground - based survey ( ultravista ) and a deep spaced - based survey ( 3dhst+candels ) .
we measure the galaxy major and minor merger fractions ( @xmath51 and @xmath52 ) . applying the merging observability timescale ( @xmath223 ) from @xcite , we infer the merger rates , as well as the evolution in stellar mass , size and number density for massive galaxies .
we summarise our findings as follows : 1 .
the merger fraction shows a steep increase from @xmath295 to 1 , with @xmath51 showing a stronger evolution than @xmath52 . using the stellar mass ratio selection ( figure [ fig : pf_massratio ] , left ) , @xmath51 and @xmath52
show a plateau at @xmath372 and diminishes beyond @xmath320 .
if the observed @xmath7-band flux ratio selection is used instead ( figure [ fig : pf_fluxratio ] , right ) , @xmath51 and @xmath52 increase monotonically with redshift .
the ultravista and 3dhst+candels show discrepant results at @xmath39 due to the magnitude limit of the ultravista dr1 survey .
2 . selecting mergers by
the observed @xmath7-band flux ratio leads to an increasing merger fraction with redshift , while selecting mergers by stellar mass ratio shows a diminishing redshift dependence .
this variation in merger selection technique is the cause of the discrepant merger fraction measurements at @xmath39 in the literature @xcite .
the discrepancy is a consequence of the @xmath350 evolution of galaxies with redshift : at high redshifts and lower @xmath359 , galaxies have higher star formation rates and lower @xmath350 ratios .
the two selections produce consistent merger fractions at @xmath373 , but the fractions diverge at @xmath374 the @xmath7-band flux ratio selection is biased towards bright , star - forming low - mass satellites at @xmath363 , and the stellar mass ratio selection is biased against low - mass satellites which have significant cold gas mass .
cold gas measurements for massive galaxies and their satellites are required to refine the merger definition using the baryon mass ratio .
our inferred merger rates using the stellar mass ratio selection is consistent with the gas - poor ( @xmath375 ) merger rates of the simulations of @xcite . on the other hand ,
our inferred merger rates using the @xmath7-band flux ratio selection is consistent with their predicted gas - rich ( @xmath376 ) ones .
we get consistent merger rates when mergers are selected from different @xmath35 bins ( 5 - 20 , 10 - 30 , 10 - 50 , 10 - 100 kpc @xmath32 ) when the relevant @xmath223 from @xcite are applied .
however , we note that the widest @xmath35 are systematically higher than the other bins , with a more noticeable discrepancy at lower redshift .
this is consistent with the pairs at 50 - 100 kpc @xmath32 probing large - scale structure formation .
the results imply that an average massive ( @xmath377 ) galaxy experiences @xmath378 major and @xmath379 minor mergers over the redshift range of @xmath5 , if mergers are selected by stellar mass ratio . there may be an additional @xmath380 major merger and @xmath360 minor merger if mergers are selected by the @xmath7-band flux ratio .
the mass - weighted average stellar mass ratio is @xmath2 3:1 - 4:1 , implying that the inferred stellar mass accretion rate is primarily driven by intermediate mass ratio mergers up to @xmath318 .
this work extend the expectations from @xmath381 to @xmath318 that major merging is the dominant process for stellar mass accretion for massive galaxies .
7 . major and minor merging combined can at most increase the sizes by a factor of two from @xmath15 to 0.1 for an average @xmath382 quiescent galaxy , if we assume that all mergers are dry .
additional mechanisms are thus required to explain the strong observed size evolution ( factor of @xmath383 ) .
the observed amount of major merging is sufficient to explain the evolution of the formation of new massive ( @xmath358 ) galaxies by number density arguments .
these very massive galaxies can only increase their stellar masses by at most @xmath384 during @xmath357 by processes in addition to major and minor merging , in order to match the observed number density evolution .
this hints that star formation and very minor merging are unlikely mechanisms responsible for the observed size evolution .
we are grateful for the contributions of the ultravista , comsos , candels and 3dhst collaborations for making the catalogs available for public use . am acknowledges tomo goto , knud jahnke , and jennifer lotz for helpful conversations at the early phase of this project .
am also thanks bo milvang - jensen for clarification of the ultravista data , and anna gallazzi for clarifying the stellar population models .
the dark cosmology centre is funded by the danish national research foundation .
we acknowledge the hpc facility at the university of copenhagen for providing the computing resources used in this work .
st and az gratefully acknowledge support from the lundbeck foundation .
this work has made use of the ultravista catalog , which is based on data products from observations made with eso telescopes at the la silla paranal observatory under eso programme i d 179.a-2005 and on data products produced by terapix and the cambridge astronomy survey unit on behalf of the ultravista consortium .
this work is in part based on observations taken by the 3d - hst treasury program ( go 12177 and 12328 ) with the nasa / esa hst , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas5 - 26555 .
the 3dhst+candels catalog is compiled using the datasets in these papers : @xcite ; almaini / foucaud in prep and dunlop et al . in prep .
in order to measure the merger fraction evolution robustly , it is essential to ensure completeness in the identification of merging satellites especially at high redshifts .
we assess the completeness of faint satellites in two aspects : 1 .
stellar mass completeness : is ultravista mass complete at high-@xmath14 for the 10:1 satellites ? 2 .
surface brightness ( sb ) : do we miss low sb faint satellites ?
we present our analysis in the following subsections .
we estimate the stellar mass ( @xmath359 ) completeness of the ultravista catalog by comparing the @xmath385-band magnitudes and photo-@xmath14 s of the detected galaxies with those of the deeper @xmath385-band selected fireworks catalog ( @xmath386 at 5@xmath34 depth , @xcite ) in the chandra deep field south . assuming that the fireworks catalog is 100% complete , we take the fractions of massive galaxies in fireworks above different @xmath359 in different redshift bins which are fainter than the ultravista survey magnitude limit as the mass completeness limits .
the results are shown in figure [ fig : mass_completeness ] . from this comparison
we estimate that for the ultravista sample , massive galaxies of log@xmath4 are @xmath387 complete at @xmath388 .
major ( @xmath389 4:1 ) satellites of log@xmath390 log@xmath391 are above 80% complete for @xmath392 .
minor satellites ( 4:1 @xmath393 10:1 ) of log@xmath394 log@xmath395 are above 80% complete for @xmath396 .
we list the @xmath387 limits in table [ table : completeness_all ] .
the candels survey is sensitive to faint objects ( @xmath397 at 5@xmath34 depth , @xcite ) .
for example quiescent galaxies with @xmath398 are 50% complete at @xmath399 ( 3.2 ) for wide and deep regions @xcite , therefore we expect the stellar mass completeness not to be an issue .
-band magnitude distribution of ultravista to the deeper fireworks catalog .
the stellar mass bins of the massive galaxies of log@xmath4 , as well as their 4:1 and 10:1 satellites are shown in different colors as indicated in the legend .
the dashed line shows the @xmath400 completeness limit .
, scaledwidth=50.0% ] ccc ultravista & @xmath401 & @xmath402 + candels & & @xmath402 + ultravista & @xmath403 & @xmath40 + candels & & @xmath42 + ultravista & @xmath40 & @xmath41 + candels & & @xmath15 + [ table : completeness_all ] the detection of objects at faint magnitudes is sensitive to their surface brightness ( sb ) profiles and the source extraction thresholds . in order to test the redshift limit up to which we are complete to detecting the faintest possible satellites , we simulate the source detection by simulating galaxies with a range of sersic profiles with magnitudes determined by a dust - free , maximally old stellar population at given @xmath359 and redshifts .
the effective half - light radii ( @xmath347 ) assumed are the extrema calculated from the observed scaling relations and/or simulations , as described in detail below . to emulate the actual observations of ultravista and candels ,
the sersic profiles are smoothed to the instrument psf and added to images with blank patches of sky , and sextractor is run with the object detection settings of the respective catalogs .
structural measurements of intermediate mass ( @xmath404 ) galaxies at @xmath355 are sparse due to their faintness .
we list the possibilities here and select the extreme sizes for our simulations . 1 .
observationally , the sizes of local elliptical or early - type galaxies scale with stellar mass as @xmath405 for @xmath406 .
the observed @xmath174 stellar mass - size relation has a similar slope @xcite .
if intermediate mass galaxies have the same stellar density as massive galaxies , then the radius scales with stellar mass as @xmath407 .
lastly , numerical simulations for merger - driven size evolution have shown that a hernquist profile in projection can be described by a sersic index of @xmath408 @xcite .
these simulations use the same scale radius for the stellar halos of the host galaxy and the satellite which has only a tenth of the host stellar mass for the `` diffuse '' case . considering the above mentioned possibilities
, we simulate the two extreme sizes of a @xmath409 quiescent galaxy : the most compact ( constant stellar density : @xmath407 ) and the most extended ( simulation : @xmath410 ) .
observations show that a @xmath411 quiescent galaxy has log@xmath412 at @xmath413 and log@xmath414 at @xmath415 @xcite , with a scatter of @xmath416 .
we scale the sizes to one - tenth of the stellar mass with the extreme scenarios , e.g. our simulated @xmath409 galaxy at @xmath15 has @xmath347 of 0.29 kpc ( compact ) to 1.95 kpc ( extended ) , equivalent to 0.035@xmath417 and 0.248@xmath417 .
we assume a maximally old , dust - free stellar population with a single burst and highest metallicity ( @xmath418=0.03 ) to compute the faintest possible magnitudes for these intermediate mass galaxies using the updated version ( 2012 ) of the stellar population synthesis model library of @xcite .
this corresponds to magnitude limits of @xmath419 and @xmath420 for a @xmath409 maximally old galaxy at @xmath15 .
we simulate different light profiles using three sersic indices ( @xmath421[0.5 , 1 , 4 ] ) , in which the latter two represent the exponential disk profile and the de vaucouleurs profile respectively .
we assume two axial ratios of @xmath422 $ ] , though we note that lower axial ratios are easier to detect when the source is closer to the sb limit . with the assumed parameters we generate sersic models according to the @xmath7 and @xmath385 limits .
we smooth the images with a gaussian beam corresponding to the psf size of the imaging surveys .
then we add them to blank regions on the candels - wide @xmath7-band and the ultravista @xmath385-band images , and we extract sources from the simulated images with the corresponding sextractor settings of the two surveys .
we outline the results of our simulation for both catalogs . for the ultravista dr1 catalog , as long as the source is brighter than @xmath385=24.2 - 24.3 mag arcsec@xmath20 , we are able to extract the sources for all the sersic models simulated .
this corresponds to @xmath40 ( 1.5 ) for using ultravisata dr1 to detect major ( minor ) satellites .
the limit for the candels wide catalog is @xmath7=26.45 mag arcsec@xmath20 , corresponding to @xmath160 ( 2.5 ) for major ( minor ) satellites .
we note that these limits are more constraining that those derived from a simple stellar mass completeness argument ( appendix [ sec : mass_completeness ] ) . from this test
we observe that the source detection for faint objects close to the sb limit depends on the following structural parameters : ( 1 ) @xmath347 : for a given integrated magnitude , the larger the @xmath347 the lower the sb per pixel . sufficient pixels ( 10 pixels following ultravista and candels settings ) above the sb threshold are required for a detection ; ( 2 ) @xmath423 : for a given integrated magnitude , the combination of a very low @xmath423 and very extended @xmath347 may lead to too low sb / pix for detection .
on the other hand , for a very high @xmath423 and very compact @xmath347 a non - detection may result due to the insufficient number of pixels above the detection threshold ; ( 3 ) @xmath424 : if the axis ratio is close to 1 , the flux densities are divided over more pixels than the case of a lower @xmath424 , resulting in an insufficient number of pixels above the detection threshold .
we note that our derived limits may be subject to change , if there are systematic uncertainties in the magnitudes and/or the stellar mass .
namely , the magnitude limits are derived from dust - free models , which may be reasonable assumptions given that the faintest possible galaxies at @xmath15 are not actively star - forming .
on the other hand , there are known systematic uncertainties in stellar masses ( @xmath425 dex ) and ages from sed fitting due to different assumptions of imf or stellar population synthesis model .
if the modeled magnitudes are actually fainter or if the stellar masses are underestimated , then our sb completeness limit may be lower than the numbers quoted here .
we only compare our results with previous merger fraction measurements using the close pair selection but not the morphological selection ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
as the morphological selection is sensitive to the imaging quality , merger fraction measurements may be subject to large systematic uncertainties beyond @xmath12 .
we refer readers to @xcite for a comprehensive review on the two methods , and focus on comparing our results with works that use the close pair method to identify mergers .
we note that for the few studies which cover a different @xmath35 range than our data points shown on figure [ fig : pf_lit ] , we use the observability timescales of @xcite to correct the merger fractions for a fair comparison .
we compare our merger fractions with @xmath426 studies using the close pair selection .
as the selection criteria vary slightly across studies , we re - run our selection according to the published studies for a fair comparison .
we compare our merger fractions with similar studies that select mergers using the stellar mass ratio @xcite . in these studies ,
the projected separation limits are @xmath35 = 13 - 30 kpc @xmath32 and 10 - 30 kpc @xmath32 respectively .
we replicate the selections by slightly modifying our criteria : we search for satellites around massive quiescent galaxies ( @xmath427 and ssfr @xmath428 ) , using a limit of @xmath35=10 - 30 kpc @xmath32 .
we note that the results of @xcite are based on satellites around quiescent galaxies at lower stellar masses ( @xmath429 ) .
we check that lowering the stellar mass cut by 0.3 dex gives consistent merger fractions within the large poisson uncertainties , as is also shown in ( * ? ? ?
* table 3 ) .
the comparison is shown in figure [ fig : pf_lit_massratio ] ( left ) .
we find our @xmath51 to be consistent with that of @xcite , and the one measured from ultravista is @xmath430 higher than that from @xcite at @xmath12 and 1.8 .
we note that in these redshift bins , @xcite show slightly higher @xmath52 than in other fields .
therefore we conclude that the combined @xmath51 and @xmath52 measured in our data and in @xcite are in good agreement .
the discrepancy of @xmath431 in the @xmath51 can be explained by the separation of major and minor mergers , as well as cosmic variance and photo-@xmath14 criterion variation .
this discrepancy does not affect the conclusions made in this work .
@xcite and @xcite search for satellites of @xmath7-band flux ratios down to 4:1 around galaxies more massive than @xmath432 , within projected separations of @xmath433 kpc ,
i.e. 21 kpc @xmath32 . in particular , @xcite impose a lower limit of @xmath434 kpc to screen out confused pairs which are likely unresolved with nicmos .
this comparison is illustrated in figure [ fig : pf_lit_fluxratio ] ( right ) .
our @xmath51 is consistent with these studies .
@xcite present the first measurement of the @xmath51 at @xmath13 in the hudf using the stellar mass ratio selection .
they use a smaller @xmath435 20 kpc @xmath32 and search for satellites around galaxies of @xmath436 , which is six times lower than our mass criteria .
this may explain why their @xmath51 to be @xmath239 higher than ours . as discussed in section [ sec : future ] , flux - limited spectroscopic surveys may lead to biased merger fractions due to mass incompleteness , slit / fiber collision , etc .
bearing in mind the difference in the merger selection , we compare our results using photometric mergers with those using spectroscopic mergers .
our results are consistent with @xcite who measure the @xmath51 and @xmath52 of @xmath437 galaxies from the spectroscopic survey of vvds up to @xmath12 using the @xmath172-band flux ratio selection .
the observed @xmath7-band corresponds approximately to the rest - frame @xmath172-band at @xmath318 and therefore our results using the flux ratio selection are directly comparable to their work . @xcite and @xcite
extend measurements of spectroscopic merger fractions to @xmath438 , in which the former use a flux ratio selection for star - forming galaxies and the latter a stellar mass ratio selection .
both works report a @xmath51 of @xmath439 .
our major merger fraction are marginally consistent with that of @xcite although we note that their primary sample consists of star - forming galaxies only , and may include more mergers if merging does trigger star formation activity .
our merger fractions are @xmath440 lower than that of @xcite .
both of these studies sample the mergers around less massive galaxies ( 0.8 - 1.7 dex lower than our mass limit ) , and we speculate that it may account for the higher fractions .
@xcite select mergers photometrically with the @xmath385-band flux ratio , and report a mildly increasing @xmath51 for massive ( @xmath442 ) galaxies from @xmath300 to 1.2 .
when compared to our @xmath51 using the @xmath7-band flux ratio for the similar @xmath359 and @xmath35 range ( figure [ fig : pf_lit_fluxratio ] , right ) , our results are in good agreement with theirs .
@xcite measure the @xmath51 and @xmath52 of massive ( @xmath442 ) galaxies in zcosmos at @xmath443 , selecting mergers by stellar mass ratio and relative velocity @xmath444 km s@xmath445 .
they find a redshift dependence of the @xmath51 as @xmath446 , and a redshift - constant @xmath52 in this redshift range .
@xcite present results for @xmath51 at @xmath447 for cosmos with similar selection criteria .
we compare to their @xmath51 for galaxies with log(@xmath448)=11 - 11.4 .
our results are consistent to these two works , as shown in figure [ fig : pf_lit_massratio ] ( left ) .
@xcite demonstrate that the variation in selecting the parent galaxy sample and the mass ratio probe leads to different redshift trends in the merger fraction .
therefore we do not compare our results directly with the pair fraction measurements at @xmath441 with different selection criteria ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we note that once the selection differences are accounted for , the merger rate per galaxy presented in section [ sec : compare_merger_rates ] of this work is consistent with those inferred from these works as presented in @xcite : both follow a monotonically increasing trend from @xmath295 to @xmath449 .
, h. , kosugi , g. , akiyama , m. , takata , t. , sekiguchi , k. , & furusawa , j. 2008 , in astronomical society of the pacific conference series , vol .
399 , panoramic views of galaxy formation and evolution , ed .
t. kodama , t. yamada , & k. aoki , 131 | we measure the merger fraction of massive galaxies using the ultravista / cosmos _ ks_-band selected catalog , complemented with the deeper , higher resolution 3dhst+candels catalog selected in the _ hst_/wfc3 _
h_-band , presenting the largest mass - complete photometric merger sample up to @xmath0 .
we find that selecting mergers using the @xmath1-band flux ratio leads to an increasing merger fraction with redshift , while selecting mergers using the stellar mass ratio causes a diminishing redshift dependence .
defining major and minor mergers as having stellar mass ratios of 1:1 - 4:1 and 4:1 - 10:1 respectively , the results imply @xmath21 major and @xmath31 minor merger for an average massive ( log@xmath4 ) galaxy during @xmath5 .
there may be an additional @xmath6 major ( minor ) merger if we use the @xmath7-band flux ratio selection .
the observed amount of major merging alone is sufficient to explain the observed number density evolution for the very massive ( log@xmath8 ) galaxies .
we argue that these very massive galaxies can put on a maximum of @xmath9 of stellar mass in addition to major and minor merging , so that their number density evolution remains consistent with observations .
the observed number of major and minor mergers can increase the size of a massive quiescent galaxy by a factor of two at most .
this amount of merging is enough to bring the compact quiescent galaxies formed at @xmath10 to lie at @xmath11 below the mean of the stellar mass - size relation as measured in some works ( e.g. * ? ? ?
* ) , but additional mechanisms are needed to fully explain the evolution , and to be consistent with works suggesting stronger evolution ( e.g. * ? ? ?
* ) . |
the merging of two galaxies will produce a binary black hole at the center of the newly formed galaxy .
if the two black holes do not stall , they will ultimately merge due to emission of gravitational wave radiation .
the gravitational waves carry away linear momentum , causing the centre of mass of the coalescing bh system to recoil in the opposite direction ( peres 1962 , bekenstein 1973 ) .
early analytical calculations predicted that mergers of non - spinning black holes can attain kicks with velocities of up to a few hundred kms@xmath0 ( e.g. , fitchett & detweiler 1984 , favata et al .
2004 , blanchet et al .
2005 , damour & gopakumar 2006 ) , recently confirmed by numerical simulations ( e.g. , baker et al .
2006 , herrmann et al . 2007a , gonzlez et al .
these velocities are above the escape velocity of dwarf galaxies , low - mass spirals , and high - redshift dark matter halos .
if many bhs were removed from their hosts in the early history of the universe , this would have profound consequences for galaxy assembly and bh growth in the early universe , and would give rise to a population of interstellar and intergalactic bhs ( e.g. , madau et al .
2004 , merritt et al .
2004 , madau & quataert 2004 , haiman 2004 , yoo & miralda - escud 2004 , volonteri & perna 2005 , volonteri & rees 2006 , libeskind et al .
2006 ) .
recent numerical relativity simulations of certain configurations of merging , _ spinning _ bhs have produced much higher recoil velocities , up to several thousand kms@xmath0 ( campanelli et al .
2007a , b , gonzlez et al .
2007b , tichy & marronetti 2007 , herrmann et al
. 2007b , dain et al .
2008 , schnittman et al .
2008 ) , scaling to an expected maximum around 4000 kms@xmath0 ( campanelli et al . 20007a , b , baker et al . 2008 ) for maximally spinning equal - mass binaries with anti - aligned spins in the orbital plane .
these kick velocities exceed the escape velocities of even massive elliptical galaxies ( fig . 2 of merritt et al .
2004 ) and therefore the new results reinforce and enhance consequences studied earlier for the smaller kicks , with potentially far - reaching implications for the early phases of bh growth from early stellar - mass precursors or later intermediate - mass precursors ( schnittman 2007 , volonteri 2007 ) and consequently for the frequency of gravitational wave signals detectable with _ lisa _ ( sesana 2007 ) , for the scatter in the @xmath1 relation ( libeskind et al .
2006 ) , and for the offsets and oscillations of recoiling bhs in galaxy cores ( gualandris & merritt 2008 ) .
the recoiling black holes will carry a fraction of nuclear gas and stars with them ( merritt et al .
2004 , 2006 , madau & quataert 2004 , loeb 2007 )
. they would be detectable spatially in the form of seyfert or quasar activity offset from the galaxy core ( madau & quataert 2004 ) , or in the form of broad emission lines kinematically offset from the narrow emission lines ( bonning et al .
2007 , komossa et al .
2008 ) . because of the broad astrophysical implications , the search for and actual identification of such recoiling black holes is of great interest , and will place important constraints on bh growth during the epoch of structure formation , on predictions of maximum recoil velocity , and on arguments suggesting that the bh spin configurations leading to maximal recoil velocities should be rare in gas - rich mergers ( bogdanovi et al . 2007 )
. bonning et al .
( 2007 ) searched for recoiled smbhs in the sloan digital sky survey ( sdss ) database , looking for systematic kinematic offsets between broad - line gas attached to the recoiling bh , and narrow - line gas left behind .
they did not find any promising candidate , and concluded that smbh recoil with large kick velocities is relatively rare . here
, we present the best candidate to date for a recoiling smbh , the quasar + sdssj092712.65 + 294344.0 ( sdssj0927 + 2943 hereafter ) . its unusual emission - line spectrum matches key predictions from the recoiled - smbh scenario .
we use a cosmology with @xmath2=70 kms@xmath0mpc@xmath0 , @xmath3=0.3 and @xmath4=0.7 throughout this letter .
sdssj0927 + 2943 at redshift @xmath5=0.713 is a luminous quasar , observed in the course of the sdss ( adelman - mccarthy et al .
2007 ) , and was found by us in a systematic search for active galactic nuclei ( agn ) with high [ oiii ] velocity shifts .
the sdss spectrum , corrected for the galactic reddening of e(b - v ) = 0.021 mag , is displayed in fig .
the underlying continuum spectral energy distribution ( sed ) was modeled as a powerlaw with a best - fit slope of @xmath6 ( where @xmath7 ) .
each emission line was fit by a single gaussian except the feii multiplets , which were modeled by templates built from i zw 1 ( @xmath8 , vron - cetty et al .
2004 ; @xmath9 , tsuzuki et al .
the redshifts of the feii lines were tied either to mgii ( the uv multiplets ) or to broad h@xmath10 ( the optical multiplets ) .
two systems of strong emission lines can be identified in the spectrum , which we refer to as the `` red '' ( r ) and `` blue '' ( b ) systems .
the red system consists of very narrow emission lines ( red nels , r - nels hereafter ) of [ oiii]5007 , [ oii]3727 , [ neiii]3869 , faint [ nev]3426 and balmer lines , all of them almost unresolved ( fwhm , obs([oiii ] ) = 230 kms@xmath0 ; the widths of the narrow lines are all very similar , and were therefore all fixed to the same value in order to derive fluxes ) .
the blue system shows classical broad balmer and mgii2798 emission lines ( bels ) , plus unusually broad nels ( blue nels , b - nels hereafter ) .
all lines of the blue system are blueshifted by about 2650 kms@xmath0 relative to the r - nels(see tab . 1 for redshifts ; the value of 2650 kms@xmath0 is the shift between broad h@xmath10 and r-[oiii ] ) .
the b - nels show broad [ nev ] with a width of fwhm([nev])=2080 kms@xmath0 , and broad [ neiii ] with fwhm([neiii])=1020 kms@xmath0 .
[ oiii ] and [ oii ] are present , too , with widths of 460 kms@xmath0 .
the bels appear in balmer lines and in mgii with fwhm(h@xmath10)=5740 kms@xmath0 and fwhm(mgii)=3530 kms@xmath0 .
line ratios indicate agn - like excitation in both systems , r - nels and b - nels .
emission - line properties are summarized in tab . 1 . in order to see whether any ( faint ) broad - line emission is also accompanying the r - nels ,
we have performed the following test .
we have first subtracted the best - fit continuum , feii multiplets , and nels from the observed spectrum , and then fit the h@xmath10 regime with two broad gaussians .
the redshift of the second gaussian was fixed to that of the r - nels , and its width constrained to be in the range @xmath11 km s@xmath0 .
no successful fit could be obtained if a contribution of the second gaussian is enforced ; its contribution is always consistent with zero .
we therefore conclude that only the b - nels are accompanied by broad - line gas at the same redshift .
the broadband sed of sdssj0927 + 2943 is rather blue with sdss magnitudes of @xmath12 , @xmath13 , @xmath14 , @xmath15 , @xmath16 , and _ galex _
( martin et al . 2005 ) magnitudes of @xmath17 and @xmath18 .
assuming that the standard broad - line region ( blr ) scaling relations ( kaspi et al . 2005 ) hold , we expect a blr radius of @xmath190.1 pc and estimate a smbh mass of sdssj0927 + 2943 of @xmath20 m@xmath21 from the width of h@xmath10 and the luminosity at 5100 .
we have searched the x - ray archives for observations of sdssj0927 + 2943 .
the quasar is serendipitously located close to the edge of two rosat hri images . the deeper exposure , of 19 ks duration , was performed in april - may 1995 .
x - ray emission from sdssj0927 + 2943 is detected with a countrate of 0.0037@xmath22 cts / s , which translates into a soft x - ray luminosity of @xmath23 erg / s ( assuming an x - ray spectrum with no intrinsic absorption , and with photon index @xmath24 ) .
the second hri image was taken in november 1994 with a duration of 10 ks .
sdssj0927 + 2943 is detected with a countrate of 0.0047@xmath25 cts / s , consistent with the other measurement .
the x - ray detection demonstrates the presence of an inner accretion disk ( relevant for the discussion in sect .
lllcc r - nel & [ oiii]5007 & 0.71279 & 170 & 10.1 + & h@xmath10 & 0.71279 & 170 & 1.0 + & h@xmath26 & 0.71279 & 170 & 0.4 + & [ neiii]3869 & 0.71279 & 170 & 0.7 + & [ oii]3727 & 0.71279 & 170 & 2.6 + & [ nev]3426 & 0.71279 & 170 & 0.3 + b - nel & [ oiii]5007 & 0.69713 & 460 & 6.7 + & h@xmath10 & 0.69713 & 460 & 1.0 + & [ neiii]3869 & 0.69678 & 1020 & 1.8 + & [ oii]3727 & 0.69801 & 460 & 1.5 + & [ nev]3426 & 0.69709 & 2080 & 4.0 + bel & h@xmath10 & 0.69770 & 5740 & 30.7 + & h@xmath26 & 0.69970 & 3880 & 10.7 + & h@xmath27 & 0.69996 & 3260 & 3.8 + & mgii 2798 & 0.69832 & 3530 & 29.5
several pairs of quasars have been detected at projected separations of 3 - 10@xmath28 . while some of them can be explained by lensing , others very likely represent real pairs ( e.g. , kochanek et al .
1999 ; review by komossa 2003 ) . is sdssj0927 + 2943 actually a binary quasar ?
we temporarily refer to this hypothetical system as sdssj0927 + 2943a , b .
the difference in velocity of the two sets of emission lines exceeds the peculiar velocities observed in galaxy clusters , and is too large for the two galaxies to form a bound merger .
their redshift difference corresponds to @xmath19100 mpc , if their redshifts are cosmological .
consequently , we would then have a very unlikely projection effect , including not just one , but two intrinsically extremely unusual agn : sdssj0927 + 2943a with exceptionally broad neon lines of [ neiii ] and [ nev ] . and
sdssj0927 + 2943b as one of the rare type-2 quasars with , in addition , exceptionally narrow emission lines [ their observed width of @xmath19230 kms@xmath0 is below that typically observed in quasar narrow - line regions ( fwhm @xmath29 kms@xmath0 ; zakamska et al . 2003 ) ] .
this rare source would have to be projected by chance behind the other unusual agn .
we are therefore let to consider the alternative hypothesis that we actually see only one agn , its smbh and bound emission - line region having been ejected from the core with a line - of - sight speed of 2650 kms@xmath0 , while the bulk of the narrow - line region ( nlr ) and other interstellar medium ( ism ) was left behind and shines in narrow emission lines .
this is the scenario actually predicted in the context of the recent black hole recoil simulations discussed in sect . 1 , and it is the observational signature bonning et al .
( 2007 ) searched for , and komossa et al .
( 2008 ) paid attention to , but did not detect .
in this picture , sdssj0927 + 2943 underwent a merger in the past . upon merging , its central smbh recoiled taking with it the blr and perhaps other very high ionization gas , leaving behind the nlr .
gas with a velocity larger than the recoil velocity will remain bound to the smbh , and a trail of partially bound gas may form .
accretion activity may have switched off temporarily in the course of the binary smbh merging , once the orbital decay time due to emission of gravitational waves had become smaller than the viscous timescale of the disk ( e.g. , liu et al .
2003 , milosavljevi & phinney 2005 ) .
after coalescence , the inner disk will re - form quickly ( loeb 2007 ) and the accreting smbh will then illuminate the bound gas , a trail of partially bound gas and swept - up ism , the surrounding ism / halo and the left - behind bulk of the nlr gas .
the nlr gas left behind will retain memory of the original ionization from the pre - merger accretion phase for only limited time , given by the light - travel time , and the hydrogen recombination timescale , @xmath30 yr , where @xmath31 is the gas temperature in 10@xmath32 k , and @xmath33 is the gas density in units of 10@xmath34 @xmath35 .
[ oiii ] will fade away more quickly than hydrogen lines ( binette & robinson 1987 ) .
however , once the recoiling smbh re - forms its inner accretion disk , it will re - illuminate parts of the nlr left behind ( from larger distance than before ) , and any surrounding ism .
emission - line ratios would depend on the average density which would be higher in the classical nlr than in other ism / halo gas , and on metal abundances which will be lower in the halo .
which aspects of this scenario do we observe in sdssj0927 + 2943 ? the blue system of emission - lines , the bels and b - nels , represents gas bound to the recoiling smbh , seen in broad balmer lines , broad mgii and broad forbidden lines , especially [ nev ] .
after coalescence and recoil , matter orbiting the smbh with a velocity much larger than the recoil velocity will remain bound to the smbh ( merritt et al .
2006 ) ; i.e. matter within a region whose size is given by @xmath36 .
in our case , @xmath37 cm , which is about a factor of 3 larger than the blr of sdssj0927 + 2943 .
the width of [ neiii ] ( of the b - nel system ) of 1020 kms@xmath0 is too narrow for this gas to be originally bound to the smbh ( except in case of projection effects ) , but as the disk keeps accreting onto the recoiling smbh , it will spread radially , and will also drive some outflows . the semi - broad [ oii ] and [ oiii ] most likely have this same origin , and/or are from swept - up ism in front of the path of the recoiling smbh emission from the gas bound to the recoiling smbh ; even though shielding geometries might be constructed in which this actually becomes possible . ] .
the red system of emission lines , the r - nels , represents the nlr gas left behind , and/or ism surrounding the recoiling smbh , not bound to it , but illuminated by its accretion disk .
all r - nels have the same profiles and are very narrow , and the emission - line ratios imply an agn ionizing continuum .
can we distinguish between nlr and other ism / halo gas , in terms of line ratios and line profiles ?
ratios _ would depend on density and distance . as the recoiling
smbh moves away from the galaxy core , illumination of the nlr from an , on average larger , distance would decrease the degree of ionization , while illumination of low - density ism surrounding the recoiling smbh would increase the degree of ionization of the line - emitting gas .
it is possible that we see a mix of these two processes .
we do not expect to see pure very low - density halo gas .
its high ionization parameter would lead to a much higher degree of ionization than we actually observe , as we have also verified by test calculations with ferland s photoionization code _ cloudy _
( s. komossa et al .
, in preparation ) .
the density - sensitive [ sii ] doublet is outside the observed wave band , but will be accessible with nir spectroscopy .
independent evidence for the origin of the r - nels comes from their _ profiles _ which are very narrow .
their width is below the stellar velocity dispersion of @xmath38=260 kms@xmath0 predicted from the @xmath39 relation ( ferrarese & ford 2005 ) .
this indicates that the r - nel gas illuminated by the recoiling smbh either does not feel the full bulge potential , or the local velocity field of the illuminated gas does not reflect the full velocity dispersion .
this latter situation is expected if the recoiling smbh was in a disk where the velocity pattern is predominantly rotational , so that we only see a fraction of the full pattern .
the whole parameter space of bh recoil velocities is still being explored ( e.g. , baker et al . 2006 , 2007 , campanelli et al . 2007a , b , choi et al .
2007 , gonzlez et al . 2007a , b , herrmann et al .
2007a , b , pollney et al .
2007 , schnittman et al .
2008 , and references therein ; see pretorius 2007 for a review ) .
several recent calculations focused on ( almost ) equal mass bhs with specific spin configurations such that kick velocities are maximized .
the line - of - sight velocity we measure is comparable to the recoil velocities predicted by the runs of gonzles et al .
( 2007b ; @xmath40 kms@xmath0 ) and by dain et al .
( 2008 ; 3300 kms@xmath0 ) .
the scaling formulae of campanelli et al .
( 2007a ) and of baker et al .
( 2008 ) predict maximal recoil velocities of @xmath193800 kms@xmath0 . in order to reach a high kick velocity ,
the pre - merger bhs of sdssj0927 + 2943 must have been rapidly spinning and of nearly equal mass ; i.e. the galaxy hosting sdssj0927 + 2943 must have undergone a major merger .
assuming that the recoiling smbh carried with it an amount of mass which is not larger than its own mass , and that most of that gas is ultimately available for accretion , we can estimate an upper limit on the duration of the quasar activity , @xmath41 .
we further assume that the gas continues to accrete at its current rate of @xmath42 , where @xmath43 is the eddington luminosity , and we use a radiative efficiency of @xmath44=0.37 which is appropriate for rapidly spinning bhs .
this implies an upper limit on the lifetime of quasar activity of @xmath45 yr . if the space velocity of the recoiling smbh was close to the maximum possible velocity , it would reach a projected separation from the core on the order of a few kiloparsec within @xmath1910@xmath46 yr .
future _ hst _ imaging of sdssj0927 + 2943 will be valuable to distinguish whether the host galaxy shows a distorted morphology as would be expected if the merger was recent .
with its spatial resolution of 0.1@xmath28 , _ hst _ would allow resolution of a spatial scale of @xmath191 kpc in the galaxy and measurement of a corresponding offset of the accreting smbh from the galaxy core .
the x - ray detection of sdssj0927 + 2943 holds promise for a deep _ chandra _ study of smbh and host galaxy .
if the recoiling smbh is a radio emitter , high - resolution radio observations would provide even more precise measurement of its location . in summary , sdssj0927 + 2943
is the best candidate to date for a recoiling supermassive black hole .
its spectrum shows the characteristic signature of two separate emission - line systems , kinematically offset by @xmath192650 kms@xmath0 broad lines from gas bound to the recoiling hole , and narrow lines from gas left behind .
further study of this source and detection of similar ones will provide important information on bh recoil , relevant timescales , and the frequency of bhs removed from their host galaxies , and therefore on simulations of galaxy formation and evolution , and on the question how many bhs grew early by merging . | we present sdssj092712.65 + 294344.0 as the best candidate to date for a recoiling supermassive black hole ( smbh ) .
sdssj0927 + 2943 shows an exceptional optical emission - line spectrum with two sets of emission lines : one set of very narrow emission lines , and a second set of broad balmer and broad high - ionization forbidden lines which are blueshifted by 2650 kms@xmath0 relative to the set of narrow emission lines .
this observation is most naturally explained if the smbh was ejected from the core of the galaxy , carrying with it the broad - line gas while leaving behind the bulk of the narrow - line gas .
we show that the observed properties of sdssj0927 + 2943 are consistent with predictions and expectations from recent numerical relativity simulations which demonstrate that smbhs can receive kicks up to several thousand kms@xmath0 due to anisotropic emission of gravitational waves during the coalescence of a binary .
our detection of a strong candidate for a rapidly recoiling smbh implies that kicks large enough to remove smbhs completely from their host galaxies do occur , with important implications for models of black hole and galaxy assembly at the epoch of structure formation , and for recoil models . |
exclusive @xmath19 decays proceed dominantly via a @xmath20 tree process and can be used to determine @xmath18 , one of the smallest and least known elements of the cabibbo - kobayashi - maskawa matrix @xcite .
however , the need to translate the observed rate to a @xmath18 value using model - dependent decay form - factors ( ff ) has resulted in large theoretical uncertainties .
the recent release of ff results for @xmath17 calculated by unquenched lattice qcd ( lqcd ) @xcite makes possible the first model - independent determination of @xmath18 .
since lqcd results are available only in the high @xmath12 region ( @xmath21gev@xmath16 ) , a clean measurement of the partial @xmath17 branching fraction in the same high @xmath12 region is needed
. there have been several measurements in the past by cleo , babar and belle for the @xmath17 , @xmath22 , @xmath23 and @xmath24 modes @xcite .
the analyses in these measurements utilize the method , originally developed by cleo , where the @xmath6 decays are reconstructed by inferring the undetected neutrino mass from missing energy and momentum ( `` @xmath25-reconstruction method '' ) @xcite . in the @xmath6-factory era
, we will improve the statistical precision by simply applying the @xmath25-reconstruction method using a large amount of data .
however , the poor signal - to - noise ratio will limit the systematic uncertainty of the measurement . in this paper we present measurements of @xmath26 and @xmath27 decays using @xmath28 decay tagging .
we reconstruct the entire decay chain from the @xmath4 , @xmath29 , @xmath30 and @xmath31 with several @xmath32 sub - modes .
the back - to - back correlation of the two @xmath6 mesons in the @xmath4 rest frame allows us to constrain the kinematics of the double semileptonic decay .
the signal is reconstructed in four modes , @xmath26 and @xmath27 .
yields and branching fractions are extracted from a simultaneous fit of the @xmath33 and @xmath34 samples in three intervals of @xmath12 , accounting for cross - feed between modes as well as other backgrounds .
we have applied this method to @xmath35 decays for the first time , and have succeeded in reconstructing these decays with significantly improved signal - to - noise ratios compared to the @xmath25-reconstruction method .
inclusion of charge conjugate decays is implied throughout this paper .
the analysis is based on data recorded with the belle detector at the kekb collider operating at the center - of - mass ( c.m . )
energy for the @xmath4 resonance @xcite .
the @xmath4 dataset that is used corresponds to an integrated luminosity of 253 fb@xmath36 and contains @xmath2 @xmath3 events .
the belle detector is a large - solid - angle magnetic spectrometer that consists of a silicon vertex detector ( svd ) , a 50-layer central drift chamber ( cdc ) , an array of aerogel threshold erenkov counters ( acc ) , a barrel - like arrangement of time - of - flight scintillation counters ( tof ) , and an electromagnetic calorimeter comprised of csi(tl ) crystals ( ecl ) located inside a super - conducting solenoid coil that provides a 1.5 t magnetic field .
an iron flux - return located outside of the coil is instrumented to detect @xmath37 mesons and to identify muons ( klm ) .
the detector is described in detail elsewhere @xcite .
two inner detector configurations were used . a 2.0 cm beam pipe and a 3-layer silicon vertex detector
was used for the first sample of @xmath38 @xmath39 pairs , while a 1.5 cm beam pipe , a 4-layer silicon detector , and a small - cell inner drift chamber were used to record the remaining 123 million @xmath39 pairs @xcite .
a detailed monte carlo ( mc ) simulation , which fully describes the detector geometry and response and is based on geant @xcite , is applied to estimate the signal detection efficiency and to study the background . to examine the ff dependence , mc samples for the @xmath35 signal decays
are generated with different form - factor models : a quark model ( isgw ii @xcite ) , light cone sum rules ( lcsr for @xmath17 @xcite and @xmath40 @xcite ) and quenched lattice qcd ( ukqcd @xcite ) .
we also use unquenched lattice qcd ( fnal @xcite and hpqcd @xcite ) for @xmath17 and a relativistic quark model ( melikhov @xcite ) for @xmath40 . to model the cross - feed from other
@xmath19 decays , mc samples are generated with the isgw ii model for the resonant components ( @xmath17 and @xmath40 components are excluded in this sample ) and the defazio - neubert model @xcite for the non - resonant component . to model the @xmath39 and continuum backgrounds , large generic @xmath39 and @xmath41 monte carlo ( based on evtgen @xcite )
samples are used .
charged particle tracks are reconstructed from hits in the svd and the cdc .
they are required to satisfy track quality cuts based on their impact parameters relative to the measured interaction point ( ip ) of the two beams .
charged kaons are identified by combining information on ionization loss ( @xmath42 ) in the cdc , herenkov light yields in the acc and time - of - flight measured by the tof system . for the nominal requirement , the kaon identification efficiency is approximately @xmath43 and the rate for misidentification of pions as kaons is about @xmath44 .
hadron tracks that are not identified as kaons are treated as pions .
tracks satisfying the lepton identification criteria , as described below , are removed from consideration .
neutral pions are reconstructed using @xmath45 pairs with an invariant mass between 117 and 150mev/@xmath46 .
each @xmath45 is required to have a minimum energy of @xmath47 mev .
@xmath48 mesons are reconstructed using pairs of tracks that are consistent with having a common vertex and that have an invariant mass within @xmath49mev/@xmath46 of the known @xmath48 mass .
electron identification is based on a combination of @xmath42 in the cdc , the response of the acc , shower shape in the ecl and the ratio of energy deposit in the ecl to the momentum measured by the tracking system .
muons are identified by their signals in the klm resistive plate counters , which are interleaved with the iron of the solenoid return yoke .
the lepton identification efficiencies are estimated to be about 90% for both electrons and muons in the momentum region above 1.2gev/@xmath50 , where leptons from prompt @xmath6 decays dominate .
the hadron misidentification rate is measured using reconstructed @xmath51 and found to be less than 0.2% for electrons and 1.5% for muons in the same momentum region . for the reconstruction of @xmath31 ,
the lepton candidate is required to have the correct sign charge with respect to the @xmath52 meson flavor and a laboratory momentum ( @xmath53 ) greater than 1.0gev/@xmath50 .
the @xmath52 meson candidates are reconstructed by using seven decay modes of @xmath54 : @xmath55 , @xmath56 , @xmath57 , @xmath58 , @xmath59 , @xmath60 , @xmath61 ; and ten decay modes of @xmath62 : @xmath63 , @xmath64 , @xmath65 , @xmath66 , @xmath67 , @xmath68 , @xmath69 , @xmath70 , @xmath71 , @xmath72 .
the candidates are required to have an invariant mass @xmath73 within @xmath74 ( @xmath75 is a standard deviation ) of the nominal @xmath52 mass , where the mass resolution @xmath75 is dependent on the decay mode .
@xmath76 mesons are reconstructed in the modes @xmath77 , @xmath78 and @xmath79 by combining a @xmath52 meson candidate and a charged or neutral pion .
each @xmath80 candidate is required to have a mass difference @xmath81 within @xmath82 of the nominal values . for the reconstruction of @xmath83 , the lepton candidate is required to have the right sign charge with respect to the @xmath84 system and @xmath53 greater than 0.8 gev/@xmath50 .
the @xmath84 system may consist of one pion or two pions ( @xmath85 or @xmath86 for a @xmath87 tag and @xmath88 or @xmath89 for a @xmath90 tag ) .
the event is required to have no additional charged tracks or @xmath91 candidates .
we also require that the residual energy from neutral clusters be less than 0.15 gev ( @xmath92gev ) .
the two leptons on the tag and the signal sides are required to have opposite charge .
the loss of signal due to @xmath93 mixing is estimated by mc simulation .
we then impose a constraint based on the kinematics of the double semileptonic decay in the @xmath4 rest frame . in the semileptonic decay on each side , @xmath94
( @xmath95 and @xmath96 ) , the angle between the @xmath97 meson and the detected @xmath98 system @xmath99 is calculated from the relation , @xmath100 ( @xmath101 : 4-momentum vector ) and the known @xmath102 ( the absolute momentum of the mother @xmath6 meson ) .
this means that the @xmath97 direction is constrained on the surface of a cone defined with the angle @xmath103 around the direction of the @xmath98 system , as shown graphically in fig .
[ fig : double_cone ] .
the back - to - back relation of the two @xmath6 meson directions then implies that the real @xmath6 direction is on the intersection of the two cones when one of the @xmath6 systems is spatially inverted .
denoting @xmath104 the angle between the @xmath105 and @xmath106 , the @xmath6 directional vector @xmath107 is given by , @xmath108 , @xmath109 , and @xmath110 with the coordinate definition in fig .
[ fig : double_cone ] , where the @xmath105 and @xmath111 are aligned along the @xmath112-axis and in the @xmath113 plane , respectively .
if the hypothesis of the double semileptonic decay is correct and all the decay products are detected except for the two neutrinos , @xmath114 must range from 0 to 1 .
events passing a rather loose cut @xmath115 are used for signal extraction at a later stage of the analysis .
since the direction of the @xmath6 meson is not uniquely determined , we calculate , @xmath12 as @xmath116 , using the beam energy ( @xmath117 ) , energy ( @xmath118 ) and momentum ( @xmath119 ) of the @xmath84 system and neglecting the momentum of the @xmath6 meson in the c.m . system .
the signal monte carlo simulation finds that the @xmath12 resolution depends on the reconstructed @xmath12 and is in the range 0.32 - 0.95gev@xmath16 .
according to monte carlo simulation , the largest backgrounds originate from @xmath120 and non - signal @xmath19 decays , where some particles escape detection .
there are sizable contributions from cross talk between the @xmath87 and @xmath34 tags .
the contribution from @xmath41 processes is found to be negligible . for events selected
as described above , the signal mc simulation indicates that the total detection efficiency ( @xmath121 ) , averaged over electron and muon channels , is @xmath122 for @xmath123 and @xmath124 for @xmath125 , @xmath126 for @xmath127 and @xmath128 for @xmath129 assuming the lcsr ff model . here , @xmath121 is defined with respect to the number of @xmath3 pairs , where one @xmath6 decays into the signal mode , and includes the loss of signal due to @xmath93 mixing .
because of the loose lepton momentum cut ( @xmath53 @xmath130gev/@xmath50 ) , the variation of efficiency with different ff models is relatively small .
table [ tbl : matrix_sig ] gives the matrix @xmath131 , the efficiency for a signal event generated with true @xmath12 in the bin @xmath132 to be reconstructed in the bin @xmath133 .
table [ tbl : summary_vub ] summarizes the results , where the first and second errors are the experimental statistical and systematic errors , respectively .
the third error is based on the error on @xmath135 quoted by the lqcd authors .
these theoretical errors are asymmetric because we assign them by taking the variation in @xmath18 when @xmath135 is varied by the quoted errors .
the values are in agreement with those from inclusive @xmath136 decays @xcite . to summarize , we have measured the branching fractions of the decays @xmath137 and @xmath138 in @xmath139 @xmath140 events using a method which tags one @xmath6 in the mode @xmath141 .
our results are consistent with previous measurements , and their precision is comparable to that of results from other experiments . the ratios of results for neutral and charged @xmath6 meson modes are found to be consistent with isospin .
the partial rates are measured in three bins of @xmath12 and compared with distributions predicted by several theories . from the rate in the region @xmath142
gev@xmath16 and recent results from lqcd calculations , we extract @xmath18 : @xmath143 the experimental precision on these values is 13% , currently dominated by the statistical error of 11% . by accumulating more integrated
luminosity , a measurement with errors below 10% is feasible . with improvements to unquenched lqcd calculations
, the present method may provide a precise determination of @xmath18 .
we thank the kekb group for the excellent operation of the accelerator , the kek cryogenics group for the efficient operation of the solenoid , and the kek computer group and the national institute of informatics for valuable computing and super - sinet network support .
we acknowledge support from the ministry of education , culture , sports , science , and technology of japan and the japan society for the promotion of science ; the australian research council and the australian department of education , science and training ; the national science foundation of china and the knowledge innovation program of chinese academy of sciencies under contract no .
10575109 and ihep - u-503 ; the department of science and technology of india ; the bk21 program of the ministry of education of korea , and the chep src program and basic research program ( grant no .
r01 - 2005 - 000 - 10089 - 0 ) of the korea science and engineering foundation ; the polish state committee for scientific research under contract no .
2p03b 01324 ; the ministry of science and technology of the russian federation ; the slovenian research agency ; the swiss national science foundation ; the national science council and the ministry of education of taiwan ; and the u.s.department of energy .
h. kakuno _ et al . _
( belle collaboration ) , phys .
lett * 92 * , 071802 ( 2004 ) ; the inclusive branching fraction used in the fitting is based on the partial branching fraction @xmath144gev/@xmath145 gev@xmath16 ) and a calculation of @xmath146 shown in hep - ex/0408115 . | we report measurements of the charmless semileptonic decays @xmath0 and @xmath1 , based on a sample of @xmath2 @xmath3 events collected at the @xmath4 resonance with the belle detector at the kekb @xmath5 asymmetric collider . in this analysis ,
the accompanying @xmath6 meson is reconstructed in the semileptonic mode @xmath7 , enabling detection of the signal modes with high purity .
we measure the branching fractions @xmath8 , @xmath9 , @xmath10 and @xmath11 , where the errors are statistical , experimental systematic , and systematic due to form - factor uncertainties , respectively . for each mode we also present the partial branching fractions in three @xmath12 intervals : @xmath13 , @xmath14 , and @xmath15gev@xmath16 . from our partial branching fractions for @xmath17 and recent results for the form factor from unquenched lattice qcd calculations ,
we obtain values of the ckm matrix element @xmath18 . 2.cm -1.5 cm belle prerpint 2006 - 10 + kek preprint 2006 - 4 , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , and , semileptonic , b decay , exclusive 12.15.hh , 12.38.gc , 13.25.hw |
wire - grid ( wg ) polarizers are widely used at mm - wavelengths as efficient polarization analyzers in polarimeters ( see e.g. @xcite ) , or as beamsplitters in martin - puplett interferometers ( mpi , see @xcite ) .
in fact , an array of metallic wires with diameter and spacing much smaller than the wavelength performs as an almost ideal polarizer at mm wavelengths ( see e.g. @xcite ) , providing an almost ideal beamsplitter for mpis . in the case of astronomical observations ,
low - background operation is required , to observe the faintest sources . in general , fourier transform spectrometers ( fts ) like the mpi can measure a very wide frequency range , often covering up to two decades in frequency .
this is an evident advantage with respect to dispersion and fabry - perot spectrometers , but comes at the cost of a higher radiative background on the detector , which is constantly illuminated by radiation from the whole frequency range covered .
for this reason the best performing instruments are cooled at cryogenic temperatures , and are operated aboard of space carriers to avoid the background and noise produced by the earth atmosphere .
a noticeable example was the cobe - firas fts @xcite , which was cooled at t=1.5k .
firas measured the spectrum of the cosmic microwave background with outstanding precision @xcite , with negligible contributions from instrumental emission .
intermediate performance can be obtained , quickly and cheaply , using stratospheric balloons . in this case
, the residual atmosphere provides a non - negligible background , and the polarimeter / mpi can be operated at room temperature , provided that its background is kept under control .
this means that the emissivity of all the optical elements of the instrument has to be as low as possible , to obtain an instrument - produced background lower than the background produced by the residual atmosphere . in figure [ fig1 ]
we provide a quantitative comparison between photon noise produced by the earth atmosphere ( quantum fluctuations only ) and photon noise produced by low - emissivity metal surfaces ( assuming a dependence on wavelength as @xmath1 , as expected for bulk metal using the hagen - rubens formula @xcite ) . as evident from the figure ,
the constraint on the emissivity of the wire - grid is not very stringent for ground based observations , while it is very stringent for balloon - borne observations at mm wavelengths .
the two dashed lines refer to a metal surface at the same temperature as the atmosphere , with emissivity 0.02 and 0.001 ( top to bottom ) at @xmath2=2 mm .
[ fig1 ] , height=226 ] while measurements of the emissivity of metallic mirrors in this frequency range are readily available , and for clean aluminum or brass surfaces are of the order of 0.3@xmath0 at @xmath2=2 mm ( see e.g. @xcite ) , the emissivity of metallic wire - grids has not been measured systematically .
photon noise is not the only problem induced by emissive optical components .
the average level of background radiation can saturate sensitive detectors , or seriously limit their responsivity .
for this reason in figure [ fig2 ] we provide a quantitative comparison similar to fig.1 but based on the integrated background power , over the range from 15 ghz to the frequency of interest .
the background power values from this figure must be used to compute the background power over the frequency coverage of the instrument , and compared to the saturation power of the detector . in the case of tes bolometers for ground based measurements in the mm atmospheric windows ,
bolometers are designed for a saturation power ranging from @xmath3 1 pw ( balloon - borne experiments , low frequencies ) to @xmath4 10 nw ( ground based experiments , high frequencies ) .
the two dashed lines refer to a metal reflector surface at the same temperature as the atmosphere , with emissivity 0.02 and 0.001 ( top to bottom ) at @xmath2=2 mm .
[ fig2 ] , height=226 ] at variance with metallic mirrors , where the surface can be cleaned , and high conductivity bulky metal pieces can be used , wire grids are built either with thin free - standing tungsten wires or with gold strips evaporated on a thin dielectric substrate . in both cases we do not expect that the effective emissivity is the same as for bulk metal . and we also expect that aging and oxidization can be an important effect , increasing the emissivity of the device with time . from the discussion around figs .
[ fig1 ] and [ fig2 ] and from the considerations above , it is evident that reliable measurements of wire - grid emissivity are in order to properly design sensitive polarimeters and mpis for astronomical use , and decide the operation temperature of the optical components . in this paper
we describe a measurement setup we have developed to measure the effective emissivity of wire grids , at temperatures close to room temperature , at mm - wavelengths .
we discuss the instrument design , report the results of measurements of different wire - grids , and discuss their application in the case of balloon - borne mpis for mm - wave astronomy .
in our approach the emissivity is measured heating the wg and detecting the change of emission .
the wg is inserted in a room - temperature blackbody cavity , with walls covered by eccosorb _
foils , 6 mm thick , so that both transmitted and reflected radiation are carefully controlled .
the radiation emitted , transmitted and reflected by the wg is modulated by a room - temperature chopper with eccosorb - coated blades , and is detected by a 0.3k bolometer .
the wire grid is mounted on a metal ring suspended in the middle of the blackbody cavity by means of kevlar cords , and can be heated by power resistors .
the design of the suspension support for the wg was optimized in order to have a value for the time constant of the heating process not too high . by this way we calculate the thermal capacitance of the wire grid support and
we assume that the two prevalent ways to disperse heat are the conductivity of the kevlar wires and the convection process on the wire grid surfaces .
the best compromise was to design a very low profile alluminium ring with a weight of only 259 grams and 32 links of 38 millimeters lenght 1.5 millimiters in diameter kevlar wire ( see fig.3 ) .
the ring is equipped with pt100 thermometers , biased by a 1 ma current and readout by a 16-bit adc .
temperature stabilization is obtained by means of a digital pid algorithm , which uses the thermometer reading to drive a pwm current through the power resistors .
we are particularly interested to the d band , so our detector bandpass ranges from 125ghz to 175ghz .
the nep of the system under room - temperature background is around @xmath5 .
this detector is the same used for the brain experiment @xcite .
the detector is ac biased at 415hz , while the chopper modulates the optical signal at 20 hz .
two lock - in amplifiers are cascaded for demodulation . a block diagram and a picture of the system
are shown in figs .
[ sys1 ] and [ sys2 ] .
coated chopper ; d : pt crycooler with @xmath6 cryostat ; e : eccosorb@xmath7 coated cavity at room temperature ; f : fresnel lens ; g : diaphragm .
[ sys1 ] , width=302 ] , width=264 ] we measured the emissivity of three different types of wire grid : 1 . 23 cm diameter wire - grid , made of tungsten filaments , 10 @xmath8 diameter , 4 years old ( wg1 ) 2 . 5 cm diameter wire - grid made of tungsten filaments , 10 @xmath8 diameter , 9 months old ( wg2 ) 3 . 5 cm diameter wire - grid made of evaporated gold wired on polipropylene film ( wg3 )
when the chopper is closed the detector receives radiation from the eccosorb - coated chopper blades , at a temperature @xmath9 and with emissivity @xmath10 .
when the chopper is open , the detector recives radiation emitted by the wg , with wires at physical temperature @xmath11 and with emissivity @xmath12 , radiation transmitted by the wg and radiation reflected and/or scattered by the wg .
the latter two come from blackbodies at @xmath13 .
working with thermal radiation in the rayleigh - jeans region , the power emitted is proportional to the physical temperature of the emitter .
so we can write the measured signal , after demodulation , as : @xmath14\ ] ] where @xmath15 is the system responsivity in @xmath16 , while @xmath17 and @xmath18 are the wire - grid trasmissivity and reflectivity .
the product @xmath19 is calibrated removing the wg and placing an eccosorb plate cooled at @xmath20 behind the chopper : in this case , in fact , @xmath21\ ] ] in addition , the eccosorb emissivity can be estimated measuring the transmissivity ( very low ) and the reflectivity ( also very low ) of one of the eccosorb slabs , using a powerful source ( a 150 ghz gunn oscillator ) and the same detector .
we get @xmath22 .
once the responsivity @xmath23 is known , then the ratio @xmath24 can be estimated .
this is a linear function of @xmath25 .
the slope of the line is equal to the wg emissivity @xmath26 , which can be estimated from a best fit of the @xmath24 measured for different @xmath25 .
in figure [ signals ] we plot the signals detected for the three wire grids . detected for different temperatures @xmath27 of the wgs .
the emissivities of the wgs are estimated from the slopes of the best fit lines .
the symbols are diamonds for wg1 , triangles for wg2 , asterisks for wg3 .
[ signals ] , width=283 ] from the slopes of the best fit lines and the calibration described above we estimate the emissivities reported in table[tab : emis ] .
one potential systematic error source is a temperature gradient along the metal wires of the wg .
these wires are heated by the metal support ring at their edges , but are cooled by ambient air at the center .
tungsten is not a very good heat conductor , and evaporated gold wires are very thin . for these reasons we investigated this issue carrying out a finite - elements thermal analysis on a model of the system .
we have used comsol multiphysics @xcite to simulate the transmission of heat in a tungsten filament of different lengths @xmath28 , in thermal contact at both edges with a thermostat at temperature @xmath29 .
the wire is suspended in still air at normal pressure and with temperature @xmath30 , and heat exchange between the wire and air is taken into account . assuming the wire center temperature is @xmath31 at the beginning , its regime center temperature @xmath32 is computed .
the results are given in table [ tab : tempwire ] .
.measured emissivity for the three wire grids described in the text .
[ tab : emis ] [ cols="^,^",options="header " , ] the estimates in table[tab : tempwire ] show that for the smaller diameter wgs ( wg2 and wg3 ) the systematic error is smaller than the statistical error of our measurements . for the large wg ( wg1 )
this systematic error can produce an underestimate of the emissivity of the order fo 10% ( taking into account that the @xmath32 is estimated in the middle of the wire , while the temperatures of the edges are correct ) .
comparing to the noise and background produced by the atmosphere plotted in fig.[fig1 ] and fig.[fig2 ] we see that room - temperature wgs contribute very significantly to the total background on the detector . in the basic design of a mpi three wgs
are needed .
while the background generated from the wg can be neglected in the case of ground based astronomical observations in the mm atmospheric windows , it is the dominant background in the case of balloon - borne observations with a room.temperature interferometer .
for example , in the case of ballon - borne observations of the cmb in the 140 and 220 ghz bands , the different contributions to the total background on the detector are summarized in figs .
[ bkg150 ] and [ bkg220 ] as a function of the band width .
it is evident that the presence of room - temperature wgs dominates the total background on the detectors .
this is specially important in the case of tes detectors , where this additional background challenges the saturation power , limiting in practice the width of the band . ] for the 220 ghz band . here
the limitation to the width of the band is very relevant .
[ bkg220 ] ] in fig .
[ sz ] we plot a simulation of a 3-hours observation of the sunyaev - zeldovich ( sz ) effect in a typical cluster of galaxies , obtained using a room - temperature differential mpi on a stratospheric balloon ( like the olimpo experiment @xcite ) , with photon - noise limited tes detectors working in bands optimized as described above .
see @xcite for details .
the potential of this method is evident , despite of the bandwidth limitation due to the emissivity of the wgs discussed here . and [ bkg220 ] .
the thick line is the best fit to the simulated observations , and the two thin lines represent the two contributing components : sz effect and interstellar dust emission .
we have measured the emissivity of different wire grids at 150 ghz .
the emissivity , of the order of a few % , is larger than the one of metallic mirrors normally used at these frequencies and dominates the emissivity budget of a mpi .
this is relevant in the case of observations carried out with a room - temperature interferometer at balloon altitude , because it dominates the radiative background of the instrument .
this work has been supported by asi ( agenzia spaziale italiana ) grants olimpo , boomerang and millimetron , and by pnra ( italian national antarctic research program ) grant brain .
we warmly acknowledge mr .
giorgio amico for carefully machining many parts of the instrument .
p. de bernardis , s. colafrancesco , g. d alessandro , et al .
, low - resolution spectroscopy of the sunyaev - zeldovich effect and estimates of cluster parameters , astronomy and astrophysics , 538 , a86 ( 2012 ) | we have measured , using a custom setup , the emissivity of metallic wire - grids , suitable for polarimeters and interferometers at mm and far infrared wavelengths .
we find that the effective emissivity of these devices is of the order of a few @xmath0 , depending on fabrication technology and aging .
we discuss their use in astronomical instruments , with special attention to martin puplett interferometers in low - background applications , like astronomical observations of the cosmic microwave background .
polarizers , interferometers , cosmic microwave background |
recently , high - t@xmath3 superconductivity with anomalous magnetic properties was reported in the weak - ferromagnetic ru-1212 system rusr@xmath0rcu@xmath0o@xmath1 ( r = sm , eu , gd , y ) with the tetragonal tlba@xmath0cacu@xmath0o@xmath16-type structure.@xcite for the ca - substituted system , a possible superconductivity was also reported in the weak - ferromagnetic compounds ruca@xmath0rcu@xmath0o@xmath1 ( r = pr - gd).@xcite the metallic weak - ferromagnetic ( wfm ) order is originated from the long range order of ru moments in the ruo@xmath17 octahedra due to strong ru-4d@xmath18-o-2p@xmath19 hybridization in this strongly - correlated electron system .
the curie temperature t@xmath20 130 k observed from magnetization measurement in the prototype compound rusr@xmath0gdcu@xmath0o@xmath1 is probably a canted g - type antiferromagnetic order with ru@xmath21 moment @xmath22 canted along the tetragonal basal plane resulting a small net spontaneous magnetic moment @xmath23 @xmath24(ru@xmath21 ) too small to be detected in neutron diffraction.@xcite the occurrence of high - t@xmath3 superconductivity with maximum resistivity onset t@xmath3(onset ) @xmath10 60 k in rusr@xmath0gdcu@xmath0o@xmath1 is related with the quasi - two - dimensional cuo@xmath0 bi - layers separated by a rare earth layer in the ru-1212 structure.@xcite broad resistivity transition width @xmath25t@xmath3 = t@xmath3(onset ) - t@xmath3(zero ) @xmath10 15 - 20 k observed is most likely originated from the coexistence and the interplay between superconductivity and weak - ferromagnetic order.@xcite the diamagnetic t@xmath3 is observed anomalously at lower temperature near t@xmath3(zero ) instead of at t@xmath3(onset ) , and a reasonable large meissner signal was reported using stationary sample magnetometer with diamagnetic t@xmath26 30 k in @xmath27 1
g applied field at zero - field - cooled ( zfc ) mode.@xcite lower t@xmath3(onset ) @xmath10 40 k and 12 k were observed for rusr@xmath0eucu@xmath0o@xmath1 and rusr@xmath0smcu@xmath0o@xmath1 , respectively.@xcite no superconductivity can be detected in rusr@xmath0rcu@xmath0o@xmath1 ( r = pr , nd).@xcite superconducting rusr@xmath0ycu@xmath0o@xmath1 phase is stable only under the high pressure.@xcite the physics is still unclear in this system , and it will be interesting to investigate the effect of the weak - ferromagnetic order on superconducting critical fields b@xmath13 and b@xmath5 , as well as on the possible existence of a spontaneous vortex state ( svs ) at higher temperature above the meissner state .
the stoichiometric rusr@xmath0gdcu@xmath0o@xmath1 samples were synthesized by the standard solid - state reaction method .
high - purity ruo@xmath0 ( 99.99 @xmath28 ) , srco@xmath29 ( 99.99 @xmath28 ) , gd@xmath0o@xmath29 ( 99.99 @xmath28 ) , and cuo ( 99.99 @xmath28 ) preheated powders with the nominal composition ratio of ru : sr : gd : cu = 1 : 2 : 1 : 2 were well mixed and calcined at 960@xmath30c in air for 16 hours .
the calcined powders were then pressed into pellets and sintered in flowing n@xmath0 gas at 1015@xmath30c for 10 hours to form rusr@xmath0gdo@xmath17 and cu@xmath0o precursors .
this step is crucial in order to avoid the formation of unwanted impurity phases .
the n@xmath0-sintered pellets were heated at 1060@xmath30c in flowing o@xmath0 gas for 10 hours to form the ru-1212 phase .
the pellets were oxygen - annealed at slightly higher 1065@xmath30c for 5 days and slowly furnace - cooled to room temperature with a rate of 15@xmath30c per hour.@xcite the powder x - ray diffraction data were collected with a rigaku rotaflex 18-kw rotating anode diffractometer using graphite monochromatized cu - k@xmath31 radiation with a scanning step of 0.02@xmath30 ( 10 second counting time per step ) in the 2 @xmath32 ranges of 5 - 100@xmath30 . the electrical resistivity and magneto - resistivity measurements were performed using the standard four - probe method with a linear research lr-700 ac ( 16hz ) resistance bridge from 2 k to 300 k with applied magnetic field up to 7 t. the magnetization , magnetic susceptibility and magnetic hysteresis measurements from 2 k to 300 k with applied fields from 1 g to 7 t were carried out with a quantum design 1-t @xmath22-metal shielded mpms2 or a 7-t mpms superconducting quantum interference device ( squid ) magnetometer .
electrical resistivity @xmath33(t ) and volume magnetic susceptibility @xmath34(t ) at 1-g field - cooled ( fc ) and zero - field - cooled ( zfc ) modes for oxygen - annealed rusr@xmath0gdcu@xmath0o@xmath1 ]
the powder x - ray diffraction pattern for the oxygen - annealed rusr@xmath0gdcu@xmath0o@xmath1 polycrystalline sample indicates close to single phase with the tetragonal lattice parameters of @xmath36 = 0.5428(5 ) nm and @xmath37 = 1.1589(9 ) nm .
the space group p4/mbm is used for rietveld refinement analysis , where neutron diffraction data indicate that a ruo@xmath17 octahedra 14@xmath30 rotation around the c - axis is needed to accommodate physically reasonable ru - o bond lengths.@xcite the refinement gives a good residual error r of 3.64 @xmath28 , weighted pattern error r@xmath38 = 6.07 @xmath28 , and bragg error r@xmath39 = 5.05 @xmath28 .
the temperature dependence of the electrical resistivity @xmath33(t ) and the volume magnetic susceptibility @xmath34(t ) at 1-g field - cooled ( fc ) and zero - field - cooled ( zfc ) modes for rusr@xmath0gdcu@xmath0o@xmath1 are shown collectively in fig .
the high temperature resistivity decreases monotonically from room temperature value of 9.2 [email protected] ( not shown ) to 6.4 [email protected] at 200 k , and extrapolated to 2.8 [email protected] at 0 k with a good resistivity ratio @xmath33(300 k)/@xmath33(0 k ) of 3.3 for the polycrystalline sample .
high temperature resistivity shows a non - fermi - liquid - like linear t - dependence down to curie temperature t@xmath2 of 131 k , then changes to a t@xmath41 behavior below t@xmath2 due to magnetic order .
the superconducting onset temperature of 56 k is determined from the deviation from t@xmath41 behavior , with a zero resistivity t@xmath3(zero ) at 39 k. the broad transition width @xmath25t@xmath3 = 17 k observed is the common feature for all reported ru-1212 resistivity data , which indicates that the superconducting josephson coupling along the tetragonal c - axis between cu - o bi - layers may be partially blocked by the dipole field b@xmath42 of ordered ru moments in the ru - o layer.@xcite the diamagnetic t@xmath3 at 39 k was observed in the 1-g zfc susceptibility measurement .
the full meissner shielding signal 4@xmath43 = 4@xmath44m / b@xmath45 -1.5 ( gaussian units ) was recorded at 5 k. this value is identical to the meissner shielding signal expected for a superconducting sphere with a demagnetization factor n of -4@xmath44/3 and in an applied field b@xmath46 well below lower critical field b@xmath5 . the large diamagnetic signal in 1-g zfc mode is the best data observed so far from various reported susceptibility measurement techniques.@xcite since our measurements were performed with the standard moving - sample squid magnetometer , it is clear that sample quality is more crucial than measuring techniques .
both zfc and fc data reveal a curie temperature t@xmath2 of 131 k. however , in 1-g fc mode , no diamagnetic field - expulsion signal can be detected below 39 k due to strong flux pinning where superconductivity coexists with weak - ferromagnetic order .
zfc volume susceptibility @xmath34(t ) for rusr@xmath0gdcu@xmath0 o@xmath1 at 1 g , 10 g , and 100 g. note that the full meissner shielding signal was observed only at low applied field and low temperature . ]
the zero - field - cooled ( zfc ) volume susceptibility @xmath34(t ) at 1 g , 10 g , and 100 g applied fields are shown collectively in fig .
all data show the same magnetic order t@xmath2(ru ) of 131 k. although the diamagnetic t@xmath3 of 39 k was still observed at 10-g zfc measurement , the diamagnetic signal at 5 k is reduced to 60@xmath28 of the full meissner signal .
consider the polycrystalline nature of sample with varying microcrystallite size and orientation , the average superconducting lower critical field b@xmath5 at 5 k is estimated to be close to 10 g. no net diamagnetic signal can be detected at 100-g zfc mode where the sample is already in the vortex glass or lattice state and the small diamagnetic signal is overshadowed by a large weak - ferromagnetic background.@xcite low field superconducting hysteresis loops m - b@xmath46 for rusr@xmath0gdcu@xmath0o@xmath1 : ( a ) at 5 k , 10 k , 15 k , 20 k ; ( b ) at 25 k , 30 k , and 35 k. ] based on this information , the low - field ( @xmath420 g ) isothermal superconducting hysteresis loops m - b@xmath46 are measured and collectively shown in fig .
3(a ) ( 5 k , 10 k , 15 k , and 20 k ) and 3(b ) ( 25 k , 30 k , and 35 k ) .
the initial magnetization curve deviates from straight line in 4 g at 5 k , 3.5 g at 10 k , 3 g at 15 k , 2 g at 20 k , and 1 g at 25 k. this is the narrow region that full meissner signals are detected and is roughly corresponding to the anisotropic lower critical field in the ab - plane b@xmath9(t ) with b@xmath9(0 ) @xmath10 4 g. the average lower critical field b@xmath8 for polycrystalline sample is determined from the peaks of initial diamagnetic magnetization curves .
b@xmath5 decreases steadily from 12 g at 5 k , 11 g at 10 k , 9 g at 15 k , 6 g at 20 k , 3 g at 25 k and below 1 g at 30 k. a simple empirical parabolic fitting gives b@xmath5(t ) = b@xmath5(0)[1 - ( t / t@xmath6)@xmath7 , with average b@xmath8(0 ) = 12 g and t@xmath6 = 30 k ( see fig .
4 ) . using the anisotropic ginzburg - landau formula b@xmath8 = [ 2b@xmath9 + b@xmath47/3 ,
c - axis b@xmath48 28 g and anisotropy parameter @xmath11 7 is estimated .
this value is close to reported anisotropic @xmath49-value for yba@xmath0cu@xmath29o@xmath16 where the 123-type structure can be written as cu-1212 cuba@xmath0ycu@xmath0o@xmath16 .
an average penetration depth @xmath50(0 ) = [ @xmath51/2@xmath44b@xmath8(0)]@xmath52 of 520 nm was derived with estimated @xmath53(0 ) = 340 nm and @xmath54(0 ) = 2400 nm from b@xmath55 = @xmath51/2@xmath56 and b@xmath9 = @xmath51/2@xmath57 , where @xmath51 is flux quantum . the lower field , low temperature superconducting phase diagram b@xmath46(t ) of rusr@xmath0gdcu@xmath0o@xmath1 .
] since t@xmath6 = 30 k is well below t@xmath3(onset ) = 56 k and t@xmath3(zero ) = 39 k in zero applied field , a spontaneous vortex state ( svs ) indeed exists between 30 k and 56 k. the low field phase diagram b@xmath46(t ) for polycrystalline sample is shown in fig .
4 , with the average b@xmath5(t ) separates the meissner state from the vortex state and a smaller b@xmath9(t ) inside the meissner region for reference .
t@xmath3(zero ) = 39 k in the broad resistive transition is the onset of vortex depinning by driving current .
this temperature is very close to the melting transition temperature t@xmath14 from the spontaneous vortex glass or lattice state to the spontaneous liquid state due to nonzero dipole field b@xmath42 of weak - ferromagnetic order .
the upper critical field b@xmath13 defined from t@xmath3(onset ) and the vortex melting field b@xmath14(t ) defined from t@xmath3(zero ) are temperature independent for small applied fields below 20 g. the internal dipole field generated by a weak - ferromagnetic order can be estimated using a simple extrapolation [ b@xmath5(0)+ b@xmath58/b@xmath5(0 ) = t@xmath3/t@xmath6 = 56 k/30 k , which results with a dipole field b@xmath42 @xmath10 10.4
g on the cuo@xmath0 bi - layers .
a small net spontaneous magnetic moment @xmath23 of @xmath10 0.11 @xmath12 per ru is estimated using b@xmath59 2@xmath23/d@xmath60 with d = @xmath37/2 = 0.58 nm is the distance between midpoint of cuo@xmath0 bi - layers and two nearest - neighbor ru moments .
if the weak - ferromagnetic structure is a canted g - type antiferromagnetic order with ru moments @xmath22 (= 1.5 @xmath12 for ru@xmath21 in t@xmath61 states ) canted along the tetragonal basal plane , the small net spontaneous magnetic moment gives a canting angle of 4@xmath30 from the tetragonal c - axis and is difficult to be detected in neutron diffraction with a resolution around 0.1 @xmath12.@xcite at 5 k , the shape of superconducting hysteresis loop with a large remanent molar magnetization m@xmath62 of 83 g.cm@xmath60/mol indicates strong pinning as well as a good indication of bulk nature of superconductivity for the oxygen - annealed sample .
the remanent m@xmath62 decreases to 4 g.cm@xmath60/mol at 30 k and 1 g.cm@xmath60/mol at 35 k , where a weak - ferromagnetic background can be clearly seen .
fluctuation in the hysteresis loop is probably also related to the weak - ferromagnetic order .
temperature dependence of magnetoresistivity @xmath33(t , b@xmath46 ) for rusr@xmath0gdcu@xmath0o@xmath1 in applied field up to 7 t. ] to study the high - field effect on superconductivity , the magnetoresistivity @xmath33(t , b@xmath46 ) for rusr@xmath0gdcu@xmath0o@xmath1 up to 7 t are collectively shown in fig .
5 . the broadening of resistive transition in magnetic fields is the common features for all high - t@xmath3 cuprate superconductors.@xcite the normal state resistivity is field independent and follows a t@xmath41-dependence below t@xmath2 , with superconducting t@xmath3(onset ) of 56 k in zero field decreases slightly to 53 k in 7-t field .
the temperature dependence of upper critical field b@xmath13(t ) can be fitted with a linear function b@xmath13(0)[1 - t / t@xmath3 ] with average b@xmath13(0 ) = 133 t.@xcite an average coherence length @xmath65 = [ @xmath51/2@xmath44b@xmath66(0)]@xmath52 of 0.5 nm with the ginzburg - landau parameter @xmath67 of 1040 and the thermodynamic critical field b@xmath3(0 ) = ( [email protected]@xmath13)@xmath52 = 0.32 t. no anisotropic @xmath68 and @xmath69 values can be estimated from present data .
the t@xmath3(zero ) decreases from 39 k in zero applied field to 32 k in 1-kg , 28 k in 5-kg , 25 k in 1-t , 22 k in 2-t , 19 k in 3-t , 17 k in 4-t , 16 k in 5-t , 15 k in 6-t , and 14 k in 7-t field . if the zero resistivity is taken as the lower bound of the vortex melting temperature t@xmath14 , then the temperature dependence of the vortex melting transition line b@xmath14(t )
can be fitted roughly by the formula b@xmath14(t ) = b@xmath14(0)[1 - t / t@xmath14]@xmath15 with b@xmath14(0 ) = 35 t and large exponent 3.5 . in the lower field region , b@xmath14(t )
rises as [ 1 - t / t@xmath14]@xmath41 as predicted by the mean - field approximation for temperature near t@xmath14 = 39 k.@xcite the full phase diagram b@xmath46(t ) of rusr@xmath0gdcu@xmath0o@xmath1 is shown in fig . 6 to exhibit both high field and low field features .
the very broad vortex liquid region with @xmath25 t = 17 k in zero field and @xmath25 t = 42 k in 7-t field is extraordinary and is most likely originated from the coexistence and the interplay between superconductivity and weak - ferromagnetic order .
this magnetic order is so weak that superconductivity can coexist with the magnetic order , but the effect of a weak spontaneous magnetic moment @xmath71 0.1 @xmath12 is detected through the appearance of a spontaneous vortex state above 30 k with a broad spontaneous vortex liquid region above t@xmath14 of 39 k. to study the broad vortex liquid region , the isothermal field - dependent magnetoresistivity @xmath33(b@xmath46 ) for t @xmath72 t@xmath3 are shown in fig . 7 , where zero resistivity gives a lower bound of vortex melting field b@xmath14 .
in the resistive vortex liquid region , there is no unique power - law fit , but the data can be roughly fitted with various powers between 1/6 and 1 , depending on temperature range . at t
= 40 k @xmath73 t@xmath14 = 39 k
, resistivity is @xmath10 b@xmath74 for b@xmath75 2 t and gradually changes to @xmath10 b@xmath76 below 1 t. at 30 k with b@xmath14 = 0.2 t , resistivity @xmath10 b@xmath77 for b@xmath75 2 t and gradually changes to linear b@xmath46 below 1 t. at 20 k with b@xmath14 @xmath10 2.4 t , similar linear variation of resistivity with b@xmath46 was observed for b@xmath75 3.5 t. full phase diagram b@xmath46(t ) of rusr@xmath0gdcu@xmath0o@xmath1 . ]
field dependence of magnetoresistivity @xmath33(b@xmath46 ) for rusr@xmath0gdcu@xmath0o@xmath1 in the vortex state at 20 k , 30 k , and 40 k. ] the last issue to be addressed is the depression of t@xmath3 by small spontaneous ru magnetic moments .
the weak - ferromagnetic order is actually a canted antiferromagnetic order that can coexist with superconductivity .
however , the observed t@xmath3 of 56 k is too low as compared with 93 k for yba@xmath0cu@xmath29o@xmath16 or 103 k for tlba@xmath0cacu@xmath0o@xmath16 .
the depression of t@xmath3 by small spontaneous magnetic moment can be partially recovered by substitution of nonmagnetic cu ions at ru site .
for example , in the ru@xmath78cu@xmath79sr@xmath0gdcu@xmath0o@xmath1 system , t@xmath3 onset up to 65 k for x = 0.1 and 72 k for x = 0.4 was reported.@xcite
the lower critical field with b@xmath5(0 ) = 12 g and t@xmath6 = 30 k indicates the existence of a spontaneous vortex state ( svs ) between 30 k and t@xmath3 of 56 k. this svs state is closely related with the weak - ferromagnetic order with a net spontaneous magnetic moment of @xmath10 0.1 @xmath12 per ru .
the broad vortex liquid region observed above vortex melting line b@xmath14(t ) is also due to the coexistence and the interplay between superconductivity and weak - ferromagnetic order .
p. w. klamut , b. dabrowski , s. m. mini , m .
maxwell , j. mais , i. felner , u. asaf , f. ritter , a. shengelaya , r. khasanov , i .
m. savic , h. keller , a. wisniewski , r. puzniak , i. m. fita , c. sulkowski , and m. matusiak , physica c * 387 * , 33 ( 2003 ) . | a spontaneous vortex state ( svs ) between 30 k and 56 k was observed for the weak - ferromagnetic superconductor rusr@xmath0gdcu@xmath0o@xmath1 with ferromagnetic curie temperature t@xmath2 = 131 k and superconducting transition temperature t@xmath3 = 56 k. the low field ( @xmath420 g ) superconducting hysteresis loop indicates a narrow meissner state region within average lower critical field b@xmath5(t)= b@xmath5(0)[1 - ( t / t@xmath6)@xmath7 , with average b@xmath8(0 ) = 12
g and t@xmath6 = 30 k. full meissner shielding signal in very low applied field indicates an ab - plane b@xmath9(0 ) @xmath10 4 g with an estimated anisotropic parameter @xmath11 7 for this layered system .
the existence of a spontaneous vortex state between 30 k and 56 k is the result of weak - ferromagnetic order with a net spontaneous magnetic moment of @xmath10 0.1 @xmath12/ru , which generates a weak magnetic dipole field around 10 g in the cuo@xmath0 bi - layers .
the upper critical field b@xmath13 varies linearly as ( 1 - t / t@xmath3 ) up to 7-t field .
the vortex melting line b@xmath14 varies as ( 1 - t / t@xmath14)@xmath15 with melting transition temperature t@xmath14 = 39 k and a very broad vortex liquid region due to the coexistence and the interplay between superconductivity and weak - ferromagnetic order . |
collisionless shocks driven by intense laser pulses attract growing attention from both laboratory astrophysics @xcite and plasma - based accelerator communities @xcite . in astrophysics , collisionless shocks
are observed in supernova remnants , gamma ray bursts , active galactic nucleus jets and so on @xcite .
particularly , collisionless shocks are relevant to the generation of extremely high energy particles , which is one of the most fundamental problems in astrophysics @xcite .
however , the direct study of collisionless shocks in astronomical environments is passive and suffers from the lack of details .
in contrast , high power laser - matter interaction can create astrophysical - like conditions and thus provides convenient test - beds for many astrophysical processes including the collisionless shock formation @xcite . with regard to plasma - based accelerator ,
ion accelerations by laser - driven collisionless electrostatic shocks have been observed in a great number of simulations @xcite and experiments @xcite . in particular ,
monoenergetic ion acceleration by collisionless shock has been proposed by using a near critical density target with a slowly decreasing density profile at the rear side @xcite . with this target configuration ,
the electric field relevant to the target normal sheath acceleration becomes smaller and nearly constant .
recently , quasi - monoenergetic ion beams have been obtained in the experiments using infrared co2 laser pulses and gas - jet targets @xcite .
the gas - jet targets have the advantage of operating at a high - repetition rate , which is a crucial requirement in many practical applications such as cancer radiotherapy @xcite . in principle , collisionless electrostatic shocks can originate from the temperature or / and density jumps of plasmas @xcite .
the kinetic theory of collisionless electrostatic shock formation in the interaction of plasma slabs with different temperatures and densities was presented by sorasio _ _ et al.__@xcite , and its relativistic generalization and ion - reflection condition were given by stockem _
_ et al.__@xcite . for a rigorous electrostatic shock
, the ion - reflection condition can be written as @xcite @xmath0 where @xmath1 is the ionic charge state , @xmath2 is the peak of electrostatic potential , @xmath3 is the ion kinetic energy in the shock frame , and for initially cold ions @xmath4 are equal to the shock speed @xmath5 . in the interaction of an intense laser pulse with a highly overdense target ,
the shock speed is mainly determined by the laser hole - boring ( or piston ) velocity @xmath6 @xcite . while laser - driven collisionless shock has been extensively studied using highly overdense targets @xcite , the laser - plasma interactions are more complicated in the near critical density regime @xcite and the collisionless shocks excited therein are rather unexplored @xcite . in this paper
, we study the formation of collisionless electrostatic shock and the subsequent ion acceleration in the interaction of intense laser pulses with near critical density plasmas .
a substantial number of electrons in such a plasma can be pushed forward due to the laser pondermotive force and form an electron density peak ahead of the laser front .
simultaneously , a strong electrostatic field is generated due to the charge separation .
consequently , some ions may be pulled by this charge separation field and be accumulated ahead of the laser front . meanwhile , the plasma will be greatly heated because of its strong coupling with laser .
therefore , a high - temperature high - density plasma flow can be produced ahead of the laser front .
when such a laser - driven plasma flow interacts with the background plasma , a collisionless electrostatic shock can be excited .
particle - in - cell simulations show that the speed of this collisionless shock can be very high since the laser pulse propagates much faster in a near critical density plasma than in a highly overdense plasma .
most important , unexpected ion reflections by this collisionless shock is clearly observed even if the conventional ion - reflection condition eq.([ref - condition ] ) are not satisfied .
to study the laser - driven collisionless shock generation in the near critical density regime , 1d3v particle - in - cell ( pic ) simulations have been performed with the code osiris @xcite .
the simulation box locates in @xmath7 and consists of 48,000 cells , and 600 macroparticles per cell are allocated in the plasma region , where @xmath8 is the laser wavelength . in the simulations ,
initially cold and uniform targets are positioned in @xmath9 .
the targets are assumed to be fully ionized , and the ions are protons . for convenience ,
we assume that the laser pulse is linearly polarized ( lp ) in the @xmath10 direction and propagates along the positive @xmath11 direction . for reference ,
the laser pulse is assumed to arrive at the target surface @xmath12 at @xmath13 , and all simulations begin at @xmath14 and end at @xmath15 , where @xmath16 is the wave period of the laser .
the intensity of the laser pulse increases to a constant value @xmath17 after a quick rising in the first five wave cycles , where @xmath18 is the normalized amplitude of laser electric field .
the target electron density @xmath19 varies from @xmath20 to @xmath21 in different simulation cases , where @xmath22 is the critical density and @xmath23 is the angular frequency of the laser .
we have found that the dynamics of laser - driven collisionless shock is very distinct in different laser - plasma parameter regimes . ,
( b ) ion density @xmath24 , and ( c ) longitudinal electric field @xmath25 in the interaction of an intense laser pulse of @xmath26 with a near critical density target of @xmath27 .
the black dashed lines in ( a ) and ( b ) mark the laser front , while the red dashed line and white dashed line in ( b ) marks the shock front .
the white , black , and red lines in ( c ) indicate the orbits of remaining , trapped , and shock - reflected ions , respectively .
, scaledwidth=50.0% ] we now analyze a representative simulation with an intense laser intensity @xmath26 and a near critical target density @xmath27 .
the relativistic - induced penetration of the laser pulse into this near critical density target is evidenced by the combination of figs . [ figrt2d](a ) and [ figrt2d](b ) , where the time evolutions of laser electric field @xmath28 and ion density @xmath24 are displayed , respectively .
the laser front is marked by the black dashed - line in fig .
[ figrt2d](a ) , from which we can divide the laser propagation roughly into two stages . in the first stage ( @xmath29 ) , the forward velocity of the laser front @xmath30 decreases gradually from @xmath31 to @xmath32 . in the second stage ( @xmath33 ) ,
the forward velocity of the laser front becomes nearly constant @xmath34 .
moreover , the pronounced trapping and reflections of ions occur in the second stage as shown in fig .
[ figrt2d](b ) . and ( b ) electrostatic potential @xmath35 at some different moments of @xmath36 , 56 , 153.2 , and 154@xmath37 in the interaction of an intense laser pulse of @xmath26 with a near critical density target of @xmath27.,scaledwidth=50.0% ] to understand the distinct plasma dynamics in these two stages
, we display the time evolution of longitudinal electric field @xmath25 in fig .
[ figrt2d](c ) . while the profiles of @xmath25 and the corresponding electrostatic potential @xmath38 at some representative moments are shown in fig .
[ figexphi](a ) and [ figexphi](b ) , respectively .
in addition , the electron and ion density profiles at some representative moments are shown in fig . [ figneni](a ) and fig.[figneni](b ) , respectively .
and ( b ) ion density @xmath24 at different moments of @xmath13 , 40 , 80 , and 120@xmath37 in the interaction of an intense laser pulse of @xmath26 with a near critical density target of @xmath27.,scaledwidth=50.0% ] since the initial laser - plasma conditions satisfy @xmath39 , the relativistic transparency(rt ) rather than the hole - boring(hb ) dominates the laser propagation at the beginning of the first stage @xcite . as a result ,
the initial forward velocity of the laser front @xmath40 is much higher than the hole - boring velocity @xmath41 .
meantime , the laser pondermotive force will push a substantial number of electrons forward as shown in fig . [
figneni](a ) , resulting in a strong charge - separation field around the laser front . fig .
[ figexphi](a ) shows that the amplitude of this charge - separation field can be as high as @xmath42 at a representative moment @xmath43 . despite of this strong electric field ,
no obvious ion reflection occurs at this moment as shown fig .
[ figneni](b ) , and the white line in fig .
[ figrt2d](c ) shows a typical ion orbit .
this is because at this moment the laser front propagates at a speed as fast as @xmath31 , so does the charge - separation field . as a result , the ion kinetic energy in the co - moving frame of the charge - separation field is about @xmath44 , which is much larger than the peak of electrostatic potential @xmath45 at this moment as shown in fig .
[ figexphi](b ) .
so no obvious ion reflection occurs at this moment .
however , the forward velocity of the laser front will gradually decrease due to the accumulation of particles ahead of the laser front and the dissipation of laser pulse in a dense plasma @xcite . as the ponderomotive force of a linearly polarized pulse
is time oscillating , it can not compress a tenuous plasma immediately if @xmath46 .
however , some electrons will accumulate continually ahead of the laser front and form an electron density peak , thus a strong unipolar electrostatic field is formed .
after that some ions can accumulate ahead of the laser front too and form an ion density peak due to this strong unipolar electrostatic field . because the thermal velocity of the electrons are much faster than that of the ions , the ion density peak can be much narrower and higher than the electron density peak .
so the morphology of the charge separation field changes from unipolar to bipolar in this process .
it results in the further accumulation of the ions ahead of the laser front . on the other hand
, the accumulation of the particles also results in the broadening of the charge - separation layer , as indicated by the blue line in fig.[figexphi ] . due to the broadening of the charge - separation layer , the peak of the electrostatic potential @xmath2 increases to about @xmath47 at this moment even if the amplitude of the charge - separation field is nearly same to the previous one . for @xmath48
, we can estimate that it will be larger than the ion kinetic energy @xmath49 in the co - moving frame of the laser front if @xmath50 . because the forward velocity of laser front @xmath30 will decrease to about @xmath32 at the end of this stage , the obvious
trapping and reflections of ions will begin to occur at a turning point in this stage . in return
, the trapping and reflections of ions will dissipate the laser pulse and slow it down further . in the second stage
, however , it seems that a balance is roughly built between the propagation and the dissipation of the laser pulse as a nearly constant forward velocity of the laser front @xmath34 is presented .
since the hole - boring rather than the relativistic transparency dominates the laser propagation in this stage @xcite , the forward velocity of the laser front is roughly equal to @xmath51 . here , the amplitude of the local laser field @xmath52 is smaller than the amplitude of the laser field in the vacuum due to the dissipation in the plasma .
more importantly , we find that in the second stage the ion reflections may take place in two different scenarios . in the first scenario , a great number of ions can be reflected at the laser - plasma interface as in the first stage .
these reflected ions will form an ion beam that propagates faster than the laser pulse . during the propagation of this ion beam in the background plasma
, a collisionless electrostatic shock can be stimulated with a strong longitudinal electric field at the beam - plasma interface as shown in fig.[figrt2d](c ) . at a representative moment
@xmath53 , the peak of the electrostatic potential is about @xmath54 as shown in fig . [ figexphi](b ) .
this potential is even larger than the kinetic energy of background ions in the co - moving frame of the beam - plasma interface .
therefore , in the second scenario , some ions in the background plasma can be reflected by the collisionless electrostatic shock that excited by the ion beam generated in the first scenario .
it is worthwhile to point out that the mean velocity of the ions reflected in the second scenario ( @xmath55 ) is about two times of that in the first scenario ( @xmath56 ) .
the mean velocity of the ions reflected in the first scenario is roughly twice the forward velocity of the laser front ( @xmath57 ) . while the mean velocity of the ions reflected in the second scenario is roughly twice the velocity of the collisionless electrostatic shock .
as indicated by the slop of red dashed - line in fig.[figrt2d](b ) , this shock velocity is about @xmath58 , which is roughly equal to the propagation velocity of the ion beam that consists of the ions reflected in the first scenario . in the interaction of an intense laser pulse of @xmath26 with a near critical density target of @xmath27.,scaledwidth=50.0% ] besides the ions being accelerated to higher energies , we notice that the longitudinal electric field is strongly time - oscillating in the interaction of an intense laser pulses with a near critical density plasma . although the time oscillation in the amplitude of the longitudinal electric field has been reported previously in the solitary or collisionless shock acceleration @xcite , here we find that not only the amplitude but also the profile shape of the longitudinal electric field oscillate violently is a near critical density plasma .
[ figexphi](a ) shows that the field amplitude nearly doubles from @xmath53 to @xmath59 , while the location of the field peak even steps backward .
consequently , the peak of the electrostatic potential also oscillates violently . as shown in fig .
[ figexphi](b ) , its value at @xmath60 is even less than half of that at @xmath53 .
this high - frequent nonlinear oscillation in the peak of the electrostatic potential is fully evidenced by fig .
[ figphimax2d ] , which also illuminates that the conventional ion - reflection condition @xmath61 can not always be satisfied after the shock formation ( @xmath33 ) .
$ ] and ( b ) the time integrals of these fields @xmath62 dt$ ] ( i.e. , the longitudinal momenta @xmath63 ) of the shock - reflected ( red lines ) and shock - trapped ( black lines ) ions , respectively , in the interaction of an intense laser pulse of @xmath26 with a near critical density target of @xmath27 .
the ion momenta obtained from the simulation are shown as the separate squares at some representative moments .
the orbits of the reflected and trapped ions are drawn in fig.[figrt2d](c ) as the red and black lines , respectively . , scaledwidth=50.0% ] however , we notice that the ion reflections by this collisionless shock are still possible even if the peak of the electrostatic potential @xmath2 is smaller than the ion kinetic energy @xmath64 in the shock frame .
the underlying physics is that the longitudinal electric field here is a time - oscillating electrostatic field .
therefore , it does not need to satisfy the conventional ion - reflection condition in eulerian coordinates .
the general ion - reflection condition in a time - oscillating electric field can be obtained from the momentum conservation in lagrangian coordinates as @xmath65 dt \geq m_i v_{shock } , \label{new - condition}\ ] ] where @xmath63 is the longitudinal ion momentum , and @xmath66 is the coordinate of the @xmath67-th ion at time @xmath68 . from the above inequality , we can see that the instantaneous electric field felt by the ion determines whether or not this ion is reflected . the red line in fig .
[ figrt2d](c ) illustrates an ion reflection by the shock at around @xmath69 , despite that the peak of the electrostatic potential @xmath2 is much smaller than the ion kinetic energy @xmath64 during its reflection in fig .
[ figphimax2d ] .
meanwhile , a lot of ions in the upstream region can not be reflected at the shock front , but they penetrate the shock and are trapped between the shock and the laser front . the black line in fig .
[ figrt2d](c ) shows the orbit of such a trapped ion , which lag behind the shock front at latter time .
the accumulation of these trapped ions can result in the broadening of the charge - separation layer , which may contribute to the increase of the average value of the electrostatic potential peak @xmath2 at around @xmath70 .
the time evolutions of the localized longitudinal electric fields felt by the above mentioned shock - reflected and trapped ions are compared in fig . [ figexpx](a ) .
the integrations of these two fields over the time result in the momenta of these two ions , which are displayed in fig . [ figexpx](b ) . as shown in fig .
[ figexpx](a ) , the trapped ion experiences an alternatively positive and negative longitudinal electric field .
therefore , each of its acceleration periods is followed by a deceleration period to some extent in fig . [
figexpx](b ) . as a result
, it can not be accelerated to very high energy and finally be trapped in the downstream . while the shock - reflected ion nearly always experiences a positive longitudinal electric field in fig .
[ figexpx](a ) , hence it can be greatly accelerated . as shown in fig .
[ figexpx](b ) , for the reflected ion the integration of the electric filed results in a momentum larger than @xmath71 , and vice versa for the trapped ion .
( blue triangle - line ) and the shock front @xmath5 ( red star - line ) as functions of the normalized amplitude of the incident laser intensity .
the target parameters are same to those in fig .
[ figrt2d].,scaledwidth=50.0% ] due to the strong dissipation of the laser pulse in a near critical density target , the real - time local laser intensity when the collisionless shock is excited will be smaller than the initial incident laser intensity .
therefore , it may be interesting to investigate the effect of the laser intensity on the collisionless shock generation in a near critical density plasma .
in particular , one may be curious about what will happen if we initially employ a laser pulse at the real - time local laser intensity .
for instance , the real - time local laser intensity is about @xmath72 when the collisionless shock is excited in the interaction of an intense laser pulse of @xmath26 with a near critical density target of @xmath27 . using the initial laser intensity @xmath73
, we find no qualitative difference in the shock generation except for a slightly lower shock velocity . in fig .
[ figvelocity ] , the forward velocities of the laser front @xmath30 ( blue triangle - line ) and the shock front @xmath5 ( red star - line ) are shown as functions of the normalized amplitude of the incident laser intensity .
one can see that the forward velocity of the laser front increases monotonically with the increasing laser intensity , and the velocity of the excited shock is roughly two times of the forward velocity of the laser front in a broad laser intensity regime ( @xmath74 ) .
it illuminates that in a near critical density plasma the collisionless shock generation is not sensitive to the laser intensity . if the laser intensity @xmath75 ; however , we find no obvious evidence of the shock generation in our simulations .
this is because an ultra - intense laser pulse will propagate too fast in a near critical density plasma due to the strong relativistic transparency @xcite , so there s no enough time for the electrons and ions to accumulate sufficiently ahead of the laser front .
as a result , the laser pulse can propagate very fast during the whole laser - plasma interaction , and the ion reflections become very difficult in this case .
on the other hand , if the initial laser intensity is too low , the forward velocity of the laser front will be not so high in comparison with the plasma thermal velocity .
so it also becomes difficult to excite a fast shock in the background plasma if @xmath76 as shown in fig .
[ figvelocity ] . for the shock formation and ( b ) the shock speed @xmath5 as functions of the initial target density .
the laser parameters are same to those in fig .
[ figrt2d].,scaledwidth=50.0% ] it is important to notice that a considerable long response time is required before the laser pulse to excite the collisionless shock in the relativistically transparent plasmas .
[ figvsts](a ) shows that the response time for the shock formation decreases monotonously with the increasing initial target density . in the target with a density @xmath77 ,
the laser propagation is dominated by the relativistic transparency at the beginning .
therefore , at first there is only the accumulation of ions by the charge separation field , and the accumulation is too rarefied to excite the shock in the background plasma . with the accumulation of the ions
ahead of the laser front , however , the laser pulse and the charge separation field will be slowed down and capable of accumulating more ions .
finally , the flow of accumulated ions will become dense enough to excite a collisionless shock in the background plasma .
the lower the initial target density is , the longer response time it takes for the laser pulse to excite the collisionless shock . for targets with densities @xmath78 , collisionless shock formation hasnt been observed in our pic simulation up to @xmath79 .
in contrast , the collisionless shock can be quickly excited in a few laser wave cycles in the regime @xmath80 dominated by the hb . taking into account the rising time of the laser pulse
, the collisionless shock can be considered to be excited immediately in this regime . due to the transition of laser propagation mechanism from the relativistic transparency to the hole - boring
, we also notice that the shock speed increases slightly ( but abnormally ) with the increasing target density in the region @xmath81 in fig . [
figvsts](b ) . in the case of @xmath27 ,
the laser pulse at first can penetrate into the target via the relativistic transparency , but such a penetration is highly dissipative @xcite .
therefore , the laser pulse has been greatly weakened in the target . while the hole - boring dominates the laser pulse propagation nearly from the beginning in the case of @xmath82 ,
so the laser pulse can not penetrate deep into the target and becomes less dissipated . as a result , the speed of the excited shock in the case of @xmath82 is faster than that in the case of @xmath27 . in the interaction of an intense laser pulse with ( a)(b ) a relativistic transparent target of @xmath27 and ( c)(d ) a solid target of @xmath83 .
the downstream and upstream regions of the shock structures are roughly separated by the blue dashed lines .
the laser parameters are same to those in fig .
[ figrt2d].,scaledwidth=50.0% ] we notice that the shock speed is as high as @xmath84 with the density @xmath27 .
this high shock speed may be attributed to the combination of the fast laser propagation and the highly heated plasma .
the former provides a fast piston to excite the shock , while the latter results in a high ion acoustic speed . as shown in fig .
[ fignete](a ) , the plasma electron temperature @xmath85 in the upstream region is about @xmath86 for the relativistic transparency case @xmath27 .
compared with this high electron temperature , the ion temperature @xmath87 in the upstream region is negligible .
consequently , the ion acoustic speed can be estimated as @xmath88 . from the ion acoustic speed and the shock speed
, we can estimate that the mach number is about @xmath89 .
in contrast , we find that the electron temperature in the upstream region is only about @xmath90 in a solid target case with @xmath83 .
so the corresponding ion acoustic speed is only about @xmath91 . considering the speed of the shock @xmath92 excited in this case
, we can get the mach number @xmath93 .
although the mach numbers of the shocks in the relativistically transparent and solid target cases are very close to each other , the shock speeds are considerably different due to the different ion acoustic speeds . substituting the mach number @xmath94 into the jump conditions for an ideal shock @xcite , we can get the density and the pressure ratios between the downstream and the upstream regions as @xmath95 where the mass density @xmath96 , the pressure @xmath97 , the subscript zero ( one ) denotes the upstream ( downstream ) variables , and @xmath98 is used with the degree of freedom @xmath99 for a 1d fully ionized plasma . as shown in figs . [
fignete](a ) and [ fignete](b ) , the electron ( ion ) number densities in the upstream and downstream regions are about @xmath100 ( @xmath101 ) and @xmath102 ( @xmath103 ) , respectively . and
the electron ( ion ) temperatures in the upstream and downstream regions are about @xmath104 ( @xmath105 ) and @xmath106 ( @xmath107 ) , respectively . from these simulation results
, we can estimate the density ratio @xmath108 and the pressure ratio @xmath109 , which roughly agree with the prediction values given by eqs .
( [ densityratio ] ) and eqs .
( [ pressureratio ] ) .
furthermore , we find that the densities , pressures , and internal energies of the upstream and the downstream obtained from the simulation roughly satisfy the following rankine - hugoniot relations @xcite @xmath110 where the shock velocity @xmath111 , the piston velocity @xmath112 , and the internal energy per unit mass @xmath113 . phase space and ( b ) ion energy spectrum at @xmath15 in the interaction of an intense laser pulse with a relativistic transparent target ; and ( c ) and ( d ) are the corresponding ion distribution and energy spectrum in the case with a solid target .
the laser and target parameters are same to those in fig .
[ fignete ] .
, scaledwidth=50.0% ] as an important consequence of the high shock speed , the ions can be accelerated to very high energies in the laser interaction with a rt plasma . as shown in figs .
[ figphase](a ) and [ figphase](b ) , the center momentum and energy of shock - accelerated ions in the rt ( @xmath27 ) case are about @xmath114 and @xmath115 mev , respectively .
while in the higher overdense target ( @xmath83 ) case they are only about @xmath116 and @xmath117 mev , respectively .
furthermore , the yield of energetic ions can also be greatly enhanced in the rt case . as indicated in figs .
[ figphase](b ) and [ figphase](d ) , the number of energetic ions with @xmath118 mev in the case of @xmath27 is about 4 times larger than the number of energetic ions with @xmath119 mev in the case of @xmath83 .
laser - driven collisionless shock formation and the consequent ion acceleration have been identified in the near critical density plasmas .
comparing with a highly overdense target , an intense laser pulse can propagate much faster in a near critical density plasma , so does the laser - driven collisionless shock .
consequently , a part of ions can be reflected and accelerated to very high energies .
however , in a near critical density plasma it may take a considerable long response time before the generation of the collisionless shock .
more important , we find that the longitudinal electric field related to the collisionless shock formation is strongly time - oscillating not only in the amplitude but also in the profile shape in this case .
with such a strongly time - oscillating electric field , the ions can be reflected even if the peak of the electrostatic potential is smaller than the ion kinetic energy in the shock frame . in the above content
, we use the dimensionless variables for the laser and target parameters . for the employed laser intensity @xmath17=10 ( @xmath120 w/@xmath121m@xmath122 )
, it has already been achieved in the experiments using nd : yag or ti : sapphire laser pulses @xcite . in order to excite a collisionless electrostatic shock in a near critical density plasma , however , the duration of the laser pulse should also be longer than a few hundred femtoseconds ( about 200 hundred wave periods ) .
therefore , the relevant experiments might be conducted only in the picosecond petawatt laser facilities , such as lfex @xcite . concerning the near critical density targets ,
they might be achieved with tunable density profiles by the advanced micro / nano - fabrication technologies @xcite .
for instance , targets with the density range of 1 - 1000 mg/@xmath123 can be fabricated by the pulsed laser deposition technique @xcite . with the idea of the layer - by - layer nanoarchitectonics ,
nanotube films have been recently structured on the nanometer scale to provide near critical density targets @xcite .
alternatively , the near critical density targets might be also realized by the combination of highly compressed gas targets and intense infrared laser pulses @xcite .
the analyses in this paper are based on one - dimensional pic simulations .
previous studies showed no qualitative difference between one- and multi - dimensional simulations of laser - driven collisionless shock formation in highly overdense targets @xcite .
however , in the near critical density regime it may take a longer response time before the shock formation in multi - dimensional simulations , since the laser transverse ponderomotive force will push particles out radially and slow down the accumulation of particles ahead the laser front .
on the other hand , radial shocks might be excited by the laser transverse ponderomotive force too @xcite .
this work was supported in part by the national basic research program of china ( grant no .
2013cba01504 ) and the national natural science foundation of china ( grant nos . 11675108 , 11405108 , 11421064 , 11129503 , 11374210 and 11375262 ) . smw and mc appreciate the supports from national 1000 youth talent project of china .
zms acknowledges the support of a leverhulme trust research project grant at the university of strathclyde .
simulations have been carried out on the pi cluster of shanghai jiao tong university .
y. kuramitsu , y. sakawa , t. morita , t. ide , k. nishio , h. tanji , h. aoki , s. dono , c. d. gregory , j. n. waugh , n. woolsey , a. dizire , a. pelka , a. ravasio , b. loupias , m. koenig , s. a. pikuz , y. t. li , y. zhang , x. liu , j. y. zhong , j. zhang , g. gregori , n. nakanii , k. kondo , y. mori , e. miura , r. kodama , y. kitagawa , k. mima , k. a. tanaka , h. azechi , t. moritaka , y. matsumoto , t. sano , a. mizuta , n. ohnishi , m. hoshino and h. takabe , plasma phys .
fusion * 54 * , 124049 ( 2012 ) .
x. liu , y. t. li , y. zhang , j. y. zhong , w. d. zheng , q. l. dong , m. chen , g. zhao , y. sakawa , t. morita , t. kuramitsu , t. n. kato , l. m. chen , x. lu , j. l. ma , w. m. wang , z. m. sheng , h. takabe , y. rhee , y. k. ding , s. e. jiang , s. y. liu , j. q. zhu and j. zhang , new j. phys . * 13 * , 093001 ( 2011 ) . d. w. yuan , y. t. li , x. liu , y. zhang , j. y. zhong , w. d. zheng , d. l. dong , m. chen , y. sakawa , t. morita , y. kuramitsu , t. n. kato , h. takabe , y. rhee , j. q. zhu , g. zhao and j. zhang , high energy density phys . * 9 * , 239 ( 2013 )
. h. ahmed , m. e. dieckmann , l. romagnani , d. doria , g. sarri , m. cerchez , e. ianni , i. kourakis , a. l. giesecke , m. notley , r. prasad , k. quinn , o. willi , and m. borghesi , phys .
lett . * 110 * , 205001(2013 ) .
m. s. wei , s. p. d. mangles , z. najmudin , b.walton , a. gopal , m. tatarakis , a. e. dangor , e. l. clark , r.g .
evans , s. fritzler , r. j. clarke , c. hernandez - gomez , d. neely , w. mori , m. tzoufras , and k. krushelnick , phys .
* 93 * , 155003 ( 2004 ) .
c. a. j. palmer , n. p. dover , i. pogorelsky , m. babzien , g. i. dudnikova , m. ispiriyan , m. n. polyanskiy , j. schreiber , p. shkolnikov , v. yakimenko , and z. najmudin , phys .
lett . * 106 * , 014801 ( 2011 ) .
k. zeil , m. baumann , e. beyreuther , t. burris - mog , t. e. cowan , w. enghardt , l. karsch , s. d. kraft , l. laschinsky , j. metzkes , d. naumburger , m. oppelt , c. richter , r. sauerbrey , m. schrer , u. schramm and j. pawelke , appl .
b * 110 * , 437 ( 2013 ) .
n. p. dover , c. a. j. palmer , m. j. v. streeter , h. ahmed , b. albertazzi , m. borghesi , d.c.carroll , j. fuchs , r. heathcote , p. hilz , k. f. kakolee , s. kar , r. kodama , a. kon , d. a. maclellan , p. mckenna , s. r. nagel , d. neely , m. m. notley , m. nakatsutsumi , r. prasad , g. scott , m tampo , m. zepf , j. schreiber and z. najmudin , new j. phys . * 18 * , 013038 ( 2016 ) . j. h. bin , w. j. ma , h. y. wang , m. j. v. streeter , c. kreuzer , d. kiefer , m. yeung , s. cousens , p. s. foster , b. dromey , x. q. yan , r. ramis , j. meyer - ter - vehn , m. zepf , and j. schreiber , phys .
* 115 * , 064801 ( 2015 ) . | laser - driven collisonless electrostatic shock formation and the subsequent ion acceleration have been studied in near critical density plasmas .
particle - in - cell simulations show that both the speed of laser - driven collisionless electrostatic shock and the energies of shock - accelerated ions can be greatly enhanced due to fast laser propagation in near critical density plasmas .
however , a response time longer than tens of laser wave cycles is required before the shock formation in a near critical density plasma , in contrast to the quick shock formation in a highly overdense target .
more important , we find that some ions can be reflected by the collisionless shock even if the electrostatic potential jump across the shock is smaller than the ion kinetic energy in the shock frame , which seems against the conventional ion - reflection condition .
these anomalous ion reflections are attributed to the strongly time - oscillating electric field accompanying laser - driven collisionless shock in a near critical density plasma . |
galaxies are gregarious systems , most of them are gathered in groups or clusters , while only 30% are isolated and 10% are binaries in the field .
nevertheless compact groups ( cg ) are quite rare and according to hickson s classification ( hickson , 1982 ) only 0.1 % of galaxies belong to cgs .
criteria of population ( initially four galaxies in the group ) , isolation ( dynamically independent systems ) and compactness ( separation between galaxies comparable to the sizes of the galaxies ) are chosen by hickson to build his catalog . with these criteria around one hundred cgs were found on the palomar observatory sky survey red prints .
compact groups are ideal sites to study the influence of strong dynamical evolution due to environment on molecular cloud formation and star formation efficiency .
they appear in projection as the densest galaxy systems known , even denser than the cores of rich clusters , and they show multiple signs of interactions . due to their high density , and relatively small velocity dispersion , these systems are unstable with regard to merging instability .
the dynamical friction time - scale is of the order of 2.10@xmath0 yrs , and n - body simulations predict their rapid evolution towards a single elliptical massive galaxy ( e.g. barnes 1989 ) .
the existence of many such compact groups is therefore a puzzle , and the physical reality of hcg has been questioned ( e.g. mamon 1986 , 1987 ) ; but evidence of galaxy - galaxy interactions in those groups , either morphologic ( hickson 1990 ; mendes de oliveira 1992 ) , or kinematic ( rubin et al . 1991 ) , speaks in favour of their reality .
latest spectroscopic observations showed that 92 of the original 100 groups have at least three galaxies with recession velocities within 1000 of each other ( hickson et al .
the presence of hot intergalactic gas , detected by x - ray emission centered on some hcgs , is a further confirmation of the existence of these compact groups ( pildis et al .
1995 , ponman et al .
+ most of galaxies that belong to groups are in fact in loose groups of 10 - 30 galaxies and about 50% of all galaxies belong to loose groups .
but loose groups are in their great majority un - bound and un - virialised ( gourgoulhon et al .
1992 ) while their true dynamical state is ambiguous ( expanding , collapsing , transient ) . clusters of galaxies are more near equilibrium , specially in their centers ( about 10% of all galaxies belong to clusters )
. however , the depth of their potential well leads to high relative velocities between galaxies that reduce the efficiency of galaxy - galaxy encounters .
the influence of environment is revealed by the high proportion of ellipticals and lenticulars , and by the hi gas deficiency of spirals ( dressler 1984 , cayatte et al .
this gas deficiency can be explained by ram - pressure as well as tidal interactions ( combes et al .
no molecular gas deficiency has been detected , either in virgo ( kenney & young 1988 ) , or in coma ( casoli et al . 1991 ) , which suggests that the inner parts of the galaxies are not affected by their environment , since the co emission essentially comes from the galaxy central regions .
however , there could be two compensating effects at play here : the enhancement of co emission in interacting galaxies ( cf braine et combes 1993 , combes et al .
1994 ) , and the outer gas stripping , stopping the gas fueling of galaxies .
+ in compact groups , some hi deficiency has also been reported ( williams & rood 1987 ) , but no co emission deficiency , according to a first study by boselli et al ( 1996 ) with the sest telescope .
it is further interesting to investigate whether hcgs are actually sampling the highest densities of galaxies in the universe .
it has been claimed that , even if the cgs are real , we are not sure of their high density , since they could correspond to loose groups with only a high _ projected _ density through chance alignment of filaments along the line of sight ( e.g. mamon 1992 ) .
but no loose groups are observed in hcg neighborhood in the majority ( 67% ) of cases ( rood & williams 1989 ) .
hickson ( 1982 ) found that the groups contain fewer spirals than a comparable sample of field galaxies .
the spiral fraction decreases from 60% in the least compact groups to 20% in the most compact .
there is also a deficiency of faint galaxies with respect to rich clusters and field .
this apparent deficiency is more severe in groups with elliptical first - ranked galaxies .
radio properties of compact groups have been studied by menon & hickson ( 1985 ) and menon ( 1991 , 1995 ) .
although the far - infrared and radio luminosities are still highly correlated as for field galaxies , the total radio emission from hcg spirals is relatively lower by a factor 2 in compact group galaxies while the nuclear radio emission is enhanced by a factor of about 10 compared to isolated galaxies .
the results suggest a scenario in which interactions among group galaxies produce inflow of gas towards the centers , elevating the star formation there , and consequently the radio and far - infrared emissions .
but at the same time the removal of gas and magnetic fields from the extended disks of the galaxies results in a decrease of total radio - emission .
williams & rood ( 1987 ) have observed 51 of the 100 hickson groups in the hi line , and detected 34 of them .
they find that on average a hickson compact group contains half as much neutral hydrogen as a loose group with a similar distribution of galaxy luminosities and morphological types .
this result supports the reality of compact groups as independent dynamical systems and not transient or projected configurations in loose groups .
the recent rosat survey of hcgs by ponman et al ( 1996 ) also confirms that the groups are intrinsically compact , and not the projection of loose groups .
they infer that more than 75% of the hcgs possess hot intragroup gas . + we present here a large co survey of hickson group galaxies with the iram 30 m telescope , and compare the relative gas content and star formation efficiency of cg galaxies with other samples belonging to widely different environments .
after describing the observations in section 2 , the sample and the data in section 3 , we discuss the main conclusions and possible interpretations in section 4 .
observations were carried out on 4 - 10 september 1995 with the 30 meter radiotelescope of the instituto de radio astronomia milimetrica ( iram ) in pico veleta , near granada in spain .
single - sideband sis receivers were tuned for the and transitions at respectively 115 and 230 ghz .
weather conditions were excellent during the run with typical effective system temperatures of 300 - 400 k ( scale ) at 115 ghz and 500 - 800 k ( scale ) at 230 ghz .
+ for each line a 512 @xmath11 mhz channel filter bank is used with a velocity resolution of 2.6 smoothed for each spectrum to 10.4 or 20.8 according the quality of each spectrum . at 115 ghz and 230 ghz
we assume a hpbw of 22@xmath2 and 11@xmath3 respectively .
pointing was done frequently on continuum sources with corrections of the offsets up to 8@xmath3 , providing an accuracy of 3@xmath4 ( greve et al , 1996 ) .
+ the temperature - scale calibration was checked on the sources w3oh , oria and irc+10216 ( mauersberger et al .
1989 ) ; except for a transient problem for the 3 mm calibration , it remained in a reasonable range providing at least a 20% calibration accuracy .
we use the wobbler with a switch cycle of 4 seconds and a beam throw of 90 - 240@xmath5 avoiding off position on another galaxy of the same compact group .
each 12 minutes a chopper wheel calibration was performed on a load at ambient temperature and on a cold load ( 77 k ) .
the line temperatures are expressed in the scale , antenna temperature corrected for atmospheric attenuation and rear sidelobes .
baselines were flat allowing us to subtract only linear polynomials out of the spectra .
our observed sample is composed of 70 galaxies towards 45 compact groups , taken from the catalog of hickson compact groups ( hcg , hickson 1982 ) .
we discarded afterwards 4 galaxies ( 11a,19b,73a,78a ) which appear not to belong to compact groups ( hickson , 1992 ) .
the galaxies , mostly spirals , are selected for their radio continuum ( menon 1995 ) and iras ( hickson et al .
1989 ) detections .
all the targets are northern sources with @xmath6 .
the redshift range from 1200 up to 18500 ( 27b ) .
if we use h@xmath7=75 for the hubble constant , as adopted in this paper , the mean distance of the sample is 95 mpc with a standard deviation of 45 mpc .
we present in fig .
[ stat_sample ] the statistical distribution of our sample for distance , type , far infrared ( ) luminosity and median projected separation .
+ in table [ tab_sample_1 ] we display the main properties of the sample .
the column headings are the following : name is taken from hickson classification , type is taken from hickson et al . ( 1989 ) , d is the distance computed with a correction for the galactic rotation using a solar galactic velocity of rotation of 250 , r is the median projected galaxy - galaxy separation , @xmath8 indicates the diameter in arcsecs at @xmath9 = 24.5 mag.arcsec@xmath10 , from hickson et al .
( 1989 ) , blue luminosity l@xmath11 is computed as follows = 12.208 - 0.4b@xmath12+log(1+z)+2log(d / mpc ) , luminosity is computed using reprocessed iras data from allam et al .
( 1996 ) and following hickson et al .
( 1989 ) derivation , is a dust temperature indicator using an emissivity dependence as @xmath13 in the 60 - 100 @xmath14 m iras- range , s@xmath15 is the total radio continuum flux density from menon ( 1995 ) , neutral hydrogen content has been found mainly from williams & rood ( 1987 ) , and some from huchtmeier & richter ( 1989 ) . concerning luminosity or
hi mass , we indicate in table [ tab_sample_1 ] the whole group emission preceded by @xmath16 , meaning that poor spatial resolution does not allow a separation per galaxy .
+ [ cols="^,<,^,^,^,^,^,^,^,^,^,^ " , ] we check the completeness of the / distribution function ( top left of fig .
[ df_mass ] ) by simulating the distribution function taking into account the threshold of detection for that quantity .
the minimum temperature detection is 4 mk in our observations .
then we consider the distance distribution for the galaxies to be uniform up to 150 mpc or gaussian with the parameters of our sample .
the linewidth distribution has been fitted to the blue luminosity with a power law ( tully - fisher - like relation with @xmath17 ) .
then inclination angle is distributed uniformly between 0 and 90 degrees .
the blue luminosity is distributed with a gaussian distribution ( @xmath18 , @xmath19 ) . for each @xmath20 bin ,
the fraction of realisations above the threshold detection is computed to estimate the completeness of our sample .
results are displayed in fig .
[ plot_completude ] for @xmath21 realisations .
gaussian and uniform distance distributions are two extremes chosen to estimate the weight of the distance parameter : the 50 % level of completeness is @xmath22 for the uniform case and -1.5 for the gaussian case . assuming that our control sample is complete , we compute a cumulative distribution biased by the completeness function of fig . [ plot_completude ] .
[ plot_cf_simul ] shows the result where it appears that the malmquist bias in our sample , spread over a large distance range , can explain part of the apparent enhancement of molecular gas in the hggs .
however the control sample , assumed to be complete , is not likely to be so for the low values of @xmath23 where the samples are the most different .
similarly the @xmath24 distribution is affected by the malmquist bias ( verter , 1993 ) , but the point is that the cumulative distribution is lower than the control distribution on the whole range of variation up to higher values , leading to a suggestion of a real enhancement of the dust material in the hcgs .
the close correlation between dust and molecular gas suggests that the content is really enhanced in the hcgs . _ as it has been shown previously , that enhancement is only significant for the most compact groups in their merging phase , confusing somewhat the question of molecular gas enhancement in the whole hcg sample .
_ what can also be seen in fig .
[ fir - h2-sfe - hi](a ) is the large dispersion of the sfe as defined by the / ratio : a large number of galaxies are deficient in co emission , leading to a depressed / ratio and large sfe .
this large dispersion is mainly due to the dwarf and elliptical samples , but also to small galaxies in the control sample ; we have checked that the objects with high sfe at moderate / have a lower than average .
this phenomenon disappears when the total gas content is considered instead of the mere h@xmath25 content , as shown in fig [ fir - h2-sfe - hi](b ) and ( d ) .
it is striking that the total normalised gas content is almost a constant , independent of /.
there is one exception for the cluster population where the stripping of the neutral and molecular content is at play .
we find a deficiency of the molecular content in these galaxies , which seems related to a lower luminosity ( horellou et al .
all that suggests the importance of in star formation as a source of fueling , through the conversion @xmath26 , and the higher reliability of the star formation indicator /(+ ) .
a least square fit yields the relation @xmath27 for all the samples gathered .
the good correlation between the normalised and could be in a large part due to the dependence of both quantities on the metallicity and temperature of the interstellar medium .
it is now well established that the co to h@xmath25 conversion ratio is strongly dependent on metallicity ( e.g. rubio et al 1993 ) , as well as the dust - to - gas ratio ( and therefore the fir luminosity ) . as shown on fig .
[ df_ratio ] , all the cumulative distributions for the molecular gas - to - dust ratio are highly correlated , (: all , except dwarf and elliptic samples , @xmath28 6 .
( probability @xmath29 0.54 ) ) . while the dependence of on temperature is direct , that of the co emission is more complex
. low brightness temperatures of the co lines are obtained either for cold gas , or diffuse gas ; the rotational levels of the molecule are excited by collision , and diffuse molecular clouds are generally sub - thermal .
the use of a standard co to h@xmath25 conversion ratio is then problematic . for thermalised dense gas
however , the co emission is directly proportional to the gas temperature . the metal abundance and the gas recycling
are closely related to the imf and evolution of the mf .
high mass stars ( m@xmath30 ) are active on short time scales ( @xmath31 yrs ) whereas low mass stars have an influence on much longer time scales ( vigroux et al .
thus the co abundance is very dependent on the star formation rate , and the co/ conversion ratio could be very variable , particularly for galaxies with recent star formation episodes ( e.g. casoli et al 1992 , henkel & mauersberger 1993 ) . given the enhancement in co emission detected in hcg with respect to a control sample and the strong correlation between the radio continuum , co and the fir for hcg spirals it would appear that the lack of enhancement of the total fir emission in hcg galaxies is due to lack of spatial resolution of iras measurements .
this is particularly important if the fir is mainly enhanced in the central regions of the galaxies .
iso observations might allow to check this assertion .
the consequence of an enhanced without fir enhancement is a lower star forming efficiency for hcg , as displayed in fig .
[ df_fir ] .
this property might appear surprising , but disappears when the total gas content is taken into account ( this result should be taken with caution , since only 14 galaxies have a well - defined content in our hcg sample ) .
[ td ] we have tested the correlations of the star formation indicators , with and without account of the atomic phase , with the dust temperature , in fig . [ survey_td ] .
both quantities correlate well with : for all samples together we find a relationship flatter than young et al .
( 1989 ) , i.e. /@xmath32 @xmath33 , but with differences and large dispersions among categories : for the hcg sample alone /@xmath32 @xmath34 and for the starburst sample /@xmath32 @xmath35 . sage ( 1993 ) pointed out that a single dust temperature is an `` average '' of the cold dust and warm dust associated respectively with the quiescent molecular clouds and the clouds with massive star forming ones .
devereux & young ( 1990 ) emphasized that high mass ( @xmath29 6 ) o and b stars are responsible for high ( @xmath36 ) and h@xmath37 luminosities , advocating a two - component model : one dust component heated by high mass stars ( @xmath38 50 - 60 k ) and the other heated by the interstellar radiation field ( @xmath38 16 - 20 k ) .
the low mean temperature of the hcg sample ( @xmath39 33 k ) suggests that the luminosity is coming from an important quiescent molecular phase , together with the `` cirrus '' phase associated with diffuse atomic hydrogen .
it could explain why the _ global _ luminosity does not fit with a high star formation population , without excluding star formation towards the center , as revealed by radio continuum .
in a recent study , lisenfeld et al .
( 1996 ) find the same /@xmath40 ratio for starburst interacting and normal galaxies . from this
, we could infer that the enhanced radio continuum emission from the center of hcg galaxies should be accompanied with an enhanced central fir luminosity .
the constancy of the /@xmath40 ratio has been interpreted as being due to a strong and fast ( @xmath41 yrs ) increase of the magnetic field at the beginning of the starburst , together with a time - scale of variation of the star - formation rate longer than some @xmath41 yrs .
the enhancement of the content in the most compact groups suggests that tidal interactions in hcg are efficient in driving the gas inwards .
this is also a confirmation that at least some groups are actually compact and not only projections along the line of sight .
these very compact groups should merge through dynamical friction on a short time scale ( a few @xmath42 yrs , cf barnes 1989 ) .
a conclusion from the present work is that the most compact of these groups have concentrated an important amount of molecular gas without initiating yet important star formation .
however they must correspond to a short duration phase just before merging and enhanced star - formation .
it is tempting to identify the next phase of this process to the ultra luminous infra red galaxies ( ulirgs ) .
the latter have their infrared luminosity powered by massive star formation ( lutz et al .
1996 ) with an important consumption of molecular gas .
sanders et al ( 1988 ) have shown that many ulirgs are interacting / merging galaxies ; recently clements & baker ( 1996 ) extended that to the vast majority of the ulirgs sample .
some of these systems , with luminous masses ranging up to few @xmath43 , should represent the remnant of some compact group which has undertaken multiple mergers on a very short time - scale ( @xmath44 yrs ) .
the fir luminosity of the final ulirgs should be a function of the spiral fraction in the parent group , the most powerful ulirgs being the result of the merging of typically four gas - rich spirals .
as pointed out earlier the spatial distribution of interstellar matter within galaxies will play an important role in the rate of fuelling of starbursts during interactions .
but does the expected rate of ulirgs formation via cg merging match the presently observed frequency of these objects ?
the answer is very uncertain , since the actual fraction of very compact groups , on the point of merging , is not known .
this fraction can not be close to 1 , since there would be too large a discrepancy between the expected and observed number of ultra luminous galaxies and their remnants ( e.g. williams & rood 1987 , sulentic & rabaa 1994 ) .
simulations of galaxy formation and large - scale structures evolution have suggested that some of the cgs could be filaments of galaxies seen end - on ( hernquist et al .
this idea has been studied further by pildis et al .
( 1996 ) : if it is true that galaxies that will form a compact group spend a large fraction of their time first in a filament , this filament will appear as a cg in projection only for less than 20% of cases .
then the galaxies will fall into a real cg , and this phase corresponds to at least 30% of their lifetime . although these figures are model dependent , they suggest that the majority of hcg in the sky are physically compact groups
. the time - scale of merging can then depend highly on initial conditions , and in particular on the elliptical fraction , which may alleviate the over - merging problem ( e.g. governato et al .
1991 , garcia - gomez et al .
1996 ) .
the threshold in mean galaxy separation for the enhancement of suggests that it corresponds to the last stage of the life of the compact group , when each galaxy is undertaking frequent and strong tidal effects . in fig .
[ histo_sepa ] we plot the histogram of the mean projected separation of the whole sample of hcgs from hickson et al .
( 1992 ) : there is a significant cut - off at short separation at approximatively 20 - 30 kpc , which must correspond to a short life time for very compact configurations .
we suggest that this cut - off corresponds to the acceleration of the merging process , and at the same time to significant inward gas flows , that account for our observations .
dynamically , such an exponential acceleration of the collapse is predicted by simulations and analytic models of satellite decay through dynamical friction ( see for example leeuwin & combes 1997 ) .
this rapid acceleration has been interpreted through the excitation of numerous high - order resonances for a satellite at about twice the primary radius ( e.g. tremaine & weinberg 1984 ) .
from our survey of co emission in 70 galaxies belonging to 45 hickson compact groups , we have detected 57 objects .
we find in average that the gas and dust contents / and / show evidence of enhancement with respect to our control sample .
this result however is somewhat weakened due to the malmquist bias in our sample .
for the most compact groups the enhancement is more clear . on the contrary
, the global far - infrared flux does not appear to be enhanced with respect to the control sample .
the fir and distribution indicates that the fir luminosity is coming essentially from a cold dust component heated by the interstellar radiation field . from the general correlation between fir and radio contiuum power we suggest that only the very centers of some groups are experiencing star formation and are sites of enhanced fir emission . iras spatial resolution is not sufficient to show this directly .
statistical tests show that the hcg gas and dust contents are closer to that of pair and starburst galaxies , revealing the efficiency of tidal interactions in driving the gas inwards in compact group galaxies .
we find a stronger enhancement for the cgs having a short mean separation ( @xmath45 kpc ) .
we suggest that these most compact , high- content groups , may be in a final merging phase , just before the starburst phase , that will lead them in a very short time - scale to the ulirgs category .
+ the comparison of the various samples suggest that the _ total _ gas content ( + ) should be taken into account to estimate the star formation efficiency
. the corresponding sfe indicator , / ( + ) , should be more reliable , and allow us to avoid some systematic effects depending on metallicity and temperature .
[ fig_k ] we follow gordon et al . ( 1992 ) to derive h@xmath25 masses from line observations .
our temperature unit is expressed in antenna temperature scale which is corrected for atmospheric attenuation and rear sidelobes .
the radiation temperature t@xmath46 of the extragalactic source is then : where @xmath48 is the observed wavelength ( 2.6 mm ) , d is the iram radio telescope diameter ( 30 m ) , k is the correction factor for the coupling of the source with the beam , @xmath49 is the apperture efficiency ( 0.55 ) at 115 ghz , t@xmath50 is an antenna temperature which is @xmath51t@xmath52 in the iram convention , explicitely t@xmath50=0.92t@xmath52 , and @xmath53 is the source size . without taking into account cosmological correction , because of low redshift , column density of molecular hydrogen it written down as where @xmath55 is the velocity interval . in equation [ equa_tr ] k@xmath56 is the factor which corrects the measured antenna temperature for the weighting of the source distribution by the large antenna beam in case of a smaller source .
we have defined the source solid angle where @xmath60 denotes the normalized antenna power pattern ( baars , 1973 ) .
experiments have shown that we can approximate @xmath61 by a gaussian beam . as mentionned in section 3.3 an exponential law of scale length h = d@xmath11/10
is taken to model the source distribution function .
as long as the source size is smaller than the beam size we have where @xmath63 and @xmath64 are the beam and the source sizes .
we present in fig .
[ fig_k ] the plot of k in the case of exponential and uniform source distributions , for the iram-30 m beam at 115 ghz ( 22 ) . in the case of a source larger than the beam size
it is more difficult to compute the coupling of the beam to the source region since contributions from error beam can be quite significant to the resulting spectra .
a simple representation of the overall beam , including the error beam , is not available to correct that coupling . since in our sample , galaxies are at most a few beam sizes large we used equation [ equa_k ] integrated on the beam size to derive the k factor for galaxies with larger optical diameters
. + if i@xmath65 is the velocity - integrated temperature in t@xmath52 scale and given the iram-30 m parameters , the total mass of h@xmath25 is then given by we are very grateful to the staff at pico veleta for their help during these observations and especially to r. moreno and r. frapolli .
we also thank the anonymous referee whose suggestions greatly improved this paper .
this work largely benefitted from the leda , ned and simbad data bases .
this investigation was also supported by a grant from the natural sciences and engineering research council of canada to t.k.m .
allam s. , assendorp r. , longo g. et al : 1996 a&as 117 , 39 baars j.w.m .
: 1973 , ieee trans .
ap-21 , 461 barnes j. : 1989 , nature 338 , 123 barnes j. , hernquist l. : 1982 , ara&a 30 , 705 boselli a. , mendes des oliveira c. , balkowski c. , cayatte v. , casoli f. : 1996 , a&a 314 , 738 braine j. , combes f. : 1992 , a&a 264 , 433 braine j. , combes f. : 1993 , a&a 269 , 7 casoli f. , boiss p. , combes f. , dupraz c. : 1991 , a&a 249 , 359 casoli f. , dupraz c. , combes f. : 1992 , a&a 264 , 55 cayatte v. , van gorkom j.h . , balkowski c. , kotanyi c. : 1990 , aj 100 , 604 clements d.l . , baker a.c . : 1996 , a&a 314 , 5 combes f. , dupraz c. , casoli f. , pagani l. : 1988 , a&a 203 , l9 combes , f. , prugniel p. , rampazzo r. , sulentic j.w . : 1994 , a&a 281,725 dressler a. : 1984 , araa 22 , 185 garcia - gomez c. , athanassoula e. , garijo a. : 1996 , a&a 313 , 363 gordon m.a . , baars j.w.m .
, cocke w.j . : 1992 , a&a 264 , 337 gourgoulhon e. , chamaraux p. , fouqu p. : 1992
, a&a 255 , 69 governato f. , chincarini g. , bhatia r. : 1991 , apj 371 , l15 greve a. , panis j - f . ,
thum c. : 1996 a&as 115 , 379 henkel c. , mauersberger r. : 1993 , a&a 274 , 730 hernquist l. , katz n. , weinberg d.h .
: 1995 , apj 442 , 57 hickson p. : 1982 , apj 255 , 382 hickson p. , kindl e. , auman
j.r : 1989 , apjs 70 , 687 hickson p. , mendes de oliveira c. , huchra j.p . , palumbo g.g.c .
: 1992 , apj 399 , 353 hickson p. : 1990
, in iau coll 124 , `` paired and interacting galaxies '' , ed .
sulentic & w.c .
keel , p. 77
hildebrand , r.h .
1983 , qjras 24 , 267 israel f.p .
, tacconi l.j . ,
baas , f. : 1995 , a&a 295 , 599 kenney j.d.p .
, young j.s . : 1988 , apjs 66 , 261 leeuwin f. , combes f. : 1997 , mnras 284 , 45 leon s. , combes f. , junqueira s. , 1997 , in preparation lisenfeld u. , voelk h - j . , xu c. : 1996 , a&a 314 , 745 lutz , d. , genzel r. , sternberg a. et al : 1996 , a&a 315 , l137 mamon g. : 1986 , apj 307 , 426 mamon g. : 1987 , apj 321 , 622 mamon g. : 1992 in `` physics of nearby galaxies : nature or nurture ? '' , ed .
thuan , c. balkowski & j. tran thanh van , editions frontires , gif - sur - yvette , p. 367
mamon g.a . : 1994 , in `` n - body problems and gravitational dynamics '' , ed .
f. combes & e. athanassoula , p. 188
mauersberger r. , gulin m. , martin - pintado j. et al : 1989 a&as 79 , 217 mendes de oliveira c. , hickson p. : 1991 , apj 380 , 30 mendes de oliveira c. , hickson p. : 1994 ,
apj 427 , 684 mendes de oliveira c. : 1992 , phd thesis , british columbia univ .
menon t.k .
: 1995 , mnras 274 , 845 menon t.k . , hickson p. : 1985 , apj 296 , 60 menon t.k
: 1991 , apj 372 , 419 mulchaey j.s . ,
davis d.s .
, mushotzky r.f .
, burstein d. 1993 , apj 413 , l81 pildis r.a .
, bregman j.n . , schombert j.m . : 1995 , aj 110 , 149 pildis r.a . ,
evrard a.e .
, bregman j.n . : 1996 , aj 112 , 378 ponman t.j . , bertram d. : 1993 , nature 363 , 51 ponman t.j . , bourner p.d.j .
, ebeling h. , bhringer h. : 1996 , mnras 283 , 690 ribeiro a.l.b . , carvalho r.r .
, coziol r. , capelato h.v . , zepf s. : 1996 , apj 463 , l5 rood h.j .
, williams b.a . : 1989 , apj 339 , 772 rubio m. , lequeux j. , boulanger f. : 1993 , a&a 271 , 9 rubin v.c . , hunter d.a .
, ford w.k . :
1991 , apjs 76 , 153 sage l.j . , salzer j.j . ,
: henkel c 1992 , a&a 265 , 19 sage l.j . : 1993 , a&a 272 , 123 sanders d.b . , soifer b.t . , elias j.h et al : 1988 , apj 325 , 74 sanders d.b
, scoville n.z .
, soifer b.t . : 1991 , apj 370 , 158 solomon p.m. , sage l.j . : 1988 , apj 613 , 334 strong a.w . , bloemen j.b.g.m .
, dame t.m .
et al : 1988 a&a 207 , 1 sulentic j.w . , de mello rabaa , d. : 1993 , apj 410 , 520 sulentic j.w . ,
rabaa , c. : 1994 , apj 429 , 531 thronson , h.a . ,
telesco , c.m .
1986 , apj 311 , 98 tinney c.g .
, scoville n.z .
, sanders d.b .
, soifer b.t . : 1990 , apj 362 , 473 verdes - montenegro l. et al : 1997 , a&a sub verter , f. , 1993 , apj , 402 , 141 vigroux l. , mirabel f. , altieri b. et al : 1996 , a&a 315 , l93 wiklind t. , combes f. , henkel c. : 1995 , a&a 297 , 643 williams b.a . , rood h.j . : 1987 , apjs 63 , 265 xu c. , sulentic j.w .
: 1991 , apj 374 , 407 young j.s . , xie s. , kenney j.d.p .
, rice w.l . : 1989 , apjs 70 , 699 young j.s . , xie s. , tacconi l. et al : 1995 , apjs 98 , 219 young j.s .
, allen l , kenney j.d.p . , lesser a. , rownd b. : 1996 , aj 112 , 1903 young , j.s . , knezek , p. 1989
, apj 347 , l55 yun m.s . , verdes - montenegro l. , del olmo a. , perea j. : 1997 , apj 457 , l21 | we have observed 70 galaxies belonging to 45 hickson compact groups in the and lines , in order to determine their molecular content .
we detected 57 galaxies , corresponding to a detection rate of 81% .
we compare the gas content relative to blue and luminosities of galaxies in compact groups with respect to other samples in the literature , including various environments and morphological types .
we find that there is some hint of enhanced / and / ratios in the galaxies from compact group with respect to our control sample , especially for the most compact groups , suggesting that tidal interactions can drive the gas component inwards , by removing its angular momentum , and concentrating it in the dense central regions , where it is easily detected .
the molecular gas content in compact group galaxies is similar to that in pairs and starburst samples .
however , the total luminosity of hcgs is quite similar to that of the control sample , and therefore the star formation efficiency appears lower than in the control galaxies .
however this assumes that the fir spatial distributions are similar in both samples which is not the case at radio frequencies .
higher spatial resolution fir data are needed to make a valid comparison . given their short dynamical friction time - scale , it is possible that some of these systems are in the final stage before merging , leading to ultra - luminous starburst phases .
we also find for all galaxy samples that the content ( derived from co luminosity and normalised to blue luminosity ) is strongly correlated to the luminosity , while the total gas content ( + ) is not . |
many workers argue that a thermodynamic anomaly underlies the onset of glassy dynamics @xcite .
evidence for this view is the rough , though not quantitative @xcite , correlation between dynamic fragility and excess heat capacity discontinuity @xcite .
this thermodynamic view contrasts with the picture we have advocated @xcite , attributing glassy behavior to dynamic heterogeneity @xcite , with growing length scales appearing in space - time , but not space alone @xcite . indeed ,
direct observations of diffusive motion in colloidal glasses reveal excitations that are local and sparse @xcite .
these findings seem consistent with excitations being local and uncorrelated , as assumed in the two - state model for the low temperature behavior of structural glass heat capacities @xcite .
spatial correlations under such conditions can arise through constraints on particle motions that are relieved only when adjacent regions have exhibited some degree of mobility @xcite .
trajectories are then correlated throughout space and time , with varying degrees of hierarchical structure determining the extent to which the system is fragile or strong @xcite . from this perspective
, a non - thermodynamic explanation emerges for correlation between heat capacity and fragility @xcite : the concentration of excitations required to support fragile hierarchical dynamics is higher than that required for strong non - hierarchical dynamics .
the juxtaposition of heat capacity discontinuities is therefore understood as a consequence of different excitation concentrations .
biroli , bouchaud and tarjus ( bb&t ) @xcite have taken issue with this explanation .
they show that the simplest treatment of defect models can not simultaneously fit the size and temperature dependence of the experimental heat capacity while simultaneously fitting dynamic properties . in general and in more current context ,
however , the terminology `` defect models '' refers to a broad class of kinetically constrained models @xcite . as a result of this generality ,
their question `` are defect models consistent with the entropy and specific heat of glass - formers ? '' has a non - trivial answer ( see also the analysis in @xcite ) .
in particular , this class of models has an assortment of possible dynamical behaviors , so that relaxation data can be fit with many different functions of excitation concentration .
in addition , kinetically constrained models are coarse grained caricatures of fluids , so that many degrees of freedom remain unspecified . here , we focus on this latter feature , the possible consequences of unspecified degrees of freedom . the starting point is the assumption that small length - scale and small time - scale features of a fluid can be integrated out leaving only simple stochastic rules for the dynamics of discrete variables on a lattice .
these rules contain constraints imagined to be the consequence of the actual intermolecular interactions , interactions that limit the space or metric for molecular motions @xcite .
the excitations or defects that survive coarse graining distinguish microscopic regions of space - time that exhibit molecular mobility from those where molecules are jammed or immobile . in particular ,
an excitation or defect is a microscopic region of space for which particle mobility emerges .
this characterization is related to coarse graining in time , because it takes time to discern whether or not mobility is exhibited .
bear in mind , `` defect '' does not distinguish disordered arrangements of atoms from those that are ordered because there are generally many disordered yet jammed configurations .
similarly , a lack of `` excitation '' does not necessarily imply low energy because some immobile regions may have the same energy as mobile regions . uncertainty concerning this terminology may be the origin of criticisms of facilitated models leveled by bb&t @xcite and also by lubchenko and wolynes @xcite . as a simple illustration ,
consider dividing space into cells , with grid spacing larger than the equilibrium correlation length of the material . at a given time frame ,
the micro states of different cells are then uncorrelated .
further , suppose there are two energy levels , @xmath0 and @xmath1 , for the states of a given cell , where the states with energy @xmath1 include those that will exhibit particle mobility after a coarse graining time @xmath2 .
there is a fraction , @xmath3 , of states with energy @xmath1 that exhibit mobility in this way ; the others remain jammed . in this case
the average excitation concentration is @xmath3 times the concentration of cells with energy @xmath1 , and thus the number of molecules contributing to energy or enthalpy fluctuations is @xmath4 times larger than that contributing to mobility fluctuations . in other words ,
when considering the thermodynamic implications of a kinetically constrained defect model , a factor @xmath4 is required to account for energetic states or degrees of freedom absent from explicit consideration .
bb&t ignore this factor , which leads them to the difficulties they describe .
in the next section we discuss this point in greater detail .
we show that facilitated or kinetically constrained models such as the one we presented in ref .
@xcite admit a variety of possibilities for partitioning micro states .
this demonstration justifies the heat capacity formulas we used in ref .
@xcite and therefore supports our argument about the juxtaposition of heat capacities for strong and fragile glass formers .
the dynamics of facilitated models we consider are governed by master equations for distribution functions of the mobility field on a lattice .
consistent with our earlier papers , we use the symbol @xmath5 to denote the value of the mobility field at lattice site @xmath6 during time period @xmath7 .
it is a binary field in that @xmath8 takes on one of two values , @xmath0 ( corresponding to a cell that is unexcited , jammed ) or @xmath9 ( corresponding to excited , mobile ) .
each cell is imagined to contain several molecules , so we expect that even after some coarse graining , there are still many ( not just two ) microstates possible for each cell at a given time @xmath7 .
we further imagine these many microstates , specified with @xmath10 , evolve according to their own master equation .
the master equation for the distribution of the mobility field @xmath11 is a contraction of that for the micro - state field @xmath10 . in this section ,
we show how this contraction works .
we begin by considering the thermodynamics and statistics of the fields @xmath11 and @xmath10 , and then consider the master equations for the distributions of these fields .
imagine we partition the liquid into a lattice with grid spacing @xmath12 , and denote the state in cell @xmath6 by @xmath10 .
imagine further that the set of available cell levels @xmath13 can be split into two groups : one subset of @xmath13 is `` unexcited '' with respect to mobility , @xmath14 , and the complementary subset is `` excited '' , @xmath15 . in practice ,
@xmath11 is most conveniently determined by observing system for a coarse graining time @xmath2 .
this splitting , or projection , of many micro states into two different mobility states ( or several , as in the generalization of @xcite ) is the the first central assumption of facilitated models @xcite . while the state @xmath16 of a cell will correspond to one of the two mobility states , @xmath17 , in general its energy , @xmath18 ,
will be distributed .
( for notational simplicity , when clarity is not diminished , we drop the subscripts on @xmath8 and @xmath19 . )
let @xmath20 and @xmath21 represent the normalized probability densities for the energy levels in states @xmath22 and @xmath23 , respectively .
the cell partition function then reads ( we set @xmath24 ) : @xmath25 where @xmath26 and @xmath27 gives the total number of levels in each of the two mobility states .
the excitation concentration is given by : @xmath28 the average energy per cell is : @xmath29 where @xmath30 the specific heat per cell is : @xmath31 where @xmath32 the quantity @xmath33 determines the concentration @xmath34 of excitations ; @xmath35 gives the average cell energy ; and @xmath36 sets the specific heat at small @xmath34 .
except in the case where @xmath37 are delta functions , these scales are in principle all different .
as a simple example , which will be useful below , consider the following model with just two energy levels .
the unexcited states @xmath22 occupy both energy levels , which we denote @xmath38 and @xmath39 , with energies @xmath40 and @xmath41 , respectively : @xmath42 where @xmath43 sets the relative degeneracy between the two levels .
the excited states @xmath23 have a single energy level , which we denote @xmath44 , with energy @xmath45 : @xmath46 in this case , the cell partition function , average energy and specific heat become : @xmath47 where @xmath48 denotes the relative occupation of the states @xmath39 with respect to states @xmath38 , @xmath49 $ ] .
the total concentration of levels with energy @xmath1 is @xmath50 where we have defined @xmath3 as the ratio of mobile cells to the total number of cells with energy @xmath1 .
the specific heat then reads @xmath51 we now show how an appropriate projection of the dynamics of the full states @xmath19 of individual cells into their mobility states @xmath52 gives a two - state facilitated model such as the fa or east models , and the slightly more general class of models introduced in ref.@xcite .
the approach we take has been used to derive approximate real - space renormalization group equations @xcite . here
, however , the approach requires no approximation .
the master equation for the dynamics can be written @xcite : @xmath53 where @xmath54 is the state vector for the probability density of the system at time @xmath7 , and @xmath55 is the liouvillian operator for the dynamics . in the direct product representation
it reads @xcite : @xmath56 where @xmath57 is the state vector for level @xmath19 in site @xmath6 .
the second assumption in facilitated models is that transitions in a cell @xmath6 are only possible if neighbouring cells @xmath58 are mobile , @xmath59 @xcite . in a system with facilitated dynamics
@xmath55 reads @xcite : @xmath60 where @xmath61 is the kinetic constraint operator imposing facilitation , and @xmath62 is the unconstrained dynamic operator at site @xmath6 .
for example , in the east facilitated model @xmath63 , and in the direct product representation : @xmath64 the off diagonal elements of @xmath62 are given by the transition rates between levels : @xmath65 , while the diagonal elements are such that the sum of columns in @xmath62 vanish .
a choice satisfying detailed balance is @xmath66 , where @xmath67 is the equilibrium distribution .
notice that we have assumed that transitions between any of the levels in the cell are only possible if facilitated by a neighbour . in order to project the dynamics of the full set of levels @xmath68 into that of only the excitation occupancies
@xmath69 we need projection , @xmath70 , and embedding , @xmath71 , operators @xcite : @xmath72 where the @xmath73 and @xmath74 s are constants .
the projection is performed independently in each cell , so these are site diagonal operators .
they must also satisfy @xmath75 .
the dynamics of the projected state @xmath76 is then given by the projected liouvillian @xmath77 where the elements of the projection and embedding matrices ( [ t1t2 ] ) are given by : @xmath78 the choice of rates in @xmath55 , which depend only on the final level , @xmath79 , ensures that the projected dynamics is markovian @xcite .
we now show explicitly how the projection procedure works in the simple the three - level example of the previous subsection . in this case @xmath80 , and the onsite operators can be represented by @xmath81 matrices : @xmath82 of the three levels @xmath68 , only @xmath83 corresponds to the excited state , and the projection matrix is : @xmath84 the corresponding embedding matrix is : @xmath85 the projected operators then read : @xmath86 these are precisely the operators for a two state facilitated model @xcite . using the facilitated models of ref .
@xcite and compared to the formula used by biroli , bouchaud and tarjus ( indicated as bb&t ) @xcite . *
top panel * : excess specific heat , @xmath87 , of supercooled otp ( left hand scale ) and 3bp ( right hand scale ) .
symbols indicate experimental data @xcite .
full curves are fits using the formulas for the heat capacity per molecule discussed in the text ( where @xmath88 ) : @xmath89 , where @xmath90 and @xmath91 . here
g \exp \left ( -j / t + j / t_{\mathrm{ref } } \right)$ ] , where the first ( approximate ) equality neglects terms of order @xmath93 , and the second introduces the notation used in our ref .
@xcite : @xmath38 is a degeneracy that we imagine measures the possible directions of mobility , and @xmath94 bounds from above the temperatures we consider .
the values of @xmath95 and @xmath96 for otp and 3bp were determined in @xcite from kinetic data . on physical grounds
we expect @xmath97 , and @xmath3 to increase with temperature .
the fitting of @xmath87 gives @xmath98 and @xmath99 for otp and 3bp , respectively .
@xmath100 at @xmath101 as argued in @xcite . clearly , it is possible to describe the size and modest change with @xmath102 of the specific heat in the vicinity of @xmath101 , in contrast to the observation of bb&t . *
bottom panel * : excess entropy , @xmath103 , of otp ( left hand scale ) and 3bp ( right hand scale ) .
symbols correspond to experimental data @xcite .
full curves are fits using the formula for the entropy per molecule which follows from eq .
( [ c1 ] ) : @xmath104 } - ( \phi^{-1 } c ) \ln{(\phi^{-1 } c ) } - ( 1- \phi^{-1 } c ) \ln{(1-\phi^{-1 } c ) } + ( \phi^{-1 }
c ) \ln{[(\omega_0 \alpha + \omega_1)/(1-\alpha ) \omega_0 ] } \}$ ] .
the shown fits correspond to the same parameters as before plus @xmath105 } , s^{-1 } \ln{[(\omega_0 \alpha + \omega_1)/(1-\alpha ) \omega_0]}=2.8 , 1.2 $ ] and @xmath106 for otp and 3bp , respectively.,title="fig:",width=325 ] using the facilitated models of ref .
@xcite and compared to the formula used by biroli , bouchaud and tarjus ( indicated as bb&t ) @xcite . *
top panel * : excess specific heat , @xmath87 , of supercooled otp ( left hand scale ) and 3bp ( right hand scale ) .
symbols indicate experimental data @xcite .
full curves are fits using the formulas for the heat capacity per molecule discussed in the text ( where @xmath88 ) : @xmath89 , where @xmath90 and @xmath91 . here
g \exp \left ( -j / t + j / t_{\mathrm{ref } } \right)$ ] , where the first ( approximate ) equality neglects terms of order @xmath93 , and the second introduces the notation used in our ref .
@xcite : @xmath38 is a degeneracy that we imagine measures the possible directions of mobility , and @xmath94 bounds from above the temperatures we consider .
the values of @xmath95 and @xmath96 for otp and 3bp were determined in @xcite from kinetic data . on physical grounds
we expect @xmath97 , and @xmath3 to increase with temperature .
the fitting of @xmath87 gives @xmath98 and @xmath99 for otp and 3bp , respectively .
@xmath100 at @xmath101 as argued in @xcite . clearly , it is possible to describe the size and modest change with @xmath102 of the specific heat in the vicinity of @xmath101 , in contrast to the observation of bb&t .
* bottom panel * : excess entropy , @xmath103 , of otp ( left hand scale ) and 3bp ( right hand scale ) .
symbols correspond to experimental data @xcite .
full curves are fits using the formula for the entropy per molecule which follows from eq .
( [ c1 ] ) : @xmath104 } - ( \phi^{-1 } c ) \ln{(\phi^{-1 }
c ) } - ( 1- \phi^{-1 } c ) \ln{(1-\phi^{-1 } c ) } + ( \phi^{-1 } c ) \ln{[(\omega_0 \alpha + \omega_1)/(1-\alpha ) \omega_0 ] } \}$ ] .
the shown fits correspond to the same parameters as before plus @xmath105 } , s^{-1 } \ln{[(\omega_0 \alpha + \omega_1)/(1-\alpha ) \omega_0]}=2.8 , 1.2 $ ] and @xmath106 for otp and 3bp , respectively.,title="fig:",width=302 ]
subsection [ thermo ] tells us that if the mobility field is the result of coarse - graining and projecting the many mobile and immobile micro states , then different energy ( or enthalpy ) scales enter into the concentration of mobile cells @xmath34 , the average energy ( or enthalpy ) , and the specific heat .
how these scales relate to each other depends on the distribution of energy levels . at equilibrium ,
the average excitation concentration , @xmath34 , is a function of thermodynamic state .
its value should be indicative of the concentration of regions with excess energy or enthalpy , but to say more requires more assumptions or detail than needed to predict dynamical trends . in order to illustrate how facilitated
models can be used to fit thermodynamic data the simple example given above of two immobile cell levels @xmath38 and @xmath39 and a single mobile level @xmath44 will suffice . from eqs .
( [ z3state])([cv3state ] ) we see that while @xmath34 gives the concentration of cells displaying mobility , both mobile and immobile cells can contribute to energy ( or enthalpy ) fluctuations . the relevant quantity to determine the specific heat is not @xmath34 but @xmath107 . from eqs .
( [ defphi ] ) and ( [ cv3state ] ) we have that the specific heat _ per cell _ is : @xmath108 the asymptotic equality of eq .
( [ c1 ] ) is , in effect , the expression we refer to in the penultimate paragraph of ref.@xcite .
the factor @xmath109 accounts for the number of molecules contributing to energy or enthalpy fluctuations , this number being larger than that for molecules contributing to mobility fluctuations .
this fact is implicit in the coarse graining procedure , as shown above . in ref .
@xcite , @xmath4 is included in the quantity we called @xmath110 . in the few lines devoted to this topic in ref .
@xcite , the meaning of @xmath110 was obscure and its description misleading . indeed , where we wrote `` @xmath110 is the number of molecules that contribute to enthalpy fluctuations per mobile cell ''
, we should have written `` @xmath110 accounts for the number of molecules that contribute to enthalpy fluctuations beyond those associated with mobile cells . '' to be perfectly explicit , if @xmath111 is the average number of molecules in a cell , then @xmath112 . with this expression , the relative values of @xmath113 for several super - cooled liquids at @xmath101 are consistent with the relative values of their heat capacity jumps at the glass transition .
this consistency is independent of the value of @xmath111 and @xmath110 , but the value of @xmath110 does play a role in the actual value of the heat capacity discontinuity .
specifically , @xmath114 is consistent with the value of the discontinuities of the heat capacity per molecule at @xmath115 when @xmath116 .
bb&t @xcite identify @xmath111 with @xmath110 , and further neglect the factor @xmath117 , and in so doing disagree with this consistency .
bb&t s second criticism is about the temperature variation of @xmath113 .
they consider the relationship @xmath118 this formula is consistent with the asymptotic equality of eq.([c1 ] ) , which requires @xmath119 to be small .
while it is small at the glass transition , @xmath120 may not be small at the higher temperatures considered .
the right - hand side of bb&t s formula is missing the requisite factor of @xmath121 present in eq .
( [ c1 ] ) .
it is also missing a possible temperature dependence of @xmath3 .
this temperature dependence would reflect that of transport in a liquid .
it is typically more pronounced for isobaric temperature variation than for isochoric temperature variation , but weak and sub - arrhenius in either case @xcite .
thus , we write @xmath122 , where the parameters @xmath123 and @xmath124 are temperature independent . in ref .
@xcite , we chose to not consider @xmath113 above @xmath101 because doing so requires the additional parameters described in the previous paragraph .
but since , bb&t raise this point , we show in fig.1 that the effects of these extra factors on the temperature dependence of @xmath113 is significant .
this figure can be contrasted with that in bb&t s @xcite .
while the asymptotic formula predicts an exponential variation with temperature , the full formula predicts modest variation with temperature that is not inconsistent with experiment .
we also show fits to the corresponding excess entropy , @xmath125 .
better agreement between theory and experiment for these properties would require modification of the models and perhaps further parameterization . in the us ,
this work was supported initially by the nsf , and more recently by doe grant no .
de - fe - fg03 - 87er13793 . in the uk
, it was supported by epsrc grants no .
gr / r83712/01 and gr / s54074/01 , and university of nottingham grant no .
fef 3024 .
see , for example , d. kivelson , s.a .
kivelson , x .-
zhao , z. nussinov , and g. tarjus , physica a * 219 * , 27 ( 1995 ) ; m. mzard and g. parisi , phys .
82 * , 747 ( 1999 ) ; x. xia and p.g .
wolynes , proc .
usa * 97 * , 2990 ( 2000 ) ; s. franz and g. parisi , j. phys .
c * 12 * , 6335 ( 2000 ) .
for reviews see : h. sillescu , j. non - cryst .
solids * 243 * , 81 ( 1999 ) ; m.d .
ediger , annu .
phys . chem . *
51 * , 99 ( 2000 ) ; s.c .
glotzer , j. non - cryst . solids , * 274 * , 342 ( 2000 ) ; r. richert , j. phys . condens .
matter * 14 * , r703 ( 2002 ) .
l. berthier and j.p .
garrahan , phys .
e * 68 * , 041201 ( 2003 ) ; s. whitelam , l. berthier and j.p .
garrahan , phys .
92 * , 185705 ( 2004 ) ; y.j .
jung , j.p .
garrahan and d. chandler , phys .
e * 69 * , 061205 ( 2004 ) ; l. berthier , d. chandler and j.p .
garrahan , europhys . lett . * 69 * , 230 ( 2005 ) . | in recent papers , we have argued that kinetically constrained coarse grained models can be applied to understand dynamic properties of glass forming materials , and we have used this approach in various applications that appear to validate this view . in one such paper [ j.p .
garrahan and d. chandler , proc .
nat .
acad .
sci .
usa * 100 * , 9710 ( 2003 ) ] , among other things we argued that this approach also explains why the heat capacity discontinuity at the glass transition is generally larger for fragile materials than for strong materials . in the preceding article , biroli , bouchaud and tarjus ( bb&t ) have objected to our explanation on this point , arguing that the class of models we apply is inconsistent with both the absolute size and temperature dependence of the experimental specific heat .
their argument , however , neglects parameters associated with the coarse graining . accounting for these parameters ,
we show here that our treatment of dynamics is not inconsistent with heat capacity discontinuities . |
several new estimation methods have been recently proposed for the linear regression model with observation error in the design .
such problems arise in a variety of applications , see @xcite . in this work
we consider the following regression model with observation error in the design : @xmath2 here the random vector @xmath3 and the random @xmath4 matrix @xmath5 are observed , the @xmath4 matrix @xmath6 is unknown , @xmath7 is an @xmath4 random noise matrix , and @xmath8 is a random noise vector .
the vector of unknown parameters of interest is @xmath9 which is assumed to belong to a given convex subset @xmath10 of @xmath11 characterizing some prior knowledge about @xmath9 ( potentially @xmath12 ) .
similarly to the recent literature on this topic , we consider the setting where the dimension @xmath13 can be much larger than the sample size @xmath14 and the vector @xmath9 is @xmath15-sparse , which means that it has not more than @xmath15 non - zero components .
+ the need for new estimators under errors in the design arises from the fact that standard estimators ( e.g. lasso and dantzig selector ) might become unstable , see @xcite . to deal with this framework ,
various assumptions have been considered , leading to different estimators . +
a classical assumption in the literature is a uniform boundedness condition on the errors in the design , namely , @xmath16 denotes the @xmath17-norm for @xmath18 .
note that this assumption allows for various dependences between the errors in the design . in this
setting , the matrix uncertainty selector ( mu selector ) , which is robust to the presence of errors in the design , is proposed in @xcite .
the mu selector @xmath19 is defined as a solution of the minimization problem @xmath20 where the parameters @xmath21 and @xmath22 depend on the level of the noises of @xmath7 and @xmath23 respectively . under appropriate choices of these parameters and suitable assumptions on @xmath6 ,
it was shown in @xcite that with probability close to 1 , @xmath24 different positive constants that do not depend on @xmath9 , @xmath15 , @xmath14 , @xmath13 , @xmath25 .
the result ( [ rtmu ] ) implies consistency as the sample size @xmath14 tends to infinity provided that the error in the design goes to zero sufficiently fast to offset @xmath26 , and the number of variables @xmath13 and the sparsity @xmath15 of @xmath27 do not grow too fast relative to the sample size @xmath14 .
+ an alternative assumption considered in the literature is that the entries of the random matrix @xmath7 are independent with zero mean , the values @xmath28 are finite , and data - driven estimators @xmath29 of @xmath30 are available converging with an appropriate rate .
this assumption motivated the idea to compensate the bias of using the observable @xmath31 instead of the unobservable @xmath32 in thanks to the estimates of @xmath30 .
this compensated mu selector , introduced in @xcite and denoted as @xmath33 , is defined as a solution of the minimization problem @xmath34 where @xmath35 is the diagonal matrix with entries @xmath36 and @xmath37 and @xmath38 are constants chosen according to the level of the noises and the accuracy of the @xmath39 .
+ rates of convergence of the compensated mu selector were established in @xcite .
importantly , the compensated mu selector can be consistent as the sample size @xmath14 increases even if the error in the design does not vanish .
this is in contrast to the case of the mu selector , where the bounds are small only if the bound on the design error @xmath0 is small . in particular , under regularity conditions , when @xmath27 is @xmath15-sparse , it is shown in @xcite that with probability close to 1 @xmath40 has been recently proposed and analyzed in @xcite .
the estimator @xmath41 is defined as the first component of any solution of the optimization problem @xmath42 where @xmath43 , @xmath21 and @xmath22 are some positive tuning constants .
akin to @xmath33 , this estimator compensates for the bias by using the estimators @xmath36 of @xmath30 .
however it exploits a combination of @xmath44 and @xmath45-norm regularization to be more adaptive .
it was shown to attain a bound as in ( [ rt13 ] ) and to be computationally feasible since it is cast as a tractable convex optimization problem ( a second order cone programming problem ) . moreover ,
under mild additional conditions , with probability close to 1 , the estimator ( [ conic ] ) achieves improved bounds of the form @xmath46 converges to @xmath47 in sup - norm with the rate @xmath48 .
it is shown in @xcite that the rate of convergence in is minimax optimal in the considered model .
+ there have been other approaches to the errors - in - variables model , usually exploiting some knowledge about the vector @xmath27 , see @xcite .
assuming @xmath49 is known , @xcite proposed an estimator @xmath50 defined as the solution of a non - convex program which can be well approximated by an iterative relaxation procedure . in the case where the entries of the regression matrix @xmath6 are zero - mean subgaussian and @xmath27 is @xmath15-sparse , under appropriate assumptions
, it is shown in @xcite that for the error in @xmath45-norm ( @xmath51 ) , @xmath52 depends on @xmath9 , so that there is no guarantee that the estimator attains the optimal bound as in . assuming that the sparsity @xmath15 of @xmath27 is known and the non - zero components of @xmath27 are separated from zero in the way that @xmath53 an orthogonal matching pursuit algorithm to estimate @xmath9 is introduced in @xcite .
focusing as in @xcite on the particular case where the entries of the regression matrix @xmath6 are zero - mean subgaussian , it is shown in @xcite that this last estimator satisfies a bound analogous to ( [ 6 ] ) , as well as a consistent support recovery result .
+ the main purpose of this work is to show that an additional regularization term based on the @xmath1-norm leads to improved rates of convergence in several situations .
we propose two new estimators for @xmath27 .
the first proposal is applicable under a new combination of the assumptions mentioned above .
namely , we assume that the components of the errors in the design are uniformly bounded by @xmath25 as in ( [ wbound ] ) , and that the rows of @xmath7 are independent and with zero mean .
however , we will neither assume that a data - driven estimator @xmath54 is available , nor that specific features of @xmath27 are known ( e.g. @xmath15 or @xmath49 ) .
the estimator is defined as a solution of a regularized optimization problem which uses simultaneously @xmath44 , @xmath45 , and @xmath1 regularization functions .
it can be cast as a convex optimization problem and the solution can be easily computed .
we study its rates of convergence in various norms in section [ first ] .
one of the conclusions is that for @xmath55 the new estimator has improved rates of convergence compared to the mu selector .
furthermore , note that the conic estimator @xmath56 studied in @xcite can be also applied .
indeed , our setting can be embedded into that of @xcite with @xmath57 being the identically zero @xmath58 matrix , which means that we have an estimator of each @xmath30 with an error bounded by @xmath59 . comparing the bounds
yields that the conic estimator @xmath56 achieves the same rate as our new estimator if @xmath25 is smaller than or of the order @xmath60 . however , there is no bound for @xmath56 available when @xmath61 .
+ the second estimator we propose applies to the same setting as in @xcite .
the idea of taking advantage of an additional @xmath1-norm regularization can be used to improve the conic estimator @xmath56 of @xcite whenever the rate of convergence of the estimator @xmath57 for @xmath30 , @xmath62 , is slower than @xmath48 .
this motivates us to propose and analyze a modification of the conic estimator .
we derive new rates of convergence that can lead to improvements . however , we acknowledge that in the case considered in @xcite , where the rate of convergence of @xmath57 is @xmath48 , there is no gain in the rates of convergence when using the additional @xmath1-norm regularization . + the paper is organized as follows .
section [ sec : assump ] contains the notation , main assumptions and some preliminary lemmas needed to determine threshold constants in the algorithms . the definition and properties of our first estimator
are given in section [ first ] whereas those of our second procedure can be found in section [ second ] .
section 5 contains simulation results .
some auxiliary lemmas are relegated to an appendix .
in this section , we introduce the assumptions which will be required to derive the rates of convergence of the proposed estimators .
one set of conditions pertains to the design matrix and the second to the errors in the model .
we also state preliminary lemmas related to the stochastic error terms .
we start by introducing some notation .
let @xmath63 be a set of integers .
we denote by @xmath64 the cardinality of @xmath65 . for a vector @xmath66 in @xmath67 , we denote by @xmath68 the vector in @xmath67 whose @xmath69th component satisfies @xmath70 if @xmath71 , and @xmath72 otherwise . for @xmath73 ,
the random variable @xmath74 is said to be _ sub - gaussian with variance parameter _
@xmath75 ( or shortly _ @xmath76-sub - gaussian _ ) if , for all @xmath77 , @xmath78\leq \text{exp}(\gamma^2t^2/2).\ ] ] a random vector @xmath79 is said to be _ sub - gaussian with variance parameter _ @xmath75 if the inner products @xmath80 are @xmath76-sub - gaussian for any @xmath81 with @xmath82 . the performance of the estimators that we consider below is influenced by the properties of the gram matrix @xmath83 we will assume that : _ _ * the matrix @xmath6 is deterministic . in order to characterize the behavior of the design matrix , we set @xmath84 where the @xmath85 are the elements of matrix @xmath6 and we consider the sensitivity characteristics related to the gram matrix @xmath86 . for @xmath87
, define the cone @xmath88 where @xmath65 is a subset of @xmath89 . for @xmath90 $ ] and an integer @xmath91 $ ] , the _ @xmath17-sensitivity _ ( cf .
@xcite ) is defined as follows : @xmath92 like in @xcite , we use here the sensitivities to derive the rates of convergence of estimators under sparsity .
importantly , as shown in @xcite , the approach based on sensitivities is more general than that based on the restricted eigenvalue or the coherence conditions , see also @xcite . in particular , under those conditions , we have @xmath93 for some constant @xmath94 , which implies the usual optimal bounds for the errors .
next we turn to the error @xmath7 in the design and the error @xmath23 in the regression equation .
we will make the following assumptions .
+ _ _ * the elements of the random vector @xmath23 are independent zero - mean sub - gaussian random variables with variance parameter @xmath95 . *
the rows @xmath96 , @xmath97 , of the noise matrix @xmath7 are independent zero - mean sub - gaussian random vectors with variance parameter @xmath98 .
furthermore , @xmath7 is independent of @xmath23 .
we now state some useful lemmas from @xcite and @xcite that provide bounds to various stochastic error terms that play a role in our analysis .
we state them here because they introduce the thresholds @xmath99 that will be used in the definition of the estimators . in
what follows , @xmath47 is the diagonal matrix with diagonal elements @xmath30 , @xmath100 , and for a square matrix @xmath101 , we denote by @xmath102 the matrix with the same dimensions as @xmath101 , the same diagonal elements , and all off - diagonal elements equal to zero . [ lem2 ]
let @xmath103 and assume ( a1)-(a3 ) . then , with probability at least @xmath104 ( for each event ) , @xmath105 where @xmath106 and for an integer @xmath107 , @xmath108 where @xmath109 are positive constants depending only on @xmath110 .
[ lem3a ] let @xmath103 , @xmath111 and assume ( a1)-(a3 ) . then , with probability at least @xmath104 , @xmath112 where @xmath113 .
in addition , with probability at least @xmath104 , @xmath114 where @xmath115 and @xmath116 are positive constants depending only on @xmath117 .
the proofs of lemmas [ lem2 ] and [ lem3a ] can be found in @xcite and @xcite respectively .
in this section , we define and analyze our first estimator .
it can be seen as a compromise between the mu selector ( [ def : mu ] ) and the conic estimator ( [ conic ] ) achieved thanks to an additional @xmath1-norm regularization . in the setting that we consider now , the estimate @xmath54 is not available but the rows of the design error matrix @xmath7 are independent with mean 0 , and its entries are uniformly bounded .
formally , in this section , we make the following assumption . _
_ * almost surely , @xmath119 .
thus , assumptions ( a1)-(a4 ) imply the assumptions in @xcite .
however , they neither imply or are implied by the assumptions in @xcite .
that is , it is an intermediary set of conditions relative to the original assumptions for the mu selector in @xcite and to those for the compensated mu selector in @xcite .
importantly , we do not assume that there are some accurate estimators of the @xmath30 .
+ we consider the estimator @xmath120 such that @xmath121 is a solution of the following minimization problem @xmath122 where @xmath123 and @xmath124 are tuning constants and the minimum is taken over @xmath125 .
this estimator @xmath120 will be further referred to as the @xmath118-mu selector .
+ the estimator above attempts to mimic the conic estimator ( [ conic ] ) without an estimator @xmath57 for @xmath30 , @xmath62 . in order to make @xmath27 feasible for ( [ conicmunew ] ) , the contribution of the unknown term @xmath126 needs to be bounded .
this is precisely the role of the extra term @xmath127 in the constraint since @xmath128 and @xmath129 almost surely .
note that the use of @xmath130 and @xmath131 instead of @xmath132 and @xmath133 in the constraint makes ( [ conicmunew ] ) a convex programming problem .
+ this new estimator exploits assumptions ( a2)-(a4 ) to achieve a rate of convergence that is intermediary relative to the rate of the mu selector and to that of the conic estimator .
+ set @xmath134 and @xmath135 .
note that @xmath21 and @xmath22 are of order @xmath48 .
the next theorem summarizes the performance of the estimator defined by solving ( [ conicmunew ] ) .
[ th : munew ] let assumptions ( a1)-(a4 ) hold .
assume that the true parameter @xmath27 is @xmath15-sparse and belongs to @xmath136 .
let @xmath103 , @xmath137 and @xmath138 , and let @xmath120 be the @xmath118-mu selector . if @xmath139 for some constant @xmath94 then , with probability at least @xmath140 , @xmath141 for some constants @xmath142 and @xmath143 ( here we set @xmath144 ) .
+ if in addition , @xmath145 for some small enough constant @xmath146 then , with the same probability we have @xmath147 for some constants @xmath142 and @xmath143 .
+ under the same assumptions with @xmath148 , the prediction error admits the following bound , with the same probability : @xmath149 we proceed in three steps .
step 1 establishes initial relations and the fact that @xmath150 belongs to @xmath151 .
step 2 provides a bound on @xmath152 .
step 3 establishes the rates of convergence stated in the theorem .
we work on the event of probability at least @xmath153 where all the inequalities in lemmas [ lem2 ] and [ lem3a ] are realized . throughout the proof ,
@xmath154 we often make use of the inequalities @xmath155 + _ step 1 .
_ we first note that @xmath156 with probability at least @xmath157 by lemma [ lem2 ] .
next , lemma [ lem3a ] and the fact that , due to ( [ wbound ] ) , we have @xmath158 imply @xmath159 combining and we get that @xmath160 is feasible for the problem ( [ conicmunew ] ) , so that @xmath161 arguments similar to lead to @xmath162 _ step 2 .
_ we have @xmath163 the results of step 1 and of lemmas [ lem2 ] and [ lem3a ] imply the following bounds @xmath164 these relations and the inequality @xmath165 yield that @xmath166 _ step 3 . _
next note that @xmath167 letting @xmath168 we have @xmath169 by the definition of the @xmath17-sensitivity , @xmath170 now , follows by combining the last two displays and the assumption on @xmath171 . to prove
, we use that @xmath172 under our conditions , @xmath173 for some @xmath174 .
thus , we have @xmath175 which implies in view of the definition of the @xmath17-sensitivity and the assumption on @xmath171 . + to show ,
note first that @xmath176 by with @xmath148 , @xmath177 combining this inequality with and proves .
we have stated theorem [ th : munew ] under assumption ( a4 ) to make the analysis streamlined with the previous literature , see @xcite .
however , inspection of the proofs shows that a more general condition can be used .
the results of theorem [ th : munew ] hold with probability at least @xmath178 if instead of assumption ( a4 ) we require @xmath7 to satisfy : @xmath179 with probability at least @xmath180 , for some @xmath181 . compared to @xcite , the results in theorem [ th : munew ] exploit the zero mean condition on the noise matrix @xmath7 . as in @xcite ,
the estimator is consistent as @xmath25 goes to zero . in order to compare the rates in theorem [ th : munew ] with those for the mu selector
, we recall that , by theorem 3 in @xcite , the mu selector satisfies @xmath182 with probability close to 1 .
while both rates share some terms , a term of order @xmath183 appears only in the rate for the mu selector whereas a term of the order @xmath184 appears only for the @xmath118-mu selector .
therefore , the improvement upon the original mu selector is achieved whenever @xmath185 .
+ if the additional condition @xmath186 holds , we can use the bound and a better accuracy is achieved by the proposed estimator . in particular , @xmath187 no longer drives the rate of convergence .
the impact of @xmath25 on this rate is in the term @xmath188 for the mu selector .
furthermore , the rate of convergence of the new estimator also has a term of the form @xmath189 .
thus the new estimator obtains a better accuracy by exploiting additional assumptions together with the fact that @xmath190 is of larger order than @xmath191 , which holds whenever @xmath185 . finally , the impact of going down from the @xmath44-norm to the @xmath45- or @xmath1-norms is not negligible neither .
for example , if all non - zero components of @xmath9 are equal to the same constant @xmath192 , we have @xmath193 while @xmath194 , and @xmath195 .
then , the comparison in is reduces to comparing @xmath196 featuring the maximum contrast between the two rates .
+ finally , note that the conic estimator @xmath56 studied in @xcite can be also applied under the assumptions of this section .
indeed , our setting can be embedded into that of @xcite with @xmath57 being the identically zero @xmath58 matrix , which means that we have an estimator of each @xmath30 with an error bounded by @xmath197 .
the results in @xcite assume @xmath198 but they do not apply to designs with @xmath199 of larger order . comparing the bound in theorem [ th : munew ] to the bound yields that the conic estimator @xmath56 achieves the same rate as our new estimator whenever @xmath25 is smaller than or of the order @xmath60 . however , there is no bound for @xmath56 available when @xmath61 .
[ sec : conic ] in this section , we discuss a modification of the conic estimator proposed in @xcite .
we introduce an additional @xmath1-norm regularization to better adapt to the estimation error in @xmath57 . as discussed in the introduction , this is beneficial when the rate of convergence of @xmath57 to @xmath47 is slower than @xmath48 , which is not covered by @xcite . here
we consider the same assumptions as in @xcite with the only difference that now we allow for any rate of convergence of @xmath57 to @xmath47 .
thus , we replace assumption ( a4 ) by the following assumption on the availability of estimators for @xmath30 , @xmath62 . _
_ * there exist statistics @xmath29 and positive numbers @xmath200 such that for any @xmath201 , we have @xmath202\leq \varepsilon.\ ] ] in what follows , we fix @xmath203 and set @xmath204 we are particularly interested in cases where @xmath205 is of larger order than @xmath48 . to define the estimator , we consider the following minimization problem : @xmath206 here , @xmath123 and @xmath124 are tuning constants and the minimum is taken over @xmath125 .
let @xmath207 be a solution of ( [ conicnew ] ) .
we take @xmath120 as estimator of @xmath9 and we call it the @xmath118-compensated mu selector . the rates of convergence of this estimator are given in the next theorem .
[ th : conic ] let assumptions ( a1)-(a3 ) , and ( a5 ) hold .
assume that the true parameter @xmath27 is @xmath15-sparse and belongs to @xmath136 .
let @xmath103 and @xmath137 .
suppose also that @xmath208 for some constant @xmath94 and that @xmath209 for some small enough constant @xmath210 .
let @xmath120 be the @xmath118-compensated mu selector . then , with probability at least @xmath211 , latexmath:[\[\begin{aligned } \label{11a } for some constants @xmath142 and @xmath143 ( here we set @xmath144 ) .
+ under the same assumptions with @xmath148 , the prediction error admits the following bound , with the same probability : @xmath213 throughout the proof , we assume that we are on the event of probability at least @xmath214 where the results of lemmas [ lem3 ] , [ lem4 ] and [ lem5 ] in the appendix hold .
property ( [ c1 ] ) in lemma [ lem4 ] implies that @xmath215 is in the cone @xmath151 , where @xmath154 therefore , by the definition of the @xmath17-sensitivity and lemma [ lem5 ] , we have @xmath216 where @xmath217 and @xmath218 are of order @xmath219 , and @xmath220 and @xmath221 are of order @xmath222 . using again ( [ c1 ] ) , we have @xmath223 it follows that @xmath224 which implies , by , @xmath225 in view of the assumptions of the theorem .
recall that @xmath226 , where @xmath192 is a constant .
therefore , since we assume that @xmath227 , ( [ 11a ] ) follows if @xmath146 is small enough .
+ to prove ( [ 11c ] ) , we use .
remark that from ( [ 11a ] ) with @xmath148 , we have @xmath228 lemma [ lem5 ] in the appendix yields @xmath229 combining the above bound for @xmath230 and ( [ sb ] ) , we get @xmath231 since @xmath232 for some constant @xmath233 under our assumptions .
this proves ( [ 11c ] ) .
theorem [ th : conic ] generalizes the results in @xcite to estimators @xmath57 that converge with rate @xmath234 of larger order than @xmath48 . at the same time , if @xmath234 is smaller than @xmath48 , both the conic estimator @xmath56 of @xcite and the @xmath118-compensated mu selector achieve the same rate of convergence .
+ for such designs that condition ( [ hyps ] ) does not hold , the conclusions of theorem [ th : conic ] need to be slightly modified as shown in the next theorem .
+ [ th : conicrelaxed ] let assumptions ( a1)-(a3 ) , and ( a5 ) hold .
assume that the true parameter @xmath27 is @xmath15-sparse and belongs to @xmath136 .
let @xmath103 and @xmath137 .
let @xmath120 be the @xmath118-compensated mu selector . then , with probability at least @xmath211 , latexmath:[\[\begin{aligned }
\label{eq : th : conicrelaxed:1 } for some constants @xmath142 and @xmath143 , and the prediction error admits the following bound , with the same probability : @xmath236 again , throughout the proof , we assume that we are on the event of probability at least @xmath214 where the results of lemmas [ lem3 ] , [ lem4 ] and [ lem5 ] in the appendix hold . property ( [ c1 ] ) in lemma [ lem4 ] implies that @xmath215 is in the cone @xmath151 , where @xmath154 since @xmath237 we obtain @xmath238 therefore @xmath239 which implies .
note also that , due to , the above displays immediately imply the bound on the prediction risk given by the second term under the minimum in .
the first term under the minimum in is obtained by combining , with @xmath148 , and .
this section aims to illustrate the finite sample performance of the proposed estimators .
we will focus on the @xmath118-compensated mu selector only .
we consider the following data generating process @xmath240 here , @xmath241 are independent and @xmath242 , @xmath243 , @xmath244 where @xmath245 is the identity matrix and @xmath246 is @xmath247 matrix with elements @xmath248 .
we consider the vector of unknown parameters @xmath249 .
we set @xmath250 , @xmath251 , and @xmath252 .
we assume that @xmath253 is known and we set @xmath254 .
the penalty parameters are set as @xmath255 , @xmath256 , for @xmath257 .
+ in our first set of simulations , we illustrate the finite sample performance of the proposed estimator by setting @xmath258 .
the @xmath118-compensated mu selector will be denoted by @xmath118 .
we compare its performance with other recent proposals in the literature , namely the conic estimator ( denoted as conic ( @xmath43 ) for @xmath259 ) , and the compensated mu selector ( cmu ) .
we also provide the ( infeasible ) dantzig selector which knows @xmath6 ( dantzig x ) and the dantzig selector that uses only @xmath5 ( dantzig z ) as additional benchmark for the performance .
.simulation results for 100 replications . for each estimator
we provide average bias ( bias ) , average root - mean squared error ( rmse ) , and average prediction risk ( pr ) .
[ cols="<,^,^,^,^,^,^ " , ] tables [ table : mc3 ] and [ table : mc4 ] show the performance for different values of @xmath43 and @xmath260 .
we note that these parameters seem to have different impact on the finite sample performance even if @xmath261 is kept constant .
importantly , we observe that the addition of safeguard constraints virtually always leads to improvements although small ( even zero sometimes ) for most of the tested parameter values . in the case
@xmath262 , using safeguard constraints makes almost no difference and overall performance of both estimators is better .
in contrast , the estimators perform worse when @xmath263 and the safeguard constraints lead to improvements .
finally , as expected , the safeguard constraints improve substantially the performance when @xmath264 . in that case , the performance becomes comparable to that of the cmu estimator . essentially , the safeguard constraints help to avoid severe underfitting .
they are very helpful when the performance is below of what can be achieved .
nonetheless , we recommend to keep them in all cases as it does not impact negatively the estimator and the additional computational burden seems minimal .
in what follows , we write for brevity @xmath265 , @xmath266 , and we set @xmath215 , @xmath154 [ lem3 ] assume ( a1)-(a3 ) and ( a5 ) .
then with probability at least @xmath267 , the pair @xmath268 belongs to the feasible set of the minimization problem ( [ conicnew ] ) .
first , note that @xmath269 is equal to @xmath270 by definition of @xmath271 and @xmath199 , with probability at least @xmath272 , we have @xmath273 where in ( [ lem3.3 ] ) and ( [ lem3.4 ] ) we have used that the considered matrices are diagonal .
also , by lemma [ lem3a ] , with probability at least @xmath157 , we have @xmath274 combining the decomposition of @xmath269 together with ( [ lem3.2])-([ineg1 ] ) , we find that @xmath275 with probability at least @xmath267 , which implies the lemma .
@xmath276 + [ lem4 ] let @xmath277 be the @xmath118-compensated mu - selector .
assume ( a1)-(a3 ) and ( a5 ) .
then with probability at least @xmath267 ( on the same event as in lemma [ lem3 ] ) , we have @xmath278 set @xmath150 . on the event of lemma [ lem3 ]
, @xmath279 belongs to the feasible set of the minimization problem ( [ conic ] ) .
consequently , @xmath280 this implies @xmath281 and so @xmath282 and ( [ c1 ] ) follows . to prove ( [ c2 ] )
, it suffices to note that ( [ c3 ] ) implies @xmath283 and the result follows since @xmath284 and @xmath285 .
similar calculations yield the bound for @xmath286 .
[ lem5 ] let @xmath277 be the @xmath118-compensated mu - selector .
assume ( a1)-(a3 ) and ( a5 ) .
then , on a subset of the event of lemma [ lem3 ] having probability at least @xmath214 , we have @xmath287 where @xmath288 , @xmath289 , @xmath290 , @xmath291 .
note that @xmath217 and @xmath218 are of order @xmath219 , and @xmath220 and @xmath221 are of order @xmath222 . throughout the proof
, we assume that we are on the event of probability at least @xmath267 where inequalities ( [ lem3.2 ] ) ( [ ineg1 ] ) hold and @xmath292 belongs to the feasible set of the minimization problem ( [ conicnew ] ) .
we have @xmath293 using the fact that @xmath207 belongs to the feasible set of the minimization problem ( [ conic ] ) together with ( [ c2 ] ) , we obtain @xmath294 using that @xmath295 , assumption ( a5 ) together with ( [ lem3.4 ] ) yields that @xmath296 now remark that @xmath297 . in view of lemma [ lem3a ] and ( [ lem3.3 ] ) , on the initial event of probability at least @xmath298
, @xmath299 moreover , we have @xmath300 therefore , @xmath301 with probability at least @xmath211 ( since we intersect the initial event of probability at least @xmath298 with the event of probability at least @xmath302 where the bounds @xmath303 and @xmath304 hold for the corresponding terms ) . next , on the same event of probability at least @xmath211 , @xmath305 to complete the proof , it suffices to plug ( [ c9 ] ) ( [ c8 ] ) in the last inequality for @xmath306 and to obtain @xmath307
the work of the third author is supported by genes , and by the french national research agency ( anr ) as part of idex grant anr -11- idex-0003 - 02 , labex ecodec ( anr - 11-labex-0047 ) , and ipanema grant ( anr-13-bsh1 - 0004 - 02 ) .
e. gautier and a. b. tsybakov ( 2013 ) pivotal estimation in high - dimensional regression via linear programming . in : _ empirical inference
a festschrift in honor of vladimir n. vapnik _ ,
b.schlkopf , z. luo , v. vovk eds . , 195 - 204 .
springer , new york e.a .
m. rosenbaum and a.b .
tsybakov ( 2013 ) improved matrix uncertainty selector . in : _ from probability to statistics and back : high - dimensional models and processes a festschrift in honor of jon a. wellner _ ,
m.banerjee et al .
eds . _ ims collections _ , vol.9 , 276290 , institute of mathematical statistics . | several new estimation methods have been recently proposed for the linear regression model with observation error in the design .
different assumptions on the data generating process have motivated different estimators and analysis .
in particular , the literature considered ( 1 ) observation errors in the design uniformly bounded by some @xmath0 , and ( 2 ) zero mean independent observation errors . under the first assumption ,
the rates of convergence of the proposed estimators depend explicitly on @xmath0 , while the second assumption has been applied when an estimator for the second moment of the observational error is available .
this work proposes and studies two new estimators which , compared to other procedures for regression models with errors in the design , exploit an additional @xmath1-norm regularization .
the first estimator is applicable when both ( 1 ) and ( 2 ) hold but does not require an estimator for the second moment of the observational error .
the second estimator is applicable under ( 2 ) and requires an estimator for the second moment of the observation error .
importantly , we impose no assumption on the accuracy of this pilot estimator , in contrast to the previously known procedures .
as the recent proposals , we allow the number of covariates to be much larger than the sample size .
we establish the rates of convergence of the estimators and compare them with the bounds obtained for related estimators in the literature .
these comparisons show interesting insights on the interplay of the assumptions and the achievable rates of convergence . |
in a previous work @xcite , the quantum hamilton - jacobi equation [ 2 - 5 ] ( qhje ) was made use of to demonstrate the existence of particle trajectories in a complex space , for different quantum states .
this complex quantum trajectory representation was obtained by modifying the de broglie - bohm ( dbb ) approach to quantum mechanics @xcite , which allows particle motion guided by the wave function .
one of the advantages of the resulting theory , which offers a new interpretation of quantum mechanics , is that it does not face the problem of stationarity of particles in bound states , encountered in the dbb representation .
another trajectory approach to quantum mechanics , which also claims the absence of this problem , is the representation developed by floyd , faraggi , matone ( ffm ) and others [ 7 - 9 ] .
the new complex trajectory representation proceeds by first substituting @xmath1 in the schrodinger equation to obtain the quantum hamilton - jacobi equation @xmath2 = \frac{i\hbar}{2 m } \frac{\partial^2 \hat{s}}{\partial x^2 } , \label{eq : qhje}\ ] ] and then postulating an equation of motion @xmath3 for the particle .
the trajectories @xmath4 of the particle in the complex @xmath5-plane are obtained by integrating this equation with respect to time @xcite .
it was observed that the above identification @xmath1 helps to utilize all the information contained in @xmath6 while obtaining the trajectory .
( the dbb approach , which uses @xmath7 does not have this advantage . )
the complex eigentrajectories in the free particle , harmonic oscillator and potential step problems and complex trajectories for a wave packet solution were obtained in @xcite .
the representation was extended to three dimensional problems , such as the hydrogen atom , by yang @xcite and was used to investigate one dimensional scattering problems and bound state problems by chou and wyatt @xcite .
later , a complex trajectory approach for solving the qhje was developed by tannor and co - workers @xcite .
the qhje was derived independently by sanz and miret - artes @xcite , who also found the complex trajectory representation useful in better understanding the nonlocality in quantum mechanics @xcite .
it is well known that the qhje as given in eq .
( [ eq : qhje ] ) was used by many physicists such as wentzel , pauli and dirac , even during the time of inception of quantum mechanics @xcite . in a commendable work in 1982 ,
leacock and padget @xcite have used the qhje to obtain eigenvalues in many bound state problems , without actually having to solve the corresponding schrodinger equation .
however , there were no trajectories in their work and it was only in @xcite that the equation of motion ( [ eq : xdot ] ) explicitly solved and the complex trajectories of particles in any quantum state obtained and drawn , for the first time .
this paper also highlighted the interpretational value of the complex quantum trajectory representation , vis - a - vis the bohmian mechanics .
eventhough this formulation requires that the wavefunction is known from a separate calculation , it is generally conceded that it was the work in @xcite which provided a complex trajectory interpretation of quantum mechanics @xcite . instead of computing the complex trajectories
@xmath4 , the complex paths @xmath8 in the above scheme can directly be found by integrating the equation @xmath9 where eq .
( [ eq : xdot ] ) shall be used . in @xcite , it was noted that even for an eigenstate , the particle can be in any one of its infinitely many possible quantum trajectories , depending on its initial position in the complex plane .
therefore , the expectation values of dynamical variables are to be evaluated over an ensemble of particles in all possible trajectories . it was postulated that the average of a dynamical variable @xmath10 can be obtained using the measure @xmath11 as @xmath12 where the integral is to be taken along the real axis @xcite . also it was noted that in this form , there is no need to make the conventional operator replacements .
the above postulate is equivalent to the born s probability axiom for observables such as position , momentum , energy , etc . , and
one can show that @xmath13 coincides with the corresponding quantum mechanical expectation values .
this makes the new scheme equivalent to standard quantum mechanics when averages of dynamical variables are computed .
one of the challenges before this complex quantum trajectory representation , which is an ontological theory of particle motion , is to explain the quantum probability axiom . in the dbb approach
, there were several attempts to obtain the @xmath0 probability distribution from more fundamental assumptions @xcite . in the present paper , we first attempt to obtain this distribution along the real line from the velocity of particles in the complex trajectory representation .
it is found that always there exists a direct relationship between the @xmath0 distribution and the imaginary component of the particle s velocity on the real line . since this distribution
is defined and used only along the real axis , the conservation equation for probability in the standard quantum mechanics is valid here also , without any modifications . at the same time , since we have the complex paths , it would be natural to consider the probability for the particle to be in a particular path .
in addition , we may consider the probability to find the particle around different points in the same path , which can also be different .
thus it is desirable to extend the probability axiom to the @xmath14-plane .
but in this case , it becomes necessary to see whether probability conservation holds everywhere in the plane . a recent paper by poirier @xcite addresses this issue and obtains some negative results for the choices made for such a distribution .
poirier tries to define an extended probability by working with @xmath6 , the solution of time - dependent schrodinger equation .
the probability density at a point @xmath5 in the complex plane is postulated as @xmath15 where @xmath16 .
the complexified flux is chosen as @xmath17 where @xmath18 is given by equation ( [ eq : xdot ] ) and prime denotes spatial differentiation .
with the help of time - dependent schrodinger equation , the author shows that , in general , @xmath19 this , arguably leads to nonconservation of probability along trajectories .
but it shall be reminded that this negative result is based on the choices made in @xcite for the probability density and flux . in the present paper , on the other hand , we first show that the born s probability density , defined along @xmath20 , is obtainable as the exponential of an integral over the real line , of @xmath21 , the imaginary part of the particle s complex velocity .
next , extending this probability axiom to the entire @xmath14-plane , we attempt to solve the appropriate equation for a conserved probability density . it is also demanded that this quantity should agree with born s probability rule along the real axis , which is the boundary condition in this case .
an important result obtained is that the characteristic curves of the conservation equation , along which the information about the solution propagates , are the same as the paths of particle in the complex trajectory representation .
we find that there exist conserved probability densities which agree with the boundary condition in most of the examples considered .
an alternative method to evaluate this in terms of a trajectory integral , in a manner similar to obtaining the @xmath0 probability distribution along the real line , is also presented .
this latter method shall be of interest while solving the qhje for the motion of individual particles .
most importantly , we show that the above conservation equation can be derived from this definition of probability , by using the complex extension of schrodinger equation .
let us recall that in this scheme , the real part of the velocity of a particle on the real line , denoted as @xmath22 , always agrees with the velocity of particle in the dbb representation @xcite . to obtain the quantum probability function from the velocities , which is our first goal , we note further that the imaginary component @xmath21 of the velocity of the particle in the complex trajectory can be written as @xmath23 } { \psi ^{\star } \psi}. \label{eq : xidot}\ ] ] this helps to write the probability density to find the particle around some point @xmath20 as @xmath24 where the integral is taken along the real axis .
this possibility of regaining the quantum probability distribution from the velocity field is a unique feature of the complex trajectory formulation .
for instance , in the dbb approach , the velocity fields for all bound eigenstates are zero everywhere and it is not possible to obtain a relation between velocity and probability . the ffm trajectory representation , on the other hand , does not claim any connection with probability . in the following ,
we limit ourselves to one dimension and to time - independent problems .
the variable @xmath5 is always assumed to be complex . as stated earlier , in this new representation , which is based on the existence of trajectories in the complex plane
, even a particle in an eigenstate can be in any one of the infinitely many possible trajectories , depending on its initial position in the complex plane .
thus it is also desirable to look for the probability of a particle to be in an area @xmath25 around some point ( @xmath26 ) in the complex plane .
let this quantity be denoted as @xmath27 .
an explicit expression for @xmath28 can be arrived at in the following way . extending the probability density to the entire @xmath14-plane demands a conservation equation of the form ( for time - independent cases ) @xmath29 to solve this partial differential equation , we may write @xmath30 , where @xmath31 is some solution of eq .
( [ eq : cons ] ) and @xmath32 is some combination of @xmath33 and @xmath34 , whose value remains a constant along its characteristic curves @xcite . substituting this form of @xmath35 into ( [ eq : cons ] )
, we see that the characteristic curves are obtained by integrating the equation @xmath36 which is found to be the same as eq .
( [ eq : path ] ) .
this demonstrates the important property that the characteristic curves for the above conservation equation are identical to the complex paths of particles in the present quantum trajectory representation .
we may now find the exact form of @xmath37 by requiring that @xmath38 agrees with the probability @xmath39 , which is the boundary condition in this case .
let the integration constant in the above equation ( [ eq : diff ] ) be the ( real ) coordinate of any one point of crossing of the trajectory on the real axis , denoted as @xmath40 .
since the characteristic curves are identical to the complex paths , one can take @xmath40 as the constant @xmath32 along the characteristic curve and let it be expressed in terms of @xmath33 and @xmath34 .
the assumed form for the extended probability distribution @xmath35 may then be written as @xmath41 now we can choose @xmath42 subject to the boundary condition .
at the point @xmath43 at which the curve @xmath44 crosses the real line , we demand ( the boundary condition ) @xmath45 and obtain @xmath42 . expressing @xmath40 in terms of @xmath33 and @xmath34 in @xmath42 , eq .
( [ eq : rhogen ] ) gives @xmath28 .
a word of caution is appropriate here .
there may be instances , as we shall see below , when the boundary condition overdetermines the problem and we are unable to find a solution .
it is observed that this happens in certain regions of the complex space where the trajectories were previously suspected to be nonviable .
we may also note that the differential in eq .
( [ eq : path ] ) , given by @xmath46 is inexact because @xmath47 unless both the partial derivatives are zero .
this is due to the cauchy - riemann conditions satisfied by the analytic function @xmath48 in the complex @xmath5-plane ( except at its singular points ) . however , this inexact differential can always be made exact by multiplying with an integrating factor @xmath49 , which obeys @xmath50 thus we see that the integrating factor @xmath51 can serve as @xmath31 in eq .
( [ eq : rhogen ] ) .
the following observation may be helpful in finding @xmath31 , or even @xmath28 , in the case of some special potentials .
let us denote @xmath52 .
then using the time - independent schrodinger equation , one can rewrite the equation of motion ( [ eq : xdot ] ) as @xmath53 where @xmath54 . for @xmath55 , a real constant
, one can integrate this to obtain @xmath56,\ ] ] from which we get the trajectory of the particle in the complex @xmath5-plane as @xmath57 .
moreover , one can see that @xmath58 but for constant potentials , @xmath59 in the complex plane .
it can be seen that the above expression ( [ eq : chisq ] ) satisfies the conservation equation and agrees with the boundary condition . in other words , the extended , conserved probability @xmath28 in this case can be written as @xmath60 and hence @xmath35 varies inversely as @xmath61 , as the particle moves along a particular trajectory with fixed @xmath62 .
this result is not contradictory to the wkb result that @xmath63 for constant potentials @xcite , since @xmath64 along a trajectory in our case is not the same as @xmath65 . in the case of harmonic oscillator potential
, we shall see below that an expression of the form @xmath66 will give the solution @xmath31 of the conservation equation , but for the particular solution which agrees with the boundary condition , one need to find @xmath42 too .
let us now first apply the procedure to the harmonic oscillator . in the @xmath67 case , we have @xmath68 and the quantum probability density is regained by using eq .
( [ eq : psistarpsi ] ) as @xmath69 to find @xmath28 , first let us note that @xmath70 , a constant , can be a solution of the conservation equation ( [ eq : cons ] ) in this case . then we find @xmath42 using eq.([eq : fxr0 ] ) as proportional to @xmath71 . to put @xmath40 in terms of @xmath33 and @xmath34 , eq .
( [ eq : path ] ) shall be integrated to obtain the paths as @xmath72 @xcite . equating this constant to @xmath73 ,
we can finally write @xmath74 this conserved probability is plotted in fig . 1
. similarly in the @xmath75 eigenstate of this case , one uses eq .
( [ eq : psistarpsi ] ) to regain @xmath76 . in the next step
, @xmath77 is found to be a solution of the conservation equation ( [ eq : cons ] ) .
( one could guess this expression from the form of @xmath78 in this case , as it appears in its denominator . )
@xmath42 can now be found as @xmath79 the complex paths in this case are given by @xmath80 , a constant .
( note that also this appears in the expression for @xmath78 .
explicitly , one obtains @xmath81 ) . equating this constant to @xmath82
, one can obtain @xmath42 and also the extended probability density @xmath35 , in terms of @xmath33 and @xmath34 .
but while taking square roots , one need to be careful . it shall be noted that for the region containing the subnests in the harmonic oscillator ( with @xmath83 in the present @xmath75 case ) @xcite , the boundary condition overdetermines the problem , resulting in there being no solution .
but for the region outside it , one can write @xmath84 and therefore , @xmath85 which is plotted in fig .
2 . now let us consider another example of a particle in an infinite square well potential , between @xmath86 and @xmath87 .
using the energy eigenfuntion @xmath88 , one can find that the complex paths are given by @xmath89 we can choose @xmath90 as a solution of the conservation equation .
( here again , we have @xmath91 . )
interestingly , here , we have @xmath92 and hence , @xmath93 the @xmath75 case of this distribution is plotted in fig .
a similar situation arises in the case of complex trajectories in the potential step problem discussed in @xcite .
again , we find @xmath91 and @xmath94 .
for the reflection constant equal to @xmath95 , we thus get , for the potential step , @xmath96 which is plotted in fig .
it may now appear that the above two approaches to probability , namely , the one which gives the born s @xmath0 probability along the real line using equation ( [ eq : psistarpsi ] ) and the other , which gives a conserved probability density over the entire @xmath14-plane by solving a conservation equation , do not have anything in common .
but in this section , we show that there is a more elegant way of obtaining @xmath28 , by uniting them .
this method may also be useful in finding the variation of probability along any trajectory .
first , we postulate that if @xmath97 , the extended probability density at some point @xmath98 is given , then @xmath28 at another point that lies on the trajectory which passes through @xmath98 , is @xmath99 .
\label{eq : rho_def}\ ] ] here , the integral is taken along the trajectory @xmath100 $ ] .
one can derive the continuity equation by using the extended version of the schrodinger equation , which gives @xmath101 since energy and time are assumed real @xcite .
this helps to write the above definition ( [ eq : rho_def ] ) as @xmath102 which in turn gives @xmath103 the last step follows from the analyticity of @xmath48 .
this leads to the continuity equation for the particle , as it moves along : i.e. , @xmath104 while evaluating @xmath35 with the help of ( [ eq : rho_def ] ) above , one needs to know @xmath97 at @xmath98 and if we choose this point as @xmath105 , the point of crossing of the trajectory on the real line , then @xmath97 may take the value @xmath106 and may be found using ( [ eq : psistarpsi ] ) . here
it shall be reminded that the integral in @xmath106 is evaluated over the real line only . to summarize the alternative method
, we note that the conserved , extended probability density is @xmath107 , \label{eq : rho_alt}\ ] ] with the integral in the first factor evaluated over the real line and that in the second factor over the trajectory of the particle .
this method is easier to implement numerically on a computer .
we can now demonstrate this trajectory integral method by applying it to the @xmath75 harmonic oscillator state . in this case
, taking the numerical values of both @xmath108 and @xmath109 to be unity ( which then also applies to @xmath110 ) , it is easily seen that @xcite @xmath111 and @xmath112 here one can write @xmath113 .
finding @xmath33 and @xmath34 using the above expressions , we have evaluated numerically @xmath114 for different @xmath62 and @xmath115 .
this was found to be the same extended distribution as that given by eq .
( [ eq : n1prob ] ) and shown in fig .
2 , with @xmath116 . more specifically , in this case , since equations ( [ eq : rhogen ] ) and ( [ eq : rho_alt ] ) must agree , the surface plots of @xmath77 and the expression @xmath117 with various values of @xmath62 , @xmath115 but plotted against @xmath33 and @xmath34 evaluated using equations ( [ eq : n1xr ] ) and ( [ eq : n1xi ] ) , must be the same .
this too is found to be true , which demonstrates that the two methods of evaluating the probability give identical results .
the complex quantum trajectory representation is worth pursuing mainly for the alternative interpretation it offers to standard quantum mechanics .
the standard quantum theory is spectacularly successful in explaining all observations made so far and there is no demand for modifications to the quantum probability axiom .
it is envisaged that the new complex trajectory representation does not differ from standard quantum mechanics , with regard to its predictions of experimental results .
this is ensured by accepting the born s probability axiom as such in the new theory , on the real line . the quantum probability density , though it is not claimed to be perfectly understood , works well in a miraculous manner . in this paper
, we obtain this distribution along the real line from the velocity field of particles in the complex trajectory representation .
its conservation on the real line is guaranteed in standard quantum mechanics by virtue of the schrodinger equation and there is no need for any separate proof for the same in the complex trajectory representation .
what we try to do next is to extend it to the complex plane , where it should obey conservation laws , as in the case of any other probability function .
our great expectation is to find such a probability density in the complex plane , in the hope of it giving us a clue to the basis of this distribution .
here we have first adopted the strategy of solving the conservation equation in the @xmath14-plane , with the velocity field given by eq .
( [ eq : xdot ] ) , applicable to one dimensional , time - independent single particle states . in this endeavor
, there is a significant improvement in our understanding of the origin of quantum probability .
first , we have noticed that the characteristic curves of the conservation equation , along which the information about the solution propagates , are the same as the paths of particle in the complex trajectory representation .
the most notable results we have obtained are that an extended , conserved probability @xmath28 , which agrees with the quantum probability rule along the real line , can exist and that it can be derived from the velocity field of particles .
there are parts of the plane in some examples where such a probability can not be found .
but in those cases where it is possible , we find that @xmath28 can be written as a product of two factors ; one [ @xmath42 ] , a constant for the given complex path and the other [ @xmath31 ] , dependent on the velocity of the particle as it moves along this path . in particular , we note that the born s probability rule @xmath120 itself , which is valid on the real line , can be written as a product as stated above . on the other hand , in an alternative method of evaluating the extended probability ,
@xmath28 is found to be obtainable as @xmath106 times an exponential factor which involves an integral with respect to time along the trajectory , of the imaginary part of the sum of the complex kinetic and potential energies of the particle .
the two methods of obtaining the extended probability density were compared and found to give identical results .
the latter trajectory integral method is easier to implement using a computer and may be particularly useful while solving the qhje for single particles .
the solutions we obtain for the complex plane are not the same as @xmath0 , except in the case of those constant potentials , discussed in sec .
it shall be noted that @xmath35 is defined only over that region of the @xmath14-plane , where characteristic curves which cross the real axis pass . at all other points , @xmath35
is assumed to be zero .
the extended probability may be normalized on this basis .
also the present representation offers the viewpoint that a particle is observed only when it surfaces on the real axis .
@xmath39 is the probability for such appearances and shall be used for computations involving measurable quantities . in this way , all the standard results in quantum mechanics continue to be unaffected .
most importantly , we have seen that the definition ( [ eq : rho_def ] ) of the extended probability density leads to the continuity equation ( [ eq : cons_gen ] ) , while using the complex extension of schrodinger equation .
this definition of @xmath35 can thus be of general validity .
but here , @xmath35 is a real function of @xmath33 and @xmath34 , which does not obey the cauchy - riemann equations , and hence is not analytic in the complex @xmath5-plane .
it may be noted that in our attempt to have a better understanding of the origin of quantum probability rule , where we could make some inroads , this nonanalyticity does not appear as a problem .
@xcite strongly advocates an analytic expression for @xmath35 , in view of its anticipated advantages in the synthetic time - dependent schrodinger equation applications .
but the definition in @xcite is shown not to lead to a continuity equation and it is presented there as a negative result . whether our definition of extended probability in this paper , which does obey a continuity equation , is helpful for synthetic applications is an open problem , and is worth pursuing . to conclude
, we summarize the positive and negative features of this possible definition of quantum probability density in the complex space , in comparison with the distribution prescribed in @xcite .
first , we see that our definition helps to obtain the born s @xmath0 probability density along the real line from the velocity field in the complex trajectory formalism , in a nontrivial way . on the other hand ,
the distribution in @xcite is simply the analytic continuation of @xmath0 and hence the agreement is trivial .
second , our definition of the probability density for the particle to be around some point in its trajectory , as it moves along , obeys a conservation equation .
in contrast , it was shown in @xcite that the probability density defined in it is not conserved , in general . furthermore , while solving the conservation equation , we have noted that @xmath35 can be written as a product between two factors , one of which is a constant for a trajectory .
this feature was deduced while attempting to solve the conservation equation and is a consequence of the fact that trajectories and the characteristic curves of the conservation equation are identical in the complex trajectory representation .
this has no parallels in the proposed probability density in @xcite .
a negative aspect of the present formalism is that the probability density is not an analytic function in the complex plane , a feature which is suspected to adversely affect its applicability in the synthetic solutions .
this is another feature that distinguishes the present distribution from the analytic one studied in @xcite .
it is noted that in our attempt to understand the origin of quantum probability rule in complex trajectory representation , this nonanalyticity does not pose any problem .
but the complex probability that appears in @xcite off of the real axis would be an undesirable feature , and could have posed some problems even if it was found to obey the continuity equation
. thus the issue of whether analyticity or real - valuedness of probability density is the more desirable feature in synthetic applications can be settled only when one progresses with such applications of the present formalism .
g. wentzel , z. phys .
38 ( 1926 ) 518 ; w. pauli , in : h. geiger , k. scheel ( ed . ) , handbuch der physik , 2nd ed .
24 , part 1 , springer - verlag , berlin , 1933 , pp .
83 - 272 ; p.a.m .
dirac , the principles of quantum mechanics , oxford univer- sity press , london , 1958 .
floyd , phys .
d 26 ( 1982 ) 1339 ; e.r .
floyd , phys .
d 25 ( 1982 ) 1547 ; e.r .
floyd , phys .
d 29 ( 1984 ) 1842 ; e.r .
floyd , phys .
d 34 ( 1986 ) 3246 ; e.r .
floyd , found .
phys . lett . 9 ( 1996 ) 489 ; e.r .
floyd , found .
13 ( 2000 ) 235 ; e.r .
floyd , int .
phys . a 14 ( 1999 ) 1111 .
a. faraggi , m. matone , phys .
b 450 ( 1999 ) 34 ; a. faraggi , m. matone , phys . lett .
b 437 ( 1998 ) 369 ; a. faraggi , m. matone , phys . lett .
b 445 ( 1999 ) 77 ; a. faraggi , m. matone , int . j. mod . phys .
a 15 ( 2000 ) 1869 . | it is shown that in the complex trajectory representation of quantum mechanics , the born s @xmath0 probability density can be obtained from the imaginary part of the velocity field of particles on the real axis . extending this probability axiom to the complex plane , we first attempt to find a probability density by solving an appropriate conservation equation .
the characteristic curves of this conservation equation are found to be the same as the complex paths of particles in the new representation .
the boundary condition in this case is that the extended probability density should agree with the quantum probability rule along the real line . for the simple , time - independent
, one - dimensional problems worked out here , we find that a conserved probability density can be derived from the velocity field of particles , except in regions where the trajectories were previously suspected to be nonviable .
an alternative method to find this probability density in terms of a trajectory integral , which is easier to implement on a computer and useful for single particle solutions , is also presented .
most importantly , we show , by using the complex extension of schrodinger equation , that the desired conservation equation can be derived from this definition of probability density . quantum hamilton - jacobi equation , trajectory representation , probability axiom , complex methods 03.65.ca |
we acknowledge partial support of fondecyt ( chile ) projects 1060627 and 1060651 , conicyt / pbct proyecto anillo de investigacin en ciencia y tecnologa act30/2006 and u.s .
national science foundation grant dms 06 - 00037 . | we establish rigorous upper and lower bounds for the speed of pulled fronts with a cutoff .
we show that the brunet - derrida formula corresponds to the leading order expansion in the cut - off parameter of both the upper and lower bounds . for
sufficiently large cut - off parameter the brunet - derrida formula lies outside the allowed band determined from the bounds . if nonlinearities are neglected the upper and lower bounds coincide and are the exact linear speed for all values of the cut - off parameter .
the reaction diffusion equation @xmath0 provides a simple description of phenomena in fields such as population dynamics , chemical reactions , flame propagation , fluids , qcd , among others @xcite .
it is one of the simplest models which shows how a small perturbation to an unstable state develops into a moving front joining a stable to an unstable state .
the reaction term @xmath1 satisfies different conditions depending on the physical problem of interest .
one of the first , and most studied cases , is the fisher reaction term @xmath2 for which the asymptotic speed of the propagating front is @xmath3 , a value determined from linear considerations .
a more general case was studied by kolmogorov , petrovskii and piscounov ( kpp)@xcite who showed that for all reaction terms which satisfy the kpp condition @xmath4 the asymptotic speed of the front joining the stable @xmath5 point to the unstable @xmath6 point is given by @xmath7 these fronts are called pulled since it is the leading edge of the front which determines the velocity of propagation . in the rest of this work
we assume that @xmath8 .
the evolution of localized initial conditions for general reaction terms , and rigorous properties of the fronts were studied by aronson and weinberger @xcite .
the asymptotic speed of the front for all reaction terms can be found from the integral variational principle @xcite @xmath9 where the supremum is taken over all positive monotonic decreasing functions @xmath10 for which the integrals exist and where @xmath11 .
the supremum is always attained for reaction terms which are not pulled .
two effects not included in the classical reaction diffusion equation ( [ rd ] ) , are the effect of noise and the effect of a finite number @xmath12 of diffusive particles .
it was shown by brunet and derrida that such effects can be simulated by introducing a cut - off in the reaction term . in the case of noise
the cut - off parameter measures the amplitude of the noise while in the case of finite number of @xmath12 diffusing particles the cut - off parameter @xmath13 .
there is substantial numerical evidence that introducing a cut - off in the reaction terms reproduces accurately the effect of noise and finiteness in the number of diffusing particles @xcite . by means of an asymptotic matching brunet and derrida showed that for a reaction term @xmath14 a small cut - off changes the speed of the front to @xmath15 in recent work it has been show that the brunet - derrida formula for the speed is correct to @xmath16@xmath17 for a wider class of reaction terms @xcite .
the purpose of this work is to show that for reaction terms of the form @xmath18 where @xmath19 satisfies the kpp condition eq .
( [ kppcondition ] ) and @xmath20 is the step function , the speed @xmath21 of the front with the cutoff satisfies @xmath22 with @xmath23 we see that for @xmath24 , @xmath25 .
the function @xmath26 depends on the nonlinear terms of the reaction function . for small @xmath27 the series expansion of the upper bound @xmath28
is @xmath29 the contribution of the nonlinearities , contained in the term @xmath26 , appears at @xmath16@xmath17 , so that the leading order terms in the expansion of the upper and lower bounds give the brunet - derrida formula .
if nonlinearities are neglected the value @xmath30 is the analog of the kpp value @xmath3 for reaction terms which satisfy the kpp condition , but with a cutoff . in what follows we derive the bounds and apply them to the fisher reaction term @xcite @xmath31 and to the reaction term studied by brunet and derrida @xmath32 .
the main tool to obtain the bounds is the variational principle for the speed . as shown in previous work @xcite
, we may perform the change variables @xmath33 where @xmath34 in eq.([vp1 ] ) and write the variational expression for the speed as @xmath35 where @xmath36 is an arbitrary parameter , @xmath37 and the supremum is taken over positive increasing functions @xmath38 such that @xmath39 , @xmath40 and for which all the integrals in ( [ newvp ] ) are finite .
therefore , for any suitable trial function @xmath38 we know that @xmath41 consider now reaction terms @xmath1 with a cut - off @xmath27 of the form @xmath42 where @xmath43 , the nonlinearity , is such that @xmath44 .
we find @xmath45 where @xmath46 assume now that @xmath1 satisfies the kpp criterion eq.([kppcondition ] ) .
since @xmath47 , it follows that @xmath48 where @xmath49 and therefore @xmath50 \equiv \sup_{u(s ) } 2 \,\frac { g(1)/s_0 + \int_0^{s_0 } g(u(s))/s^2 d\,s}{\int_0^{s_0 } \left ( d u /d s\right)^2 d\,s}.\ ] ] one can prove ( rigorous details will be given elsewhere ) that @xmath51 is bounded above and that there exists a function @xmath52 for which the supremum is attained .
this function is the monotonic increasing solution to the euler - lagrange equation for @xmath51 satisfying the boundary conditions @xmath53 .
one can also prove that the variational parameter @xmath54 is finite and @xmath55 .
in summary , the maximizing function for @xmath51 is the solution of @xmath56 @xmath57 subject to the boundary conditions @xmath58 with the function and its derivative continuous at @xmath59 .
the solution to this problem is given by @xmath60 with @xmath61 where @xmath62 is the first positive solution of @xmath63 the maximum of @xmath64 $ ] can be calculated easily .
we obtain after performing the integrals , @xmath65 = 4 \sin^2 ( \phi _ * ) \equiv c^2_{up}. \label{top}\ ] ] to obtain the lower bound we shall use the optimizing function @xmath52 as a a suitable trial function in eq .
( [ lowerbound ] ) .
we obtain @xmath66 \label{bottom}\ ] ] since @xmath67 is negative , we may combine eqs.([top ] ) and ( [ bottom ] ) and write our main result as given in eq .
( [ main ] ) . as an example
consider the reaction term studied by brunet and derrida , @xmath32 .
the lower bound can be written explicitly as @xmath68 the integral has a long analytic expression which we omit here . from the explicit expression above it
is not difficult to show that the contribution of the two last terms , which arise from the nonlinear terms , are of @xmath69 . in figures 1 and 2
we show the bounds together with the brunet - derrida formula as a function of @xmath27 .
the solid lines correspond to the upper and lower bounds .
the dashed line is the brunet - derrida formula . .
the solid lines correspond to the bounds , the dots to the brunet - derrida formula .
, height=188 ] as a second example we consider the fisher reaction term @xmath70 with a cut - off .
the lower bound becomes @xmath71 again , the integral can be done analytically and we do not show it here . in fig .
3 we show the upper and lower bounds and the brunet - derrida formula . in this case
the brunet - derrida formula leaves the allowed band at larger value of @xmath27 . in general for reaction terms @xmath72 , the gap between the upper and lower bounds becomes narrower and the brunet - derrida formula valid for a smaller range of @xmath27 . .
lines as in fig . 1 ,
height=188 ] in summary , we have studied the effect of a cut - off on reaction terms which satisfy the kpp condition eq.([kppcondition ] ) .
we have found upper and lower bounds valid for all values of the cut - off parameters , which allow to assess the accuracy of the brunet derrida formula .
if we consider only the linear terms , the upper and lower bounds coincide and give the exact linear value for the speed , of which the two leading order terms are the brunet - derrida formula . |
it is now widely accepted that folding of small single - domain proteins follows `` nucleation - condensation '' mechanism @xcite whereby relatively small fragment of protein structure is formed in the transition state between unfolded and folded states .
residues belonging to this fragment constitute specific folding nucleus ( sfn ) .
considerable experimental @xcite and theoretical @xcite effort has been devoted to identification of folding nuclei in real proteins and various models as well as factors that determine its location in structure and in sequence .
one of the most intriguing aspect of nucleation - condensation mechanism of protein folding is its relation to protein evolution .
indeed residues constituting folding nucleus can be metaphorically considered `` accelerator pedals '' of folding @xcite since mutations in those positions affect folding rate to a much greater extent than elsewhere in a protein .
one can conclude that if there is evolutionary control of folding rate it should have resulted in additional pressure applied on folding nucleus residues , and such pressure can be manifested in noticeable additional conservation of nucleus residues .
this idea was first proposed in @xcite where it was applied to prediction of nucleus residues from protein structure .
many sequences were designed to fit the structure of chymotripsin inhibitor 2 ( ci2 ) with low energy .
positions conserved among the designed sequences were identified as a putative nucleus . this way _ blind _ predictions of folding nucleus in ci2 were made that were verified in independent experiments @xcite
. in related studies papers ptitsyn studied conservatism in distant yet related by sequence homology members of cytochrome c @xcite and myoglobin @xcite families . in both cases
he found conserved clusters of residues without an obvious functional role which he suggested to belong to folding nucleus of those proteins .
michnick and shakhnovich @xcite carried out an analysis of conservation in natural and designed sequences for families of three structurally related proteins - ubiquitin , raf and ferredoxin and predicted possible folding nucleus for those proteins .
neverteheless the notion of folding nucleus conservation has drawn some controvercy in the lietrature . while earlier papers @xcite suggested conservation of folding nucleus in some proteins , a more recent paper by plaxco and coauthors @xcite argued to the opposite .
these authors looked at conservatism profile in several protein families for which protein engineering analysis of folding transition states has been carried out , and did not observe correlation between conservation and experimentally measured @xmath0-values .
this made them conclude that there is no evolutionary pressure to control the folding rates . in this work
we study evolutionary conservation of the folding nucleus for several homologous proteins .
conservation of the folding nucleus is systematically compared with the conservation in the rest of the protein sequence .
in contrast to previous studies , we perform rigorous statistical test to assess significance of higher conservation in the folding nucleus .
the main result of this study is that for all studied proteins , except acp , folding nucleus is significantly more conserved than the rest of the protein .
we explain the difference between our thorough statistical analysis and that of plaxco et al @xcite by pointing out to some technical shortcomings in the earlier work @xcite .
to study evolutionary conservation of the folding nucleus we turn to nine proteins for which nucleus has been experimentally identified from protein engineering analysis : ci2 , fkbp12 , acbp , chey , tenascin , cd2.d1 , u1a , acp and ada2h .
for each of them we obtain a multiple sequence alignment from hssp database @xcite ( or pfam @xcite database if hssp contains too few sequences ) .
we compute variability at position @xmath1 of the alignment as @xmath2 where @xmath3 is the frequency of residues from class @xmath4 in position @xmath1 .
we use six classes of residues to reflect physical - chemical properties of amino acids and their natural pattern of substitutions : aliphatic [ a v l i m c ] , aromatic [ f w y h ] , polar [ s t n q ] , basic [ k r ] , acidic [ d e ] , and special ( reflecting their special conformational properties ) [ g p ] . as a result of this classification mutations within a class are ignored ( e.g. @xmath5 ) , while mutations that change the class are taken into account .
figure 1 presents variability profile for studied proteins with nucleation positions marked by filled circles .
importantly , we defined the folding nucleus as it was identified by the original experimental groups ( table 1 ) .
figure 2 clearly shows that nucleus residues are almost always among the most conserved ones for all studied proteins .
it also shows that nucleus residues are not the only conserved ones : many other residues ( predominantly in the cores of the proteins ) are also conserved . in order to evaluate statistical significance of nucleus conservation we compare evolutionary conservation of the folding nucleus with the conservation of all residues in the protein using the following statistical test .
we start from the null hypothesis h0 that nucleus residues are * no * more conserved than the whole protein sequence . to test this hypothesis
we compute median variability of the nucleus residues ( @xmath6 $ ] ) and compare it with the distribution of medians variability of the same number of residues randomly chosen in the same protein ( @xmath7)$ ] ) .
the distribution @xmath7)$ ] is obtained by choosing @xmath8 random sets of @xmath9 residues ( @xmath9 is the number of residues in the nucleus ) .
then the fraction of instances with @xmath10<{\mbox{med}}[s_{nuc}]$ ] gives the probability @xmath11 of accepting h0 .
in other words , @xmath11 is the probability that observed lower variability of the folding nucleus is obtained by chance .
hence , @xmath12 indicates statistically significant strong evolutionary conservation of the folding nucleus .
below we use confidence level @xmath13 .
table 2 presents computed @xmath11 values . _
the main result of this work is that in all proteins , except acp , residues in the folding nucleus are significantly more conserved than the rest of the protein . _
next we study how obtained results depend on the way amino acids are grouped into classes ( see table 2 ) . when classification scheme from @xcite ( bt ) is used , still all proteins except acp exhibit significant conservation of the folding nucleus .
this clearly demonstrates that observed conservation of the folding nucleus is not a consequence of a particular choice of the classification scheme . however , when amino acids are _ not _ grouped into classes , nucleus exhibits significant conservation only in four out of nine proteins .
taken together these results indicate that substitutions in the folding nucleus may occur , but they are limited to residues that belong to the same class ( i.e. have similar physical - chemical properties @xcite ) . to study what physical - chemical properties are conserved in the folding nucleus we used various classification schemes . starting from all 20 amino acids , we grouped some of them into classes and repeated the analysis , including the statistical tests ( see table 2 ) . the goal is to find a _
minimal _ classification ( i.e. grouping the _ minimal _ number of amino acids together ) that provides statistically significant conservation of the folding nucleus .
our results show that classification where only i , l , and v are grouped in one class while all other amino acids each represent their own class satisfies this requirement ( see table 2).this classification provides significant conservation of the nucleus for all proteins except acp with @xmath14 , and for all proteins except acp and fkbp12 with @xmath13 .
this result demonstrates that @xmath15 are the most common substitutions in the nucleus ( and in the protein core in general @xcite ) .
these substitutions are tolerated in the nucleus as they do not change much neither stability of the native fold nor the folding rate .
analysis of available experimental data ( l.li unpublished ) shows that changes in stability upon @xmath16 mutations are in average @xmath17kcal mol@xmath18 for the native state and @xmath19 kcal mol@xmath18 for the transition state . note that grouping of residues into classes to assess conservation is similar to the use of substitution matrices in sequence alignment techniques . the underlying idea for both methods is to take into account natural physical - chemical similarity between amino acids and their substitution patterns .
plaxco et all used all 20 types of amino acids and failed to identify strong conservation of the folding nucleus @xcite .
similarly , a method that relies on simple sequence identity can not detect distant homology .
however distant homology between sequences can be detected using proper substitution matrices @xcite .
the use of substitution matrices is physically meaningful since they weight , e.g. , @xmath20 match higher then @xmath21 , while a method that relays on percentage of sequence identity weights @xmath20 and @xmath21 equally .
likewise , our amino acid classification scheme does not count @xmath22 as a mutation , while it certainly considers substitutions like @xmath23 as mutations to be counted .
although , on average , nucleus is more conserved , than the rest of the protein , not all nucleating residues are strongly conserved .
for example , in chey two out of ten nucleation residues are not conserved . in ada2h two out of five and in tenascin one out
of four residues are not conserved .
some nucleus residues may be less conserved because they belong to `` extended nucleus '' @xcite or because of limitation of our residues classification scheme that puts aromatic and aliphatic residues into two different groups , while aromatic - aliphatic substitutions may occur in the core of some proteins ( i.e. tenascin , ada2h ) usually as a result of correlated mutations that are not treated properly in this approach ( but are taken into account in the conservation - of - conservation approach @xcite ) .
another interesting observation is that the only protein that exhibits no preferential conservation of the folding nucleus is acp , which is the slowest folding protein among all studied two - state folding proteins ( @xmath24 ) .
perhaps , this protein did not undergo evolutionary selection for faster folding and hence its folding nucleus is under no additional pressure to be conserved .
note that , as expected , several other residues in studied proteins are as conserved as the nucleating ones .
( see fig.2 ) those are the residues of the active site , core hydrophobic residues responsible for stabilization of the native structure and others .
this suggests that although folding nucleus is conserved it can not be uniquely identified just by analysis of _ a single _ protein family as a pattern of conservation is dominated by residues conserved for protein stability and function ( see @xcite ) .
thus a consistent analysis should discriminate between residues that are conserved for functional reasons , for stability reasons and for kinetic reasons ( folding nucleus ) , like it was done in a more detailed conservation - of - conservation analysis in @xcite .
why do results of our analysis differ from those of plaxco et al @xcite ?
first , we took into account physical - chemical properties of amino acids and their natural substitution patterns to group amino acids into classes .
as we showed , substitutions of large aliphatic residues ( i , l , v ) are frequent in folding nuclei and this confused previous analysis that did not apply any amino acid classification scheme . while plaxco et al claimed in their paper @xcite
( without providing a supporting evidence ) that grouping of amino acids into classes did not change their conclusions , our analysis shows that proper classification of amino acids is crucial for detecting conservation in the folding nucleus .
second , plaxco et al used a different method to compute sequence variability : @xmath25 \label{eq : plaxco}\ ] ] this equation differs from eq.([eq : s ] ) , used in this study , in normalization by @xmath26 - the `` background '' frequency of residue type @xmath4 in all proteins .
although the difference may seem technical , equations ( [ eq : s ] ) and ( [ eq : plaxco ] ) are based on two different models of evolution .
we argue that while equation [ eq : plaxco ] may be adequate for dna sequence analysis @xcite it is not appropriate for analysis of protein evolution .
equation [ eq : plaxco ] implicitly assumes that amino acid composition @xmath26 is fixed _ a priori _ in each protein .
hence equation ( [ eq : plaxco ] ) tends to underestimate conservation of `` frequent '' amino acids ( l , a , s etc ) , while overestimating conservation of less frequent amino acids ( w , c , h etc ) .
in contrast , equation ( [ eq : s ] ) assumes that conservation requirement itself affects the composition , i.e. higher conservation of an amino acid leads to its higher frequency in proteins . to illustrate this point consider a toy protein that consists of two types of residues : hydrophobic
h and polar p. assume that @xmath27 of amino acids in this proteins are in the core and @xmath28 are in the loops .
also assume that in the toy world selection for stability requires a 100% conservation of h amino acids in the core , while loops are under no evolutionary pressure and h and p are equally probable in the loops
. then @xmath29 and @xmath30 . at
conserved core positions @xmath31 , while in the loops @xmath32 .
hence , the use of equation ( [ eq : plaxco ] leads to a counterintuitive and apparently wrong result @xmath33 , i.e. that loops are more conserved than 100% conserved core !
clearly this result shows inadequacy of equation ( [ eq : plaxco ] ) as applied to protein evolution with unconstrained composition .
similarly , application of equation [ eq : plaxco ] to real proteins leads to unreasonably low conservation of the hydrophobic core as compared to exposed loops ( data not shown ) .
a possible way to compensate for variations in amino acid composition of proteins is to define the sequence entropy as in @xcite : @xmath34 where the second term gives the `` background '' variability due to amino acid composition .
this term however does not depend on @xmath1 and hence does not change the relative variability .
interestingly , the use of equation ( [ eq : plaxco ] ) by plaxco et al @xcite gave rise to a surprising result that active sites in proteins are generally no more conserved than the rest of the protein ( see fig.2 of @xcite ) .
conservation of known active sites was used as a control in @xcite for their method of analysis based on equation [ eq : plaxco ] which it apparently failed . finally , plaxco et al did not study conservation of the folding nucleus .
instead , they focused on the residues that featured high @xmath0-values in protein engineering experiments and compared them with low @xmath0-value residues . as we explained above residues in the folding nucleus do not necessarily exhibit high @xmath0-values , and many low @xmath0-value residues are conserved in evolution as they contribute to stabilization of the native structure .
comparison with low @xmath0-value residues instead of comparison with the _ whole _ protein also confused previous analysis since most of @xmath0-values have been measured for amino acids located in the the core of a protein and hence these amino acids are on average more conserved . here
, in contrast , we used the folding nucleus as it was identified for each protein by the original experimental group and compared its conservation with the conservation of all amino acids in the protein . in summary
, we showed that folding nucleus is indeed conserved in most of the proteins whose folding transition states are known from protein engineering analysis .
that does not mean that folding nucleus residues are the _ the only _ conserved ones in any family of homologous proteins .
that also may not mean that folding nucleus is _ more _ conserved than other residues in the protein core , as nucleus is equally important for protein stability and for fast folding .
our result show that the folding nucleus is more conserved than the rest of the protein .
as stated earlier it is difficult to uniquely identify folding nucleus by looking at a conservation profile in just one family of homologous sequences .
nevertheless conservation of folding nucleus found in this paper and in other works @xcite points out to an exciting possibility that folding rates may be of biological significance .
biological significance of this fact needs to be assessed in future studies .
abagyan , r. & batalov , s. ( 1997 ) .
do aligned sequences share the same fold ? , 273:35568 .
abkevich , v. , gutin , a. , & shakhnovich , e. ( 1994 ) .
specific nucleus as the transition state for protein folding : evidence from the lattice model .
, 33:1002610036 .
bateman , a. , birney , e. , durbin , r. , eddy , s. , howe , k. , & sonnhammer , e. ( 2000 ) .
the pfam protein families database .
, 28:2636 .
benner , s. , cohen , m. , & gonnet , g. ( 1994 ) .
amino acid substitution during functionally constrained divergent evolution of protein sequences .
, 7:132332 .
branden , c. & tooze , j. ( 1998 ) . .
garland publishing , inc .
, new york .
brenner , s. , chothia , c. , & hubbard , t. ( 1998 ) . assessing sequence comparison methods with reliable
structurally identified distant evolutionary relationships .
, 95:60738 .
chiti , f. , taddei , n. , white , p. , bucciantini , m. , magherini , f. , stefani , m. , & dobson , c. ( 1999 ) .
mutational analysis of acylphosphatase suggests the importance of topology and contact order in protein folding .
, 6:10051009 .
dodge , c. , schneider , r. , & sander , c. ( 1998 ) .
the hssp database of protein structure - sequence alignments and family profiles . , 26:3135 .
dokholyan , n. , buldyrev , s. , stanley , h. , & shakhnovich , e. ( 2000 ) . identifying the protein folding nucleus using molecular dynamics .
, 296:11831188 .
fersht , a. ( 1997 ) .
nucleation mechanism of protein folding .
, 7:1014 .
guo , z. & thirumalai , d. ( 1995 ) .
nucleation mechanism for protein folding and theoretical predictions for hydrogen - exchange labelling experiments .
, 35:137139 .
hamill , s. , steward , a. , & clarke , j. ( 2000 ) .
the folding of an immunoglobulin - like greek key protein is defined by a common - core nucleus and regions constrained by topology .
, 297:165168 .
henikoff , s. & henikoff , j. ( 1992 ) .
amino acid substitution matrices from protein blocks . , 89:109159 .
itzhaki , l. , otzen , d. , & fersht , a. ( 1995 ) .
the structure of the transition state for folding of chymotrypsin inhibitor 2 analyzed by protein engineering methods : evidence for a nucleation - condensation mechanism for protein folding .
, 254:260288 .
klimov , d. & thirumalai , d. ( 1998 ) .
lattice models for proteins reveal multiple folding nuclei for nucleation - collapse mechanism .
, 282:471492 .
kragelund , b. , osmark , p. , neergaard , t. , schiodt , j. , kristiansen , k. , knudsen , j. , & poulsen , f. ( 1999 ) .
the formation of a native - like structure containing eight conserved hydrophobic residues is rate limiting in two - state protein folding of acbp .
, 6:594601 .
li , l. , mirny , l. , & shakhnovich , e. ( 2000 ) .
kinetics , thermodynamics and evolution of non - native interactions in protein folding nucleus . , 7:336341 .
lopez - hernandez , e. & serrano , l. ( 1996 ) .
structure of the transition state for folding of the 129 aa protein chey resembles that of a smaller protein , ci2 .
, 1:4355 .
lorch , m. , mason , j. , clarke , a. , & parker , m. ( 1999 ) .
effects of core mutations on the folding of a beta - sheet protein : implications for backbone organization in the i - state . , 38:137785 .
main , e. , fulton , k. , & jackson , s. ( 1999 ) .
folding pathway of fkbp12 and characterisation of the transition state .
, 291:429444 .
martinez , j. , pissabarro , t. , & serrano , l. ( 1998 ) .
obligatory steps in protein folding and the conformational diversity of the transition state .
, 5:721729 .
michnick , s. & shakhnovich , e. ( 1998 ) . a strategy for detecting the conservation of folding - nucleus residues in protein superfamilies .
, 3:239251 .
mirny , l. , abkevich , v. , & shakhnovich , e. ( 1998a ) .
how evolution makes proteins fold quickly .
, 95:49764981 .
mirny , l. , abkevich , v. , & shakhnovich , e. ( 1998b ) .
how evolution makes proteins fold quickly .
, 95:497681 .
mirny , l. & ei , s. protein folding theory : from lattice to all - atom models . , 30:in press .
mirny , l. & shakhnovich , e. ( 1999 ) . universally conserved residues in protein folds .
reading evolutionary signals about protein function , stability and folding kinetics .
, 291:177196 .
pande , v. , grosberg , a. , rokshar , d. , & tanaka , t. ( 1998 ) .
pathways for protein folding : is a `` new view '' needed ? , 8:6879 .
plaxco , k. , larson , s. , ruczinski , i. , riddle , d. , buchwitz , b. , davidson , a. , & baker , d. ( 2000 ) . evolutionary conservation in protein folding kinetics .
, 298:303312 .
ptitsyn , o. ( 1998 ) .
protein folding and protein evolution : common folding nucleus in different subfamilies of c - type cytochromes ?
, 278:655666 .
ptitsyn , o. & ting , k. ( 1999 ) .
non - functional conserved residues in globins and their possible role as a folding nucleus . , 291:671682 .
schneider , t. ( 1999 ) .
measuring molecular information [ letter ] .
, 201:8792 .
shakhnovich , e. ( 1997 ) .
theoretical studies of protein - folding thermodynamics and kinetics . , 7:2940 .
shakhnovich , e. , abkevich , v. , & ptitsyn , o. ( 1996 ) . conserved residues and the mechanism of protein folding . , 379:9698 .
stormo , g. ( 1998 ) .
information content and free energy in dna protein interactions . , 195:1357 .
ternstrom , t. , mayor , u. , akke , m. , & oliveberg , m. ( 1999 ) . from snap - shot to movie : phi - value analysis of protein folding transition states taken one step further .
, 96:1485414859 .
thompson , m. & goldstein , r. ( 1996 ) .
constructing amino acid residue substitution classes maximally indicative .
, 25:2837 .
vilegas , v. , martinez , j. , avilez , f. , & serrano , l. ( 1998 ) .
structure of the transition state in the folding process of human procarboxypeptidase a2 activation domain .
, 283:10271036 .
* fig.1 * variability profiles ( sequence entropy ) for nine different proteins computed using ms residue classes .
circles indicate positions at which @xmath0-values have been experimentally measured .
residues forming the folding nucleus are shown by filled circles . *
fig.2 * nine studied proteins with c@xmath35 atoms colored according to the degree of their conservation ( evaluated in fig.1 ) : from blue ( high conservation ) to light - blue , green , yellow and red ( no conservation ) . folding nucleus residues
are shown by twice as large spheres .
notice conserved ( blue ) cores of the proteins and non - conserved ( yellow and red ) surfaces .
also notice several conserved non - nucleus residues in the protein core . | in this communication we present statistical analysis of conservation profiles in families of homologous sequences for nine proteins whose folding nucleus was determined by protein engineering methods .
we show that in all but one protein ( acp ) folding nucleus residues are significantly more conserved than the rest of the protein .
two aspects of our study are especially important : 1 ) grouping of amino acids into classes according to their physical - chemical properties and 2 ) proper normalization of amino acid probabilities that reflects the fact that evolutionary pressure to conserve some amino acid types may itself affect concentration of various amino acid types in protein families
. neglect of any of those two factors may make physical and biological `` signals '' from conservation profiles disappear .
epsf running title : conservation of folding nucleus submitted to _
journal of molecular biology _ harvard university , department of + chemistry and chemical biology + 12 oxford street , cambridge ma 02138 + |
here we will provide the formulas describing the time - evolution of the mechanical oscillator along with the formula for the qfi of the steady state . assuming that the mechanical oscillator is continuously monitored with efficiency @xmath19 , the evolution is described by the following stochastic master equation @xmath83 \ : dt+ ( \gamma_{\sf env } + \gamma_{\sf fun } ) \ : \mathcal{d}[\hat{x}]\varrho \ : dt \nonumber \\ & \qquad + \sqrt{\eta \gamma_{\sf env } } \mathcal{h}[\hat{x } ] \varrho \ : dw \label{eq : smesm}\end{aligned}\ ] ] where @xmath6 , @xmath7\varrho = o\varrho o^\dag - ( o^\dag o \varrho + \varrho o^\dag o)/2 $ ] and @xmath20\varrho = o \varrho + \varrho o^\dag - \tr[(o+o^\dag)\varrho]$ ] . this equation can be translated in the following equations for first moments and covariance matrix , fully describing the evolution for gaussian quantum states @xmath84 where @xmath85 is a vector of wiener increments such that @xmath86 and the matrices read @xmath87 the steady state covariance matrix can be derived analytically as @xmath88 where @xmath89 notice that , typically , the steady state above is a squeezed state , in the sense that its smallest eigenvalue will be smaller than one . obtaining the decomposition , in terms of diagonal single - mode squeezers and orthogonal phase shifters , of the symplectic operation that relates the vacuum state to this steady state is a straightforward task , that just requires one to diagonalise the matrix @xmath46 .
the corresponding quantum fisher information can be easily evaluated by using the formula @xcite @xmath90}{1+\mu_\phi^2 } + 2 \frac{(\mu_{\sf ss}^{\prime})^2}{1-\mu_{\sf ss}^4 } \ : , \label{eq : gaussqfi}\end{aligned}\ ] ] where @xmath91 = 1/\sqrt{\det[\sigmacm_{\sf ss}]}$ ] represents the purity of the state , and primed quantities corresponds to derivative with respect to the parameter @xmath10 .
one then obtains @xmath92 + \gamma_{\sf fun } \left ( \omega_m - 3 \upsilon\right ) } { 8\upsilon ( \gamma_{\sf env } + \gamma_{\sf fun } ) \left [ \eta^2 \gamma_{\sf env}^2 - ( \gamma_{\sf env } + \gamma_{\sf fun})^2 \right ] } .\end{aligned}\ ] ] | inspired by the notion that environmental noise is in principle observable , whilst fundamental noise due to spontaneous localisation would not be , we study the estimation of the diffusion parameter induced by wave function collapse models under continuous monitoring of the environment .
we take into account finite measurement efficiencies and , in order to quantify the advantage granted by monitoring , we analyse the quantum fisher information associated with such a diffusion parameter , identify optimal measurements in limiting cases , and assess the performance of such measurements in more realistic conditions . _
introduction . _ - spontaneous localization models @xcite , in their many flavours and variations , were introduced from the late eighties primarily as an attempt to unify the dynamics of microscopic and macroscopic systems , encompassing measurement apparata , which customary quantum mechanics only describes through ad hoc prescriptions that can not be relied to the fundamental dynamical principles .
while such models reproduce quantum and classical mechanics in the extreme regimes of few ( @xmath0 ) and very many ( @xmath1 ) elementary constituents , they do deviate substantially from standard quantum mechanics in the intermediate mesoscopic regime . as molecular interferometry @xcite and quantum opto - mechanics , especially in the levitating paradigm @xcite , are swiftly advancing into this mesoscopic middle ground , there is currently a lively interest in designing and carrying out experiments that would falsify either standard quantum mechanics or its spontaneously localized variants @xcite . in a nutshell ,
spontaneous localization models postulate the presence of an additional stochastic term in the schrdinger equation , that would be responsible for the wave - function collapse and the perceived discontinuous dynamics of quantum projective measurements .
this would essentially imply the existence of a source of `` fundamental '' decoherence , in the form of momentum dissipation , acting on a mesoscopic system , such as a levitating opto - mechanical nanosphere .
it has hence been recently noted that , if the sources of `` environmental '' decoherence due to the interaction and entanglement with the environment are well known , the additional fundamental decoherence could be directly observed by tracking the system s dynamics @xcite .
the detection of fundamental effects over the background of environmental ones is however obviously difficult , as the two may take the same form and imply qualitatively similar effects .
the primary intent of this work is emphasising that a possible distinction between fundamental and environmental decoherence is that , while the former is unavoidable and beyond repair , the latter can in principle be reversed through measurements : if the physical degrees of freedom of the environment are completely or partially accessible , one can perform measurements on them that partly restore information about the quantum state @xcite .
drawing from this notion , we will hence consider the estimation of the free parameter of qmupl ( `` quantum mechanics with universal position localization '' ) or of an equivalent function of the two parameters of the csl ( `` continuous spontaneous localization '' ) wavefunction collapse model , under time - continuous measurements on the environment of a quantum degree of freedom , such as the centre of mass of a levitated nanosphere @xcite .
the latter will be our system of reference , bearing in mind that similar results would apply to more general settings .
the monitoring we consider , aided by markovian linear feedback , has the added bonus of stabilising the dynamics @xcite , so that we will be in a position to base our investigation entirely on steady state properties and not on the features of the transient dynamics , which may be more elusive to record in practice . as a further element of novelty
, we will not just consider specific empirical signatures of the different values of the collapse parameter but instead address their systematic , ultimate discrimination by applying quantum estimation techniques and deriving the quantum fisher information ( qfi ) associated with such a parameter @xcite .
thus , we will quantify exactly the advantage provided by continuous monitoring as a decrease in the achievable uncertainty on the parameter estimation , and hence on the discrimination between different theories .
we shall derive analytical expressions for both the qfi and the optimal final measurement for parameter discrimination in limiting instances , and show the latter performs remarkably well in realistic situations too . _
the dynamics . _ - to fix ideas , we shall consider a single noisy continuous variable quantum degree of freedom subject to a positive definite harmonic hamiltonian and to momentum diffusion , as would be the case for a trapped nanosphere undergoing heating via photon scattering , background gas collisions and blackbody radiation @xcite . in order to simplify our treatment , we shall not include the typically smaller effects of position diffusion and friction @xcite , which could be accounted for promptly within our formalism but would not add much conceptual insight .
the additional stochastic term acting on the state vector according to the qmupl model is equivalent to a momentum diffusion lindblad superoperator entering the master equation for the quantum state @xmath2 .
the same is approximately true for the centre of mass of motion of mesoscopic objects in the csl model , since the position fluctuations are expected to be much smaller than the model localization length and one can perform a first order expansion of the superoperator @xcite .
hence , the overall dynamics we shall consider is the following : @xmath3 + \gamma \ : \mathcal{d}[\hat{x}]\varrho \ ; , \label{eq : me}\end{aligned}\ ] ] where @xmath4=i$ ] ( @xmath5 ) , @xmath6 and @xmath7\varrho = o\varrho o^\dag - ( o^\dag o \varrho + \varrho o^\dag o)/2 $ ] . the momentum diffusion rate is the sum of two contributions : @xmath8 , where @xmath9 is due to environmental effect , while @xmath10 is fundamental .
our aim is analysing the estimation of @xmath10 .
notice that such a parameter is equivalent to the only fundamental parameter of qmupl and constrains the parameters of the csl models through the formula @xmath11 @xcite , where @xmath12 is the mass of the object , @xmath13 is a factor that depends on its geometry , and the parameters @xmath14 and @xmath15 characterise the model ( the value for the intrinsic length scale is typically chosen at @xmath16 nm , while bounds on the collapse rate are currently placed at @xmath17 @xcite ) . since we will consider only one mode , we will assume that after having trapped the nanosphere and cooled its motion down by sideband cooling , one will either turn off or detune the driving field in order to decouple the nanosphere motion and the cavity field .
hence , we will focus on the evolution of the mechanical oscillator alone .
as already argued , in principle one can always counter the environmental decoherence by monitoring the environment . here , we suppose to monitor the nanosphere position through the scattered light , obtaining a conditional dynamics described by the following conditional master equation @xcite : @xmath18 \varrho \ : { \rm d}w \ : , \label{eq : sme}\end{aligned}\ ] ] where @xmath19 denotes the monitoring efficiency , @xmath20\varrho = o \varrho + \varrho o^\dag - \tr[\varrho ( o + o^\dag)]$ ] and @xmath21 represents a standard wiener increment . as the hamiltonian @xmath22 is quadratic in the position and momentum operators , the evolution described by eq .
( [ eq : sme ] ) sends gaussian states into gaussian states , and thus we can fully describe it by looking at the evolution of the covariance matrix @xmath23 and the first moments vector @xmath24 of the quantum state @xmath2 , defined in components as @xmath25 $ ] and @xmath26 $ ] for the operator vector @xmath27 . remarkably , due to a very specific property of quantum and classical conditional gaussian statistics , one obtains @xcite that , while the evolution of the first moments is , as expected , stochastic , _
i.e. _ depends on the results of the measurement performed on the environment , the covariance matrix evolves deterministically . in formulae : @xmath28 where @xmath29 is a vector of independent wiener increments .
the matrices @xmath30 , @xmath31 and @xmath32 are given in the supplemental material @xcite as functions of the dynamical parameters .
notice that the efficiency parameter @xmath33 $ ] will allow us to describe realistic situations where the environmental degrees of freedom are only partially accessible ( as would be the case for the imperfect collection of light scattered by a nanosphere ) .
note also that setting @xmath34 obviously yields the original , unmonitored dynamics .
thanks to the deterministic evolution of the second moments , one can prove that real - time linear feedback ( _ i.e. _ real - time displacement in phase space depending on the measurement current ) can be applied in order to obtain a steady state with zero first moments and thus remove the stochasticity of the evolution @xcite .
this may not be the optimal strategy in terms of the estimation of @xmath10 , but we will nonetheless consider this regime in what follows and set the first moments to zero , as it does allow for a deterministic steady state and will let us illustrate the advantage granted by the monitoring with a very compact , entirely analytical treatment .
also , in an experiment , a stable steady state is certainly much more desirable than a stochastically fluctuating one , on which one would have to perform an optimal discriminating measurement that would also fluctuate stochastically .
now , if all the other dynamical parameters are known , the steady state solutions of eq .
( [ eq : evolutioncm ] ) yield a family of gaussian states with zero first moments parametrized by the different values of @xmath10 . to assess the effectiveness of our strategy , in the following we will derive the ultimate limits on the estimation of @xmath10 , quantified by the quantum and classical cramr - rao bound evaluated on such a family of states . _ quantum estimation theory . _
- let us consider a family of quantum states @xmath35 parametrized by a parameter @xmath36 that we want to estimate .
if one performs a measurement described by a positive operator valued measure ( povm ) @xmath37 , the ultimate limit on the precision of any unbiased estimator for the parameter @xmath36 is set by the cramr - rao bound @xcite @xmath38\ : , \end{aligned}\ ] ] where @xmath39 is the number of measurements performed , @xmath40 is the so - called ( classical ) fisher information ( fi ) , and @xmath41 $ ] denotes the conditional probability describing the whole measurement process . by optimizing over all the possible povms ,
one derives the quantum cramr - rao bound ( qcrb ) @xcite @xmath38 \geq 1/[m h(\gamma ) ] \ : , \label{eq : qcrb}\end{aligned}\ ] ] where @xmath42 $ ] is the quantum fisher information ( qfi ) and @xmath43 is the symmetric logarithmic derivative ( sld ) that is implicitly defined by the equation @xmath44 . as apparent from eq .
( [ eq : qcrb ] ) , the qfi quantifies with how much precision one can estimate the parameter @xmath36 independently from the specific measurement performed .
geometrically , the qfi corresponds to the bures metric in the hilbert space @xcite : large values of the qfi correspond to large bures distances between two quantum states @xmath35 and @xmath45 obtained via an infinitesimal variation of the parameter @xmath36 .
we also remark that for single - parameter estimation it is guaranteed that an optimal povm saturating the quantum cramr - rao bound always exists @xcite .
_ fundamental diffusion estimation . _ - as detailed above , we want to assess the estimation of the parameter @xmath10 , whose information is encoded in the gaussian steady state obtained through the monitoring described by eq .
( [ eq : sme ] ) .
in particular , we will focus only on the steady state covariance matrix solution @xmath46 of the riccati equation ( [ eq : evolutioncm ] ) , as we can assume that the first moments will be equal to zero .
it should be remarked here that the linear driving needed to set the first moments to zero does depend on the parameter @xmath10 we need to estimate . and so will the final optimal quantum measurement to be performed on the steady state .
however we can invoke , as customary in _ local _ quantum estimation problems , a multi - step adaptive protocol in order to solve this possible conundrum : one can apply the optimal protocol valid for an initial rough guess of the parameter @xmath10 ( say , in our case , @xmath47 ) , estimate the parameter through the measurement just performed , and then refine the operation and optimal measurement to be implemented given the latest estimate of the parameter .
it has been shown in several cases that , after a few adaptive steps , one obtains an estimator giving the true value of the parameter and saturating the cramr - rao bound @xcite . as a function of the monitoring efficiency @xmath19 , for @xmath48 , @xmath49 and and for different values of the estimated fundamental diffusion @xmath10 . from top to bottom : @xmath50 .
[ f : qfi ] right : ratio between the classical fi for the povm introduced in the main text ( optimal for perfect monitoring and no collapses ) , and the qfi @xmath51 as a function of the monitoring efficiency @xmath19 and of the ratio between fundamental decoherence @xmath10 and environmental decoherence @xmath9 ( the environmental decoherence is kept fixed to @xmath49 ) .
[ f : optimalpovm],title="fig : " ] as a function of the monitoring efficiency @xmath19 , for @xmath48 , @xmath49 and and for different values of the estimated fundamental diffusion @xmath10 . from top to bottom : @xmath50 .
[ f : qfi ] right : ratio between the classical fi for the povm introduced in the main text ( optimal for perfect monitoring and no collapses ) , and the qfi @xmath51 as a function of the monitoring efficiency @xmath19 and of the ratio between fundamental decoherence @xmath10 and environmental decoherence @xmath9 ( the environmental decoherence is kept fixed to @xmath49 ) .
[ f : optimalpovm],title="fig : " ] under these assumptions ( gaussian steady state with zero first moments ) , the qfi can be evaluated analytically from the covariance matrix @xmath46 , as described in @xcite .
the general analytical formulae for the steady state covariance matrix @xmath46 and the qfi @xmath51 are rather cumbersome and are reported in the supplemental material @xcite .
we start our analysis by discussing the general properties of the qfi by looking at fig .
[ f : qfi ] .
firstly , it is apparent from the graph that the qfi increases monotonically with the detection efficiency @xmath19 , providing one with a quantitative confirmation that countering the environmental decoherence through continuous measurements would help in the estimation of the collapse - induced diffusion .
also , the qfi decreases monotonically with @xmath10 , implying that , in principle , smaller values of the parameter can be estimated more efficiently in terms of absolute error .
next , it is instructive to consider the limiting case of perfect monitoring efficiency ( @xmath52 ) , where the qfi takes the compact analytical form @xmath53 one can easily check that in the limit of both perfect monitoring and zero fundamental decoherence ( @xmath47 ) , the qfi diverges .
this can be intuitively understood by looking at how the steady state @xmath54 changes in the hilbert space by varying the parameter of interest : as we have already remarked , for perfect monitoring and zero fundamental decoherence , the steady state is pure ; however , increasing the value of @xmath10 from zero introduces a diffusion that can not be neutralized by monitoring the environment and , therefore , a mixed steady state ( the same reasoning applies to the case of fixed @xmath47 and measurement efficiency @xmath19 decreasing from the maximum value ) .
the abrupt change from pure to mixed states is responsible for the diverging qfi , and also yields insight as to the identification of the optimal quantum measurement saturating the qcrb . in point of fact , this argument singles out a dichotomic measurement corresponding to projecting either on the steady state itself @xmath55 or on the rest of the hilbert space , in order to be sensitive to the change from a pure to a mixed state .
the corresponding povm , described by the operators @xmath56 and @xmath57 , is indeed optimal , as it can be shown to achieve the quantum cramr - rao bound .
this povm can be realized by first applying the symplectic operation that sends the vacuum state into the pure steady state @xmath55 ( this operation involves some squeezing @xcite , that could be obtained by modulating the trap potential @xcite ) , and then by performing a vacuum projection ( which , in optomechanics , could in principle be achieved by a mapping of the mechanical state onto the light mode through red sideband driving , followed by a measurement of the latter with an avalanche photodiode , that distinguishes between zero and any positive number of photons ) .
more revealing than the quantum fisher information itself is the associated signal to noise ratio @xmath58 quite simply , the ratio between the estimated value and its standard deviation which can be easily determined from eq .
( [ eta1 ] ) in the case of perfect monitoring , and reads @xmath59 which , in the regime @xmath60 , goes like @xmath61 ( let us remind the reader that @xmath39 is the number of estimation runs ) , and hence vanishes as @xmath10 vanishes .
the limits of perfect continuous monitoring of the environment and zero fundamental decoherence are clearly idealizations .
however , the continuity of the qfi assures us that very high precision can be obtained in their neighbourhood , _
i.e. _ for high but not perfect efficiency @xmath62 and for the interesting case of small csl decoherence , when @xmath63 .
we have in fact also investigated the performance of the povm just described also for different values of @xmath19 and @xmath10 , where we know it is no longer optimal .
as apparent from fig .
[ f : optimalpovm ] , the ratio between the classical fi and the qfi is still above @xmath64 for a reasonably large region of values of @xmath19 and @xmath10 .
this is particularly relevant since the optimal measurement , as is often the case , depends on the parameter to be estimated ; given these results , one can indeed apply the optimal measurement for the case of zero fundamental decoherence , and still achieve a very high precision in the most interesting region of small , but not zero , values of @xmath10 .
the most convincing evidence for the advantage granted by the continuous monitoring comes from considering the number of measurements needed to achieve a signal to noise ratio of one in the plausible scenario of continuous spontaneous localization with @xmath65 , for a hypothetical nanosphere of radius @xmath66 , mechanical frequency @xmath67 khz and subject environmental diffusion @xmath68 as a function of the monitoring efficiency @xmath19 , as shown in fig .
[ f : snr ] .
it can be seen that the number of runs goes from around one million for the unmonitored case to around @xmath69 for perfect monitoring .
hence , one concludes that environmental monitoring would help in designing experiments able to improve the existing bounds on @xmath10 . as a further piece of analysis , one can consider the dependence of the signal to noise ratio @xmath70 on the mechanical frequency @xmath71 in the csl model , where @xmath71 and the oscillator mass @xmath12 determine @xmath10 . for perfect efficiency
, this reads @xmath72 with @xmath73 .
this is a decreasing function of @xmath74 , confirming that oscillators at lower frequencies ( which , however , are more challenging to cool down to the quantum regime ) would prove advantageous in this context , as already indicated in proposals such as @xcite , where an ion trap , rather than optical tweezers , was considered . and for @xmath65 ( other parameters are set considering a nanosphere of radius @xmath66 , with mechanical frequency @xmath67 khz and with environmental diffusion @xmath68 kzhz ) .
[ f : snr ] right : ratio between the qfi @xmath75 obtained for the state at time @xmath76 and the steady state qfi @xmath51 as a function of the evolution time @xmath76 , for @xmath77 khz , @xmath68 khz and @xmath78 .
[ f : finitetime],title="fig : " ] and for @xmath65 ( other parameters are set considering a nanosphere of radius @xmath66 , with mechanical frequency @xmath67 khz and with environmental diffusion @xmath68 kzhz ) . [
f : snr ] right : ratio between the qfi @xmath75 obtained for the state at time @xmath76 and the steady state qfi @xmath51 as a function of the evolution time @xmath76 , for @xmath77 khz , @xmath68 khz and @xmath78 .
[ f : finitetime],title="fig : " ] _ finite time analysis . _ - all the results reported above have been derived considering the mechanical oscillator steady state . in order to validate such an analysis
, we investigate the transient dynamics of the qfi and in particular its ratio with the qfi obtained at steady state .
the results are plotted for experimentally reasonable values of the parameters in fig .
[ f : finitetime ] .
we focus in particular on small values of the fundamental decoherence parameter @xmath10 , but we have numerical evidence that similar results are obtained for larger values .
as can be seen from the graph , the ratio goes indeed to one in a relatively small time ( around @xmath79 with our parameters ) , _ i.e. _ the precision on the estimation of the parameter @xmath10 can be safely obtained at finite time . while in the case of perfect monitoring ( @xmath80 ) , the ratio increases monotonically to @xmath81 , in the case of @xmath82 one observes a maximum , indicating the possibility of a better estimation at finite times .
however , this marginal advantage would come at the price of having to perform fast measurements at a certain fixed time : as argued above , in a laboratory it may be more expedient to wait for the stable steady state . _ discussion and outlook .
_ - it should be noted that our estimation analysis assumed perfect knowledge of all the dynamical parameters other than our target @xmath10 .
this may be particularly delicate , especially in regard to the environmental diffusion @xmath9 , which would have to be inferred from other parameters through theoretical considerations @xcite .
quantum estimation theory could however be adapated to allow for uncertainties in dynamical parameters other than the estimated one ( see e.g. @xcite ) . in the optomechanical paradigm
, one could also take into account the coupled light field : the extension of the present study to the full , two - mode optomechanical system , and the identification of associated optimal global detection strategies , will be an interesting development of this line of inquiry .
nevertheless , as it stands , our study clearly highlights the substantial advantage that monitoring environmental decoherence , whenever and to whatever extent possible , would grant .
this notion , we argue , should hence inform the design of experiments aimed at setting bounds on wave function collapse models . here
monitoring would be , in a sense , an active way of `` putting aside the impediments of matter '' that hinder the detection of fundamental effects , much in the same fashion as friction was standing in the way of galileo s analysis of free fall motion @xcite .
+ _ note added _ : after the completion of this work , we became aware of a related analysis where the qfi is employed to assess non - interferometric tests of wave - function collapse models ( see @xcite ) .
_ acknowledgments . _ - we thank p. barker and d. goldwater for discussions .
mgg and as acknowledge support from epsrc through grant ep / k026267/1 .
mgg acknowledges support from the marie skodowska - curie action h2020-msca - if-2015 .
osd acknowledge support from grant 2015/12747 - 9 , so paulo research foundation ( fapesp )
. 99 a. bassi , k. lochan , s. satin , t.p .
singh , and h. ulbricht , rev .
mod .
phys . * 85 * , 471 ( 2013 ) .
s. gerlich _
et al . _ , nat .
commun .
* 2 * , 263 ( 2011 ) .
markus arndt and klaus hornberger , nat .
phys .
* 10 * , 271 ( 2014 ) .
d. chang _
et al . _ ,
proc .
natl .
acad .
sci .
u.s.a . *
107 * , 1005 ( 2010 ) .
o. romero - isart , m. juan , r. quidant , and j.i .
cirac , new j. phys .
* 12 * , 033015 ( 2010 ) .
j. gieseler , b. deutsch , r. quidant , and l. novotny , phys .
rev . lett . * 109 * , 103603 ( 2012 ) . o. romero - isart , phys .
rev .
a * 84 * , 052121 ( 2011 ) .
m. bahrami , m. paternostro , a. bassi and h. ulbricht , phys .
rev .
lett . * 112 * , 210404 ( 2014 ) .
s. nimmrichter , k. hornberger , k. hammerer , phys .
rev .
lett . * 113 * , 020405 ( 2014 ) .
a. vinante , m. bahrami , a. bassi , o. usenko , g. wijts , and t. h. oosterkamp , phys .
rev .
lett . * 116 * , 090402 ( 2016 ) .
j. li , s. zippilli , j. zhang , d. vitali , phys .
rev .
a * 93 * , 050102(r ) ( 2016 ) .
d. goldwater , m. paternostro and p. f. barker , arxiv:1506.08782 [ quant - ph ] .
m. abdi , p. degenfeld - schonburg , m. sameti , c. navarrete - benlloch , and m.j .
hartmann , arxiv:1602.07922 [ quant - ph ] .
let us mention here that another , possibly complementary , strategy to distinguish between fundamental and environmental quantum noise could be offered by dynamical decoupling techniques @xcite . c. arenz , r. hillier , m. fraas , and d. burgarth , phys .
rev .
a * 92 * , 022102 ( 2015 ) .
r. hillier , c. arenz , and d. burgarth , j. physics a : math . and
theoretical * 48 * , 155301 ( 2015 ) .
m. a. taylor , j. janousek , v. daria , j. knittel , b. hage , h .- a .
bachor and w. p. bowen , nature photonics * 7 * , 229 ( 2013 ) .
m. g. genoni _
et al . _ ,
new j. phys . ,
* 17 * 073019 ( 2015 ) .
m. g. a. paris , int .
j. quant
. inf .
* 7 * , 125 ( 2009 ) . o. romero - isart _ et al .
_ , phys .
rev .
a * 83 * , 013803 ( 2011 ) .
b. rodenburg , l. p. neukirch , a. n. vamivakas , and m. bhattacharya , optica * 3 * , 318 ( 2016 ) . s. l. adler , j. phys .
a * 40 * , 2935 ( 2006 ) . c. gardiner and p. zoller ,
_ quantum noise _
( springer , berlin heidelberg , 2004 ) . h. m. wiseman and g. j. milburn , _ quantum measurement and control _
( cambridge university press , new york , 2010 ) . m. g. genoni , l. lami and a. serafini , contemporary physics ( 2016 ) . h. m. wiseman , a. c. doherty , phys . rev
. lett . * 94 * , 070405 ( 2005 ) .
see supplemental material h. cramer , _ mathematical methods of statistics _ , ( princeton university press , 1946 ) c. w. helstrom , r. s. kennedy , ieee trans . inf .
th . *
20 *
, 16 ( 1974 ) ; s. braunstein and c. caves , phys . rev . lett . * 72 * , 3439 ( 1994 ) .
h - j .
sommers et al . , j. phys .
a * 36 * , 10083 ( 2003 ) .
d. brivio , s. cialdi , s. vezzoli , b. teklu , m. g. genoni , s. olivares and m. g. a. paris , phys .
rev .
a * 81 * , 012305 ( 2010 ) .
a. berni , t. gehring , b. m. nielsen , v. handchen , m. g. a. paris and u. l. andersen , nature photonics * 9 * , 577 ( 2015 ) .
a. monras , arxiv : 1303.3682 [ quant - ph ]
. o. pinel , p. jian , n. treps , c. fabre and d. braun , phys .
rev .
a * 88 * , 040102(r ) ( 2013 ) .
d. j. heinzen and d. j. wineland , phys . rev .
a * 42 * , 2977 ( 1990 ) .
d. e. chang , c. a. regal , s. b. papp , d. j. wilson , j. ye , o. painter , h.j .
kimble , and p. zoller , proc .
natl .
acad .
sci .
usa * 107 * , 1005 ( 2010 ) .
m. d. vidrighin , g. donati , m. g. genoni , x .-
m .
jin , w. s. kolthammer , m. s. kim , a. datta , m. barbieri and i. walmsley , nature communcations * 5 * , 3532 ( 2014 ) .
m. szczykulska , t. baumgratz and a. datta , arxiv:1604.02615 [ quant - ph ] .
g. galilei , _ two new sciences _ ( elsevier , leiden , 1638 ) .
english translation by h. crew and a. del savio ( macmillan , new york , 1914 ) , reprinted in s. hawking , _ on the shoulders of giants _ ( penguin , 2003 ) .
s. mcmillen et al .
( to appear ) . |
here we provide more details about the derivation of the effective two body ( four fermion ) interaction introduced in eq . in the main text .
the procedure to obtain this term for the dimer model is to identify the dimer hilbert space with a subspace of the @xmath2 model hilbert space and project the @xmath2 hamiltonian eq . accordingly .
we identify the state with zero dimers of any kind with the state with zero electrons for the @xmath2 model .
the rest of the hilbert space for the dimers can be introduced via the operators @xmath79 and @xmath80 .
the phases @xmath81 represent a gauge choice and we shall follow the one by rokhsar and kivelson @xcite and define @xmath82 and @xmath83 , here @xmath84 is the y - component of the dimer co - ordinate . given the conventional inner product for the electron hilbert space ,
the dimer basis is not orthonormal and therefore does not serve as a convenient basis to calculate matrix elements .
this can be resolved by gram - schmidt orthogonalization . in general ,
if we denote the basis elements of the dimer hilbert space by @xmath125 and the overlap matrix between states @xmath126 , then an orthonormal basis for the hilbert space is given by @xcite : @xmath127 it is not too hard to check that the matrix @xmath128 is a real symmetric matrix @xmath129 and therefore @xmath130 . from this
it follows that @xmath131 , i.e. , the new states are orthonormal .
the hamiltonian projected onto this basis is given by @xcite : @xmath132 to leading order , @xmath133 as the dimers are nearly orthogonal . to show this consider two states @xmath134 and @xmath135 .
we can form the loop graph of @xmath134 and @xmath135 by deleting all the dimers that @xmath134 and @xmath135 have in common .
the rest of the dimers will form loops ( with dimers from state @xmath134 and state @xmath135 alternating along a loop ) . if there is a loop of length 2 , that is two dimers of different type on the same link then @xmath136 , so we have @xmath137 for those states . assuming there are no such links we have that all loops are at least length four .
now the overlap of @xmath134 and @xmath135 is the product of overlaps over all loops .
furthermore it is known that the overlap between two loops is exponential in the length of the loop @xcite .
since all loops are of at least length four ( rather long ) to leading order we may set the overlap matrix to zero if there is at least one loop or the states @xmath134 and @xmath135 are different
. now the states are normalized to unity so we have @xmath133 . the pairing term ( four - fermion interaction )
comes from the spin - spin term in the @xmath2 model , namely @xmath85 .
let us focus on a single plaquette term and consider eight relevant states for the dimers on this plaquette , @xmath86 and @xmath87 .
we notice that the hamiltonian @xmath89 is zero in the triplet channel .
this means that the effective hamiltonian for the dimers @xmath138 is also zero in the triplet channel .
indeed , the spins of the dimers are the same as the spins of the electrons for the @xmath2 model , so the projected hamiltonian has the same spin structure .
as such we might as well restrict the spins of the dimers to lie in a singlet ( there are two such states per plaquette with the two dimers lying either along the x - axis or along the y - axis ) .
moreover , the projected hamiltonian is diagonal in this basis .
indeed the bare hamiltonian @xmath89 contains no hopping terms for the electrons , only spin flip terms .
as such the only terms that could contribute to off diagonal matrix elements come from states of the form @xmath139 and @xmath140 ( and linear combinations thereof ) which belong to both dimer configurations ( along the x - axis and along the y - axis ) .
however the hamiltonian annihilates such states and the projected hamiltonian has no corresponding hopping terms . by symmetry
the hamiltonian when restricted to the singlet subspace is a multiple of the identity matrix .
its value is given by : @xmath141 and , within the spin - singlet channel , the hamiltonian is @xmath91 .
correspondingly , we can add to our hamiltonian in eq .
the term : @xmath142 this is a spin spin hamiltonian for the fermionic dimers .
particular realization of the dirac cones for @xmath143 , @xmath144 , @xmath145 , and @xmath146 ( top ) , @xmath147 ( middle ) , and @xmath148 ( bottom panel ) .
the superconducting paring was chosen to optimize the zero temperature gibbs free energy @xmath149 , respectively .
the whole brillouin zone from @xmath150 to @xmath151 is shown on the left and a cross section along the major diagonal on the right , which highlights @xmath152 of the dirac cones . ] a particular realization of the dirac cones .
we chose @xmath143 , @xmath144 , @xmath145 , and @xmath153 , the superconducting paring was chosen to optimize the zero temperature gibbs free energy @xmath154 .
the whole brillouin zone is shown on the left and a zoomed in version on two of the cones on the right .
, title="fig : " ] a particular realization of the dirac cones .
we chose @xmath143 , @xmath144 , @xmath145 , and @xmath153 , the superconducting paring was chosen to optimize the zero temperature gibbs free energy @xmath154 .
the whole brillouin zone is shown on the left and a zoomed in version on two of the cones on the right .
, title="fig : " ] a particular realization of the dirac cones .
we chose @xmath143 , @xmath144 , @xmath145 , and @xmath155 , the superconducting paring was chosen to optimize the zero temperature gibbs free energy @xmath156 .
the whole brillouin zone is shown on the left and a zoomed in version on two of the cones on the right .
, title="fig : " ] a particular realization of the dirac cones .
we chose @xmath143 , @xmath144 , @xmath145 , and @xmath155 , the superconducting paring was chosen to optimize the zero temperature gibbs free energy @xmath156 .
the whole brillouin zone is shown on the left and a zoomed in version on two of the cones on the right .
, title="fig : " ] a particular realization of the dirac cones .
we chose @xmath143 , @xmath144 , @xmath145 , and @xmath157 , the superconducting paring was chosen to optimize the zero temperature gibbs free energy @xmath158 .
the whole brillouin zone is shown on the left and a zoomed in version on two of the cones on the right .
, title="fig : " ] a particular realization of the dirac cones .
we chose @xmath143 , @xmath144 , @xmath145 , and @xmath157 , the superconducting paring was chosen to optimize the zero temperature gibbs free energy @xmath158 .
the whole brillouin zone is shown on the left and a zoomed in version on two of the cones on the right .
, title="fig : " ] to verify the existence and robustness of the dirac cones predicted in the main text , we plot numerically the function @xmath159 near the values of the hopping matrix elements @xmath143 , @xmath144 , @xmath145 , for various values of @xmath160 near the @xmath161 that minimizes the free energy .
we focus on the case of large @xmath162 ( see fig . [
fig : dirac_cones ] ) , and we find eight cones in all cases , even though the exchange coupling is taken to be @xmath163 times larger then the relevant value for cuprate superconductors . to get an estimate of the ratio of @xmath164 we note that for the cuprates @xmath165 ev , the nearest neighbor hopping @xmath166 ev @xcite , @xmath167 @xcite , and in general @xmath168 , as any boson bilinear must be less then the largest occupation number @xmath169 .
recalling that @xmath170 , we get for @xmath171 that @xmath172 , likely of the order of @xmath173 .
the considerations given in the main text can be extended to multiple species ( @xmath174 ) and different occupation number ( @xmath128 ) of dimers .
the meanfield studied in the main text becomes arbitrarily accurate in this limit .
we consider the case of the square lattice but generalizations to different lattice geometries is straightforward .
we consider @xmath174 different species of bosonic dimers @xmath175 living on the links of the lattice and @xmath176 species of fermionic dimers @xmath177 also living on the same links .
we represent the dimers using @xmath174 species of bosons living on the links of the lattice @xmath178 and @xmath176 species of fermions @xmath179 .
the dimer hilbert space can be mapped onto a subspace of the boson / fermion hilbert space via the correspondence where @xmath180 bosonic dimers of species @xmath181 living on the link @xmath6 are identified with the state for the slave particles where the occupation number of the boson @xmath181 on link @xmath6 is given by @xmath180 .
similarly for the fermions . rather than enforcing the constraint of at most one dimer per link ( which is redundant in the physical case where @xmath182 ) , we introduce the constraint @xmath183 here , for convenience of notation , @xmath12 labels the four links @xmath13 that are attached to vertex @xmath14 . in the dimer language this corresponds to the constraint that the total number of dimers of any species on all the links touching the vertex @xmath14 is given by @xmath184 .
@xmath128 can be an arbitrary number and in the limit where @xmath128 becomes large the bosonic part of the dimer model becomes semiclassical . in the path integral formulation
, the constraint can be written as @xmath185\,,\label{eq : constarint_integral}\ ] ] where @xmath186 is the length of a time slice .
there is no prescription to write a hamiltonian for a large @xmath174 theory .
however , in order to proceed further we need to write down hamiltonians for our slave bosons / fermions that reduce to the hamiltonian given in the main text in the case @xmath187 and are amenable to large @xmath174 expansion @xcite .
all the hamiltonians written in this section have a direct interpretation in terms of hamiltonians for the dimers ( they all correspond to various dimer hopping terms and terms that count the number of flippable dimer plaquettes ) .
note that it does not matter whether the hamiltonians we produce respect the constraint in eq .
as in the path integral formulation we insert projectors onto the physical space @xmath188 at every time slice . the hint towards how to do this extension comes from the expressions derived around eqs . and in the main text .
indeed in order to apply a hs transformation to our expressions we need to write our hamiltonian ( when restricted to a single plaquette ) schematically in the form @xmath189 where @xmath190 and @xmath191 are single particle operators for one species of dimer , either bosonic or fermionic ( see the main text e.g. eqs . , and ) .
the main idea is to replace @xmath192 .
here @xmath193 , @xmath194 are single particle operators either bosonic or fermionic which are identical to @xmath190 and @xmath191 except they now carry an index @xmath181 . as such each hs transformation given in the main text corresponds to a different large @xmath174 hamiltonian . in the large @xmath174 limit , with this extension , we present models where the hs mean - field becomes arbitrarily accurate .
qualitatively we expect mean - field theory to become more and more accurate as each particle interacts with @xmath174 particles with an interaction strength that is attenuated by @xmath195 .
we now proceed to give several examples of this procedure .
we note that none of the hamiltonians have any dependence on @xmath128 .
in particular we can replace @xmath196 where @xmath28 is arbitrary and each value of @xmath28 produces a different hamiltonian .
similarly we have @xmath197 where again @xmath37 is arbitrary .
the bose / fermi part may be extended to the large @xmath174 limit in a similar manner .
for instance @xmath198 the four fermion term in the hamiltonian ( see eq . ) is given by spin spin interactions and has a variety of large @xmath174 extensions which were previously tabulated @xcite .
we will not repeat this procedure here .
we note that , to produce superconductivity , the optimum extension is given by the @xmath199 formalism @xcite .
it is possible to do a hs decoupling on all these terms , based on the following identity @xmath200\,.\label{eq : hubbard_stratonovich_large_n}\ ] ] the mean - field equations ( stationary points of this integral ) are given by @xmath201 and @xmath202 . from this
we see that the mean - field is simply a sum of @xmath174 copies of the mean - field for a single species problem and we can replace @xmath203 and @xmath204 , where @xmath190 and @xmath191 are single particle operators .
there is one saddle point which reduces to @xmath174 copies of the mean - field theories found in the main text . in the limit where @xmath174 goes to infinity ,
the mean - field results become exact , see ref .
( section 17.2 ) . the derivation given in ref
. works verbatim for our case .
indeed an integration over the bosonic and fermionic fields that appear in the partition function of our theory ( for which the hamiltonian is quadratic ) may be performed to obtain : @xmath205\,,\label{eq : action_large_n}\ ] ] where @xmath206 are all the possible hs fields which we may introduce .
the only dependence on @xmath174 is the overall scaling @xmath207 of the action . using an argument identical to ref .
( section 17.2 ) it is possible to show that higher loop corrections to the partition function in eq .
vanish as @xmath208 , where @xmath209 is the number of propagators and @xmath210 is the number of loops .
as such all the higher loop corrections vanish when @xmath211 , making mean - field exact .
to discuss the various symmetries of our systems we focus for simplicity on the case when @xmath212 for arbitrary @xmath128 ( this does not entail any additional complexity beyond the physical case of @xmath213 ) .
consider the gauge transformation where we assign a @xmath16 phase to each vertex of our system , namely where each boson and fermion operator @xmath8 and @xmath9 on each link transforms as : @xmath214 any hamiltonian that preserves the constraint @xmath215 is automatically invariant under the @xmath16 gauge transformation in eq . .
indeed any hamiltonian that preserves the constraint in eq . can be written as a sum of monomials each of which is a product of creation and annihilation operators . in order to preserve the constraint eq .
we mush have the same number of creation and annihilation operators for the bosons / fermions at every vertex ( otherwise the constraint is no longer satisfied ) . under this condition , the total phase from phase factors in eq .
associated with each vertex vanishes , leading to an invariant hamiltonian . in particular , one can explicitly check that the hamiltonians in eqs . in the main text are invariant under the gauge transformation given in eq . .
is gauge compatible ( here @xmath217 is an arbitrary constant ) .
one simply has to transform the variables @xmath218 and @xmath219 in the opposite way as the gauge transformation for the bosons . focusing on a single plaquette and labeling the lower left corner site as @xmath14 with the other sites @xmath220 , @xmath221 and @xmath222 , the gauge transformation for the hs bosons is given by : @xmath223z_{i , j}\nonumber \\ w_{i1/2 } & \rightarrow\exp\left[i\left(\theta_{i}-\theta_{i+\hat{x}+\hat{y}}\right)\right]w_{i1/2}\nonumber \\
w_{i3/4 } & \rightarrow\exp\left[i\left(\theta_{i+\hat{x}}-\theta_{i+\hat{y}}\right)\right]w_{i3/4}\,.\label{eq : gauge_transformation_hubbard_stratanovich}\end{aligned}\ ] ] similar considerations can be made about the boson fermion and the fermion fermion part of the hamiltonian in the main text ( in particular the hs transformations in eqs . and
are completely gauge compatible ) . from this
it would appear that the solutions to our mean - field equations lead to a large degeneracy of mean - fields .
indeed it would seem that any gauge transformation of the mean - field solutions leads to a different solution with the same energy and as such a different ground state .
however , this is not the case : there is only one state of the physical system that can be obtained from two different states that differ by a gauge transformation .
more precisely , after we project onto the physical subspace via the projection operators @xmath224 , eq . , two states that differ by a gauge transformation given in eq
. project onto the same state up to an unobservable overall phase . to show this , without loss of generality ,
assume that under the gauge transformation in eq .
the state with no bosons / fermions transforms into itself , @xmath225 .
now an arbitrary state may be written as a linear combination of terms of the form : @xmath226 after projection we may as well assume that there are exactly @xmath128 @xmath227 s at every vertex @xmath14 in the expression in eq . .
under a gauge transformation , we can group the phases associated with every vertex together and , because of the constraint that there are exactly @xmath128 bosons / fermions at any vertex , this gauge transformation becomes a state independent phase : @xmath229 here we consider the projective symmetry group ( psg ) construction , again focusing on the case where @xmath212 .
the main idea behind the psg ( which was originally introduced for spin systems @xcite ) is that in order for a mean - field state to be invariant under a symmetry transformation of the system ( e.g. , a translation or a rotation ) , the mean - field ansatz needs not remain invariant under the transformation but needs only be invariant following a gauge transformation of the form in eq .
( which does not change the state of the systems as discussed previously ) .
using this observation we may define the psg of a mean - field ansatz @xcite as the set of all lattice transformations followed by gauge transformations ( as in eq . ) which leave the mean - field ansatz invariant .
an important subgroup of the psg is the igg ( invariant gauge group ) which is the set of all gauge transformations in eq . that leave the mean field ansatz invariant .
we will now proceed to calculate the igg for various lattices .
we only focus on hs generated mean - fields of the form in eq . which are gauge transformation compatible via eq . and ignore all other mean - fields .
it can be checked directly that the mean - fields obtained by considering the fermion fermion or fermion boson part of the hamiltonian do not change the value of the igg .
we begin with bipartite lattices , with sublattices labeled by @xmath190 and @xmath191 . the first constraint we obtain by the gauge transformation in eq .
is that @xmath230 , i.e. , it is gauge invariant and therefore it gives no further restrictions on the form of the gauge transformations .
the second constraint is that @xmath231w_{i1/2}\nonumber \\ w_{i3/4 } & = \exp\left[i\left(\theta_{i+\hat{x}}-\theta_{i+\hat{y}}\right)\right]w_{i3/4}\,.\label{eq : constraint_regular}\end{aligned}\ ] ] we assume that the terms @xmath219 are non - zero ( otherwise the mean - field lattice loses connectivity and breaks up into one dimensional sublattices ) . in this case we have that @xmath232 and @xmath233 .
therefore the igg for a bipartite lattice ( assuming the terms @xmath218 are zero ) is simply @xmath234 where there are two different phases living on the two sublattices @xmath190 and @xmath191 . when the decoupling fields @xmath218 are non - zero , we have the constraint @xmath235=\\ = \exp\left[2i\left(\theta_{a}+\theta_{b}\right)\right]=1\ ; , \end{array}\label{eq : constraint_superconducting}\ ] ] so that @xmath236 and the igg becomes @xmath237 .
adding fermions does not change the @xmath237 .
one simply gets another copy of the same set of equations .
qualitatively this is because the hs fields transform in a way determined by the transformation properties of fermion or boson bilinears under eq . .
since the bosons and fermions transform in the same away under eq . , it does not add any new `` information '' ( or constraints ) to consider fermions .
if we consider non - bipartite lattices , the constraint @xmath230 again does not effect the igg . on the contrary , the constraints in eq .
insure that the phases @xmath18 and @xmath238 for any two sites that can be reached by a finite number of translations along lines joining second nearest neighbors along a plaquette is the same . since any two sites on a non - bipartite lattice
may be joined that way , the igg when the @xmath239 is @xmath16 , i.e. , the same phase for every site . when the terms @xmath240 do not vanish , we have that @xmath235=\exp\left(4i\theta\right)=1\;,\end{array}\label{eq : non_bipartite}\ ] ] which means that the igg is now given by @xmath241 . adding fermions does not change the @xmath241 , as once again we get multiple copies of the same equations .
the psg for the boson / fermion system may be computed directly .
for example one can check that the algebraic psg for the dimer system is the same as the algebraic psg for a bosonic spin liquid on the same lattice with the same igg @xcite . indeed
the gauge degrees of freedom are identical and the `` commutator '' constraint equations @xcite are of the same form for the same lattice and the same igg . | we present a slave - particle mean - field study of the mixed boson+fermion quantum dimer model introduced by punk , allais , and sachdev [ pnas * 112 * , 9552 ( 2015 ) ] to describe the physics of the pseudogap phase in cuprate superconductors .
our analysis naturally leads to four charge @xmath0 fermion pockets whose total area is equal to the hole doping @xmath1 , for a range of parameters consistent with the @xmath2 model for high temperature superconductivity . here
we find that the dimers are unstable to d - wave superconductivity at low temperatures .
the region of the phase diagram with d - wave rather than s - wave superconductivity matches well with the appearance of the four fermion pockets . in the superconducting regime ,
the dispersion contains eight dirac cones along the diagonals of the brillouin zone . _
[ introduction]introduction . _ the rokhsar - kivelson quantum dimer model ( qdm ) was originally introduced to describe a possible magnetically - disordered phase the resonating valence bond ( rvb ) phase in high - temperature superconducting materials @xcite .
the arena where the qdm has been deployed has greatly expanded since its inception , and the model has taken on a key role in the study of a variety of magnetic quantum systems .
quantum dimers show up prominently in the study of hard - core bosons hopping on frustrated lattices @xcite , of arrays of josephson junctions with capacitative and josephson couplings @xcite , of frustrated ising models with an external field or with perturbative xy couplings @xcite , of various types of gauge theories @xcite , and of models with large spin - orbit couplings @xcite and various cold atom setups @xcite .
the study of qdms led to an abundance of new phenomena including deconfined quantum criticality and new routes to deconfinement @xcite .
it also provided one of the earliest known examples of topologically ordered states in a lattice model @xcite .
recently qdms have been revisited as models of high - temperature superconductivity @xcite .
this was motivated by the need to reconcile transport experiments @xcite and photoemission data @xcite in the underdoped region of cuprate superconductors : while photoemission data show fermi arcs enclosing an area @xmath3 ( with @xmath1 being the doping ) , transport measurements indicate plain fermi - liquid properties consistent with an area @xmath1 . in order to resolve this issue and produce a fermi liquid which encloses an area @xmath1 , the authors of refs .
@xcite introduced a model for the pseudogap region of the cuprate superconductors which consists of two types of dimers : one spinless bosonic dimer representing a valence bond between two neighboring spins and one spin @xmath4 fermionic dimer representing a hole delocalized between two sites .
fig .
[ fig : configurations ] shows an example of a boson+fermion dimer covering of the square lattice and depicts the dimer moves dictated by the quantum hamiltonian in eq . .
the boson+fermion qdm ( bfqdm ) was introduced and studied numerically in ref . using exact diagonalization , supporting the existence of a fractionalized fermi liquid enclosing an area @xmath1 . in this work we present a slave boson and fermion formulation of the bfqdm .
we find that four symmetric fermion pockets , located in the vicinity of @xmath5 in the brillouin zone , naturally appear at mean - field level .
the total area of the four pockets is given by the hole ( fermionic ) doping .
we find that the system is unstable to d - wave superconductivity at low temperatures .
the region of the phase diagram with d - wave superconductivity matches well the region with four fermion pockets . in the superconducting phase ,
the fermionic dimers ( holes ) acquire a dirac dispersion at eight points along the diagonals of the brillouin zone . the boson+fermion quantum dimer model of ref .
( a ) a particular dimer configuration . the lattice is shown in black .
the bosonic dimers representing the valence bonds are shown in blue while the spinful fermionic dimers representing a single electron delocalized over two sites are shown in green .
( b ) diagrams representative of the various terms in the dimer hamiltonian eq .
.,title="fig : " ] the boson+fermion quantum dimer model of ref . .
( a ) a particular dimer configuration .
the lattice is shown in black .
the bosonic dimers representing the valence bonds are shown in blue while the spinful fermionic dimers representing a single electron delocalized over two sites are shown in green .
( b ) diagrams representative of the various terms in the dimer hamiltonian eq .
.,title="fig : " ] [ sec : mapping - onto - slave ] _ mapping onto slave boson / fermion model . _
the quantum dimer model can be mapped exactly onto a slave boson+fermion model by considering a secondary hilbert space where we assign to each link ( @xmath6 ) of the lattice ( @xmath7 ) a bosonic mode @xmath8 and a spinful fermionic mode @xmath9 ( @xmath10 ) .
we associate the number of dimers on a link with the occupation numbers of the bosons or fermions on that link .
as such we have embedded the dimer hilbert space in a larger boson / fermion hilbert space .
the constraint that each site of the lattice has one and exactly one dimer attached to it may be rephrased in the boson / fermion language as : @xmath11 here , for convenience of notation , @xmath12 labels the four links @xmath13 that are attached to vertex @xmath14 .
any hamiltonian for the dimers has a boson / fermion representation ; in particular the terms illustrated in fig .
[ fig : configurations]b can be written as : @xmath15 where we included a chemical potential for the holes ( fermionic dimers ) , which is important for the connection with doped high - temperature superconductors @xcite .
the terms not written explicitly in eq .
are simply obtained from those shown by translational symmetry , four fold rotational symmetry , and reflection symmetry about the two diagonals .
this hamiltonian also has a local @xmath16 gauge symmetry @xmath17 with a phase @xmath18 associate to each vertex @xmath14 .
any hamiltonian that preserves the constraint given in eq .
is invariant under this gauge transformation @xcite .
a slave boson / fermion formulation of the bfqdm is obtained by introducing a lagrange multiplier : a real field @xmath19 that enforces the dimer constraint eq . at all times
@xmath20 , and shifting the action by @xmath21 .
[ sec : a - variety - of]_slave boson / fermion mean - field decoupling .
_ a systematic mean - field approach can be obtained by taking the saddle point with respect to the lagrange multiplier field @xmath22 , with a time - independent value @xmath23 that enforces the average constraint @xmath24 .
this procedure is accompanied by hubbard stratonovich ( hs ) transformations of every term in the hamiltonian in eq . separately .
we begin with the purely bosonic potential term : @xmath25 where @xmath26 and @xmath27 are auxiliary fields to be integrated over and @xmath28 is arbitrary . at mean - field level
we can drop the integrals over the auxiliary fields and replace them with their saddle point values @xmath29 , @xmath30 and @xmath31 .
the hopping term may be decoupled in a similar manner : @xmath32 where , again , at mean - field level we use the saddle point values @xmath33 , @xmath34 , @xmath35 , @xmath36 and @xmath37 is arbitrary .
other hs decouplings , and linear combinations thereof , are also possible .
we can make substantial progress in understanding the fermionic component of the theory without detailed analysis of the bosonic component .
indeed , any translationally invariant ( liquid - like ) bosonic ansatz , naturally expected in @xmath16 gauge theories coupled to fermions with a fermi surface @xcite , yields similar fermionic effective theories .
the fermionic mean - field hamiltonian reads @xmath38 which is effectively a tight - biding model with renormalized hoppings @xmath39 , @xmath40 and @xmath41 . the resulting model is defined on the bipartite checkerboard lattice that is medial to the original square lattice .
the horizontal ( @xmath42 ) and vertical ( @xmath43 ) links make up the two sublattices where the fermions reside . we define ( in momentum space )
the spinor that encodes these two flavors as @xmath44 and @xmath45 where : @xmath46 the eigenvalues are given by @xmath47 , where @xmath48 and @xmath49 . for hole doping @xmath1 ( the number of fermions in our model )
the lower band @xmath50 will be partially occupied .
the total area enclosed by the fermi surface in the lower band is equal to the hole doping @xmath1 ( multiplied by @xmath51 ) .
the hamiltonian eq .
has four - fold rotational symmetry , @xmath52 and @xmath53 , and reflection symmetry about the diagonals @xmath52 and @xmath54 as well as @xmath55 and @xmath53 . depending on the relative values of @xmath56 , @xmath57 and @xmath58
, the band minima will be located at different points in the brillouin zone , and the fermi surface topology will vary accordingly . in fig .
[ fig : k - minima]a we show the position of the minima along the @xmath59 directions ( or @xmath60 line ) , as a function of the ratios @xmath61 and @xmath62 .
we identify two regions in parameter space , where the dispersion minima are ( i ) at the @xmath63 point ( blue - colored region ) , and ( ii ) in between the @xmath63 and @xmath64 points , varying continuously with @xmath65 ( faded region ) .
an example of dispersion where the minima are at @xmath66 is shown in the bottom inset of fig .
[ fig : k - minima]a . case ( ii ) is clearly conducive to the appearance of four fermi pockets in an appropriate range of the chemical potential .
( a ) location of the band minima as a function of @xmath61 and @xmath62 .
the color scale corresponds to the distance along the @xmath60 line in the brillouin zone : blue corresponds to the @xmath63 point , @xmath67 , and red corresponds to the @xmath64 point , @xmath68 .
the insets show contours of the dispersion of the lower band of the hamiltonian eq . for specific choices of parameters in the corresponding regions .
( b ) dominant superconductivity instability as a function of @xmath61 and @xmath62 for doping @xmath69 and @xmath70 : d - wave ( white ) _ vs. _ s - wave ( black ) .
note the good correlation between d - wave superconductivity and the appearance of four band minima . ]
_ d - wave superconductivity_. to study superconducting instabilities we need to include four - fermion terms in the hamiltonian , i.e. , go beyond the model introduced in refs . and summarized in fig .
[ fig : configurations]b and eq . .
consider the @xmath2 hamiltonian on the square lattice @xcite , @xmath71 subject to the constraint that @xmath72 . here
@xmath73 and @xmath74 are the electron creation and annihilation operators ( @xmath75 ) of the @xmath2 model , @xmath76 ( with @xmath77 summed over ) , and @xmath78 .
we can identify the dimer hilbert space with a subspace of the hilbert space for the @xmath2 model , where the zero dimers state corresponds to the state with zero electrons , and the rest of the hilbert space can be introduced via the operators @xmath79 and @xmath80 .
the phases @xmath81 represent a gauge choice and we shall follow the one by rokhsar and kivelson @xcite and define @xmath82 and @xmath83 , where @xmath84 is the @xmath43-component of the 2d square lattice site index @xmath14 . given the conventional inner product for the electron hilbert space , the dimer basis is not orthonormal .
this issue can be addressed in general by gram - schmidt orthogonalization @xcite ; however , it is customary to use the leading order approximation and to assume that the dimer states are orthogonal ( and normalized ) @xcite .
the relevant hamiltonian can then be determined by projecting eq . onto this basis .
the pairing term ( four - fermion interaction ) comes from the spin - spin term in the @xmath2 model , namely @xmath85 .
let us focus on a single plaquette term and consider eight relevant states for this plaquette , @xmath86 and @xmath87 , @xmath88 .
the hamiltonian @xmath89 is non - zero only in the singlet channel and therefore we must restrict the spins @xmath90 to be in a singlet state , thereby the effective hamiltonian for the dimers is given by @xmath91 @xcite . as such we add to our hamiltonian in eq .
the term : @xmath92 for convenience we define @xmath93 and @xmath94 , whereby d - wave pairing corresponds to @xmath95 ( which in turn can be chosen real with an appropriate choice of phase ) . using a hs transformation on eq . , @xmath96 , where @xmath97 and @xmath98 .
here @xmath99 and @xmath100 .
the eigenvalues of this hamiltonian are given by @xmath101 where @xmath102\\ \lambda_{\vec{k } } & = \frac{1}{4}\left[(\xi_{\vec{k}}^{x})^{2}-(\xi_{\vec{k}}^{y})^{2}+(\delta_{\vec{k}}^{x})^{2}-(\delta_{\vec{k}}^{y})^{2}\right]^{2}\\ \xi_{\vec{k } } & = |\gamma_{\vec{k}}|^{2}\left[(\xi_{\vec{k}}^{x}+\xi_{\vec{k}}^{y})^{2}+(\delta_{\vec{k}}^{x}-\delta_{\vec{k}}^{y})^{2}\right].\end{aligned}\ ] ] when @xmath56 , @xmath57 and @xmath58 are such that there are four fermi pockets ( in the absence of superconductivity ) , there are eight dirac points in the dispersion , i.e. , there are eight nodes where the gap @xmath103 .
these points are located along the diagonals of the brillouin zone . when @xmath59 , @xmath104 vanishes , and the gap closing condition @xmath105 is equivalent to @xmath106 , where @xmath48 and @xmath107 .
notice that the fermi surface in the absence of superconductivity is given by @xmath108 .
therefore , whenever there are four fermi pockets , for a range of @xmath109 there will be two nodes for each pocket , slightly shifted along the diagonal from the original fermi surface @xcite .
using self consistent equations for the superconducting order parameter , we can then compare s - wave and d - wave instabilities .
up to an unimportant constant energy shift , the gibbs free energy is given by @xmath110.\nonumber \end{aligned}\ ] ] minimizing the free energy with respect to @xmath111 , we obtain : @xmath112\right\ } \nonumber \end{aligned}\ ] ] and similarly for @xmath113 . from the symmetries of this equation
we see that there are two solutions , @xmath114 , corresponding to d - wave and extended s - wave superconductivity .
we numerically compare the two solutions at zero temperature and find that d - wave superconductivity wins for a large range of ratios @xmath62 and @xmath61 , as illustrated in fig .
[ fig : k - minima]b .
the correlation between the region with fermion pockets depicted in fig [ fig : k - minima]a and the region with d - wave superconductivity in fig .
[ fig : k - minima]b is evident .
this can be qualitatively understood as the largest change in the gibbs free energy upon entering the superconducting state comes from the contribution of the integral around the fs . since the shape of the four fermi pockets follows largely the nodal lines of the s - wave order parameter , and it anti - correlates with the d - wave nodal lines , one expects the appearance of the pockets to favor d - wave superconductivity .
whereas the horizontal boundaries match very well in the two panels in fig .
[ fig : k - minima ] , the vertical boundaries less so . indeed , along the horizontal boundary the dispersion transitions smoothly from having a single minimum at the @xmath63 point to having four minima along the @xmath60 direction in the brillouin zone , i.e. , the minima move continuously away from the @xmath63 point ( which thus becomes a maximum ) . on the other hand , along the vertical boundary ,
the minima jump discontinuously from the @xmath63 point to the new four minima , as four local minima at finite momenta dip down to become the global minima . depending on the value of the chemical potential ,
there is a region in the @xmath62 _ vs. _ @xmath61 plane near the vertical boundary where the fermi surface has five sheets , four pockets coexisting with a surface surrounding the @xmath63 point .
the latter favors s - wave superconductivity as it has no nodes at the @xmath63 point , and it is therefore expected to shift the position of the boundary between d - wave and s - wave superconductivity , as observed . _
conclusions .
_ we presented a slave particle formulation of a mixed boson+fermion quantum dimer model recently proposed in the context of high - t@xmath115 superconductors @xcite .
a key finding of this work is that substantial progress can be made using a mean - field analysis that simply assumes a translational and rotational invariant ( liquid ) state for the bosonic component .
we analyze the effective theory for the remaining fermionic degrees of freedom , and distinguish between two regimes of fermi surface topology , depending on the effective couplings obtained from both microscopic parameters and correlations of the bosonic liquid state .
the two regimes correspond to one fermi surface around the @xmath63 point , or four fermi pockets centered along the @xmath60 lines . by including additional interactions that arise from the @xmath2 model , we find that the system is unstable to superconductivity .
the symmetry of the superconducting order parameter , s - wave _ vs. _ d - wave , is shown to correlate strongly with the fermi surface topology , with d - wave being favored when four fermi pockets are present .
* acknowledgements : * this work was supported in part by engineering and physical sciences research council ( epsrc ) grants no .
ep / g049394/1 ( c.ca . ) and no .
ep / m007065/1 ( c.ca . and g.g . ) , by doe grant def-06er46316 ( c.ch . ) , and by the epsrc network plus on `` emergence and physics far from equilibrium '' .
statement of compliance with the epsrc policy framework on research data : this publication reports theoretical work that does not require supporting research data . c.ca . and g.g .
thank the bu visitor program for its hospitality .
10 d. s. rokhsar and s. a. kivelson , phys .
rev .
lett . * 61 * , 2376 ( 1988 ) .
a. sen , k. damle and t. senthil phys .
rev .
b * 76 * , 235107 ( 2007 ) ; d. poilblanc , k. penc , and n. shannon phys . rev .
b * 75 * 220503 ( 2007 ) ; d. poilblanc , phys . rev
. b. * 76 * , 115104 ( 2007 ) .
a. f. albuquerque , h. g. katzgraber , m. troyer and g. blatter phys .
rev .
b * 78 * , 014503 ( 2008 ) .
r. moessner and s. l. sondhi , phys .
rev .
b * 68 * , 054405 ( 2003 ) .
r. moessner , s. l. sondhi and e. fradkin , phys .
rev .
b * 65 * , 024504 ( 2002 ) .
f. vernay , a. ralko , f. becca and f. mila , phys .
rev .
b * 74 * , 054402 ( 2006 ) .
h. p. buchler , m. hermele , s. d. huber , m. p. a. fisher and p. zoller , phys .
rev .
lett . * 95 * , 040402 ( 2005 ) .
e. fradkin , d. a. huse , r. moessner , v. ognasian and s. l. sondhi , phys .
rev .
b * 69 * , 224415 ( 2004 ) ; a. vishwanath , l. balents and t. senthil , phys . rev .
b * 69 * , 224415 ( 2004 ) .
x. g. wen , _ quantum field theory of many - body systems _ , ( oxford university press , oxford 2004 ) .
m. punk , a. allais and s. sachdev , pnas * 112 * , 9552 ( 2015 ) .
d. chowdhury and s. sachdev _ the enigma of the pseudogap phase in the cuprate superconductors _ in _ quantum criticality in condensed matter : phenomena , materials and ideas in theory and experiment _
j. jedrzejewski eds .
( word scientific publishing co , singapore 2016 ) .
a. a. patel , d. chowdhury , a. allais , and s. sachdev , arxiv 1602.05954 .
y. ando , y. kurita , s. komiya , s. ono , and k. segawa , phys .
rev .
lett . * 92 * , 197001 ( 2004 ) .
j. orenstein , g. a. thomas , a. j. millis , s. l. cooper , d. h. rapkine , t. timusk , l. f. schneemeyer , and j. v. waszczak , phys .
rev .
b * 42 * , 6342 ( 1990 ) .
s. uchida , t. ido , h. takagi , t. arima , y. tokura , and s. tajima , phys .
rev .
b * 43 * , 7942 ( 1991 ) .
p. a. lee , n. nagaosa and x. g. wen , rev .
mod .
phys .
* 78 * ( 2006 ) .
a. damascelli , z. hussain , and z .- x .
shen , rev .
mod .
phys .
75 , 473 ( 2003 ) .
k. m. shen , f. ronning , d. h. lu , f. baumberger , n. j. c. ingle , w. s. lee , w. meevasana , y. kohsaka , m. azuma , m. takano , h. takagi , and z .- x .
shen , science * 307 * , 901 ( 2005 ) .
h .- b .
yang , j. d. rameau , z .- h . pan , g. d. gu , p. d. johnson , h. claus , d. g. hinks , and t. e. kidd , phys .
rev .
lett . * 107 * , 047003 ( 2011 ) .
j. gonsalez , m. a. martin - delgado , g. sierra , a. h. vozmediano , _ quantum electron liquids and high - tc superconductivity _ ( springer , berlin 1995 ) .
j. klamut , b. w. veal , b. m. dabrowski , p. w. klamut , m. kazimerski eds .
_ recent developments in high temperature superconductivity _
( springer verlag , berlin 1995 ) . at first glance
it appears that there are more gauge transformations that preserve the hamiltonian .
in particular it looks like the transformation @xmath116 and @xmath117 , subject to the constraint that the phases around a plaquette add up to a multiple of @xmath118 @xmath119 , is another set of gauge transformations . here
@xmath120 is an integer and @xmath121 , @xmath122 , @xmath123 and @xmath124 are four links arranged counter clockwise around a plaquette
. however one can verify ( by explicitly constructing the gauge transformations ) that every gauge transformation of this form can be written as a gauge transformation of the form in eq . .
see supplementary online information .
s. s. lee , phys .
rev .
b * 78 * , 085129 ( 2008 ) m. hermele , t. senthil , m. p. a. fisher , p. a. lee , n. nagaosa , and x. g. wen , phys . rev .
b * 70 * , 214437 ( 2004 ) .
k .- s .
kim , phys .
rev .
b * 72 * , 245106 ( 2005 )
. m. ogata , and h. fukuyama , rep .
. phys . *
71 * , 036501 ( 2008 ) .
k. s. raman , r. moessner and s. l. sondhi , phys .
rev .
b * 72 * , 064413 ( 2005 ) . |
it is usually believed that galactic supernova remnants ( snrs ) are the main accelerators of cosmic rays ( crs , mainly protons ) and can boost particles up to the knee " energy of @xmath23 @xmath3 ev through the diffusive shock acceleration ( dsa ) .
an important probe of this conjecture is the hadronic @xmath1-ray emission ( namely , @xmath4-decay @xmath1-rays produced in the inelastic collision between the accelerated protons and the target baryons ) from the snrs that interact with molecular clouds ( mcs ) .
fortunately , great progress has been made in understanding the snr paradigm by the detection of a characteristic @xmath5-decay bump " in the interacting snrs ic 443 @xcite and w44 @xcite . yet
, multifaceted investigations of hadronic interaction are still needed to increase and strengthen the evidence for the paradigm in a large sample of snr - mc association systems ( e.g. , * ? ? ?
it is , however , often uneasy to distinguish the nature of the @xmath1-rays between the hadronic scenario and leptonic ( inverse compton [ ic ] and/or bremsstrahlung ) scenario even in the snr - mc systems , such as snr rx j1713.7@xmath03946 , and deep explorations for the spectral properties and emitting mechanism of them in the gev tev range are essential .
snr rx j1713.7@xmath03946 ( [email protected] ) was discovered by _
x - ray observations @xcite and suggested to be the remnant of the historical supernova ad 393 ( @xcite ; hereafter an age of the remnant @xmath6 yr will be used ) .
its shell - like x - ray emission is dominated by a nonthermal component and a lack of thermal line features ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
it is found to be confined in a molecular cavity , and the elevated gas temperature and broad molecular line wings in a few molecular cloudlets are ascribed to the high - energy events of the snr @xcite .
the molecular cavity can be reasonably understood as the product of the energetic stellar wind and ionizing photons of the massive progenitor star ( which is suggested to be no later than b0b1 type , with a mass @xmath715 @xmath8 ; @xcite ) . in the radio band , it shows faint emission and has an average angular diameter of @xmath9 , corresponding to an average radius @xmath10 pc at distance @xmath11 kpc ( e.g. * ? ? ?
* ; * ? ? ?
it is suggested that the snr is in the free expansion evolutionary phase ( e.g. * ? ? ?
* ; * ? ? ?
* ) and that the forward shock has not yet reached the wall enclosing the wind - blown cavity in view of the high velocity currently measured @xcite . since the detection of tev @xmath1-ray emission from it @xcite , snr rx j1713.7@xmath03946 has garnered an enormous amount of attention and aroused constant debates on the hadronic versus leptonic origin of its @xmath1-ray emission ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the debates seemed to have been concluded when the 2yr _
fermi_-lat observations revealed a hard gev spectrum with a power - law photon index @xmath12 @xcite , which appears to support the leptonic scenario ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
subsequently , however , a strong correlation of azimuthal distribution was found between the tev @xmath1-ray flux and the column density of total interstellar protons , which is in favor of a hadronic origin of the @xmath1-rays @xcite .
some hadronic interaction models have also been elaborated to interpret the hard gev emission @xcite . both scenarios in terms of relativistic leptons and protons accelerated via standard
dsa can explain the hard spectrum with their advantages , but meanwhile also present difficulties . in the context of pure leptonic processes
, the ic scattering seems to naturally explain the hard gev @xmath1-ray spectrum without any extra assumptions and can be well compatible with the lack of thermal x - ray line emission . a leptonic origin of the tev emission
is strongly suggested by @xcite using a spectral inversion method independent of the particle acceleration model .
the morphology in tev band closely matches the nonthermal shell in the radio and x - ray bands @xcite , also suggesting an origin from the same electron population . on the other hand , it is pointed out that single - zone distribution of elections encounters a difficulty in providing a good fit to its broadband spectral energy distribution ( sed ; * ? ? ?
also , if the @xmath1-ray emission is leptonic dominated , it remains perplexity why it does not contain any significant hadronic contribution , given the observational fact that the snr is environed by molecular gas . in the context of pure hadronic processes , an inhomogeneous shocked medium with dense clumps @xcite and a dense shell outside the snr without contact @xcite are invoked . in the former case ,
@xcite consider that , in the downstream of the shock wave , the penetration depth regulated by energy dependence of bohm diffusion is generally shorter than the thickness of clumps , which gives the mass of the proton - illuminated gas depending on the energy of incident protons @xmath13 and hence a @xmath1-ray index 0.5 smaller than that of the parent protons . alternatively , @xcite suggest that , because of turbulent layers surrounding the clumps , only protons above a minimum energy can effectively diffuse into clumps , which may result in a hard proton spectrum below that energy in clumps .
it is , however , difficult in this case to explain why no thermal x - ray line emission from the shocked dense medium is observed .
possibly , the dense clumps swept up by the shock wave can survive and still remain a temperature too low to produce thermal x - rays @xcite . in the latter case , the dense shell ( within a thickness of @xmath20.3 pc ) outside the snr
is effectively illuminated by tev protons in the cr precursor region , giving rise to a hard @xmath1-ray spectrum @xcite . in this case
, the shock does not interact with dense material , and thus the difficulty with the lack of thermal x - rays can be avoided ( although the model fluxes in @xmath210gev1tev appear to be underestimated ) . here
we explore the role played by the diffusive relativistic protons that escape from the snr shock wave and hit the surrounding dense gas in the @xmath1-ray emission of rx j1713.7@xmath03946 , considering that the blast wave is propagating in a molecular cavity , in which there may be a mount of shocked clumps .
most recently , the tev @xmath1-ray emission has been found to stretch noticeably beyond the nonthermal x - ray emission region , also indicating a contribution of the escaping protons to the @xmath1-ray emission ( * ? ? ? * and references therein ) .
currently , there are two types of mechanisms in essence suggested for hadronic emission : @xmath5 decay not only can take place in shock - crushed dense clouds ( e.g. , * ? ? ?
* ; * ? ? ?
* ) but also can occur in the adjacent mcs illuminated by diffusive escaping protons ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in the following ,
we show that the gev
tev @xmath1-ray spectrum of rx j1713.7@xmath03946 with updated 5yr _
fermi_-lat data can be perfectly interpreted with a two - zone scenario , in which the @xmath4-decay @xmath1-rays resulting from the diffusive protons will substantially contribute to the gev @xmath1-rays while the leptonic process or the @xmath14@xmath14 interaction in the shocked clumps will be responsible for the tev @xmath1-rays . in the calculation , we present a simplified algorithm for the accumulative diffusion " model for escaping protons developed by @xcite and apply the markov chain monte carlo ( mcmc ) method to constrain the model parameters .
here we consider two @xmath1-ray - emitting zones for snr rx j1713.6@xmath03946 .
the first zone is the cavity wall ( the dense matter at the cavity boundary ) , which is bombarded by the protons that escaped from the shock wave during the history of expansion .
the second zone is inside the snr , and the emission could be either ic scattering off the accelerated electrons ( model ia ) or the hadronic emission from the shocked clumps ( model iia ) .
it is suggested that cr protons that escape from snrs can illuminate the nearby massive clouds and produce @xmath1-rays @xcite .
some possible evidence for this scenario has been obtained by the observation of diffuse @xmath1-ray emission from the vicinity of snr shells , such as in w44 @xcite , with which massive mcs are spatially coincident . the model has successively been improved and applied to a number of snr - mc systems to explain the origin of the @xmath1-ray emission ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
in the case of rx j1713.7@xmath03946 , although the blast wave may not have hit upon the cavity wall ( e.g. * ? ? ?
* ) , a considerable fraction of the escaping protons have arrived at the wall .
it can be expected that the @xmath14@xmath14 collision between these protons with the molecules in the wall gives birth to @xmath5 decay @xmath1-rays .
this emission , together with the leptonic emission from the electrons and/or the hadronic emission from the shocked clumps , will be calculated below to jointly interpret the spectral and spatial behavior of the @xmath1-rays from the snr . for the continuous injection of the accelerated protons that escape from the propagating shock wave surface
, @xcite established an accumulative diffusion " model , in which the protons at a given position are a collection of the diffusive protons escaping from the entire shock surface at different radii throughout the history of the snr expansion .
they derived the proton distribution by integrating the solution of the diffusion equation given by @xcite for the impulsive point - like source injection ( see equations .
( 1)(2 ) in * ? ? ?
actually , the algorithm of the model can be simplified by directly solving the diffusion equation of the escaping protons as follows .
ignoring protons energy losses , the diffusion equation in the spherically symmetric case is written as @xmath15 where @xmath16 is the distribution of the escaping protons with energy @xmath17 at a given distance @xmath18 from the snr center at a given time @xmath19 from the explosion time , @xmath20 is the source injection function , and @xmath21 is the diffusion coefficient .
here we assume that the diffusion coefficient has the form of @xmath22 , where @xmath23 is the correction factor of slow diffusion around the snr and @xmath240.7 @xcite is the index of diffusion coefficient .
following the treatment in @xcite , we define a new function @xmath25 , and then equation reads as @xmath26 the green function for eq.([eq : dif2 ] ) , namely , the solution for point - like impulsive injection source @xmath27 , is @xcite @xmath28- \exp \left [ -\left(\frac{r+\xi } { r_{\rm d}}\right)^2 \right ] \right\},\end{aligned}\ ] ] where @xmath29 is the diffusion radius .
the injection function of the protons that escape from the shock front is @xmath30 where @xmath31 is the injection rate from the spherical surface @xmath32 and @xmath33 is the shock radius .
the solution of equation is @xmath34- \exp\left[-\left(\frac{r+r_{s}(t_i)}{r_{\rm d}}\right)^{2}\right ] \right\}{\rm d } t_i . \end{split}\ ] ] this solution , with a single integral , is equivalent to equation ( 2 ) in @xcite , which is a triple integral , and will be reduced to equation ( 8) in @xcite for the case of point - like continuous injection as @xmath35 .
the shock - accelerated particles are assumed to obey a power - law distribution with a high - energy cutoff : @xmath36 where @xmath37 , @xmath38 is the particle kinetic energy , @xmath39 is the power - law spectral index and is taken as @xmath40 , and @xmath41 is the cutoff energy and will be set to the cr knee " energy for the protons , leaving @xmath42 as a free parameter .
the normalization parameter @xmath43 is determined from the total kinetic energy of particles above 1 gev . in our calculation ,
two physical parameters are adopted instead of the normalization parameters : the energy conversion efficiency @xmath44 , namely , the faction of explosion energy converted to cr energy , and the number ratio between the accelerated electrons and protons at 1 gev , @xmath45 .
the energy conversion efficiency ( @xmath44 ) is always difficult to determine directly from the acceleration theory , and there is usually not a consensus in the value of the energy conversion efficiency . in the _ cr_-_hydro_-_nei _ modeling for snr rx j1713.7@xmath03946 , it was suggested that about 16% of the explosion energy has converted into the accelerated particles @xcite
. a lower value of 6% was obtained via modeling the @xmath1-ray emission @xcite . for electron - to - proton number ratio ( @xmath46 ) ,
an order of 0.01 was derived from the crs measured at earth and a lower limit , @xmath7@xmath47 , was implied according to the radio observations of the snrs in nearby galaxies @xcite .
the distribution of the escaping protons , @xmath48 , is obtained from equation , where @xmath49 is the radius of the cavity wall and is set to be @xmath210 pc . in the calculation , following @xcite , we assume that the snr radius conforms to the self - similar solution for the free expansion phase @xcite , @xmath50 , where @xmath51 and @xmath52 are the power - law indices of the initial outer density profile in the ejecta and of the initial ambient medium density profile . here
we take @xmath53 and @xmath54 ( for a shocked wind - dominated medium ) , but taking @xmath55 or @xmath56 will produce a similar result in our calculation .
we adopt the average remnant radius @xmath57pc at present time @xmath58yr .
the injection rate is given by @xmath59 .
based on the calculations of @xcite , we assume that about 20% of the accelerated protons escape from the snr . the explosion energy of the snr is assumed to be @xmath60erg . for the hadronic @xmath1-ray emission , we use the analytic photon emissivity developed by @xcite for @xmath14@xmath14 interaction , including the enhancement factor of 1.84 due to the contribution from heavy nuclei @xcite .
we calculate the leptonic emission following @xcite .
in addition to the cosmic microwave background , the seed photons also include the galactic interstellar infrared photons at 30 k with an energy density of 0.3 ev@xmath61 @xcite .
two cases will be considered for the @xmath1-ray emission arising from the inside of the snr : ( i ) it is dominated by the ic process of the accelerated electron ; and ( ii ) it is dominated by the collision of the accelerated protons with the shocked clumps . on the inner side of the shock front , the @xmath1-ray emission is assumed to be mainly from the shock - accelerated electrons , which are also producing the synchrotron emission in the spectral range from radio to x - rays .
the leptonic process in this snr has often been considered as the origin of the @xmath1-ray emission from it , chiefly owing to the consistency with the hard gev @xmath1-ray spectrum , the lack of thermal x - ray emission , and its location in a low - density cavity of molecular gas . in this case , we adopt a typical electron - to - proton number ratio @xmath46 = 0.01 .
( with this parameter , the @xmath1-ray emission from the possibly shocked clumps will be insignificant , as can be seen in the following results . )
there are seven parameters in this case ( for the combined contribution from the outside diffusive protons and inside electrons ; model ia ) : @xmath44 , @xmath62 , @xmath63 , @xmath23 , @xmath42 , @xmath64 , and @xmath65 , where @xmath64 is the average magnetic field at the shock and @xmath65 is the mass of the baryon targets in the cavity wall bombarded by the escaping protons . to constrain the model parameters ,
we employ the mcmc approach , which , based on the bayesian statistics , is better than the grid approach , with a more efficient sampling of the parameter space of interest , especially for high dimensions . a brief introduction to the basic procedure of the mcmc sampling can be found in @xcite , and more details can be found in @xcite , @xcite , and @xcite . in our mcmc fitting calculation ,
all parameters are set to be free , but @xmath63 is confined in the range 0.30.7 , and the observational data with only upper limits are excluded .
the one - dimensional ( 1d ) probability distributions of model parameters and the best - fit sed are shown in figure [ fig : casea ] .
the 1d probability distribution of each parameter converges well with a single peak ( left panel of figure [ fig : casea ] ) .
the best - fit parameters with 1@xmath66 statistical uncertainties and the @xmath67 values are listed in tables [ tab : params ] and table [ tab : chi2 ] , respectively .
the model @xmath1-ray sed ( see right panel of figure [ fig : casea ] ) shows that this model perfectly explains the @xmath1-ray data . in the energy range
@xmath23500 gev , the @xmath1-rays are mainly contributed by the hadronic emission with a peak flux therein ; in the remaining regime , they are dominated by the leptonic emission .
the fluxes above 0.25tev of the hadronic ( outer ) and leptonic ( inner ) components are about @xmath68 and @xmath69 , respectively .
the fluxes in the range from 0.3 to 40tev of the outer and inner components are about @xmath70 and @xmath71 , respectively . for the escape process
, we got @xmath72 , which is similar to the previous estimates for snr environs ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and @xmath73 , which is in the normal range .
the mass of the baryon targets , @xmath74 , seems reasonable as compared to the estimate of the surrounding interstellar protons reservoir , @xmath75 @xcite .
in addition , the @xmath1-ray emission from the possibly shocked clumps ( see the next subsection for description ) in this case is also plotted by gray dotted line in figure [ fig : casea ] , implying an insignificant contribution to the total @xmath1-ray emission . for comparison
, we also test the notion of the pure leptonic process as the origin of the gev
tev @xmath1-rays , incorporating the updated _ fermi _ data ( model ib ) .
the best - fit seds are displayed by the gray solid line in figure [ fig : casea ] .
the model parameters and @xmath67 values of the best fit are also listed in tables [ tab : params ] and [ tab : chi2 ] , respectively . the @xmath67 values in various wavebands
are obviously higher than those in model ia except in the radio .
the best - fit sed shows that the pure leptonic model underestimates the gev flux , similar to the results of previous multiband fits with simple leptonic models ( e.g. , * ? ? ?
* ; * ? ? ?
in addition , we also calculate the @xmath67 value in the gev band using our best - fit sed and the previous _ fermi _ data @xcite , which are listed in table [ tab : chi2 ] in parentheses . in this case , the @xmath1-ray emission in the downstream are mainly from the decay of the neutral pions generated in the inelastic collision between the relativistic protons and shocked dense clumps .
it is pointed out that there may be some dense clumps shocked by the blast wave , which can act as a target for the bombardment of energetic protons ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . as mentioned in section section [ sec
: intr ] , a few hadronic models have been proposed to interpret the observed gev
tev @xmath1-ray spectrum .
the ic @xmath1-ray generation may be suppressed owing to the high magnetic field amplified by the turbulence induced by the shock cloud interaction ( e.g. * ? ? ?
this case requires a relatively small amount of electrons ( for the synchrotron ) , and thus we adopt a low value of the electron - to - proton number ratio , @xmath76 . following @xcite
, the total mass of the gas in the clumps directly bombarded by the accelerated protons , @xmath77 , can be obtained via three steps : ( 1 ) estimate the average mass density @xmath78 for each clump with physical parameters given by ( * ? ? ?
* see their tables 3 and 4 ) .
( 2 ) estimate the mass in the outer layer of each clump that is penetrated by relativistic protons using @xmath79 ( see * ? ? ?
* ) , where @xmath80 is the clump radius and @xmath81 is the penetration depth of the accelerated protons into the clumps during diffusion time @xmath82 .
the diffusion time is estimated as @xmath83 , where @xmath84 ( e.g. , * ? ? ?
* ; * ? ? ?
* ) is the shock velocity and @xmath851pc is the mean length that the shock moves forward from the clumps , estimated from the spatial distribution of the clumps ( sano et al .
( 3 ) scale the mass of the gas collided by protons according to the area fraction covered by x - rays for each clump and sum the mass of the clumps .
hence , we obtain @xmath86 where @xmath87 is the degree of magnetic field fluctuations and @xmath88 is the magnetic field in the dense region of mcs ( see * ? ? ?
we assume that the trapped accelerated protons which uniformly distribute in the downstream have a power - law energy distribution , which can also be described by equation . in order to account for the hess data ,
however , the proton cutoff energy needs to be a free parameter to be decided .
there are eight parameters for the combined contribution from the outside diffusive protons and inside trapped protons ( model iia ) : @xmath44 , @xmath23 , @xmath62 , @xmath63 , @xmath64 , @xmath42 , @xmath89 , and @xmath65 , which are constrained by using the mcmc method .
the 1d probability distributions of model parameters and the best - fit sed are shown in figure [ fig : caseb ] .
the best - fit parameters with 1@xmath66 statistical uncertainties and the @xmath67 values are also listed in tables [ tab : params ] and [ tab : chi2 ] , respectively .
similar to that in model ia , the model @xmath1-ray sed ( see right panel in figure [ fig : caseb ] ) also is the sum of two components , the @xmath1-ray emission from the shocked clumps penetrated by the trapped protons and the cavity wall illuminated by the escaping protons . in the energy range
@xmath22200 gev , the @xmath1-rays are mainly contributed by the escaping protons with a peak flux therein ; in the remaining range , they are dominated by the contribution of the trapped protons . the fluxes above 0.25tev from the escaping ( outer ) and trapped ( inner ) protons are about @xmath90 and @xmath91 , respectively .
the 0.340tev fluxes for the outer and inner components are about @xmath92 and @xmath93 , respectively .
parameters @xmath23 and @xmath63 ( for the escaping protons ) are similar to those in model ia .
the ic @xmath1-ray emission is trivial in this case , as displayed with gray dotted line in figure [ fig : caseb ] .
we also consider the downstream proton
clump interaction as the only mechanism of the gev
tev @xmath1-ray emission ( referred to model iib ) , with the spectral fit parameters also listed in tables [ tab : params ] and [ tab : chi2 ] .
the best - fit sed is displayed by the gray solid line in figure [ fig : caseb ] and shows that the contribution from the interaction between the trapped protons and the shocked clumps alone will also underestimate the gev flux .
as stated above , although the pure ic scattering or the pure shock clump interaction inside the snr can explain the low spectral index at gev band , such simple models tend to underestimate the gev - band flux . in view of the expansion of the snr in a molecular cavity and the presence of the dense molecular gas surrounding the snr , we additionally take into account the role of the energetic protons that escaped from the blast shock during the expansion history . these protons
can contribute the @xmath1-ray emission due to illumination " of the nearby dense gas due to @xmath14@xmath14 collision .
considering that the snr shock is still propagating in the low - density medium ( in which there may be interspersal dense clumps ) , we have calculated this illumination effect on the broadband spectrum of snr rx j1713.7@xmath03946 in a two - zone model . from the fitted sed and the @xmath67 value
, we can see that both model ia and model iia give similarly good results and the hadronic @xmath1-ray emission from the escaping protons plays an important role in the explanation of the spectral data , especially in the gev band .
the flux of this component , peaking around @xmath94gev , naturally compensates the underestimate at the gev band in the aforementioned simple models . from the point of view of spectral shape ,
the hadronic spectrum from the inside clumps ( in model iia ) actually has a similar effect on the total @xmath1-ray emission to the leptonic one ( in model ia ) .
given the presence of the shocked clumps , the dominant mechanism of @xmath1-ray emission in the downstream depends on the electron - to - proton number ratio , @xmath46 : it was found that the hadronic emission from the inside clumps dominates for @xmath46@xmath954@xmath96 while the leptonic component dominates for @xmath46@xmath974@xmath96 .
both models give a small energy transfer fraction @xmath98 for this young snr , similar to those suggested for snrs cas a ( e.g. * ? ? ?
* ) and tycho ( e.g. * ? ? ?
the two models entail different cutoff energies @xmath89 of shock - accelerated protons .
it is constrained as @xmath99tev by the current hess data in model iia , while it is fixed as the knee " energy in model ia .
this leads to the difference of the @xmath1-ray spectrum above @xmath100tev between the two cases ( see figure [ fig : casea ] and [ fig : caseb ] ) .
if we consider bohm diffusion in the acceleration process , the maximum energy of the shock - accelerated protons can be estimated as @xmath101tev ( e.g. * ? ? ?
* ) , which is close to the knee " energy and obviously higher than the fitted value in model iia . therefore , the @xmath1-ray observation above 100tev ( e.g. , the lhaaso project ) is expected to give a constraint on this parameter and then possibly make a distinction between the two cases .
our models imply two emission zones , but , unfortunately , the site of @xmath2gev emission at the cavity wall can not be resolved by the _
fermi_-lat owing to the large point spread function ( psf ) .
however , the spatial distribution of the two - zone emission can be partly reflected in the radial surface brightness profile of the tev emission . with the help of the angular resolution ( better than @xmath102 ) and the increased data set of 150 hr of hess observation
, it is found that the tev @xmath1-ray emission appears significantly more extended than the x - ray one ( * ? ? ?
* and references therein ) .
it may provide evidence of escape of protons that would emit hadronic @xmath1-rays when interacting with surrounding matter @xcite . here
we present a crude comparison of the observed tev @xmath1-ray radial brightness profile with that generated from our two - zone model .
we fit the tev profile ( above 0.25tev ) of the sector - shaped region 3 " given in @xcite , assuming uniform emissivities in the two emission zones . for model ia , a flux ratio ( above 0.25tev ) between the outer and inner components of @xmath24.1:5.9
is adopted from the above model calculation for the @xmath1-ray emission of the entire remnant . in figure
[ fig : pro](a ) , the fitting profile line ( black solid line , which is the sum of the leptonic [ green solid line ] and hadronic [ blue solid line ] components ) can reproduce both the tev brightness peak , which is coincident with the nonthermal x - ray peak , and the broad profile wing that extends outside the x - ray - emitting region .
this fit gives the inner radius of the cavity wall @xmath103 and the thicknesses of both the outer and inner emission regions @xmath104 , where @xmath105pc ( or @xmath106 ) for region 3 . "
the fitted @xmath107 of region 3 " is consistent with the outer edge of the southwestern x - ray arc ( in region 3 " ) shown in the _ xmm - newton _ and _ suzaku _ images @xcite .
this profile fit for region 3 " illustrates that the outstretched @xmath1-ray emission can be ascribed to a contribution from an outer emitting region . for model iia
, the @xmath1-ray emission in the downstream is suggested to mainly come from the shocked clumps . according to sano et al .
( 2013 ) , most of the interacting clumps are distributed on the boundary of the snr .
although their 3-dimensional spacial distribution may be complicated , for simplicity , we assume that they are all close to the snr edge . according to the flux ratio ( above 0.25tev ) between the outer ( escape ) and inner ( trapped ) components , @xmath22.8:7.2
, we fit the tev profile , as in model ia , and obtain the inner radius of the cavity wall @xmath108 and the thicknesses of the outer and inner emission regions @xmath109 and @xmath110 , respectively .
the fitting profile line is plotted by the black dashed line in figure [ fig : pro](a ) and also can reproduce the observed features , including the outer extending part .
similar fittings to the earlier 0.340tev hess data of the entire snr @xcite are also made with model ia and model iia ( see figure [ fig : pro](b ) ) . for model ia , according to the 0.340 tev flux ratio between the outer and inner components , @xmath23.9:6.1 , we obtain the thicknesses of the outer and inner emission zones @xmath111 and @xmath112 , respectively . for model iia , based on the 0.340tev flux ratio between the two zones , @xmath22.7:7.3 , the thicknesses of the two zones are fitted as @xmath113 and @xmath114 , respectively . for both models , @xmath115pc and @xmath116pc are used .
the fitting profiles match the observational data as well .
a dense shell located just outside the shock front without contact has recently been suggested to explain the gev
tev @xmath1-ray emission via a @xmath14@xmath14 interaction mechanism @xcite . in their model ,
the @xmath1-ray emission arises from the cr precursor region , where the accelerated protons directly interact with the target particles without an escape process .
both gev and tev @xmath1-ray emission are dominated by hadronic emission and should be outside the remnant . especially for tev @xmath1-rays ,
the emission is mainly concentrated in the dense shell with a thickness of 0.3pc ( corresponding to @xmath21@xmath117 at a distance of 1kpc ) , which may be difficult to match the broad tev radial profile even if the psf of hess and the projected effect are taken into account . in our model , the emission arises from two zones : the emission from the inside of the snr can be either due to ic scattering or collision of the trapped protons with the shocked clumps , and the outer emission is ascribed to the collision of the protons that escaped from the shock front during the previous expansion with the surrounding dense matter .
the outer emission mainly contributes in the gev band and the inner emission dominates in the tev band .
motivated by the intriguing gev
tev @xmath1-ray spectrum of snr rx j1713.7@xmath03946 and the tev @xmath1-ray radial profile , we have explored the role played by the diffusive relativistic protons that escaped from the snr shock wave in the @xmath1-ray emission , apart from the high - energy particles emission from the inside of the snr . in the scenario that the snr shock is propagating in a molecular cavity , in which there may be interspersal shocked dense clumps , we consider that the @xmath1-ray emission from the inside of the snr may arise either from the ic scattering ( model ia ) or from the interaction between the trapped energetic protons and the clumps ( model iia ) .
the dominant origin between them depends on the electron - to - proton number ratio , above or below @xmath24@xmath118 .
the surrounding molecular cavity wall is considered to also produce @xmath1-ray emission due to the illumination " by the diffusive protons that escaped from the shock wave during the expansion history .
we simplify the algorithm for the @xcite accumulative diffusion " model for diffusive escaping protons .
this two - zone model is fit to the broadband spectrum of the snr that incorporates the updated 5yr fermi data , with application of the mcmc method .
the broadband fluxes can be well explained by the two - zone model , in which the @xmath1-ray emission from the inside governs the tev band , while the outer emission component substantially contributes to the gev @xmath1-rays and naturally compensates for the underestimate of the gev flux of the inner component . in the meantime
, we show that the two - zone model can also reproduce the tev @xmath1-ray radial brightness profile that , at a resolution better than @xmath102 , significantly extends outside the nonthermal x - ray - emitting region .
is indebted to hui li for the helpful discussion on the accumulative diffusion model for escaping protons , to qiang yuan for providing the mcmc code , which is adapted from the cosmomc package @xcite , and to ke - ping qiu and yang su for the helpful discussion .
we thank the support of nsfc grant 11233001 , 973 program grant 2015cb857100 , grant 20120091110048 from the educational ministry of china , and the program b for outstanding phd candidate of nanjing university .
llcccccccc ia & @xmath119 & @xmath120 & @xmath121 & @xmath122 & @xmath123 & @xmath124 & @xmath125 & @xmath126 & @xmath127 + ib & @xmath119 & @xmath128 & @xmath129 & @xmath130 & @xmath131 & @xmath124 & @xmath124 & @xmath124 & @xmath124 + iia & @xmath132 & @xmath133 & @xmath134 & @xmath135 & @xmath136 & @xmath137 & @xmath138 & @xmath139 & @xmath140 + iib & @xmath132 & @xmath141 & @xmath142 & @xmath143 & @xmath144 & @xmath145 & @xmath124 & @xmath124 & llccccc ia & 3.8 & 364.6 & 10.2 ( 13.5@xmath146 ) & 37.1 & 1.78(233 ) + ib & 0.1 & 374.1 & 36.3 ( 5.5@xmath146 ) & 120.0 & 2.25(236 ) + iia & 3.7 & 364.6 & 9.1 ( 7.0@xmath146 ) & 36.4 & 1.78(232 ) + iib & 0.8 & 368.6 & 23.9 ( 3.0@xmath146 ) & 41.7 & 1.85(235 ) 3946 according to the radio @xcite , x - ray @xcite , gev ( _ fermi _ : @xcite , @xcite ) , and tev ( hess : @xcite ) data .
the gray solid line represents the fitting sed for model ib . , title="fig:",height=204 ]
3946 according to the radio @xcite , x - ray @xcite , gev ( _ fermi _ : @xcite , @xcite ) , and tev ( hess : @xcite ) data .
the gray solid line represents the fitting sed for model ib . , title="fig:",height=211 ] but for model iia .
the gray solid line in the right panel represents the fitting sed for model iib . ,
title="fig:",height=204 ] but for model iia .
the gray solid line in the right panel represents the fitting sed for model iib . , title="fig:",height=211 ] -ray radial surface brightness profile of sector region 3 . "
the hess flux data and the _ xmm - newton _ x - ray brightness profile convolved with the hess psf of @xmath102 ( gray line ) are adopted from @xcite .
the radial profile of region 3 " is fitted with a two - zone emission ( in black ) , which is the sum of the inner ( in green ) and outer ( in blue ) components .
the solid and dashed lines represent the profiles in model ia and model iia , respectively .
all of the fitting profile lines have been convolved with the hess psf of @xmath102 .
( b ) tev @xmath1-ray radial surface brightness profile for the entire remnant .
the hess flux data are adopted from @xcite .
the meanings of all lines convolved with the hess psf of @xmath147 are the same as those in ( a ) . , title="fig:",height=211 ] -ray radial surface brightness profile of sector region 3 . "
the hess flux data and the _ xmm - newton _ x - ray brightness profile convolved with the hess psf of @xmath102 ( gray line ) are adopted from @xcite .
the radial profile of region 3 " is fitted with a two - zone emission ( in black ) , which is the sum of the inner ( in green ) and outer ( in blue ) components .
the solid and dashed lines represent the profiles in model ia and model iia , respectively .
all of the fitting profile lines have been convolved with the hess psf of @xmath102 .
( b ) tev @xmath1-ray radial surface brightness profile for the entire remnant .
the hess flux data are adopted from @xcite .
the meanings of all lines convolved with the hess psf of @xmath147 are the same as those in ( a ) .
, title="fig:",height=211 ] | rx j1713.7@xmath03946 is a prototype in the @xmath1-ray - bright supernova remnants ( snrs ) and is in continuing debates on its hadronic versus leptonic origin of the @xmath1-ray emission .
we explore the role played by the diffusive relativistic protons that escape from the snr shock wave in the @xmath1-ray emission , apart from the high - energy particles emission from the inside of the snr . in the scenario that the snr shock propagates in a clumpy molecular cavity , we consider that the @xmath1-ray emission from the inside of the snr may arise either from the inverse compton scattering or from the interaction between the trapped energetic protons and the shocked clumps .
the dominant origin between them depends on the electron - to - proton number ratio .
the diffusive protons that escaped from the shock wave during the expansion history can provide an outer hadronic @xmath1-ray component by bombarding the surrounding dense matter .
the broadband spectrum can be well explained by this two - zone model , in which the @xmath1-ray emission from the inside governs the tev band , while the outer emission component substantially contributes to the gev @xmath1-rays .
the two - zone model can also explain the tev @xmath1-ray radial brightness profile that significantly stretches beyond the nonthermal x - ray - emitting region . in the calculation
, we present a simplified algorithm for @xcite s ( 2010 ) accumulative diffusion " model for escaping protons and apply the markov chain monte carlo method to constrain the physical parameters . |
it has been known for some time that interactions on the 2.73 k blackbody cosmic microwave background ( cmb ) will severely degrade the energies of cosmic ray nucleons with energies beyond @xmath2 the greisen - zatsepin - kuzmin ( gzk ) cutoff @xcite .
it was therefore very surprising when the fly s eye atmospheric fluorescence detector reported the observation of an extremely high energy cosmic ray ( ehecr ) event with an energy of @xmath3 @xcite .
this was followed by the detection of a @xmath4 event by the agasa air shower array @xcite .
these discoveries substantiated earlier claims from the volcano ranch @xcite , haverah park @xcite and yakutsk @xcite air shower arrays that cosmic rays do exist beyond the gzk cutoff .
about a dozen such events are now known .
detailed accounts of the data may be found in recent reviews @xcite . in figure [ fig1 ]
we show the ehecr spectrum for energies exceeding @xmath5 @xcite ; note that the fluxes have been multiplied by @xmath6 .
it is believed that cosmic rays with energies up to @xmath7 , the so - called ` ankle ' , are predominantly of galactic origin , possibly accelerated by the fermi mechanism in supernova remnants @xcite . above this energy ,
the spectrum flattens and the composition changes from being mostly heavy nuclei to mostly protons .
such a correlated change in the spectrum and composition was first established by the fly s eye experiment @xcite and figure [ fig1 ] shows their suggested two - component fit to the data . the new component which dominates at energies beyond @xmath7 is isotropic and
therefore can not possibly originate in the galactic disk @xcite .
however it also extends well beyond the gzk cutoff raising serious problems for hypothetical extragalactic sources .
because of the rapid energy degradation at these energies through photo - pion production on the cmb , such sources must exist within @xmath8 , in fact within @xmath9 for the highest energy fly s eye event @xcite . for heavy nuclei ,
the energy loss is less severe according to a revised calculation @xcite so the range may extend upto @xmath10 .
general arguments @xcite provide correlated constraints on the magnetic field strength and spatial extent of the region necessary to accelerate particles to such high energies and these requirements are barely met by likely astrophysical sites such as active galactic nuclei and the ` hot spots ' of radio galaxies @xcite
. moreover there are few such sources close to us and no definite correlations have been found between their locations and the arrival directions of the most energetic events @xcite .
it has been speculated that gamma - ray bursts which too are isotropically distributed , may be responsible for ehecrs @xcite .
however since these are at cosmological distances , one would expect to see the gzk cutoff in the cosmic ray spectrum contrary to observations ( cf .
ref.@xcite ) . some of the above arguments may be evaded if the ehecr events are due not to nucleons but neutral particles such as photons and neutrinos .
although high energy photons also suffer energy losses in traversing the cmb and the extragalactic radio background , there is no threshold effect which would cause a cutoff near the gzk value @xcite .
however the observed shower profile of the highest energy fly s eye event @xcite argues against the primary being a photon since it would have interacted on the geomagnetic field and started cascading well before entering the atmosphere @xcite .
the observed events are also unlikely to be initiated by neutrinos as they all have incident angles of less than @xmath11 from the zenith and thus too small a path length in the atmosphere for interactions @xcite .
this argument may be evaded if neutrinos become strongly interacting at high energies due to new physics beyond the standard model @xcite , but such proposals are found not to be phenomenologically viable @xcite ( although this is disputed @xcite ) .
( alternatively , the propagating high energy neutrinos could annihilate on the relic cosmic neutrino background , assumed to have a small mass of @xmath12 ev , to make hadronic jets within the gzk zone @xcite . )
other exotic possibilities have been suggested , e.g. monopoles @xcite , stable supersymmetric hadrons @xcite and loops of superconducting cosmic string ( ` vortons ' ) @xcite .
however these possibilities have many phenomenological problems @xcite and we do not discuss them further .
thus one is encouraged to seek ` top - down ' explanations for ehecrs in which they originate from the decay of massive particles , rather than being accelerated up from low energies .
the most discussed models in this connection are based on the annihilation or collapse of topological defects such as cosmic strings or monopoles formed in the early universe @xcite . when topological defects are destroyed their energy
is released as massive gauge and higgs bosons which are expected to have masses of @xmath13 if such defects have formed at a gut - symmetry breaking phase transition .
the decays of such particles can generate cascades of high energy nucleons , @xmath14-rays and neutrinos .
a more recent suggestion is that ehecrs arise from the decays of metastable particles with masses @xmath15 which constitute a fraction of the dark matter @xcite .
these authors suggest that such particles can be produced during reheating following inflation or through the decay of hybrid topological defects such as monopoles connected by strings , or walls bounded by strings .
the required metastability of the particle is ensured by an unspecified discrete symmetry which is violated by quantum gravity ( wormhole ) effects .
another suggestion is that the long lifetime is due to non - perturbative instanton effects @xcite . in ref.@xcite , a candidate metastable particle
is identified in a @xmath16 gut .
a generic feature of these ` top - down ' models is that the ehecr spectrum resulting from the decay cascade is essentially determined by particle physics considerations .
of course the subsequent propagation effects have astrophysical uncertainties but since the decays must occur relatively locally in order to evade the gzk cutoff @xcite , they are relatively unimportant .
thus although the proposal is speculative , it is possible , in principle , to make reliable calculations to confront with data . in this work
we consider another possible candidate for a relic metastable massive particle @xcite whose decays can give rise to the observed highest energy cosmic rays .
first we discuss ( [ crypton ] ) why this candidate , which arises from the hidden sector of supersymmetry breaking , is perhaps physically better motivated than the other suggested relics .
we then undertake ( [ decay ] ) a detailed calculation of the decay cascade using a monte carlo event generator to simulate non - perturbative qcd effects .
this allows us to obtain a more reliable estimate of the cosmic ray spectrum than has been possible in earlier work on both topological defect models @xcite and a decaying particle model @xcite .
we confront our results with observations and identify the mass and abundance / lifetime required to fit the data .
we conclude ( [ concl ] ) with a summary of experimental tests of the decaying particle hypothesis .
soon after the discovery of the anomaly - free heterotic superstring theory in ten dimensions based on the gauge group @xmath17 , it was pointed out @xcite that in the physical low energy theory ( where a grand unified @xmath18 or @xmath19 group is broken by wilson lines ) , the minimum value of magnetic charge is not the dirac quantum @xmath20 but an integral multiple thereof .
conversely , the minimum electric charge is smaller than the electron charge @xmath21 by the same ratio . where @xmath22 is four - dimensional minkowski space and @xmath23 is some compactified six - manifold .
such fractionally charged states exist because @xmath23 is not simply connected these are states in which a closed string wraps around a non - contractible loop in @xmath23 .
] this was found to be a generic feature of all superstring models based on a level - one kac - moody algebra @xcite . in view of the severe experimental upper bounds on the relic abundance of fractional charges
@xcite , this posed a potential embarrassment for superstring phenomenology @xcite .
a simple solution to the problem of fractional charges ( with an obvious historical analogue in quarks and qcd ) is to confine them and it was shown that this can be done in the hidden sector of supersymmetry breaking in the framework of the @xmath24 unification model @xcite . in this model ,
all fractionally charged states have charges @xmath25 and are placed in * 4 * or * 6 * representations of a hidden @xmath26 gauge group which becomes strong at a scale @xmath27 and in * 10 * representations of a hidden @xmath28 group which becomes strong at a scale @xmath29 .
this results in integer - charged particles ` cryptons ' which may be 2-constituent mesons , 3-constituent baryons or 4-constituent ` tetrons ' @xcite
. some of these mesons could be light ( in analogy to the pion of qcd ) but most of the states should be heavy with masses of order the confinement scale @xmath30 .
( other possibilities for stable superstring relics have been discussed in ref.@xcite . )
the constituent fields have very few renormalizable ( @xmath31 ) superpotential interactions , so most of these states can only decay via higher - order ( @xmath32 ) superpotential terms .
generically , crypton lifetimes are expected to be @xcite @xmath33 where , @xmath34 gev is the normalized planck scale , giving @xmath35 for @xmath26 and @xmath28 bound states respectively .
thus @xmath36 for @xmath37 and @xmath38 for @xmath39 .
detailed studies of the possible effects of decays of relic cryptons on primordial nucleosynthesis and the cmb spectrum @xcite , as well as on the diffuse @xmath14-ray background @xcite have established that such particles , if they survive as relics of the big bang , must either decay well before nucleosynthesis or have lifetimes longer than the age of the universe ( @xmath40 ) . in the latter case , if such particles make an interesting contribution to the dark matter , their lifetime is further required to exceed @xmath41 in order to respect experimental bounds on the flux of high energy neutrinos expected from their decays @xcite .
it is seen from eq.([tau10tau4 ] ) that these constraints favour @xmath26 mesons over their @xmath28 counterparts as possible constituents of the dark matter .
it is then natural to contemplate the possibility that such cryptons with a mass of @xmath42 and a lifetime @xmath43 are also responsible for the observed highest energy cosmic rays .
recently the above discussion has been extended to other massive metastable particle candidates in superstring / m - theory @xcite .
these authors discuss constructions with higher - level kac - moody algebras ( necessary to accommodate adjoint higgs representations in ( unified ) models other than @xmath24 ) and note that similar metastable bound states occur in such models .
they go on to consider other candidate particles in m - theory such as kaluza - klein states associated with extra dimensions but find that these are not as attractive , being either too heavy or too unstable .
they suggest that although the @xmath44 model @xcite discussed above was constructed in the weak coupling limit , it may be elevated to an m - theory model in the strong coupling limit .
the @xmath26 tetrons are then still the most likely candidates for massive metastable dark matter with the modification that the planck scale @xmath45 in eq.([taucrypton ] ) may be replaced by a somewhat smaller scale . the main reason why this possibility was not seriously entertained earlier concerns the expected relic abundance of such massive particles .
if cryptons were maintained in chemical equilibrium in the early universe through self - annihilations , their present energy density is given by the usual ` freeze - out ' calculation as inversely proportional to the ( velocity - averaged ) annihilation cross - section @xcite .
estimating this to be @xmath46 we see that equilibrium would have been established if the annihilation rate exceeded the hubble expansion rate ( @xmath47 ) , i.e. at temperatures @xmath48 the relic abundance is then simply estimated as the equilibrium value at decoupling : @xmath49 this is the basis for the conclusion that no stable relic particle may have a mass in excess of @xmath50 without ` overclosing ' the universe , i.e. contributing @xmath51 @xcite .
this does not necessarily apply to cryptons since a period of inflation should have diluted their abundance to essentially zero , along with monopoles and other such supermassive relics .
if the reheating temperature following inflation is restricted to be @xmath52 in order not to produce too many gravitinos @xcite , cryptons would not have been generated afterwards .
however it has been recently recognized that in supersymmetric cosmology , there is likely to be a late stage of ` thermal inflation ' @xcite due to symmetry breaking along flat directions at intermediate scales @xcite .
this would adequately dilute the abundance of thermally generated gravitinos following inflation so the bound quoted above on the reheating temperature is no longer valid and the value of @xmath53 may be much higher . of the scalar ` inflaton ' field
is constrained to be @xmath54 by the anisotropy in the cmb observed by cobe , where the slope parameter @xmath55 is required to be @xmath56 to permit inflation to occur @xcite .
( the number of e - folds of expansion until the end of inflation is just @xmath57 and this should exceed @xmath58 in order to solve the flatness and homogeneity problems of the standard cosmology . ) the reheat temperature @xmath59 can , in principle , have been as high as @xmath60 although it is usually considerably smaller since the inflaton field is very weakly coupled in most inflationary models . ] in that case cryptons even as massive as @xmath61 may well have been brought back into thermal equilibrium during reheating after inflation and survived with the huge relic abundance ( [ crypdens ] ) .
however thermal inflation would also have diluted this to an acceptable level as was noted in ref.@xcite ; to obtain @xmath62 , the number of e - folds of thermal inflation required is just @xmath63 this fits in well with the expectation that @xmath64 for the intermediate scale @xmath65 in the range @xmath66 @xcite . of course given the uncertainty in the value of @xmath67 ( and indeed the possibility that there may be more than one such epoch ) , @xmath68 could well have been reduced to a negligibly small value .
another possibility is that massive particles such as cryptons were never in thermal equilibrium but were created with a cosmologically interesting abundance due to the varying gravitational field during ( primordial ) inflation @xcite .
a cosmologically interesting relic abundance then arises for @xmath69 where @xmath70 is the likely hubble parameter during inflation @xcite .
this is certainly very encouraging but it should be remembered that a later stage of thermal inflation would dilute such an abundance to a negligible level , as discussed above .
it is clear that the relic abundance of massive particles such as cryptons will necessarily be very uncertain given our ignorance of the thermal history of the universe prior to nucleosynthesis .
however as the above discussion illustrates , there are two complementary ways in which a cosmologically interesting abundance may result so we may reasonably consider such particles as candidates for the dark matter .
we now move on to discuss whether relic cryptons can indeed be the source of the ehecr by determining the expected spectrum of high energy particles from their decays .
to calculate the expected flux of cosmic rays from decays of massive particles such as cryptons , we must consider the contribution from both decaying particles in the halo of our galaxy as well as those elsewhere in the universe .
since such massive particles would behave as cold dark matter and cluster efficiently in all gravitational potential wells , their abundance in our galactic halo would be enhanced above their cosmological abundance by a factor @xmath71 note that @xmath72 where @xmath73 is the critical density in terms of the present hubble parameter @xmath74kmsec@xmath75mpc@xmath75 .
if for simplicity we assume a spherical halo of uniform density , @xmath76 then @xmath77 and the number density of cryptons in the halo is @xmath78 the actual density of dark matter in the halo must of course fall off as @xmath79 to account for the flat rotation curve of the disk but we do not consider it necessary at this stage to investigate realistic mass models . thus the universal density of cryptons is smaller than the halo density by about the same numerical factor by which the distance to the horizon ( @xmath80 ) exceeds the halo radius , so the extragalactic contribution to the ehecr flux from decaying cryptons can not exceed the halo contribution .
in particular , the gzk cutoff scale for protons @xcite or heavy nuclei @xcite are all much smaller than the horizon distance , so only the halo contribution need be considered , as was emphasized in ref.@xcite . only for neutrinos
would the extragalactic component be comparable in magnitude @xcite .
henceforth we restrict ourselves to considering crypton decays in the halo alone .
now the injection spectrum from particle decay is , to a good approximation , @xmath81 for lifetimes longer than the age of the universe ( @xmath82 ) . here
@xmath83 is a measure of particle energy ( assuming 2-body decays ) and the fragmentation function @xmath84 is the average number of particles @xmath85 released per decay , per unit interval of @xmath86 , at the value @xmath86 .
the flux at earth is then @xmath87 the final state particles which interest us most are ` protons ' and neutrinos / antineutrinos where the former includes other nucleons , e.g. antiprotons and neutrons , since they all interact similarly in the earth s atmosphere . to compare with observations
we multiply the fluxes by @xmath6 and define @xmath88 for photons and electrons / positrons , propagation energy losses are substantial even within the halo and we do not attempt to determine these
. however their injection spectra from particle decay are given by the same computation as for protons and neutrinos , to which we now turn .
heavy particles , whether gut - scale bosons ( in topological defect models ) , cryptons or other hypothetical massive particles , will decay into quarks and leptons .
the quarks will hadronize producing jets of mostly pions with a small admixture of nucleons and antinucleons .
the neutral pions will decay to give photons while charged pion decays will yield neutrinos and antineutrinos in addition to leptons .
thus the final spectrum of the decay produced particles will be essentially determined by the ` fragmentation ' of quarks / gluons into hadrons .
this is a non - perturbative qcd process and it has not been possible to calculate it by analytic means .
usually phenomenologically motivated approximations are used to model experimental data on inclusive jet multiplicities and scaling violations @xcite .
so far , authors of proposals involving heavy particle decay , e.g. in the context of topological defect models @xcite , have employed a hadronic fragmentation function suggested by hill @xcite @xmath89(1-x)^2}{x\sqrt{\ln(1/x ) } } .\ ] ] it is further _ assumed _ that 3% of the hadronic jets from massive relic particle decays turn into nucleons , while the other 97% are pions which decay into photons and neutrinos .
this was based on the leading logarithm approximation of qcd @xcite applied to experimental data from petra on jet production in @xmath90 collisions at tens of gev .
the estimated jet multiplicity from gluon fragmentation was convoluted with the gluon distribution to determine the total hadron yield ; to estimate the spectrum , it was assumed that the first moment of the distribution is normalized to unity and the large @xmath86 behaviour was guessed to be @xmath91 @xcite .
as we shall see , the hill fragmentation function ( [ hillfrag ] ) significantly _ overestimates _ the yield of high @xmath86 final states from the decay of very massive particles and , moreover , photons and neutrinos are actually produced with a spectrum quite different from that of nucleons . thus the decay spectra derived using eq.([hillfrag ] ) for topological defect models @xcite are inaccurate .
subsequently , another form called the modified leading logarithm approximation ( mlla ) which gives a better description of data at low @xmath86 has been proposed @xcite ; a gaussian approximation to this is @xmath92 , \ ] ] where @xmath93 is a constant and @xmath94^{3/2 } , \ ] ] with @xmath95 and @xmath96 .
this fragmentation function is employed by the authors of ref.@xcite to compute the spectrum from relic particle decays ; they determine @xmath93 by requiring that the integral of @xmath97 over the range @xmath98 $ ] be equal to the fraction of the energy transferred to hadrons .
however this procedure is not exact as the form ( [ mlla ] ) is inapplicable for large @xmath86 and therefore can not be normalized in this manner .
thus the shape of the cosmic ray spectrum computed @xcite by this method for decaying particles is only reliable for small @xmath86 and its normalization uncertain . given the importance of determining the energy spectrum accurately , we decided to improve on these approximate formulations by using the standard tool employed by experimental high energy physicists to study qcd fragmentation , viz . a monte carlo event generator . here
the non - perturbative hadronization process is simulated on a computer by a well tested phenomenological model @xcite .
although this requires extensive numerical calculations , it is the only means by which successful contact can be made between theory and experimental data .
we chose the programme herwig @xcite ( hadron emission reactions with interfering gluons ) which incorporates the cluster model for hadronization and is based on a shower algorithm of the ` jet calculus ' type @xcite . to check our results
we also ran the jetset programme @xcite and found good agreement over the energy range where comparison was possible .
however for the very high energies studied in this work , herwig proved to be more suitable for reasons of computing time @xcite . even so the calculations described here took several months on a digital alpha workstation . for definiteness , we assume the heavy particles to decay into a quark - antiquark pair with unit branching ratio .
the quark and antiquark , each carrying away energy @xmath99 , form jets which lead to the generation of many particles through cascading , hadronization and decays of some of the generated particles .
this can be simulated by herwig via the annihilation process @xmath100 with center - of - mass energy @xmath101 , where @xmath102 stands for all six kinematically allowed quark flavours .
the event generator outputs kinematical details of all final state particles , e.g. protons , photons and leptons ( electrons , positrons and neutrinos ) .
we divided the @xmath86-range into 100 bins of width @xmath103 .
after each event simulation the number of protons , neutrinos , photons as well as electrons and positrons per energy bin was counted .
we ran 10000 events for each of the masses @xmath104 and @xmath105 .
after all events had been run , the particle numbers in the bins are divided by the bin width and the number of events , in order to obtain the fragmentation functions @xmath84 .
apart from altering some relevant parameters in the computer code to allow it to run at the high energies studied here , we also switched off initial state radiation since it is not relevant for the present study .
unfortunately , it was not feasible to study the high @xmath86 behaviour of the fragmentation functions for decaying particle masses higher than @xmath105 because of numerical convergence problems in the computer code .
( already for masses exceeding @xmath106 quadruple precision had to be used . )
hence we have had to extrapolate the fragmentation functions to high @xmath86 for very heavy masses as described later .
first we show the proton fragmentation function obtained from the herwig runs in figure [ fig2 ] to illustrate that it depends on the decaying particle mass , contrary to the approximation ( [ hillfrag ] ) employed in previous work on topological defects @xcite . rather than being constant ,
it decreases with increasing @xmath107 for @xmath108 , while at very low @xmath86 it increases with increasing @xmath107 .
the large fluctuations at @xmath109 are due to the fact that relatively few particles are produced with such high energies despite the 10000 events per simulation .
we note also that the shape differs significantly at high @xmath86 from the approximation used in ref.@xcite . in figure [ fig3 ] the fragmentation functions for protons , photons , neutrinos and electrons
are compared for @xmath110 .
it is seen that at very low @xmath86 there are more photons and neutrinos generated by the particle decay than electrons and protons . in the regime @xmath111 , photons , neutrinos and protons are generated with roughly equal abundances .
however , for @xmath109 , photons and neutrinos again outnumber protons , in particular protons cut off at @xmath112 whereas neutrinos and photons are generated in the cascades with energies up to @xmath113 .
these differences will lead to different shapes of the expected fluxes @xmath114 as can be seen from eq.([fluxi ] ) .
we now compare our proton fragmentation function with the commonly used hill approximation @xcite in figure [ fig4 ] .
although his form provides a good fit for a low decaying particle mass , viz .
@xmath115 , it no longer does so for a high mass , viz .
this is understandable given that the numerical co - efficients in eq.([hillfrag ] ) were chosen to match relatively low energy collider data .
however the functional form itself is well motivated and using our herwig runs we can determine new numerical co - efficients appropriate to heavier mass particles .
another advantage of the present approach is that the spectrum of neutrinos and photons is determined separately from that of the protons and not simply assumed to be proportional as in previous work @xcite . to study the highest energy cosmic ray events we need to consider particle masses beyond @xmath105 but this is difficult to do directly with herwig for technical reasons as mentioned earlier .
we therefore resort to an extrapolation procedure as follows . for the range
@xmath116 $ ] the fragmentation functions are smooth and evolve monotonically with @xmath107 so the fragmentation functions for a @xmath117 particle is obtained from simple linear extrapolation of the lower energy fragmentation functions in each individual energy bin . for @xmath118 $ ]
we first fit the calculated fragmentation functions to the form @xmath119(1-x)^2}{x\sqrt{\ln(1/x ) } } , \ ] ] for protons , and the form @xmath120(1-x)^2}{x\ln(1/x ) } , \ ] ] which proves more suitable for photons , neutrinos and electrons .
the numerical co - efficients @xmath121 and @xmath122 are determined for particle masses less than @xmath105 by minimizing @xmath123 in the fit to the actual herwig runs . in figures
[ fig5 ] and [ fig6 ] we show these fits for @xmath124 to the proton and neutrino fragmentation functions corresponding to masses of @xmath125 and @xmath126 .
then we determine the appropriate co - efficients for heavier masses by extrapolation .
an example , for @xmath127 , is shown in the figures . for @xmath128 ,
statistical fluctuations become too severe so we extrapolate the fitting functions between the value at @xmath129 and a cutoff which is taken to be @xmath130 for protons and @xmath131 for neutrinos , based on the observed behaviour for masses upto @xmath105 shown in figures [ fig2 ] and [ fig3 ] . finally , we mention the continuation of the proton fragmentation function for very low @xmath86 , viz .
@xmath132 , which is relevant at high masses e.g. @xmath127 .
since it proved impractical to have additional binning intervals at very small @xmath86 , we employ the fragmentation function ( [ mlla ] ) in this regime , normalized to our computations at @xmath133 . with the fragmentation functions obtained above , we can now calculate the expected fluxes of protons and neutrinos from decays of particles such as cryptons in the halo .
we normalize the calculated proton flux ( [ fluxi ] ) to the observed cosmic ray flux at @xmath134 @xcite : @xmath135 = 24.32 .\ ] ] note that the corresponding neutrino flux @xmath136 is then a _ prediction _ as the fragmentation function for neutrinos is computed independently .
the expected proton fluxes are shown in figure [ fig7 ] .
we see that a crypton with @xmath110 fits the flat power law well but can not explain the events beyond @xmath137 . although this is easily achieved for @xmath127 , the decays of such a massive particle would overproduce protons for @xmath138 .
thus a crypton with mass @xmath139 provides the best compromise although it too predicts a spectrum somewhat flatter than the one indicated observationally .
( the reader is reminded that all differential fluxes have been multiplied by @xmath6 in eq.([fluxi ] ) . )
an interesting signature for forthcoming experiments is the predicted ratio of the proton to neutrino flux @xcite . in figure [ fig8 ]
we compare the expected flux of protons and neutrinos for @xmath140 .
( we also show the photon flux to illustrate the difference from the prediction in ref.@xcite but emphasize that this will be degraded through interactions with photon backgrounds during travel to earth . ) as can be seen , the neutrino flux exceeds the proton flux for @xmath141 and also for @xmath142 , as may have been anticipated from the comparison of their respective fragmentation functions .
thus the ratio @xmath143 has a characteristic peak at about @xmath144 as shown in figure [ fig9 ] .
this could be an useful diagonistic of the decaying particle hypothesis for future experiments such as the pierre auger project .
note that taking the extragalactic contribution into account would boost the neutrino flux by a factor of @xmath145 over that shown in the figures .
the abundance and lifetime of decaying particles such as cryptons are related through the spectrum normalization ( [ fenorm ] ) as : @xmath146 where @xmath147 for crypton masses @xmath148 respectively . for a given crypton mass ,
a higher lifetime must be compensated for by a higher relic abundance , as illustrated in figure [ fig10 ] .
so for example , if @xmath149 , cryptons with a mass of @xmath140 are required to have a lifetime of @xmath150 if they are to explain the ehecr flux .
if the enhancement in the halo is @xmath151 as expected for cold dark matter , then the lifetime may be increased to @xmath152 if @xmath153 ; alternatively , for the same lifetime one could tolerate a lower relic abundance @xmath154 . with regard to the fluxes of electrons and photons
, both species would generate electromagnetic cascades on the prevalent radiation backgrounds through pair production and inverse compton - scattering .
a thorough analysis of such propagation effects and the resulting modifications of the injected photon and electron spectra has been performed @xcite .
it was found that the relic decaying particles with @xmath155 would contribute excessively to the diffuse @xmath14-ray background and are therefore ruled out .
hence , the mass range we favour , viz .
@xmath156 , does not lead to any conflict with observations .
this conclusion is strengthened by the fact that according to our calculations the previous estimate @xcite of the @xmath14-ray flux from decaying particles was too high .
although the positrons released in the decays may be accumulated in the galactic halo , the astrophysical uncertainty in the containment time does not allow a restrictive constraint to be derived from limits on the positron flux in cosmic rays @xcite .
with regard to the neutrino background , the predicted flux at high energies is well below the upper limits derived from consideration of horizontal air showers @xcite , again because the decaying crypton mass is restricted to be less than about @xmath117 .
it is also interesting to consider the flux at lower energies of @xmath157 where experiments such the forthcoming antares detector @xcite will be most sensitive . as seen in figure [ fig8 ] the predicted neutrino flux dominates over the proton flux at low energies , thus the bulk of the energy released by the decaying cryptons ends up as neutrinos . therefore we expect the neutrino flux at tev energies to be at least @xmath158 times larger than the ehecr flux at @xmath159 .
moreover the neutrinos should be well correlated in both time and arrival direction with the cosmic rays since the path length in the galactic halo is @xmath160 .
this is in contrast to the case of other suggested cosmologically distant sources such as gamma - ray bursts where the relative time delay can be upto @xmath161 @xcite .
we have investigated the hypothesis that the highest energy cosmic rays , in particular those observed beyond the gzk cutoff , arise from the decay of massive metastable relic particles which constitute a fraction of the dark matter in the galactic halo .
to simplify computations ( using the herwig monte carlo event generator ) we have considered only decays into @xmath162 pairs with unit branching ratio .
comparison with experimental data indicates that a decaying particle mass of @xmath163 is required to fit the spectral shape while the absolute flux requires a lifetime of @xmath164 if such particles contribute the critical density .
the predicted decay spectra may be somewhat altered if 3-body decays and other final states ( e.g. supersymmetric particles @xcite ) are considered .
however our conclusions regarding the preferred mass and relic abundance / lifetime of the decaying particle are unlikely to be affected .
in particular it would appear that the approximations used to calculate the particle spectra in previous studies of decaying topological defects @xcite and hypothetical massive particles @xcite were not sufficiently accurate .
our work indicates that the topological defect model is disfavoured unless the mass of the decaying gauge bosons is less than about @xmath117 , which is well below the unification scale of @xmath165 .
( a similar conclusion is arrived at by independent arguments in refs.@xcite . ) by contrast , cryptons from the hidden sector of supersymmetry breaking have a mass of the required order , as well as a decay lifetime which is naturally suppressed .
however their relic abundance is difficult to estimate reliably , although we have argued that it may be cosmologically interesting .
the primary intention of this work is to attempt to quantify the decaying particle hypothesis in a manner which is of interest to experimentalists .
we have therefore computed the expected neutrino to proton ratio as a function of energy since this is an important test of competing hypotheses for forthcoming experiments , in particular the pierre auger project @xcite .
of course our cleanest prediction is that the cosmic ray spectrum should cut off just below the mass of the decaying crypton , at @xmath166 .
moreover , with sufficient event statistics it should be possible to identify the small anisotropy which should result from the distribution of the decaying particles in the galactic halo @xcite .
thus although the hypothesis investigated here is very speculative , it is nevertheless testable .
perhaps nature has indeed been kind to us and provided a spectacular cosmic signature of physics well beyond the standard model .
efimov et al .
icrr symp . on astrophysical aspects of the most energetic cosmic rays , eds .
m. nagano and f. takahara ( world scientific , 1990 ) p.20 ; + b.n .
afanasiev et al . , proc .
24th intern .
cosmic ray conf . 2 ( 1995 ) 756 .
watson , nucl .
suppl . ) 22a ( 1991 ) 116 ; proc .
dpf summer study on high energy physics , snowmass , 1994 , eds .
e.w . kolb and r.d .
peccei ( world scientific , 1995 ) p.126 ; + p. sokolsky , p. sommers and b.r .
dawson , phys . rep . 217
( 1992 ) 225 ; + j.w .
cronin , nucl .
suppl . ) 28b ( 1992 ) 213 ; + s. yoshida and h. dai , astro - ph/9802294 .
aharonian , b.l .
kanewski and v.a .
sahakian , j. phys .
g17 ( 1991 ) 1989 ; + h.p .
vankov and p.v .
stavrev , phys .
lett . b226 ( 1991 ) 178 ; + f. halzen , r. vazquez , t. stanev and h.p .
vankov , astropart .
phys . 3 ( 1995 ) 151 ; + t. stanev and h.p .
vankov , phys .
d55 ( 1997 ) 1365 .
hill , d.n .
schramm and t.p .
walker , phys .
d36 ( 1987 ) 1007 ; + p.
bhattacharjee , c.t . hill and d.n .
schramm , phys .
69 ( 1992 ) 567 ; + f.a .
aharonian , p. bhattacharjee and d.n .
schramm , phys .
d46 ( 1992 ) 4188 ; + g. sigl , space sci .
rev . 75 ( 1996 ) 375 .
g. lazarides , c. panagiotakopoulos and q. shafi , phys .
56 ( 1986 ) 557 ; nucl .
b307 ( 1988 ) 937 ; + k. enqvist , d.v .
nanopoulos and m. quiros , phys .
169b ( 1986 ) 343 ; + o. bertolami and g.g .
ross , phys .
lett . 183b ( 1987 ) 163 ; + j. ellis , k. enqvist , d.v .
nanopoulos and k. olive , phys .
b188 ( 1987 ) 415 , b225 ( 1989 ) 313 .
l. dokshitzer , v.a .
khoze , a.h .
mueller and s.i .
troyan , _ basics of perturbative qcd _
( editions frontieres , 1991 ) ; + r.k .
ellis , w.j .
stirling and b.r .
webber , _ qcd and collider physics _ ( cambridge university press , 1996 ) ; + v.a .
khoze and w. ochs , int .
a12 ( 1997 ) 2949 . | the expected proton and neutrino fluxes from decays of massive metastable relic particles are calculated using the herwig qcd event generator .
the predicted proton spectrum can account for the observed flux of extremely high energy cosmic rays beyond the greisen - zatsepin - kuzmin cutoff , for a decaying particle mass of @xmath0 gev .
the lifetime required is of @xmath1 yr if such particles constitute all of the dark matter ( with a proportionally shorter lifetime for a smaller contribution ) .
such values are plausible if the metastable particles are hadron - like bound states from the hidden sector of supersymmetry breaking which decay through non - renormalizable interactions .
the expected ratio of the proton to neutrino flux is given as a diagonistic of the decaying particle model for the forthcoming pierre auger project . -3cm^-3 |
the most interesting features of perfect metallic magnetic multilayer are the oscillatory behavior of bilinear exchange coupling energy between ferromagnets @xcite due to friedel oscillations , and large magnetoresistance in small magnetic fields @xcite . the magnetic structure of adjoining magnetic layers in perfect multilayered structure oscillates between ferro and antiferro - magnetic states with increasing the spacer thickness @xmath1 . disorder in layers contributes to biquadratic exchange coupling .
it was shown that fluctuations of thickness of the paramagnetic layer give rise to biquadratic exchange coupling , often leading to non - collinear magnetic ordering @xcite . in ref .
the role of scattering of conducting electrons by impurities in metallic magnetic multilayer was studied .
it was pointed out that in the case of small ( compared to thickness of the layers ) mean free path of conducting electrons , when average friedel oscillations are exponentially suppressed , the exchange coupling energy due to random friedel oscillations and correlation effects can have biquadratic form .
transition to the non - collinear phase in disordered structure with increasing @xmath1 was experimentally observed in ref . .
it is established that the magnetoresistance of perfect metallic magnetic multilayered structure is related to spin - depended scattering of conducting electrons at the interfaces between the layers @xcite . in the case of disorder ,
scattering of conducting electrons by impurities suppresses the effect of spin - dependent scattering on magnetoresistance .
moreover , it was theoretically shown that the magnitude of the magnetoresistance decreases exponentially in the case of the current - in - plane ( cip ) geometry , when thickness of the nonmagnetic spacer exceeds the mean free path of conducting electrons @xcite . in the present paper
we study the effect of electron - electron interactions on cip conductivity and exchange coupling energy of disordered metallic multilayered structure consisting of two ferromagnetic layers with paramagnetic spacer .
we consider the case when mean free path of conducting electrons is smaller than thicknesses of layers .
it is known that in the disordered conductors electron - electron interactions result in anomalous contributions to conductivity , thermodynamic quantities and negative magnetoconductivity @xcite .
physics behind the effect of electron - electron interactions in disordered conductors is the electron scattering by the random friedel oscillations @xcite .
freidel oscillations in magnetic multilayered structure do depend on the angle @xmath2 between directions of magnetization in magnetic layers .
the study of this dependence is the subject of this paper .
the scattering by freidel oscillations exists in any type of ferromagnetic structure : itinerant ( _ d- _ type ) or localized ( _ f- _ type ) . in what follows , the most relevant factors of disordered magnetic multilayered structure are the characteristic of the disorder - the thouless energy @xmath3 ( @xmath4 is the conducting electron diffusion constant ) , and ferromagnetic splitting of conducting electrons . because of that we consider the model of localized _ s - f _ magnetism that contains these parameters in the most transparent way .
we propose that each ferromagnetic layer is described by the homogeneous magnetization , and that directions of magnetizations of different layers make an angle @xmath2 .
the conductivity and exchange coupling energy in magnetic multilayered structure are periodic functions of @xmath2 .
we show that depending on the ratio of introduced above characteristic energies , the angle @xmath2 dependent contributions of electron - electron interactions experience the @xmath5 periodicity transition .
magnitudes of contributions are estimated .
we suppose that our results might be relevant to the series of works related to transport properties of mesoscopic ferromagnets with domain walls , where it was experimentally @xcite and theoretically @xcite shown that effects of electron interactions and weak localization are important .
we would like to mention that results of this paper complement the theory of electron - electron interactions in disorders conductors @xcite as a study of effect of electron - electron interactions in spatially inhomogeneous effective magnetic field that is imposed by the magnetization in layers .
the system of our study is disordered metallic magnetic multilayer structure .
it consists of two ferromagnetic layers of the @xmath6-type located at @xmath7 and a paramagnetic layer between ( @xmath8 ) .
the hamiltonian of conducting electrons of this system is : @xmath9 here @xmath10 is the hamiltonian of free electrons in a random field @xmath11 . treating scattering of free electrons we use the standard diagram technique @xcite , which assumes @xmath12 and @xmath13 . here
@xmath14 is the density of states at the fermi level per one spin , @xmath15 is the electron mean free time , and we have set @xmath16 .
we assume the hamiltonian of free electrons to be the same in every layer . the second term @xmath17 is the hamiltonian of the coulomb interaction between conducting electrons .
@xmath18 , where @xmath19 and @xmath20 are the electron creation and annihilation operators correspondingly .
we will treat coulomb interaction within the random phase approximation .
the third term @xmath21 is the hamiltonian of @xmath22 exchange in the ferromagnetic layers . at temperatures much lower than the curie temperature
one might neglect electron - magnon interaction , therefore @xmath23 here @xmath24 is a spin of localized @xmath6-electrons , @xmath25 is their density .
@xmath26 is the @xmath22 exchange interaction .
@xmath27 is the ferromagnetic layer magnetization direction unit vector .
integration here is over the ferromagnetic layers . neglecting the contribution of electron - magnon interaction to the conductivity and exchange coupling energy , we assume that the coulomb energy per electron is larger than the ferromagnetic splitting @xmath28 .
finally we consider the following hamiltonian of a disordered metallic magnetic multilayer : @xmath29 this hamiltonian describes a fermi liquid in an effective inhomogeneous magnetic field that acts only on electron spin .
this magnetic field results in the spin splitting with energy @xmath30 in the ferromagnetic layers , and in zero splitting in the paramagnetic layer .
direction of this magnetic field is defined by the ferromagnetic layer magnetization direction unit vector @xmath31 : @xmath32at @xmath33 and @xmath34at @xmath35 , with @xmath36 .
now we are ready to study the effect of electron interaction on the cip conductivity and exchange coupling energy between the ferromagnetic layers
. we will be interested in its dependence on the angle between the direction of magnetization @xmath2 , ferromagnetic splitting energy @xmath37 , and paramagnetic thickness @xmath1 .
we use kubo linear response formalism in order to study the effect of electron interaction on conductivity . treating electron interaction as a perturbation ,
we come up with corrections to conventional drude formula for conductivity .
we study only the first order perturbation correction .
the magnetic field dependent , or in our case magnetization dependent , correction to conductivity comes from the interaction between two electrons in the triplet state ( so - called the hartree - type correction ) @xcite .
disorder averaged correction can be described with the help of a feynman diagram , shown in the fig .
black squares in the figure stand for the diffusion ladder that represents electrons charge and spin densities propogation .
calculations of presented diagrams for the homogeneous magnetic field are carried out in ref . .
we make use of the bilinear representation of the diffusion ladder in order to calculate these diagrams for inhomogeneous magnetic field .
our calculations show that in the case of the cip geometry all diffusion ladder combinations that appear in the figure can be simplified in to the following expression : @xmath39 where @xmath40 is the electron interaction constant that is discussed in ref . .
we assume that interaction constants might be different in ferromagnetic and paramagnetic layers .
@xmath41 is the temperature , @xmath42 is the diffusion ladder . for the cip geometry
it is convenient to consider conductance @xmath43 as a sum over the paramagnetic and ferromagnetic layers .
contribution to the multilayer thermodynamic potential , which depends on the relative orientation of the ferromagnetic magnetizations , is also due to interaction between two electrons in the triplet state .
this correction is presented in the fig . @xmath44 and given by an expression @xcite @xmath45 where @xmath46 is the matsubara frequency , @xmath47 is the area of the system .
the angle - dependent part of @xmath48 determines exchange coupling energy of adjoint ferromagnetic layers .
the angle @xmath2 dependence in expressions ( [ cip ] ) and ( [ coupling ] ) is due to the diffusion ladder .
graphical equation for the diffusion ladder @xmath49 is shown in the fig .
@xmath50 , and in the case of diffusion approximation ( @xmath51 , @xmath52 , and @xmath53 with @xmath54 being the fermi velocity ) satisfies the following differential equation @xmath55 here @xmath56 is the diffusion constant .
we assume it to be the same in every layer .
we notice that in expressions ( [ cip ] ) and ( [ coupling ] ) only trace of the diffusion ladder appears .
since the trace is invariant under unitary transformations , we may choose the ferromagnetic layers magnetization direction unit vector in the convenient for further calculations : in @xmath57 the direction is @xmath58 , and in @xmath35 it is @xmath59 .
solution of the equation ( [ dif ] ) has to satisfy the boundary conditions @xmath60 at @xmath61 , and the continuity conditions at @xmath62 .
here we present our main results by skipping all derivations that will be given in the appendix to the paper .
we consider only angle - dependent parts of ( [ cip ] ) and ( [ coupling ] ) .
characteristic energies of a magnetic multilayer are the thouless energy of the paramagnetic layer @xmath3 and ferromagnetic splitting energy @xmath64 .
we present results for the cases of large ferromagnetic thickness @xmath65 : @xmath66 and for small temperatures : @xmath67 .
our main finding is the following : depending on the ratio of the thouless energy and ferromagnetic splitting energy , contribution to the cip conductance ( [ cip ] ) and exchange coupling energy between ferromagnetic layers ( [ coupling ] ) might have @xmath68 or @xmath69 periodicity as a function of @xmath2 .
@xmath70 transition occurs when ratio @xmath71 increases .
\(i ) in the case of large thouless energy @xmath72 the leading term the in angle - dependent parts of the cip conductance and exchange coupling energy is proportional to @xmath73 . one can show that ferromagnetic layers give the main contribution to the conductance as well as to exchange coupling energy : @xmath74 @xmath75 correction to conductance originating from the paramagnetic layer is reduced by a factor of @xmath76 .
\(ii ) in the limit of @xmath77 the leading term in ( [ cip ] ) and ( [ coupling ] ) is @xmath78 , and it is independent on the ferromagnetic splitting energy @xmath64 : @xmath79@xmath80 the physics behind this transition is the following . according to ref . the magnetic field dependent contributions to the conductivity and exchange coupling energy are due to interaction of electrons with total spin @xmath81 and projection @xmath82 .
in our case when direction of @xmath83 varies in space , projections @xmath84 are defined only locally . at small @xmath64
the leading contribution is due to interaction of electrons with total spin @xmath81 and projections @xmath82 , and the dependence of the conductivity and of the exchange coupling energy is proportional to @xmath85 . in the case of large @xmath64 diffusion modes with total spin @xmath81 and projection @xmath82 do not penetrate into ferromagnetic layers and their contributions to the conductivity and the exchange coupling energy are suppressed .
the main contribution in this case is due to the diffusion mode with @xmath81 , @xmath86 and proportional to @xmath78 .
let us note , that in this case the electron - magnon interaction , which creates @xmath82 diffusion modes in ferromagnetic layers might be neglected even if the coulomb energy per electron is smaller than the ferromagnetic splitting . at temperatures @xmath87 the contribution to the conductance decreases as @xmath88 \cos ^{2}\varphi .
\label{9a}\ ] ] and exchange coupling energy decreases with the temperature as @xmath89 . in the fermi liquid theory
the interaction constants are considered as some parameters . in the case of weakly non ideal 3d fermi gas @xmath90 , where @xmath91 and @xmath92 are the inverse screening length and the fermi momentum , correspondingly @xcite . for the coulomb interaction @xmath93 and @xmath94 are both positive . in both limits , ( i ) and
( ii ) , contribution to conductance has a minimum at @xmath95 .
exchange coupling energy has a minimum at @xmath96 in the first limit , and at @xmath97 in the second limit .
let us estimate the value of exchange coupling energy .
taking @xmath98 , @xmath99 ( where @xmath100 is the boltzmann constant ) , we obtain that crossover occurs at the paramagnetic layer thickness of @xmath101 .
plugging these values into equations ( [ 6a ] ) and ( [ 8 ] ) , and assuming @xmath102 we obtain @xmath103 .
correction to conductance is of the order of percents of @xmath104 .
in the present paper we have considered properties of a disordered magnetic metallic multilayer in the case of small electron mean free path compared to paramagnetic layer thickness .
the angle @xmath2 dependence of the cip conductance and of the exchange coupling energy is determined by the electron - electron interactions .
depending on the ratio of the thouless and ferromagnetic splitting energies these quantities might be @xmath105 as well as @xmath106 periodic functions of @xmath2 .
obtaining our results we have neglected the effect of spin - orbit scattering .
this scattering smears out the magnetic dependence of electron interaction corrections to thermodynamic and kinetic properties@xcite .
however , this effect become crucial for paramagnetic thickness ( larger than the spin relaxation length due to spin - orbital scattering ) when values of calculated quantities already become too small
. we would like to point out that in ref .
contribution to exchange coupling energy due to the interaction was considered only in the paramagnetic layer . in this paper
we have shown that the contribution of ferromagnetic layers is larger than that of paramagnetic layer .
this work is supported by the russian fund for fundamental research grants 05 - 02 - 17816 and 06 - 02 - 17047 .
here we describe the main steps of derivation .
we solve the differential equation ( [ dif ] ) for diffusion ladder . according to the expressions ( [ cip ] ) , ( [ coupling ] )
the solutions of interest are those with @xmath107 and @xmath108 lying in the same layer .
\(i ) let us first derive the diffusion ladder when @xmath109 is in ferromagnetic layer .
it is convenient to present solution in paramagnetic layer ( @xmath110 ) in the following form : @xmath111 and in ferromagnetic layer at @xmath112 solution that satisfies the boundary condition at @xmath113 is convenient to present in the following form : @xmath114 where @xmath115where @xmath116 are projectors on the spin up and down states correspondingly ; @xmath117 , @xmath118 , and @xmath119 is the matrix of rotation along the @xmath120 axis .
@xmath121 are matrixes - constants to be determined with the help of continuity conditions .
it is convenient to obtain closed set equations for matrixes @xmath122 with the help of the continuity conditions for rotated diffusion ladder ( [ diffpara1 ] ) @xmath123 at @xmath124 : @xmath125 where we have used the differential operator @xmath126 and continuity conditions for diffusion ladder ( [ diffpara1 ] ) at @xmath127 are @xmath128 with @xmath129 .
@xmath131 } { w_{+}}+\left ( \lambda\rightarrow-\lambda,\lambda_{1}\rightarrow-\lambda_{1}\right ) \right ] \frac{1}{d}\left [ \frac{\sinh\left ( 2qd\right ) } { q}+2d\right ] \nonumber\\ & + \frac{8q}{d}\operatorname{re}\left [ \frac{2+\lambda+\lambda\lambda _ { 1}+\left ( 2\lambda_{1}+\lambda\lambda_{1}+\lambda\right ) \cos\varphi } { 2q_{1}\left ( q^{2}-q_{1}^{2}\right ) w_{+}}+\left ( \lambda\rightarrow -\lambda,\lambda_{1}\rightarrow-\lambda_{1}\right ) \right],\end{aligned}\ ] ] \(ii ) the same procedure can be applied for the case when @xmath133 is in paramagnetic layer .
now the solution of equation @xmath134 when @xmath107 is in paramagnetic layer is convenient to present in the following form : @xmath135 in this case the continuity conditions for the rotated diffusion ladder ( [ diffpara2 ] ) @xmath123 at @xmath124 are the same as ( [ bc1 ] ) . at @xmath127 continuity conditions for the diffusion ladder ( [ diffpara2 ] ) change as @xmath136 solving for matrixes @xmath130 and subtracting angle independent part we obtain the diffusion ladder for paramagnetic layer ( [ diffpara2 ] ) , whose trace is @xmath137 -\frac{l\operatorname{re}\lambda^{\ast}\left ( 1-\lambda\right ) } % { 2dqw_{-}}\left [ \left ( -\cos\varphi+\lambda_{1}\right ) + \left ( 1-\lambda_{1}\cos\varphi\right ) \frac{\sinh{ql}}{ql}\right ] \nonumber\\ & + \lambda_{1}l\frac{\operatorname{re}\left [ \left ( 1+\lambda^{\ast } \right ) \left ( 1+\lambda\cos\varphi\right ) \right ] \frac{\sinh{ql}}{ql}-\operatorname{re}\left [ \left ( 1+\lambda^{\ast}\right ) \left ( \lambda+\cos\varphi\right ) \right ] } { 2dqw_{+ } } + \lambda_{1}l\frac{\operatorname{re}\left [ \left ( 1-\lambda^{\ast } \right ) \left ( 1-\lambda\cos\varphi\right ) \right ] \frac{\sinh{ql}}{ql}-\operatorname{re}\left [ \left ( 1-\lambda\right ) \left ( \lambda^{\ast } -\cos\varphi\right ) \right ] } { 2dqw_{-}}.%\end{aligned}\ ] ] the results ( [ 5a])@xmath138([9a ] ) will be recovered if one will plug diffusion ladder traces ( [ ferr ] ) and ( [ para ] ) into the expressions for the conductivity ( [ cip ] ) and exchange coupling energy ( [ coupling ] ) .
the integrals over the frequency in resulting expressions can be evaluated as following : in the case of @xmath139 the main contribution is due to frequencies @xmath140 where @xmath141 . in this case
expressions ( [ cip ] ) and ( [ coupling ] ) are even functions of @xmath142 . in region from @xmath143 @xmath144 the main contribution is due to @xmath145 in this case @xmath146 , and expressions ( [ cip ] ) and ( [ coupling ] ) have the 2@xmath106 periodicity .
p. grunberg , r. schreiber , y. pang , m.b .
brodsky , and h. sowers , phys .
* 57 * , 2442 ( 1986 ) .
parkin , n. more , and k.p .
roche , phys .
* 64 * , 2304 ( 1990 ) .
baibich , j.m .
broto , a. fert , f. nguyen van dau , f. petroff , p. eitenne , g. creuzet , a. friederich , and j. chazelas , phys .
lett . * 61 * , 2472 ( 1988 ) .
g. binasch , p. grunberg , f. saurenbach , and w. zinn , phys .
b * 39 * , 4828 ( 1989 ) . j.c .
slonczewski , phys.rev.lett . * 67 * , 3172 ( 1991 ) .
a. zyuzin , b.spivak , i.vagner , and p. wyder , phys.rev.b * 62 * , 13899 ( 2000 ) .
a.yu . zyuzin , jetp lett . * 76 * , 495 ( 2002 ) .
p. fuchs , u. ramsperger , a. vaterlaus , and m. landolt , phys .
b * 55 * , 12546 ( 1997 ) . r.e .
camley and j. barnas , phys .
lett . * 63 * , 664 ( 1989 ) .
s. zhang , p.m. levy , and a.fert , phys . rev .
b * 45 * , 8689 ( 1992 ) .
nesbet , j. phys .
* 6 * , l449 ( 1994 ) .
a. barthelemy , a.fert , phys.rev .
b , * 43 * , 13124 ( 1991 ) .
b.l . altshuler and a.g .
aronov in electron - electron interactions in disordered systems , edited by a.l .
efros and m. pollak ( north - holland , amsterdam,1985 ) .
p.a . lee and t.v .
ramakrishnan , rev .
phys . * 57 * , 287 ( 1985 ) .
a. m. rudin , i. l. aleiner , and l. i. glazman , phys .
b * 55 * , 9322 ( 1997 ) kiming hong and n. giordano j.phys . : condens . matter * 10 * , l401 ( 1998 ) g. tatara and h. fukuyama , phys .
lett . * 78 * , 3773 ( 1997 ) yuli lyanda - geller , i.l .
aleiner , and paul m. goldbart , phys .
lett . * 81 * , 3215 ( 1998 ) a.a .
abrikosov , l.p .
gorkov , and i.e. dzyaloshinski , `` methods of quantum field theory in statistical physics '' , dover publications ( 1975 ) . | we consider the effect of electron - electron interactions on the current - in - plane ( cip ) conductivity and exchange coupling energy of a disordered metallic magnetic multilayer .
we analyze its dependence on the value of ferromagnetic splitting of conducting electrons and ferromagnetic layers relative magnetizations orientation .
we show that contribution to the cip conductivity and exchange coupling energy as a periodic function of the angle of magnetizations relative orientation experience @xmath0 transition depending on the characteristic energies : ferromagnetic splitting of the conducting electrons and the thouless energy of paramagnetic layer . |
the primary purpose of the heavy ion programme at the cern sps is the search for a transient deconfined state of strongly interacting matter during the early stage of nucleus nucleus collisions @xcite .
when a sufficiently high initial energy density is reached , the formation of such a state of quasi free quarks and gluons , the quark gluon plasma ( qgp ) , is expected .
a key problem is the identification of experimental signatures of qgp creation @xcite .
numerous proposals were discussed in the past @xcite , but the significance of these signals has come under renewed scrutiny . a possible , promising strategy is a study of the energy dependence of pion and strangeness yields .
it was suggested @xcite that the transition may lead to anomalies in this dependence : a steepening of the increase of the pion yield and a non monotonic behaviour of the strangeness to pion ratio .
the space time evolution of the created fireball @xcite and the event by event fluctuations @xcite may also be sensitive to crossing the transition region .
first experimental results from pb+pb ( au+au ) collisions at top sps ( 158 @xmath0gev ) and ags ( 11 @xmath0gev ) energies have suggested @xcite that anomalies in pion and strangeness production may be located between these energies .
the study of this hypothesis is the motivation for a dedicated energy scan at the cern sps @xcite . within this ongoing project na49
has recorded central pb+pb collisions at 40 and 80 @xmath0gev during the heavy ion runs in 1999 and 2000 , respectively .
the data at the top sps energy ( 158 @xmath0gev ) were taken in previous sps runs . in this paper
we report final results on the energy dependence of charged pion and kaon production .
a preliminary analysis was presented in refs .
pseudorapidity spectra of charged particles produced in pb+pb collisions at 40 and 158 @xmath0gev were recently published by the na50 experiment @xcite .
the energy scan programme at the cern sps will be completed in 2002 by taking data at 20 and 30 @xmath0gev .
the na49 experimental set up
@xcite is shown in fig .
[ setup ] .
it consists of four large volume time projection chambers ( tpcs ) .
two of these , the vertex tpcs ( vtpc-1 and vtpc-2 ) , are placed in the magnetic field of two super - conducting dipole magnets .
this allows separation of positively and negatively charged tracks and a precise measurement of the particle momenta @xmath27 with a resolution of @xmath28 ( gev / c)@xmath29 .
the other two tpcs ( mtpc - l and mtpc - r ) , positioned downstream of the magnets were optimised for high precision measurement of the ionization energy loss @xmath30 ( relative resolution of about 4% ) which provides a means of determining the particle mass .
the particle identification capability of the mtpcs is augmented by two time of flight ( tof ) detector arrays with a resolution @xmath31 60 ps .
the acceptance of the na49 detector is illustrated in fig .
[ acceptance ] where we show the distribution of reconstructed unidentified particles versus the total momentum @xmath27 and transverse momentum @xmath32 for the 80 @xmath0gev data .
the light ( dark ) shaded region shows the coverage of the mtpcs ( tof ) detectors .
the resolution of the @xmath30 measurement allows hadron identification in the mtpcs for @xmath33 gev . at each incident energy the tof acceptance for kaons
was kept at midrapidity by lowering the nominal magnetic field ( @xmath34(vtx1 , 2 ) @xmath35 t at 158 @xmath0gev ) in proportion to the beam energy .
data were taken for both field polarities .
the target , a thin lead foil ( 224 mg/@xmath36 @xmath371% of the interaction length ) , was positioned about 80 cm upstream from vtpc-1 .
central collisions were selected by a trigger using information from a downstream calorimeter ( vcal ) , which measured the energy of the projectile spectator nucleons .
the geometrical acceptance of the vcal calorimeter was adjusted for each energy in order to cover the projectile spectator region by a proper setting of a collimator ( coll ) @xcite .
raw @xmath3 and @xmath38 yields were extracted from fits of the distributions of @xmath30 and @xmath39 ( where available ) in narrow bins of momentum and transverse momentum .
the spectra at midrapidity are obtained using the combined @xmath30 and @xmath39 information ( @xmath40 analysis ) .
the resulting distributions were corrected for geometrical acceptance , losses due to in flight decays and reconstruction efficiency .
the first two corrections are calculated using the detector simulation package geant @xcite .
the procedure used for the efficiency calculation is discussed below , in the context of the pion analysis . for pions at midrapidity
the acceptance of the @xmath30 and @xmath40 methods is limited to the high @xmath32 region ( see fig . [ acceptance ] ) . to obtain the @xmath41 spectra yields of all negatively charged particles
were determined as a function of rapidity ( calculated assuming the @xmath42mass ) and @xmath32 .
the contamination of @xmath38 , @xmath43 and @xmath44 from the interaction vertex as well as non vertex hadrons originating from strange particle decays and secondary interactions was subtracted using two different methods . in the first method
each track measured in the tpcs was extrapolated back to the target plane .
the distance between the track and the interaction vertex was calculated in this plane ( track impact parameter ) .
the impact parameter distributions were used to establish cuts for selection of tracks from the interaction vertex and to estimate the contribution of remaining non
vertex tracks . a correction for @xmath38 contamination
was calculated using parametrised @xmath38 spectra . in the second method
all tracks fitted to the interaction vertex were accepted .
the necessary corrections were calculated based on a venus @xcite simulation of central pb+pb collisions .
the venus events were processed by geant @xcite and the na49 software which simulates the tpc response and produces files in raw data format .
these events were reconstructed and the reconstructed tracks were matched to the venus input .
the obtained correction was scaled by a factor which matches the simulated venus and the measured hadron yields .
the corrections determined using both methods are compatible .
the venus based correction , amounting to 2025% , is used for the data presented in this paper .
the @xmath4 spectra were not analysed , because the positive particles have a large and uncertain contribution of protons .
the resulting @xmath45 and @xmath41 spectra were corrected for geometrical acceptance , and losses due to inefficiencies of the tracking algorithms and quality cuts .
these losses were determined using the ` embedding ' method .
events containing a few tracks were generated and processed by the simulation software .
the resulting raw data were embedded into real events .
the combined raw data were reconstructed and the input tracks were matched with the reconstructed ones .
the calculated losses are about 5% . in order to reduce the systematic errors ,
the analysis has been restricted to regions of phase space where the background and efficiency corrections are small and approximately uniform .
the systematic errors were estimated to be below 10% .
this estimate is based on the comparison of results obtained using different detectors ( tof , tpcs ) , and varying cuts and correction strategies ( see above ) .
additionally , data taken at the two magnetic field polarities were analysed and the results were found to agree within 5% .
note that the same experimental procedure was used to obtain results at all three energies .
thus to a large extent the systematic uncertainties are common for the na49 measurements .
the average number of wounded nucleons @xcite @xmath46 ( the notation @xmath47 will be used to denote the mean multiplicity in full phase space throughout the paper ) as given in table i was not directly measured but was calculated using the fritiof model @xcite . in an unbiased sample of generated inelastic interactions a subsample of central events was selected by applying a cut in the number of projectile spectators .
this cut was adjusted such that the selected fraction was equal to the fraction of all inelastic interactions accepted by the central trigger in the experiment .
the value of @xmath46 was then calculated for the fritiof pb+pb collisions selected in this way .
finally it was verified that the @xmath46 value for central pb+pb collisions at 158 @xmath0gev agrees ( within several per cent ) with the total number of net baryons determined from the participant domain ( in the rapidity interval @xmath48 ) for these collisions @xcite .
note that in this paper @xmath49 denotes the rapidity of a particle in the collision center of mass system .
table i summarises the parameters characterising the data samples used in this analysis .
spectra of transverse mass @xmath50 ( @xmath51 is the rest mass of the particle ) for @xmath3 , @xmath38 ( @xmath40 analysis ) and @xmath41 mesons produced near midrapidity ( @xmath52 for kaons and @xmath53 for pions ) in central pb+pb collisions at 40 , 80 and 158 @xmath0gev are shown in fig .
[ dndmt ] .
the full lines indicate a fit of the function @xmath54 to the data in the range 0.2 gev @xmath55 0.7 gev .
the values obtained for the inverse slope parameter @xmath56 are presented in table ii . the @xmath56 parameter is smaller for pions than for kaons .
a weak increase of @xmath56 with increasing energy is suggested by the pion data , whereas no significant change is seen for kaons .
the rapidity distributions @xmath57 plotted in fig . [ dndy ] were obtained by summing the measured @xmath58 spectra and using the fitted exponential function ( eq . [ mt ] ) to extrapolate to full @xmath58 . for most bins
the necessary correction is small ( @xmath59 ) .
the values of @xmath57 at midrapidity ( @xmath60 ) are given in table ii .
an increase of rapidity density with energy is seen for all particles .
the rapidity spectra were parameterised by the sum of two gauss distributions placed symmetrically with respect to midrapidity : @xmath61.\ ] ] the results of the fits are indicated by the full lines in fig .
[ dndy ] and the obtained values of the parameters @xmath62 and @xmath63 are given in table iii . since both @xmath64 and @xmath63 increase with increasing energy , the width of the observed rapidity distributions clearly increases with energy .
the mean multiplicities in full phase space were calculated by integrating the parametrised rapidity spectra .
the resulting numbers are given in table ii .
the mean multiplicity of @xmath4 mesons given in table ii was calculated by scaling @xmath65 by the @xmath66 ratio ( 0.91 , 0.94 and 0.97 at 40 , 80 and 158 @xmath0gev , respectively ) measured in the region where both @xmath30 and @xmath39 measurements are available .
similar ratios ( within 2% ) are predicted by the venus model @xcite .
we have also checked , with the venus model , that the @xmath66 ratio of total multiplicities is , within 1.5% , equal to the ratio in the @xmath40 acceptance .
results on @xmath57 and inverse slope parameters near midrapidity at 158 @xmath0gev are compatible with previously published measurements @xcite .
the energy dependence of the mean pion multiplicity @xmath67 is shown in fig .
[ energy ] . in this figure
the ratio @xmath68 is plotted as a function of the collision energy , expressed by fermi s measure @xcite : @xmath69 , where @xmath70 is the c.m.s .
energy per nucleon
nucleon pair and @xmath71 the rest mass of the nucleon .
measurements by na49 are compared to results from other experiments on central nucleus
nucleus collisions @xcite and to a compilation of data from @xmath1(@xmath43 ) interactions ( see references in @xcite ) .
one observes that the mean pion multiplicity in @xmath1(@xmath43 ) interactions is approximately proportional to @xmath72 ; the dashed line in fig .
[ energy ] indicates a fit of the form @xmath73 to the data , yielding @xmath74 gev@xmath75 . for central a+a collisions the dependence is more complicated and
it can not be fitted by a single linear function ( @xmath76 ) . below 40 @xmath0gev
the ratio @xmath77 in a+a collisions is lower than in @xmath1 interactions ( pion suppression ) , while at higher energies @xmath77 is larger in a+a collisions than in @xmath1(@xmath43 ) interactions ( pion enhancement ) . in the region between ags and the lowest sps energy ( 1540 @xmath0gev )
the slope changes from @xmath78 gev@xmath75 ( @xmath79 ) for the fit to the points up to top ags energy to @xmath80 gev@xmath75 ( @xmath81 ) for the fit to the top sps energy and the rhic data points @xcite .
the fit to the top sps and rhic points is indicated by the full line in fig .
[ energy ] . the transition from pion suppression to pion enhancement
is demonstrated more clearly in the insert of fig .
[ energy ] , where the difference between @xmath77 for a+a collisions and the straight line parametrisation of the @xmath1 data is plotted as a function of @xmath72 up to the highest sps energy .
midrapidity and full phase space kaon to pion ratios are shown as a function of @xmath70 in figs .
[ energy_mid ] and [ energy_4pi ] @xcite , respectively .
a monotonic increase with @xmath70 of the @xmath82 ratio is measured . for the @xmath83 ratio , a very different behaviour
is observed : a steep increase in the low ( ags @xcite ) energy region is followed by a maximum around 40 @xmath0gev .
the measurement at rhic indicates that the @xmath83 ratio stays nearly constant starting from the top sps energy . for comparison the results on the @xmath84 ratio in @xmath1 interactions @xcite are also shown in fig .
[ energy_4pi ] .
the @xmath1 data have rather large experimental uncertainties @xcite , but suggest a monotonic increase of the ratio .
it should be noted that the @xmath85 ratio is expected to be similar ( within about 10% ) for @xmath1 , @xmath86 and @xmath87 interactions @xcite .
the energy dependence of the @xmath88 ratio at midrapidity is shown in fig .
[ energy_ratio ]
. the ratio increases with @xmath70 from about 0.15 at low ags energy @xcite to about 0.5 at sps energies and reaches about 0.9 at rhic @xcite .
the difference between the dependence of the @xmath3 and @xmath38 yields on @xmath70 can be attributed to their different sensitivity to the baryon density .
kaons ( @xmath3 and @xmath89 ) carry a dominant fraction of all produced @xmath90quarks ( more than 95% in pb+pb collisions at 158 @xmath0gev if open strangeness is considered ) .
therefore the @xmath3 yield ( @xmath91 in approximately isospin symmetric collisions of heavy nuclei ) is nearly proportional to the total strangeness production and only weakly sensitive to the baryon density . as a significant fraction of @xmath92quarks (
e.g. about 50% in central pb+pb collisions at 158 @xmath0gev ) is carried by hyperons , the number of produced antikaons ( @xmath38 and @xmath93 ) is sensitive to both the strangeness yield and the baryon density . in fig .
[ es ] an alternative measure of the strangeness to pion ratio , @xmath94 , is plotted as a function of @xmath72 for a+a collisions @xcite and @xmath1 interactions
@xcite . for a+a collisions
the @xmath95 multiplicity , when not published , ( e.g. for the na49 points ) was estimated as @xmath96)/0.8 , based on strangeness conservation and approximate isospin symmetry of the colliding nuclei @xcite .
the wealth of data on @xmath95 and @xmath97 ( @xmath98 ) production in @xmath1 interactions @xcite allows a much more precise determination of @xmath99 ( open circles in fig .
[ es ] ) than of the @xmath100 ratio ( open circles in fig .
[ energy_4pi ] ) . by construction
, @xmath99 should be almost independent ( an expected variation of several per cent ) of the charge composition of colliding nuclei .
one may conclude from figs .
[ energy_4pi ] and [ es ] that a non monotonic energy dependence ( or a sharp turnover ) of the total strangeness to pion ratio appears to be a special property of heavy ion collisions , which is not observed in elementary interactions .
the energy dependence of pion and strangeness yields was discussed within various approaches to nucleus nucleus collisions . in this section
we compare our results with published model predictions .
it was suggested @xcite that a transition to a deconfined state of matter may cause anomalies in the energy dependence of pion and strangeness production .
this led to the formulation of the statistical model of the early stage ( smes ) @xcite which is based on the assumption that the system created at the early stage ( be it confined matter or a qgp ) is in equilibrium and a transition from a reaction with purely confined matter to a reaction with a qgp at the early stage occurs when the transition temperature @xmath101 is reached .
for @xmath101 values of 170200 mev the transition region ranges between 1560 @xmath0gev @xcite . due to
the assumed generalised fermi
landau initial conditions @xcite the @xmath77 ratio ( a measure of entropy per baryon ) increases approximately linear with @xmath72 outside the transition region .
the slope parameter is proportional to @xmath102 @xcite , where @xmath103 is an effective number of internal degrees of freedom at the early stage . in the transition region
a steepening of the pion energy dependence is expected , because of activation of a large number of partonic degrees of freedom .
this is , in fact , observed in the data on central pb+pb ( au+au ) collisions , where the steepening starts in the range 1540 @xmath0gev ( see fig . [ energy ] ) .
the linear dependence on @xmath72 is obeyed by the data at lower and higher energies ( including rhic ) .
an increase of the slope by a factor of about 1.3 is measured ( see sect .
v ) , which corresponds to an increase of the effective number of internal degrees of freedom by a factor of 1.3@xmath104 @xmath37 3 , within the smes @xcite . in
the smes model the @xmath105 and @xmath99 ratios are roughly proportional to the total strangeness to entropy ratio which is assumed to be preserved from the early stage till freeze out . at low collision energies
the strangeness to entropy ratio steeply increases with collision energy , due to the low temperature at the early stage ( @xmath106 ) and the high mass of the carriers of strangeness ( @xmath107 mev , the kaon mass ) in the confined state . when the transition to a qgp is crossed ( @xmath108 ) , the mass of the strangeness carriers is significantly reduced ( @xmath109 mev , the strange quark mass ) . due to the low mass ( @xmath110 ) the strangeness yield becomes ( approximately ) proportional to the entropy , and the strangeness
to entropy ( or pion ) ratio is independent of energy .
this leads to a `` jump '' in the energy dependence from the larger value for confined matter at @xmath101 to the qgp value .
thus , within the smes , the measured non
monotonic energy dependence of the strangeness to entropy ratio is followed by a saturation at the qgp value ( see figs .
[ energy_4pi ] and [ es ] ) which is a direct consequence of the onset of deconfinement taking place at about 40 @xmath0gev .
numerous models have been developed to explain hadron production in reactions of heavy nuclei without explicitly invoking a transient qgp phase .
the simplest one is the statistical hadron gas model @xcite where independent of the collision energy the hadrochemical freeze out creates a hadron gas in equilibrium @xcite . the temperature , baryon chemical potential and hadronization volume are free parameters of the model , which are fitted to the data at each energy . in this formulation
, the hadron gas model does not predict the energy dependence of hadron production . recently
, an extension of the model was proposed , in which it is assumed that the values of the thermal parameters ( temperature and baryon chemical potential ) evolve smoothly with the collision energy @xcite .
the energy dependence calculated within this extended hadron gas model for the @xmath85 ratio is compared to the experimental results in fig .
[ rqmd ] . due to its construction , the prevailing trend in the data
is reproduced by the model , but the decrease of the ratio between 40 and 158 @xmath0gev is not well described .
the measured strangeness to pion yield in central pb+pb collisions at 158 @xmath0gev is about 25% lower than the expectation for the fully equilibrated hadron gas @xcite .
several dynamical hadron
string models have been developed to study hadron production in a+a collisions .
these models treat the elementary collisions with a string hadronic framework as a starting point .
the models are then extended with effects which are expected to be relevant in a+a collisions ( such as string string interactions and hadronic rescattering ) .
the predictions of the rqmd @xcite and the urqmd @xcite models are shown in fig .
it is seen that rqmd , like the hadron gas model , fails to describe the decrease of the @xmath100 ratio in the sps energy range .
the urqmd model predicts a ratio which , above @xmath111 gev , does not show any sizable energy dependence and which is significantly lower ( e.g. by about 40% at 40 @xmath0gev ) than the data .
this is mainly due to the fact that urqmd overestimates pion production at sps energies by more than 30% @xcite .
the rqmd prediction of the energy dependence of the @xmath83 ratio at midrapidity is shown in fig [ hsd ] .
the model also fails to reproduce the experimental data both in shape and magnitude .
in addition , fig .
[ hsd ] presents the prediction of the hsd model @xcite ( another version of the dynamical hadron string approach ) which shows a monotonic increase of the @xmath83 ratio with energy .
this trend is very different from the measured one .
results on charged pion and kaon production in central pb+pb collisions at 40 , 80 and 158 @xmath0gev are presented .
these are compared with data at lower and higher energies as well as with results from @xmath1 interactions .
the mean pion multiplicity per wounded nucleon increases approximately linearly with @xmath2 with a change of slope starting in the region 1540 @xmath0gev .
the change from pion suppression with respect to @xmath1 interactions , as observed at low collision energies , to pion enhancement at high energies occurs at about 40 @xmath0gev .
a non monotonic energy dependence of the ratio of @xmath3 to @xmath4 yields is observed , with a maximum close to 40 @xmath0gev and an indication of a nearly constant value at higher energies .
this characteristic energy dependence is not observed in elementary interactions and seems to be an unique feature of heavy ion collisions .
the measured dependences can be related to an increase of the entropy production and a decrease of the strangeness to entropy ratio in central pb+pb collisions in the low sps energy range .
they can be understood within the statistical model of the early stage of nucleus nucleus collisions which assumes that a transient state of deconfined matter is created in pb+pb collisions for energies larger than about 40 @xmath0gev .
currently available models without this assumption do not reproduce the measured energy dependence of pion and strangeness production equally well .
this work was supported by the director , office of energy research , division of nuclear physics of the office of high energy and nuclear physics of the us department of energy ( de - aco3 - 76sfooo98 and de - fg02 - 91er40609 ) , the us national science foundation , the bundesministerium fur bildung und forschung , germany , the alexander von humboldt foundation , the uk engineering and physical sciences research council , the polish state committee for scientific research ( 5 p03b 13820 and 2 p03b 02418 ) , the hungarian scientific research foundation ( t14920 and t23790 ) , hungarian national science foundation , otka , ( f034707 ) , the ec marie curie foundation , and the polish - german foundation . for review
see proceedings of the 14th international conference on ultra
relativistic nucleus nucleus collisions , quark matter 99 , eds . : l. riccati , m. masera and e. vercellin , nucl .
a661 * , 1c ( 1999 ) .
l. ahle et al .
( e802 collab . ) , phys .
* c57 * , 466 ( 1998 ) , + l. ahle et al .
( e802 collab . ) , phys .
rev . * c58 * , 3523 ( 1998 ) , + l. ahle et al .
( e802 collab ) , phys .
rev . * c60 * , 044904 ( 1999 ) , +
l. ahle et al .
( e866 collab . and e917 collab . ) , phys .
* b476 * , 1 ( 2000 ) , + l. ahle et al .
( e866 collab . and
e917 collab . ) , phys .
* b490 * , 53 ( 2000 ) , + j. barrette et al .
( e877 collab . ) , phys .
rev . * c62 * , 024901 ( 2000 ) , + d. pelte et al .
( fopi collab . ) , z. phys .
* a357 * , 215 ( 1997 ) .
b. b. back et al . ( phobos collab . ) , phys . rev . lett . * 85 * , 3100 ( 2000 ) and phys .
lett . * 87 * , 102303 ( 2001 ) and phys .
lett . * 88 * , 022302 ( 2002 ) , + _ the mean charged particle multiplicity in central au+au collisions at @xmath70 = 200 gev was obtained by integrating the published pseudo - rapidity distribution . at @xmath70 = 56 gev the measurement at midrapidity was extrapolated to @xmath112 assuming scaling of the shape of the distribution with the beam rapidity .
the charged hadron multiplicities were corrected for contributions of charged kaons and protons . _ r. hagedorn , cern report cern - th-7190 - 94 and proceedings of nato advanced study workshop on hot hadronic matter : theory and experiment , divonne - les - bains , switzerland , 27 jun - 1 jul 1994 , edited by j. letessier , h. gutbrod and j. rafelski , [ hot hadronic matter , v. 346 , 13 ( 1994 ) ] , + j. cleymans and h. satz , z. phys . * c57 * ( 1993 ) 135 , + j. sollfrank , m. gadzicki , u. heinz and j. rafelski , z. phys .
* c61 * ( 1994 ) 659 , + p. braun
munzinger , j. stachel , j. wessels and n. xu , phys .
b365 * ( 1996 ) 1 , + g. d. yen , m. i. gorenstein , w. greiner , s.n .
yang , phys . rev . * c56 * ( 1997 ) 2210 , + g. d. yen and m. i. gorenstein , phys
* c59 * ( 1999 ) 2788 .
numbers of analysed events , cross sections of selected central interactions as percentage of total inelastic cross section ( @xmath113 b ) and mean numbers of wounded nucleons for the selected central pb+pb collisions at 40 , 80 and 158 @xmath0gev .
the first error given in the last column is statistical , the second systematic .
@xmath114 [ % ] & @xmath46 & & & 40 & 4@xmath115 10@xmath116 & 7.2 & 349 @xmath117 1 @xmath117 5 80 & 3@xmath115 10@xmath116 & 7.2 & 349 @xmath117 1 @xmath117 5 158 & 4@xmath115 10@xmath116 & 5.0 & 362 @xmath117 1 @xmath117 5 inverse slope parameters @xmath56 of the transverse mass spectra fitted in the interval 0.2 gev @xmath55 0.7 gev at midrapidity ( @xmath52 for kaons in the @xmath40 analysis , and @xmath118 for pions ) , rapidity densities @xmath57 averaged over the interval @xmath119 as well as total mean multiplicities of @xmath41 , @xmath4 , @xmath38 and @xmath3 mesons produced in central pb+pb collisions at 40 , 80 and 158 @xmath0@xmath115gev .
the first error is statistical , the second systematic .
note that @xmath120 is not directly measured .
@xmath0gev & 158 @xmath0gev & & & @xmath121(mev ) & 169 @xmath117 2 @xmath117 10 & 179 @xmath117 3 @xmath117 10 & 180 @xmath117 3 @xmath117 10 @xmath122(mev ) & 232 @xmath117 3 @xmath117 6 & 230 @xmath117 5 @xmath117 6 & 232 @xmath117 4 @xmath117 6 @xmath123(mev ) & 226 @xmath117 3 @xmath117 6 & 217 @xmath117 3 @xmath117 6 & 226 @xmath117 9 @xmath117 6 @xmath124 & 106.1 @xmath117 0.4 @xmath117 6 & 140.4 @xmath1170.5 @xmath117 7 & 175.4 @xmath117 0.7 @xmath117 9 @xmath125 & 96.6 @xmath117 0.4 @xmath117 6 & 132.0 @xmath1170.5 @xmath117 7 & 170.1 @xmath117 0.7 @xmath117 9 @xmath126 & 20.1 @xmath117 0.3 @xmath117 1.0 & 24.6 @xmath117 0.2 @xmath117 1.2 & 29.6 @xmath117 0.3 @xmath117 1.5 @xmath127 & 7.58 @xmath117 0.12 @xmath117 0.4 & 11.7 @xmath117 0.10 @xmath117 0.6 & 16.8 @xmath117 0.2 @xmath117 0.8 @xmath65 & 322 @xmath117 3 @xmath117 16 & 474 @xmath117 5 @xmath117 23 & 639 @xmath117 17 @xmath117 31 @xmath120 & 293 @xmath117 3 @xmath117 15 & 446 @xmath117 5 @xmath117 22 & 619 @xmath117 17 @xmath117 31 @xmath128 & 59.1 @xmath117 1.9 @xmath117 3 & 76.9 @xmath117 2 @xmath117 4 & 103.0 @xmath117 5 @xmath117 5 @xmath129 & 19.2 @xmath117 0.5 @xmath117 1.0 & 32.4 @xmath117 0.6 @xmath117 1.6 & 51.9 @xmath117 1.9 @xmath117 3 fitted parameters of the two gauss parametrization ( see text , eq .
2 ) of rapidity distributions measured for @xmath41 , @xmath38 and @xmath3 mesons produced in central pb+pb collisions at 40 , 80 and 158 @xmath0@xmath115gev . only statistical errors are given .
@xmath0gev & 158 @xmath0gev & & & @xmath130 & 74.0 @xmath117 0.5 & 97.0 @xmath117 0.7 & 107.6 @xmath117 1.8 @xmath131 & 16.2 @xmath117 0.4 & 19.3 @xmath117 0.3 & 23.4 @xmath117 0.6 @xmath132 & 6.03 @xmath117 0.13 & 9.16 @xmath117 0.12 & 12.8 @xmath117 0.3 @xmath133 & 0.872 @xmath117 0.005 & 0.974 @xmath117 0.007 & 1.18 @xmath117 0.02 @xmath134 & 0.725 @xmath117 0.016 & 0.792 @xmath117 0.018 & 0.88 @xmath117 0.04 @xmath135 & 0.635 @xmath117 0.011 & 0.705 @xmath117 0.010 & 0.81 @xmath117 0.02 @xmath136 & 0.666 @xmath117 0.006 & 0.756 @xmath117 0.006 & 0.72 @xmath117 0.02 @xmath137 & 0.694 @xmath117 0.008 & 0.742 @xmath117 0.008 & 0.839 @xmath117
0.012 @xmath138 & 0.569 @xmath117 0.010 & 0.668 @xmath117 0.005 & 0.727 @xmath117 0.010 | measurements of charged pion and kaon production in central pb+pb collisions at 40 , 80 and 158 @xmath0gev are presented .
these are compared with data at lower and higher energies as well as with results from @xmath1 interactions .
the mean pion multiplicity per wounded nucleon increases approximately linearly with @xmath2 with a change of slope starting in the region 1540 @xmath0gev .
the change from pion suppression with respect to @xmath1 interactions , as observed at low collision energies , to pion enhancement at high energies occurs at about 40 @xmath0gev .
a non monotonic energy dependence of the ratio of @xmath3 to @xmath4 yields is observed , with a maximum close to 40 @xmath0gev and an indication of a nearly constant value at higher energies .
the measured dependences may be related to an increase of the entropy production and a decrease of the strangeness to entropy ratio in central pb+pb collisions in the low sps energy range , which is consistent with the hypothesis that a transient state of deconfined matter is created above these energies .
other interpretations of the data are also discussed . = -1.5 cm * energy dependence of * * pion and kaon production * * in central pb+pb collisions * * the na49 collaboration * * the na49 collaboration * s.v .
afanasiev@xmath5,t .
anticic@xmath6 , d. barna@xmath7 , j. bartke@xmath8 , r.a .
barton@xmath9 , m. behler@xmath10 , l. betev@xmath11 , h. biakowska@xmath12 , a. billmeier@xmath11 , c. blume@xmath13 , c.o .
blyth@xmath9 , b. boimska@xmath12 , m. botje@xmath14 , j. bracinik@xmath15 , r. bramm@xmath11 , r. brun@xmath16 , p. buni@xmath17 , v. cerny@xmath15 , j.g .
cramer@xmath18 , p. csat@xmath7 , p. dinkelaker@xmath11 , v. eckardt@xmath19 , p. filip@xmath19 , z. fodor@xmath7 , p. foka@xmath13 , p. freund@xmath19 , v. friese@xmath10 , j. gl@xmath7 , m. gadzicki@xmath11 , g. georgopoulos@xmath20 , e. gadysz@xmath8 , s. hegyi@xmath7 , c. hhne@xmath10 , g. igo@xmath21 , p.g .
jones@xmath9 , k. kadija@xmath22 , a. karev@xmath19 , v.i .
kolesnikov@xmath5 , t. kollegger@xmath11 , m. kowalski@xmath8 , i. kraus@xmath13 , m. kreps@xmath15 , m. van leeuwen@xmath14 , p. lvai@xmath7 , a.i .
malakhov@xmath5 , s. margetis@xmath23 , c. markert@xmath13 , b.w .
mayes@xmath24 , g.l .
melkumov@xmath5 , a. mischke@xmath13 , j. molnr@xmath7 , j.m .
nelson@xmath9 , g. plla@xmath7 , a.d .
panagiotou@xmath20 , k. perl@xmath25 , a. petridis@xmath20 , m. pikna@xmath15 , l. pinsky@xmath24 , f. phlhofer@xmath10 , j.g .
reid@xmath18 , r. renfordt@xmath11 , w. retyk@xmath25 , c. roland@xmath26 , g. roland@xmath26 , a. rybicki@xmath8 , t. sammer@xmath19 , a. sandoval@xmath13 , h. sann@xmath13 , n. schmitz@xmath19 , p. seyboth@xmath19 , f. siklr@xmath7 , b. sitar@xmath15 , e. skrzypczak@xmath25 , g.t.a .
squier@xmath9 , r. stock@xmath11 , h. strbele@xmath11 , t. susa@xmath6 , i. szentptery@xmath7 , j. sziklai@xmath7 , t.a .
trainor@xmath18 , d. varga@xmath7 , m. vassiliou@xmath20 , g.i .
veres@xmath7 , g. vesztergombi@xmath7 , d. vrani@xmath13 , a. wetzler@xmath11 , c. whitten@xmath21 , i.k .
yoo@xmath10 , j. zaranek@xmath11 , j. zimnyi@xmath7 @xmath14nikhef , amsterdam , netherlands .
+ @xmath20department of physics , university of athens , athens , greece .
+ @xmath9birmingham university , birmingham , england .
+ @xmath15comenius university , bratislava , slovakia .
+ @xmath7kfki research institute for particle and nuclear physics , budapest , hungary .
+ @xmath26mit , cambridge , usa .
+ @xmath8institute of nuclear physics , cracow , poland .
+ @xmath13gesellschaft fr schwerionenforschung ( gsi ) , darmstadt , germany .
+ @xmath5joint institute for nuclear research , dubna , russia .
+ @xmath11fachbereich physik der universitt , frankfurt , germany .
+ @xmath16cern , geneva , switzerland .
+ @xmath24university of houston , houston , tx , usa . + @xmath23kent state university , kent , oh , usa . + @xmath21university of california at los angeles , los angeles , usa .
+ @xmath10fachbereich physik der universitt , marburg , germany .
+ @xmath19max - planck - institut fr physik , munich , germany .
+ @xmath12institute for nuclear studies , warsaw , poland .
+ @xmath25institute for experimental physics , university of warsaw , warsaw , poland .
+ @xmath18nuclear physics laboratory , university of washington , seattle , wa , usa . + @xmath6rudjer boskovic institute , zagreb , croatia . |
self - propelled particles ( spps ) are attracting more and more interest as a representative example of out - of - equilibrium systems@xcite .
examples of spps range from the microscopic scale , with algae and bacteria , to the macroscopic scale , which includes all animals , and recently , even artificial spps have been constructed , such as active janus particles@xcite . among them , microswimmers are of particular interest , since they are suited to well - controlled lab experiments@xcite , and have many potential applications , for example , as targeted drug delivery systems .
it is known that nontrivial motion , like the collective alignment of the swimming direction or the dynamic clustering can be observed , even for systems where particles interact with each other only through hydrodynamic interactions and excluded volume effects@xcite .
such collective motion is mainly due to the complicated hydrodynamic interactions , and can not be predicted only from knowledge of the single particle dynamics . that these hydrodynamic interactions are strongly affected by the presence of confining walls is well known .
experimentally , das _
_ reported that when a single janus particle swims in the vicinity of a wall , it tends to swim along the wall@xcite .
_ investigated the collective dynamics of a bacterial dispersion inside a circular confinement , and reported that the system spontaneously shows an anomalous double - vortex motion , the outermost agents swimming along the wall@xcite .
while most numerical studies have focused on the dynamics in bulk , several works have reported on the dynamics of microswimmers near walls or under confinement@xcite . in the case of single - particle dynamics under confinement ,
most works consider flat parallel walls or cylindrical pipes@xcite . as to the dynamics of many particle systems , they are usually studied under confinement of flat parallel walls . in particular , li and ardekani
have investigated the static structure and shown evidence for the accumulation of particles near the wall@xcite .
zttl and stark have studied a similar system@xcite , but under extreme confinement , and observed a dense - dilute phase separation , which is not seen in bulk .
our previous work@xcite has focused on the dynamic properties of a microswimmer dispersion confined between flat parallel walls with relatively large separations ( compared to the size of the particles ) . in ref .
@xcite , we observed a traveling wave - like collective motion for a specific range of swimming parameters and densities .
though such a motion would seem to be a consequence of confinement , we clarified that it can be understood as the manifestation of the pseudo - acoustic properties of the system , which is already observed in bulk .
as shown here , even for the same confinement geometry , the dynamic behavior can be considerably different depending on the strength of the confinement ( i.e. , the wall separation ) .
therefore , how the dynamics changes under confinement is a very difficult question to answer .
although we can find several works on the collective dynamics in bulk or between flat parallel walls , the many particle dynamics of swimmers in pipes has not been extensively studied so far .
this is the focus of the present work .
intuitively , we can expect that for big enough pipes , the dynamics will be the same as for bulk systems .
therefore , we focus on the dynamics in pipes with diameters comparable to that of the particle , where we can expect nontrivial behaviors .
in fact , in the work by zhu _
_ , which deals with the single particle dynamics in a pipe with diameter three times the particle diameter , it is reported that different dynamical modes can be observed depending on the swimming type and strength@xcite .
taking into account the possibility that the size and the shape of the pipe affect the dynamic properties , in this work , we investigated the collective alignment effects , known as the polar order formation , varying the four main parameters , namely , the pipe size , the pipe shape , the volume fraction and the type of the swimmers . regarding the polar order formation in many particle systems ,
the behavior in pipes is mostly the same as in bulk .
however , for a specified region of the parameter space , we observed clear wall effects , which we investigated by measuring the structural information of the dispersion .
in addition , we observed the pipe size dependent order / disorder phase transition only for the parameters at which clustering behaviors have been reported in bulk and between flat walls@xcite . in this work
, we also measured the bulk structural information and obtained indirect evidences which states that the clustering is important for the collective alignment for a specific range of parameters .
as the numerical model for microswimmers , we employed the squirmer model@xcite . in this model ,
the microswimmers are expressed by rigid particles with a prescribed flow field on their surface .
the general squirmer model is expressed in the form of an infinite expansion , with components in the tangential , radial , and azimuthal directions . however , utilizing only the first two modes of the tangential field , following eq .
( [ eq : sq_2 ] ) , already enables us to model different types of swimmers , namely pushers , pullers and the neutral swimmers .
@xmath0 where , @xmath1 denotes the surface flow field , @xmath2 is a unit vector directed from the center of the particle to a point on its surface , @xmath3 the polar angle between @xmath2 and the swimming direction @xmath4 , and @xmath5 is the tangential unit vector at @xmath2 .
this simplified squirmer model has been widely used and is known to lead to a wide variety of nontrivial phenomena@xcite .
the coefficient of the first mode , @xmath6 , determines the swimming velocity of an isolated squirmer ( @xmath7 ) , and that of the second mode , @xmath8 , determines the stresslet@xcite .
the ratio between the coefficients of the two modes , @xmath9 in eq .
( [ eq : sq_2 ] ) , determines the swimming type and strength .
negative values of @xmath10 represents pusher - type swimmers , which swim with an extensile flow field in the swimming direction , and positive values describe puller - type swimmers , which swim with a contractile flow . a value of @xmath11 stands for a neutral swimmer , which generates a potential flow field .
in this model , the prescribed flow field is assumed to be an axisymmetric pure tangential one and the particle is rigid and spherical .
these are the only assumptions employed to derive the model .
therefore , it is thought that this model can capture the features of the artificial microswimmers like janus particles even though it was originally intended to describe ciliary propulsion of micro - organisms . in this work
, we investigate the dynamics of particles swimming in a viscous fluid . for this
, we must solve the combined fluid - solid problem . in order to do this efficiently , we employed the smoothed profile method ( spm ) . in the spm
, we do nt treat the sharp boundary between the solid and fluid phases explicitly . instead
, a diffuse interface with thickness of @xmath12 is introduced , and the phase boundaries are expressed by using a continuous order parameter @xmath13 , which takes the value of one in the solid domain , zero in the fluid domain , and intermediate values within the diffuse interface region . employing this methods , all the physical quantities can be expressed in terms of continuous fields defined over the whole computational domain .
thus , the calculation cost can be reduced by orders of magnitude .
see refs.@xcite for more detailed information about this method . in this work , the dynamics of swimmers in pipes are considered . for this
, we have to model solid walls numerically . within the spm scheme
, the walls can be expressed by assemblies of particles which are pinned and not allowed to translate or rotate . by this
, the stick boundary at the wall surface is guaranteed .
the wall particles are placed uniformly and have the same diameter as the swimmer particles . because we consider various shapes and sizes of pipes , the number of particles composing the walls varies depending on the pipe .
we note that the surface of the confinement has a non - zero roughness because the walls are represented by a collection of discrete particles . using the calculation method presented above
, we conducted three dimensional direct numerical simulations ( dns ) of microswimmer dispersions confined in pipes .
the diameter @xmath14 and interface thickness @xmath12 of particles are set to be @xmath15 and @xmath16 respectively , where @xmath17 stands for the grid spacing . the excluded volume effect between particles ( both wall - particle and particle - particle )
is implemented by a truncated lenard - jones type @xmath18 potential ( @xmath19 ) . regarding the parameters in eq .
( [ eq : sq_2 ] ) , we set @xmath20 for all the calculations and varied @xmath10 to study the @xmath10 dependency of the dynamics .
this value of @xmath6 gives a particle reynolds number of one for an isolated swimmer .
the symmetry axis of the pipe is defined to be parallel to the @xmath21 axis . in what follows
, we refer to the confinement with a circular cross section as a pipe , and the one with a rectangular cross section as a duct .
regardless of the size or shape of confinement , the length of the pipe @xmath22 is set to be @xmath23 and the periodic boundary condition is set in this direction ( in other directions , the system is enclosed by the confinement , and the system size in those directions depends on the confinements ) . unless stated otherwise ,
thermal fluctuations are ignored throughout this paper .
we will consider the pipe with diameter @xmath24 to be our reference confinement system .
the diameter @xmath25 is measured with respect to the center of the particles composing the walls .
for the definition of the particle volume fraction we used : @xmath26 where @xmath27 the volume of one particle and @xmath28 the volume in which particle centers can move freely . in our systems , @xmath28 for a pipe is calculated following eq .
( [ eq : vm ] ) because of the excluded volume effect between swimmers and wall composing particles , @xmath29 where , @xmath30 . for ducts ,
a similar correction is employed .
typical trajectories of the three different dynamical modes .
( a ) pusher s rotational orbit ( @xmath31 ; represented by blue ) ( b ) neutral swimmer s spiral trajectory ( @xmath11 ; represented by gray ) ( c ) puller s rectilinear path ( @xmath32 ; represented by red ) .
note that these three trajectories are obtained using the same initial position and orientation .
only the trajectories in steady states are shown .
white particles represent the particles composing the pipe . to facilitate the visualization , only half of the pipe - particles are drawn . ]
the single - particle dynamics in a pipe has been studied by zhu _
et al . _ , who reported that distinct dynamical modes are observed depending on the swimming parameter @xmath10@xcite . in order to confirm the qualitative agreement within our calculation scheme , we performed calculations with one particle for various values of @xmath10 . here
, we consider only a pipe with diameter @xmath24 . after a sufficiently long time , all the cases achieve steady states .
typical trajectories of the steady states are shown in fig .
( [ fig : trajectories ] ) . here , the result of the cases with @xmath33 are shown .
the dynamical mode at the steady state depends only on the sign of @xmath10 and is independent of the absolute value : a pusher gets trapped in an orbit and shows constant rotation at a fixed @xmath21 position , while a puller shows rectilinear translation along the pipe axis at a constant distance from the wall .
a neutral swimmer achieves the combined dynamics of the two modes : spiral motion along the pipe axis ( rotation + translation ) .
these results are consistent with the results in ref.@xcite ; although the trajectories of the pushers are not analyzed in detail in this work .
as shown here , the single - particle dynamics are affected by the confinement .
the intriguing thing is that the dynamical mode is determined only by the sign of @xmath10 , independent of the absolute value .
regarding the bulk collective properties , it is known that such an absolute value dependence exists .
for example , the polar order @xmath34 ( eq . ( [ eq : bulk_q ] ) ) , used to measure the collective alignment , strongly depends on the absolute value of @xmath10 .
@xmath35 where , @xmath36 stands for the number of particles in the system .
this quantitative measure of the degree of order , @xmath34 , takes a value of one when the system is completely aligned , and @xmath37 when the system is completely isotropic .
the value of @xmath34 decreases as the absolute value of @xmath10 increases@xcite .
we next performed calculations for many particle systems in a pipe of diameter @xmath24 .
it is already known that the polar order @xmath34 is a function of both the swimming parameter @xmath10 , and the volume fraction of the particles @xmath38 .
first of all , let us recapitulate the behaviors in bulk systems@xcite . if we measure the polar order as a function of @xmath10 at a certain value of the volume fraction , the value of @xmath34 decreases with the increase in the absolute value of @xmath10 , as stated above .
this tendency is asymmetry with respect to positive and negative @xmath10 values , with negative values ( pushers ) , showing a faster decay . if we regard @xmath34 as a function of @xmath38 for a certain value of @xmath10 , the behavior is most easily understood in terms of the deviations from the volume fraction dependence for neutral swimmers .
this reference system ( @xmath11 ) shows two regimes : for @xmath39 , the values are constant ; for @xmath40 there is an abrupt drop in the order parameter to @xmath41 . a similar behavior is observed for pushers ( negative values of @xmath10 ) . at small volume fractions , they show only a weak decrease with increasing volume fraction , and then exhibit an abrupt drop at a critical volume fraction @xmath42 .
pushers show a decrease both in the degree of ordering and the value of @xmath42 .
pullers , on the other hand , show a decrease only in the degree of ordering , and systems with @xmath43 even maintain the order at very high volume fractions , where all other systems show @xmath41 . specifically , we stress that in bulk , the system with @xmath32 shows a nonzero value of @xmath34 up to very high volume fractions
. we conducted simulations over a wide range of combinations of @xmath10 and @xmath13 in the pipe . from the particle trajectories obtained from these simulations
, we then calculated the polar order parameter along the @xmath21 direction ( the direction of the pipe elongation ) , @xmath44 where @xmath45 is the @xmath21-component of the swimming direction @xmath4 of particle @xmath46 and the angular brackets denote a time average . in fig .
[ fig : bulk_pipe ] , the values of @xmath47 are shown as a function of @xmath10 . here , the values of @xmath34 in bulk are also shown .
the volume fraction is @xmath48 for all the cases ( @xmath49 in the pipe , @xmath50 in bulk ) . in fig .
[ fig : bulk_pipe ] , we see a good match between results in bulk and those in pipe over a wide range of @xmath10
. however , for @xmath51 , we see clear differences . in pipes , over this @xmath10 range , the values of @xmath34 are reduced compared to the bulk results , with the gap increasing with an increase in the absolute value of @xmath10 .
finally , at @xmath52 , the dynamics show a qualitatively different behavior : in the pipe , @xmath53 which means the order is completely lost . the polar order parameter @xmath47 in a pipe with diameter @xmath54 as a function of the swimming parameter @xmath10 . the polar order @xmath34 in bulk systems is also shown as a light - colored line .
the volume fraction is @xmath55 in both cases .
the dashed line stands for the value of @xmath56 . ] the polar order parameter @xmath57 in a pipe with diameter @xmath54 as a function of volume fraction and swimming parameter .
the polar order @xmath34 in bulk systems is also shown using dashed lines . ] in fig .
[ fig : order_pipe_big ] , detailed results showing the volume fraction dependency are shown for four different values of @xmath10 , which exhibit qualitatively different behavior , @xmath58 . for the volume fraction ,
we consider six different values in the range of @xmath59 ( corresponding to @xmath60 ) .
as before , the bulk results are shown as dashed lines in fig .
[ fig : order_pipe_big ] .
although the results in the pipe are qualitatively the same as those in bulk for some of the cases , several differences can be perceived . in the case with @xmath11 , the value of @xmath34 at the smallest volume fraction , @xmath61 shows an obvious reduction in the pipe compared to the bulk . at other volume fractions ,
the values match reasonably well .
in the case of @xmath62 , we observe a similar reduction at the lowest volume fraction .
in addition , in this case , we observe a rise in @xmath57 at intermediate volume fraction ( @xmath63 ) .
finally , in the case with @xmath32 , we see no polar order ( @xmath64 ) at any of the volume fractions considered . in order to understand these differences from the view point of the structures inside the pipe , we measured the following two functions , the local density @xmath65 and the local polar order along the @xmath21-axis @xmath66 .
these are defined as a function of the perpendicular distance @xmath67 from the symmetry axis of the pipe : @xmath68 where @xmath69 is the number of particles at a distance @xmath70 from the pipe axis , @xmath71 the average number density of the whole system , @xmath72 the width of the cylindrical bins , and the summation in eq .
( [ eq : qinpipe ] ) is taken over all the particles in a given bin .
the local density of the shell matches the overall average value when @xmath73 .
the results are shown in fig .
[ fig : dis_pipe ] for the same values of @xmath10 shown in fig .
[ fig : order_pipe_big ] . from this figure
, we can explain the differences seen in fig .
[ fig : order_pipe_big ] . in the case of @xmath74 and @xmath75 ,
when the volume fraction is small ( @xmath76 , ( a ) ) , particles show a strong tendency to accumulate in the vicinity of the wall and almost no particles are observed within the central region ( @xmath77 ) .
such a wall accumulation effect has been reported for active systems in general@xcite , and several works on squirmers have also shown this behavior@xcite . at the same time , we can see that the local polar order is reduced in the vicinity of the wall .
in other words , the pipe can be divided into two distinct regions from the view point of @xmath78 : an inner region where the local order has a value similar to that of the corresponding bulk dispersion , and an outer region ( near to the walls ) where the order is reduced . in the following ,
we specify the characteristic size of this outer region by @xmath79 , the distance from the wall to the point where the inner region starts .
because most particles accumulate in this outer region , where the order is reduced , the overall polar order is smaller than that in bulk .
then , as @xmath38 is increased , the particles will become more uniformly distributed , and @xmath80 is reduced . at the same time , due to the wall accumulation effect , the effective volume fraction in the center region is smaller than the global volume fraction . taking into account the fact that the polar order increases with decreasing volume fraction ,
this reduction in the volume fraction leads to a higher local polar order at the center , compared to the bulk value at the same global volume fraction .
as a result of these two effects , namely , the decrease in @xmath80 and the increase of the local order due to the diluteness at the center , at @xmath81 , the value of @xmath47 recovers the bulk result . in the case of @xmath62 , at @xmath82 , the value of @xmath47 is even larger than that in bulk . in the case of @xmath11 , such an overshoot
is not observed .
this is because for @xmath11 , the volume fraction dependency of
the polar order is so small that the order is determined only by the value of @xmath83 . for @xmath32 ,
the value of @xmath83 becomes very large and we do nt see any ordering throughout the pipe , although the system with @xmath32 does show a nonzero value for polar order parameter @xmath34 in bulk . in the limit when the diameter of the pipe goes to infinity , we expect the dynamics should converge to that observed in bulk
. therefore , we should see a qualitative transition at a critical diameter .
this will be investigated in more in detail in sec .
c .. we also note that the behavior of @xmath84 for @xmath32 is very similar to that of @xmath85 at all volume fractions , although the order formation tendencies are completely different . at the highest volume fraction @xmath86 , because of the screening of the hydrodynamic interactions , the excluded volume effects become dominant in determining the structure .
consequently , the structure is the same for all values of @xmath10 .
however , it is interesting to note that only in the case of @xmath85 , do we observe a nonzero value of @xmath47 .
we have clarified that in many particle systems , the wall effects can be characterized by the accumulation of particles over a distance of @xmath80 close to the walls .
these can lead to both enhancement and inhibition of the ordering , depending on the degree of accumulation and the size of @xmath80 .
while for some cases we do observe a rectification effect due to the existence of walls , the origin for such effects seems to have nothing to do with the causes responsible for the single particle trajectories presented above . here , we show results for the pipe diameter dependence on the many - particle dynamics .
first , we present the results for the single - particle systems in a pipe with a `` small '' diameter . here , we use @xmath87 , which corresponds to the system in ref.@xcite . for these single - particle systems
, we have observed the same qualitative results as those for @xmath24 ( which we refer to as the `` big '' pipe ) , except for pushers .
first , let us recapitulate the single particle motion in the big pipe . in this case
, the three different swimmers ( puller , pusher and neutral swimmer ) show qualitatively different trajectories .
a puller swims parallel to the pipe axis , while a pusher is trapped in a closed orbit at a constant @xmath21-value .
finally , a neutral swimmer shows a behavior in between that of a pusher and a puller , namely a spiral motion . in the small pipe
, only a pusher shows a different dynamical mode from that observed in the big pipe . in the small pipe , a pusher follows a spiral trajectory , just like a neutral swimmer .
while a pusher in the big pipe also shows a spiral trajectory at the beginning of the simulations , the pitch will gradually become shorter , until the pusher becomes trapped in a closed orbit . in the small pipe
the pitch of the spiral is constant and does nt show this decay .
nevertheless , we still see a distinct difference between a pusher and a neutral swimmer , if we look at the projection onto the @xmath88-plane ( see fig . [
fig : tra_pro ] ) .
the pusher s trajectory shows a perfect circular shape while the neutral swimmer shows a wave - like pattern .
the trajectory we have obtained for the neutral swimmer is neither regularly patterned nor closed , unlike the results report in ref@xcite .
we believe this discrepancy is due to the roughness of the wall surface .
the projection of the trajectories in a small pipe onto the @xmath88-plane .
( a ) pusher with @xmath31 .
( b ) neutral swimmer .
the lines represent the projections of the trajectories , with color intensity used to represent the time . ] the time evolution of the temporal order @xmath89 for systems with @xmath90 in the small pipe with @xmath91 at @xmath92 .
the horizontal axis gives the time normalized by @xmath93 , where @xmath94 is the steady state velocity of an isolated swimmer .
the vertical axis represents the value of temporal order . ] the dependency on @xmath10 of the polar order @xmath95 in pipes .
the dark lines are for the pipe with @xmath91 , while the light lines are for @xmath24 .
the volume fraction is @xmath55 for both cases .
the shaded region represents the region where the anomalous behavior was reported in ref.@xcite . ]
next , we investigated the polar order of many particle systems in the small pipe .
we present the results for a pipe with @xmath91 , as a basic reference system , but we have confirmed that the same qualitative results are obtained for @xmath87 , for both the single- and many - particle dynamics . in the small pipe , over a specific range of parameters ,
the value of the polar order does nt show a stable nonzero value . in fig .
[ fig : t_e ] , the time evolution of the temporal order @xmath96 , defined in eq .
( [ eq : ave_e ] ) , is given for two cases to illustrate both the steady and unsteady behaviors .
@xmath97 both results are for a system with @xmath91 and @xmath98 , and the steady results are for @xmath99 ( the weak pullers ) , the unsteady one for @xmath85 ( intermediate pullers ) .
as seen here , in the unsteady system , even the sign of @xmath96 , which represents the direction of motion , changes in time ; while in the systems with stable order , such change is never observed .
this change of direction shows the absence of any long - time persistent order in the pipe . in order to evaluate such direction - flipping behavior , we define an alternative in - pipe polar order parameter that can also indicate the persistence of the motion : @xmath100 only the order of taking the average and the absolute value
is changed with respect to the standard definition of the polar order @xmath47 , eq .
( [ eq : pipe_q ] ) . for the cases
in which the direction does nt change , the values @xmath47 and @xmath95 are essentially the same .
we measured @xmath95 in the small pipe for various values of @xmath10 and compared the results with those obtained for the big pipe .
the results are shown in fig .
[ fig : order_pipe_small ] at a volume fraction of @xmath63 .
we observe a remarkable differences over the range @xmath101 . in the small pipe
, @xmath96 has no steady state value and changes directions , as described above .
actually , for such systems , the values of @xmath95 are around @xmath102 , which means that the system has no order .
we consider this collapse of the order as an order / disorder phase transition .
the gray - shaded region in fig .
[ fig : order_pipe_small ] indicates the range of @xmath10 at which anomalous behaviors were reported in preceding studies@xcite . in ref.@xcite , over this range of @xmath10 , particles form big cluster which exhibit traveling wave - like motion , bouncing back and forth between walls .
interestingly , the shaded area corresponds very well to the region over which we observe the pipe - size dependent order / disorder transition in the present work .
we refer to pullers in these regions as `` intermediate '' pullers . for @xmath10 values other than those of intermediate pullers ,
the results are roughly the same between @xmath24 and @xmath91 .
we note that we have confirmed that the same qualitative results are obtained for different values of @xmath38 , at least over the range we have considered ( @xmath103 ) .
we would also like to stress that in bulk , we have observed a similar order / disorder phase transition when we change the volume fraction@xcite : when the volume fraction becomes very high , none of the swimmers except intermediate pullers are able to maintain the polar order . in contrast , in pipes , only intermediate pullers lose their polar ordering when the size of the pipe becomes small . the pipe diameter dependency of the in - pipe polar order @xmath95 .
the horizontal axis represents the pipe diameter @xmath104 normalized by the particle diameter .
the volume fraction is @xmath55 for all cases . ]
we have also investigated the detailed size dependency of the polar order for intermediate pullers . in fig .
[ fig : size_d_05 ] , the values of the polar order @xmath105 are shown as a function of the pipe diameter . here , the results are shown not only for intermediate pullers @xmath85 , but also for weak pullers @xmath99 , which do nt show the phase transition .
the volume fraction is @xmath55 for all cases ( @xmath106 , depending on the size of the pipe ) .
we see an abrupt drop of @xmath95 at the critical diameter @xmath107 , though at higher @xmath108 , the values are constant . in systems with @xmath109 , the value of @xmath95 does not achieve the steady state , similar to the situations seen in fig . [
fig : order_pipe_small ] .
in other words , in such small pipes , intermediate pullers can not maintain the polar order anymore , though temporally they can exhibit order .
we believe this sensitivity to the size suggests that intermediate pullers need clusters bigger than a characteristic size @xmath110 to stabilize the polar order . in order to specify the size of the cluster in bulk , we performed bulk simulations and measured the generalized radial distribution functions @xmath111@xcite , @xmath112 where @xmath113 represents the distance from a reference particle @xmath46 to a second particle @xmath114 , @xmath115 the relative angle between the swimming directions of both particles , and @xmath116 is the @xmath117-th degree legendre polynomial .
the zero - order function @xmath118 is the standard radial distribution function , and the first - order function @xmath119 represents the local degree of order . in the limit when @xmath120 , we expect @xmath121 and @xmath122 .
therefore , we can define the excess local polar order @xmath123 , in such a way that it will converge to zero in the limit of @xmath120 .
the bulk system considered here has a cubic domain with a linear dimension of @xmath124 .
periodic boundary conditions are used in all directions , and the particles have the same diameter and interface thickness as in the simulations discussed above .
the volume fraction is tuned by changing the number of particles , @xmath125 .
the calculated results for @xmath118 , @xmath119 and @xmath126 are shown in fig .
[ fig : gen_dis ] , for @xmath127 and @xmath58 , corresponding to the confined systems in fig .
[ fig : dis_pipe ] .
in addition , results for @xmath118 for passive colloidal systems are shown to aid in comparison .
for these passive systems , the temperature is set so that @xmath128 , where @xmath129 is the energy unit in our system , @xmath130 the fluid mass density and @xmath131 the viscosity .
the exact value of the temperature does not affect the radial distribution function , provided that it is small enough . to begin the analysis of the structure , we discuss the behavior of @xmath132 for the reference system of passive colloids . at the smallest volume fraction ( @xmath76 )
, it shows only a small first peak around @xmath133 , and quickly decays , converging to one around @xmath134 .
the intensity of this first peak becomes larger as the volume fraction increases . at a volume fraction of @xmath135
, we observe a shallow undershoot after the first peak .
the correlation length becomes longer and the value converges to one at around @xmath136 . at @xmath82 ,
it starts to show a second peak at a position @xmath137 , though the intensity is very small .
finally , at @xmath86 , the system is in a mixed state of ordered and disordered phases@xcite , and shows four clearly distinguishable peaks .
now , we would like to present the results for the active systems . in the case of @xmath11 , for all the volume fractions but the highest one ,
the first peak is smaller or shifted to larger distances compared to the passive system .
in addition , the first peak does nt even appear at the lowest volume fraction , @xmath76 .
these behaviors may indicates that there is an effective repulsive interaction between particles caused by the flow field . due to this repelling flow field ,
neutral swimmers tend to get distributed uniformly .
similar effects can be seen in the pipe , as shown in fig .
[ fig : dis_pipe ] where neutrals swimmers show a much more uniform distribution compared to the pushers .
the correlations beyond the first peak follow closely those of the passive system , while other swimmers show longer ranged correlations .
this tendency may reflect the fact that the flow field around a neutral swimmer ( decaying as @xmath138 ) is more localized , because of the lack of the force dipole term ( decaying as @xmath139 ) .
regarding @xmath140 , the qualitative behavior follows that of @xmath132 .
we note that the neutral swimmers exhibit a unique behavior around @xmath136 , where @xmath141 shows negative values for intermediate values of @xmath142 . in the case of @xmath62 and @xmath143 ,
the results are quite similar for all the volume fractions , although the peak heights show slight differences . in these cases ,
the particles seem to show attractive tendencies due to the flow field , resulting in higher peaks for @xmath118 . as before
, @xmath140 follows the qualitative behavior seen in @xmath118 . besides the higher peaks ,
the correlation length is longer than that seen for neutral swimmers .
however , the bulk values are reached by @xmath144 in all cases . in the case with @xmath85 , or intermediate pullers , many unique behaviors can be observed .
first of all , for intermediate pullers @xmath132 shows much more outstanding peaks than in any of the other cases . at @xmath135 ,
only intermediate pullers show a definite second peak , and at @xmath82 , the second peak is much higher than in the other systems , and a third peak even appears .
a similar peak behavior can be observed also in the plot of @xmath145 .
in addition , we see a very long tail , which is not seen in the other cases .
all these unique behaviors for intermediate pullers can be understood as indirect evidence for their already known dynamic clustering tendency@xcite .
the high peaks reflect the clustering behavior , and the long tail in @xmath119 shows the strongly localized ordering .
the nonzero order at high volume fractions helps explain how intermediate pullers can swim collectively , even under such extreme conditions .
we note that the critical diameter of the pipe @xmath110 is bigger than any of the peaks detected here . at the highest volume fraction , @xmath86 , the shapes of @xmath118 are the same for all @xmath10 , while only the passive system shows a different shape .
regarding the neutral swimmer systems , @xmath11 , this is surprising because this system shows quite similar shapes at all the other volume fractions .
the shape difference of @xmath132 between the passive and active systems at @xmath86 may be due to the phase transition of the passive system .
for passive hard - spheres , the volume fraction in the range @xmath146 is known as the transient state of the ordered state and the disordered state@xcite . in the passive system at @xmath86 , shown in fig .
[ fig : gen_dis ] , such an ordered state seems to appear , while the active systems are still completely in the fluid phase . at this volume fraction , because the excluded volume effect is the dominant factor to determine the structure , rather than the flow field , all the active systems show the same shape for @xmath147 . in summary , although we could not obtain a direct connection between the bulk structural information and the in - pipe order / disorder phase transition , we believe our results for the structure present indirect evidence for the importance of the dynamic clustering on the stabilization of the polar order for intermediate pullers systems . in order to verify the exact connection between these two quantities
, we have to conduct a more detailed analysis on the dynamic clustering in bulk .
our previous results @xcite support this view on the importance of dynamic clustering . in this work , we investigated whether or not the polar order seen in bulk can be explained only by the repetition of binary collisions .
we found that this is the case , except for systems of intermediate pullers , which means that many particle interactions might be important in this case . finally , we have also considered pipes with diameters larger than @xmath54 . as the pipe size increases
, we can expect that the dynamics approaches that of seen in bulk systems .
indeed , for systems with @xmath148 , we see in fig . [ fig : order_pipe_big ] that the order parameters in confinement agree very well with the bulk values .
thus , for these values of @xmath10 , @xmath24 seem to be big enough . therefore , we have only considered larger diameters for @xmath32 , to investigate whether a finite value of @xmath34 is recovered for a large enough pipe diameter .
we considered diameters up to @xmath149 but were not able to recover the bulk values in this case .
thus , the hypothesis stated in section iii .
b. is still not solved within the range of parameters considered in the present study .
although it is possible that the dynamics can change for even larger diameters , we will not discuss this problem any more , as it lies outside the scope of the present work .
the dynamics of fluids in pipes is known to depend on the shape of the cross section .
therefore , we have also investigated the effect of the shape dependency on the dynamics of the swimmers .
so far , most previous investigations have concerned dynamics in cylindrical pipes , with circular cross sections . here , we consider pipes with a rectangular cross section .
we refer to such confinement geometries as `` ducts '' .
we considered two sizes of ducts , specified by a lateral size of @xmath150 and @xmath54 , respectively . regarding the single - particle dynamics , the trajectories are qualitatively different from those in pipes for the case of pushers and neutral swimmers .
pullers will swim parallel to the duct axis , as they do in pipes . in ducts ,
regardless of the size , the trajectories of the pushers are spiral .
of course , reflecting the duct shape , the projections onto the @xmath88-plane are no longer circular ( fig .
[ fig : pro_tra_duct ] ) .
the neutral swimmer shows a similar spiral trajectory , but it has a strong dependency on the initial state and sometimes the trajectories can become more complicated ( fig .
[ fig : pro_tra_duct ] ) . considering all the results for the single - particle system of pushers
, we can conclude that the emergence of the circular closed orbits depends strongly on the curvature of the confining walls .
therefore , if we increase the diameter of the pipe continuously from @xmath54 , at some point we will find a critical value for pushers to lose their orbit - type characteristics .
such a detailed study is beyond our scope .
the projection of the trajectories of a swimmer in duct onto the @xmath88-plane .
( a ) pusher with @xmath32 .
( b ) neutral swimmer .
the lines represent the projections of the trajectories , with color intensity used to represent the time . ] with regards to the many particle dynamics , we do nt see any big differences from those in pipes .
namely , the results are almost the same as those in bulk when the duct size is big enough , and the ordering seen for intermediate pullers collapses when the size is decreased . as a result , intriguingly , from the view point of the polar order of the many particle dispersion , the results are qualitatively the same between pipes and ducts
. however , it is possible that we one can observe unique dynamics in a skewed confinement .
we investigated the polar order formation in pipes , conducting three - dimensional direct numerical simulation with fully resolved hydrodynamics . as a result of the investigation over a broad range of parameters
, we confirmed that in most cases the dynamics of many particle systems in pipes match very well those in bulk .
we have only observed considerable differences over a small range of parameters .
we have presented an explanation for the emergence of wall effects by measuring the structural information .
they are mostly due to the wall accumulating effects and the reduction in the degree of order in the vicinity of the walls .
also , we have studied the effects of the change in pipe size .
we have clarified that for a specific range of the swimming parameter , the system undergoes an order / disorder phase transition when the size of the pipe shrinks .
considering the results of the structural analysis both in pipes and in bulk , we believe that this is due to the intrinsic clustering tendency of the swimmers .
there seems to exist a minimum cluster size which is necessary to maintain the polar order of intermediate pullers .
if we change the shape of the confinement , we found no qualitatively differences , from the view point of the polar order .
these results can be utilized to address the transport problem of microswimmers , however , further studies , like in winding pipes or in pipes with abrupt expansion / contraction are desired for realistic applications .
we thank h. ito for enlightening discussions . this work was supported by the japan society for the promotion of science ( jsps ) kakenhi grant no .
26247069 and also by a grant - in - aid for scientific research on innovative areas dynamical ordering of biomolecular systems for creation of integrated functions ( no .
16h00765 ) from the ministry of education , culture , sports , science , and technology of japan .
we also acknowledge the supporting program for interaction - based initiative team studies ( spirits ) of kyoto university . | we investigate the dynamics of model microswimmers under confinement , in cylindrical geometries , by means of three dimensional direct numerical calculations with fully resolved hydrodynamics .
such swimmers are known to show collective alignment in bulk , and we confirmed that similar alignment can be observed even in pipes , although the volume fraction dependency shows qualitative differences . by comparing the structural information , we investigated the cause of such differences .
we found an order / disorder phase transition for a specific type of swimmers , as the size of the pipe becomes comparable to the size of the swimmers .
such dynamics are found not to depend on the geometry of confinement . |
since the initial observation of exciton - polaritons in a strongly coupled microcavity in 1992 @xcite , a wide range of quantum many - body effects have been observed in polariton fluids such as bose - einstein condensation @xcite , and superfluidity exhibiting quantized vortices @xcite and solitons@xcite .
most of these results have been interpreted in terms of nonequilibrium bose gas theory , because the thermalization of the polaritons has been limited by their short cavity lifetime , on the order of 10 ps , compared to a thermalization time of the order of 1 ps .
our recent results @xcite have indicated that we can now produce structures which allow much longer lifetime , of the order of 200 ps . here
we report on accurate measurements of this lifetime using a unique method in which we inject polariton pulses at finite momentum into a microcavity and track their motion in time and space .
this allows us to observe `` slow reflection , '' in which renormalized light slows down to zero velocity , turns around , and goes back the other way .
in addition to providing a measure of the lifetime , the long - distance propagation seen here allows the possibility of beam - like polariton - interaction experiments and all - optical switching methods over long distances .
as the technology of microcavity polaritons is now well established , much attention has turned to increasing the lifetime of the polaritons , to allow better thermalization and to allow propagation over longer distances .
the lifetime of polaritons is a function of the intrinsic photon lifetime of the cavity and the fraction of photon in the polariton states . as amply discussed elsewhere@xcite
, a polariton state @xmath2 is a superposition of an exciton state @xmath3 and a photon state @xmath4 , @xmath5 where @xmath6 and @xmath7 are the @xmath8-dependent hopfield coefficients .
the @xmath0 signs indicate that there are two superpositions , known as the upper and lower polaritons ; in the experiments reported here we focus entirely on the lower polariton branch . at resonance , @xmath9 , while far from resonance the polariton can be nearly fully photon - like or exciton - like .
this implies that the @xmath8-dependent lifetime @xmath10 of the polaritons is given by @xmath11 for polaritons in our gaas - based samples , the rate of nonradiative recombination @xmath12 is negligible , so the lifetime is essentially entirely determined by the photon fraction and the cavity lifetime . in early polariton experiments@xcite , the cavity lifetime was on the order of 1 ps while the polariton lifetime was at most 10 - 15 ps , even well into excitonic detunings .
this implied that polaritons would only scatter a few times on average before decaying . in recent experiments@xcite
, the polariton lifetime has been extended to about 30 ps .
our previous work @xcite has given estimates of the polariton lifetime in new samples of the order of 100 - 200 ps , allowing polaritons to propagate hundreds of microns to millimeters within the cavity and to show a sharp transition to a superfluid state . because of the propagation of the polaritons to long distances away from the excitation spot , the configuration of those experiments made it difficult to get an accurate measure of the lifetime .
a measurement spatially restricted to the laser excitation spot would give a severe underestimate of the lifetime , because the polaritons do not stay put
they feel a force due to the cavity thickness gradient that pulls them to one side , leading them to travel hundreds of microns from the excitation spot . therefore , to accurately measure the lifetime , a measurement must track the polaritons in space as they move .
the measurements reported here do just that .
these measurements confirm the earlier estimates of the lifetime but considerably reduce the uncertainty .
the sample was arranged such that the gradient was aligned with the streak camera time slit , and then polaritons were injected at a large angle such that they moved directly against the gradient .
the experimental setup is shown in figure [ fig1 ] .
we used an objective with a wide field of view in addition to a large numerical aperture .
a resonantly injected picosecond pulse of polaritons was tracked as it entered the field of view , turned around and traveled away .
this occurs because the sample has a cavity thickness variation that leads to an energy gradient of the polariton . in simple terms
, one can think of the motion of the polaritons as governed by energy conservation with the following hamiltonian , which is just the same as that of a massive object moving in a potential gradient : @xmath13 here @xmath14 is the effective mass of the lower polariton branch that we observe , which depends weakly on @xmath8 , and is equal approximately to @xmath15 times the vacuum electron mass in these experiments .
the force @xmath16 is given by the gradient in space of the @xmath17 cavity resonance energy , and is approximately equal to @xmath18 mev / mm for the section of the microcavity studied here .
we will refer to `` uphill '' as moving toward higher cavity resonance energy ( narrower cavity width ) and `` downhill '' as moving to lower energy ( wider cavity width ) .
this experimental setup utilizes the fact that the polaritons in these high-@xmath1 samples flow over a great spatial distance and change in - plane momentum rapidly .
the lifetime of shorter - lived polaritons is more difficult to directly observe by streak camera measurements due to the overlap of any emission with the injecting laser . upon resonantly injecting polaritons ,
the created population is in the same state as the exciting laser .
the initial polariton population therefore will have the same characteristics as the exciting laser and can not be separated from it . observing any other state ( for example by looking at cross - polarized emission ) will inherently measure the scattering time of the polaritons to enter that state . in this experiment , we rely on the fact that polaritons will flow ballistically from the point of injection to the point of detection in order to separate the observed luminescence from the reflected laser . to the extent that this motion is ballistic , integrating the population over the observed spatial region
will directly yield the population decay of the polaritons . unlike the case of observing luminescence from a different energy or polarization state than the initial population , this method directly follows the decay of a single population rather than relying on an average over many @xmath8-states .
the momentum of the injected polaritons is controlled by the angle of the laser which generates them .
the angle of incidence used here was @xmath19 , corresponding to an initial wavevector of @xmath20 @xmath21 .
after propagating uphill for over two millimeters , the polaritons enter our spatial field of view and optical collection angle . observing the polaritons
far from the injection point reduces collection of scattered light from the laser excitation , and injecting polaritons at a large angle ensures that the reflected laser is outside the collection angle of the lens . as the polaritons flow against the cavity gradient they lose momentum , effectively exchanging in - plane kinetic energy for confinement energy , similar to a ball rolling uphill exchanging kinetic energy for gravitational potential energy .
since polaritons have a one - to - one relationship of in - plane momentum to external angle of emission for emitted photons , watching the luminescence change emission angle while the gas of polaritons propagates gives us a direct observation of their slowing . because the entire process is energy conserving , the injection laser , the ballistic polaritons and the emission all have the same wavelength .
once the polaritons reach a turn - around point , they flow back downhill and the emission angle increases to the negative direction . figure [ fig2 ] shows time - integrated real space emission intensity from the microcavity near the turn - around point of the polariton gas .
the coordinates in this image are such that the injection point is at roughly ( 0,0 ) , and the force due to the cavity gradient is nearly directly toward @xmath22 . while polaritons were injected primarily in the @xmath23-direction , the initial narrow spread of momenta in the y - direction led to a spread in real space after propagation over a long distance . at roughly @xmath24 mm ,
polaritons are seen entering the field of view , which also corresponds to the acceptance angle of the optics . the brightest streak ,
directly horizontal at y=0 mm , is the trajectory of the most intense part of the injected population which was peaked at zero momentum in the @xmath25-direction .
other bright streaks can be seen arcing to @xmath26 , and the entire range of states reach their respective turn - around points at @xmath27 mm .
the fact that there are bright streaks in this image rather than a smooth cloud suggests that the injection of the polaritons into the cavity occurs unevenly in momentum space .
the asymmetry of the cloud between @xmath28 and @xmath29 may be due to a slight misalignment between the cavity gradient and the injection direction . , @xmath25)=(0,0 ) , and the gradient is approximately toward @xmath22 .
polaritons approach this field of view from the left and turn around at @xmath30 mm before flowing back to @xmath22 .
the sharp cutoff at @xmath311.7 mm is due to clipping in the spectrometer.,width=326 ] to measure the lifetime , the bright jet of polaritons was time - resolved using a hamamatsu streak camera . to facilitate this , the sample was initially installed such that the gradient was aligned with the horizontal time slit on the streak camera .
this enabled us to track a single jet of polaritons while they propagate against the gradient , turn around , and travel backwards , as shown in figure [ fig3](a ) .
the vertical distance axis in this figure corresponds to the horizontal @xmath32-axis in figure [ fig2 ] .
the trajectory of the polaritons is easily seen in the data , which in this region is well described by a parabolic fit , as expected for the hamiltonian ( [ ham ] ) , which is equivalent to that of a ball moving with a constant force due to gravity . indeed
, these data directly demonstrate the in - plane velocity and acceleration of the polaritons during their trajectory .
one should note that this region of observation is already more than a millimeter and nearly 200 ps from the injection point , indicating that these polaritons are propagating farther and persisting longer than those in earlier samples , even without confinement in 1d structures , such as used in ref .
. 150 @xmath33
m @xmath34 150 @xmath33 m of fig .
( a ) intensity vs @xmath32-distance vs time of the propagating polaritons .
the dashed red line is a fit to the polariton motion as they feel a constant acceleration of @xmath35 mm / ns@xmath36 .
this acceleration is in good agreement with the expected value based on the known cavity gradient and the effective mass .
( b ) the polariton intensity of ( a ) summed in the @xmath32-dimension to highlight the exponential decay of the population .
the data are well fit by a single exponential decay with lifetime of @xmath37 ps.,width=326 ] a simple analysis of this data yields the polariton lifetime after summing in the spatial dimension , as shown in figure [ fig3](b ) .
the data are well fit by a single exponential with a lifetime of @xmath38 ps . for the region of the sample
observed in figure [ fig3 ] , the detuning of the polariton corresponds to the lower polariton approximately 75% photonic .
( although the polaritons move long distances , their detuning does not change much because they stay at the same energy . ) from this we estimate that the cavity photon lifetime is approximately 135 ps , which corresponds to a @xmath1-factor of over 320,000 .
it should be noted that this lifetime measurement may still be an underestimation of the lifetime .
close inspection of figure [ fig2 ] reveals that individual jets of polaritons are still spreading out from the central jet .
a population with some spread in initial momenta perpendicular to the cavity gradient must spread out horizontally while propagating uphill .
the fraction of polaritons that move out of our field of view will lead to an underestimation of the lifetime .
this error can be compensated for by using a narrower time slit to cut out adjacent jets at early times ; however , narrow slit widths can result in errors that will either underestimate or overestimate the polariton lifetime if the entirety of the main jet is not aligned with the time slit . in this experiment ,
data was collected over a range of slit widths from 50 to 300 @xmath33 m with consistent results .
in another set of experiments , we generated polaritons via a 200-fs pump pulse tuned to one - half the energy of the polaritons .
such experiments will be presented in more depth elsewhere , but we can see here that the propagation of long - lived polaritons can be used to probe the dynamics of two - photon generated polaritons .
we conducted two experiments .
first , we recreated geometry of the one - photon experiments discussed above by injecting polaritons uphill and observing the turn - around point , as shown in figure [ fig4](a ) .
second , we injected polaritons at normal incidence , directly at the point of observation , as shown in figure [ fig4](b ) .
figure [ fig4](a ) clearly shows characteristics similar to the one - photon resonant injection data presented in figure [ fig3 ] , with polaritons entering the field of view at @xmath39 200 ps after the injecting laser pulse and following a roughly parabolic trajectory .
however , the trajectory is very broad with a poorly defined turn - around point .
this can be understood by recognizing that the initial population is created in a range of energy states due to the spectral width of the exciting femtosecond pulse .
since there is a broad range of photon energy and momenta in the pump , a range of polariton states can satisfy energy and momentum conservation in the two photon absorption .
this method of observing polariton propagation can therefore enable us to characterize the initial polariton population . in fig .
[ fig4](b ) we present data from a different setup employing two - photon generation of polaritons . in this case
, we pump at normal incidence directly through the microscope objective used to image the luminescence .
since the pump laser has wavelength far from the polariton wavelength , there is no difficulty with scattered laser light . since the entire na of the objective was used , the angle of incidence of the pump light ranged over @xmath0 20@xmath40 even though the intensity was maximum at 0@xmath40 and the laser was spectrally tuned to the @xmath17 state .
the data indicate that polaritons created directly via two - photon absorption are peaked at an initial wavevector uphill , with no @xmath17 polaritons created initially .
the long - lived population stationary at @xmath41 is substrate luminescence excited by two - photon absorption .
figure [ fig4](c ) shows the same data as figure [ fig4](b ) , except that the image is centered to eliminate the clipping at @xmath42 in ( b ) .
the majority of the injected polaritons have finite @xmath8 , even though the pump light was centered at @xmath17 .
this supports the view that that two - photon absorption of the polaritons , which should be forbidden at @xmath17 due to the selection rules , becomes allowed at finite @xmath8 , due to valence - band mixing with the higher - lying light - hole states .
these results show clearly that two - photon resonant generation of polaritons is possible .
one can expect very strong nonlinear effects from microcavity polaritons due to the strong interaction of the cavity mode with the quantum well exciton@xcite .
we point out that there is no comparable one - photon excitation experiment such an experiment will not work because the exciting laser will be reflected directly back into the imaging system .
polaritons can be viewed as `` renormalized photons , '' especially in the region of the cavity where the the detuning makes the polaritons mostly photon - like .
as mentioned above , the behavior we have seen here can thus be viewed as a type of `` slow light , '' or `` slow reflection , '' in which the photons decelerate from @xmath39 3.5% of the speed of light to a full stop and then go back the other way .
this behavior is expected for light in a wedge - shaped cavity , without any need for the excitonic part of the polaritons .
however , it has been hard to directly observe , because one must have very high @xmath1 and fast time resolution to track the motion of the photons .
these measurements show that the photons can truly be viewed as having effective mass and feeling a force . with such long distance propagation and long lifetime
, it is now possible to construct experiments in which two or more beams are used and caused to interact .
this could be used to directly measure the polariton - polariton interactions and also for schemes of optical gating using polaritons , as presented e.g. in ref .
@xcite .
additionally , we have shown that two - photon injection of polaritons is a very relevant phenomena in gaas microcavities , but there is a noticeable dependence of the absorption strength on the angle of incidence . by temporally- and spatially - resolving the propagation of polaritons after generation
we can probe the mechanisms by which different states are occupied by this nonlinear absorption .
the work at the university of pittsburgh was supported by the national science foundation under grants phy-1205762 and eccs-1243778 .
the work at princeton university was partially funded by the gordon and betty moore foundation as well as the national science foundation mrsec program through the princeton center for complex materials ( dmr-0819860 ) . 10 c. weisbuch , m. nishioka , a. ishikawa , and y. arakawa , `` observation of the coupled exciton - photon mode splitting in a semiconductor quantum microcavity , '' physical review letters * 69 * , 33143317 ( 1992 ) .
j. kasprzak , m. richard , s. kundermann , a. baas , p. jeambrun , j. m. j. keeling , f. m. marchetti , m. h. szymaska , r. andr , j. l. staehli , v. savona , p.
b. littlewood , b. deveaud , and l. s. dang , `` bose einstein condensation of exciton polaritons , '' nature * 443 * , 40914 ( 2006 ) .
r. balili , v. hartwell , d. snoke , l. pfeiffer , and k. west , `` bose - einstein condensation of microcavity polaritons in a trap , '' science * 316 * , 10071010 ( 2007 ) .
k. g. lagoudakis , f. manni , b. pietka , m. wouters , t. c. h. liew , v. savona , a. v. kavokin , r. andr , and b. deveaud - pldran , `` probing the dynamics of spontaneous quantum vortices in polariton superfluids , '' physical review letters * 106 * , 115301 ( 2011 ) .
a. amo , s. pigeon , d. sanvitto , v. g. sala , r. hivet , i. carusotto , f. pisanello , g. lemenager , r. houdre , e. giacobino , c. ciuti , and a. bramati , `` polariton superfluids reveal quantum hydrodynamic solitons , '' science * 332 * , 11671170 ( 2011 ) .
b. nelsen , g. liu , m. steger , d. w. snoke , r. balili , k. west , and l. pfeiffer , `` dissipationless flow and sharp threshold of a polariton condensate with long lifetime , '' physical review x * 3 * , 041015 ( 2013 ) .
m. steger , g. liu , b. nelsen , c. gautham , d. w. snoke , r. balili , l. pfeiffer , and k. west , `` long - range ballistic motion and coherent flow of long - lifetime polaritons , '' physical review b * 88 * , 235314 ( 2013 ) .
a. v. kavokin , j. j. baumberg , g. malpuech , and f. p. laussy , _ microcavities _ ( oxford university press , new york , 2007 ) .
f. tassone and y. yamamoto , `` lasing and squeezing of composite bosons in a semiconductor microcavity , '' physical review a * 62 * , 063809 ( 2000 ) .
h. deng , g. weihs , d. snoke , j. bloch , and y. yamamoto , `` polariton lasing vs. photon lasing in a semiconductor microcavity , '' proceedings of the national academy of sciences of the united states of america * 100 * , 1531815323 ( 2003 ) .
g. tosi , g. christmann , n. g. berloff , p. tsotsis , t. gao , z. hatzopoulos , p. g. savvidis , and j. j. baumberg , `` sculpting oscillators with light within a nonlinear quantum fluid , '' nature physics * 8 * , 190194 ( 2012 ) .
e. wertz , l. ferrier , d. solnyshkov , r. johne , d. sanvitto , a. lematre , i. sagnes , r. grousson , a. v. kavokin , p. senellart , g. malpuech , and j. bloch , `` spontaneous formation and optical manipulation of extended polariton condensates , '' nature physics * 6 * , 860864 ( 2010 ) .
v. pellegrini , r. colombelli , i. carusotto , f. beltram , s. rubini , r. lantier , a. franciosi , c. vinegoni , and l. pavesi , `` resonant second harmonic generation in znse bulk microcavity , '' applied physics letters * 74 * , 19451947 ( 1999 ) .
s. lei , y. yao , z. li , t. yu , and z. zou , `` design and theoretical analysis of resonant cavity for second - harmonic generation with high efficiency , '' applied physics letters * 98 * , 13 ( 2011 ) .
d. ballarini , m. de giorgi , e. cancellieri , r. houdr , e. giacobino , r. cingolani , a. bramati , g. gigli , and d. sanvitto , `` all - optical polariton transistor . ''
nature communications * 4 * , 1778 ( 2013 ) . * supplemental information
this sample is the same as used in refs @xcite .
a @xmath43 microcavity contains three sets of four gaas quantum wells located at the antinodes of the cavity mode .
the qws are 70 pure gaas embedded in pure alas barriers at least 30 thick .
the optical mode is confined between distributed bragg reflectors made of alas / al@xmath44ga@xmath45as with 32 pairs on the top surface and 40 pairs on the bottom surface .
molecular beam epitaxial growth of the sample leads to an inherent wedge to the cavity thickness , resulting in a gradient of the cavity as well as exciton energies .
the polariton exhibits a rabi coupling of 6 mev at 5 k ; the cavity mode gradient is 13 mev / mm and the exciton gradient is 1.5 mev / mm .
the sample was held in a cold - finger cryostat at 5 k for all experiments .
emission was collected using a n.a.=0.42 microscope objective .
a preliminary imaging lens permitted spatial filtering of the real space image data , and a subsequent iris in the fourier image plane permitted filtering of the emission angle .
secondary lenses could be exchanged to image either the real - space or angle - resolved emission .
luminescence was imaged through a spectrometer onto either a standard ccd or onto a hamamatsu streak camera .
in the resonant injection experiment , polaritons were resonantly injected at @xmath46 nm with a picosecond laser far on the photonic side of the sample .
the injected state had a detuning of approximately -2.6 mev and corresponded to an external angle of roughly 42@xmath47 ( @xmath48 @xmath21 ) .
the sample and pump laser were arranged such that the polaritons were moving anti - parallel to the cavity gradient , which was aligned with the time slit .
the angle of incidence was larger than the collection angle of the optics , so the reflected beam was not collected . additionally , the pump spot was spatially outside the field of view such that scattered light was not collected . at a distance of approximately 2 mm from the injection point ,
emission entered the collection range of the optics . at the turn - around point , the polaritons are more photonic with a detuning of -7.4 mev which corresponds to a photon fraction of 75% .
for the two - photon injection experiment , a 200-fs pulse generated by a coherent opa system was tuned to one - half the energy of the desired polariton transition . for the 40@xmath40 injection case ,
the laser was tuned to excite at @xmath46 nm . for the 0@xmath40 injection case , the laser was tuned to excite at @xmath49 nm .
the x - distance from the injection point for the 40@xmath40 injection cases was estimated as follows : the @xmath32-distance in figs . [ fig2],[fig3](a ) , and [ fig4](a ) was determined from the fit of the polariton @xmath32 vs @xmath50 trajectory presented in fig . [
fig3](a ) .
extrapolation of this fit back to time @xmath51 as determined by locating scattered laser light determines the initial position of excitation .
this initial excitation position is consistent with the sample parameters and injection conditions .
this method assumes that the acceleration of the polaritons is strictly constant from creation to turn around .
variation in the acceleration due to a non - constant energy gradient in addition to the changing mass of the polariton implies uncertainty on the overall offset of this axis , but the spatial magnification was measured directly . | we resonantly inject polaritons into a microcavity and track them in time and space as they feel a force due to the cavity gradient .
this is an example of `` slow reflection , '' as the polaritons , which can be viewed as renormalized photons , slow down to zero velocity and then move back in the opposite direction .
these measurements accurately measure the lifetime of the polaritons in our samples , which is 180 @xmath0 10 ps , corresponding to a cavity leakage time of 135 ps and a cavity @xmath1 of 320,000 .
such long - lived polaritons propagate millimeters in these wedge - shaped microcavities .
additionally , we generate polaritons by two - photon excitation directly into the polariton states , allowing the possibility of modulation of the two - photon absorption by a polariton condensate . |
numerical simulation is one of the most powerful methods for searching thermal equilibrium states of the system .
it has elucidated numerous novel properties that one could hardly deal with analytically .
one of the standard techniques is the markov - chain monte carlo sampling on the basis of the metropolis algorithm , e.g. , the single - spin flip update used for classical spin systems .
a common difficulty in such numerical simulations is that the state is often frozen into some configuration and hardly updated , especially when the system has a peculiar energy landscape with multiple local minima or almost flat energy dependence in the phase space .
for instance , in some classical spin systems , the spin configuration is frozen in a multiple - domain state , and hardly escapes from the metastable state by local spin - flip update . for avoiding such freezing problems ,
a variety of the simulation techniques have been proposed , such as the simulated annealing @xcite , cluster update @xcite , and replica exchange methods @xcite . in this paper
, we propose an efficient technique , which will be useful for preventing the system from being frozen into multiple - domain states .
our method is similar to the floating zone method for growing a single crystal in experiments .
specifically , a part of the system is heated up to ` melt ' the frozen structure , and the molten region is shifted smoothly to sweep away local defects or domain boundaries during the simulation .
the procedure helps to grow a single - domain state efficiently , similar to the crystal growth in experiments .
that is the reason why we call this method the simulated floating zone ( sfz ) method " .
we demonstrate the efficiency of the sfz method by applying it to a complicated magnetic ground state in the kondo lattice model with classical localized spins .
we show that the multiple - domain structures are quickly swept away and the system reaches a single - domain state .
we find that the efficiency of the sfz method is optimized by setting the local heating temperature comparable to the characteristic energy scale of the ordered phase .
the rest of the paper is organized as follows . in sec .
[ sec : method ] , we introduce the fundamental idea of the sfz method . in sec .
[ sec : appl ] , we demonstrate the efficiency of the sfz method by applying it to the langevin dynamics simulation for the kondo lattice model .
we summarize our results in sec .
[ sec : summary ] .
the idea of the sfz method is simple . it follows the floating zone method used in the single crystal growth . in the floating zone method ,
a narrow region of the crystal is melted by heating , and this molten region is shifted along the crystal in a rod shape .
the procedure suppresses the grains and domains efficiently , and results in a purified single crystal .
we can introduce the similar technique in numerical simulations for obtaining spatially - uniform equilibrium state .
thus , the sfz method consists of the following steps : ( 1 ) a part of the system is heated up so as to melt the metastable configuration , ( 2 ) the molten region is moved smoothly along the system , and ( 3 ) the procedures are repeated a sufficient number of times , until a spatially - uniform stable state is reached . figure [ fig : sfz_schematic ] shows a schematic picture of the sfz method .
the red region is the heated region , whose temperature is set at a higher temperature ( @xmath0 ) than the other parts of the system ( @xmath1 ) .
the heated region is shifted along the system in the direction of the arrow .
the size and shape of the heated region as well as the speed and direction of its movement are arbitrary : one can tune them so as to reach the stable configuration quickly .
the sfz method is easy to implement and generally compatible with many simulation techniques .
it is also applicable to a wide range of systems , both classical and quantum , in any spatial dimensions .
schematic picture of the sfz method .
the red region is the heated region , which is smoothly shifted in the direction of the arrow .
, width=302 ]
in this section , we demonstrate the efficiency of the sfz method in the numerical search of the stable ground state , where domain formation makes difficult to reach a spatially - uniform state . as a typical example
, we here consider a noncoplanar spin texture emergent from the interplay between charge and spin degrees of freedom in electrons .
specifically , we focus on a magnetically - ordered state composed of a periodic arrangement of spin vortices ( meron , i.e. , half - skyrmion ) , which we recently found as a candidate for the ground state of the kondo lattice model with classical localized spins on a square lattice @xcite .
the hamiltonian is given by @xmath2 where @xmath3 is a creation ( annihilation ) operator of an itinerant electron with spin @xmath4 at site @xmath5 , @xmath6 is the vector representation of the pauli matrices , and @xmath7 is the classical localized spin at site @xmath5 with @xmath8 . the first term in eq .
( [ eq : klm ] ) is the hopping term of itinerant electrons with transfer integral @xmath9 : here , we consider only the nearest- and third - neighbor hoppings , @xmath10 and @xmath11 , respectively , on the square lattice . the second term in eq .
( [ eq : klm ] ) is the onsite hunds - rule coupling with the coupling constant @xmath12 .
we set the lattice constant @xmath13 , the reduced planck constant @xmath14 , and the boltzmann constant @xmath15 .
the vortex crystal state that we found in ref .
@xcite exhibits stripes of the spin scalar chirality in the diagonal direction of the square lattice .
the spin scalar chirality is defined for each square plaquette @xmath16 as @xmath17 , \label{eq : chi}\ ] ] where sites @xmath18 are the vertices of the plaquette @xmath16 in the counterclockwise direction .
thus , the vortex crystal state has trivial twofold degeneracy with respect to the directions of chirality stripes , which are connected by in - plane fourfold rotation , in addition to the global rotation in the spin space .
it is this twofold degeneracy that leads to domain formation in the current system , as discussed later in fig .
[ fig : d_sweeps](b ) .
we employ the sfz method in the numerical simulation using the langevin dynamics for the update of the configurations of localized spins @xcite .
the langevin dynamics of the localized spins is described by the landau - lifshitz - gilbert equation @xcite as @xmath19 where the first ( second ) term describes a precession ( dumping ) of spins .
@xmath20 is the internal effective magnetic field acting on @xmath7 , which is given by @xmath21 here , @xmath22 is the grand potential for a given spin configuration @xmath23 , which is calculated by using the kernel polynomial method @xcite ; @xmath24 describes thermal fluctuations in the langevin dynamics , which satisfies @xmath25 [ @xmath26 is the time average and @xmath27 is the @xmath28 component of @xmath24 ( @xmath29 , @xmath30 , @xmath31 ) ] . in this thermal - fluctuation term
, we introduce the sfz method .
the system within a narrow diagonal strip is heated up to a temperature @xmath0 , and moved along another diagonal direction ; see the schematic picture in fig .
[ fig : d_sweeps](a ) ; we take the heated strip in the diagonal direction because the chiral stripes appear along the diagonal ( or @xmath32 rotated ) direction , as shown below .
( a ) schematic of the sfz method applied to the model in eq .
( [ eq : klm ] ) on the square lattice .
the red region represents the heated strip .
the arrow indicates the direction of the movement of the strip : the strip moves from the bottom - left to top - right corner , and restarts from the bottom - left corner in the next cycle .
( b)-(h ) real - space distributions of the spin scalar chirality [ eq . ( [ eq : chi ] ) ] in the simulation with the use of the sfz method after the sweeps indicated in each figure .
the simulation was performed at @xmath10 , @xmath11 , and @xmath33 on the square lattice with @xmath34 sites .
, width=604 ] figure [ fig : d_sweeps ] shows the results of the langevin dynamics simulation accelerated by the sfz method .
we consider the model in eq .
( [ eq : klm ] ) with @xmath35 on the square lattice with @xmath34 sites and periodic boundary conditions .
we set the electron filling so as to realize the vortex crystal with ordering vectors @xmath36 @xcite .
the langevin dynamics simulation is performed by taking @xmath37 and the time interval as @xmath38 . in the kernel polynomial method
, we take the chebyshev order @xmath39 and the number of random vectors @xmath40 .
we use the heated strip with four - site width ( we do not impose the periodic boundary conditions on the strip shape ) , and set the temperature at @xmath41 , while taking @xmath42 .
the heated strip is shifted in a constant speed along the system so that one sweep from the bottom - left to top - right corner takes a time of @xmath43 : the langevin dynamics update is performed for @xmath44 times during the single sweep .
the next sweep starts from the same bottom - left corner just after the previous sweep ends .
figure [ fig : d_sweeps](b ) shows a typical multiple - domain state obtained by the langevin dynamics simulation without the sfz method staring from a random spin configuration .
we here plot the real - space distribution of the scalar chirality @xmath45 defined in eq .
( [ eq : chi ] ) .
the system has several domains ; each domain has chiral stripes running in the diagonal - upward or downward direction . in this state ,
the spin configuration is almost frozen : the domain walls are hardly removed by the update in the langevin dynamics .
once we turn on the sfz method , however , the multiple domains are smeared and the system reaches a single - domain state .
figures [ fig : d_sweeps](c)-[fig : d_sweeps](h ) show the evolution of the real - space distributions of @xmath45 during the simulation with the use of the sfz method .
as shown in the figures , when we repeat the sweep , one type of the two degenerated domains grows and merges into larger domains .
after eight cycles , we successfully obtain the spatially - uniform state , as shown in fig .
[ fig : d_sweeps](h ) .
we note that the direction of the chiral stripes in the final uniform state is always parallel to the heated strip : when we rotate the strip and its motion by @xmath46 in the sfz simulation , the other domain with @xmath46-rotated stripes dominates the system in the final uniform state .
such a selection of energetically - degenerate domains might occur in the sfz method , depending on the microscopic energetics in the system as well as the details of the floating zone . in the current case ,
we note that the selected domain has the dominant spin helix parallel to the heated strip @xcite .
the reason why this domain is chosen is presumably because the disturbance of electron motion along the spin helix leads to a substantial energy loss in the spin state .
time evolution of the largest peak ( a ) in the spin structure factor , @xmath47 , and ( b ) in the chirality structure factor , @xmath48 , divided by the system size for different temperatures of the heated region in the sfz method , @xmath49 , @xmath50 , @xmath51 , @xmath52 , and @xmath53 . , width=604 ] let us discuss the efficiency of the sfz method while changing the heating temperature @xmath54 . in order to measure the efficiency , we monitor the time evolution of the dominant peaks of the structure factors for spin and chirality degrees of freedom .
the structure factors are defined by @xmath55 where @xmath56 is the system size and @xmath57 is the real - space coordinate of site @xmath5 .
figure [ fig : sweep_ts ] shows the time evolution of the largest peaks of @xmath47 and @xmath48 in the sfz simulation for several different @xmath54 by starting from the same initial state shown in fig .
[ fig : d_sweeps](b ) . here
, we plot the data at every sfz sweep .
when we set @xmath58 , both peaks grow very slowly and do not change substantially from the initial values after the long simulation time .
this is presumably because @xmath58 is too low to melt the frozen domain structure . for @xmath59 ,
both peaks grow during the simulation , and finally saturate at the same values , which are expected for the single - domain state , after the sufficiently long simulation time .
when further increasing @xmath54 , however , the time necessary for the convergence becomes longer . for both @xmath47 and @xmath48 , @xmath60-@xmath51 optimizes the efficiency of the sfz simulation in the current case .
we note that the optimal value of @xmath54 is comparable to the energy scale of the magnetic ordering . in the kondo lattice model in the small @xmath12 region ,
the energy scale is set by the so - called ruderman - kittel - kasuya - yosida ( rkky ) interaction , which is given by the second - order perturbation in terms of @xmath12 @xcite . in the current case ,
the energy scale is about @xmath61 , which is close to the optimal @xmath62-@xmath51 .
thus , the result suggests that the sfz method works most efficiently when the heating temperature is comparable to the melting temperature of the corresponding ordered phase .
this is presumably because the small part of the system included in the heated strip is easily aligned in a spatially - uniform ordered state owing to the growing correlation length near the melting temperature .
in summary , we have introduced the sfz method as an efficient technique to obtain the thermal equilibrium state in numerical simulations . in this method , a part of the system is locally heated up and shifted smoothly all over the system .
the main purpose of using this method is to accelerate the numerical convergence , preventing the state from being frozen in a local minimum or a flat energy landscape in the phase space .
the advantage is that this method is simple and applicable to many situations . in this paper , we have applied the sfz method to the langevin dynamics simulations for the square kondo lattice model as an example where domain formation is often problematic for obtaining the ground state in the simulations without the sfz method .
we have demonstrated that the introduction of the sfz method sweeps away the domain structures and substantially accelerates the convergence to the spatially - uniform state .
in addition , we have found that the convergence is most accelerated by setting the local heating temperature comparable to the characteristic energy scale of the ordered state .
the sfz methods based on the langevin dynamics simulations were carried out at the supercomputer center , institute for solid state physics , university of tokyo .
is supported by the japan society for the promotion of science through a research fellowship for young scientists and the program for leading graduate schools ( alps ) .
this research was supported by kakenhi ( no . 24340076 , 26800199 ) , the strategic programs for innovative research ( spire ) , mext , and the computational materials science initiative ( cmsi ) , japan .
10 url # 1#1urlprefix[2][]#2 kirkpatrick s 1984 _ j. stat . phys . _ * 34 * 975 | this paper provides the simulated floating zone ( sfz ) method that is an efficient simulation technique to obtain thermal equilibrium states , especially useful when domain formation prevents the system from reaching a spatially - uniform stable state . in the sfz method ,
the system is heated up locally , and the heated region is steadily shifted , similar to the floating zone method for growing a single crystal with less lattice defect and impurity in experiments .
we demonstrate that the sfz method suppresses domain formation and accelerates the optimization of the state , taking an example of a magnetic vortex crystal state realized in itinerant magnets .
we show that the efficiency is maximized when the local heating temperature is tuned to be comparable to the characteristic energy scale of the ordered state . |
the first extra solar planet was discovered in the early 1990s by @xcite . today
the statistics of the observations show that the architecture of our solar system seems to be unique compared with exoplanetary systems . at the moment we know about 2000 exoplanets in more than 1200 planetary systems , among them more than 100 exoplanets are in binary - star systems and two dozen are in multiple - star systems .
the data of all planets are collected in the exoplanet - catalogue maintained by j. schneider ; whereas the binary and multiple - star systems can be found separately in the catalogue of exoplanets in binary star systems maintained by r. schwarz , which we will also introduce in this paper . + approximately 70 percent of the main- and pre - main - sequence stars are members of binary or multiple star systems : 67 % for g - m star , e.g. @xcite ; and approximately 70 % for o - b stars ( e.g. @xcite , @xcite )
. statistics of solar - type dwarfs were studied by @xcite with a distance - limited sample of 4847 targets .
a field population was found of about 54% for single stars , 33% binary stars , 8% triple systems , 4% for quadrupole systems , 1% for systems
observational evidence indicates that many of these systems contain potentially planet - forming circumstellar or circumbinary discs , implying that planet formation may be a common phenomenon in and around binary stars ( e.g. @xcite , @xcite , @xcite , @xcite ) .
this fact led many research groups to examine the planetary formation and evolution and dynamical stability in binary star systems , either in general or for selected systems @xcite . despite many theoretical studies on the planetary formation in double star systems ,
the formation processes are not entirely understood @xcite .
+ from the dynamical point of view the binary star systems as well as multiple star systems are particularly interesting . according to the work of @xcite
one can distinguish three types of planetary orbits in a binary star system : 1 .
s - type or circumstellar motion , where the planet orbits one of the two stars ; 2 .
p - type or circumbinary motion , where the planet orbits the entire binary ; 3 .
t - type : a planet may orbit close to one of the two equilibrium points @xmath1 and @xmath2 ; we call them trojan planets
. the dynamical study of @xcite could show with a few real binary systems that the t - type configuration is not only of theoretical interest and @xcite could show that t - type orbits can be detected with etv signals . the graphic representation of the different dynamical scenarios is given in fig .
[ fig1 ] . the first planet in p - type motion , was detected in 2009 ( hw vir ( ab ) c , * ? ? ?
since that time planets in well detached binary systems become more and more attractive , especially tight coplanar circumbinary planets around short - period binaries @xcite .
further p - type planets were discovered in the following years , where especially the space - mission kepler was very successful . among them are also multiplanetary circumbinary systems , like hw virginis or kepler 47 @xcite . + from the observational point of view well detached binary star systems with separations smaller than 3 au are more interesting than wide binary systems because the observation time for the latter ones is much longer .
furthermore , well detached binaries offer reasonable signal - to - noise ratio ( s / n ) values for photometry and radial velocity ( rv ) amplitudes @xcite .
a first study of test particles in circumbinary orbits was presented by @xcite , @xcite and @xcite .
@xcite studied the dynamics of binary star systems with a circumbinary planet , and calculated its eclipse timing variations ( etvs ) for different values of the mass ratio and orbital elements of the binary and the perturbing body .
+ most observations of planets in binaries are focused on @xmath3 ( stars have similar masses ) and are restricted to sun - like stars . in fig .
[ fig2 ] we show a distribution of the mass ratios of all detected exoplanets in binaries and we found that the most common mass ratios @xmath4 are @xmath5 and 0.5 .
therefore we use different mass ratios for our simulations for p- and s - type systems .
this paper is divided into three parts : the first part is devoted to the possible detection of exoplanets in well detached binary star systems in p- and s - type motion by the help of eclipse timing variations ( etv ) . in the second part we prepare statistics for well detached binary star systems from several catalogues and discussed the possibility of further candidates .
the actual statistics of planets in binaries and multiple star systems are taken from the catalogue of exoplanets in binary star systems which we introduce in the chapter 6 .
( upper graph ) and the separation ( @xmath6 , shown in the lower graph ) of all binary star systems with exoplanets ( including the binaries in multiple star systems , see figs .
[ triple ] , [ quad ] ) , taken from the binary catalogue of exoplanets ( http://www.univie.ac.at/adg/schwarz/multiple.html ) . to see the well detached binary systems we zoom into the histogram of @xmath6 in the inset.,width=219 ] ( upper graph ) and the separation ( @xmath6 , shown in the lower graph ) of all binary star systems with exoplanets ( including the binaries in multiple star systems , see figs .
[ triple ] , [ quad ] ) , taken from the binary catalogue of exoplanets ( http://www.univie.ac.at/adg/schwarz/multiple.html ) . to see the well detached binary systems we zoom into the histogram of @xmath6 in the inset.,title="fig:",width=328 ]
the photometric detection of extrasolar planets is of particular interest for the discoveries in eclipsing binaries .
we investigated well detached binary star systems , where the initial separation of the stars is 0.5 to 3 au . from the dynamical point of view
these initial separations are very interesting , because planets in s- and p - type orbits are possible and they are supported by the first kepler discoveries of a long ( @xmath7 ) eclipsing binary with a pulsating red giant component @xcite we studied the planar full three - body problem ( 3bp ) with numerical integrators . in this problem two finite bodies , the primaries ( @xmath8 @xmath9 , @xmath10 @xmath9 ) revolve about their common center of mass , starting with three different eccentricities ( @xmath11 ) . a third body @xmath12 moves around @xmath13 for s - type or around both stars for p - type motion in the same plane as @xmath14 .
we have regarded all the celestial bodies involved as point masses and integrated the equations of motion for a time up to @xmath15 yrs for the etv - maps shown in section [ s ] and [ p ] and @xmath16 yrs for the stability limits . for our simulations we used a gauss radau integrator with an adaptive step size
( see * ? ? ?
* and references therein ) for the etvs and for the stability the lie - method with an automatic step size control to solve the equations of motion . in the s - type configuration we also investigated close - in planets ,
therefore we considered the general relativity in our calculations , @xcite .
we considered well detached eclipsing binaries , with distances between the stars of 0.5 , 1 and 3 au . to ensure that the effect of the variations of the mass ratio @xmath17 of the binary is included in our study , we considered the following three models : * * model 1 : * @xmath13 = @xmath14 = 1@xmath18 , corresponds to @xmath19 * * model 2 : * @xmath13 = 1@xmath18 and @xmath14 = 0.5@xmath18 , corresponds to @xmath20 * * model 3 : * @xmath13 = 1@xmath18 and @xmath14 = 0.1@xmath18 , corresponds to @xmath21 as shown in fig .
[ fig2 ] the mass ratio of our models is quite common , when we look at the histogram of the detected exoplanets in binaries and multiple star systems .
we changed the other two models ( @xmath20 and @xmath21 ) because the statistics of other binary catalogues ( section[statistics ] ) are equally - distributed . to get a good estimation about occurring perturbations on the secondary star ( to measure etvs ) we used planets with different masses @xmath22 : earth which corresponds to @xmath23 @xmath18 ] , neptune which corresponds to @xmath25 @xmath18 ] , and jupiter which corresponds to @xmath26 @xmath18 ] .
for the analysis of the orbit we used the method of the maximum eccentricity @xmath27 . in former studies we found a good agreement with chaos indicators like the lyapunov characteristic indicator ( lci ) ( e.g. * ? ? ?
* ; * ? ? ?
the @xmath27 method uses as an indication of stability a straightforward check based on the maximum value of the planet s eccentricity reached during the total integration time ( @xmath28 ) .
if the planet s orbit becomes parabolic ( @xmath29 ) the system is considered to be unstable .
the @xmath27 is defined as follows : @xmath30 since the first exoplanets in p - type motion were detected , the investigation of the eclipse timing variation became more and more important .
the etv signal of the secondary star will be induced by an additional planet .
this gravitational perturbation affects the motions of the two stars and cause their orbits to deviate from keplerian . in an eclipsing binary , these deviations result in variations over time and duration of the eclipse .
this method is particularly important in the case when the planet s orbit is not in the line of sight , which causes the absence of a transit signal .
however , such planets cause perturbations in the orbit of the transiting star , leading to detectable etvs .
similar investigations for transit timing variations ( ttvs ) were done in several articles like e.g. @xcite ; @xcite and @xcite .
the feasibility of the detection of extrasolar planets by the partial occultation on eclipsing binaries was investigated by @xcite .
the goal of our work was to show which planet sizes for the s- and p - type configurations are detectable in the etv signal of the secondary star with current observational equipment . in order to approximate the detectability of possible extrasolar planets by means of etvs we used the work of @xcite who investigated the sensitivity of the eclipse timing technique for the ground and space - based photometric observations .
they showed in a best - case scenario ( excluding e.g. star spots or pulsations ) , that the typical photometric error ( detectable timing amplitude dt ) for corot is about dt = 4 sec for a brightness ( l ) of 12 [ mag ] and dt = 16 sec for l=15.5 [ mag ] .
kepler has a dt = 0.5 sec for l=9 [ mag ] and a dt = 4 sec for l=14.5 [ mag ] .
future space missions will support the effort to detect smaller planets , like for example : * plato ( planetary transits and oscillations of stars ) will monitor relatively nearby stars to hunt for sun - earth analogue systems @xcite .
* tess ( transiting exoplanet survey satellite ) space mission is dedicated to detect nearby earth or super - earth - size planets on close - in orbits around the brightest m dwarfs @xcite .
* cheops ( characterising exoplanets satellite ) will examine transiting exoplanets of known bright and nearby host stars @xcite . for our investigations we will use as detection criterion the photometric precision of corot @xmath31 sec as well as that of kepler @xmath32 sec .
we determined the etvs by calculating the amplitude for the perturbed case , where the planet - induced constant rate of apsidal precession is removed by a linear fit .
we also took into account the long - term effects caused by the binaries motion around the systems center of mass and the light travel time effect @xcite .
several observations of exoplanets in detached binaries motivated us to investigate the possible detection of exoplanets with etv s .
therefore we used the configuration `` primary star - planet - secondary star '' with the following initial conditions : * * masses : * as shown in section [ models ] we used different masses for the primary and the secondary star ( model 1 , 2 and 3 ) , which we think represents a quite common mass ratio for binaries and might be useful for future observations of different stars ( see discussion section [ statistics ] ) .
+ for the planets we used three different masses : jupiter ( @xmath33 ) , neptune ( @xmath34 ) and earth ( @xmath35 ) .
* * semi - major axis : * we considered eclipsing binaries , with separations between the stars of @xmath6=0.5 , 1 and 3 au .
the outermost stability border for the possible exoplanets were taken from the literature @xcite and verified by numerical integrations . within the stability borders we integrated 80 equally distributed configurations in case of @xmath6=0.5 au and @xmath6=1 au . for @xmath6=3
au we used 160 equally distributed configurations . *
* eccentricity : * the eccentricity of the planet was varied between 0.0 and 0.5 ( and divided into 80 data points ) . the binary s eccentricity was set to 0 , 0.2 and 0.4 .
* all other orbital elements were set to zero ( @xmath36 = @xmath37 = m = 0 ) . for our computations of the etv - maps , we changed the distance from the planet to the primary star and the eccentricity of the exoplanet .
the grid size of the etv maps were changed from 80x80 ( @xmath38x@xmath39 ) for @xmath6 0.1 and 1 au and extended to 160x80 for @xmath6=3 au .
an example is given in fig .
[ etv1 ] for model 1 ( @xmath40 @xmath18 ) for 3 different masses of the planets : 1 @xmath33 ( fig .
[ etv1 ] upper left graph ) , 1 @xmath34 ( upper right graph ) and 1 @xmath35 ( lower graph ) .
the separation of the binaries is @xmath41 au and they have non eccentric orbits .
the stability border for the separation of @xmath6=1 au for all planets ( @xmath33 , @xmath34 and @xmath35 ) is roughly a=0.3 au .
close or outside the border , the influence of the secondary becomes too large and the planets escape ( see fig .
[ etv1 ] white region ) . as one can see the stability border shrinks for higher eccentricities of the planet ( @xmath42 ) .
the timing amplitude dt of planets with @xmath33 is 10 times larger ( @xmath43 ) than for @xmath34 ( @xmath44 ) and very small for the earth . for the @xmath32 sec it is possible to detect earth - like planets as well as by future space missions , which will have a better time resolution or photometric errors .
+ tables [ stype_nep ] and [ stype_earth ] summarise our results .
the table for planets with 1 @xmath33 is not shown , because for all stable initial conditions the etv - maps would be 100 percent detectable within the typical photometric error of corot ( @xmath45 ) and kepler ( @xmath46 ) .
the neptune - sized planets are detectable for almost all stable orbits in the etv - map for @xmath46 , whereas for @xmath45 especially the etvs for small separation of the binaries ( @xmath6=0.5 ) and larger eccentricities ( @xmath47 and @xmath48 ) are not detectable .
this is similar for earth - like planets .
however , much more stable orbits are not detectable for both photometric errors .
for @xmath45 almost no etv signals are detectable ( only for @xmath49 with the model 2 and model 3 for low values of @xmath50 ) . [ cols="^,^,^,^,^ " , ] [ table : catalogs ] an early example of a comprehensive data collection on binary star systems includes the work of @xcite which is based on the finsen - worley catalogue " published in @xcite . the
4th catalogue of orbits of visual binaries " contains orbital elements of about 930 objects in 847 systems , whereby triples are counted as two systems .
the statistical distribution of the semi - major axis illustrates that almost 96% of the visual binaries are located at a angular separation of less than 5 arcsec . in 1988 and 1989 , the 15th complementary catalogue of sbs " was published by @xcite and @xcite , respectively .
this database contains the orbital data and the derived masses of 436 spectroscopic binaries .
the statistical analysis of the available data of 310 semi - major axes shows that just above 92% of the stars have an @xmath51 @xmath52 200 au ( minimum distance ) .
the catalogue of eclipsing binaries parameters " of @xcite and @xcite respectively , not only includes orbital parameters , masses , and luminosities but also photometric orbit data of 44 eclipsing binary systems whereby all values of @xmath53 are smaller than 0.5 au .
@xcite and @xcite respectively , published the `` catalogue of dms - type eclipsing binaries '' which contains information of 113 binaries with photometric and spectroscopic parameters .
the semi - major axis of the detached main - sequence - type eclipsing binaries is illustrated in fig .
[ fig : a_bin_1999_svechnikov ] .
the catalogue of semi - detached eclipsing binaries " of @xcite from the year @xcite is a collection of slightly more than 230 semi - detached eclipsing binary star systems with known photometrical orbital elements .
the distribution of the semi - major axes illustrates a significant accumulation of systems with an @xmath53 smaller than 0.1 au .
just over 85% of the eclipsing binaries are located in that area . the semi - major axis of 155 spectroscopic binaries with red - giant primaries in open clusters @xcite are shown in the left - hand panel of fig .
[ fig:2007_mermilliod ] .
orbital periods range from 2.07 to 689 days ( 1.89 years ) .
it is apparent from this distribution that almost 62% of the spectroscopic binaries have a semi - major axis smaller than 100 au with a nearly exponential increase of binaries towards smaller @xmath53 .
the statistical evaluation for @xmath53 @xmath52 20 au is presented in the right - hand panel of fig .
[ fig:2007_mermilliod ] .
we found out that most of the detected exoplanets in binary - star systems are stars with masses like our sun or slightly smaller ( as shown in the catalogue of exoplanets in binary star systems ) .
but , @xcite showed that close pairs of white dwarfs are very common in our galaxy , with the order of 100 - 300 million .
it is of particular interest for our study to find binaries within a separation of @xmath53 @xmath52 3 au .
based on our statistics , we have found a set of 462 candidates that fit our requirements . with regard to the planning of possible future catalogues , a uniform classification of the semi - major axis ( e.g. in au )
would be preferable in order to evaluate statistical analyses and to provide useful information .
the extrasolar planets catalogue was the first online catalogue and is available since february 1995 at http://exoplanet.eu @xcite .
the catalogue has been upgraded in 2005 by additional graphical and statistical online services @xcite .
other databases followed some years later : the california and carnegie planet search table at http://exoplanet.org @xcite and the geneva extrasolar planet search programmes table ( mayor , queloz , udry and naef ) providing first hand data from the observers using the radial velocity and transit method .
the advantage of the extrasolar planets encyclopaedia @xcite maintained by the exoplanet team is that it lists all detection methods ( astrometry , pulsar timing , microlensing , imaging etc . ) . cataloguing the data of exoplanetary systems
becomes more and more important , due to the fact that they conclude the observations and support the theoretical studies .
since planets in binary star systems were detected they become more important . in 2013
we started to compile a catalogue for binary and multiple star systems because at that time there did not exist a list of exoplanets in binary star systems .
now also the open exoplanet catalogue shows exoplanets in binary and multiple star systems , which is a community driven and decentralised astronomical database and available at http://www.openexoplanetcatalogue.com/ @xcite . at the beginning of our catalogue
we wanted to supplement the `` extrasolar planets encyclopedia '' , in agreement and with the support of j. schneider and his team . in case of binary and multiple star systems
the challenge is big due to the fact that the observations are more complicated and the data have much more errors than for single star systems .
what concerns us primarily are the statistics , that is why we do not present the errors in our list .
if more details are needed , we made a link to the extrasolar planets encyclopedia ( the link is contained in the name of the system ) .
another purpose of our catalogue is to present review statistics of other binary catalogues , which is a big challenge because the catalogues present only very special stars or regions of our galaxy and are non - uniform .
the catalogue is described as it was in the year 2016 , organised in 12 columns and will be updated monthly .
we distinguish detection from discovery , because some planets for example are discovered by radial velocity and detected by transit afterwards .
one can sort in two directions : ascending , meaning from the lowest value to the highest , or descending .
for example , by clicking on the header e.g. discovery the list will be sorted after the largest value , when you click again it will be sorted after the smallest value .
the list is originally sorted by the distance between the binaries ( @xmath6 ) , all rows can be sorted in the same way except the comments .
in addition an introduction and help is also given in the menu bar of the catalogue including an example list .
to make your own statistics the data is available as .csv file .
all systems are linked to the extrasolar planets encyclopedia some of the systems to the open exoplanet catalogue where one can find references and additional data on the systems .
this part of the catalogue represents only the stellar data of the system ( see fig . [ star ] ) . +
* system * + name or designation of the system and the structure of the system , where capital letters refer to a star , and small letters refer to a planet .
example : dp leo ab b `` ab b '' or `` a bb '' referred to a s - type planet , while `` ab b '' refer to a p - type planet as marked in the column on the planetary motion . + * discovery * + gives the year of the first discovery .
+ * spectral type * + this shows the spectral types of the stars .
unfortunately the data for some systems is incomplete . + * distance [ parsec ] * + distance from the sun to the system in units of parsecs ( 1 parsec = 3.26 light - years ) . +
* mass ratio ( @xmath17 * ) + given as dimensionless proportion @xmath4 , where @xmath13 is the mass of the first star and @xmath14 is the secondary star s mass . + * a@xmath54 [ au ] * + represents the distance between the double stars given in astronomical units .
if the semi - major is not given , a minimum of a@xmath54 will be approximated trigonometrically by the published separation angle @xmath55 given in [ arcsec ] and the distance from the sun to system @xmath56 given in [ parsec ] .
@xmath57 + * eccentricity ( e@xmath58 * ) + represents the eccentricity of the second star .
this parameter is very rarely known . + * number of planets * + systems with one planet are dominant , but multiplanet systems become more and more frequent . + * planet motion s - type , p - type * + as shown in the introduction see also fig .
[ fig1 ] . + * mass : @xmath13 [ @xmath18 ] and @xmath14 [ @xmath18 ] * + mass of the first and the second star given in units of the masses of our sun .
+ here we present the data of the planets , where the first two columns are similar to the star data ( see fig . [ planet ] ) .
+ * mass m x sin i * + the portion of a distant planet s mass that is detectable is determined by its line of sight , when observed from earth . if the angle of inclination from the `` face - on '' position is `` i '' , then the component which is in line with the earth is given by sin(i ) . + * semi - major axis [ au ] * + represents the semi - major axis of the planet s orbit given in astronomical units .
if the semi - major is not given , it will be derived from the published orbital period and from the mass of the host star through the kepler law .
@xmath59 { \frac { g \cdot m _ { * } \cdot p^2 } { 4 \cdot \pi^{2}}}$ ] + for s - type : @xmath60 is the mass @xmath13 of star 1 or @xmath14 of star 2 depending on the planet s orbit . for p - type : @xmath60 is the sum of the masses @xmath13 and @xmath14 .
these approximations are strongly influenced by the star s masses .
+ * orbital period [ d]*. + represents the orbital period of the planet given in days .
+ * eccentricity * + represents the eccentricity of the planet . +
* argument of perihelion [ deg ] * + represents the angle from the body s ascending node to its periapsis , measured in the direction of motion . +
* radius [ @xmath61 * + represents the planet s radius given in units of one jupiter radius . +
* inclination * + this value does not always represent the orbital inclination of the planet , especially for transiting planets it shows only the inclination relative to the line of sight . + * detection method * + shows the different detection methods which were used for the observations .
+ beside binary star systems also multiple star systems may harbour exoplanets .
the different possibilities for triple star systems are shown in fig.[triple ] , whereas quadruple star systems are presented in fig.[quad ] . in this list
the first three columns are similar to the list of the binary star systems .
+ * a@xmath62 or a@xmath63 [ au ] * + distance in au of the third star from the inner binary , or of the two binaries from each other . + * a@xmath64 [ au ] * + separation of the inner binary in case of a triple star system . + * a@xmath65 [ au ] * + separation of the other binary in case of a quadruple star system .
+ * number of planets * + systems with one planet are dominant , but multiplanet systems become more and more frequent . + * number of stars * + total number of detected stars in a multiple star system .
+ * mass of @xmath13 - @xmath66 [ @xmath18 ] * + mass of the @xmath67 , @xmath68 , @xmath69 and @xmath70 star given in units of the mass of our sun .
the list is completely identical with the planet data list of binary star systems .
in this article we used the statistics of the binary catalogue of exoplanets , which we introduce in the appendix .
we prepared statistics of exoplanets in well detached binary systems . in the second part of the article we enlarged the statistics by the investigation of well detached binary star systems from several catalogues and discussed the possibility of further candidates .
finally we investigated the possibility to detect exoplanets in well detached binary systems with eclipse timing variations . in the statistics of the binary catalogue of exoplanets
we could show that the separation which we used for our calculations are not only of theoretical interest ( fig .
[ fig2 ] , lower graph ) .
this also applies for the mass ratios which we used in the models 1,2 and 3 ( see section [ models ] and fig .
[ fig2 ] , upper graph ) .
we enlarged our investigation with further studies of well detached binary star systems from several catalogues and discussed the possibility of further candidates .
these investigations resulted in 462 candidates have star separations not larger than 3 au .
this is the separation which we investigated in the etv study . in this paper
we studied the circumstances favourable to detect s- and p - type planets in well detached binary - star - systems using eclipse timing variations ( etvs ) . to determine the probability of the detection of such variations with ground based telescopes and space telescopes , we investigated the dynamics of well detached binary star systems with a star separation in the range of @xmath71au .
we performed numerical simulations by using the full three - body problem as dynamical model .
the stability and the etvs are investigated by computing etv maps for different masses of the secondary star ( model 1 - 3 ) , separations ( @xmath72 and @xmath73 au ) and eccentricities ( @xmath74=0 , 0.2 and 0.4 ) .
in addition we changed the planet s mass ( earth , neptune and jupiter size ) eccentricities ( @xmath39=0 - 0.5 ) and semi - major axis ( depending on the configuration s- or p - type ) . for our investigations we used as detection criterion the photometric precision of corot @xmath31 sec as well as that of kepler @xmath32 sec , which we think is a realistic limit
in general the etv amplitude @xmath75 depends mainly on the eccentricity , the semi - major axis and the mass of the planet .
the stars separation and eccentricity ( @xmath74 ) mainly restricts the stable region of the planets .
we conclude that many amplitudes of etvs are large enough to detect exoplanets with neptune and jupiter - sizes in s - type and p - type configurations . whereas for the s - type configuration also
earth - size planets provide detectable etv signals .
we can conclude that possible terrestrial - like planets are detectable in binary star systems by the help of eclipse timing variations with restrictions .
however , future space missions will have a better precision which will enlarge the number of detectable etv signals .
r. schwarz , b. funk and .
bazs want to acknowledge the support by the austrian fwf project p23810-n16 .
akeson , r.l . , koerner , d.w . ,
jensen , e.l.n .
, 1998 , `` a circumstellar dust disk around t tau n : sub - arcsecond millimeter imaging '' , american astronomical society , 193rd aas meeting , # 73.08 ; bulletin of the american astronomical society , 30 , 1362 le sidaner p. , dubernet m .-
l . , simon g. et al .
, sf2a-2007 : proceedings of the annual meeting of the french society of astronomy and astrophysics held in grenoble , france , july 2 - 6 , 2007 , eds . : j. bouvier , a. chalabaev , and c. charbonnel , p.71 .
orosz , j.a . ,
welsh , w.f . ,
carter , j.a . ,
fabrycky , d.c . ,
cochran , w.d . ,
endl , m. , ford , e.b . , haghighipour , n. , macqueen , p.j . , mazeh , t. , sanchis - ojeda , r. , short , d.r .
, torres , g. , agol , e. , buchhave , l.a . , doyle , l.r . ,
isaacson , h. , lissauer , j.j . ,
marcy , g.w . , shporer , a. , windmiller , g. , barclay , t. , boss , a.p . ,
clarke , b.d . ,
fortney , j. , geary , j.c .
, holman , m.j . ,
huber , d. , jenkins , j.m .
, kinemuchi , k. , kruse , e. , ragozzine , d. , sasselov , d. , still , m. , tenenbaum , p. , uddin , k. , winn , j.n . ,
koch , d.g .
, borucki , w.j . , 2012 ,
science , 337 , 1511 | this paper is devoted to study the circumstances favourable to detect circumstellar and circumbinary planets in well detached binary - star - systems using eclipse timing variations ( etvs ) .
we investigated the dynamics of well detached binary star systems with a star separation from 0.5 to 3 au , to determine the probability of the detection of such variations with ground based telescopes and space telescopes ( like former missions corot and kepler and future space missions plato , tess and cheops ) .
for the chosen star separations both dynamical configurations ( circumstellar and circumbinary ) may be observable .
we performed numerical simulations by using the full three - body problem as dynamical model .
the dynamical stability and the etvs are investigated by computing etv maps for different masses of the secondary star and the exoplanet ( earth , neptune and jupiter size ) .
in addition we changed the planet s and binary s eccentricities .
we conclude that many amplitudes of etvs are large enough to detect exoplanets in binary star systems . as an application , we prepared statistics of the catalogue of exoplanets in binary star systems which we introduce in this article and compared the statistics with our parameter - space which we used for our calculations .
in addition to these statistics of the catalogue we enlarged them by the investigation of well detached binary star systems from several catalogues and discussed the possibility of further candidates .
[ firstpage ] methods : numerical catalogues planets and satellites : detection (stars ) : planetary systems ( stars):binaries : general stars : statistics |
rare b decays are induced by flavor changing neutral currents ( fcnc ) at loop level
. therefore they are phenomenologically rich and provide a comprehensive information about the theoretical models and the existing free parameters .
the forthcoming experiments at slac , kek b - factories , hera - b and possible future accelerators stimulate the study of such decays since the large number of events can take place and various branching ratios , cp - violating asymmetries , polarization effects , etc .
, can be measured @xcite . among b decay modes ,
inclusive @xmath0 is interesting since it is theoretically clean and sensitive to new physics beyond the sm , like two higgs doublet model ( 2hdm ) @xcite , minimal supersymmetric standard model ( mssm ) @xcite , etc .
there are various studies on this process in the literature .
the branching ratio ( @xmath4 ) of @xmath0 decay in the sm is @xmath5 for on - shell gluon @xcite .
this ratio can be enhanced with the addition of qcd corrections or by taking into account the extensions of the sm .
the enhanced @xmath6 is among the possible explanations for the semileptonic branching ratio @xmath7 and the average charm multiplicity .
the theoretical predictions of @xmath7 @xcite are slightly different than the experimental measurements obtained at the @xmath8 and @xmath9 resonance @xcite . further the measured charm multiplicity @xmath10 is smaller than the theoretical result .
the enhancement of @xmath11 and therefore @xmath6 rate would explain the missing charm and @xmath7 problem @xcite .
further , @xmath12 reported by cleo @xcite stimulates to study on the enhancement of @xmath6 . in @xcite ,
the enhancement of @xmath13 was obtained less than one order compared to the sm case in the framework of the 2hdm ( model i and ii ) for @xmath14 and @xmath15 .
the possibility of large @xmath4 in the supersymmetric models was studied in @xcite . in @xcite @xmath4
was calculated in the model iii and the prediction of the enhancement , at least one order larger compared to the sm one , makes it possible to describe the results coming from experiments @xcite . in the case of time - like gluon , namely @xmath16 decay
, @xmath4 should be consistent with the cleo data @xcite @xmath17 and in @xcite , it was showed that the model iii enhancement was not contradict with this data for light - like gluon case .
the calculation of @xmath13 with the addition of next to leading logarithmic ( nll ) qcd corrections was done in @xcite and it was observed that this ratio enhanced by more than a factor of 2 .
cp violating asymmetry ( @xmath18 ) is another physical parameter which can give strong clues for the physics beyond the sm .
the source of cp violating effects in the sm are complex cabbibo - cobayashi - maskawa ( ckm ) matrix elements .
@xmath18 for the inclusive @xmath0 decay vanishes in the sm and this forces one to go beyond the sm to check if a measurable @xmath18 is obtained . in this work ,
we study the decay width @xmath19 and @xmath18 of @xmath0 decay in the 3hdm and model iii version of 2hdm . in these models , it is possible to enhance @xmath19 and to get a measurable @xmath18 . since the yukawa couplings for new physics can be chosen complex and the addition of nll corrections @xcite brings additional complex quantities into the amplitude , theoretically , it is possible to get a considerable @xmath18 , at the order of the magnitude @xmath20 .
this effect is due to new physics beyond the sm , 3hdm and model iii in our case .
the paper is organized as follows : in section 2 , we give a brief summary of the model iii and @xmath21 and present the expressions appearing in the calculation of the decay width of the inclusive @xmath0 decay .
further we calculate the cp asymmetry @xmath18 of the process .
section 3 is devoted to discussion and our conclusions .
in this section , we study nll corrected @xmath0 decay width and the cp violating effects in the framework of the multi higgs doublet models ( model iii version of 2hdm and 3hdm ) in the sm and model i and ii 2hdm , the flavour changing neutral current at tree level is forbidden .
however , they are permitted in the general 2hdm , so called model iii with new parameters , i.e. yukawa couplings .
the yukawa interaction in this general case reads as @xmath22 where @xmath23 and @xmath24 denote chiral projections @xmath25 , @xmath26 , for @xmath27 , are the two scalar doublets , @xmath28 are quark doublets , @xmath29 and @xmath30 are quark singlets , @xmath31 and @xmath32 are the matrices of the yukawa couplings .
the flavor changing ( fc ) part of the interaction is given by @xmath33 the choice of @xmath34 and @xmath35 @xmath36\ , ; \phi_{2}=\frac{1}{\sqrt{2}}\left(\begin{array}{c c } \sqrt{2 } h^{+}\\ h_1+i h_2 \end{array}\right ) \,\
, , \label{choice}\end{aligned}\ ] ] and the vacuum expectation values , @xmath37 allows us to carry the information about new physics in the doublet @xmath35 .
further , we take @xmath38 , @xmath39 as the mass eigenstates @xmath40 , @xmath41 respectively .
note that , at tree level , there is no mixing among cp even neutral higgs particles , namely the sm one , @xmath42 , and beyond , @xmath40 . in eq.([lagrangianfc ] ) the couplings @xmath43 for the fc charged interactions are @xmath44 where @xmath45 is defined by the expression @xmath46 where @xmath45 is denoted as @xmath47 . here
the charged couplings are the linear combinations of neutral couplings multiplied by @xmath48 matrix elements ( see @xcite for details ) . in the case of the general 3hdm
, there is an additional higgs doublet , @xmath49 , and the yukawa interaction can be written as @xmath50 where @xmath51 is the new yukawa matrix having complex entries , in general .
the similar choice of higgs doublets @xmath52\ , , \nonumber \\ \\ \phi_{2}=\frac{1}{\sqrt{2}}\left(\begin{array}{c c } \sqrt{2 } h^{+}\\ h^1+i h^2 \end{array}\right ) \,\ , , \,\ , \phi_{3}=\frac{1}{\sqrt{2}}\left(\begin{array}{c c } \sqrt{2 } f^{+}\\ h^3+i h^4 \end{array}\right ) \,\
, , \nonumber \label{choice3h}\end{aligned}\ ] ] with the vacuum expectation values , @xmath53 can be done and the information about new physics is carried beyond the sm in the last two doublets , @xmath35 and @xmath49 .
further , we take @xmath38 , @xmath39 , @xmath54 and @xmath55 as the mass eigenstates @xmath40 , @xmath41 , @xmath56 , @xmath57 where @xmath56 , @xmath57 are new neutral higgs bosons due to the additional higgs doublet in the 3hdm ( see @xcite ) . the yukawa interaction for the flavor changing ( fc )
part is @xmath58 where the charged couplings @xmath59 and @xmath60 are @xmath61 and @xmath62 since there exist additional charged higgs particles , @xmath63 , and neutral higgs bosons @xmath64 , @xmath65 in the 3hdm , we introduce a new global @xmath66 symmetry in the higgs sector , considering three higgs scalars as orthogonal vectors in a new space , which we call higgs flavor space and we denote the higgs flavor index by `` @xmath67 '' , where @xmath68 .
the transformation reads @xmath69 where @xmath70 is the global parameter , which represents a rotation of the vectors @xmath35 and @xmath49 along the axis that @xmath34 lies , in the higgs flavor space .
this symmetry ensures that the new particles are mass degenerate with their counterparts existing in model iii ( see @xcite for details ) .
further the yukawa lagrangian ( eq.([lagrangian3h ] ) ) is invariant under this transformation if the yukawa matrices satisfy the expressions @xmath71 and we get @xmath72 therefore , it possible to parametrize the yukawa matrices @xmath73 and @xmath74 as : @xmath75 where @xmath76 are real matrices satisfy the equation @xmath77 and the angle @xmath2 is the source of cp violation .
here @xmath78 with @xmath79 and @xmath80 denotes transpose operation . in eq .
( [ yukpar ] ) , we take @xmath81 complex to carry all cp violating effects in the third higgs scalar .
now , we would like to continue the study of the inclusive process @xmath0 .
our starting point is the recent calculation of nll corrected decay width @xcite @xmath82 where @xmath83 with @xmath84 and @xmath85 is the result for the finite part of bremsstrahlung corrections @xmath86 where @xmath87\,m_b^2\,e_q\ , e_r \nonumber \,\ , , \\
\tau^+_{28}&= & \frac{16\,\,56}{3}\ , \hat{c}_2 \ , re[c_8^{0\,eff\,*}\ , \bar\delta i_{23}]\,m_b^2\,e_q\ , e_r \nonumber \,\ , , \\
\tau^-_{28}&= & -96\,\hat{c}_2 \
, re [ c^{0,eff\,*}_8 \,\bar\delta i_{17 } ] \,m_b^4\ , ( m_b\ , ( e_q + e_r ) - 2\ , ( e^2_q + e_r^2 + e_q\ , e_r)\nonumber \\ & + & 4\,\frac{e_q\ , e_r\,(e_q+e_r)}{m_b}]/ ( e_q\ , e_r ) \label{taudefn}\end{aligned}\ ] ] here @xmath88 and @xmath89 , and @xmath90 .
( see @xcite for details ) . in eqs .
( [ dwidthd2 ] ) and ( [ taudefn ] ) the wilson coefficients @xmath91 and @xmath92 ( eq . ( [ c12mu ] ) ) includes ll corrections and new physics effects enter into the expressions through the coefficients @xmath91 and @xmath93 ( see eq .
( [ parametr ] ) ) . the symbol @xmath94 is defined as @xmath95 and @xmath96 .
the vectors @xmath97 , appearing during qcd corrections , and the wilson coefficients @xmath98 , @xmath99 and @xmath100 , the functions @xmath101 and @xmath102 in eqs .
( [ taudefn ] ) , @xmath103 and the numbers @xmath104 in eq .
( [ dwidthd2 ] ) are given in @xcite .
now , we would like to start with the calculation of cp asymmetry for the inclusive decay underconsideration .
the possible sources of cp violation in the model iii ( 3hdm ) are the complex yukawa couplings .
our procedure is to neglect all yukawa couplings except @xmath105 and @xmath106 ( @xmath107 and @xmath108 ) ( see eqs .
( [ yukpar ] , [ yukpareq ] ) and discussion section ) in the model iii ( @xmath21 ) .
therefore , in the model iii ( @xmath21 ) , only the combination @xmath109 ( @xmath110 ) is responsible for @xmath18 . using the definition of @xmath18 @xmath111 we get @xmath112 \,\frac{\omega^{d}+\omega^{br } } { \lambda^{d}+\lambda^{br}}\ , \ , , \label{acp2}\ ] ] in the model iii where @xmath113 and @xmath114 are the contributions coming from d - part ( bremsstrahlung - part ) and they read as @xmath115 \nonumber \,\ , , \\
\omega^{br}&= & 2\ , c_{2 } \int de_{q } de_{r } ( b_{5}\,im [ \bar\delta_{23}]+ b_{6}\,im [ \bar\delta i_{17 } ] ) \nonumber \,\ , , \\
\lambda^{d}&= & 2\ , c_{1 } \
{ |a_{6}|^2 + |\bar\xi^{d}_{n , bb}|^2 \ , \lambda^{br}&= & 2\ , c_{2 } \int de_{q}\ , de_{r } \ { b_{4 } + re[\bar\xi^{d}_{n , bb}]\ , ( b_{5}\ , re[\bar\delta i_{23 } ] + b_{6}\ , re[\bar\delta i_{17 } ] ) \}\ , .
\label{omlam}\end{aligned}\ ] ] the functions @xmath116 and @xmath117 are defined as @xmath118 + r_{1 } ] + c_2^{0}(\mu)\ , [ l_{2 } + ln[\frac{m_{b}}{\mu } ] + r_{2 } ] ) \nonumber % \\ & + & a_{3}\ , [ ( l_{8 } + 8 + \beta_{0 } ) \,ln[\frac{m_{b}}{\mu } ] + r_{8 } - % \frac{16}{3 } ] \nonumber \,\ , , \\ a_{6 } & = & ( \eta^{14/23 } a_{1}+a_{3 } ) + \frac{\alpha_{s}(\mu)}{4 \pi}\ , [ a_4 + \chi a_1-\frac{16}{3 } \eta^{14/23}\ , a_1+a_3\nonumber \\ & + & ( \eta^{14/23}\ , a_1+a_3 ) \ , [ ( l_{8}+8+\beta_{0})\ , ln[\frac{m_{b}}{\mu } ] + r_{8}]+a_5 ] \nonumber \,\ , , \\ a_{7 } & = & \eta^{14/23}\ , a_{2}\ , \ { 1 + \frac{\alpha_{s}(\mu)}{4 \pi}\ , [ \eta^{-14/23 } \chi -\frac{16}{3 } + ( l_{8}+8+\beta_{0})\ , ln[\frac{m_{b}}{\mu } ] + r_{8 } ] \}\ , , \label{a567}\end{aligned}\ ] ] and @xmath119 + b_3\ , ( \eta^{14/23 } a_1 + a_3)\ , re[\bar\delta i_{17 } ] \nonumber \,\ , , \\
b_5&=&b_2\ , \eta^{14/23 } a_2 \nonumber \,\ , , \\
b_6&= & b_{3}\ , \eta^{14/23 } a_2 \,\ , , \label{b456}\end{aligned}\ ] ] @xmath120 appearing in eq .
( [ b456 ] ) read @xmath121 \nonumber \,\ , , \\
b_{2 } & = & 32\ , m_b^2\ , e_q\ , e_r\ , [ 8\ , \hat{c}_1 + \frac { 28}{3}\ , \hat{c}_2 ] \nonumber \,\ , , \\ b_{3}&= & \frac{\tau^-_{28}}{re[c_8^{0 , eff\,*}\ , \bar{\delta}_{i17 } ] } \,\ , .
\label{b23}\end{aligned}\ ] ] here we use the parametrizations @xmath122 with @xmath123 and the wilson coefficients @xmath124 with @xmath125 the ll corrected wilson coefficients @xmath126 and @xmath127 are @xmath128 and @xmath129 in eq .
( [ a567 ] ) the parameter @xmath130 is given by @xmath131 in our calculations we take only @xmath132 complex , @xmath133 , where @xmath2 is the cp violating parameter which is restricted by the experimental upper limit of the neutron electric dipole moment eq .
( [ neutrdip ] ) . for @xmath21 ,
it is necessary to make the following replacements : @xmath134 & \rightarrow & \bar{\epsilon}^{d}_{n , bb}\ , sin^2\,\theta \nonumber \,\ , , \\
re[\bar{\xi}^{d}_{n , bb } ] & \rightarrow & \bar{\epsilon}^{d}_{n , bb}\ , cos^2\,\theta \nonumber \,\ , , \\ \label{ksirhopar}\end{aligned}\ ] ]
the general 3hdm model contains large number of free parameters , such as masses of charged and neutral higgs bosons , complex yukawa matrices , @xmath135 , @xmath136 with quark family indices @xmath137 .
first , a new global @xmath66 symmetry is introduced in the higgs flavor space to connect the yukawa matrices in the second and third doublet and to keep the masses of new charged ( neutral ) higgs particles in the third doublet degenerate to the ones in the second doublet @xcite .
second , the yukawa couplings , which are entries of yukawa matrices , is restricted using the experimental measurements , namely , @xmath138 mixing , the @xmath139 parameter @xcite and the cleo measurement @xcite , @xmath140 the constraints for the fc couplings from @xmath138 processes and @xmath139 parameter for the model iii were investigated without qcd corrections @xcite and the following predictions are reached : @xmath141 where @xmath142 is up ( down ) quark and @xmath137 are the generation numbers and further @xmath143 in the analysis , the ansatz proposed by cheng and sher , @xmath144 is used .
respecting these constraints and using the measurement by the cleo @xcite collaboration we neglect all yukawa couplings except @xmath145 , @xmath146 in the model iii . in @xmath21 ,
the same restrictions are done by taking into account only the couplings @xmath147 and @xmath148 .
this section is devoted to the study of the cp parameter @xmath149 and the scale @xmath1 dependencies of the decay width @xmath19 and cp asymmetry of @xmath18 for the inclusive decay @xmath0 , in the framework of the model iii and @xmath21 . in our analysis , we restrict the parameters @xmath2 , @xmath145 and @xmath150 ( @xmath147 and @xmath151 ) in the model iii ( @xmath21 ) , using the constraint for @xmath152 , @xmath153 , coming from the cleo data eq .
( [ br2 ] ) ( see @xcite ) . here
@xmath154 is the effective magnetic dipole type wilson coefficient for @xmath155 vertex .
the above restriction allows us to define a constraint region for the parameter @xmath145 ( @xmath147 ) in terms of @xmath146 ( @xmath148 ) and @xmath2 in the the model iii , ( @xmath21 ) .
further , in our numerical calculations we respect the constraint for the angle @xmath2 , due to the experimental upper limit of neutron electric dipole moment , namely @xmath156 which places an upper bound on the couplings with the expression in the model iii ( @xmath21 ) : @xmath157 ( @xmath158 ) for @xmath159 gev @xcite .
[ gammasin ] ( [ gammasin3h ] ) is devoted to the @xmath161 dependence of @xmath19 for @xmath162 , @xmath163 ( @xmath164 ) and @xmath165 ( @xmath166 ) in the model iii ( @xmath167 ) . here
@xmath19 is restricted between solid ( dashed ) lines for @xmath168 ( @xmath169 ) . as shown in fig .
[ gammasin ] , the decay width @xmath19 can reach @xmath170 in the region @xmath171 for @xmath168 and the possible enhancement , a factor of @xmath172 compared to the sm one ( @xmath173 @xcite can be reached . for @xmath167 , the upper range for the decay width
@xmath19 is @xmath174 in the region @xmath171 for @xmath168 and this leads to an enhancement , a factor of @xmath175 compared to the sm one .
@xmath19 decreases with increasing @xmath149 for @xmath168 and it can get larger values compared to the @xmath169 case , in both models . the @xmath176 dependence of @xmath19 is weak for @xmath169 and for this case , it takes slightly smaller values in the @xmath21 compared to the ones in the model iii . in our numerical calculations , we observe that the contribution of bremsstrahlung corrections are almost one order smaller as a magnitude compared to the rest . further , the restriction regions for @xmath168 and @xmath169 become more seperated with increasing values of the scale @xmath1 and this behaviour is strong in the @xmath21 . the scale dependence of @xmath19 is weak for the values @xmath177 and almost no dependence is observed for the large values of @xmath1 scale for both models .
( see figs .
[ gammamu ] and [ gammamu3h ] ) . in fig .
[ acpsin ] and [ acpsin3h ] , we present the @xmath176 dependence of @xmath18 for @xmath162 , @xmath163 ( @xmath164 ) and @xmath178 in the model iii ( @xmath21 ) . here
@xmath18 is restricted in the region bounded by solid ( dashed ) lines for @xmath168 ( @xmath169 ) . as shown in figures
, @xmath179 reaches @xmath180 for @xmath181 and all possible values of @xmath18 are negative .
however , for @xmath169 , the allowed region becomes broader and @xmath18 can take both signs , even can vanish .
for this case , @xmath179 reaches almost @xmath182 as an upper limit in both models .
further @xmath18 is more sensitive to @xmath176 in the @xmath167 compared to the model iii .
[ acpmu ] and [ acpmu3h ] represent the scale @xmath1 dependence of @xmath18 for @xmath183 , @xmath184 and @xmath178 in both models underconsideration .
the scale dependence of @xmath18 is also weak for the values @xmath177 similar to that of @xmath19 . here
the increasing values of @xmath176 cause to increase the size of restriction region . at this stage
we give the numerical values of @xmath19 and @xmath18 for @xmath185 @xmath186 and @xmath162 in the range @xmath187 , for model iii @xmath188 : @xmath189 and @xmath190 now we would like to present our conclusions : * @xmath19 can reach @xmath191 in the model iii ( @xmath21 ) and this is an enhancement a factor of @xmath192 compared to the sm one . *
a measurable cp asymmetry @xmath18 exists with the addition of nll qcd corrections and choice of complex yukawa coupling @xmath193 ( @xmath194 ( see section 2 ) ) in the model iii ( @xmath21 ) .
@xmath179 can be obtained at the order of the magnitude of @xmath195 .
this physical parameter is coming from the new physics effects and it can give strong clues about the physics beyond the sm . 1 , babar collaboration , slac report , no . slac - r 504 , 1998 + f. takasaki , _ hep - ex/ _ 9912004 ; + p.ecrola , _ nucl . instrum
. meth . _ * a446 * ( 2000 ) 407 , _ hep - ex/ _ 9910067 .
j. ellis , _ hep - ph/ _ 9910404 ; y. f. zhou and y. l. wu , _ mod . phys .
* a15 * ( 2000 ) 185 ; + x. g. he , _ hep - ph/ _ 0001313 ; c. d. lu .
_ hep - ph/ _ 0001321 .
s. glashow and s. weinberg , _ phys .
_ * d15 * ( 1977 ) 1958 .
m. misiak , s.pokorski and j. rosiek , _ hep - ph/ _ 9703442 , to appear in `` heavy flavor ii '' , eds . a.
j. buras and m. lindner , world scientific publishing co. singapore .
s. a. abel , w. n. cottinggham and i. b. whittingham , _ phys . rev . _ * d58 * ( 1998 ) 073006 .
m. ciuchini , e. franco , g. martinelli , l. reina and l. silvestrini , _ phys .
* b344 * ( 1994 ) 137 ; c. s. gao , c. d. lu and z. m. qiu , _ z. phys . _
* c69 * ( 1995 ) 113 .
j. e. bagan , p. ball , v. braun and p. gosdzinsky , _ nucl .
phys . _ * b432 * ( 1994 ) 3 ; e. bagan , p. ball , v. m. braun and p. gosdzinsky , _ phys .
_ * b342 * ( 1995 ) 362 ; [ erratum : _ phys .
lett . _ * b374 * ( 1996 ) 363 ] ; e. bagan , p. ball , b. fiol and p. gosdzinsky , _ phys . lett . _
* b351 * ( 1995 ) 546 ; m. neubert and c. t. sachrajda , _ nucl .
* b483 * ( 1997 ) 339 , m. neubert , _ hep - ph/0001334 _ and references therein . h. yamamoto , _ hep - ph/ _ 9912308 and references therein . a. kagan , _ hep - ph/ _ 9806266 , @xmath196 international symposium on heavy flavor physics , santa barbara , california , july 1997 , heavy flavor phys .
215 - 243 .
t. e. browder et.al . , ( cleo collaboration ) , _ phys . rev
* d81 * ( 1998 ) 1786 .
chao - qiang geng , p. turcotte and w. s. hou , _ phys .
lett . _ * b339 * ( 1994 ) 317 .
a. l. kagan , _ phys .
_ * d51 * ( 1995 ) 6196 .
s. bertolini , f. borzumati and a. masiero , _ nucl .
* b294 * ( 1987 ) 321 ; m. ciuchini , + e. gabrielli and g. f. giudice , _ phys .
_ * b388 * ( 1996 ) 353 ; a. l. kagan and m. neubert , _ phys .
_ * d58 * ( 1998 ) 094012 .
zhenjun xiao , chong sheng li and kuang - ta chao , _ phys .
lett . _ * b473 * ( 2000 ) 148 .
collaboration , t. e. coan _ et .
_ , _ phys .
* 80 * ( 1998 ) 1150 .
c. greub and p. liniger , _ phys .
_ * b494 * ( 2000 ) 237 .
e. iltan , _ phys .
* d61 * ( 2000 ) 054001 .
d. atwood , l. reina and a. soni , _ phys .
_ * d55 * ( 1997 ) 3156 .
m. s. alam collaboration , to appear in ichep98 conference ( 1998 ) .
t. m.aliev , e. iltan
_ j. phys . _
* g25 * ( 1999 ) 989 .
d. bowser - chao , k .
cheung and w. y. keung , _ phys .
rev . _ * d59 * ( 1999 ) 11006 . | we study the decay width and cp asymmetry of the inclusive process @xmath0 ( g denotes gluon ) in the multi higgs doublet models with complex yukawa couplings , including next to leading qcd corrections .
we analyse the dependencies of the decay width and cp asymmetry on the scale @xmath1 and cp violating parameter @xmath2 .
we observe that there exist an enhancement in the decay width and cp asymmetry is at the order of @xmath3 . |
in this section , we will briefly describe the group cohomology theory . as we are focusing on 2d spt phases
, we will be interested in the third cohomology group .
for a group @xmath2 , let @xmath56 be a g - module , which is an abelian group ( with multiplication operation ) on which @xmath2 acts compatibly with the multiplication operation ( the abelian group structure ) on @xmath56 : @xmath102 for the cases studied in this paper , @xmath56 is simply the @xmath1 group and @xmath32 an @xmath1 phase . the multiplication operation @xmath103 is the usual multiplication of the @xmath1 phases . the group action is trivial : @xmath104 , @xmath105 , @xmath106 .
let @xmath107 be a function of @xmath108 group elements whose value is in the g - module @xmath56 .
in other words , @xmath109 .
let @xmath110=\{\om_n \}$ ] be the space of all such functions .
note that @xmath110 $ ] is an abelian group under the function multiplication @xmath111 .
we define a map @xmath112 from @xmath113 $ ] to @xmath114 $ ] : @xmath115 let @xmath116=\ { \om_n| \om_n = d_{n-1 } \om_{n-1}| \om_{n-1 } \in \cc^{n-1}[g , m ] \}\end{aligned}\ ] ] and @xmath117=\ { \om_{n}|d_n \om_n=1 , \om_{n } \in \cc^{n}[g , m ] \}\end{aligned}\ ] ] @xmath118 $ ] and @xmath119 $ ] are also abelian groups which satisfy @xmath118 \subset \cz^n[g , m]$ ] where @xmath120\equiv \ { 1\}$ ] . @xmath119 $ ] is the group of @xmath108-cocycles and @xmath118 $ ] is the group of @xmath108-coboundaries .
the @xmath108th cohomology group of @xmath2 is defined as @xmath121= \cz^n[g , m ] /\cb^n[g , m ] \end{aligned}\ ] ] in particular , when @xmath122 , from @xmath123 we see that @xmath124=\ { \om_3| \\ & \ \ \ \frac { \om_3(g_2,g_3,g_4 ) \om_3(g_1,g_2g_3,g_4)\om_3(g_1,g_2,g_3 ) } { \om_3(g_1g_2,g_3,g_4)\om_3(g_1,g_2,g_3g_4 ) } = 1 \ } .
\nonumber \end{aligned}\ ] ] and @xmath125=\ { \om_3| \om_3(g_1,g_2,g_3)=\frac { \om_2(g_2,g_3 ) \om_2(g_1,g_2g_3)}{\om_2(g_1g_2,g_3)\om_2(g_1,g_2 ) } \ } , \label{3coboundary}\end{aligned}\ ] ] which give us the third cohomology group @xmath3=\cz^3[g , u(1)]/\cb^3[g , u(1)]$ ] .
in the symmetry operators on the edge of bosonic spt phases were represented in the matrix product operator formalism from which their connection to group cohomology is revealed and the non - existence of gapped symmetric states was proved . in this section , we review the matrix product representation of the unitary symmetry operators and how the corresponding cocycle can be calculated from the tensors in the representation . a matrix product operator acting on a 1d system is given by,@xcite o=_\{i_k},\{i_k}tr(t^i_1,i_1t^i_2,i_2 ... t^i_n , i_n)|i_1i_2 ... i_ni_1i_2 ... i_n| where for fixed @xmath23 and @xmath126 , @xmath127 is a matrix with index @xmath128 and @xmath129 .
here we are interested in symmetry transformations , therefore we restrict @xmath130 to be a unitary operator @xmath131 . using matrix product representation ,
@xmath131 does not have to be an onsite symmetry .
@xmath131 is represented by a rank - four tensor @xmath132 on each site , where @xmath23 and @xmath126 are input and output physical indices and @xmath128 , @xmath129 are inner indices .
if @xmath133 s form a representation of group @xmath2 , then they satisfy @xmath134 . correspondingly , the tensors @xmath135 and @xmath136 should have a combined action equivalent to @xmath137 .
however , the tensor @xmath138 obtained by contracting the output physical index of @xmath136 with the input physical index of @xmath135 , see fig .
[ p12 ] , is usually more redundant than @xmath137 and can only be reduced to @xmath137 if certain projection @xmath139 is applied to the inner indices ( see fig .
[ p12 ] ) . and @xmath135 into @xmath137 . ]
@xmath139 is only defined up to an arbitrary phase factor @xmath140 .
if the projection operator on the right side @xmath139 is changed by the phase factor @xmath140 , the projection operator @xmath141 on the left side is changed by phase factor @xmath142 . therefore the total action of @xmath139 and @xmath141 on @xmath138 does not change and the reduction procedure illustrated in fig.[p12 ] still works . in the following discussion , we will assume that a particular choice of phase factors have been made for each @xmath139 .
nontrivial phase factors appear when we consider the combination of three symmetry tensors @xmath135 , @xmath136 and @xmath143 .
[ p123 ] . ,
@xmath136 and @xmath135 into @xmath144 .
only the right projection operators are shown .
their combined actions differ by a phase factor @xmath145 . ]
there are two different ways to reduce the tensors
. we can either first reduce the combination of @xmath135 , @xmath136 and then combine @xmath143 or first reduce the combination of @xmath136,@xmath143 and then combine @xmath135 .
the two different ways should be equivalent .
more specifically , they should be the same up to phase on the unique block of @xmath146 which contributes to matrix contraction along the chain .
denote the projection onto the unique block of @xmath146 as @xmath147 .
we find that l q_g_1,g_2,g_3(i_3p_g_1,g_2)p_g_1g_2,g_3= + ( g_1,g_2,g_3 ) q_g_1,g_2,g_3(p_g_2,g_3i_1)p_g_1,g_2g_3
> from this we see that the reduction procedure is associative up to a phase factor @xmath145 .
if we then consider the combination of four symmetry tensors in different orders , we can see that @xmath145 forms a 3-cocycle of group @xmath2 .
that is , @xmath145 satisfies = 1 the arbitrary phase factor of @xmath139 contributes a coboundary term to @xmath145 .
that is , if we change the phase factor of @xmath139 by @xmath148 , then @xmath145 is changed to ( g_1,g_2,g_3 ) = ( g_1,g_2,g_3 ) @xmath149 still satisfies the cocycle condition and belongs to the same cohomology class as @xmath145 .
in this section , we discuss the property of the symmetry operator @xmath68 given in eqn.([umn ] ) .
first we show that @xmath68 indeed generates a @xmath0 symmetry .
next from its matrix product unitary operator representation we find that the transformation among the tensors are indeed related to the @xmath56th element in the cohomology group @xmath49 $ ] .
the calculation of cohomology class goes as described in the previous section .
we repeat the definition of @xmath68 here u^(m)_n = _ i cp^(m)_i , i+1 _
i e^i2l_i / n [ umn_a ] where @xmath58 acts on two neighboring rotors and depends on @xmath56 as @xmath59 note that @xmath63 represents @xmath61 to be confined within @xmath62 $ ] . as @xmath69 rotates all the @xmath150 s by the same angle and @xmath75 only depends on the difference between neighboring @xmath54 s , the two parts in the symmetry operator commutes .
therefore @xmath151 as @xmath152 and @xmath153 , @xmath68 indeed generators a @xmath0 symmetry on the 1d rotor system .
the matrix product representation of @xmath68 is given by @xmath154 and the tensors representing @xmath155 , @xmath156 are given by @xmath157following the calculation described in the previous section , we find that the projection operation when combining @xmath158 and @xmath159 into @xmath160 is @xmath161 where @xmath162 means addition modulo @xmath48 .
when combining @xmath158 , @xmath159 and @xmath163 , the phase angle in combining @xmath164 with @xmath165 first and then combining @xmath162 with @xmath166 is @xmath167 the phase angle in combining @xmath165 with @xmath166 first and then combining @xmath164 with @xmath168 is @xmath169 therefore , the phase difference is @xmath170 we can check explicitly that @xmath171 satisfies the cocycle condition @xmath172 also we see that @xmath173 , @xmath57 , form a @xmath0 group generated by @xmath174
. therefore , the tensor @xmath175 corresponds to the @xmath56th element in the cohomology group @xmath49 $ ] .
similar calculation holds for the @xmath1 symmetry generated by @xmath99 , @xmath176 .
the cohomology class is labeled by @xmath101 .
the free boson theory given in eqn.([lboson ] ) can be fermionized and the low energy effective action of the symmetry discussed here can be reinterpreted in terms of a free dirac fermion . in particular , the fermionized theory has lagrangian density _
i=1,2 ^l_i(_t + _ x)^l_i + ^r_i(_t - _ x)^r_i [ lfermion ] where @xmath177 and @xmath178 are two real fermions , out of which a complex fermion can be defined @xmath179 .
note that in order to have a state to state correspondence between the boson and fermion theory , the fermion theory contains both the periodic and anti - periodic sectors .
since the @xmath8 symmetry in the nontrivial @xmath8 spt phase only act on , say , the right moving sector , one may naively guess that only @xmath180 change sign , while @xmath181 , @xmath182 , and @xmath183 do not change under the @xmath8 transformation : @xmath184 . in this case , the fermion mass term , such as @xmath185 , will be allowed by the @xmath8 symmetry .
such a mass term will reduce the @xmath89 edge state to a @xmath186 edge state without breaking the @xmath8 symmetry . in the following
, we will show that the @xmath8 symmetry is actually realized in a different way .
the @xmath89 edge state is stable if the @xmath8 symmetry is not broken .
so the @xmath89 edge state represents the minimal edge state for the @xmath8 ( as well as the @xmath0 and @xmath1 ) spt phases .
the situation is best illustrated with explicit jordan - wigner transformation of the @xmath15 model in eqn.([h_z2_xy ] ) .
consider a system of size @xmath187 , @xmath188 .
after the jordan wigner transformation @xmath189 the hamiltonian becomes @xmath190 where @xmath191 is the total fermion parity in the chain and @xmath192 is the boundary term which depends on @xmath193 .
therefore , the fermion theory contains two sectors , one with an even number of fermions and therefore anti - periodic boundary condition and one with an odd number of fermions and periodic boundary condition . without terms mixing the two sectors , we can solve the free fermion hamiltonian in each sector separately .
after fourier transform , the hamiltonian becomes h=_k ( ) _ k^_k where @xmath194 takes value @xmath34 , @xmath55 , ... , @xmath195 in the periodic sector and @xmath196 , @xmath197 , ...
@xmath198 in the anti - periodic sector .
the ground state in each sector has all the modes with energy @xmath199 filled .
note that with this filling the parity constraint in each sector is automatically satisfied .
the ground state energy in the periodic sector is higher than in the anti - periodic sector and the difference is inverse proportional to system size @xmath48 . the @xmath200 operator or the @xmath201 operator in the boson theory ( as shown in fig .
[ fig : xy_z2x ] ) corresponds to changing the boundary condition of the dirac fermion from periodic to anti - periodic .
such operators would totally gap out the edge states .
however , from eqn .
( [ umn ] ) and eqn .
( [ umn1 ] ) , we see that both operators carry nontrivial quantum number in all @xmath0 ( and @xmath1 ) spt phases , therefore it is forbidden by the symmetry .
the @xmath202 operator in the boson theory corresponds to the pair creation operator @xmath203 in the fermion theory .
its combination with the @xmath204 operator ( @xmath205 in the fermion theory ) would gap out the system , but due to the existence of the two sectors the ground state would be two fold degenerate . to see this more explicitly ,
consider the @xmath15 model again where the combination of @xmath206 and @xmath207 operators can be realized with an anisotropy term h^xy_(2,0 ) = _ i x_i-1x_i - z_i-1z_i under jordan wigner transformation , it is mapped to the p - wave pairing term @xmath208 again , we have period boundary condition for @xmath209 and anti - periodic boundary condition for @xmath210 .
after fourier transform , the hamiltonian at each pair of @xmath194 and @xmath211 is @xmath212 the bogoliubov modes changes smoothly with @xmath213 and the ground state parity remains invariant .
the ground state energy is @xmath214 and explicit calculation shows that the energy difference of the two sectors ( with @xmath215 int . and @xmath216 int .
@xmath217 ) becomes exponentially small with nonzero @xmath213 .
therefore , upon adding the @xmath206 and @xmath207 terms , the ground state becomes two fold degenerate .
such an operator does carry trivial quantum number in the nontrivial @xmath8 spt phase and renders the gapless edge unstable .
however , a two fold degeneracy would always be left over in the ground states , indicating a spontaneous @xmath8 symmetry breaking at the edge .
the @xmath218 operator in the boson theory corresponds to a scattering term between the left and right moving fermions @xmath219 .
its combination with the @xmath220 operator ( @xmath221 in the fermion theory ) would gap out the system .
unlike the @xmath222 operator , there is no degeneracy left in the ground state . in the @xmath15 model
, this corresponds to a staggered coupling constant h^xy_(0,1 ) = _ i(-1)^i ( x_i-1x_i + z_i-1z_i ) mapped to fermions , the hamiltonian at @xmath194 and @xmath223 becomes @xmath224 for each pair of @xmath194 and @xmath225 , there is one positive energy mode and one negative energy mode and we want to fill the negative energy mode with a fermion to obtain to ground state . for the anti - periodic sector ,
such a construction works since there is a @xmath226 = even number of negative energy modes , and the anti - periodic sector contains an even number of fermions .
however , for the periodic sector , such a construction fails since there is a @xmath226 = even number of negative energy modes , and the periodic sector must contain an odd number of fermions .
so we have to add an fermion to a positive energy mode ( or have a hole in a negative energy mode ) , to have an odd number of fermions .
therefore , the ground state in the periodic sector has a finite energy gap above the anti - periodic one and the ground state of the whole system is nondegenerate .
however , because this term carries nontrivial quantum number in any nontrivial @xmath0 ( and @xmath1 ) spt phases , it is forbidden by the symmetry . for the trivial @xmath8 spt phase ,
the @xmath227 operators are @xmath8 symmetric operators , and can be added to the edge effective hamiltonian .
the presence of the @xmath227 operators will gap the edge state and remove the ground state degeneracy . | symmetry protected topological ( spt ) states are short - range entangled states with symmetry , which have symmetry protected gapless edge states around a gapped bulk .
recently , we proposed a systematic construction of spt phases in interacting bosonic systems , however it is not very clear what is the form of the low energy excitations on the gapless edge . in this paper , we answer this question for two dimensional bosonic spt phases with @xmath0 and @xmath1 symmetry . we find that while the low energy modes of the gapless edges are non - chiral , symmetry acts on them in a chiral way , i.e. acts on the right movers and the left movers differently . this special realization of symmetry
protects the gaplessness of the otherwise unstable edge states by prohibiting a direct scattering between the left and right movers .
moreover , understanding of the low energy effective theory leads to experimental predictions about the spt phases .
in particular , we find that _ all _ the 2d @xmath1 spt phases have _ even _ integer quantized hall conductance .
_ introduction _ a recent study shows that gapped quantum states belong to two classes : short - range entangled and long - range entangled.@xcite the long - range entanglement ( i.e. the topological order@xcite ) in the bulk of states is manifested in the existence of gapless edge modes or degenerate edge sectors .
the short - range entangled states are trivial and all belong to the same phase if there is no symmetry . however , with symmetry , even short - range entangled states can belong to different phases .
those phases are called symmetry protected topological ( spt ) phases . the symmetric short - range entanglement ( i.e. the spt order )
is also manifested in the existence of gapless edge modes around a gapped bulk if the symmetry is not broken .
for example , two and three dimensional topological insulators@xcite have a gapped insulating bulk but host gapless fermion modes with special spin configurations @xcite on the edge under the protection of time reversal symmetry .
the experimental detection of such edge modes@xcite has attracted much attention and a lot of efforts have been put into the exploration of new spt phases .
recently , we presented a systematic construction of spt phases in bosonic systems@xcite , hence extending the understanding of spt phases from free fermion systems like topological insulators to systems with strong interactions .
we showed that there is a one - to - one correspondence between 2d bosonic spt phases with symmetry @xmath2 and elements in the third cohomology group @xmath3 $ ] .
moreover , we proved that@xcite due to the existence of the special effective non - onsite symmetries on the edge of the constructed spt phases which are related to the nontrivial elements in @xmath3 $ ] , the edge states must be gapless as long as symmetry is not broken .
however , it is not clear what is the form of the gapless edge states , especially the experimentally more relevant low energy part .
a low energy effective edge theory is desired because it could provide a simple understanding of why the gapless edge is stable in these spt phases .
for example , understanding of the low energy ` helical ' edge@xcite in 2d topological insulators enables us to see that some of the relevant gapping terms are prohibited due to time reversal symmetry .
moreover , low energy excitations are directly related to the response of the spt phases to various experimental probes , which has led to many proposals about detecting the exotic properties of topological insulators@xcite .
such an understanding is hence also important for the experimental realization of bosonic spt phases . in this paper
, we study the low energy effective edge theory of the 2d bosonic spt phases with @xmath0 and @xmath1 symmetry .
we find that the gapless states on the 1d edge is non - chiral , as it should be due of the lack of intrinsic topological order@xcite in the system .
the special feature of the edge states lies in the way symmetry is realized .
in particular , we find that symmetry is realized chirally at low energy , i.e. in an inequivalent way on the right and left movers . because of the existence of this chiral symmetry , the direct scattering between the left and right moving branches of the low energy excitations is prohibited which provides protection to the gapless edge .
we would like to mention that people have used @xmath4 chern - simons theory@xcite and @xmath5 non - linear @xmath6-model@xcite to construct the edge states of the @xmath1 spt phases .
however , it is not clear whether we have obtained the edge states for all of the @xmath1 spt phases using those field theory approaches .
the construction presented in this paper has the advantage of having a direction connection to the third cohomology group @xmath7 $ ] .
so we are sure that we have obtained the edge states for all of the @xmath1 spt phases .
we would also like to point out that the chiral symmetry leads to a chiral response of the system to externally coupled gauge field even though the edge state as a whole is non - chiral . in particular , we find that all of the @xmath1 spt phases have an even - integer quantized hall conductance .
show that , due to the short range entanglement in spt phases , the edge of the systems exists as a purely local 1d system with a special non - onsite symmetry related to group cohomology .
this enables us to study the edge physics in 1d without worrying about the 2d bulk .
we will start with an exact diagonalization of the edge hamiltonian in the @xmath8 spt phase constructed in .
insights from this model are then generalized to construct a 1d rotor model with different symmetries realizing the edge states of all @xmath0 and @xmath1 spt phases
. some useful formulas of the third group cohomology @xmath3 $ ] are reviewed in appendix a[gcoh ] .
_ edge state of @xmath8 spt phase _ in we presented an explicit construction of a nontrivial bosonic spt phase with @xmath8 symmetry .
the edge hilbert space is identified as a local 1d spin @xmath9 chain .
the spin chain satisfies a @xmath8 symmetry constraint given by u_2=_i x_i _
i cz_i , i+1 where @xmath10,@xmath11 and @xmath12 are the pauli matrices and @xmath13 acts on two spins as @xmath14 .
we showed in that this non - onsite symmetry operator is related to the nontrivial element in the third cohomology group of @xmath8 and hence the edge state must be gapless if symmetry is not broken . here
we study one possible form of the edge hamiltonian which satisfies this symmetry h_2=_i x_i+z_i-1x_iz_i+1 this hamiltonian is gapless because we can map this model to an @xmath15 model .
the mapping proceeds as follows : conjugate the hamiltonian with @xmath13 operators on spin @xmath16 and @xmath17 and then change between @xmath10 and @xmath12 basis on every @xmath18th spin .
the hamiltonian then becomes h_2=_i x_i-1x_i + z_i-1z_i [ h_z2_xy ] therefore , the low energy effective theory of this model is that of a compactified boson field @xmath19 with lagrangian density = [ lboson ] this is a simple gapless state with both left and right movers and can be easily gapped out with a mass term such as the magnetic field in the @xmath20 direction @xmath21 .
however , such a term is no longer allowed when the transformed @xmath8 symmetry operation is taken into account : u_2=_2i cx_2i,2i-1 _
2i z_2i-1x_2i _
2i cx_2i,2i+1 where @xmath22 acts on spin @xmath23 and @xmath24 as @xmath25 .
this symmetry constraint prevents any term from gapping the hamiltonian without breaking the symmetry . to see
more clearly how this symmetry protects the gaplessness of the system , we study how it acts on the low energies modes .
we perform an exact diagonalization of the @xmath15 hamiltonian eqn.([h_z2_xy ] ) for a system of @xmath26 spins and identify the free boson modes .
then we calculate the @xmath8 quantum number on these modes as shown in fig .
[ fig : xy_z2x ] . note that @xmath27 is not translational invariant and does not commute with the @xmath1 symmetry of the @xmath15 model @xmath28 , therefore the free boson modes are not exact eigenstates of the @xmath8 symmetry .
however , at low energy , the @xmath8 quantum number becomes exact as the system size gets larger and in fig .
[ fig : xy_z2x ] we plot the asymptotic @xmath8 quantum number of the low energy states .
model @xmath29(eqn.([h_z2_xy ] ) ) .
@xmath30-axis is lattice momentum @xmath31 , where @xmath32 is the lattice spacing .
@xmath33-axis is energy with ground state energy set to @xmath34 and first excited state energy normalized to @xmath35 .
@xmath36 represents positive @xmath8 quantum number and @xmath37 represents negative @xmath8 quantum number .
total boson number @xmath38 and winding number @xmath39 are labeled as @xmath40 for each primary field , represented by the shaded @xmath36 or @xmath37 .
states in the same conformal tower have the same @xmath38 and @xmath39 . ] from fig .
[ fig : xy_z2x ] , we can see that the @xmath8 quantum number of each state only depends on the quantum number of the zero modes .
the zero modes are described by two integers : the total boson number @xmath38 and the winding number @xmath39 , and the @xmath8 symmetry at low energy acts as @xmath41 . from fig .
[ fig : xy_z2x ] , we also see that the primary fields labeled by @xmath40 have the following left- and right - scaling dimensions : @xmath42 . for the trivial @xmath8 spt phase , the onsite @xmath8 transformation at low energy acts as @xmath43 , which is a non - chiral action . for the non - trivial @xmath8 spt phase
, we see that the non - onsite @xmath8 transformation at low energy acts as @xmath44 .
we call such an @xmath39-dependent @xmath27 a chiral symmetry operation . from the chiral symmetry operation
, we can have a simple ( although not general ) understanding of why some of the gap opening perturbations can not appear in this edge theory .
for example , the simplest mass term in the free boson theory @xmath45 contains a direct scattering term @xmath46 between the left and right movers which carries a nontrivial quantum number under this @xmath8 symmetry and is hence not allowed .
this result is consistent with that obtained by levin @xmath47 gu@xcite .
_ edge state of @xmath0 spt phase _
understanding of how symmetry acts chirally on the edge state of the @xmath8 spt phase suggests that similar situations might appear in other spt phases as well . in this section
we are going to show that it is indeed the case for @xmath0 bosonic spt phases . from the group cohomology construction , we know that there are @xmath48 @xmath0-spt phases which form a @xmath0 group among themselves .
we are going to construct 1d rotor models to realize the edge state in each spt phase which satisfies certain non - onsite symmetry related to the nontrivial elements in @xmath49 $ ] . from these models
we can see explicitly how the symmetry acts in a chiral way on the low energy states .
taking the limit of @xmath50 in @xmath0 will lead to the understanding of the edge states in @xmath1 spt phases which we discuss in the next section .
note that the choice of the local hilbert space on the edge , here a quantum rotor , is arbitrary and will not affect the universal physics of the spt phase as long as the effective symmetry belongs to the same cohomology class . consider a 1d chain of quantum rotors describe by @xmath51 $ ] with conjugate momentum @xmath52 .
the dynamics of the chain is given by hamiltonian h_r=_i ( l_i)^2 + v ( _ i-_i-1 ) [ hr ] when @xmath53 , the system is in the gapless superfluid phase . at low energy , @xmath54 varies smoothly along the chain .
the gapless low energy effective theory is again described by a compactified boson field @xmath19 with compactification radius @xmath55 .
the low energy excitations contain both left and right moving bosons .
the generator of the non - onsite @xmath0 symmetry related to the @xmath56th element ( @xmath57 ) of the cohomology group , hence the @xmath56th spt phase with @xmath0 symmetry , takes the following form in this rotor chain : u^(m)_n = _ i cp^(m)_i , i+1 _ i e^i2l_i / n [ umn ] where @xmath58 acts on two neighboring rotors and depends on @xmath56 as @xmath59 here we need to be careful with the phase factor @xmath60 because it is not a single - valued function
. we confine @xmath61 to be within @xmath62 $ ] and denote it as @xmath63 .
then @xmath64 becomes single valued but also discontinuous when @xmath65 .
the discontinuity will not be a problem for us in the following discussion .
note that it is important that @xmath66 , because otherwise the symmetry factors into onsite operations and becomes trivial .
we show in appendix c[umn_coh ] that @xmath67 indeed generates a @xmath0 symmetry .
moreover from its matrix product unitary operator representation we find that the transformation among the representing tensors are indeed related to the @xmath56th element in the cohomology group @xmath49 $ ] .
therefore , the 1d rotor model represents one possible realization of the edge states in the corresponding spt phases .
( the matrix product unitary operator formalism and its relation to group cohomology was studied in and we review the main results in appendix b[mpuo ] ) .
the symmetry operator @xmath68 has a complicated form but its physical meaning will become clear if we consider its action on the low energy states of the rotor model in eqn.([hr ] ) .
first the @xmath69 part rotates all rotors by the same angle @xmath70 , which can be equivalently written as @xmath71 with @xmath72 being the total angular momentum of the rotors . at low energy
, @xmath73 is the total boson number with integer values @xmath38 .
moreover , at low energy @xmath54 varies smoothly along the chain therefore @xmath74 and @xmath58 adds a phase factor to the differential change in @xmath54 along the chain
. multiplied along the whole chain @xmath75 is equal to @xmath76 where @xmath39 is the winding number of the boson field @xmath19 along the chain .
put together we find that the symmetry acts on the low energy modes as [ umn1 ] u^(m)_n ~e^i2(l+mm)/n if @xmath56 is zero , this symmetry comes from a trivial spt phase and @xmath77 depends only on @xmath38 which involves the left and right movers equally , as one can see from right- and left - scaling dimensions @xmath78 .
however , when @xmath56 is nonzero , this symmetry comes from a nontrivial spt phase and @xmath68 depends on @xmath79 which involves the left and right mover in an unequal way .
put it differently , the symmetry on the edge of nontrivial @xmath0 spt phases acts chirally .
in particular , when @xmath80 , the symmetry will act only on the right movers .
similar to the discussion in the @xmath8 case , we can see that the chiral symmetry protects the gaplessness of the edge by preventing direct scattering between the left and right branches .
one may notice that @xmath81(eqn.([hr ] ) ) does not exactly commute with the symmetry @xmath68 , but this will not be a problem for our discussions .
we note that the potential term @xmath82 does commute with @xmath68 .
the kinetic term @xmath83 commute with the part that rotates @xmath54 but not the phase factor @xmath64 .
however , at low energy , @xmath84 , therefore this term becomes irrelevant locally and commutation between the hamiltonian and the symmetry operator is restored . at high energy , in order for the hamiltonian to satisfy the symmetry , we can change the kinetic term to @xmath85 .
the high energy dynamics will be changed . however , because @xmath53 and we know that the modified hamiltonian does not break the @xmath1 symmetry of the rotor model and the system can not be gapped ( due to the nontrivial cohomology class related to the symmetry ) , the system remains in the superfluid phase .
the change in the kinetic term does not affect our discussion about low energy effective physics .
_ edge state of @xmath1 spt phase _ taking the limit of @xmath86
, we can generalize our understanding of @xmath0 spt phases to @xmath1 spt phases . as we show in this section , the chiral symmetry action on the low energy effective modes on the edge of the @xmath1 spt phases leads to a chiral response of the system to externally coupled @xmath1 gauge field , even though the low energy edge state is non - chiral .
we calculate explicitly the quantized hall conductance in these spt phases from the commutator of local density operators on the edge and find that they are quantized to even integer multiples of @xmath87 . in these spt phases , a nonzero @xmath1 hall conductance exists despite a zero thermal hall conductance .
> from group cohomology , we know that there are infinite 2d bosonic spt phases with @xmath1 symmetry which form the integer group @xmath88 among themselves .
generalizing the discussion in the previous section we find that the low energy effective theory can be a @xmath89 free boson theory and the @xmath1 symmetry acts on the low energy modes as @xmath90 , where @xmath91 , @xmath38 is the total boson number , @xmath39 is the winding number and @xmath92 labels the @xmath1 spt phase .
the local density operator of this @xmath1 charge is given by @xmath93 with @xmath94 being the conjugate momentum of the boson field @xmath19 , because the spatial integration of this density operator gives rise to the generator of the @xmath1 symmetry @xmath95 .
the commutator between local density operators is given by @xmath96 = -i\frac{2m}{2\pi}\delta'(x - x').\end{aligned}\ ] ] this term will give rise to a quantized hall conductance along the edge when the system is coupled to an external @xmath1 gauge field . compared to the commutator between local density operators of a single chiral fermion @xmath97 = -i\frac{1}{2\pi}\delta'(x - x'),\end{aligned}\ ] ] we see that the hall conductance is quantized to even integer @xmath98 multiples of @xmath87 . as a consistency check
we see that the quantized hall conductance is a universal feature of the edge states in the bosonic @xmath1 spt phases and does not depend on the particular form the @xmath1 symmetry is realized on the edge .
indeed , the @xmath1 symmetry can be realized as @xmath99 , with arbitrary @xmath100 . from the group cohomology calculation ( reviewed in appendix b[mpuo ] ) we find that it belongs to the cohomology class labeled by @xmath101 . from the calculation of the commutator between local density operators
, we see that the magnitude of the commutator is proportional also to @xmath101 .
therefore , the hall conductance depends only on the cohomology class hence the spt phase the system is in and not on the details of the dynamics in the system .
_ discussion _ in this paper , we have construct the gapless edge states for _ each _ of the bosonic @xmath0 or @xmath1 spt phases in 2d .
we show that those edge states are described by a @xmath89 non - chiral free boson theory where the symmetry acts chirally on the low energy modes .
the chiral realization of the symmetry not only prevents some simple mass terms from gapping out the system but also leads to a chiral response of the system to external gauge fields .
we demonstrate this by constructing explicit 1d lattice models constrained by a non - onsite symmetry related to _
each _ nontrivial cohomology classes .
our result indicates that the field theory approach based on the @xmath4 chern - simons theory@xcite and @xmath5 non - linear @xmath6-model@xcite indeed produce all of the @xmath1 spt phases .
we want to emphasize that although we have focused exclusively on the 1d edge , a 2d bulk having the 1d chain as its edge always exists and can be constructed by treating a 1d ring as a single site and then putting the sites together .
note that while the stability and chiral response of the edge in spt phases are very similar to that of the edge in quantum hall systems , the underlying reason is very different .
the quantum hall edge states are chiral in its own , which remains gapless without the protection of any symmetry and leads to a nonzero thermal hall conductance .
finally , we want to point out that the edge theory constructed in this paper is only one possible form of realization .
it is possible that other gapless theories can be realized on the edge of spt phases , for example with central charge not equal to @xmath55
. it would be interesting to understand in general what kind of gapless theories are possible and what their universal features are .
we would like to thank zheng - cheng gu and senthil todadri for discussions .
this work is supported by nsf dmr-1005541 and nsfc 11074140 .
26 natexlab#1#1bibnamefont # 1#1bibfnamefont # 1#1citenamefont # 1#1url # 1`#1`urlprefix[2]#2 [ 2][]#2 , , , * * , ( ) . ,
* * , ( ) . ,
* * , ( ) , . , * * , ( ) . , * * , ( ) , . , * * , ( ) , .
, , , * * , ( ) , . , , , * * , ( ) . , , , * * , ( ) . , *
* , ( ) , issn . , , , , , , , ,
* * , ( ) . , , , , , , , , , ,
, * * , ( ) , issn . , , , , , , , , , , ,
* * , ( ) . , , , ( ) , . , , , * * , ( )
. , * * , ( ) . , * * , ( ) . , , , * * , ( ) . , , , * * , ( ) . , , , * * , ( ) . , * * , ( ) . , ( ) , .
( ) , .
( ) , .
( ) , . , , , ,
* * , ( ) . |
oxygen is the third most common element in the universe .
it is copiously produced when massive stars explode as type ii supernova .
this distinguishes it from fe , which is also made in type ia sn , the accretion - induced explosions of white dwarfs .
the [ o / fe ] ratio therefore reflects the mix of stars that have contributed to the enrichment of a system .
it has been used to diagnose the source of metals in x - ray gas in galaxies @xcite and in damped ly@xmath3 systems @xcite .
because type ii sn begin to explode more quickly than type ia sn after stars are formed , the o / fe ratio after star formation begins is large at first , then declines as fe , but little o , is contributed by the type ia sne @xcite .
this fact has been exploited to argue that bulge formation lasted @xmath4 1 gyr @xcite and star formation for dwarf galaxies happened in bursts @xcite .
the fact that the oldest stars in our galaxy have supersolar [ o / fe ] ratios must be considered when measuring the ages of globular clusters @xcite .
in particular , the [ o / fe ] ratios in metal - poor stars in the milky way are important because they provide a look at the chemical evolution of the early galaxy .
we can use the o and fe abundances to derive yields from type ii sne , to adopt the correct isochrones for globular clusters , and to calculate the timescale for the formation of the halo .
the [ o / fe ] ratios in old milky way stars also provide a starting point for interpreting the abundances seen in high - redshift systems .
unfortunately , the lines available in late - type stars are not ideal abundance indicators .
the strength of the forbidden lines at 6300 and 6363 are gravity - dependent and are very weak in dwarfs and subgiants .
the triplet of permitted lines at 7771 - 7774 have excitation potentials of 9.14 ev and therefore are weak in cool giants .
for some evolutionary stages the permitted lines are also affected by nlte effects @xcite .
the oh lines in the ultraviolet and infrared regions of the spectrum are measurable in dwarfs and subgiants .
however , oh is a trace species in these stars , and is particularly sensitive to inhomogeneities in temperature @xcite .
many studies using these abundance indicators show disagreement in the [ o / fe ] vs. [ fe / h ] relationship for stars with [ fe / h ] @xmath5 ( see figure 1 for an incomplete , but demonstrative , summary ) . because [ o i ] lines are stronger in giants and o i lines in dwarfs , studies using different indicators also use data from different types of stars . in general
, the studies using permitted o i lines @xcite and the uv oh lines @xcite in dwarfs and subgiants find a steep linear increase in [ o / fe ] with decreasing [ fe / h ] .
@xcite combined o i and uv oh measurements and found a slope of @xmath6 .
in contrast , the [ o i ] lines in giants and subgiants give [ o / fe ] values that plateau at @xmath7 for [ fe / h]@xmath8 @xcite .
more recent analyses @xcite show instead a slight slope , but a difference of @xmath9 dex between the indicators at [ fe / h ] @xmath10 remains .
the o abundances measured from the infrared oh lines in dwarfs , subgiants , and giants produce similar values to the [ o i ] lines @xcite .
it is possible that the differences cited above are the result of intrinsic variations in the oxygen abundance between giants and dwarfs .
however , studies of small samples of dwarfs with @xmath11[fe / h ] @xmath12 ( @xcite , 7 stars ; @xcite , 2 stars ) showed that the [ o i ] line in these stars gave an oxygen abundance 0.4 - 0.7 dex lower than that derived from the permitted lines in the same stellar spectra .
thus the discrepancy between forbidden and permitted lines can not be ascribed alone to different intrinsic oxygen abundances in giants and dwarfs .
there have been many attempts to find another solution and to reconcile the results produced by the different sets of lines , either through finding the same slope and intercept in the [ o / fe ] vs. [ fe / h ] relation for different samples of stars or through finding the same o abundance using different lines in the same star .
oxygen abundances are sensitive to the adopted stellar parameters , so several studies have argued for improved methods for finding the parameters .
@xcite constructed new color - t@xmath2 scales that produced effective temperatures that were 150200 k hotter than those used by other investigators .
these higher temperatures decreased the derived o abundance from the permitted lines so that they gave the same [ o / fe ] ( @xmath9 dex ) at low metallicities seen in giants . @xcite also found that temperatures that were hotter by 150 k than their original temperature scale would erase the discrepancy in five turnoff dwarfs and subgiants with [ fe / h ] @xmath5 . recently , the gravities , rather than the temperatures , have come under scrutiny . @xcite
re - evaluated the [ o / fe ] values for metal - poor dwarfs from @xcite and @xcite , in light of nlte effects on fe i @xcite .
king adopted gravities from @xcite and @xcite which were based on fe i / fe ii ionization balance , but with nlte corrections included for fe i , and based the [ fe / h ] scale on fe ii instead of fe i. when this is done , the o i abundances show the same slight slope as the [ o i ] abundances , though they were still higher . for five unevolved stars with both [ o i ] and o i measurements ,
the o i - based abundances exceeded the [ o i ] by @xmath13 dex .
@xcite analyzed 40 stars ( 7 with [ fe / h ] @xmath14 ) with measured o i and [ o i ] lines , ranging from dwarfs to giants .
the o i abundances were corrected for nlte effects using the results of @xcite , and they observed no difference between the two indicators on average , with the exception of the cool giants .
the tendency of the permitted lines of giants to give higher abundances than the forbidden was attributed to deficiencies in the @xcite models that were used in the analysis .
@xcite obtained high - resolution , very high s / n ( @xmath15 400 ) data for 18 dwarfs and subgiants with @xmath16[fe / h ] @xmath17 .
their equivalent width measurements have errors of @xmath18 m for the forbidden and @xmath19 m for the permitted lines .
the quality of their data allowed the forbidden lines to be measured in higher - gravity metal - poor stars than before .
when they used 1-d model atmospheres and nlte corrections , the [ o i ] , o i triplet and uv oh lines gave the same [ o / fe ] vs. [ fe / h ] relation . however , consideration of 3-d effects , in particular granulation , only reduced the oxygen abundance derived from the [ o i ] lines , and a disagreement remained at the level of 0.3 dex .
@xcite compared their [ o / fe ] values in dwarfs with those in giants of the same metallicity .
while the [ o i ] lines gave satisfactory agreement , the o i triplet lines in giants gave higher abundances than those seen in dwarfs and subgiants .
one metal - poor subgiant , bd + 23 3130 , has been subjected to intense scrutiny by several authors .
@xcite found [ o / fe ] @xmath20 for this star using the uv oh lines and o i triplet .
@xcite argued that this was incompatible with the weakness of the [ o i ] line at 6300 , which yielded [ o / fe ] @xmath21 .
@xcite observed the 6300 line of this star at s / n @xmath22 900 and measured an equivalent width of [email protected] and found [ o / fe]@xmath24 , halfway between the @xcite and @xcite values .
@xcite revised the analysis using a log @xmath25 value 0.4 dex higher than their previous study . with this analysis , they found agreement among the uv lines , the [ o i ] line and the o i triplet .
@xcite used similar atmospheric parameters , but osmarcs models also achieved agreement between the [ o i ] and uv oh lines with 1-d atmospheres , but not 3-d atmospheres .
the studies of @xcite and @xcite suggest that a solution may be found in the application of correct stellar parameters and a consistent analysis of fe and o using those parameters .
while using different indicators for different samples of stars increases the number of possible targets , especially at low metallicity , we will be looking instead at a sample that have both [ o i ] and o i lines . using both sets of lines in the same star
is important because star - to - star variations exist for oxygen and other element - to - iron ratios in metal - poor stars @xcite .
the most glaring example of this is the subgiant bd + 80 245 , whose permitted o i lines give a sub - solar [ o / fe ] ratio at [ fe / h ] @xmath26 @xcite . also , such a tactic avoids the question of whether oxygen has been depleted by internal mixing in giants , which means that they can be included in the sample .
therefore , a more rigorous way to insure both the permitted and forbidden oxygen lines truly give the same results is to use both lines in the same stars .
the recent studies of @xcite and @xcite showed that agreement between the oxygen abundance given by o i and [ o i ] could be reached , at least for turnoff dwarfs and subgiants , for their particular choices of 1-d atmospheres .
however , because of the weakness of the [ o i ] line in dwarfs and subgiants , there are only six stars in these two papers that have both [ o i ] and o i measurements .
we chose to focus on subgiants and giants to obtain a large , homogeneous sample of stars with both sets of lines measured , including a number with [ fe / h ] @xmath27 .
this will also test whether the successes with the dwarfs and subgiants can be replicated , or whether , like @xcite , the analysis of cool giants will produce different oxygen abundances .
we have taken advantage of the very high resolution ( r @xmath28 ) gecko spectrograph on cfht to obtain equivalent widths for the [ o i ] lines for a sample of 55 stars , mostly subgiants and giants , with [ fe / h ] between @xmath292.7 and solar .
additional spectra and literature sources have been included so that all 55 stars have measurements of both the permitted and forbidden oxygen lines .
the data set presented here provides a strong test of any attempts to reconcile the indicators .
we will begin our analysis by measuring the magnitude of the difference in these two oxygen abundance indicators when we adopt atmosphere parameters with t@xmath2 from colors and log @xmath25 from isochrones .
we find that the familiar pattern of o i lines giving higher o abundances than the [ o i ] lines reasserts itself .
next , we examine whether changing the assumptions of the analysis , in particular the temperature scale , eliminates the measured difference .
we then use the knowledge of the behavior of the lines to create an ad hoc parameter set that , within the assumptions of the analysis , results in agreement between the indicators .
finally , we discuss whether this ad hoc parameter scale is realistic when compared to other observables for the target stars .
an exhaustive study of all the possible solutions to the oxygen abundance problem is beyond the scope of a single paper .
we therefore concentrate on following up the apparent successes of parameter - based solutions in the recent works mentioned in the introduction .
our analysis follows the following assumptions : \1 ) the atmospheres of stars can be described by one - dimensional , plane - parallel models in lte . for most of this work , we will use kurucz ( 1995 ) models .
we use the moog stellar abundance package @xcite for the analysis .
we adopt @xmath30 and @xmath31 .
the later value is based on the reanalysis of the solar [ o i ] by @xcite , which takes in account the contamination of the 6300 [ o i ] line by a ni i weak line .
\2 ) non - lte effects limit the usefulness of fe i lines @xcite in metal - poor stars .
non - lte conditions also affect the permitted o i lines , but for the purposes of this experiment we will assume the abundance corrections of @xcite adequately compensate for the departures from lte .
we also assume the lines of fe ii are free of non - lte effects and will be used as the primary fe abundance indicator .
\3 ) within the assumptions above , we will assume the solution to the problem can be found by the application of the correct atmospheric parameters for the stars .
this assumption is similar to the solution put forth by @xcite .
the methods employed here are similar to those taken by @xcite and @xcite , but their samples contain only warm ( t@xmath2@xmath32 k ) turn - off and subgiant stars , and only have 11 stars between them with both forbidden and permitted lines .
the gecko echelle spectrograph on the canada - france - hawaii telescope ( cfht ) delivers spectra with resolution of @xmath28 , with a dispersion of @xmath33 per 13.5 @xmath34 m pixel .
only one spectral order is observable at a time ( selected by a narrow - band filter ) , meaning only a @xmath35 region is observed per exposure with the thinned 2048x4096 pixel detector .
the gecko data were obtained over 6 nights in april and september 2001 . in both runs ,
the spectra covered the wavelength range from 62906370 , covering both the 6300 and 6363 lines .
the primary candidate list was created in a similar manner as the target list of @xcite , using literature lists of known metal - poor stars , such as @xcite , @xcite , etc .
the list was then culled of stars where we estimated that one of the two sets of lines would be undetectable .
the spectra were reduced using normal iraf routines .
although previous work has shown that the scattered light effect in gecko is less than 1% , special care was taken in its removal .
a wavelength solution was applied using thar lamps taken at the beginning and end of the night .
the variation in the wavelength of the telluric o@xmath36 features between spectra is less than 100 m s@xmath37 .
details of the individual observations are given in table 1 . because the 6300 feature lies within a band of telluric o@xmath36 lines , it was sometimes necessary to remove these telluric lines from the spectrum . during the observation runs , spectra of bright , rapidly - rotating ( vsini @xmath38 km s@xmath37 ) , spectral type b or a stars were observed .
these spectra were used to divide out the telluric o@xmath36 features ( some sample spectra are shown in the appendix ) . for most stars the division was cosmetic , because the [ o i ] line was not contaminated and the very high resolution and dispersion of the gecko spectrograph lessens the probability of contamination .
@xcite suggested circumstellar [ o i ] emission may play a role in the discrepancy .
we did not observe any sign of stellar [ o i ] emission in our spectra , a point which we examine this point further in the appendix .
additional data to measure the fe and o i lines were obtained with the lick 3-m and kpno 4-m with their respective echelle spectrographs .
the new data taken with the hamilton spectrograph at lick were obtained in the same way as the previous data ( see @xcite and @xcite for more details ) .
the echelle data from kpno were obtained in january 2002 with a resolution of @xmath39 and covers the wavelength range from 4480 to 7850 . included in the kpno data
is a spectrum of the asteroid vesta , which provides a solar spectrum taken as if the sun was a point source .
the equivalent widths ( ew ) values of the 6300 and 6363 [ o i ] lines were measured using both gaussian fits and integrations . the ew of the 6300 line for bd + 23 3130 was adopted from @xcite .
the ew values for the permitted o i lines were measured in the non - gecko data or taken from literature sources or a combination of both .
table 2 gives the oxygen ew values for the stars analyzed in this paper .
the ew of the 6300 [ o i ] line has been corrected for contamination by the 6300.34 ni i line ( see section 6.1 ) .
we adopt the @xcite @xmath40-values for the 6300 and 6363 forbidden lines ( @xmath41 and @xmath42 , respectively ) , and the @xcite @xmath40-values for the 7772 , 7774 , and 7775 permitted lines ( @xmath43 , @xmath44 , and @xmath45 , respectively ) .
the fe line list of @xcite is used here . the atomic data for the fe i lines are from @xcite and the oxford group ( * ? ? ?
* and references therein ) , while the fe ii lines have data from @xcite and @xcite .
slight modifications have been made to the @xmath40-values to improve consistency between sources , as described in detail by @xcite .
many of the target stars have been analyzed by the authors before @xcite .
we adopt the fe ew values from those papers .
the fe ew values for the previously - unpublished stars are given in table 3a and 3b ( available in the electronic version only ) .
the oxygen abundances found for the solar analysis are larger than the adopted solar oxygen abundance of @xcite by 0.14 ( forbidden ) and 0.10 ( permitted ) dex .
we could change the @xmath40-values of the lines to reflect these differences , but the value of 8.69 comes from a 3-d analysis , which we do not do here .
allende prieto et al .
report that using a one - dimensional model would increase the resulting solar oxygen abundance by 0.08 dex , in reasonable agreement with the solar abundance derived by the 1-d analysis conducted here .
therefore , we choose not to do a differential abundance analysis . for this paper ,
the ratio of the abundances given by the two oxygen indicators is more important than the absolute abundance . using the present @xmath40-values
, the solar analysis yields a forbidden line oxygen abundance 0.04 dex larger than that obtained permitted lines .
however , if we assume that all of the uncertainty is from line measurements error , we get an uncertainty for the ratio of 0.06 dex ( dominated by the uncertainty in the ew of the 6300 line ) . therefore , we believe that a change in the @xmath40-values is not warranted by the analysis . if the @xmath40-values were changed to force agreement , the needed @xmath46t@xmath2 vaules in section 7 would be increased by about 35 k.
initially , we will analyze the oxygen and iron abundances using stellar parameters derived from two photometric temperature scales : alonso et al . scale ( 1996 for dwarfs and 1999 for giants ) and
the alonso scales are based on the infrared flux method ( irfm ) of @xcite , while the houdashelt scale is based on synthetic spectra with zero points based on observations .
the input photometry for these relationships came from a variety of literature sources .
the b@xmath29v and v@xmath29i data are from the hipparcos / tycho catalog , and the ubyv@xmath47 photometry is from @xcite .
k colors were taken from the papers of alonso et al .
( 1994 and 1999 ) , @xcite , @xcite , and the 2mass point source catalog .
many v - r colors were taken from @xcite , while others come from @xcite and @xcite .
measurements of the reddening were taken from literature sources such as @xcite and @xcite .
other reddenings were derived using ubvy@xmath47 photometry and the calibration of @xcite , although the limits of that calibration exclude many evolved stars .
we adopt e(b@xmath29v ) @xmath48 e(b@xmath29y ) and the transformations of @xcite .
for the 8 stars for which we could not find or derive reddening estimates , we assume zero reddening .
the final dereddened colors are given in table 4 , while the calculated and adopted t@xmath2 values are given in tables 5 and 6 .
in all cases the t@xmath2 values were only accepted if the star s parameters were within the limits of a given color s calibration . because both alonso and houdashelt give different calibrations for giants and dwarfs , stars with derived log @xmath25 @xmath49
were considered dwarfs while the remaining stars were considered giants . while we initially intended to adopt the mean t@xmath2 value for the analysis ,
the t@xmath2 values derived from the ( v@xmath29i ) colors were consistently higher . additionally , the spread between the results for the different t@xmath2@xmath29color relations for individual stars was sometimes very large , most likely due to problems with the photometric data .
therefore we ignored most of the results from the ( v@xmath29i ) t@xmath2@xmath29color relation and other discrepant points when deciding which t@xmath2 value to adopt for each star .
the final values have been rounded to the nearest 25 k increment . for convenience
, we list the final log @xmath25 , [ m / h ] , and @xmath50 values for each star in tables 5 and 6 .
we have assigned a measure of the uncertainty in t@xmath2 to each star . in most cases , the value is the standard deviation of the t@xmath2 values used in the final calculation of the adopted value . while the agreement between the individual t@xmath2@xmath29color relationships for some stars is quite good ,
we believe that the uncertainty in the photometric data and the calibrations of the t@xmath2@xmath29color relationships place a lower limit of 75 k on the t@xmath2 uncertainty .
many traditional abundance analyses derive surface gravities from forcing agreement in the abundances derived from the fe i and fe ii .
* hereafter ti99 ) and @xcite both present evidence that the fe i lines in very metal - poor stars suffer from non - lte effects
. therefore lte analyses of these lines do not give reliable abundances .
we will derive surface gravities for our stars from the mass ( @xmath51 ) , absolute v magnitude ( m@xmath52 ) , a bolometric correction ( @xmath53 ) , and effective temperature ( t@xmath2 ) : @xmath54 we adopt @xmath55 k , @xmath56 and @xmath57 . the adopted stellar masses were based mostly on the star s assumed position on the appropriate - metallicity 12 gyr @xcite isochrones ( adopting a 10 or 14 gyr isochrone results in negligible differences ) .
however , several stars are likely to have evolved beyond the first - ascent giant branch and probably have undergone some form of mass loss . for these stars we adopt @xmath58 m@xmath59 ( see below ) .
bolometric corrections were calculated from alonso et al .
( @xcite for dwarfs and @xcite for giants ) . the adopted m@xmath52 magnitude , especially for giants , can be fairly uncertain . for stars for whose hipparcos parallax value
has @xmath60 , we adopt the hipparcos m@xmath52 value . many of the remaining stars , especially the giants , have poor hipparcos parallax determinations .
however , @xcite and @xcite derive m@xmath52 values for many giants and subgiants .
@xcite used hipparcos parallax data to improve the m@xmath52 values derived by @xcite , which themselves were based on fits to globular cluster color - magnitude diagrams .
@xcite derived distances using strmgren photometry and @xcite relationships between m@xmath52 , [ fe / h ] and color . for all the non - horizontal branch ( hb ) stars , we also derived estimates for the m@xmath52 value by using their dereddened colors to place them on the 12-gyr @xcite isochrone appropriate for their estimated [ fe / h ] value . for stars with
estimated [ fe / h ] values lower than the @xmath61 limit of the isochrone grid , the [ fe / h ] @xmath62 isochrone was used . a number of the target stars are hb or asymptotic giant branch ( agb ) stars , which affects their adopted m@xmath52 and mass values .
following @xcite , we identify potential post - rgb candidates in the distance - independent @xmath63 vs. @xmath64 plane ( see figure 2 ) .
the locus of subgiant and first - ascent giants is traced by ubyv isochrones kindly provided by clem & vandenberg ( private communication ) .
for the stars that fell into the hb or agb regions of the diagram , we adopt a mass of 0.6 m@xmath59 following @xcite .
the adopted m@xmath52 value for the assumed hb stars near the zahb locus follows m@xmath65[fe
/ h]@xmath66 @xcite . for the evolved hb and agb stars , the method of determining
the m@xmath52 value was no different than other stars , although a lower limit to the m@xmath52 value was placed by the appropiate m@xmath67 value . the m@xmath52 value for stars without high quality hipparcos parallaxes or stars not on the hb was based on a combination of the @xcite , @xcite and isochrone - derived values .
table 7 lists the m@xmath52 values from the various sources , as well as the final adopted m@xmath52 and stellar mass values .
the errors in m@xmath52 were calculated from the hipparcos parallax or estimated from the source papers . for the hb stars ,
an error of 0.2 mag was adopted to account for uncertainties in the m@xmath52@xmath29[fe / h ] calibration and any evolution above the horizontal branch .
a color absolute magnitude diagram for the final adopted values is shown in figure 3 . for reference ,
a 0.5 mag error in m@xmath52 contributes a 0.2 dex uncertainty in log @xmath25 .
an estimate of the [ fe / h ] value for each star was taken from the literature source of the star .
the adopted atmospheric [ m / h ] value to denote the adopted abundance scaling for the model atmosphere , while [ fe / h ] is the derived fe abundance .
while the values are usually similar , because we have adopted atmospheres with solar abundance ratios , [ m / h ] is an input value , while [ fe / h ] is an output value .
] was @xmath68@xmath69 dex higher because most metal - poor stars have enhancements in the so - called @xmath3-elements ( o , mg , si , ca , etc . ) , which provide more free electrons than are accounted for in the solar ratio models .
after the first abundance analysis iteration of fe lines , the [ m / h ] value was based on the [ fe / h ] value derived from the fe ii lines .
we use the @xmath50 value that gave a flat distribution of derived fe i abundances as a function of line strength .
errors in the adopted @xmath50 value have negligible effect on the derived oxygen abundances because most of the oxygen lines are weak .
we use the nlte corrections of @xcite .
the grid of corrections only includes stars warmer than 4500 k , but our sample includes stars several hundred degrees cooler than that limit . for stars outside the takeda et al .
grid , we have calculated corrections using an extrapolation from the nearest grid points .
@xcite and @xcite also derived nlte corrections for oxygen lines .
a comparison of the gratton et al . and
takeda et al .
corrections for the 7772 o i line is shown in figure 4 .
we calculated the comparisions by assuming the same o i line strength using the listed [ o / fe]@xmath70 value for each metallicity .
the only large difference between the two calculations is for the hot , low - surface gravity stars . adopting the gratton et al .
correction for the hb stars in our sample would result in a reduction of the nlte correction by less than 0.15 dex .
the 6300.31 [ o i ] line is blended with a ni i line at 6300.34 .
while the ni line is fairly weak , it does affect the derived oxygen abundance in the sun @xcite . as a correction
, we have subtracted the estimated strength of the ni i line ( calculated using the adopted alonso t@xmath2 scale models and assuming [ ni / fe ] = 0 ) from the measured 6300 ew .
in general the correction was only a small fraction ( @xmath71 ) of the adopted ew value , only being significant in metal - rich stars like the sun .
the ew values for the 6300 line given in table 2 reflect the corrected values .
the weakness of [ o i ] and o i lines in metal - poor stars means that ew errors can dominate the error budget and need to be considered carefully .
@xcite gives a useful derivation of the error in ew .
we write the ew as @xmath72 where @xmath73 is the dispersion in / pix , @xmath74 is the value of the continuum , and @xmath75 is the intensity at pixel @xmath76 .
this is summed over the @xmath77 pixels which contain absorption in the line . in practice , we summed over 15 pixels in the gecko case and 7 or 8 pixels for the kpno or lick data , respectively .
the error in the ew , taking into account that the errors in @xmath74 are completely correlated is @xmath78 we used the s / n of the 6300 region as the measure of @xmath79 .
@xmath80 is also based on the s / n , but because we averaged @xmath22 50 pixels around the oxygen lines to locate the continuum , the error in continuum is @xmath81 .
we determined the s / n by actually measuring the s.d . in the spectra , rather than relying on the photon statistics , though in practice they were the same .
using the s / n ignores other sources of error , in particular scattered light , but as discussed in 3 , the gecko spectrograph set up minimizes the impact of scattered light .
we can check our calculations of the ew error in two ways .
first , for 38 stars , we measured all three o i permitted lines and then determined the expected error in the average abundance both by using the errors derived from equation 3 and by calculating the standard deviation in the mean for those three lines .
there was encouraging agreement , usually to within 0.02 to 0.03 dex .
second , we compared our ews to previous published measurements ( figure 5 ) .
for the 6300 line , we find an average offset of @xmath82 m with a rms scatter of 3.9 m .
the average offset for the 6363 is 0.0 m with an rms scatter of 2.2 m .
a number of @xcite ew values are higher than the gecko observations , while the one discordant 6363 value , from bd + 30 2611 ( = hip 73960 ) , has a 6300 ew that agrees with the @xcite value .
the ratio between our ew values for the 6300 and 6363 lines in this star do not follow the expected 3-to-1 ratio ( 37.7 m vs. 17.6 m ) , but the gecko spectrum does not show any indication of the source of the error .
the lines joining the ew values derived for the same star indicate many cases of large scatter ( up to 50% ) among studies even when our data are not considered . the uncertainty in our ew values from our statistical calculation ( generally on the order of 1 m or less ) can not explain the rms deviations seen in figure 5 .
the most likely cause for much of the scatter seen is the lower quality of the previous data .
the gecko data presented here are at higher resolution and dispersion than the previous data , have negligible scattered light , and have minimal o@xmath36 contamination problems .
to measure the effects of systematic parameter errors on the oxygen and iron abundances , we ran a series of models for each star : one with the t@xmath2 value raised by 200 k , one with the log @xmath25 value raised by 0.3 dex , one with the [ m / h ] value raised by 0.3 dex , and one with the @xmath50 value raised by 0.3 km s@xmath37 .
the abundances from each of these individual runs were compared against the results of the original run .
the overall effect of the parameter changes on the whole sample of stars is given in table 8 .
both the permitted and forbidden lines are affected by the t@xmath2 and log @xmath25 values .
the choice of t@xmath2 is the most important parameter affecting the [ o / fe ] ratios and the difference between the two oxygen indicators .
the [ m / h ] value is slightly significant , but it is unlikely that a 0.3 dex systematic error in the [ m / h ] value would occur in practice .
the effects of various parameter changes on the forbidden and permitted oxygen and fe ii abundances as a function of t@xmath2 , log @xmath25 , and [ fe / h ] are given in plots 68 for each star in the sample .
the most important feature in the plots is that a t@xmath2 change has opposite effects on the permitted and forbidden line abundances , while the other parameter changes affect the two indicators in similar ways .
the figures also show that a systematic error in the surface gravity affects the abundance indicators by the same amount , but a systematic t@xmath2 error affects giants more than dwarfs , and metal - poor stars more than metal - rich stars .
an error in [ m / h ] affects metal - rich stars more than metal - poor ones . fe
ii mostly behaves like [ o i ] when [ m / h ] is changes , but more like o i when t@xmath2 is altered .
these results indicate that great care must be taken when comparing the results from different evolutionary status .
systematic parameter problems affect some stars , such as metal - poor giants , more than others . while systematic effects may be the most important factors in resolving the the disagreement between the forbidden and permitted lines , it is important to know the random errors associated with the o abundances to determine the significance of discrepancies
. we consider random errors in ews , t@xmath2 , log @xmath25 , and [ m / h ] of the model .
the abundance error due to @xmath50 is less than 0.03 dex and will be not considered in our analysis .
we modify the formula from @xcite : @xmath83}}}\right)^2 \sigma^2_{{\rm{[m / h ] } } } + { } { } 2\biggl[\left({\partial log\epsilon\over\partial t}\right ) \left({\partial log\epsilon \over\partial { \rm{log}g}}\right)\sigma_{t{\rm{log}g } } + \nonumber\\ \left({\partial log\epsilon\over\partial { \rm{[m / h]}}}\right ) \left({\partial log\epsilon\over\partial { \rm{log}g}}\right ) \sigma_{{\rm{log}g}{\rm{[m / h ] } } } + { } \nonumber\\ { } \left({\partial log\epsilon\over\partial { \rm{[m / h]}}}\right)\left({\partial log\epsilon\over t}\right ) \sigma_{t { \rm{[m / h ] } } } \biggr],\end{aligned}\ ] ] where @xmath84 , for example , is defined as @xmath85 the partial derivatives were calculated in section 6.2.2 .
equation 1 shows that log @xmath25 is dependent on t@xmath2 .
the extent to which our uncertainties in t@xmath2 and log @xmath25 are correlated depends on the magnitude of the error in m@xmath52 . therefore , for each star in our sample , we did a monte carlo experiment where we allowed t@xmath2 and m@xmath52 to vary based on their errors , then calculated log @xmath25 using equation 1 , then placed into equation 5 to calculate @xmath86 .
other @xmath87 values were determined in the same manner .
the errors in ratios such as [ [ o i]/fe ] can be appreciably smaller than the addition in quadrature of [ o i ] error and fe ii error , because they have similar sensitivities to changes in atmospheric parameters .
we used equation a20 from @xcite , modified to include [ m / h ] errors , to calculate abundance ratio errors .
the error bars in figure 10 are calculated using this formula .
no errors were calculated for the sun .
the abundance results from the alonso and houdashelt parameter scales are shown in figures 9 and 10 . in figure 9 , neither scale results in both sets of oxygen lines giving the same abundance for all stars , although the warmer houdashelt scale does a better job .
the unweighted mean value of [ o@xmath88/o@xmath89 ( @xmath90 ) for the alonso scale is @xmath91 ( sdom ) , and @xmath92 ( sdom ) for the houdashelt scale . in figure 10 ,
[ o@xmath88/o@xmath89 is shown as a function of the stellar parameters .
the ratio [ o@xmath88/o@xmath89 is larger for the cooler , lower surface gravity giants than for the warmer , higher gravity subgiants and dwarfs .
both least - squares and spearman rank - order tests confirm that there are highly significant anti - correlations between [ o@xmath88/o@xmath89 and these two parameters .
these same tests do not support a correlation between [ o@xmath88/o@xmath89 and [ fe / h ] .
this is in contrast to previous studies ( see figure 1 ) in which the value of [ o@xmath88/o@xmath89 increases with decreasing [ fe / h ] . in these earlier studies ,
the forbidden oxygen abundances came from giants and the permitted abundances came from dwarfs .
our result suggests that the growth of [ o@xmath88/o@xmath89 with decreasing [ fe / h ] is at least partially due to comparing stars of different evolutionary status . in figure 10
, it appears that the [ o@xmath88/o@xmath89 distribution for the houdashelt scale is bimodal , with some stars clustered at [ o@xmath88/o@xmath89 @xmath93 and the majority around [ o@xmath88/o@xmath89 @xmath94 .
indeed , a kmm test @xcite finds that there is a 96% probability that two gaussians fit the distribution better than a single gaussian .
the best fit model would place 12 stars in a group with a mean of @xmath95 , and the remaining 43 in a group with a mean of @xmath96 . while this is only a two - sigma result , understanding why the 12 stars ( all with [ o@xmath88/o@xmath89 @xmath97 ) are outliers may yield clues to the origin of the overall problem .
unfortunately , a detailed investigation into the properties these 12 stars found nothing striking about these stars except that they all have log @xmath25 @xmath98 and [ fe / h ] @xmath14 .
these 12 stars are also among the stars with the highest [ o@xmath88/o@xmath89 values when the alonso t@xmath2 scale is applied , so the origin of the high [ o@xmath88/o@xmath89 value may be unrelated to the houdashelt scale .
checks for binarity , evolutionary status , systematic errors with the photometry , ew measurements , m@xmath52 and reddening determinations , etc . did not yield any noticeable pattern for the 12 stars , especially one that would lead to such a tight clustering of outliers .
it can be concluded here that both the alonso and houdashelt scales fail to totally resolve the discrepancy . while the warmer houdashelt scale comes closer than the alonso scale ,
there are still several giant stars that have large [ o@xmath88/o@xmath89 values .
two possible reasons for the failure are : first , there is missing input physics in the analysis and an additional correction to the abundance results is necessary , or , second , the physics of the analysis is adequate , but the input parameters for the models are incorrect . full exploration of the first option is beyond the scope of this paper , but one simple explanation is that the nlte corrections adopted here are simply wrong . to correct the differences seen in figure 10 , nlte corrections would have to be much larger for low - gravity stars .
the corrections of @xcite are nearly identical to those of @xcite ( see figure 4 ) for this type of star , so the choice of nlte correction does not to affect the results . in section 6.4 , we will look at the effect of changing our choice of stellar atmospheres . in section 7
, we will assume the second option is correct and calculate stellar parameters that reconcile the indicators .
as mentioned above , the analyses to this point have been done using kurucz atmospheres .
the marcs grid of stellar atmospheres @xcite are an independent calculation of one - dimensional , plane - parallel atmospheres . to test whether the adopted atmosphere grid makes a significant difference ,
we re - analyzed the measured ew values through atmospheres using the dereddened alonso temperature scale parameters .
the comparison between the results from kurucz and marcs models is shown in figure 11 .
the abundances derived from the permitted and forbidden oxygen and fe ii lines are all slightly larger for the kurucz models than for the marcs models .
these tendencies are enhanced at lower metallicities .
the marcs - derived oxygen abundances show a similar discrepancy between the permitted and forbidden lines on average ( @xmath99 for marcs compared to @xmath100 for the kurucz models ) .
therefore , the use of marcs models instead of kurucz models will not solve the problem . however , the marcs models show a lower discrepancy for metal - poor giant stars , while the kurucz model results show lower discrepancies for more metal - rich , less - evolved stars .
if we used the most favorable atmospheric model for a given star , the difference between the oxygen abundance indicators could be reduced by up to @xmath68 dex in some cases
. however , there is no justification for such a selective use of atmospheres .
in section 6.2.2 we analyzed the behavior of the derived oxygen abundances as a function of the various stellar parameters .
we can use that knowledge to derive an ad hoc parameter scale that forces the two oxygen indicators to agree .
we will then examine the resulting parameters for their validity . to derive the parameters , we assume that the changes in the abundances with respect to parameter changes are all linear that is , the first partial derivatives are all constant
. therefore we can use the values from section 6.2.2 in the calculation . if we define @xmath101 then @xmath102 } } \delta { \rm [ m / h ] } + \frac{\partial{x}}{\partial{v_t } } \delta v_t\end{aligned}\ ] ] from table 8
, it is clear that the difference in the oxygen abundance indicators is most sensitive to the t@xmath2 value .
therefore we will cast the above equation as a function of t@xmath2 and the partials derived in section 6.2.2 . because it was found that the variation in oxygen abundances due to changes in @xmath50 is small
, we will assume @xmath103 and drop that term .
we will also assume that the dependency of m@xmath52 and the bolometric correction on t@xmath2 is small and adopt @xmath104 ( derived from equation 1 ) .
the adopted atmospheric [ m / h ] value is just the [ fe ii / h ] value . that value changes like any other abundance , so it is itself described by equation 6 , but in this case @xmath105 } = \delta { \rm [ m / h]}$ ] . if that substitution is made , then it is possible to solve for @xmath106}$ ] as a function of @xmath46t@xmath2 .
when these substitutions are made into equation 6 , we have an equation that describes the change in the abundance of element x as only a function of @xmath46t@xmath2 .
therefore , we can solve for the value of t@xmath2 which will force an agreement between the abundances derived for the forbidden and permitted oxygen lines .
the value of [ o@xmath88/o@xmath89 was reduced to less than 0.01 dex in one to three iterations of the above procedure .
table 11 gives the total value of @xmath46t@xmath2 and the final stellar parameters , while table 12 gives the resulting abundances .
the mean value of @xmath46t@xmath2 derived from the alonso scale results is @xmath107 k ( s.d . ) ; figure 12 plots the @xmath46t@xmath2 value as a function of other parameters .
there is a trend of increasing @xmath46t@xmath2 with decreasing log @xmath25 , which is expected because the giants have the largest [ o@xmath88/o@xmath89 values . when the same method is applied using the houdashelt t@xmath2 scale results , the final t@xmath2 values are the same as the values calculated for the alonso scale results .
the final mean difference between the `` ad hoc '' scale and the houdashelt scale is @xmath108 k ( s.d . ) .
if the stars are split into the two groups based on the houdashelt results discussed in section 6.3 , the 12 stars with houdashelt - scale [ o@xmath88/o@xmath89 values of @xmath97 need their houdashelt t@xmath2 values increased , on mean , by @xmath109 k ( s.d . ) , while the remaining 43 stars require a mean change of @xmath110 k ( s.d . ) .
the resulting [ o / fe ] vs. [ fe / h ] plot is shown in figure 13 .
the calculation of the random errors given in tables 11 and 12 and shown in figure 12 and 13 required changes to the method used in section 6.2 .
we determined the random error in t@xmath2 by considering the uncertainties in the oxygen equivalent widths and in m@xmath52 .
however , when calculating the errors for the o@xmath111 and o@xmath112 abundances , we needed to add three terms to equation 4 that took into account the correlation between the error in the oxygen equivalent widths on one hand , and t@xmath2 , log @xmath25 and [ m / h ] on the other .
a comparison between our oxygen abundances for all three temperature scales and those of several earlier works is given in figure 14 .
with a few exceptions , the points lie on parallel tracks to the 45-degree line .
the effect of changes in the temperature , for example , can be seen by comparing the top , middle , and lower panels .
as the temperature increases , the forbidden line oxygen abundances shift to higher abundances , while the permitted line abundances shift to lower abundances .
@xcite proposed a temperature scale that would resolve the discrepancy between the forbidden and permitted lines .
his calibration of t@xmath2@xmath29color relationships is valid for stars with @xmath113 k , so there are only 5 stars within our sample that can be compared to the @xcite scale .
the differences , t@xmath114 t@xmath115 , are @xmath116 k , @xmath117 k , @xmath118 k , @xmath119 k , and @xmath120 k for a mean difference of @xmath121 k. king provides specific t@xmath2 values for three more stars in common with this sample . if all eight stars are considered , the ad hoc scale is @xmath122 k warmer , and there is only a weak significance to the correlation between the two scales . while the overlap in samples is small , the evidence suggests that the king and ad hoc temperature scales are not in general agreement , even though both scales agree that increased t@xmath2 values can resolve the oxygen problem .
the ad hoc temperature scale was picked to solve the oxygen problem , but we must examine whether the scale reasonable when compared to other observations or the predictions of stellar evolution . in figure 15 , we plot the log @xmath25 vs. t@xmath2 plane for all three parameter scales . also plotted are 10 and 12 gyr @xmath3-enhanced isochrones from @xcite for [ fe / h ] @xmath123 , @xmath124 , and @xmath61 .
this range spans the observed [ fe / h ] range for most of the target stars ; thus , most of the stars should lie between the isochrones .
the mean [ fe / h ] values for the alonso , houdashelt , and ad hoc scales are @xmath125 , @xmath126 and @xmath124 , respectively .
many of the stars in the warmer ad hoc scale lie outside the range defined by the isochrones .
this is not a metallicity effect , because the mean [ fe / h ] values of all three scales are similar .
agreement could be re - established by increasing the log @xmath25 values by @xmath127 dex , because this would have a small effect on [ o@xmath88/o@xmath89 ( see table 8) .
the largest change due to a gravity increase would be the @xmath128 dex increase in [ fe / h ] , but the net affect to [ o@xmath88/o@xmath89 , on average , would be less than 0.05 dex .
however , an increase in log @xmath25 of @xmath9 dex would imply that the adopted m@xmath129 values were too bright by @xmath130 magnitudes .
an error of this size would be noticeable in the comparison of isochrones to globular cluster sequences .
the strength and profiles of the balmer lines are dominated by the stark effect and , theoretically , are very good temperature indicators @xcite .
this indicator is insensitive to errors in the reddening and surface gravity , and is a reasonably independent source of t@xmath2 values .
recent works that include stars studied here include ( * ? ? ?
* hip 57939 ) ( * ? ? ?
* hip 57939 and hip 104659 ) , and ( * ? ? ?
* hip 30668 , hip 49371 , hip 98532 , and hip 104659 ) including all measurements , the mean value of t@xmath131 - t@xmath132 is @xmath133 k ( sdom ) , while the mean value of t@xmath131 - t@xmath134 is @xmath135 k ( sdom ) .
again , like the fe i nlte test above , the comparison stars are mostly dwarfs and subgiants , so a more extensive study of balmer line - based t@xmath2 values would be welcome .
if the reddening estimates assumed in section 5.1 were too low , the resulting t@xmath2 values would be too cool .
if the ad hoc temperature scale was the correct one for the stars , then the t@xmath2@xmath29color relations could be inverted to give the intrinsic colors of the star .
the reddening could then be determined by comparison with the observed colors . when this is done with the alonso calibrations , we find that the required mean increase in e(b@xmath29v ) needed to account for the temperature change ranges from 0.05 to 0.12 mag , depending on the color used ( greatest for b@xmath29v , smallest for v@xmath29k )
. the mean measured value of e(b@xmath29v ) for the sample is @xmath136 , so the additional reddening required overall is larger than the original value .
the star - to - star scatter in the values is large .
the star that requires the largest increase in reddening is bd + 30 2611 ( = hip 73960 ) , which was found to need 0.32 mag of additional reddening , but the measured e(b@xmath29v ) = 0.00 .
similarly , the closest sample star , hd 103095 ( = hip 57939 ) , also with measured e(b@xmath29v ) = 0.00 , would require 0.12 mag of additional reddening .
for the 8 sample stars with the most reliable hipparcos parallaxes ( @xmath137 , with a mean distance of 42 pc ) , the mean additional reddening is 0.07 @xmath138 mag , while the mean measured reddening was 0.01 mag ( five have measured e(b@xmath29v ) values of 0.00 ) .
the mean additional reddening necessary for the 19 giant stars with m@xmath52 @xmath139 ( mean distance of @xmath22 750 pc ) is 0.19 @xmath23 0.08 mag , while the mean measured e(b@xmath29v ) is 0.04 .
the overall increase of the reddening value , especially for nearby stars that should not be heavily reddened , strongly indicates that additional reddening is not the source of the temperature difference .
reddening for individual stars can be very uncertain , and may be the cause for some of the random scatter , but it is unlikely it is the cause of the systematic difference .
studies of giants within globular clusters , for which the distance and reddening can be better determined than for individual field stars , could be used to help settle this issue . for all the tests attempted here , the alonso scale produced a better match to the observations than the warmer ad hoc scale .
the houdashelt scale lies between the other two .
therefore , it is hard to justify a major change in stellar parameters just to improve the oxygen abundance situation . if the alonso parameters are the correct ones to adopt , then we have to accept that a 1-d , lte analysis with the presently available nlte corrections adopted here is not sufficient to analyze oxygen abundances .
recently , @xcite and @xcite found that same oxygen abundance was derived using either the permitted and forbidden lines in dwarfs and subgiants .
they used analyses similar to our first attempt to derive abundances , i.e , they calculated temperatures from photometry , log @xmath25 from equation 1 , etc .
our analysis of giants and subgiants shows that the two abundance indicators have not been reconciled for all stars . as indicated by figure 10 the greatest values for [ o@xmath88/o@xmath89 are for the low - gravity , low - temperature giants .
for both the alonso and houdashelt scales , we find [ o@xmath88/o@xmath89 @xmath140 for the parameter space explored by nissen et al . and israelian et al .
( 6000 @xmath23 100k , log @xmath25 @xmath22 4.0 dex and [ fe / h ] @xmath141 ) . however , there are some outstanding problems remaining even with the subdwarf and subgiant analyses .
kurucz and marcs models do not give the same answers .
nissen et al . calculated [ o / fe ] @xmath142 for hd 189558 (= hip 98532 ) using osmarcs models .
despite using similar atmospheric parameters and equivalent widths , we found [ o / fe ] @xmath143 when we used kurucz model atmospheres .
@xcite derived a smaller difference between the permitted o lines and forbidden lines in bd + 23 3130 ( hip 85855 ) than we derive here , mainly because of the different electron densities in the different sets of kurucz models used .
as mentioned in the introduction , @xcite ascribed their difficulties with cool giants in part to their use of the @xcite models .
@xcite showed that the use of 3-d model atmospheres could alter the oxygen and iron abundances in metal - poor dwarfs .
the correction to [ o@xmath144/fe ] for the metal - poor dwarf hd 140283 was @xmath145 dex , and it is no longer clear whether the permitted and forbidden lines would produce the same oxygen abundance . three - dimensional model atmospheres are not yet available for giants , and the one - dimensional models , especially for the coolest giants , do not result in the agreement seen in the higher gravity stars . we have discussed above why one possible solution , a higher temperature scale , is not a good one .
@xcite and @xcite suggest that [ o i ] is filled in by emission , and this option is discussed and eliminated in the appendix .
there are still several solutions that can solve this discrepancy .
giants have large convection zones , so granulation may play an even larger role than in the dwarfs .
giants have thin atmospheres that are penetrable by uv radiation , so nlte corrections for the permitted lines are important .
one - dimensional nlte calculations may not work if a three - dimensional model is needed to describe the true conditions in the atmosphere . until three - dimensional models become widely available , there are still tests that can be done using traditional methods . for example
, the gravities of giant stars have larger uncertainties than dwarfs due to the uncertain distance to the stars .
@xcite discusses whether , like dwarfs , the lte fe i / fe ii ionization balance can no longer be used to derive surface gravities in metal - poor giants . as seen in table 7 ,
hipparcos parallaxes are of little use to individual giants .
although the changing the surface gravity is not the solution to resolving the [ o@xmath88/o@xmath89 controversy , log @xmath25 is crucial for calculating the absolute o abundance .
thus , until more reliable data is available from gaia or sim , a study of permitted vs. forbidden lines in cluster stars with accurately known distances would be helpful .
the chemical homogeneity of most clusters also makes it possible to use the abundances other heavy elements to help constrain the parameters .
we have analyzed the forbidden and permitted oxygen lines in 55 stars , including dwarfs and giants and spanning [ fe / h ] values from solar to @xmath1 in an attempt to understand the discrepancy in these oxygen abundance indicators .
we first tried a standard analysis using the temperature scales of alonso and houdashelt .
these models produced @xmath4[op / of]@xmath15 values of @xmath7 and @xmath146 , respectively .
the discrepancy was largest for cool giants , but evolved stars of all types favor high [ o@xmath88/o@xmath89 values .
the [ o@xmath88/o@xmath89 ratio is most sensitive to temperature of all the atmospheric parameters , and it is the only one where the the effect of a change in the parameter is opposite for the two indicators . using our understanding of the effects of parameter changes on the abundances , we calculated a new parameter scale that would bring the two sets of oxygen lines into agreement .
these parameters , however , disagree with other temperature diagnostics , such as colors , the fits to the balmer lines , and the bolometric luminosities .
we conclude that either improved nlte corrections for the permitted lines or other phenomena , perhaps associated with convection and granulation , are needed to solve the oxygen problem .
jpf and jaj would like to thank the staffs at the canada - france - hawaii telescope , mauna kea observatories , lick observatory , and kitt peak for their invaluable assistance with the observations for this project .
we also like to thank poul nissen and garak israelian for their insightful correspondences on bd + 23 3130 and robert kraft , chris sneden , and james hesser for their comments on drafts of this paper .
finally , we gladly thank the anonymous referee for his valuable and insightful comments .
this research has made use of the simbad database , operated at cds , strasbourg , france .
this publication makes use of data products from the two micron all sky survey , which is a joint project of the university of massachusetts and the infrared processing and analysis center , funded by the national aeronautics and space administration and the national science foundation .
@xcite suggested that emission from circumstellar shells could fill in the [ o i ] lines in giant stars .
these shells are the result of mass loss on the giant branch .
the resulting lower ew values would then lead to the discrepancy in the oxygen abundance indicators .
the @xcite model proposes that a mass loss rate of a few @xmath147 m@xmath59 yr@xmath37 could create an h i region of about 32 au around the giant .
if the temperature of this region was about the same as the giant ( 4500 k in the model ) and the density is about @xmath148 @xmath149 , the amount of photons emitted by the 6300 [ o i ] line from the h i region would reduce the measured ew by 20 m .
@xcite admits that the required mass loss rate is a factor of about five too high than expected by theory , but the remaining assumptions are not wildly unreasonable .
we therefore examined the 6300.31 region of the 16 stars with m@xmath150 observed with gecko for signs of emission .
the stellar absorption lines of our sample are resolved at the spectral resolution of gecko .
for example , in the 16 giants examined here , the [ o i ] 6300.31 absorption lines have a mean fwhm of @xmath151 ( @xmath152 pixels ) .
the telluric [ o i ] emission lines in these same spectra have a mean fwhm of @xmath153 ( @xmath154 pixels ) .
the dominant line - broadening mechanism in the h i region is thermal doppler broadening , which for this case would be 0.045 , or less than the instrumental profile of gecko .
therefore , any emission from an h i region surrounding the giant should be a narrow feature .
regions of 1 ( @xmath155 pixels ) , centered at 6300.31 for these 16 giants are shown in figures 16 and 17 .
no binning or smoothing has been applied to the spectra . as can be seen , no significant emission is present . finally , if the 6300.31 [ o i ] line was producing significant emission , other emission lines may be present .
@xcite estimates that the chromospheric h@xmath156 emission ( which dominates over the h@xmath156 emission from the h i region ) from the model system would be several ngstroms in equivalent width .
therefore , we examined the h@xmath156 lines of the 16 giants in the lower resolution spectra used to measure the fe and permitted o i lines . of these giants ,
only six show any sign of having asymmetric h@xmath156 profiles ( hip 17639 is not among these six stars ) . of these six , only two , bd + 30
2611 (= hip 73960 ) and hd 165195 (= hip 88527 ) show any sign of h@xmath156 emission .
for these two giants , the 6300.31 [ o i ] profile is deep , symmetric , and free of obvious emission . we therefore conclude that emission from an h i region as described by @xcite does not affect the equivalent width of the [ o i ] lines to any significant amount .
rrrccccccccl sun & & & kpno & 2002 jan 5 & 1800 & 400 & same & & & 375 & + 434 & 20 & & gecko & 2001 oct 1 & 1800 & 145 & kpno & 2002 jan 6 & 1800 & 100 & 1 + 484 & 97 & -20 6718 & gecko & 2001 oct 1 & 1800 & 145 & kpno & 2002 jan 7 & 1800 & 100 & + 2413 & 2665 & + 56 70 & gecko & 2001 sep 29 & 900 & 250 & f00 & & & 150 & 1 + 2463 & 2796 & -17 70 & gecko & 2001 sep 29 & 900 & 250 & ham .
& 2000 aug 12 & 900 & 120 & + & & & & & & & kpno & 2002 jan 8 & 900 & 200 & + 3985 & 4906 & + 18 111 & gecko & 2001 sep 30 & 900 & 225 & ham .
& 2000 aug 12 & 1800 & 110 & + 4933 & 6268 & @xmath157 322 & gecko & 2001 sep 30 & 1800 & 250 & kpno & 2002 jan 8 & 600 & 200 & + 5445 & 6755 & + 60 170 & gecko & 2001 sep 29 & 900 & 230 & ham . & 2000 aug 8 & 900 & 110 & + & & & & & & & kpno & 2002 jan 5 & 600 & 200 & + 5458 & 6833 & + 53 236 & gecko & 2001 sep 29 & 300 & 190 & f00 & & & 60 & + & & & & & & & kpno & 2002 jan 5 & 300 & 260 & + 6710 & 8724 & + 16 149 & gecko & 2001 sep 29 & 900 & 190 & f00 & & & 70 & + 14086 & 18907 & & gecko & 2001 sep 29 & 600 & 320 & f00 & & & 90 & + 16214 & 21581 & @xmath158 552 & gecko & 2001 sep 29 & 900 & 190 & f00 & & & 190 & 1 + 17639 & 23798 & & gecko & 2001 sep 30 & 900 & 200 & kpno & 2002 jan 7 & 900 & 80 & 1 + 18235 & 24616 & & gecko & 2001 sep 29 & 600 & 200 & f00 & & & 100 & + 18995 & 25532 & + 22 626 & gecko & 2001 sep 29 & 900 & 175 & f00 & & & 120 & + 19378 & 26297 & @xmath159 791 & gecko & 2001 sep 29 & 600 & 175 & f00 & & & 100 & + 21648 & 29574 & @xmath160 942 & gecko & 2001 sep 29 & 900 & 170 & j02 & & & 75 & + 27654 & 39364 & @xmath161 1211 & gecko & 2001 sep 29 & 180 & 400 & f00 & & & 200 & + 29759 & & + 37 1458 & gecko & 2001 apr 7 & 2x1800 & 430 & f00 & & & 120 & 1,4,5 + & & & gecko & 2001 sep 30 & 1800 & & & & & & + 29992 & 44007 & @xmath162 1399 & gecko & 2001 apr 6 & 900 & 225 & f00 & & & 90 & 3,4 + & & & gecko & 2001 sep 29 & 900 & & & & & & + 30668 & 45282 & + 3 1247 & gecko & 2001 sep 29 & 900 & 250 & f00 & & & 90 & 1,4 + 38621 & 63791 & + 62 959 & gecko & 2001 apr 5 & 900 & 275 & j02 & & & 100 & + 43228 & 74462 & + 67 559 & gecko & 2001 apr 5 & 900 & 135 & ham . & 2001 may 5 & 1800 & 100 & 1 + 49371 & 87140 & + 55 1362 & gecko & 2001 apr 5 & 1800 & 180 & j02 & & & 100 & 4,5 + 57850 & 103036 & @xmath163 3155 & ham . & 2001 apr 7 & 1800 & 90 & same & & & 90 & + 57939 & 103095 & + 38 2285 & gecko & 2001 apr 5 & 1800 & 400 & f00 & & & 200 & 1,5 + 58514 & 233891 & + 52 1601 & gecko & 2001 apr 6 & 900 & 110 & ham .
& 2001 may 5 & 1800 & 70 & + 60719 & 108317 & + 6 2613 & gecko & 2001 apr 5 & 1800 & 310 & j02 & & & 70 & 4 + 62235 & 110885 & + 1 2749 & gecko & 2001 apr 6 & 900 & 120 & ham . & 2001 may 6 & 2100 & 70 & 1 + 62747 & 111721 & @xmath164 3709 & gecko & 2001 apr 5 & 900 & 160 & f00 & & & 140 & 1,4 + 64115 & 114095 & @xmath165 3742 & ham . & 1999 may 5 & 900 & 170 & same & & & 100 & + 65852 & & + 3 2782 & gecko & 2001 apr 6 & 1800 & 130 & ham .
& 2001 may 5 & 3600 & 120 & + 66246 & 118055 & @xmath166 3695 & gecko & 2001 apr 7 & 1800 & 100 & f00 & & & 130 & + 68594 & 122563 & + 10 2617 & gecko & 2001 apr 5 & 900 & 400 & j02 & & & 70 & 4 + 71087 & & + 18 2890 & gecko & 2001 apr 5 & 1800 & 120 & j02 & & & 70 & + 73960 & & + 30 2611 & gecko & 2001 apr 6 & 1800 & 135 & f00 & & & 100 & + 74491 & 135148 & + 12 2804 & ham . & 2001 may 5 & 2700 & 140 & same & & & 125 & + 85487 & & + 17 3248 & gecko & 2001 apr 5 & 1800 & 300 & j02 & & & 80 & + & & & gecko & 2001 sep 30 & 1800 & & & & & & + 85855 & & + 23 3130 & & & & & f00 & & & 220 & 2,4,6 + 88527 & 165195 & + 3 3579 & gecko & 2001 apr 7 & 900 & 170 & j02 & & & 140 & 3 + 88977 & 166161 & @xmath167 4566 & gecko & 2001 apr 7 & 900 & 200 & ham .
& 2001 may 5 & 900 & 100 & 4 + & & & gecko & 2001 sep 30 & 900 & & & & & & + 91182 & 171496 & & ham .
& 2000 aug 14 & 1800 & 200 & same & & & 100 & + 92167 & 175305 & + 74 792 & gecko & 2001 apr 6 & 900 & 250 & f00 & & & 110 & 1,3,4 + & & & gecko & 2001 sep 29 & 600 & & & & & & + 94931 & & + 41 3306 & ham . & 2000 aug 11 & 1800 & 170 & same & & & 140 & + 96248 & 184266 & @xmath159 5359 & gecko & 2001 apr 5 & 900 & 350 & ham .
& 2000 aug 12 & 600 & 90 & 3 + & & & gecko & 2001 sep 30 & 600 & & & & & & + 97023 & 186379 & + 24 3849 & ham . & 1998 sep 8 & 450 & 175 & same & & & 125 & + 97468 & 187111 & @xmath164 5540 & gecko & 2001 sep 30 & 1800 & 160 & f00 & & & 60 & 1 + 98532 & 189558 & & gecko & 2001 sep 30 & 900 & 250 & f00 & & & 75 & 4 + 104659 & 201891 & + 17 4519 & gecko & 2001 sep 29 & 600 & 400 & f00 & & & 120 & 1 + 106095 & 204543 & @xmath163 5460 & gecko & 2001 sep 29 & 1200 & 200 & ham . &
2000 aug 11 & 1200 & 110 & + 107337 & 206739 & @xmath164 6080 & gecko & 2001 sep 29 & 1200 & 150 & ham . & 2000 aug 11 & 1200 & 120 & + 109390 & 210295 & @xmath162 6222 & gecko & 2001 sep 30 & 1800 & 140 & f00 & & & 100 & + & & & gecko & 2001 oct 1 & 1800 & & & & & & + 112796 & 216143 & @xmath168 5873 & gecko & 2001 sep 29 & 600 & 180 & j02 & & & 175 & + 114502 & 218857 & @xmath169 6692 & gecko & 2001 sep 29 & 600 & 250 & j02 & & & 75 & + 115949 & 221170 & + 29 4940 & gecko & 2001 sep 29 & 600 & 190 & f00 & & & 100 & + rrrrrrrr sun & 4.5 & & 0.6 & 70.8 & 61.0 & 47.2 & 1.1 + 434 & 4.9 & & 1.3 & 44.9 & 46.4 & 47.1 & 4.0 + 484 & 10.1 & 4.2 & 1.3 & 26.4 & 17.3 & 13.7 & 3.1 + 2413 & 3.4 & & 0.8 & 7.7 & 6.5 & 4.4 & 2.0 + 2463 & 6.0 & 2.3 & 0.8 & 12.4 & & & 2.0 + 3985 & 5.9 & 2.7 & 0.9 & 40.7 & 32.6 & 20.9 & 2.7 + 4933 & 5.1 & & 0.8 & 7.2 & 6.3 & & 2.0 + 5445 & 5.6 & 1.7 & 0.8 & 17.3 & 14.3 & 9.3 & 1.9 + 5458 & 32.8 & 13.4 & 1.0 & 22.6 & 17.3 & 12.8 & 1.5 + 6710 & 17.4 & 6.3 & 1.0 & 18.3 & 10.4 & & 4.3 + 14086 & 9.1 & 3.4 & 0.6 & 34.0 & 32.6 & 22.5 & 3.3 + 16214 & 10.0 & 4.1 & 1.0 & 15.8 & 12.1 & 9.3 & 3.7 + 17639 & 22.5 & 8.3 & 1.0 & 8.9 & 7.9 & 3.9 & 1.5 + 18235 & 12.6 & 4.4 & 1.0 & 31.1 & 28.5 & 21.6 & 3.0 + 18995 & 13.3 & & 1.1 & 102.1 & 84.8 & 63.7 & 2.5 + 19378 & 26.8 & 10.6 & 1.1 & 8.7 & 7.2 & & 3.0 + 21648 & 45.0 & 15.0 & 1.1 & 11.7 & 9.6 & 8.3 & 4.0 + 27654 & 29.0 & 10.6 & 0.5 & 31.2 & 28.0 & 21.2 & 1.8 + 29759 & 1.7 & & 0.5 & 14.5 & 11.1 & 8.0 & 2.0 + 29992 & 13.5 & 4.0 & 0.9 & 19.8 & 18.3 & 11.4 & 2.2 + 30668 & 4.6 & 1.7 & 0.8 & 20.1 & 12.2 & 12.1 & 2.3 + 38621 & 12.9 & 4.4 & 0.7 & 16.0 & 11.8 & & 3.0 + 43228 & 23.0 & 8.0 & 1.4 & 18.9 & 14.3 & 10.3 & 3.0 + 49371 & 5.7 & 1.7 & 1.1 & 21.1 & 13.8 & 11.3 & 2.0 + 57850 & 47.0 & 23.6 & 2.1 & 22.3 & 16.3 & & 3.3 + 57939 & 1.3 & & 0.5 & 6.9 & 5.6 & 3.7 & 1.5 + 58514 & 23.7 & 8.2 & 1.8 & 40.6 & 27.2 & 24.7 & 4.3 + 60719 & 1.8 & & 0.6 & 7.2 & 8.9 & 4.0 & 2.2 + 62235 & 7.4 & & 1.6 & 85.2 & 66.9 & 58.3 & 4.3 + 62747 & 11.4 & & 1.2 & 27.6 & 24.3 & 16.3 & 2.1 + 64115 & 22.3 & 9.2 & 1.1 & 29.7 & 26.2 & 18.5 & 3.0 + 65852 & 14.3 & & 1.5 & 16.8 & & & 2.5 + 66246 & 37.3 & 12.4 & 1.9 & 9.6 & 6.7 & 5.6 & 2.3 + 68594 & 6.9 & & 0.5 & 4.4 & 2.3 & & 3.0 + 71087 & 8.8 & 3.6 & 1.6 & 22.6 & & & 4.3 + 73960 & 37.7 & 17.6 & 1.4 & 20.1 & & 11.0 & 3.0 + 74491 & 32.7 & 12.2 & 1.4 & 11.4 & 9.1 & & 2.4 + 85487 & 4.5 & 1.2 & 0.6 & 19.3 & 12.0 & & 3.7 + 85855 & 1.5 & & 0.5 & 8.9 & 6.0 & 3.4 & 1.4 + 88527 & 24.7 & 8.8 & 1.1 & 6.3 & 4.9 & 2.8 & 2.1 + 88977 & 19.2 & & 1.0 & 77.0 & 61.1 & 45.8 & 3.0 + 91182 & 22.0 & 5.5 & 1.0 & 48.0 & 44.4 & 25.5 & 3.0 + 92167 & 9.5 & 2.9 & 0.8 & 23.6 & 20.5 & 14.4 & 2.7 + 94931 & 4.1 & & 1.1 & 32.9 & 19.5 & 12.9 & 3.0 + 96248 & 4.4 & & 0.6 & & 83.6 & 60.2 & 3.3 + 97023 & 4.0 & & 1.1 & 76.7 & 69.2 & 55.8 & 2.1 + 97468 & 36.0 & 12.9 & 1.2 & 16.6 & 11.6 & & 5.0 + 98532 & 1.6 & & 0.8 & 39.7 & 32.8 & 26.4 & 4.0 + 104659 & 2.8 & 0.8 & 0.5 & 47.2 & 38.9 & 29.6 & 2.5 + 106095 & 23.9 & 7.0 & 1.0 & 21.4 & 16.2 & & 2.7 + 107337 & 18.0 & 5.7 & 1.3 & 20.7 & 17.9 & &
2.5 + 109390 & 21.1 & 5.7 & 1.4 & 31.5 & 23.0 & & 3.0 + 112796 & 13.7 & 4.2 & 1.1 & 9.0 & 7.5 & 5.5 & 1.7 + 114502 & 3.7 & & 0.8 & 13.7 & 10.1 & 6.7 & 4.0 + 115949 & 15.6 & 5.6 & 1.0 & 20.7 & 16.8 & 10.5 & 3.0 + llrrrrrrrrrrrrrr fei & 4531.15 & & 90 & & 79 & & 85 & & & & fei & 4592.66 & & & & 65 & & 69 & & & & fei & 4595.36 & 38 & & 50 & 8.9 & & & 24 & & 22 & fei & 4602.01 & 71 & 32 & 68 & 19 & 68 & 19 & 41 & & 72 & 75 fei & 4602.94 & & & & 74 & & 89 & & & & llrrrrrrrrrrrrrr [ tab : table3b ] fei & 4531.15 & & & & & & & & & & fei & 4592.66 & & & & & & & & 62 & & fei & 4595.36 & & 24 & & & 36 & & 59 & & & fei & 4602.01 & 73 & 35 & 57 & & 51 & 89 & 72 & & 66 & fei & 4602.94 & & & & & & & & 75 & & rrrrrrrrrrr 434 & 9.04 & 0.01 & 0.68 & 0.427 & & & 0.73 & 0.94 & 1.79 & 0.473 + 484 & 9.66 & 0.00 & 0.79 & 0.513 & 0.68 & 0.47 & 0.82 & 1.05 & 2.16 & 0.349 + 2413 & 7.73 & 0.07 & 0.68 & 0.497 & 0.67 & 0.46 & 0.67 & 0.85 & 2.07 & 0.350 + 2463 & 8.49 & 0.03 & 0.68 & 0.520 & 0.69 & 0.48 & 0.72 & 0.93 & 2.14 & 0.490 + 3985 & 8.76 & 0.04 & 0.74 & 0.453 & & & 0.76 & 0.97 & & 0.291 + 4933 & 8.09 & 0.02 & 0.80 & 0.583 & & & 0.81 & 1.04 & & 0.514 + 5445 & 7.72 & 0.03 & 0.67 & 0.464 & 0.59 & 0.40 & 0.70 & 0.90 & 1.96 & 0.302 + 5458 & 6.75 & & 1.14 & 0.735 & 0.95 & 0.66 & 1.12 & 1.44 & 2.88 & 0.487 + 6710 & 8.31 & 0.04 & 0.92 & 0.656 & 0.86 & 0.59 & 0.90 & 1.15 & 2.57 & 0.441 + 14086 & 5.88 & 0.05 & 0.74 & 0.472 & & & 0.82 & 1.05 & & 0.304 + 16214 & 8.71 & 0.05 & 0.74 & 0.523 & 0.71 & 0.49 & 0.76 & 0.97 & 2.15 & 0.298 + 17639 & 8.29 & 0.00 & 1.03 & 0.741 & & & 1.00 & 1.29 & 2.80 & 0.640 + 18235 & 6.68 & 0.00 & 0.82 & 0.513 & & & 0.84 & 1.08 & & 0.318 + 18995 & 8.22 & 0.07 & 0.59 & 0.433 & 0.60 & 0.41 & 0.61 & 0.78 & 1.76 & 0.500 + 19378 & 7.74 & 0.00 & 1.09 & 0.737 & 0.93 & 0.64 & 1.04 & 1.34 & 3.16 & 0.609 + 21648 & 8.34 & 0.05 & 1.25 & 0.916 & 1.12 & 0.78 & 1.16 & 1.49 & 3.33 & 0.739 + 27654 & 3.76 & 0.03 & 0.95 & 0.588 & 0.84 & 0.58 & 1.00 & 1.29 & 2.36 & 0.436 + 29759 & 8.92 & 0.00 & 0.61 & 0.435 & 0.58 & 0.40 & 0.68 & 0.87 & 1.70 & 0.222 + 29992 & 8.05 & 0.09 & 0.74 & 0.488 & 0.68 & 0.47 & 0.71 & 0.90 & 2.15 & 0.351 + 30668 & 8.00 & 0.02 & 0.68 & 0.436 & 0.58 & 0.40 & 0.71 & 0.91 & 1.83 & 0.274 + 38621 & 7.89 & 0.05 & 0.83 & 0.575 & 0.76 & 0.52 & 0.85 & 1.09 & & 0.415 + 43228 & 8.71 & 0.02 & 0.95 & 0.640 & 0.83 & 0.58 & 0.99 & 1.27 & & 0.407 + 49371 & 8.97 & & 0.72 & 0.479 & 0.64 & 0.44 & 0.77 & 0.99 & 2.02 & 0.279 + 57850 & 8.19 & & 1.27 & 0.900 & 1.03 & 0.72 & 1.26 & 1.62 & 3.13 & 0.740 + 57939 & 6.42 & 0.00 & 0.75 & 0.483 & 0.65 & 0.45 & 0.88 & 1.13 & 2.03 & 0.155 + 58514 & 8.80 & 0.00 & 0.80 & 0.557 & 0.72 & 0.50 & 0.86 & 1.10 & 2.21 & 0.488 + 60719 & 8.03 & 0.00 & 0.61 & 0.440 & 0.60 & 0.41 & 0.66 & 0.84 & 1.89 & 0.306 + 64115 & 8.35 & 0.01 & 0.93 & 0.590 & 0.73 & 0.50 & 0.99 & 1.28 & 2.39 & 0.457 + 62235 & 9.18 & 0.00 & 0.67 & 0.420 & 0.57 & 0.39 & 0.73 & 0.94 & & 0.492 + 62747 & 7.97 & 0.01 & 0.79 & 0.504 & & & 0.81 & 1.04 & 2.15 & 0.299 + 65852 & 9.70 & & 1.09 & 0.672 & 0.83 & 0.57 & 1.05 & 1.35 & & 0.538 + 66246 & 8.86 & 0.05 & 1.22 & 0.810 & 0.98 & 0.68 & 1.02 & 1.31 & & 0.640 + 68594 & 6.18 & 0.00 & 0.85 & 0.640 & 0.82 & 0.57 & 0.87 & 1.12 & 2.49 & 0.543 + 71087 & 9.84 & & 0.82 & 0.506 & 0.66 & 0.45 & 0.84 & 1.08 & 2.11 & 0.382 + 73960 & 9.13 & 0.00 & 1.25 & 0.810 & 0.96 & 0.67 & 1.21 & 1.56 & 3.00 & 0.551 + 74491 & 9.49 & & 1.39 & 0.896 & 0.99 & 0.69 & 1.36 & 1.75 & 3.14 & 0.472 + 85487 & 9.38 & & 0.67 & 0.492 & & & 0.76 & 0.97 & 2.08 & 0.451 + 85855 & 8.94 & 0.00 & 0.61 & 0.470 & & & 0.68 & 0.87 & 2.02 & 0.275 + 88527 & 7.31 & 0.13 & 1.10 & 0.823 & 0.96 & 0.67 & 1.25 & 1.61 & 2.92 & 0.704 + 88977 & 8.13 & 0.28 & 0.59 & 0.478 & 0.67 & 0.46 & 0.59 & 0.76 & 2.00 & 0.459 + 91182 & 8.49 & 0.26 & 0.82 & 0.572 & & & 0.62 & 0.80 & 2.21 & 0.402 + 92167 & 7.18 & 0.03 & 0.73 & 0.473 & 0.64 & 0.43 & 0.84 & 1.08 & & 0.286 + 94931 & 8.87 & 0.00 & 0.81 & 0.488 & & & 0.83 & 1.06 & 2.14 & 0.268 + 96248 & 7.59 & 0.04 & 0.51 & 0.395 & 0.56 & 0.38 & 0.55 & 0.70 & 1.62 & 0.605 + 97023 & 6.87 & 0.00 & 0.57 & 0.377 & & & 0.65 & 0.83 & & 0.350 + 97468 & 7.72 & 0.11 & 1.06 & 0.749 & 0.94 & 0.65 & 0.94 & 1.21 & 2.90 & 0.600 + 98532 & 7.72 & 0.01 & 0.56 & 0.379 & & & 0.62 & 0.80 & 1.57 & 0.283 + 104659 & 7.37 & 0.00 & 0.53 & 0.353 & & & 0.59 & 0.75 & 1.36 & 0.262 + 106095 & 8.29 & 0.03 & 0.86 & 0.613 & 0.79 & 0.55 & 0.83 & 1.07 & 2.42 & 0.552 + 107337 & 8.55 & 0.03 & 0.97 & 0.602 & 0.79 & 0.55 & 0.93 & 1.20 & 2.43 & 0.434 + 109390 & 9.55 & & 0.89 & 0.591 & & & 0.90 & 1.16 & 2.38 & 0.440 + 112796 & 7.82 & 0.02 & 0.93 & 0.673 & 0.86 & 0.59 & 0.90 & 1.15 & 2.60 &
0.564 + 114502 & 8.94 & 0.03 & 0.69 & 0.479 & 0.66 & 0.46 & 0.72 & 0.93 & 1.99 & 0.326 + 115949 & 7.69 & 0.06 & 0.97 & 0.683 & 0.88 & 0.61 & 0.90 & 1.16 & 2.66 & 0.555 + rccccccccrrrr 434 & @xmath170 & 5089 & 5334 & & 5426 & 5366 & 5304 & 5375 & 75 & @xmath171 & 2.41 & 1.90 + 484 & @xmath172 & 4922 & 4983 & 4907 & 5172 & 4914 & 4980 & 4950 & 75 & @xmath6 & 2.68 & 1.40 + 2413 & @xmath173 & 5042 & 4800 & 4769 & 5639 & 5024 & 5055 & 5025 & 75 & @xmath6 & 2.12 & 1.45 + 2463 & @xmath174 & 4996 & 4833 & 4806 & 5462 & 4944 & 5008 & 4950 & 90 & @xmath175 & 2.13 & 1.80 + 3985 & @xmath176 & 5143 & 5293 & & 5449 & & 5295 & 5225 & 106 & @xmath177 & 3.66 & 0.80 + 4933 & @xmath178 & 4791 & 4671 & & 5207 & & 4890 & 4750 & 155 & @xmath179 & 1.42 & 2.10 + 5445 & @xmath180 & 5109 & 5080 & 5145 & 5526 & 5150 & 5202 & 5175 & 75 & @xmath181 & 2.88 & 1.50 + 5458 & @xmath182 & 4385 & 4326 & 4270 & 4487 & 4282 & 4350 & 4300 & 81 & @xmath183 & 1.63 & 1.55 + 6710 & @xmath184 & 4638 & 4440 & 4378 & 4972 & 4512 & 4588 & 4500 & 112 & @xmath185 & 1.45 & 1.70 + 14086 & @xmath186 & 5144 & 5199 & & 5253 & & 5199 & 5175 & 75 & @xmath187 & 3.60 & 1.10 + 16214 & @xmath188 & 4889 & 4766 & 4710 & 5346 & 4929 & 4928 & 4900 & 103 & @xmath189 & 2.24 & 1.45 + 17639 & @xmath190 & 4477 & 4340 & & 4728 & 4332 & 4469 & 4375 & 82 & @xmath191 & 1.12 & 2.20 + 18235 & @xmath186 & 4962 & 5041 & & 5118 & & 5040 & 5050 & 75 & @xmath192 & 3.30 & 0.90 + 18995 & @xmath193 & 5415 & 5101 & 4980 & & 5401 & 5224 & 5400 & 75 & @xmath194 & 2.34 & 1.90 + 19378 & @xmath188 & 4413 & 4327 & 4299 & 4643 & 4099 & 4356 & 4350 & 138 & @xmath195 & 1.46 & 1.65 + 21648 & @xmath184 & 4207 & & & 4414 & 4005 & 4209 & 4200 & 119 & @xmath196 & 0.78 & 2.00 + 27654 & @xmath197 & 4682 & 4657 & 4456 & 4725 & 4702 & 4644 & 4675 & 75 & @xmath198 & 2.38 & 1.50 + 29759 & @xmath174 & 5234 & 5322 & 5265 & 5600 & 5511 & 5386 & 5400 & 145 & @xmath187 & 3.13 & 1.25 + 29992 & @xmath188 & 4878 & 4781 & 4700 & 5514 & 4930 & 4961 & 4975 & 103 & @xmath6 & 2.24 & 2.20 + 30668 & @xmath180 & 5072 & 5245 & 5180 & 5509 & 5308 & 5263 & 5300 & 161 & @xmath177 & 2.96 & 1.15 + 38621 & @xmath188 & 4774 & 4618 & 4585 & 5091 & & 4767 & 4725 & 101 & @xmath179 & 1.81 & 1.95 + 43228 & @xmath199 & 4612 & 4507 & 4461 & 4756 & & 4584 & 4600 & 113 & @xmath185 & 1.68 & 1.50 + 49371 & @xmath190 & 4910 & 5098 & 5045 & 5316 & 5076 & 5089 & 5050 & 84 & @xmath200 & 2.91 & 1.75 + 57850 & @xmath184 & 4188 & 4059 & 4099 & 4251 & 4115 & 4142 & 4150 & 75 & @xmath201 & 0.84 & 2.75 + 57939 & @xmath180 & 4965 & 5040 & 5090 & & 4987 & 5021 & 5025 & 75 & @xmath200 & 4.61 & 0.50 + 58514 & @xmath199 & 4841 & 4777 & 4786 & 5064 & 4862 & 4866 & 4825 & 75 & @xmath202 & 1.91 & 1.90 + 60719 & @xmath82 & 5237 & 5303 & 5179 & & 5257 & 5244 & 5275 & 75 & @xmath181 & 2.83 & 1.20 + 62235 & @xmath199 & 5092 & 5397 & 5276 & 5438 & & 5301 & 5300 & 84 & @xmath187 & 2.42 & 2.05 + 62747 & @xmath180 & 4868 & 4968 & & 5190 & 4925 & 4988 & 5000 & 135 & @xmath203 & 2.53 & 1.45 + 64115 & @xmath186 & 4764 & 4721 & 4762 & 4742 & 4685 & 4735 & 4750 & 75 & @xmath204 & 2.58 & 1.10 + 65852 & @xmath190 & 4403 & 4483 & 4502 & 4623 & & 4503 & 4500 & 79 & @xmath185 & 1.87 & 1.75 + 66246 & @xmath173 & 4250 & & & 4685 & & 4468 & 4225 & 131 & @xmath201 & 1.09 & 2.65 + 68594 & @xmath205 & 4715 & 4655 & 4552 & 5038 & 4594 & 4711 & 4650 & 75 & @xmath185 & 1.24 & 1.85 + 71087 & @xmath199 & 4810 & 4984 & 4990 & 5118 & 4969 & 4974 & 4975 & 86 & @xmath6 & 2.15 & 1.30 + 73960 & @xmath193 & 4223 & 4173 & 4237 & 4330 & 4198 & 4232 & 4225 & 75 & @xmath206 & 1.18 & 2.10 + 74491 & @xmath173 & 4061 & 4077 & 4178 & 4108 & 4107 & 4106 & 4100 & 75 & @xmath207 & 1.18 & 2.30 + 85487 & @xmath190 & 5045 & 5036 & & 5346 & 5009 & 5109 & 5025 & 75 & @xmath6 & 2.23 & 1.80 + 85855 & @xmath1 & 5224 & 5171 & & & 5102 & 5166 & 5175 & 75 & @xmath200 & 2.96 & 1.40 + 88527 & @xmath174 & 4393 & & & 4263 & 4243 & 4300 & 4300 & 75 & @xmath208 & 1.15 & 2.70 + 88977 & @xmath193 & 5384 & 4435 & 4376 & & 5102 & 4824 & 5100 & 280 & @xmath192 & 2.09 & 1.80 + 91182 & @xmath209 & 4991 & 4282 & & 5795 & 4870 & 4985 & 4900 & 86 & @xmath210 & 1.52 & 1.40 + 92167 & @xmath180 & 4928 & 5041 & 4977 & 5111 & & 5014 & 5050 & 75 & @xmath211 & 2.57 & 1.35 + 94931 & @xmath212 & 5048 & 5123 & & 5224 & 4871 & 5067 & 5075 & 79 & @xmath187 & 4.62 & 0.80 + 96248 & @xmath199 & 5656 & 5370 & 5195 & & 5605 & 5457 & 5525 & 153 & @xmath213 & 2.44 & 2.20 + 97023 & @xmath214 & 5784 & 5792 & & 5829 & & 5802 & 5800 & 75 & @xmath215 & 3.93 & 1.20 + 97468 & @xmath188 & 4449 & & & 4856 & 4260 & 4522 & 4300 & 134 & @xmath208 & 1.16 & 1.90 + 98532 & @xmath172 & 5612 & 5706 & & 5935 & 5631 & 5721 & 5650 & 75 & @xmath216 & 3.80 & 1.30 + 104659 & @xmath217 & 5766 & 5866 & & 6080 & 5974 & 5922 & 5875 & 104 & @xmath218 & 4.26 & 1.10 + 106095 & @xmath173 & 4721 & 4565 & 4538 & 5139 & 4646 & 4722 & 4650 & 83 & @xmath219 & 1.56 & 2.10 + 107337 & @xmath199 & 4579 & 4586 & 4546 & 4885 & 4632 & 4646 & 4600 & 75 & @xmath185 & 1.90 & 1.55 + 109390 & @xmath193 & 4730 & 4686 & & 4962 & 4677 & 4764 & 4700 & 75 & @xmath179 & 2.31 & 1.45 + 112796 & @xmath174 & 4606 & 4466 & 4408 & 4968 & 4482 & 4586 & 4500 & 83 & @xmath185 & 1.17 & 2.55 + 114502 & @xmath173 & 5008 & 4996 & 4893 & 5462 & 5116 & 5095 & 5050 & 91 & @xmath203 & 2.43 & 1.55 + 115949 & @xmath174 & 4560 & 4387 & 4295 & 4952 & 4434 & 4526 & 4475 & 110 & @xmath220 & 1.15 & 2.40 + rcccccccrrrr 434 & @xmath170 & 5590 & & 5506 & 5385 & 5494 & 5475 & 103 & @xmath194 & 2.45 & 1.90 + 484 & @xmath172 & 5327 & 5053 & 5252 & 4955 & 5147 & 5125 & 172 & @xmath181 & 2.74 & 1.40 + 2413 & @xmath173 & 5603 & 5087 & & 5057 & 5249 & 5075 & 75 & @xmath200 & 2.21 & 1.45 + 2463 & @xmath174 & 5588 & 5016 & 5543 & 4979 & 5282 & 5225 & 329 & @xmath192 & 2.27 & 1.80 + 3985 & @xmath176 & 5456 & & 5544 & & 5500 & 5500 & 75 & @xmath221 & 3.80 & 0.80 + 4933 & @xmath178 & 5294 & & 5286 & & 5290 & 5300 & 75 & @xmath181 & 1.63 & 2.10 + 5445 & @xmath180 & 5619 & 5403 & 5606 & 5180 & 5452 & 5300 & 158 & @xmath177 & 2.95 & 1.50 + 5458 & @xmath182 & 4573 & 4327 & 4538 & 4321 & 4440 & 4425 & 134 & @xmath222 & 1.64 & 1.55 + 6710 & @xmath184 & 5020 & 4541 & 5046 & 4566 & 4793 & 4750 & 277 & @xmath223 & 1.57 & 1.70 + 14086 & @xmath186 & 5447 & & 5350 & & 5399 & 5400 & 75 & @xmath213 & 3.73 & 1.10 + 16214 & @xmath188 & 5453 & 4965 & 5427 & 4970 & 5204 & 5150 & 280 & @xmath192 & 2.37 & 1.45 + 17639 & @xmath190 & 4780 & & 4792 & 4382 & 4651 & 4550 & 281 & @xmath224 & 1.17 & 2.20 + 18235 & @xmath186 & 5248 & & 5196 & & 5222 & 5225 & 75 & @xmath145 & 3.37 & 0.90 + 18995 & @xmath193 & 5843 & 5334 & & 5420 & 5532 & 5500 & 272 & @xmath213 & 2.46 & 1.90 + 19378 & @xmath188 & 4671 & 4376 & 4702 & 4143 & 4473 & 4400 & 265 & @xmath225 & 1.48 & 1.65 + 21648 & @xmath184 & 4371 & 4029 & 4462 & 4053 & 4229 & 4200 & 220 & @xmath196 & 0.78 & 2.00 + 27654 & @xmath197 & 4946 & 4592 & 4787 & 4750 & 4769 & 4775 & 145 & @xmath226 & 2.44 & 1.50 + 29759 & @xmath174 & & 5430 & & 5494 & 5462 & 5475 & 75 & @xmath177 & 3.20 & 1.25 + 29992 & @xmath188 & 5445 & 5053 & 5593 & 4971 & 5266 & 5200 & 253 & @xmath227 & 2.41 & 2.20 + 30668 & @xmath180 & 5588 & 5414 & 5587 & 5329 & 5480 & 5450 & 129 & @xmath171 & 3.04 & 1.15 + 38621 & @xmath188 & 5215 & 4811 & 5169 & & 5065 & 5075 & 221 & @xmath200 & 1.97 & 1.95 + 43228 & @xmath199 & 4959 & 4599 & 4819 & & 4792 & 4850 & 99 & @xmath185 & 1.79 & 1.50 + 49371 & @xmath190 & 5496 & 5205 & 5397 & 5105 & 5301 & 5275 & 178 & @xmath187 & 2.99 & 1.75 + 57850 & @xmath184 & 4343 & 4163 & 4295 & 4160 & 4240 & 4225 & 93 & @xmath201 & 0.87 & 2.75 + 57939 & @xmath180 & 5419 & 5329 & 5184 & 5109 & 5260 & 5225 & 112 & @xmath187 & 4.68 & 0.50 + 58514 & @xmath199 & 5301 & 4924 & 5142 & 4908 & 5069 & 5000 & 130 & @xmath203 & 1.98 & 1.90 + 60719 & @xmath82 & & 5334 & & 5259 & 5297 & 5300 & 75 & @xmath181 & 2.85 & 1.20 + 62235 & @xmath199 & 5616 & 5458 & 5518 & & 5531 & 5525 & 80 & @xmath213 & 2.45 & 2.05 + 62747 & @xmath180 & 5320 & & 5269 & 4967 & 5185 & 5125 & 214 & @xmath192 & 2.61 & 1.45 + 64115 & @xmath186 & 5004 & 4898 & 4806 & 4728 & 4859 & 4800 & 119 & @xmath204 & 2.61 & 1.10 + 65852 & @xmath190 & 4664 & 4603 & 4681 & & 4649 & 4650 & 75 & @xmath219 & 1.93 & 1.75 + 66246 & @xmath173 & 4432 & 4258 & 4746 & & 4479 & 4450 & 117 & @xmath228 & 1.07 & 2.65 + 68594 & @xmath205 & 5171 & 4637 & 5115 & 4636 & 4890 & 4650 & & @xmath185 & 1.26 & 1.85 + 71087 & @xmath199 & 5253 & 5150 & 5196 & 5009 & 5152 & 5125 & 104 & @xmath192 & 2.28 & 1.30 + 73960 & @xmath193 & 4381 & 4301 & 4375 & 4239 & 4324 & 4300 & 75 & @xmath208 & 1.21 & 2.10 + 74491 & @xmath173 & 4161 & 4245 & 4157 & 4153 & 4179 & 4175 & 75 & @xmath207 & 1.21 & 2.30 + 85487 & @xmath190 & 5614 & & 5427 & 5042 & 5361 & 5275 & 272 & @xmath227 & 2.32 & 1.80 + 85855 & @xmath1 & & & & 5108 & 5108 & 5125 & & @xmath203 & 2.93 & 1.40 + 88527 & @xmath174 & 4644 & 4301 & 4307 & 4290 & 4386 & 4475 & 172 & @xmath220 & 1.22 & 2.70 + 88977 & @xmath193 & & 5092 & & 5136 & 5114 & 5125 & 75 & @xmath181 & 2.36 & 1.80 + 91182 & @xmath209 & 5251 & & 5862 & 4904 & 5339 & 5200 & 245 & @xmath187 & 1.80 & 1.40 + 92167 & @xmath180 & 5460 & 5220 & 5191 & & 5290 & 5300 & 148 & @xmath177 & 2.67 & 1.35 + 94931 & @xmath212 & 5277 & & 5322 & 4994 & 5198 & 5150 & 200 & @xmath187 & 4.66 & 0.80 + 96248 & @xmath199 & & 5491 & 6136 & & 5814 & 5575 & 228 & @xmath213 & 2.49 & 2.20 + 97023 & @xmath214 & 5992 & & 5900 & & 5946 & 5950 & 75 & @xmath229 & 3.98 & 1.20 + 97468 & @xmath188 & 4724 & 4342 & 4925 & 4307 & 4575 & 4525 & 231 & @xmath219 & 1.21 & 1.90 + 98532 & @xmath172 & 6011 & & 5994 & 5675 & 5893 & 5850 & 189 & @xmath218 & 3.88 & 1.30 + 104659 & @xmath217 & 6134 & & 6121 & 5986 & 6080 & 6075 & 82 & @xmath230 & 4.31 & 1.10 + 106095 & @xmath173 & 5160 & 4708 & 5219 & 4701 & 4947 & 4900 & 281 & @xmath175 & 1.68 & 2.10 + 107337 & @xmath199 & 4909 & 4716 & 4954 & 4687 & 4817 & 4800 & 135 & @xmath204 & 2.00 & 1.55 + 109390 & @xmath193 & 5082 & & 5036 & 4730 & 4949 & 4900 & 191 & @xmath175 & 2.38 & 1.45 + 112796 & @xmath174 & 4993 & 4537 & 5041 & 4534 & 4776 & 4725 & 279 & @xmath198 & 1.28 & 2.55 + 114502 & @xmath173 & 5575 & 5121 & 5543 & 5142 & 5345 & 5300 & 247 & @xmath227 & 2.54 & 1.55 + 115949 & @xmath174 & 4918 & 4481 & 5025 & 4484 & 4727 & 4675 & 286 & @xmath219 & 1.25 & 2.40 + rrrrrrrrrllr 434 & 1.40 & 0.94 & @xmath213 & 2.04 & + 0.7 & + 0.73 & 0.71 & & + 0.7 & 0.2 & 0.60 + 484 & 0.26 & 5.42 & @xmath231 & 11.77 & + 1.4 & + 1.28 & & + 2.6 & + 1.4 & 0.3 & 0.87 + 2413 & 2.11 & 0.43 & @xmath232 & 0.93 & + 0.2 & + 0.66 & agb ? & + 0.2 & + 0.2 & 0.3 & 0.83 + 2463 & 0.06 & 17.7 & @xmath233 & 38.37 & @xmath209&@xmath234 & 0.56 & & + 0.6 & 0.2 & 0.60 + 3985 & 9.04 & 0.15 & + 3.54 & 0.33 & & & & + 3.6 & + 3.6 & 0.3 & 0.97 + 4933 & 1.92 & 0.54 & @xmath185 & 1.17 & @xmath197&@xmath235 & agb & & @xmath182 & 0.3 & 0.60 + 5445 & 7.74 & 0.12 & + 2.16 & 0.26 & + 1.8 & 1.47 & & + 1.7 & + 1.8 & 0.3 & 0.84 + 5458 & 4.35 & 0.18 & @xmath236 & 0.39 & @xmath209 & & & @xmath186&@xmath214 & 0.3 & 0.89 + 6710 & 3.04 & 0.31 & + 0.72 & 0.67 & @xmath182&@xmath237 & & @xmath182&@xmath182 & 0.2 & 0.83 + 14086 & 32.94 & 0.02 & + 3.47 & 0.04 & & & & + 3.5 & + 3.47 & 0.04 & 0.97 + 16214 & 4.27 & 0.28 & + 1.86 & 0.61 & & + 0.61 & & + 0.3 & + 0.6 & 0.3 & 0.84 + 17639 & 0.71 & 1.34 & @xmath238 & 2.91 & & @xmath239 & & @xmath199&@xmath188 & 0.3 & 0.83 + 18235 & 15.87 & 0.05 & + 2.68 & 0.11 & & + 2.17 & & + 2.7 & + 2.68 & 0.11 & 0.94 + 18995 & 4.39 & 0.28 & + 1.43 & 0.63 & + 0.2 & + 0.79 & 0.73 & & + 0.7 & 0.2 & 0.60 + 19378 & 1.28 & 0.79 & @xmath240 & 1.72 & @xmath209&@xmath241 & & @xmath170&@xmath176 & 0.3 & 0.84 + 21648 & 0.66 & 1.56 & @xmath242 & 3.39 & @xmath243&@xmath244 & & @xmath174&@xmath174 & 0.2 & 0.83 + 27654 & 29.05 & 0.02 & + 1.08 & 0.04 & & & & + 0.0 & + 1.08 & 0.04 & 0.89 + 29759 & 5.78 & 0.23 & + 2.73 & 0.50 & & & & + 1.8 & + 2.3 & 0.4 & 0.83 + 29992 & 5.17 & 0.20 & + 1.62 & 0.43 & + 0.6 & + 1.83 & & @xmath245 & + 0.6 & 0.4 & 0.84 + 30668 & 7.34 & 0.13 & + 2.33 & 0.28 & & + 1.89 & & + 1.7 & + 1.9 & 0.3 & 0.84 + 38621 & 1.75 & 0.57 & @xmath246 & 1.24 & @xmath247 & & & @xmath212&@xmath247 & 0.3 & 0.84 + 43228 & 1.55 & 0.75 & @xmath203 & 1.63 & @xmath212&@xmath248 & & @xmath176&@xmath214 & 0.3 & 0.84 + 49371 & 4.38 & 0.30 & + 2.18 & 0.65 & & & & + 0.5 & + 2.0 & 0.6 & 0.83 + 57850 & 0.72 & 1.51 & @xmath249 & 3.28 & & & & @xmath190&@xmath190 & 0.5 & 0.84 + 57939 & 109.21 & 0.01 & + 6.61 & 0.02 & & & & + 6.7 & + 6.61 & 0.02 & 0.61 + 58514 & 0.41 & 2.73 & @xmath250 & 5.93 & + 0.1 & + 0.13 & agb & & + 0.1 & 0.3 & 0.60 + 60719 & 4.53 & 0.23 & + 1.31 & 0.50 & + 1.6 & + 0.52 & & + 1.8 & + 1.6 & 0.2 & 0.82 + 62235 & 0.01 & 134.00 & @xmath251 & 291.00 & & + 0.74 & 0.67 & & + 0.7 & 0.2 & 0.60 + 62747 & 3.29 & 0.34 & + 0.56 & 0.74 & & + 1.16 & & + 2.2 & + 1.2 & 0.3 & 0.84 + 64115 & 4.73 & 0.24 & + 1.72 & 0.52 & & & & + 1.2 & + 1.5 & 0.5 & 0.94 + 65852 & 3.90 & 0.36 & + 2.66 & 0.78 & & & & + 0.4 & + 0.4 & 0.5 & 0.60 + 66246 & 0.66 & 1.88 & @xmath252 & 4.08 & & @xmath241 & & @xmath178&@xmath184 & 0.4 & 0.83 + 68594 & 3.76 & 0.19 & @xmath253 & 0.41 & @xmath172&@xmath254 & agb & & @xmath172 & 0.3 & 0.60 + 71087 & 1.52 & 0.95 & + 0.75 & 2.06 & + 0.3 & & & + 0.1 & + 0.3 & 0.3 & 0.84 + 73960 & 3.45 & 0.38 & + 1.82 & 0.83 & @xmath170&@xmath237 & & @xmath180&@xmath170 & 0.3 & 0.86 + 74491 & 2.11 & 0.65 & + 1.11 & 1.41 & @xmath217 & & & @xmath82&@xmath217 & 0.4 & 0.83 + 85487 & 3.67 & 0.41 & + 2.21 & 0.89 & + 0.1 & + 0.65 & 0.59 & & + 0.6 & 0.2 & 0.60 + 85855 & 4.29 & 0.27 & + 2.10 & 0.59 & & & & + 1.9 & + 2.0 & 0.4 & 0.82 + 88527 & 2.20 & 0.47 & @xmath255 & 1.02 & @xmath180&@xmath256 & & @xmath257&@xmath180 & 0.4 & 0.82 + 88977 & 3.25 & 0.37 & + 0.69 & 0.80 & @xmath197 & + 0.79 & 0.73 & & + 0.8 & 0.2 & 0.60 + 91182 & 5.57 & 0.24 & + 2.22 & 0.52 & @xmath217 & + 0.75 & agb ? & + 0.4 & @xmath217 & 0.3 & 0.95 + 92167 & 6.18 & 0.09 & + 1.13 & 0.20 & + 1.5 & + 1.79 & & + 0.8 & + 1.3 & 0.2 & 0.84 + 94931 & 28.28 & 0.03 & + 6.13 & 0.07 & & & & + 6.0 & + 6.13 & 0.07 & 0.95 + 96248 & 3.28 & 0.29 & + 0.17 & 0.63 & + 0.8 & & 0.67 & & + 0.7 & 0.2 & 0.60 + 97023 & 22.10 & 0.04 & + 3.59 & 0.09 & & & & + 4.0 & + 3.59 & 0.09 & 0.94 + 97468 & 1.99 & 0.55 & @xmath258 & 1.19 & @xmath199&@xmath124 & & @xmath188&@xmath199 & 0.2 & 0.84 + 98532 & 14.76 & 0.08 & + 3.57 & 0.17 & & & & + 3.5 & + 3.57 & 0.17 & 0.85 + 104659 & 28.26 & 0.04 & + 4.63 & 0.09 & & & & + 5.2 & + 4.63 & 0.09 & 0.76 + 106095 & 0.24 & 5.75 & @xmath259 & 12.49 & + 0.0 & @xmath260 & abg & @xmath197&@xmath214 & 0.4 & 0.60 + 107337 & 0.69 & 1.80 & @xmath261 & 3.91 & + 0.2 & @xmath200 & & @xmath197 & + 0.0 & 0.4 & 0.84 + 109390 & 2.54 & 0.55 & + 1.57 & 1.19 & & & & + 0.1 & + 0.8 & 0.5 & 0.86 + 112796 & 3.14 & 0.38 & + 0.30 & 0.83 & @xmath172&@xmath262 & & @xmath180&@xmath193 & 0.3 & 0.60 + 114502 & 3.51 & 0.40 & + 1.67 & 0.87 & + 0.7 & + 0.81 & & + 0.8 & + 0.8 & 0.3 & 0.83 + 115949 & 2.30 & 0.37 & @xmath263 & 0.80 & @xmath199&@xmath264 & & @xmath184&@xmath199 & 0.3 & 0.82 + lrrrrrrrccc t@xmath265 k & + 0.12 & + 0.07 & @xmath266&@xmath216&@xmath215 & + 0.23 & @xmath267 & + 0.12 & @xmath268&@xmath227 + log @xmath25 + 0.3 dex & + 0.11 & + 0.08 & + 0.12 & + 0.11 & + 0.09 & @xmath45 & + 0.12 & @xmath267 & @xmath45 & + 0.01 + [ m / h ] + 0.3 dex & + 0.09 & + 0.07 & 0.00 & @xmath45&@xmath45&@xmath45 & + 0.05 & + 0.03 & @xmath236&@xmath269 + v@xmath270 + 0.3 km s@xmath37 & 0.00 & 0.00 & @xmath267&@xmath45&@xmath45&@xmath236&@xmath271 & + 0.03 & + 0.02 & @xmath45 + rccccccccccccccc sun & @xmath272 & & @xmath273 & & @xmath274 & & @xmath275 & & @xmath45 & @xmath276 & & @xmath277 & & @xmath278 & + 434 & @xmath279 & @xmath280 & @xmath281 & @xmath276 & @xmath282 & @xmath283 & @xmath284 & @xmath285 & @xmath286 & @xmath287 & @xmath283 & @xmath288 & @xmath289 & @xmath290 & @xmath291 + 484 & @xmath292 & @xmath293 & @xmath294 & @xmath283 & @xmath295 & @xmath296 & @xmath297 & @xmath283 & @xmath298 & @xmath299 & @xmath300 & @xmath301 & @xmath280 & @xmath302 & @xmath303 + 2413 & @xmath304 & @xmath305 & @xmath306 & @xmath307 & @xmath308 & @xmath300 & @xmath309 & @xmath277 & @xmath310 & @xmath311 & @xmath312 & @xmath313 & @xmath287 & @xmath314 & @xmath315 + 2463 & @xmath316 & @xmath317 & @xmath318 & @xmath305 & @xmath319 & @xmath280 & @xmath320 & @xmath321 & @xmath322 & @xmath323 & @xmath324 & @xmath325 & @xmath326 & @xmath327 & @xmath328 + 3985 & @xmath329 & @xmath296 & @xmath330 & @xmath293 & @xmath331 & @xmath296 & @xmath332 & @xmath283 & @xmath333 & @xmath334 & @xmath300 & @xmath335 & @xmath296 & @xmath336 & @xmath324 + 4933 & @xmath337 & @xmath338 & @xmath339 & @xmath340 & @xmath341 & @xmath293 & @xmath341 & @xmath324 & @xmath342 & @xmath299 & @xmath303 & @xmath343 & @xmath344 & @xmath345 & @xmath346 + 5445 & @xmath347 & @xmath280 & @xmath348 & @xmath349 & @xmath350 & @xmath321 & @xmath351 & @xmath300 & @xmath352 & @xmath353 & @xmath283 & @xmath354 & @xmath349 & @xmath355 & @xmath356 + 5458 & @xmath357 & @xmath324 & @xmath358 & @xmath340 & @xmath359 & @xmath296 & @xmath360 & @xmath361 & @xmath362 & @xmath289 & @xmath283 & @xmath301 & @xmath296 & @xmath363 & @xmath361 + 6710 & @xmath364 & @xmath305 & @xmath365 & @xmath296 & @xmath366 & @xmath321 & @xmath367 & @xmath276 & @xmath368 & @xmath369 & @xmath289 & @xmath325 & @xmath344 & @xmath363 & @xmath370 + 14086 & @xmath371 & @xmath372 & @xmath373 & @xmath300 & @xmath331 & @xmath296 & @xmath337 & @xmath285 & @xmath374 & @xmath375 & @xmath372 & @xmath376 & @xmath285 & @xmath377 & @xmath283 + 16214 & @xmath378 & @xmath303 & @xmath379 & @xmath380 & @xmath381 & @xmath300 & @xmath382 & @xmath283 & @xmath383 & @xmath384 & @xmath303 & @xmath385 & @xmath338 & @xmath386 & @xmath387 + 17639 & @xmath388 & @xmath277 & @xmath389 & @xmath307 & @xmath390 & @xmath296 & @xmath391 & @xmath277 & @xmath392 & @xmath393 & @xmath361 & @xmath394 & @xmath276 & @xmath395 & @xmath396 + 18235 & @xmath397 & @xmath296 & @xmath398 & @xmath285 & @xmath399 & @xmath283 & @xmath400 & @xmath285 & @xmath401 & @xmath402 & @xmath285 & @xmath375 & @xmath283 & @xmath355 & @xmath277 + 18995 & @xmath403 & @xmath277 & @xmath404 & @xmath317 & @xmath405 & @xmath321 & @xmath406 & @xmath372 & @xmath407 & @xmath408 & @xmath280 & @xmath409 & @xmath317 & @xmath410 & @xmath356 + 19378 & @xmath411 & @xmath324 & @xmath412 & @xmath303 & @xmath413 & @xmath321 & @xmath414 & @xmath340 & @xmath126 & @xmath415 & @xmath324 & @xmath416 & @xmath361 & @xmath44 & @xmath417 + 21648 & @xmath418 & @xmath296 & @xmath419 & @xmath312 & @xmath420 & @xmath277 & @xmath421 & @xmath280 & @xmath422 & @xmath402 & @xmath277 & @xmath423 & @xmath361 & @xmath424 & @xmath307 + 27654 & @xmath425 & @xmath321 & @xmath426 & @xmath317 & @xmath427 & @xmath321 & @xmath428 & @xmath340 & @xmath248 & @xmath429 & @xmath340 & @xmath430 & @xmath280 & @xmath431 & @xmath317 + 29759 & @xmath432 & @xmath324 & @xmath279 & @xmath276 & @xmath433 & @xmath296 & @xmath320 & @xmath296 & @xmath322 & @xmath434 & @xmath296 & @xmath435 & @xmath436 & @xmath327 & @xmath417 + 29992 & @xmath437 & @xmath334 & @xmath438 & @xmath361 & @xmath439 & @xmath340 & @xmath440 & @xmath349 & @xmath441 & @xmath323 & @xmath442 & @xmath443 & @xmath444 & @xmath146 & @xmath429 + 30668 & @xmath347 & @xmath361 & @xmath445 & @xmath361 & @xmath295 & @xmath296 & @xmath446 & @xmath283 & @xmath447 & @xmath384 & @xmath296 & @xmath353 & @xmath303 & @xmath448 & @xmath307 + 38621 & @xmath449 & @xmath324 & @xmath450 & @xmath312 & @xmath451 & @xmath283 & @xmath452 & @xmath300 & @xmath453 & @xmath299 & @xmath361 & @xmath454 & @xmath317 & @xmath455 & @xmath380 + 43228 & @xmath456 & @xmath283 & @xmath457 & @xmath276 & @xmath414 & @xmath296 & @xmath458 & @xmath340 & @xmath459 & @xmath415 & @xmath296 & @xmath460 & @xmath296 & @xmath461 & @xmath303 + 49371 & @xmath462 & @xmath276 & @xmath463 & @xmath303 & @xmath464 & @xmath296 & @xmath465 & @xmath277 & @xmath466 & @xmath467 & @xmath296 & @xmath435 & @xmath324 & @xmath468 & @xmath349 + 57850 & @xmath469 & @xmath361 & @xmath470 & @xmath276 & @xmath471 & @xmath340 & @xmath421 & @xmath361 & @xmath422 & @xmath472 & @xmath283 & @xmath473 & @xmath300 & @xmath474 & @xmath361 + 57939 & @xmath348 & @xmath372 & @xmath475 & @xmath285 & @xmath476 & @xmath300 & @xmath477 & @xmath478 & @xmath479 & @xmath480 & @xmath372 & @xmath454 & @xmath285 & @xmath481 & @xmath283 + 58514 & @xmath365 & @xmath340 & @xmath482 & @xmath324 & @xmath414 & @xmath321 & @xmath483 & @xmath300 & @xmath484 & @xmath429 & @xmath300 & @xmath343 & @xmath293 & @xmath485 & @xmath436 + 60719 & @xmath486 & @xmath283 & @xmath316 & @xmath285 & @xmath487 & @xmath340 & @xmath319 & @xmath285 & @xmath488 & @xmath434 & @xmath372 & @xmath489 & @xmath300 & @xmath314 & @xmath296 + 62235 & @xmath490 & @xmath280 & @xmath491 & @xmath277 & @xmath414 & @xmath300 & @xmath492 & @xmath285 & @xmath493 & @xmath346 & @xmath283 & @xmath494 & @xmath293 & @xmath495 & @xmath312 + 62747 & @xmath496 & @xmath277 & @xmath497 & @xmath277 & @xmath295 & @xmath296 & @xmath284 & @xmath300 & @xmath286 & @xmath353 & @xmath300 & @xmath498 & @xmath361 & @xmath395 & @xmath317 + 64115 & @xmath499 & @xmath361 & @xmath371 & @xmath340 & @xmath500 & @xmath280 & @xmath501 & @xmath361 & @xmath502 & @xmath402 & @xmath503 & @xmath384 & @xmath285 & @xmath377 & @xmath285 + 65852 & @xmath504 & @xmath361 & @xmath505 & @xmath277 & @xmath506 & @xmath321 & @xmath367 & @xmath280 & @xmath368 & @xmath387 & @xmath300 & @xmath507 & @xmath300 & @xmath495 & @xmath293 + 66246 & @xmath508 & @xmath291 & @xmath438 & @xmath436 & @xmath509 & @xmath283 & @xmath510 & @xmath280 & @xmath511 & @xmath480 & @xmath349 & @xmath512 & @xmath289 & @xmath513 & @xmath514 + 68594 & @xmath515 & @xmath436 & @xmath516 & @xmath307 & @xmath517 & @xmath283 & @xmath518 & @xmath285 & @xmath519 & @xmath323 & @xmath289 & @xmath520 & @xmath442 & @xmath386 & @xmath521 + 71087 & @xmath318 & @xmath280 & @xmath522 & @xmath293 & @xmath414 & @xmath296 & @xmath523 & @xmath321 & @xmath524 & @xmath313 & @xmath296 & @xmath288 & @xmath293 & @xmath525 & @xmath312 + 73960 & @xmath526 & @xmath280 & @xmath527 & @xmath283 & @xmath528 & @xmath321 & @xmath529 & @xmath296 & @xmath254 & @xmath326 & @xmath321 & @xmath394 & @xmath300 & @xmath530 & @xmath277 + 74491 & @xmath531 & @xmath340 & @xmath532 & @xmath277 & @xmath533 & @xmath280 & @xmath534 & @xmath280 & @xmath535 & @xmath311 & @xmath300 & @xmath325 & @xmath300 & @xmath536 & @xmath293 + 85487 & @xmath537 & @xmath361 & @xmath538 & @xmath324 & @xmath539 & @xmath300 & @xmath391 & @xmath372 & @xmath392 & @xmath480 & @xmath280 & @xmath540 & @xmath303 & @xmath541 & @xmath417 + 85855 & @xmath542 & @xmath283 & @xmath543 & @xmath283 & @xmath544 & @xmath283 & @xmath545 & @xmath300 & @xmath546 & @xmath460 & @xmath478 & @xmath547 & @xmath285 & @xmath548 & @xmath283 + 88527 & @xmath549 & @xmath277 & @xmath550 & @xmath283 & @xmath551 & @xmath296 & @xmath509 & @xmath296 & @xmath552 & @xmath323 & @xmath300 & @xmath553 & @xmath300 & @xmath548 & @xmath277 + 88977 & @xmath294 & @xmath349 & @xmath554 & @xmath402 & @xmath555 & @xmath283 & @xmath556 & @xmath296 & @xmath557 & @xmath384 & @xmath307 & @xmath494 & @xmath415 & @xmath558 & @xmath559 + 91182 & @xmath560 & @xmath436 & @xmath470 & @xmath561 & @xmath562 & @xmath280 & @xmath563 & @xmath380 & @xmath564 & @xmath289 & @xmath380 & @xmath384 & @xmath375 & @xmath565 & @xmath498 + 92167 & @xmath566 & @xmath280 & @xmath567 & @xmath285 & @xmath405 & @xmath300 & @xmath568 & @xmath300 & @xmath569 & @xmath402 & @xmath285 & @xmath323 & @xmath283 & @xmath570 & @xmath277 + 94931 & @xmath571 & @xmath285 & @xmath273 & @xmath277 & @xmath572 & @xmath340 & @xmath573 & @xmath296 & @xmath574 & @xmath315 & @xmath296 & @xmath472 & @xmath285 & @xmath575 & @xmath293 + 96248 & @xmath576 & @xmath307 & @xmath577 & @xmath300 & @xmath578 & @xmath296 & @xmath579 & @xmath277 & @xmath580 & @xmath521 & @xmath324 & @xmath581 & @xmath303 & @xmath582 & @xmath356 + 97023 & @xmath398 & @xmath478 & @xmath583 & @xmath478 & @xmath537 & @xmath321 & @xmath584 & @xmath478 & @xmath585 & @xmath291 & @xmath503 & @xmath338 & @xmath478 & @xmath586 & @xmath478 + 97468 & @xmath279 & @xmath293 & @xmath281 & @xmath417 & @xmath464 & @xmath300 & @xmath381 & @xmath324 & @xmath587 & @xmath315 & @xmath436 & @xmath553 & @xmath324 & @xmath290 & @xmath380 + 98532 & @xmath450 & @xmath283 & @xmath588 & @xmath277 & @xmath589 & @xmath296 & @xmath590 & @xmath372 & @xmath591 & @xmath417 & @xmath300 & @xmath467 & @xmath277 & @xmath345 & @xmath289 + 104659 & @xmath592 & @xmath296 & @xmath593 & @xmath340 & @xmath594 & @xmath300 & @xmath589 & @xmath372 & @xmath595 & @xmath434 & @xmath321 & @xmath596 & @xmath277 & @xmath45 & @xmath303 + 106095 & @xmath378 & @xmath293 & @xmath597 & @xmath346 & @xmath598 & @xmath321 & @xmath599 & @xmath361 & @xmath600 & @xmath601 & @xmath324 & @xmath343 & @xmath396 & @xmath7 & @xmath602 + 107337 & @xmath603 & @xmath293 & @xmath604 & @xmath312 & @xmath421 & @xmath300 & @xmath458 & @xmath293 & @xmath459 & @xmath369 & @xmath280 & @xmath325 & @xmath324 & @xmath424 & @xmath307 + 109390 & @xmath605 & @xmath361 & @xmath398 & @xmath436 & @xmath606 & @xmath300 & @xmath607 & @xmath361 & @xmath608 & @xmath429 & @xmath340 & @xmath609 & @xmath324 & @xmath610 & @xmath349 + 112796 & @xmath542 & @xmath312 & @xmath611 & @xmath317 & @xmath612 & @xmath321 & @xmath509 & @xmath283 & @xmath552 & @xmath402 & @xmath276 & @xmath423 & @xmath307 & @xmath558 & @xmath514 + 114502 & @xmath613 & @xmath338 & @xmath611 & @xmath340 & @xmath614 & @xmath280 & @xmath615 & @xmath436 & @xmath616 & @xmath384 & @xmath293 & @xmath617 & @xmath291 & @xmath548 & @xmath346 + 115949 & @xmath618 & @xmath312 & @xmath419 & @xmath396 & @xmath506 & @xmath321 & @xmath619 & @xmath300 & @xmath620 & @xmath299 & @xmath312 & @xmath621 & @xmath417 & @xmath622 & @xmath623 + rccccccccccccccc sun & @xmath272 & & @xmath273 & & @xmath274 & & @xmath275 & & @xmath45 & @xmath276 & & @xmath277 & & @xmath278 & + 434 & @xmath462 & @xmath280 & @xmath624 & @xmath276 & @xmath625 & @xmath296 & @xmath284 & @xmath285 & @xmath286 & @xmath514 & @xmath283 & @xmath385 & @xmath289 & @xmath626 & @xmath291 + 484 & @xmath605 & @xmath293 & @xmath627 & @xmath283 & @xmath628 & @xmath277 & @xmath297 & @xmath283 & @xmath298 & @xmath384 & @xmath300 & @xmath429 & @xmath280 & @xmath629 & @xmath303 + 2413 & @xmath630 & @xmath305 & @xmath613 & @xmath307 & @xmath631 & @xmath300 & @xmath632 & @xmath277 & @xmath633 & @xmath369 & @xmath312 & @xmath299 & @xmath287 & @xmath146 & @xmath315 + 2463 & @xmath274 & @xmath317 & @xmath634 & @xmath305 & @xmath309 & @xmath276 & @xmath420 & @xmath321 & @xmath635 & @xmath498 & @xmath324 & @xmath376 & @xmath326 & @xmath171 & @xmath328 + 3985 & @xmath397 & @xmath296 & @xmath593 & @xmath293 & @xmath613 & @xmath293 & @xmath636 & @xmath283 & @xmath637 & @xmath408 & @xmath300 & @xmath623 & @xmath296 & @xmath215 & @xmath324 + 4933 & @xmath638 & @xmath338 & @xmath639 & @xmath340 & @xmath631 & @xmath349 & @xmath319 & @xmath324 & @xmath488 & @xmath640 & @xmath303 & @xmath287 & @xmath344 & @xmath574 & @xmath346 + 5445 & @xmath437 & @xmath280 & @xmath462 & @xmath349 & @xmath534 & @xmath283 & @xmath641 & @xmath300 & @xmath642 & @xmath354 & @xmath283 & @xmath429 & @xmath349 & @xmath643 & @xmath356 + 5458 & @xmath588 & @xmath324 & @xmath527 & @xmath340 & @xmath644 & @xmath277 & @xmath645 & @xmath361 & @xmath234 & @xmath444 & @xmath283 & @xmath408 & @xmath296 & @xmath395 & @xmath361 + 6710 & @xmath646 & @xmath305 & @xmath538 & @xmath296 & @xmath381 & @xmath283 & @xmath452 & @xmath276 & @xmath453 & @xmath384 & @xmath289 & @xmath559 & @xmath344 & @xmath647 & @xmath370 + 14086 & @xmath648 & @xmath372 & @xmath329 & @xmath300 & @xmath486 & @xmath277 & @xmath427 & @xmath285 & @xmath649 & @xmath301 & @xmath372 & @xmath328 & @xmath285 & @xmath266 & @xmath283 + 16214 & @xmath566 & @xmath303 & @xmath318 & @xmath380 & @xmath650 & @xmath280 & @xmath381 & @xmath283 & @xmath587 & @xmath596 & @xmath303 & @xmath651 & @xmath338 & @xmath194 & @xmath387 + 17639 & @xmath652 & @xmath277 & @xmath653 & @xmath307 & @xmath654 & @xmath293 & @xmath655 & @xmath277 & @xmath656 & @xmath288 & @xmath361 & @xmath561 & @xmath276 & @xmath267 & @xmath396 + 18235 & @xmath398 & @xmath296 & @xmath593 & @xmath285 & @xmath657 & @xmath340 & @xmath658 & @xmath285 & @xmath659 & @xmath353 & @xmath285 & @xmath369 & @xmath283 & @xmath216 & @xmath277 + 18995 & @xmath660 & @xmath277 & @xmath661 & @xmath317 & @xmath662 & @xmath321 & @xmath663 & @xmath372 & @xmath664 & @xmath472 & @xmath280 & @xmath435 & @xmath317 & @xmath548 & @xmath356 + 19378 & @xmath665 & @xmath324 & @xmath566 & @xmath303 & @xmath631 & @xmath300 & @xmath350 & @xmath340 & @xmath666 & @xmath480 & @xmath324 & @xmath472 & @xmath361 & @xmath667 & @xmath417 + 21648 & @xmath418 & @xmath296 & @xmath419 & @xmath312 & @xmath420 & @xmath277 & @xmath421 & @xmath280 & @xmath422 & @xmath402 & @xmath277 & @xmath423 & @xmath361 & @xmath424 & @xmath307 + 27654 & @xmath668 & @xmath321 & @xmath398 & @xmath317 & @xmath360 & @xmath321 & @xmath669 & @xmath340 & @xmath670 & @xmath335 & @xmath340 & @xmath596 & @xmath280 & @xmath146 & @xmath317 + 29759 & @xmath388 & @xmath324 & @xmath418 & @xmath276 & @xmath413 & @xmath340 & @xmath655 & @xmath296 & @xmath656 & @xmath671 & @xmath296 & @xmath520 & @xmath436 & @xmath575 & @xmath417 + 29992 & @xmath672 & @xmath334 & @xmath462 & @xmath361 & @xmath673 & @xmath361 & @xmath382 & @xmath349 & @xmath383 & @xmath674 & @xmath442 & @xmath408 & @xmath444 & @xmath177 & @xmath429 + 30668 & @xmath526 & @xmath361 & @xmath279 & @xmath361 & @xmath556 & @xmath280 & @xmath446 & @xmath283 & @xmath447 & @xmath354 & @xmath296 & @xmath675 & @xmath303 & @xmath221 & @xmath307 + 38621 & @xmath603 & @xmath324 & @xmath676 & @xmath312 & @xmath297 & @xmath293 & @xmath367 & @xmath300 & @xmath368 & @xmath393 & @xmath361 & @xmath675 & @xmath317 & @xmath177 & @xmath380 + 43228 & @xmath412 & @xmath283 & @xmath456 & @xmath276 & @xmath590 & @xmath277 & @xmath677 & @xmath340 & @xmath241 & @xmath467 & @xmath296 & @xmath429 & @xmath296 & @xmath678 & @xmath303 + 49371 & @xmath348 & @xmath276 & @xmath456 & @xmath303 & @xmath641 & @xmath280 & @xmath421 & @xmath277 & @xmath422 & @xmath460 & @xmath296 & @xmath385 & @xmath324 & @xmath267 & @xmath349 + 57850 & @xmath469 & @xmath361 & @xmath470 & @xmath276 & @xmath471 & @xmath340 & @xmath421 & @xmath361 & @xmath422 & @xmath472 & @xmath283 & @xmath473 & @xmath300 & @xmath474 & @xmath361 + 57939 & @xmath412 & @xmath372 & @xmath560 & @xmath285 & @xmath406 & @xmath296 & @xmath677 & @xmath478 & @xmath241 & @xmath467 & @xmath372 & @xmath416 & @xmath285 & @xmath679 & @xmath283 + 58514 & @xmath605 & @xmath340 & @xmath457 & @xmath324 & @xmath297 & @xmath296 & @xmath650 & @xmath300 & @xmath680 & @xmath467 & @xmath300 & @xmath559 & @xmath293 & @xmath355 & @xmath436 + 60719 & @xmath681 & @xmath283 & @xmath682 & @xmath285 & @xmath683 & @xmath340 & @xmath319 & @xmath285 & @xmath488 & @xmath684 & @xmath372 & @xmath498 & @xmath300 & @xmath647 & @xmath296 + 62235 & @xmath412 & @xmath280 & @xmath685 & @xmath277 & @xmath297 & @xmath283 & @xmath492 & @xmath285 & @xmath493 & @xmath415 & @xmath283 & @xmath498 & @xmath293 & @xmath686 & @xmath312 + 62747 & @xmath365 & @xmath277 & @xmath624 & @xmath277 & @xmath687 & @xmath280 & @xmath606 & @xmath300 & @xmath688 & @xmath354 & @xmath300 & @xmath684 & @xmath361 & @xmath667 & @xmath317 + 64115 & @xmath499 & @xmath361 & @xmath371 & @xmath340 & @xmath500 & @xmath280 & @xmath501 & @xmath361 & @xmath502 & @xmath402 & @xmath503 & @xmath384 & @xmath285 & @xmath377 & @xmath285 + 65852 & @xmath652 & @xmath361 & @xmath597 & @xmath277 & @xmath632 & @xmath296 & @xmath689 & @xmath280 & @xmath690 & @xmath375 & @xmath300 & @xmath553 & @xmath300 & @xmath410 & @xmath293 + 66246 & @xmath611 & @xmath291 & @xmath279 & @xmath436 & @xmath691 & @xmath283 & @xmath692 & @xmath280 & @xmath693 & @xmath596 & @xmath349 & @xmath385 & @xmath289 & @xmath694 & @xmath514 + 68594 & @xmath515 & @xmath436 & @xmath516 & @xmath307 & @xmath517 & @xmath283 & @xmath518 & @xmath285 & @xmath519 & @xmath323 & @xmath289 & @xmath520 & @xmath442 & @xmath386 & @xmath521 + 71087 & @xmath538 & @xmath280 & @xmath456 & @xmath293 & @xmath695 & @xmath280 & @xmath696 & @xmath321 & @xmath697 & @xmath472 & @xmath296 & @xmath323 & @xmath293 & @xmath146 & @xmath312 + 73960 & @xmath412 & @xmath280 & @xmath698 & @xmath283 & @xmath699 & @xmath321 & @xmath662 & @xmath296 & @xmath700 & @xmath701 & @xmath321 & @xmath702 & @xmath300 & @xmath558 & @xmath277 + 74491 & @xmath531 & @xmath340 & @xmath532 & @xmath277 & @xmath533 & @xmath280 & @xmath534 & @xmath280 & @xmath535 & @xmath311 & @xmath300 & @xmath325 & @xmath300 & @xmath703 & @xmath293 + 85487 & @xmath543 & @xmath361 & @xmath676 & @xmath324 & @xmath433 & @xmath296 & @xmath433 & @xmath372 & @xmath704 & @xmath354 & @xmath280 & @xmath559 & @xmath303 & @xmath705 & @xmath417 + 85855 & @xmath706 & @xmath283 & @xmath449 & @xmath283 & @xmath707 & @xmath283 & @xmath708 & @xmath300 & @xmath709 & @xmath385 & @xmath478 & @xmath507 & @xmath285 & @xmath710 & @xmath283 + 88527 & @xmath432 & @xmath277 & @xmath711 & @xmath283 & @xmath712 & @xmath293 & @xmath712 & @xmath296 & @xmath713 & @xmath288 & @xmath300 & @xmath520 & @xmath300 & @xmath694 & @xmath277 + 88977 & @xmath660 & @xmath349 & @xmath714 & @xmath402 & @xmath284 & @xmath283 & @xmath687 & @xmath296 & @xmath715 & @xmath353 & @xmath307 & @xmath716 & @xmath415 & @xmath717 & @xmath559 + 91182 & @xmath567 & @xmath436 & @xmath718 & @xmath561 & @xmath306 & @xmath293 & @xmath719 & @xmath380 & @xmath720 & @xmath721 & @xmath380 & @xmath338 & @xmath375 & @xmath629 & @xmath498 + 92167 & @xmath722 & @xmath280 & @xmath348 & @xmath285 & @xmath594 & @xmath280 & @xmath534 & @xmath300 & @xmath535 & @xmath376 & @xmath285 & @xmath415 & @xmath283 & @xmath268 & @xmath277 + 94931 & @xmath571 & @xmath285 & @xmath273 & @xmath277 & @xmath572 & @xmath340 & @xmath573 & @xmath296 & @xmath574 & @xmath315 & @xmath296 & @xmath472 & @xmath285 & @xmath575 & @xmath293 + 96248 & @xmath389 & @xmath307 & @xmath723 & @xmath300 & @xmath367 & @xmath296 & @xmath606 & @xmath277 & @xmath688 & @xmath701 & @xmath324 & @xmath724 & @xmath303 & @xmath725 & @xmath356 + 97023 & @xmath398 & @xmath478 & @xmath583 & @xmath478 & @xmath537 & @xmath321 & @xmath584 & @xmath478 & @xmath585 & @xmath291 & @xmath503 & @xmath338 & @xmath478 & @xmath586 & @xmath478 + 97468 & @xmath603 & @xmath293 & @xmath726 & @xmath417 & @xmath727 & @xmath340 & @xmath452 & @xmath324 & @xmath453 & @xmath393 & @xmath436 & @xmath640 & @xmath324 & @xmath728 & @xmath380 + 98532 & @xmath672 & @xmath283 & @xmath419 & @xmath277 & @xmath563 & @xmath340 & @xmath589 & @xmath372 & @xmath595 & @xmath326 & @xmath300 & @xmath328 & @xmath277 & @xmath728 & @xmath289 + 104659 & @xmath729 & @xmath296 & @xmath730 & @xmath340 & @xmath731 & @xmath283 & @xmath628 & @xmath372 & @xmath732 & @xmath454 & @xmath321 & @xmath601 & @xmath277 & @xmath177 & @xmath303 + 106095 & @xmath348 & @xmath293 & @xmath665 & @xmath346 & @xmath727 & @xmath277 & @xmath381 & @xmath361 & @xmath587 & @xmath559 & @xmath324 & @xmath376 & @xmath396 & @xmath733 & @xmath602 + 107337 & @xmath450 & @xmath293 & @xmath718 & @xmath312 & @xmath284 & @xmath296 & @xmath579 & @xmath293 & @xmath580 & @xmath601 & @xmath280 & @xmath596 & @xmath324 & @xmath314 & @xmath307 + 109390 & @xmath734 & @xmath361 & @xmath735 & @xmath436 & @xmath663 & @xmath296 & @xmath673 & @xmath361 & @xmath736 & @xmath354 & @xmath340 & @xmath454 & @xmath324 & @xmath737 & @xmath349 + 112796 & @xmath339 & @xmath312 & @xmath738 & @xmath317 & @xmath309 & @xmath280 & @xmath509 & @xmath283 & @xmath552 & @xmath354 & @xmath276 & @xmath671 & @xmath307 & @xmath647 & @xmath514 + 114502 & @xmath388 & @xmath338 & @xmath316 & @xmath340 & @xmath696 & @xmath361 & @xmath739 & @xmath436 & @xmath252 & @xmath596 & @xmath293 & @xmath335 & @xmath291 & @xmath740 & @xmath346 + 115949 & @xmath638 & @xmath312 & @xmath726 & @xmath396 & @xmath741 & @xmath340 & @xmath615 & @xmath300 & @xmath616 & @xmath467 & @xmath312 & @xmath742 & @xmath417 & @xmath290 & @xmath623 + rrrrrrr sun & @xmath743 & @xmath744 & @xmath745 & @xmath746 & @xmath747 & + 434 & @xmath748 & @xmath749 & @xmath172 & @xmath750 & @xmath751 & 65 + 484 & @xmath752 & @xmath753 & @xmath217 & @xmath754 & @xmath755 & 60 + 2413 & @xmath756 & @xmath757 & @xmath173 & @xmath758 & @xmath759 & 90 + 2463 & @xmath760 & @xmath757 & @xmath190 & @xmath761 & @xmath762 & 60 + 3985 & @xmath763 & @xmath764 & @xmath209 & @xmath460 & @xmath765 & 70 + 4933 & @xmath766 & @xmath767 & @xmath243 & @xmath768 & @xmath769 & 70 + 5445 & @xmath770 & @xmath771 & @xmath170 & @xmath772 & @xmath773 & 65 + 5458 & @xmath774 & @xmath775 & @xmath176 & @xmath776 & @xmath777 & 35 + 6710 & @xmath778 & @xmath767 & @xmath188 & @xmath779 & @xmath780 & 85 + 14086 & @xmath781 & @xmath764 & @xmath209 & @xmath553 & @xmath782 & 45 + 16214 & @xmath783 & @xmath784 & @xmath199 & @xmath758 & @xmath785 & 55 + 17639 & @xmath786 & @xmath787 & @xmath199 & @xmath788 & @xmath789 & 40 + 18235 & @xmath790 & @xmath791 & @xmath209 & @xmath288 & @xmath792 & 40 + 18995 & @xmath793 & @xmath749 & @xmath182 & @xmath750 & @xmath794 & 25 + 19378 & @xmath795 & @xmath767 & @xmath170 & @xmath796 & @xmath797 & 75 + 21648 & @xmath798 & @xmath799 & @xmath184 & @xmath800 & @xmath801 & 60 + 27654 & @xmath802 & @xmath749 & @xmath197 & @xmath772 & @xmath803 & 15 + 29759 & @xmath804 & @xmath791 & @xmath190 & @xmath805 & @xmath806 & 75 + 29992 & @xmath807 & @xmath784 & @xmath180 & @xmath788 & @xmath808 & 40 + 30668 & @xmath809 & @xmath810 & @xmath180 & @xmath811 & @xmath812 & 70 + 38621 & @xmath813 & @xmath814 & @xmath199 & @xmath815 & @xmath816 & 60 + 43228 & @xmath817 & @xmath818 & @xmath170 & @xmath772 & @xmath819 & 60 + 49371 & @xmath820 & @xmath771 & @xmath199 & @xmath821 & @xmath822 & 85 + 57850 & @xmath823 & @xmath824 & @xmath173 & @xmath825 & @xmath826 & 45 + 57939 & @xmath827 & @xmath828 & @xmath170 & @xmath328 & @xmath829 & 135 + 58514 & @xmath830 & @xmath831 & @xmath172 & @xmath750 & @xmath832 & 50 + 60719 & @xmath833 & @xmath834 & @xmath174 & @xmath835 & @xmath836 & 95 + 62235 & @xmath837 & @xmath749 & @xmath172 & @xmath838 & @xmath839 & 65 + 62747 & @xmath840 & @xmath753 & @xmath193 & @xmath758 & @xmath841 & 35 + 64115 & @xmath842 & @xmath753 & @xmath209 & @xmath553 & @xmath792 & 45 + 65852 & @xmath843 & @xmath831 & @xmath184 & @xmath821 & @xmath844 & 75 + 66246 & @xmath845 & @xmath824 & @xmath184 & @xmath846 & @xmath847 & 55 + 68594 & @xmath848 & @xmath849 & @xmath257 & @xmath850 & @xmath851 & 105 + 71087 & @xmath852 & @xmath757 & @xmath180 & @xmath853 & @xmath854 & 80 + 73960 & @xmath855 & @xmath787 & @xmath193 & @xmath768 & @xmath856 & 55 + 74491 & @xmath857 & @xmath858 & @xmath199 & @xmath859 & @xmath860 & 40 + 85487 & @xmath861 & @xmath757 & @xmath184 & @xmath761 & @xmath862 & 70 + 85855 & @xmath863 & @xmath771 & @xmath243 & @xmath754 & @xmath803 & 75 + 88527 & @xmath864 & @xmath787 & @xmath178 & @xmath865 & @xmath866 & 75 + 88977 & @xmath867 & @xmath784 & @xmath182 & @xmath761 & @xmath868 & 20 + 91182 & @xmath869 & @xmath818 & @xmath197 & @xmath754 & @xmath870 & 75 + 92167 & @xmath840 & @xmath753 & @xmath193 & @xmath871 & @xmath872 & 50 + 94931 & @xmath873 & @xmath874 & @xmath214 & @xmath460 & @xmath875 & 75 + 96248 & @xmath876 & @xmath749 & @xmath172 & @xmath788 & @xmath877 & 45 + 97023 & @xmath878 & @xmath879 & @xmath247 & @xmath835 & @xmath880 & 100 + 97468 & @xmath881 & @xmath787 & @xmath188 & @xmath750 & @xmath882 & 100 + 98532 & @xmath883 & @xmath884 & @xmath197 & @xmath853 & @xmath885 & 150 + 104659 & @xmath886 & @xmath887 & @xmath182 & @xmath553 & @xmath888 & 100 + 106095 & @xmath889 & @xmath890 & @xmath180 & @xmath768 & @xmath891 & 40 + 107337 & @xmath892 & @xmath831 & @xmath170 & @xmath776 & @xmath794 & 45 + 109390 & @xmath893 & @xmath784 & @xmath172 & @xmath758 & @xmath894 & 45 + 112796 & @xmath895 & @xmath858 & @xmath178 & @xmath896 & @xmath897 & 50 + 114502 & @xmath898 & @xmath749 & @xmath173 & @xmath776 & @xmath899 & 85 + 115949 & @xmath900 & @xmath858 & @xmath190 & @xmath901 & @xmath839 & 45 + rccccccccccccccc sun & @xmath272 & & @xmath272 & & @xmath275 & & @xmath418 & & @xmath902 & @xmath324 & & @xmath324 & & @xmath902 & + 434 & @xmath379 & @xmath283 & @xmath566 & @xmath280 & @xmath590 & @xmath289 & @xmath625 & @xmath285 & @xmath903 & @xmath299 & @xmath300 & @xmath651 & @xmath277 & @xmath45 & @xmath293 + 484 & @xmath627 & @xmath296 & @xmath567 & @xmath296 & @xmath556 & @xmath317 & @xmath904 & @xmath300 & @xmath905 & @xmath313 & @xmath503 & @xmath353 & @xmath277 & @xmath586 & @xmath293 + 2413 & @xmath486 & @xmath296 & @xmath486 & @xmath324 & @xmath906 & @xmath442 & @xmath632 & @xmath321 & @xmath633 & @xmath623 & @xmath280 & @xmath623 & @xmath276 & @xmath902 & @xmath303 + 2463 & @xmath388 & @xmath300 & @xmath504 & @xmath361 & @xmath907 & @xmath436 & @xmath320 & @xmath372 & @xmath322 & @xmath443 & @xmath283 & @xmath908 & @xmath361 & @xmath694 & @xmath324 + 3985 & @xmath425 & @xmath340 & @xmath527 & @xmath280 & @xmath909 & @xmath312 & @xmath658 & @xmath283 & @xmath659 & @xmath299 & @xmath277 & @xmath375 & @xmath293 & @xmath910 & @xmath324 + 4933 & @xmath911 & @xmath340 & @xmath573 & @xmath277 & @xmath683 & @xmath317 & @xmath912 & @xmath300 & @xmath913 & @xmath335 & @xmath277 & @xmath467 & @xmath293 & @xmath45 & @xmath324 + 5445 & @xmath347 & @xmath340 & @xmath914 & @xmath293 & @xmath579 & @xmath436 & @xmath641 & @xmath300 & @xmath642 & @xmath915 & @xmath277 & @xmath472 & @xmath361 & @xmath448 & @xmath276 + 5458 & @xmath294 & @xmath280 & @xmath505 & @xmath280 & @xmath658 & @xmath277 & @xmath916 & @xmath361 & @xmath917 & @xmath623 & @xmath303 & @xmath402 & @xmath289 & @xmath355 & @xmath324 + 6710 & @xmath318 & @xmath293 & @xmath918 & @xmath324 & @xmath439 & @xmath380 & @xmath451 & @xmath283 & @xmath919 & @xmath353 & @xmath324 & @xmath335 & @xmath276 & @xmath586 & @xmath436 + 14086 & @xmath398 & @xmath321 & @xmath371 & @xmath296 & @xmath911 & @xmath293 & @xmath639 & @xmath372 & @xmath920 & @xmath921 & @xmath300 & @xmath472 & @xmath340 & @xmath679 & @xmath280 + 16214 & @xmath665 & @xmath296 & @xmath538 & @xmath361 & @xmath414 & @xmath436 & @xmath440 & @xmath300 & @xmath441 & @xmath335 & @xmath280 & @xmath915 & @xmath324 & @xmath278 & @xmath276 + 17639 & @xmath432 & @xmath283 & @xmath922 & @xmath293 & @xmath391 & @xmath303 & @xmath907 & @xmath300 & @xmath923 & @xmath924 & @xmath280 & @xmath702 & @xmath361 & @xmath925 & @xmath324 + 18235 & @xmath499 & @xmath300 & @xmath926 & @xmath300 & @xmath572 & @xmath280 & @xmath400 & @xmath285 & @xmath401 & @xmath315 & @xmath283 & @xmath601 & @xmath296 & @xmath705 & @xmath340 + 18995 & @xmath593 & @xmath285 & @xmath927 & @xmath321 & @xmath928 & @xmath300 & @xmath590 & @xmath372 & @xmath591 & @xmath385 & @xmath321 & @xmath393 & @xmath300 & @xmath271 & @xmath283 + 19378 & @xmath389 & @xmath280 & @xmath389 & @xmath289 & @xmath929 & @xmath291 & @xmath641 & @xmath296 & @xmath642 & @xmath408 & @xmath361 & @xmath408 & @xmath317 & @xmath902 & @xmath312 + 21648 & @xmath469 & @xmath300 & @xmath611 & @xmath303 & @xmath699 & @xmath349 & @xmath692 & @xmath300 & @xmath693 & @xmath443 & @xmath296 & @xmath596 & @xmath289 & @xmath236 & @xmath289 + 27654 & @xmath425 & @xmath372 & @xmath698 & @xmath285 & @xmath500 & @xmath285 & @xmath719 & @xmath478 & @xmath720 & @xmath559 & @xmath285 & @xmath301 & @xmath321 & @xmath45 & @xmath300 + 29759 & @xmath930 & @xmath340 & @xmath931 & @xmath293 & @xmath932 & @xmath436 & @xmath655 & @xmath283 & @xmath656 & @xmath430 & @xmath293 & @xmath454 & @xmath324 & @xmath45 & @xmath303 + 29992 & @xmath450 & @xmath296 & @xmath526 & @xmath296 & @xmath579 & @xmath277 & @xmath696 & @xmath296 & @xmath697 & @xmath434 & @xmath277 & @xmath385 & @xmath277 & @xmath448 & @xmath277 + 30668 & @xmath665 & @xmath296 & @xmath490 & @xmath280 & @xmath650 & @xmath317 & @xmath446 & @xmath300 & @xmath447 & @xmath408 & @xmath280 & @xmath915 & @xmath293 & @xmath705 & @xmath361 + 38621 & @xmath933 & @xmath296 & @xmath550 & @xmath293 & @xmath677 & @xmath349 & @xmath452 & @xmath300 & @xmath453 & @xmath915 & @xmath280 & @xmath376 & @xmath361 & @xmath448 & @xmath324 + 43228 & @xmath379 & @xmath296 & @xmath379 & @xmath361 & @xmath934 & @xmath442 & @xmath282 & @xmath283 & @xmath935 & @xmath353 & @xmath280 & @xmath353 & @xmath276 & @xmath936 & @xmath276 + 49371 & @xmath437 & @xmath324 & @xmath437 & @xmath324 & @xmath528 & @xmath937 & @xmath421 & @xmath277 & @xmath422 & @xmath684 & @xmath289 & @xmath684 & @xmath436 & @xmath936 & @xmath312 + 57850 & @xmath450 & @xmath277 & @xmath438 & @xmath276 & @xmath606 & @xmath289 & @xmath692 & @xmath361 & @xmath693 & @xmath325 & @xmath289 & @xmath938 & @xmath312 & @xmath355 & @xmath317 + 57939 & @xmath939 & @xmath340 & @xmath438 & @xmath276 & @xmath589 & @xmath380 & @xmath350 & @xmath372 & @xmath666 & @xmath416 & @xmath280 & @xmath354 & @xmath303 & @xmath45 & @xmath436 + 58514 & @xmath624 & @xmath296 & @xmath419 & @xmath277 & @xmath529 & @xmath436 & @xmath295 & @xmath300 & @xmath940 & @xmath335 & @xmath280 & @xmath416 & @xmath361 & @xmath694 & @xmath361 + 60719 & @xmath549 & @xmath296 & @xmath634 & @xmath277 & @xmath941 & @xmath307 & @xmath942 & @xmath283 & @xmath261 & @xmath443 & @xmath277 & @xmath430 & @xmath361 & @xmath705 & @xmath324 + 62235 & @xmath419 & @xmath300 & @xmath943 & @xmath296 & @xmath944 & @xmath436 & @xmath662 & @xmath372 & @xmath700 & @xmath921 & @xmath283 & @xmath353 & @xmath340 & @xmath448 & @xmath280 + 62747 & @xmath597 & @xmath283 & @xmath597 & @xmath283 & @xmath945 & @xmath277 & @xmath458 & @xmath300 & @xmath459 & @xmath559 & @xmath340 & @xmath559 & @xmath340 & @xmath902 & @xmath280 + 64115 & @xmath499 & @xmath361 & @xmath946 & @xmath277 & @xmath573 & @xmath324 & @xmath400 & @xmath324 & @xmath401 & @xmath315 & @xmath317 & @xmath384 & @xmath436 & @xmath355 & @xmath436 + 65852 & @xmath603 & @xmath361 & @xmath389 & @xmath303 & @xmath727 & @xmath356 & @xmath578 & @xmath277 & @xmath947 & @xmath684 & @xmath289 & @xmath434 & @xmath317 & @xmath267 & @xmath338 + 66246 & @xmath646 & @xmath277 & @xmath274 & @xmath361 & @xmath510 & @xmath317 & @xmath413 & @xmath340 & @xmath948 & @xmath385 & @xmath324 & @xmath596 & @xmath303 & @xmath267 & @xmath289 + 68594 & @xmath949 & @xmath293 & @xmath950 & @xmath396 & @xmath544 & @xmath356 & @xmath951 & @xmath285 & @xmath952 & @xmath671 & @xmath293 & @xmath460 & @xmath396 & @xmath45 & @xmath338 + 71087 & @xmath538 & @xmath280 & @xmath445 & @xmath276 & @xmath297 & @xmath396 & @xmath953 & @xmath285 & @xmath264 & @xmath376 & @xmath280 & @xmath559 & @xmath303 & @xmath910 & @xmath436 + 73960 & @xmath412 & @xmath280 & @xmath954 & @xmath361 & @xmath406 & @xmath324 & @xmath955 & @xmath340 & @xmath125 & @xmath354 & @xmath324 & @xmath596 & @xmath303 & @xmath910 & @xmath303 + 74491 & @xmath469 & @xmath340 & @xmath665 & @xmath276 & @xmath696 & @xmath277 & @xmath956 & @xmath280 & @xmath957 & @xmath323 & @xmath324 & @xmath443 & @xmath436 & @xmath377 & @xmath436 + 85487 & @xmath508 & @xmath283 & @xmath508 & @xmath293 & @xmath958 & @xmath312 & @xmath433 & @xmath372 & @xmath704 & @xmath959 & @xmath296 & @xmath959 & @xmath293 & @xmath902 & @xmath324 + 85855 & @xmath537 & @xmath340 & @xmath960 & @xmath361 & @xmath961 & @xmath436 & @xmath962 & @xmath283 & @xmath963 & @xmath343 & @xmath293 & @xmath964 & @xmath276 & @xmath694 & @xmath303 + 88527 & @xmath965 & @xmath293 & @xmath966 & @xmath324 & @xmath487 & @xmath305 & @xmath539 & @xmath296 & @xmath967 & @xmath288 & @xmath324 & @xmath498 & @xmath303 & @xmath448 & @xmath317 + 88977 & @xmath532 & @xmath285 & @xmath592 & @xmath321 & @xmath359 & @xmath300 & @xmath687 & @xmath285 & @xmath715 & @xmath434 & @xmath300 & @xmath443 & @xmath300 & @xmath586 & @xmath283 + 91182 & @xmath968 & @xmath296 & @xmath968 & @xmath340 & @xmath306 & @xmath324 & @xmath916 & @xmath321 & @xmath917 & @xmath444 & @xmath340 & @xmath444 & @xmath280 & @xmath902 & @xmath293 + 92167 & @xmath560 & @xmath300 & @xmath672 & @xmath300 & @xmath406 & @xmath303 & @xmath625 & @xmath321 & @xmath903 & @xmath408 & @xmath296 & @xmath384 & @xmath340 & @xmath902 & @xmath340 + 94931 & @xmath969 & @xmath280 & @xmath571 & @xmath276 & @xmath970 & @xmath276 & @xmath399 & @xmath296 & @xmath971 & @xmath408 & @xmath293 & @xmath353 & @xmath303 & @xmath679 & @xmath436 + 96248 & @xmath463 & @xmath285 & @xmath567 & @xmath321 & @xmath650 & @xmath280 & @xmath972 & @xmath285 & @xmath973 & @xmath376 & @xmath300 & @xmath416 & @xmath300 & @xmath705 & @xmath300 + 97023 & @xmath491 & @xmath283 & @xmath398 & @xmath296 & @xmath549 & @xmath442 & @xmath304 & @xmath478 & @xmath223 & @xmath937 & @xmath296 & @xmath417 & @xmath280 & @xmath45 & @xmath293 + 97468 & @xmath411 & @xmath283 & @xmath445 & @xmath317 & @xmath458 & @xmath380 & @xmath974 & @xmath296 & @xmath239 & @xmath684 & @xmath280 & @xmath674 & @xmath344 & @xmath377 & @xmath344 + 98532 & @xmath722 & @xmath293 & @xmath627 & @xmath324 & @xmath975 & @xmath326 & @xmath589 & @xmath372 & @xmath595 & @xmath311 & @xmath361 & @xmath521 & @xmath276 & @xmath448 & @xmath317 + 104659 & @xmath592 & @xmath283 & @xmath482 & @xmath296 & @xmath976 & @xmath307 & @xmath589 & @xmath478 & @xmath595 & @xmath434 & @xmath283 & @xmath393 & @xmath296 & @xmath267 & @xmath277 + 106095 & @xmath437 & @xmath296 & @xmath490 & @xmath277 & @xmath528 & @xmath324 & @xmath381 & @xmath296 & @xmath587 & @xmath301 & @xmath293 & @xmath416 & @xmath361 & @xmath267 & @xmath324 + 107337 & @xmath726 & @xmath280 & @xmath412 & @xmath293 & @xmath607 & @xmath303 & @xmath677 & @xmath280 & @xmath241 & @xmath353 & @xmath324 & @xmath467 & @xmath303 & @xmath705 & @xmath303 + 109390 & @xmath734 & @xmath293 & @xmath734 & @xmath293 & @xmath594 & @xmath324 & @xmath650 & @xmath293 & @xmath680 & @xmath596 & @xmath436 & @xmath596 & @xmath436 & @xmath902 & @xmath436 + 112796 & @xmath977 & @xmath283 & @xmath549 & @xmath277 & @xmath614 & @xmath303 & @xmath509 & @xmath321 & @xmath552 & @xmath559 & @xmath296 & @xmath323 & @xmath293 & @xmath448 & @xmath361 + 114502 & @xmath978 & @xmath277 & @xmath364 & @xmath324 & @xmath979 & @xmath291 & @xmath739 & @xmath296 & @xmath252 & @xmath301 & @xmath324 & @xmath323 & @xmath303 & @xmath694 & @xmath436 + 115949 & @xmath922 & @xmath300 & @xmath652 & @xmath277 & @xmath440 & @xmath276 & @xmath619 & @xmath321 & @xmath620 & @xmath674 & @xmath296 & @xmath561 & @xmath293 & @xmath902 & @xmath293 + | we present oxygen abundances derived from both the permitted and forbidden oxygen lines for 55 subgiants and giants with @xmath0}$ ] values between @xmath1 and solar with the goal of understanding the discrepancy in the derived abundances . a first attempt , using t@xmath2 values from photometric calibrations and surface gravities from luminosities , obtained agreement between the indicators for turn - off stars , but the disagreement was large for evolved stars .
we find that the difference in the oxygen abundances derived from the permitted and forbidden lines is most strongly affected by t@xmath2 , and we derive a new t@xmath2 scale based on forcing the two sets of lines to give the same oxygen abundances .
these new parameters , however , do not agree with other observables , such as theoretical isochrones or balmer - line profile based t@xmath2 determinations .
our analysis finds that one - dimensional , lte analyses ( with published nlte corrections for the permitted lines ) can not fully resolve the disagreement in the two indicators without adopting a temperature scale incompatible with other temperature indicators .
we also find no evidence of circumstellar emission in the forbidden lines , removing such emission as a possible cause for the discrepancy . |
the standard model ( sm ) is very successful in explaining almost all experimental data to date , culminating in the recent discovery of the long awaited higgs boson at the cern large hadron collider ( lhc ) @xcite . in the sm
, all elementary particles acquire mass from a single higgs doublet that generates spontaneous electroweak symmetry breaking ( ewsb ) .
all charged fermions have their masses and yukawa couplings to the higgs boson as correlated but free parameters . furthermore , there are no flavor changing neutral currents ( fcnc ) mediated by gauge interactions , nor by higgs interactions ( fcnh ) , at the tree level
. the most important goals of the lhc , at run 2 and beyond , are the study of higgs properties and the search for signals , direct or indirect , of new physics beyond the sm .
as the most massive particle ever discovered , the top quark might provide clues to better understand the mechanism of ewsb . a possible explanation for its heaviness could be provided by a special two higgs doublet model for the top quark ( t2hdm ) @xcite , where it is the only fermion that couples to a higgs doublet with a large vacuum expectation value ( vev ) .
the second heaviest particle is the newly discovered higgs boson ( @xmath4 ) . with @xmath7
, it opens up the possibility of top quark decays into @xmath4 plus a charm quark .
the branching fraction of @xmath8 in sm at one loop level is approximately @xmath9 @xcite for @xmath10 gev .
if this decay is detected , it would indicate a large effective fcnh coupling of tree - level origins @xcite , or very large enhancement from beyond sm loop effects @xcite . in flavor conserving two higgs doublet models , a discrete symmetry @xcite
is often imposed to distinguish the su(2 ) doublet fields @xmath11 from @xmath12 . without such a discrete symmetry ,
a general two higgs doublet model ( 2hdm ) should possess fcnh vertices . to study such interactions , we adopt the following lagrangian involving higgs bosons and fermions @xcite , @xmath13 h^0 + \left [ \kappa^fc_{\beta-\alpha}-\rho^fs_{\beta-\alpha } \right ] h^0 - i \ , { \rm sgn}(q_f)\rho^f a^0 \right\ } p_r f \nonumber \\ & & -\bar{u } \left [ v \rho^d p_r - \rho^{u\dagger } v p_l \right ] d h^+ -\bar{\nu } \left [ \rho^l p_r \right ] l h^+ + { \rm h.c . } \ , , \end{aligned}\ ] ] where @xmath14 , @xmath15 , @xmath16 , @xmath17 , and @xmath18 is the mixing angle between neutral higgs scalars in the type ii ( 2hdm - ii ) notation @xcite .
@xmath19 matrices are diagonal and fixed by fermion masses to @xmath20 with @xmath21 gev , while @xmath22 matrices are free and have both diagonal and off - diagonal elements .
we adopt a cp conserving higgs model and choose @xmath22 matrices to be real but not necessarily hermitian .
@xmath23 , @xmath24 , @xmath25 and @xmath26 are vectors in flavor space ( @xmath27 , etc . ) . @xmath4 and @xmath2 are cp - even scalars ( @xmath28 ) , while @xmath3 is a cp - odd pseudoscalar . with the advent of the lhc ,
theoretical interest in search of fcnh top decays ( @xmath29 ) picked up @xcite , and the atlas and cms experiments have already placed the branching fraction limit @xmath30 @xcite , implying @xmath31 . for lhc at @xmath32 tev and integrated luminosity of @xmath33 fb@xmath34 ,
the atlas experiment expects @xcite to reach @xmath35 , i.e. probing down to @xmath36 .
the flavor changing heavy higgs decay ( @xmath37 ) is complementary to fcnh top decay ( @xmath29 ) , since the coupling @xmath38 is proportional to @xmath5 while @xmath39 .
higgs boson data from the lhc favor the decoupling limit @xcite or the alignment limit @xcite of a 2hdm . in this limit ,
fcnh couplings of @xmath4 are naturally suppressed by small @xmath5 , while off - diagonal couplings of @xmath2 , @xmath3 are sustained by @xmath40 . in this letter , we study the discovery potential of the lhc in the search for heavy higgs bosons @xmath2 or @xmath3 that decay into a top quark and a charm quark .
the top quark then decays into a b quark , a charged lepton ( @xmath41 or @xmath42 ) , and a neutrino . taking lhc higgs data and @xmath43 physics constraints into account
, we evaluate production rates with full tree - level matrix elements for both signal and background .
we optimize the acceptance cuts to effectively reduce the latter with realistic @xmath44-tagging and mistag efficiencies .
promising results are presented for the lhc with @xmath45 tev as well as @xmath46 tev .
in this section , we apply the latest results from lhc higgs measurements , as well as from @xmath43 physics , to constrain the parameters @xmath47 , @xmath48 , @xmath49 , @xmath50 , and @xmath5 of a general 2hdm that are relevant for observing flavor changing decays of heavy higgs bosons at the lhc .
run 1 of lhc at @xmath51 and 8 tev has provided us with information on the couplings of the higgs boson @xmath4 , by measuring the event rates relative to the sm signal strength .
even with our general 2hdm , the light higgs boson @xmath4 constitutes a narrow resonance , and the signal strength for a production channel @xmath52 and final state @xmath53 can be written as @xmath54 atlas and cms often show the signal strength of measurements in two dimensions by grouping gluon fusion and @xmath55 production on one axis ( ggf ) , and vector boson fusion and associated production on the other axis ( vbf ) .
these contours can be used to draw constraints on 2hdm s @xcite .
we follow a simpler approach and consider signal strengths for final states with the largest statistics , namely @xmath56 , @xmath57 , @xmath58 , and @xmath59 , where the dominant production mode is gluon fusion , as well as the signal strength for the @xmath60 final state from the associated production @xmath61 with @xmath62 or @xmath63 .
table [ tab : signalstrength ] shows the average signal strengths obtained by the experimental groups in run 1 .
we combine the atlas and cms results and show both the combined values with their uncertainties in the last column .
l3.4cm||c2.3cm|c2.3cm|c2.3 cm final state & @xmath42(atlas ) & @xmath42(cms ) & @xmath42(comb . )
+ @xmath64 & @xmath65 @xcite & @xmath66 @xcite & @xmath67 + @xmath68 & @xmath69 @xcite & @xmath70 @xcite & @xmath71 + @xmath72 & @xmath73 @xcite & @xmath74 @xcite & @xmath75 + @xmath76 & @xmath77 @xcite & @xmath78 @xcite & @xmath79 + @xmath80 & @xmath81 @xcite & @xmath82 @xcite & @xmath83 to find the allowed regions of 2hdm parameter space compatible with atlas and cms data , we take the discovered higgs boson as the lightest cp even state ( @xmath4 ) of a general two higgs doublet model and scan over the following sets of parameters : @xmath84 in sm , the most important contributions to the higgs total width come from @xmath60 , @xmath58 , @xmath85 , @xmath59 , @xmath86 and @xmath57 .
the same channels are expected to contribute in a general 2hdm . in our analysis , we include all the relevant parameters in eq .
( [ eq : scanpars ] ) that affect the total width , gluon fusion cross section , and the associated higgs production ( @xmath61 ) .
we cover a wide range of values for each free parameter , and require that all yukawa couplings of the mass eigenstates ( @xmath4 , @xmath2 , @xmath3 , @xmath87 ) stay perturbative , and that the constraints given in table [ tab : signalstrength ] are satisfied .
the signal strength for each production and decay channel can be expressed in terms of scale factors which are couplings of the higgs boson to fermions and gauge bosons normalized to their standard model values @xcite .
these scale factors can then be expressed in terms of the parameters of a specific 2hdm .
negative results from heavy higgs searches provide us further insights on the parameter space of an extended higgs sector .
one of the strongest results comes from the search for a heavy higgs decaying into the @xmath88 and @xmath89 final states , excluding a heavy higgs boson with sm like couplings all the way up to 1 tev @xcite .
we use these results to further constraint the parameter space of 2hdm . in fig .
[ fig : atlas_cms ] , we present the 68% ( 95% ) confidence level ( c.l . )
regions in dark ( light ) color that are compatible with lhc constraints from the light higgs boson ( @xmath4 ) alone as well as constraints from both the light higgs boson and the heavy higgs boson ( @xmath2 ) decaying into @xmath88 and @xmath89 @xcite for a general 2hdm and and a type - ii 2hdm .
this figure shows that a large value of @xmath90 is still a possibility within a general model for @xmath91 .
this is due to the lack of a strong constraint on the @xmath44-quark coupling of the higgs boson . to be consistent with the sm higgs cross section from gluon fusion ,
a small value of @xmath5 is favored for @xmath92 with @xmath93 .
experimental data from run 2 with higher energy and higher luminosity will provide much better guidance for parameters such as @xmath47 and @xmath5 .
@xmath47 plane at 68% ( 95% ) c.l . of lhc higgs data in a general 2hdm [ green , ( light cyan ) ] and type ii 2hdm [ red , ( light magenta ) ] , with constraints from ( a ) light higgs boson ( @xmath4 ) , ( b ) both light higgs boson ( @xmath4 ) and heavy higgs boson ( @xmath2 ) .
[ fig : atlas_cms ] , title="fig:",width=272 ] @xmath47 plane at 68% ( 95% ) c.l . of lhc higgs data in a general 2hdm [ green , ( light cyan ) ] and type ii 2hdm [ red , ( light magenta ) ] , with constraints from ( a ) light higgs boson ( @xmath4 ) , ( b ) both light higgs boson ( @xmath4 ) and heavy higgs boson ( @xmath2 ) .
[ fig : atlas_cms ] , title="fig:",width=272 ] the fcnh coupling @xmath49 affects the @xmath94 couplings ( @xmath95 ) through @xmath96 .
this effect contributes to fcnc processes in down - type quark sector via @xmath97 and @xmath98 loops . for simplicity ,
let us assume @xmath99 is negligible .
recasting the 2hdm - ii expression @xcite , we estimate the modifications to the @xmath100-@xmath101 ( @xmath102 ) mixing amplitude ( @xmath103 ) from the box diagrams with internal @xmath104 and @xmath98 by @xmath105_{\rm sm } } & = 1 + \frac{i_{wh}(y^w , y^h , x ) + i_{hh}(y^h)}{i_{ww}(y^w ) } , \label{eq : b - mix}\end{aligned}\ ] ] where @xmath106 ( @xmath107 ) , @xmath108 , and @xmath109 , \\ & i_{hh } \simeq \left(\frac{\rho_{tt}^*}{\kappa_t}+\frac{v_{cb}\rho_{ct}^*}{v_{tb}\kappa_t}\right)^2 \left(\frac{\rho_{tt}}{\kappa_t}+\frac{v_{cq}^*\rho_{ct}}{v_{tq}^*\kappa_t}\right)^2 y^h \left [ \frac{1+y^h}{(1-y^h)^2}+\frac{2y^h\ln y^h}{(1-y^h)^3 } \right ] .
\label{eq : b - mix - loop } \end{split}\end{aligned}\ ] ] we adopt the following intervals from the summer 2014 results by utfit @xcite , @xmath110 , \quad \phi_{b_d}\in [ -8.0^\circ,~4.4^\circ ] , \notag\\ c_{b_s}&\in [ 0.9,~1.23 ] , \quad \phi_{b_s}\in [ -3.0^\circ,~4.7^\circ ] , \label{eq : utfit}\end{aligned}\ ] ] at 95% probability , where @xmath111_{\rm sm}$ ] .
the constraints from @xmath112 mixing data are shown in fig .
[ fig : b_physics](a ) on the @xmath113 plane with @xmath114 gev .
shaded regions are excluded by the 95% probability ranges in eq .
( [ eq : utfit ] ) .
the constraint from @xmath115 ( pink regions ) is slightly tighter than the @xmath116 exclusion ( blue - shaded regions ) . combining them with constraint from the @xmath117-violating phase @xmath118 ( light - green regions )
, we obtain the upper limit @xmath119 , regardless of @xmath49 . the parameter @xmath49 is strongly constrained since its effect in eq .
( [ eq : b - mix - loop ] ) is enhanced by the ckm factor @xmath120 ( @xmath102 ) .
once @xmath47 is fixed within this range , we obtain a constraint on @xmath49 . for @xmath121
, we have @xmath122 .
furthermore , the sizable phase in @xmath123 makes @xmath49 sensitive to the @xmath117-violating phase @xmath118 , even if @xmath49 is real . for @xmath124 gev ,
the constraints become : @xmath125 regardless of @xmath49 , and @xmath126 for @xmath127 .
@xmath49 plane from @xmath112-mixing , with shaded regions excluded by eq .
( [ eq : utfit ] ) , i.e. blue ( pink ) regions by @xmath128 , light - green regions by @xmath118 ; ( b ) @xmath47@xmath48 plane from @xmath129 with @xmath130 and @xmath131 gev , with dark - green shaded regions excluded by @xmath132 in eq .
( [ eq : b2sgam - exp ] ) at @xmath133 , while pink regions by @xmath134 .
blue - solid lines in ( b ) mark @xmath135 with @xmath136 .
[ fig : b_physics ] , title="fig:",width=272 ] @xmath49 plane from @xmath112-mixing , with shaded regions excluded by eq .
( [ eq : utfit ] ) , i.e. blue ( pink ) regions by @xmath128 , light - green regions by @xmath118 ; ( b ) @xmath47@xmath48 plane from @xmath129 with @xmath130 and @xmath131 gev , with dark - green shaded regions excluded by @xmath132 in eq .
( [ eq : b2sgam - exp ] ) at @xmath133 , while pink regions by @xmath134 . blue - solid lines in ( b ) mark @xmath135 with @xmath136 .
[ fig : b_physics ] , title="fig:",width=272 ] we now turn to the @xmath137 constraint .
the @xmath97-@xmath98 loop affects this process via the wilson coefficients @xmath138 at leading - order ( lo ) , which are , at the matching scale @xmath139 , given by @xmath140 here , the operator basis and the definition of @xmath141 follow ref .
we follow the procedure in ref .
@xcite and calculate first the ratio @xmath142 with world average of measurements @xmath143 @xcite and the next - to - next - to lo ( nnlo ) prediction in sm , @xmath144 @xcite .
we then require the similar ratio based on our lo calculation , i.e. @xmath145 , to be within the range allowed by @xmath146 .
choosing the matching scale and the low - energy scale as @xmath147 and @xmath148 , we reproduce the nnlo results of ref .
@xcite , @xmath149 gev , with 95% c.l . in the type - ii 2hdm gev ) at 95% c.l . ] .
[ fig : b_physics](b ) shows the allowed region in the @xmath150 plane from @xmath151 and @xmath43 mixing for @xmath114 gev and @xmath130 .
the dark - green shaded regions are excluded by the @xmath133 experimental error of @xmath132 in eq .
( [ eq : b2sgam - exp ] ) , with the theoretical uncertainty linearly added .
the constraint on @xmath47 by @xmath115 is also shown . for @xmath152 ,
@xmath48 is constrained to be within @xmath153 .
note that this touches the region of @xmath154 .
we set @xmath130 as it is already strongly constrained by @xmath112 mixing . for @xmath124 gev ,
the @xmath151 constraint on @xmath48 becomes : @xmath155 for @xmath156 .
typically , @xmath157 constrains @xmath48 more strongly than @xmath47 , as the effect of @xmath48 is enhanced by the chiral factor @xmath158 in eq .
( [ eq : c7 ] ) .
the contribution from charm loop has a mild dependence on @xmath50 through the @xmath97 couplings , @xmath159 ( @xmath160 ) . in general
, @xmath50 may be very different from @xmath49 . since the charm quark in the loop is light and
there is no ckm enhancement , the constraint on @xmath50 is expected to be much weaker .
the constraint on @xmath50 has been analyzed in ref .
@xcite . with @xmath161
, it is found from @xmath162 mixing that @xmath163 for @xmath131 gev and @xmath164 .
this constraint implies that @xmath165 for @xmath131 gev .
we note that a large @xmath50 can enhance @xmath166 rates via the @xmath97 coupling @xmath167 and can explain the 3.4@xmath168 discrepancy between the sm prediction and babar data @xcite .
for real valued @xmath50 and @xmath169 , the solution space shown in ref .
@xcite reads @xmath170 .
additional flavor constraints can be obtained from @xmath171 mixing ( @xmath172 ) @xcite and @xmath173 mixing ( @xmath174 ) @xcite for @xmath175 gev .
the value of ( @xmath176 ) is expected to be very small , thus @xmath177 mixing provides a better limit for @xmath50 . combining experimental limits from lhc higgs data and @xmath43 physics , and assuming perturbativity
, we consider @xmath178 , @xmath179 , and @xmath180 .
we now discuss the prospects of discovering fcnh interactions at the lhc through @xmath2 and @xmath3 decays .
the number of free parameters in a general 2hdm is too large for a comprehensive collider study of the fcnh signal , so we make some assumptions that are motivated by experiment . the latest experimental results point to a higgs sector
with the light cp even state behaving like the sm higgs , indicating that @xmath5 should be small .
we consider sample cases with @xmath181 and @xmath182 , which imply @xmath40 . in our case study , we choose the heavier states ( @xmath2 , @xmath3 and @xmath183 ) to be degenerate for simplicity , which is also in accordance with the decoupling limit @xcite , and we set @xmath184 in the higgs potential @xcite .
we also set @xmath185 and choose @xmath186 such that the scalar potential satisfies stability , tree - level unitarity , and perturbativity up to large masses .
to fix the yukawa couplings , we assume that @xmath187 while @xmath188 . this is in good agreement with both @xmath43 physics constraints as well as
lhc run 1 constraints . for the off - diagonal parts that generate the flavor changing signal ,
@xmath49 is constrained to small values by @xmath43 physics but we assume that @xmath50 can have larger values . in the massless limit for charm ,
the signal cross section is , to a very good approximation , only a function of a single effective coupling @xmath189 , but also very weakly depends on @xmath190 .
the contribution to the cross section from terms with @xmath191 is at least 98% without cuts and more than 93% with all cuts for @xmath192 with @xmath193 tev , and it is even more dominating for a lower higgs mass . with these experimentally motivated choices , gluon fusion is the dominant production mode for @xmath2 and @xmath3 states , and @xmath194 becomes the dominant final state at high mass ( @xmath195 tev ) .
we display in fig .
[ fig : br13 ] the branching fractions for ( a ) the heavier scalar @xmath2 , and ( b ) the pseudoscalar @xmath3 , as functions of higgs mass with @xmath181 and @xmath196 .
the computer code 2hdmc @xcite is employed to scan over @xmath197 tev and @xmath198 with sets of parameters that satisfy potential stability , tree - level unitarity , and perturbativity .
this gives rise to the `` bands '' in fig .
we also display the branching fraction @xmath199 for the aforementioned choice of @xmath200 and @xmath186 in our lhc case study with a dashed curve in fig .
we note that with large branching fraction in most of the parameter space , @xmath201 might offer great promise to discover higgs pairs at the lhc . and ( b ) higgs pseudoscalar @xmath3 versus @xmath202 , with @xmath181 , @xmath196 , and @xmath203 for diagonal couplings .
we show the allowed regions when @xmath200 and @xmath186 are varied . branching fraction @xmath199 for the lhc case study
is shown as a dashed curve .
[ fig : br13 ] , title="fig:",width=283 ] and ( b ) higgs pseudoscalar @xmath3 versus @xmath202 , with @xmath181 , @xmath196 , and @xmath203 for diagonal couplings .
we show the allowed regions when @xmath200 and @xmath186 are varied . branching fraction @xmath199 for the lhc case study
is shown as a dashed curve .
[ fig : br13 ] , title="fig:",width=283 ] our signal is the production of a heavy higgs boson from gluon fusion , with subsequent flavor changing decays into a charm quark and a top quark , and the top decays semileptonically .
more explicitly , we consider @xmath204 ( @xmath205 or @xmath3 ) , followed by @xmath206 with @xmath207 or @xmath42 . unless explicitly specified
, @xmath208 generally denotes a quark ( @xmath208 ) or an anti - quark ( @xmath209 ) and @xmath210 represents a lepton ( @xmath211 ) or anti - lepton ( @xmath212 ) .
we calculate the matrix elements analytically , and compute the signal cross section with the parton distribution functions of mstw2008 @xcite .
the factorization and renormalization scales are chosen to be @xmath213 .
in addition , to estimate the nlo cross section for @xmath214 , we use the computer code higlu @xcite to calculate @xmath215 ( @xmath216 ) , including both top and bottom quark loops to find a @xmath217-factor .
the dominant physics background to the final state of @xmath218 comes from @xmath219 , as well as @xmath220- and @xmath98-channel single top ( @xmath221 ) .
another important background is @xmath194 production where either one of the two leptons is missed for both top quarks decaying semileptonically , or two of the four jets are missed when only one of top quarks decays semileptonically .
we employ the programs madgraph @xcite and helas @xcite to evaluate the exact matrix elements for the background processes .
the factorization and the renormalization scales are chosen to be @xmath222 for @xmath223 and @xmath224 , @xmath225 for @xmath220- and @xmath98-channel single top , and @xmath226 for @xmath194 .
we use mcfm @xcite to calculate the nlo @xmath217-factors for our background processes .
let us present our strategy for full reconstruction of each event with the help of intermediate on - shell particles .
for each event , we require one @xmath44 jet and one non-@xmath44 jet , identified through @xmath44-tagging .
in addition , we require a single isolated lepton and missing transverse energy from the neutrino in the semileptonic decay of the top quark in our fcnh signal . for lepton momentum @xmath227 and neutrino momentum @xmath228 , the invariant mass constraint for an on - shell @xmath229 , @xmath230 , can be solved for the longitudinal component of the neutrino momentum ( @xmath231 ) , which is the only unknown in the event .
we obtain two solutions @xmath232 if @xmath233 hence @xmath234 complex , the event is vetoed . for @xmath235 with @xmath234
real , we choose the solution that minimizes the reconstructed top mass @xmath236 or @xmath237 .
systematics can be the limiting factor for new physics searches at high luminosities .
precise determination of the background needs to include systematics in experiments .
since our signal is a sharp peak over a smoothly falling background , the precise knowledge of background cross section at percent level is not required for a @xmath238 discovery .
an uncertainty of 30% in the background estimation might shift the limit on @xmath239 by 10% without affecting our results . to study the discovery potential , we employ three sets of realistic cuts and tagging efficiencies . for low luminosity ( ll ) :
( a ) lhc run 1 ( @xmath240 tev ) with @xmath24125 fb@xmath34 @xcite ; and ( b ) full cm energy ( @xmath45 or 14 tev ) with @xmath242 fb@xmath34 . for high luminosity ( hl ) : ( c ) full cm energy ( @xmath32 tev ) with @xmath243 fb@xmath34 or @xmath244 fb@xmath34@xcite .
we require that in every event there should be ( a ) exactly 2 jets that have @xmath245 gev ( 30 gev for hl ) and @xmath246 , and one of them must be tagged as a @xmath44-jet ; ( b ) exactly one isolated lepton that has @xmath245 gev and @xmath246 ; ( c ) at least 20 gev ( 40 gev for hl ) of missing transverse energy ( @xmath247 ) . after reconstructing the longitudinal component of the neutrino momentum
, we further require that ( d ) the reconstructed invariant mass of the top satisfies @xmath248 ; ( e ) the reconstructed invariant mass of the higgs boson satisfies @xmath249 .
we consider a further powerful acceptance cut on non-@xmath44-tagged jet momentum . in the higgs boson decay frame ,
the charm quark momentum from @xmath250 is approximately given by @xmath251}.\ ] ] since the higgs boson from gluon fusion has little transverse momentum , the @xmath252 distribution has both a kinematic cut - off and a peak at the above @xmath253 value .
we require that the transverse momentum of the non-@xmath44-tagged jet satisfies @xmath254 .
to simulate detector effects based on atlas @xcite and cms @xcite specifications , we apply gaussian smearing of momenta : @xmath255 with individual terms added in quadrature ( @xmath256 ) . in lhc run 1 at @xmath240 tev , the @xmath44-tagging efficiency ( @xmath257 ) is taken to be @xmath258 , the probability that a @xmath259-jet is mistagged as a @xmath44-jet ( @xmath260 ) is @xmath261 and the probability that any other jet is mistagged as a @xmath44-jet ( @xmath262 ) is taken to be @xmath263 . at the full cm energy ( @xmath45 or 14 tev ) with @xmath264 30 fb@xmath34 , we follow the tagging and mistag efficiencies in the atlas technical design report @xcite : @xmath265 , @xmath266 and @xmath267 . for the full cm energy ( @xmath45 or 14 tev ) with hl of 300 fb@xmath34 or 3000 fb@xmath34 , the tagging and mistag efficiencies are taken to be @xmath268 , @xmath266 and @xmath267 .
( solid ) @xmath269 at the lhc versus @xmath270 for ( a ) @xmath240 tev and ( b ) @xmath32 tev , with @xmath271 and @xmath272 . also shown
are the background cross sections ( dashed ) from single top ( @xmath273 and @xmath274 ) , @xmath229+jets ( @xmath223 and @xmath275 ) and @xmath194 with @xmath217-factors , acceptance cuts , and tagging efficiencies .
[ fig : sigma ] , title="fig:",width=272 ] ( solid ) @xmath269 at the lhc versus @xmath270 for ( a ) @xmath240 tev and ( b ) @xmath32 tev , with @xmath271 and @xmath272 . also shown
are the background cross sections ( dashed ) from single top ( @xmath273 and @xmath274 ) , @xmath229+jets ( @xmath223 and @xmath275 ) and @xmath194 with @xmath217-factors , acceptance cuts , and tagging efficiencies .
[ fig : sigma ] , title="fig:",width=272 ]
we present the signal and background cross sections at the lhc for @xmath240 tev and @xmath46 tev in fig .
[ fig : sigma ] . all tagging efficiencies and @xmath217-factors discussed above are included .
we observe that the largest contributions to the sm background come from single - top and @xmath229+jets processes , which is to be expected , since they can both produce very similar kinematics to our signal process .
in contrast , the @xmath194 background is substantially lower because of the requirement on the number of jets and leptons passing our cuts . [ cols="^,^ " , ] to estimate the discovery potential , we obtain the lower limit on @xmath276 by requiring that the 99.4%-confidence - level ( cl ) upper limit on the background is smaller than the 99.4%-cl lower limit on the signal plus background @xcite with statistical fluctuations .
this leads to the condition , @xmath277,\ ] ] where @xmath278 is the signal ( background ) cross section and @xmath25 the integrated luminosity . choosing the parameter @xmath279 corresponds to 5@xmath168 significance . for a large number of events ( @xmath280 ) ,
this requirement is equivalent to the statistical significance @xmath281 where @xmath282 is the number of signal ( background ) events .
we show in fig .
[ fig : reach ] the discovery reach in the @xmath202@xmath191 plane for the fcnh signal @xmath283 at the lhc with @xmath45 ( 14 ) tev , for dashed ( solid ) contours and for @xmath181 and @xmath182 . fig .
[ fig : reach](a , c ) are for the heavier scalar @xmath2 alone , whereas fig . [ fig : reach](b , d ) are for the degenerate case , for which the scalar @xmath2 and pseudoscalar @xmath3 signals are added together . the fcnh decay of the heavy higgs will be observable for @xmath272 and @xmath284 up to @xmath285 gev with 3000 fb@xmath34 data .
a larger value of @xmath50 will enhance the cross section , hence statistical significance , of this fcnh signal .
in a general two higgs doublet model , there could be flavor changing neutral higgs interactions with fermions .
strong limits exist for these fcnh interactions , except those involving the third generation quarks .
it is of great interest to study the relation between the most massive elementary particle ( the top quark ) and the higgs bosons .
the lhc has discovered a higgs boson lighter than the top , which makes the rare decay @xmath8 kinematically possible . in a general 2hdm ,
the decay width of @xmath8 is proportional to @xmath286 , while that of @xmath287 is proportional to @xmath288 .
therefore , they are complementary to each other in the search for new physics beyond the standard model .
we investigated the prospects for discovering @xmath289 at the lhc , where the heavy scalar @xmath2 and pseudoscalar @xmath3 are produced via gluon fusion , which are facilitated by the extra @xmath290 couplings .
the primary physics background comes from @xmath223 , @xmath274 , @xmath275 , @xmath273 , and @xmath194 .
both signal and background processes are studied with realistic acceptance cuts as well as tagging and mistag efficiencies .
promising results have been found for the lhc with a center of mass energy of 13 tev and 14 tev .
the fcnh decay of the heavy higgs will be observable for @xmath272 and @xmath284 up to @xmath291 gev with 3000 fb@xmath34 of integrated luminosity .
this result is robust against a small @xmath286 , independent of the @xmath29 search , which becomes diminished .
if c - tagging efficiency can be improved @xcite , the discovery potential of this fcnh signal will be greatly enhanced .
if @xmath292 , @xmath293 can become comparable to @xmath294 or surpass it .
recently , it was suggested that next - to - leading order qcd and electroweak corrections might swamp the signal of higgs decays into top quark pairs @xcite .
a recent analysis shows that @xmath295 with sm couplings can be very difficult to observe at the lhc @xcite . in that case
the fcnh decay of @xmath296 might offer a promising opportunity to observe the heavier higgs bosons .
we have not emphasized the @xmath297 lepton sector .
recently , the cms collaboration reported unexpected @xmath298 events @xcite that might be explained by neutral higgs boson decay @xcite .
if this can be confirmed by the atlas collaboration in the near future , or at lhc run 2 , it will be exciting new physics for fcnh interactions , and @xmath299 , unsuppressed by decoupling ( i.e. small @xmath286 ) , could help discover the exotic scalars .
ck thanks the academia sinica , national taiwan university , and national center for theoretical sciences for excellent hospitality .
ba thanks kyu jung bae for discussions . the computing for this project
was performed at the ou supercomputing center for education & research ( oscer ) at the university of oklahoma ( ou ) .
this research was supported in part by the u.s .
department of energy under grant no .
de - fg02 - 13er41979 ( ck , ba , and bm ) ; the academic summit grant most 103 - 2745-m-002 - 001-asp , as well as grant ntu - epr-103r8915 ( wsh ) ; and grant nsc 102 - 2112-m-033 - 007-my3 ( mk ) .
g. eilam , j. l. hewett and a. soni , phys .
d * 44 * , 1473 ( 1991 ) ; [ erratum - ibid .
d * 59 * , 039901 ( 1999 ) ] .
b. mele , s. petrarca and a. soddu , phys .
b * 435 * , 401 ( 1998 ) .
j. a. aguilar - saavedra , acta phys .
b * 35 * , 2695 ( 2004 ) .
hou , phys .
b * 296 * , 179 ( 1992 ) .
s. l. glashow and s. weinberg , phys .
d * 15 * , 1958 ( 1977 ) .
s. davidson and h. e. haber , phys .
d * 72 * , 035004 ( 2005 ) [ phys .
d * 72 * , 099902 ( 2005 ) ] .
f. mahmoudi and o. stal , phys .
d * 81 * ( 2010 ) 035016 .
j. a. aguilar - saavedra and g. c. branco , phys
. lett .
b * 495 * , 347 ( 2000 ) . c. kao , h .- y .
cheng , w .- s .
hou and j. sayre , phys .
b * 716 * ( 2012 ) 225 .
chen , w .- s .
hou , c. kao and m. kohda , phys .
b * 725 * ( 2013 ) 378 .
d. atwood , s. k. gupta and a. soni , jhep * 1410 * ( 2014 ) 57 .
g. belanger , b. dumont , u. ellwanger , j. f. gunion and s. kraml , phys .
d * 88 * , 075008 ( 2013 ) .
g. aad _ et al . _ [ atlas collaboration ] , phys .
d * 90 * , 112015 ( 2014 ) .
v. khachatryan _ et al . _
[ cms collaboration ] , eur .
j. c * 74 * , 3076 ( 2014 ) .
g. aad _ et al .
_ [ atlas collaboration ] , phys .
d * 91 * , 012006 ( 2015 ) .
s. chatrchyan _ et al . _
[ cms collaboration ] , phys .
d * 89 * , 092007 ( 2014 ) .
g. aad _ et al . _ [ atlas collaboration ] , arxiv:1412.2641 [ hep - ex ] .
s. chatrchyan _ et al . _
[ cms collaboration ] , jhep * 1401 * , 096 ( 2014 ) .
g. aad _ et al . _ [ atlas collaboration ] , jhep * 1504 * , 117 ( 2015 ) .
s. chatrchyan _ et al . _
[ cms collaboration ] , jhep * 1405 * , 104 ( 2014 ) .
g. aad _ et al . _ [ atlas collaboration ] , jhep * 1501 * , 069 ( 2015 ) .
s. chatrchyan _ et al .
_ [ cms collaboration ] , phys .
d * 89 * , 012003 ( 2014 ) .
a. david _ et al .
_ [ lhc higgs cross section working group ] , arxiv:1209.0040 [ hep - ph ] ; and references therein . c.q .
geng and j.n .
ng , phys . rev .
d * 38 * , 2857 ( 1988 ) [ erratum - ibid .
d * 41 * , 1715 ( 1990 ) ] .
m. bona _ et al . _ [ utfit collaboration ] , phys .
lett . * 97 * , 151803 ( 2006 ) ; jhep * 0803 * , 049 ( 2008 ) ; see http://www.utfit.org/ for updates .
m. ciuchini , g. degrassi , p. gambino and g.f .
giudice , nucl .
b * 527 * , 21 ( 1998 ) .
a. crivellin , a. kokulu and c. greub , phys .
d * 87 * , 094031 ( 2013 ) . y. amhis _ et al .
_ [ heavy flavor averaging group ( hfag ) collaboration ] , arxiv:1412.7515 [ hep - ex ] ; see http://www.slac.stanford.edu/xorg/hfag for updates .
m. misiak _ et al .
_ , phys .
lett . * 98 * , 022002 ( 2007 ) . t. hermann , m. misiak and m. steinhauser , jhep * 1211 * , 036 ( 2012 ) .
s. fajfer , j.f .
kamenik , i. nisandzic and j. zupan , phys .
lett . * 109 * , 161801 ( 2012 ) .
a. crivellin , c. greub and a. kokulu , phys .
d * 86 * , 054014 ( 2012 ) .
lees _ et al .
_ [ babar collaboration ] , phys .
lett . * 109 * , 101802 ( 2012 ) . | a general two higgs doublet model ( 2hdm ) is adopted to study the signature of flavor changing neutral higgs ( fcnh ) decay @xmath0 , where @xmath1 could be a cp - even scalar ( @xmath2 ) or a cp - odd pseudoscalar ( @xmath3 ) .
measurement of the light 125 gev neutral higgs boson ( @xmath4 ) couplings at the large hadron collider ( lhc ) favor the decoupling limit or the alignment limit of a 2hdm , in which gauge boson and diagonal fermion couplings of @xmath4 approach standard model values . in such limit , fcnh couplings of @xmath4 are naturally suppressed by a small mixing parameter @xmath5 , while the off - diagonal couplings of heavier neutral scalars @xmath1 are sustained by @xmath6 .
we study physics background from dominant processes with realistic acceptance cuts and tagging efficiencies .
promising results are found for the lhc running at 13 or 14 tev collision energies . |
a fundamental challenge in hadronic physics is trying to understand the structure of mesons and baryons in terms of their quark - gluon constituents , as given by the underlying theory of the strong interaction .
this theory is known as quantum chromodynamics ( qcd ) .
form factors of hadrons play an important role in this description , because they provide information about the internal structure of the hadron .
one of the simplest hadronic systems available for study is the pion , whose valence structure is a bound state of a quark and an antiquark .
its electromagnetic structure is parameterized by a single form factor , 2 , which depends on @xmath4 , where @xmath5 is the four - momentum squared of the virtual photon .
@xmath6 is well determined up to values of @xmath0 of 0.28 gev@xmath1 by elastic @xmath7 scattering @xcite , from which the charge radius has been extracted .
determining 2 at larger values of @xmath0 requires the use of pion electroproduction from a nucleon target .
the longitudinal part of the cross section for pion electroproduction , , contains the pion exchange ( @xmath3-pole ) process , in which the virtual photon couples to a virtual pion inside the nucleon .
this process is expected to dominate at small values of the mandelstam variable @xmath8 , thus allowing for the determination of .
pion electroproduction data have previously been obtained for values of of 0.18 to 9.8 at the cambridge electron accelerator ( cea ) and at cornell @xcite , and at the deutsches elektronen - synchrotron ( desy ) @xcite .
most of the high @xmath0 data have come from experiments at cornell . in these experiments , @xmath6 was extracted from the longitudinal cross sections , which were isolated by subtracting a model of the transverse contribution from the unseparated cross sections .
pion electroproduction data were also obtained at desy @xcite for values of @xmath0 of 0.35 and 0.7 gev@xmath1 , and longitudinal ( l ) and transverse ( t ) cross sections were extracted using the rosenbluth l / t separation method . with the availability of the high - intensity , continuous electron beams and well - understood magnetic spectrometers at the thomas jefferson national accelerator facility ( jlab ) it became possible to determine l / t separated cross sections with high precision , and thus to study the pion form factor in the regime of = 0.5 - 3.0 .
in 1997 , high - precision pion electroproduction data for values of @xmath0 between 0.60 and 1.60 were acquired at jlab at a value of the invariant mass of the photon - nucleon system of @xmath2=1.95 gev .
the main results were published in @xcite , with an updated analysis of these data published in @xcite . in 2003 ,
the range @xmath0=1.60 - 2.45 was studied at a value @xmath2=2.22 gev @xcite . at each value of @xmath0 ,
cross sections were obtained at two different values of the virtual photon polarization , @xmath9 , allowing for the separation of the longitudinal and transverse components of the cross section .
the purpose of this work is to describe the experiment and analysis in detail and to present and discuss additional results .
the discussion has been split into two parts .
this paper describes the experiment and analysis , presents the measured cross sections , including the separation into the structure functions , along with a detailed discussion of the systematic uncertainties , and compares them with previous l / t separated data and with theoretical calculations for the cross section .
the paper immediately following @xcite discusses the determination of and presents the resulting values , including all uncertainties .
these values are then compared to various theoretical predictions .
this division was chosen to separate the determination of the cross section , with its various experimental issues , from the extraction of from the measured cross sections , which is model dependent .
if more advanced or other models will become available , new values for may be extracted from the same cross sections .
this paper is organized as follows : in the next section the basic formalism of the reaction is presented . in section [ sec : experiment ] the experiment performed at jlab is described , including the experimental set - up and the calibrations of the spectrometers . the data analysis and a discussion of the various efficiencies that play a role are presented in section [ sec : analysis ] .
the determination of the unseparated cross sections and the separation of these cross section into the four different structure functions is described in section [ sec : xsecdet ] .
the results are presented in section [ sec : results ] .
the global features of the separated cross sections are discussed and a comparison is made with the results of theoretical calculations . in this discussion the data from @xcite are also included . the paper is concluded with a short summary .
the kinematics of the reaction are displayed in fig . [
fig : eepi_kinematics ] .
the three - momentum vectors of the incoming and scattered electrons are denoted by @xmath10 and @xmath11 , respectively .
together they define the scattering plane .
the corresponding four - momenta are @xmath12 and @xmath13 .
the electron scattering angle is denoted by @xmath14 .
the four - momentum of the transferred virtual photon , @xmath15 , is given by @xmath16 .
as usual , the variable is defined as the negative of the transferred four - momentum squared : @xmath17 .
the three - momentum @xmath18 and the three - momentum vector of the pion @xmath19 together define the reaction plane .
the angle between the scattering plane and the reaction plane is denoted by @xmath20 , while the angle ( in the lab system ) between @xmath19 and @xmath18 is @xmath21 .
the missing energy and missing momentum are defined as : @xmath22 the missing mass of the recoil system can then be expressed as @xmath23 . in the case of the reaction the missing mass is given by the neutron mass @xmath24 .
the reaction can conveniently be described using three lorentz invariants .
in addition to @xmath0 , we use the invariant mass of the virtual photon - nucleon system , @xmath2 , which can be expressed as @xmath25 , where @xmath26 is the proton mass , and the mandelstam variable @xmath27 .
the latter can be expanded into latexmath:[\[\label{eq : def_t } t=(e_\pi-\omega)^2 - |\mathbf{p}_\pi|^2 - |\mathbf{q}|^2 + 2 \hspace{1 mm } reaction @xmath3 is always negative .
the minimum value @xmath29 of @xmath8 is reached for @xmath30 .
the minimum value of @xmath8 increases for increasing values of and decreasing values of @xmath2 .
describing the incoming and outgoing electrons by plane waves , the cross section for the reaction can be written in the one - photon exchange approximation as @xmath31 here @xmath32 is the virtual photon flux factor @xmath33 where @xmath34 is the fine structure constant , the factor @xmath35 is the equivalent real - photon energy , which is the laboratory energy a real photon would need to produce a system with invariant mass @xmath2 , and @xmath36 is the polarization of the virtual photon . the two - fold differential cross section can be written in terms of an invariant cross section as @xmath37 where @xmath38 is the jacobian for the transformation from @xmath39 to @xmath40 .
the cross section can be decomposed into four structure functions corresponding to the polarization states of the virtual photon , a longitudinal one ( l ) , a transverse one ( t ) , and two interference terms ( lt and tt ) @xcite : @xmath41 where the @xmath42 depend on , @xmath2 and @xmath3 .
the dependence of the interference structure functions on @xmath21 features the following leading order behavior @xcite : @xmath43 and @xmath44
. therefore the interference structure functions are zero in parallel kinematics ( @xmath30 ) , i.e. , at @xmath45 .
the four structure functions can be isolated if data are taken at different values of @xmath9 and @xmath20 , while @xmath2 , and @xmath3 are kept constant .
the photon polarization @xmath9 can be varied by changing the electron energy and scattering angle ( the so - called `` rosenbluth- '' or l / t - separation ) .
the angle @xmath20 can be varied by measuring the pion left and right of the @xmath18-vector , and out of the scattering plane .
the two @xmath6 experiments were carried out in 1997 ( fpi - 1 @xcite ) and 2003 ( fpi-2 @xcite ) in hall c at jlab .
the unpolarized electron beam from the cebaf accelerator was incident on a liquid hydrogen target .
two moderate acceptance , magnetic focusing spectrometers were used to detect the particles of interest .
the produced charged pions were detected in the high momentum spectrometer ( hms ) , while the scattered electrons were detected in the short orbit spectrometer ( sos ) .
the choice of kinematics for the two experiments was based on maximizing the range in @xmath0 for a value of the invariant mass @xmath2 above the resonance region . for each @xmath0 ,
data were taken at two values of the virtual photon polarization , @xmath9 , with @xmath46 typically @xmath47 0.25 .
this allowed for a separation of the longitudinal , transverse , lt , and tt cross sections .
constraints on the kinematics were imposed by the maximum available electron energy , the maximum central momentum of the sos , and the minimum hms angle .
the central kinematics for the two experiments are given in table [ tab : kinematics ] .
in parallel kinematics , , when the pion spectrometer is situated in the direction of the @xmath18 vector , the acceptances of the two spectrometers do not provide a uniform coverage in @xmath20 .
thus , to attain full coverage in @xmath48 , additional data were taken with the hms at a slightly smaller and larger angle compared to the central angle for the high @xmath9 settings . at low @xmath9
only the larger angle setting was possible .
ccccccccc @xmath0 & @xmath2 & @xmath49 & @xmath50 & @xmath14 & @xmath51 & @xmath52 & @xmath53 & + ( gev@xmath1 ) & ( gev ) & ( gev@xmath1 ) & ( gev ) & ( deg ) & ( gev ) & ( deg ) & ( gev ) & + + 0.60 & 1.95 & 0.030 & 2.445 & 38.40 & 0.567 & 9.99 denotes the angle of the momentum transfer @xmath54 .
the actual hms angle was 10.49@xmath55 . ] & 1.856 & 0.37 + 0.60 & 1.95 & 0.030 & 3.548 & 18.31 & 1.670 & 14.97 & 1.856 & 0.74 + 0.75 & 1.95 & 0.044 & 2.673 & 36.50 & 0.715 & 11.46 & 1.929 & 0.43 + 0.75 & 1.95 & 0.044 & 3.548 & 21.01 & 1.590 & 15.45 & 1.929 & 0.70 + 1.00 & 1.95 & 0.071 & 2.673 & 47.26 & 0.582 & 10.63 & 2.048 & 0.33 + 1.00 & 1.95 & 0.071 & 3.548 & 25.41 & 1.457 & 15.65 & 2.048 & 0.65 + 1.60 & 1.95 & 0.150 & 3.005 & 56.49 & 0.594 & 10.49 & 2.326 & 0.27 + 1.60 & 1.95 & 0.150 & 4.045 & 28.48 & 1.634 & 16.63 & 2.326 & 0.63 + + 1.60 & 2.22 & 0.093 & 3.779 & 43.10 & 0.786 & 9.53 . ] & 2.931 & 0.33 + 1.60 & 2.22 & 0.093 & 4.709 & 43.10 & 1.650 & 12.54 & 2.931 & 0.58 + 2.45 & 2.22 & 0.189 & 4.210 & 51.48 & 0.771 & 9.19 . ] & 3.336 & 0.27 + 2.45 & 2.22 & 0.189 & 5.246 & 29.43 & 1.740 & 12.20 & 3.336 & 0.54 + the experiments made use of the unpolarized , continuous wave ( cw , 100% duty factor ) electron beam provided by the jlab accelerator @xcite .
the beam has a microstructure that helps in the identification of coincident events , which is further described in section [ sec : analysis ] .
beam currents were between 10 and 100 @xmath56a . in order to precisely determine the kinematics ,
the beam position and angle on target were monitored using beam position monitors ( bpm ) .
the accuracy of the position measurement was about 0.5 mm and about 0.2 mrad for the incident angle . in the fpi experiments ,
the beam current was measured by two beam current monitors ( bcm1 and bcm2 ) . to minimize drifts in the gain ,
both bcms are calibrated to an absolute reference .
the calibration is performed using an unser current monitor @xcite , which has an extremely stable gain , but suffers from large drifts in the offset on short time scales .
the run - to - run uncertainty in the current as measured by bcm1 and bcm2 was found to be about 0.2% at 100 @xmath56a .
adding the normalization uncertainty from the unser monitor , which is estimated to be 0.4% , results in an absolute uncertainty for the charge measurement of 0.5% . a more detailed description of the beam current monitors can be found in @xcite in order to reduce local density reductions of the liquid targets ,
the beam was rastered using a pair of fast raster magnets to a [email protected] mm@xmath1 pattern during fpi-1 and to a 2@xmath572 mm@xmath1 profile during fpi-2 .
the raster position was recorded event by event .
a more detailed description of the fast raster system can be found in @xcite the energy of the electron beam in hall c is measured using the deflection of the electron beam in a known magnetic field in the hall c arc . including the uncertainty in the field integral and the angular uncertainty , the beam energy can be determined with a precision of @xmath58 @xmath59 @xmath60 .
a detailed description of the beam energy measurement using the arc method is available in @xcite .
the two fpi experiments used a three - loop cryogenic target stack , mounted together with a special optics target assembly .
the cryogenic targets use the same coolant supply and are cooled on the cryotarget ladder simultaneously .
two different cryogenic target cell types were used . in the fpi-1 experiment ,
a 4.5 cm long cylindrical cell with the axis mounted horizontally and parallel to the beam direction was used ( horizontal flow `` beer can '' design ) . in the fpi-2 experiment , a 4.0 cm diameter cylindrical cell with vertical axis (
vertical flow `` tuna can '' design ) was used .
the cell walls are made from aluminum alloy t6061 with a thickness of 0.0127 cm ( the beer can front wall is half as thick ) .
the cryogenic targets are typically kept at a nominal operating temperature about 2 k below the boiling point .
the hydrogen target was kept at a temperature of 19 k , giving a density of 0.0723 @xmath61 0.0005
g/@xmath62 @xcite .
cell temperatures were kept constant to within 100 mk during the experiment .
since the uncertainty in temperature gives a negligible contribution , the uncertainty in the target density is completely due to the equation of state . _
( color online ) schematic view of the optics target assembly `` quintar '' ( not to scale)._,width=336 ] the optics target assembly was mounted beneath the cryogenic target ladder .
it consists of five carbon foils ( quintar ) and two aluminum foils .
a schematic design of the quintar is shown in fig .
[ fig : quintar ] . by moving the target stack vertically
, the five targets can be moved into the beam individually or simultaneously .
the solid carbon foils are used with the beam incident on two or five ( `` quintar '' ) foils simultaneously for the purpose of calibrating the vertex position ( @xmath63 ) along the beam direction ( see section [ sec : optics ] ) . during fpi-1 , the quintar @xmath63-positions ( relative to the nominal target center ) were @[email protected] cm , @xmath613.0 cm , 0 cm , while in fpi-2 they were @[email protected] cm , @xmath613.8 cm , 0 cm .
the two aluminum foils situated at [email protected] cm ( fpi-1 ) or [email protected] cm ( fpi - 2 ) constitute the `` dummy target '' used to measure the contribution of the aluminum cell wall to the cryotarget yields .
the material of the aluminum dummy targets is al - t7075 ( @xmath64=2.795 g / cm@xmath1 ) , a higher strength alloy .
the dummy target foils are approximately seven times thicker than the cryotarget cell walls .
further details on the mechanical aspects of the cryotargets can be found in @xcite . a schematic overhead view of the hall c spectrometers is shown in fig . [ fig_hallc_spectrometers ] . both spectrometers have a relatively large momentum and solid angle acceptance and
are equipped with similar and highly versatile detector packages .
the short orbit spectrometer ( sos ) , which was optimized for the detection of short - lived particles , has a relatively short flight path of about 7.4 m and a maximum central momentum of 1.74 gev / c .
the high momentum spectrometer ( hms ) has a 26 m pathlength and a maximum central momentum of 7.5 gev / c . _
( color online ) schematic view of the hall c spectrometers with the target and beamline._,width=336 ] the hms consists of three superconducting quadrupole magnets and a 25@xmath66 vertical - bend dipole magnet used in a point - to - point tune for the central ray .
the momentum acceptance of the hms is about @xmath61 10% .
all magnets are mounted on a common carriage , which can be moved on rails around a rigidly mounted central bearing .
a detailed description of the spectrometer hardware is given in @xcite .
the design specifications are given in table [ tab : hmsspec ] .
.[tab : hmsspec ] _ the nominal specifications for the high momentum spectrometer ( hms ) and the short orbit spectrometer ( sos ) . _ [ cols="<,<,<",options="header " , ]
the separated cross sections are listed in table [ tab : xsech ] and shown in figs . [
fig : xsec1h ] ( , ) and [ fig : xsec2h ] ( , ) . in the following subsections , the global dependences of and will be reviewed , and the data compared to model calculations for the reaction . _ the separated cross sections , ( circles ) and ( squares ) at central values of @xmath0=0.60 , 0.75 , 1.00 , 1.60 gev@xmath1 ( @xmath2=1.95 gev ) , and @xmath0=1.60 , 2.45 gev@xmath1 ( @xmath2=2.22 gev ) .
the values of @xmath67 and @xmath68 are different for each @xmath8-bin . the error bars for indicate the statistical and uncorrelated systematic uncertainties in both @xmath9 and @xmath8 combined in quadrature .
the error band denotes the correlated part of the systematic uncertainty by which all data points move collectively for .
the error bars for represent the total uncertainty .
the curves denote regge calculations ( vgl , @xcite ) for ( solid line ) and ( dashed line ) for @xmath69=0.462 gev@xmath1 and @xmath70=1.5 gev@xmath1 .
also shown is a calculation for ( dashed - dotted line ) using a gpd model @xcite including power corrections._,width=336 ] _ the interference terms , ( circles ) and ( squares ) at central values of @xmath0=0.60 , 0.75 , 1.00 , 1.60 gev@xmath1 ( @xmath2=1.95 gev ) , and @xmath0=1.60 , 2.45 gev@xmath1 ( @xmath2=2.22 gev ) .
the curves denote regge calculations ( vgl , @xcite ) for ( solid line ) and ( dashed line ) with @xmath69=0.462 gev@xmath1 and @xmath70=1.5 gev@xmath1._,width=336 ] @xmath67 & @xmath8 & & & & + ( gev@xmath1 ) & ( gev ) & ( gev@xmath1 ) & ( @xmath56b / gev@xmath1 ) & ( @xmath56b / gev@xmath1 ) & ( @xmath56b / gev@xmath1 ) & ( @xmath56b / gev@xmath1 ) + + 0.526 & 1.983 & 0.026 & 31.360 @xmath61 1.602 , 1.927 & 8.672 @xmath61 1.241 & 1.982 @xmath61 0.491 & -0.187 @xmath61 0.71 + 0.576 & 1.956 & 0.038 &
24.410 @xmath61 1.119 , 1.774 & 10.660 @xmath61 1.081 & 1.581 @xmath61 0.288 & -2.034 @xmath61 0.427 + 0.612 & 1.942 & 0.050 & 20.240 @xmath61 1.044 , 1.583 & 10.520 @xmath61 1.000 & 0.409 @xmath61 0.255 & -3.811 @xmath61 0.406 + 0.631 & 1.934 & 0.062 & 14.870 @xmath61 1.155 , 1.366 & 10.820 @xmath61 0.992 & -0.745 @xmath61 0.302 & -5.117 @xmath61 0.524 + 0.646 & 1.929 & 0.074 &
11.230 @xmath61 1.469 , 1.210 & 10.770 @xmath61 1.097 & -1.020 @xmath61 0.390 & -6.966 @xmath61 0.816 + + 0.660 & 1.992 & 0.037 & 20.600
@xmath61 1.976 , 1.895 & 9.812 @xmath61 1.532 & 0.565 @xmath61 0.393 & 0.208 @xmath61 0.623 + 0.707 & 1.961 & 0.051 & 16.280 @xmath61 1.509 , 1.788 & 10.440 @xmath61 1.344 & 1.135 @xmath61 0.268 & -0.454 @xmath61 0.420 + 0.753 & 1.943 & 0.065 & 14.990 @xmath61 1.270 , 1.573 & 8.580 @xmath61 1.150 & 0.618 @xmath61 0.206 & -1.910 @xmath61 0.378 + 0.781 & 1.930 & 0.079 & 11.170 @xmath61 1.214 , 1.416 & 9.084 @xmath61 1.091 & -0.409 @xmath61 0.197 & -2.547 @xmath61 0.419 + 0.794 & 1.926 & 0.093 & 9.949 @xmath61 1.376 , 1.277 & 8.267 @xmath61 1.110 & -0.827
@xmath61 0.220 & -3.474 @xmath61 0.534 + + 0.877 & 1.999 & 0.060 & 14.280 @xmath61 1.157 , 1.103 & 7.084 @xmath61 0.791 & 1.049 @xmath61 0.294 & -0.794 @xmath61 0.474 + 0.945 & 1.970 & 0.080 & 11.840 @xmath61 0.887 , 0.978 & 6.526 @xmath61 0.657 & 1.339 @xmath61 0.205 & -1.584 @xmath61 0.329 + 1.010 & 1.943 & 0.100 & 9.732 @xmath61 0.773 , 0.837 & 5.656 @xmath61 0.572 & 0.719 @xmath61 0.164 & -0.582 @xmath61 0.302 + 1.050 & 1.926 & 0.120 & 7.116 @xmath61 0.789 , 0.747 & 5.926 @xmath61 0.570 & 0.331 @xmath61 0.158 & -1.277 @xmath61 0.360 + 1.067 & 1.921 & 0.140 & 4.207 @xmath61 1.012 , 0.612 & 5.802 @xmath61 0.656 & 0.087 @xmath61 0.187 & -0.458 @xmath61 0.471 + + 1.455 & 2.001 & 0.135 & 5.618 @xmath61 0.431 , 0.442 & 3.613 @xmath61 0.294 & 0.537 @xmath61 0.125 & -0.022 @xmath61
0.200 + 1.532 & 1.975 & 0.165 & 4.378 @xmath61 0.356 , 0.390 & 3.507 @xmath61 0.257 & 0.356 @xmath61 0.095 & -0.268 @xmath61 0.156 + 1.610 & 1.944 & 0.195 & 3.191 @xmath61 0.322 , 0.351 & 3.528 @xmath61 0.241 & 0.143 @xmath61 0.081 & -0.126 @xmath61 0.153 + 1.664 & 1.924 & 0.225 & 2.357 @xmath61 0.313 , 0.310 & 3.354 @xmath61 0.228 & -0.028 @xmath61 0.076 & -0.241 @xmath61 0.167 + 1.702 & 1.911 & 0.255 & 2.563 @xmath61 0.356 , 0.268 & 2.542 @xmath61 0.227 & -0.100 @xmath61 0.085 &
-0.083 @xmath61 0.196 + + 1.416 & 2.274 & 0.079 & 6.060
@xmath61 0.464 , 0.564 & 2.802 @xmath61 0.27 & 0.195 @xmath61 0.073 & -0.346 @xmath61 0.177 + 1.513 & 2.242 & 0.112 & 4.470 @xmath61 0.342 , 0.457 & 2.459 @xmath61 0.21 & 0.370 @xmath61 0.081 & -0.500 @xmath61 0.169 + 1.593 & 2.213 & 0.139 & 3.661 @xmath61 0.303 , 0.397 & 2.198 @xmath61 0.19 & 0.334 @xmath61 0.089 & -0.481 @xmath61 0.139 + 1.667 & 2.187 & 0.166 & 2.975 @xmath61 0.294 , 0.358 & 2.124 @xmath61 0.18 & 0.235 @xmath61 0.081 & -0.469 @xmath61 0.139 + 1.763 & 2.153 & 0.215 & 1.630
@xmath61 0.292 , 0.315 & 2.369 @xmath61 0.19 & 0.247 @xmath61 0.087 & -0.823 @xmath61 0.300 + + 2.215 & 2.308 & 0.145 & 2.078 @xmath61 0.180 , 0.229 & 1.635 @xmath61 0.11 & 0.217 @xmath61 0.034 & -0.060 @xmath61 0.163 + 2.279 & 2.264 & 0.202 & 1.365 @xmath61 0.125 , 0.179 & 1.395 @xmath61 0.08 & 0.168 @xmath61 0.025 & -0.199 @xmath61 0.066 + 2.411 & 2.223 & 0.245 & 0.980
@xmath61 0.110 , 0.159 & 1.337 @xmath61 0.08 & 0.159 @xmath61 0.023 & -0.163 @xmath61 0.045 + 2.539 & 2.181 & 0.288 & 0.786 @xmath61 0.114 , 0.150 & 1.304 @xmath61 0.08 & 0.128 @xmath61 0.018 & -0.187 @xmath61 0.120 + 2.703 & 2.127 & 0.365 & 0.564 @xmath61 0.123 , 0.137 & 1.240 @xmath61 0.08 & 0.161 @xmath61 0.020 & -0.234 @xmath61 0.109 + at all values of @xmath0 , the longitudinal cross section shows the characteristic fall - off with @xmath8 due to the pion pole .
its magnitude ( at constant @xmath2 ) drops with increasing @xmath0 , mainly because the value of @xmath71 increases with .
the transverse cross section is largely flat with @xmath8 , while its magnitude drops with increasing .
the interference term is rather small , while the value of , which clearly shows the behavior of going to 0 at @xmath45 , drops rapidly with increasing .
with the availability of our precision separated cross sections over an extended kinematic range , it is interesting to look into the global dependences of the longitudinal and transverse cross sections upon @xmath2 , @xmath0 and @xmath3 .
a similar study was done in @xcite with the more limited data then available . for the purpose of this study , both our cross sections and those of @xcite were used . for , the photoproduction data of @xcite were also used .
the @xmath2 dependences of the earlier and data were observed @xcite to follow @xmath72 , where @xmath73 is the nucleon mass .
our @xmath0=1.60 data at @xmath2=1.95 , 2.22 gev are consistent with this within about 10% . _
( color online ) the @xmath3 dependence of the longitudinal @xmath74 cross section .
the data from @xcite are scaled in @xmath2 and @xmath0 ( see the text ) to common values of @xmath2=2.19 gev and @xmath750.7 ._,width=336 ] because is dominated by the pion - pole contribution , its @xmath0-dependence is largely given by @xmath76 .
[ fig - sigl - tdep - all ] shows the results for , where all cross sections have been scaled to @xmath2=2.19 gev according to @xmath72 , and to @xmath0=0.70 gev@xmath1 using the factor @xmath76 , where @xmath6 was assumed to follow the monopole form @xmath77 .
although overall the data follow an almost exponential @xmath3 dependence , upon close inspection it is observed that at constant the data deviate from that curve , i.e. , the and @xmath3-dependences do not factorize completely , and at both high and low @xmath8 deviations from a pure exponential are observed .
fitting the data with an exponential @xmath78 results in a slope parameter @xmath79=10.5 @xmath61 1.8 gev@xmath80 , and a normalization factor @xmath81=19.0 @xmath61 2.0 .
such a form describes all data within about 50% .
no simple prediction exists for the -dependence of .
[ fig - sigt - q2dep - all ] shows the @xmath0 dependence of the data at @xmath8=0.08 and 0.2 gev@xmath1 , scaled to @xmath2=2.19 gev .
the data show a clear dependence on @xmath0 , which is reasonably - well described by a factor of the form @xmath82 . _
( color online ) the @xmath0 dependence of the transverse @xmath74 cross section at @xmath8=0.08 and 0.2 gev@xmath1 .
the cross sections are scaled to @xmath2=2.19 gev .
the photoproduction point is from @xcite .
the curve indicates a parameterization for of the form @xmath82._,width=336 ] fig .
[ fig - sigt - tdep - all ] displays the electro- and photo - production data scaled to @xmath2=2.19 gev using the functional form @xmath72 , and to @xmath0=0.7 gev@xmath1 according to @xmath82 , where @[email protected] and @[email protected] .
an exponential in @xmath3 analogous to the form used for describes the photoproduction , and electroproduction data from fpi-1 and fpi-2 to within 30% , while the desy data are overpredicted by a factor of about two .
an exponential fit results in a slope parameter of @xmath79=2.3 @xmath61 1.5 gev@xmath80 , and a normalization factor of @xmath81=5.4 @xmath61 1.4 .
though the slope is less steep than for it is clear that is not independent of @xmath3 and @xmath0 in these kinematics .
this is different from the conclusions of @xcite . _
( color online ) the @xmath3 dependence of the transverse @xmath74 cross section .
the data from @xcite are scaled to common values of @xmath85 gev and @xmath0=0.7 gev@xmath1 ( see the text)._,width=336 ] fig .
[ fig - siglt - sigtt - tdep - all ] shows the @xmath86 dependence of and .
the data are scaled to common values of @xmath2=2.19 gev and @xmath0=0.70 gev@xmath1 using a factor @xmath87 .
the data were scaled in @xmath0 analogous to . in both cases ,
no overall trend could be identified due to the large scatter of the data . _
( color online ) the @xmath86 dependence of the interference terms .
both are scaled to common values of @xmath2=2.19 gev and @xmath0=0.70 gev@xmath1 , see the text .
the plotting symbols are the same as in fig .
[ fig - sigt - tdep - all]._,width=336 ] in @xcite , vanderhaeghen , guidal and laget ( vgl ) developed a regge model for pion production , in which the pole - like propagators of born term models are replaced with regge propagators , i.e. , the interaction is effectively described by the exchange of a family of particles with the same quantum numbers instead of a single particle . for forward pion production ,
the dominant exchanges are the @xmath88 and @xmath64 trajectories .
these determine the @xmath3-dependence of the cross section without the use of a @xmath89 factor .
since the @xmath3-channel @xmath88 diagram is by itself not gauge invariant , in the vgl model the @xmath90-channel ( for @xmath74 production ) or @xmath91-channel ( for @xmath92 production ) nucleon exchange diagram was also reggeized , to ensure gauge invariance of their sum .
the model is parameter free , as the coupling constants at the vertices ( such as @xmath93 ) are well determined by precise studies and analyses in the resonance region .
the vgl model was first applied to pion photoproduction @xcite .
the model gave a good and consistent description of the @xmath2- and @xmath3-dependences of the available @xmath74 and @xmath92 photoproduction data including the spin asymmetries .
the fact that both the @xmath88 ( unnatural - parity ) and the @xmath64 ( natural - parity ) trajectories are incorporated in the model proved to be essential to explain the different behaviors of @xmath74 and @xmath92 photoproduction . in @xcite
, the model was extended to pion electroproduction .
as the @xmath88- and @xmath64-exchange amplitudes are separately gauge invariant , two different electromagnetic form factors were introduced for the @xmath88 and @xmath64 exchanges without violating the gauge invariance of the model . in both cases ,
monopole forms are used .
form factors of monopole type were taken for the @xmath88 and @xmath64 exchanges : @xmath94^{-1}.\ ] ] the model gave a good description of @xmath74 electroproduction data out to large values of @xmath8 at @xmath2-values of 2.15 and 3.1 gev for = 1.2 @xcite , and of the @xmath95 ratio at @xmath85 gev , = 0.7 and 1.35 @xcite .
the vgl model is compared to our electroproduction data in figs .
[ fig : xsec1h ] and [ fig : xsec2h ] .
the vgl cross sections were evaluated at the same @xmath96 and @xmath97 values as the data . over the range of @xmath8 covered by this work ,
is completely determined by the @xmath88 trajectory , while , and are also sensitive to the @xmath64 exchange contribution .
comparison of the model calculations to previous data gave a value for of about @xmath98 .
here , calculations with a common value of = 0.462 are shown .
this is the same value as is used in @xcite .
the value of is more poorly known . here ,
calculations with = 1.500 are shown , where this upper value is determined from the application of the vgl model to kaon electroproduction @xcite . with a single value of = 0.462
, the vgl model does an overall good job of describing the magnitude , and @xmath3 , @xmath2 and @xmath0-dependences of our data .
however , as shown in fig .
[ fig : xsec1h ] , the description of the @xmath3-dependence is not as good for @xmath99 , @xmath100 gev , where the model prediction is too flat in comparison to the experimental data .
the model also strongly underestimates , almost independent of the value of , and this underestimation appears to grow with , the fall - off of the data with being less than that of the model .
this deficiency is also reflected in a too - small prediction for .
please note that vgl s definition of differs from ours by a minus sign , which has been included here .
the calculations at @xmath101 gev are generally satisfactory , but the agreement with the data is much worse at the lower value @xmath100 gev , the data getting smaller or even becoming negative at larger values of @xmath8 .
recently , the vgl model was extended @xcite by including , apart from a slightly different way to handle the gauge invariance , a hard scattering between the virtual photon and a quark , followed by hadronization of the system into a pion plus residual nucleon . with plausible assumptions a good description of
was obtained , with no influence on .
a more recent development is the effective lagrangian model of faessler , gutsche , lyubovitskij and obukhovsky ( fglo , @xcite ) .
this is a modified born term model , in which an effective lagrangian is used to describe nucleon , pion , @xmath64 , and photon degrees of freedom .
the ( combined ) effect of @xmath90- and @xmath91-channel contributions , which interferes with the pion @xmath3-pole , is modeled using a constituent quark model .
the authors pay special attention to the role of the @xmath64 meson in @xmath74 electroproduction and show that the @xmath64 @xmath3-pole contribution is very important for obtaining a good description of the magnitude of .
when comparing vector and tensor representations of the @xmath64 contribution , the latter was found to give better results . unlike the vgl model , the cross section depends here also on the @xmath64 exchange , because of the interference of the @xmath88 and tensor @xmath64 exchange contributions .
the model contains a few free parameters , such as the renormalization constant of the kroll - ruderman contact term used to model the @xmath102-channel , and @xmath3-dependent strong meson - nucleon vertices , which are parameterized in monopole form , as are the electromagnetic form factors .
the corresponding parameters were adjusted so as to give overall good agreement with our and data .
_ separated @xmath74 electroproduction cross sections , ( circles ) and ( squares ) at central values of @xmath0=0.60 , 0.75 , 1.00 , 1.60 gev@xmath1 ( @xmath2=1.95 gev ) , and @xmath0=1.60 , 2.45 gev@xmath1 ( @xmath2=2.22 gev ) in comparison to the predictions of the vgl regge @xcite ( solid line ) and to the fglo effective lagrangian @xcite ( dashed line ) model . a common value of = 0.462 is used .
note that the average values of @xmath2 and @xmath0 are different for each @xmath8-bin .
the error bars denote statistical and @xmath3 uncorrelated systematic uncertainties combined in quadrature .
in addition , there is a @xmath3 and @xmath9 correlated systematic uncertainty of 4 - 6% , by which all data points move collectively._,width=336 ] the fglo model calculation is compared to our data in fig .
[ fig : xsec_obu ] . a common value of = 0.462 is used throughout .
the other model parameters were fixed at the values assigned by the authors .
generally , the agreement of the fglo model with the data is rather good , but the model gives a too - flat @xmath3-dependence at = 0.60 , @xmath100 gev . while on average the model calculation is in agreement with the data , it fails to describe the @xmath0- and @xmath2-dependences .
for example , the model under - predicts the = 1.60 , @xmath100 gev data by about a factor of two , while those at = 1.60 , @xmath101 gev are reproduced , and the = 2.45 , @xmath101 gev data under - predicted again by 20 - 60% .
no calculations are available for the interference cross sections .
vanderhaeghen , guichon , and guidal ( vgg ) @xcite have performed a calculation for using generalized parton distributions ( gpds ) .
this approach is based on a soft - hard factorization theorem @xcite . since the 1-gluon perturbative diagram severely underestimates the value of the pion form factor at the relevant @xmath0 , power corrections due to intrinsic transverse momenta and soft overlap contributions
were included in the calculation , thereby increasing the calculated cross sections by an order of magnitude .
the vgg gpd model is compared to our electroproduction data in fig .
[ fig : xsec1h ] .
the gpd calculation gives a rather good description of the @xmath3-dependence of the @xmath103 2 gev data , while the @xmath0-dependence is also described fairly well .
the determination of the onset of this regime remains one of the great challenges in contemporary gpd studies .
measurements to address this issue , approved for data taking after the completion of the jlab upgrade @xcite , may be expected to place a constraint on the value of @xmath0 for which one can reliably apply perturbative qcd concepts and extract generalized parton distributions .
precision data for the reaction were obtained in order to study the pion form factor in the regime = 0.5 - 3.0 .
the data were acquired at jlab making use of the high - intensity , continuous cebaf electron beams and the magnetic spectrometers in hall c. the cross sections were measured for values of the mandelstam variable @xmath3 close to its minimum value @xmath45 for ( central ) four - momentum transfers ranging from @xmath0=0.60 to 2.45 gev@xmath1 , at an invariant mass of the photon - nucleon system of @xmath2=1.95 or 2.22 gev .
since is to be determined from the longitudinal part , , of the cross section , the measured cross sections were decomposed into the four structure functions , , , and at every @xmath0 .
this required measuring the cross section at two values of the virtual photon polarization , @xmath9 , and as function of the azimuthal angle @xmath104 of the produced pion . in the analysis ,
a monte carlo simulation of the whole experimental setup was used .
the simulation included a model cross section fitted to the data , and thus allowed for accurate acceptance corrections .
good control of the systematic uncertainty is extremely important in l / t separations , as the error bars are inflated by @xmath46 .
therefore , all parts of the experimental setup and the analysis procedures were carefully inspected and calibrated .
this included the optical properties of the spectrometers , the tracking and particle identification methods , and the various efficiencies .
as a result , the total systematic uncertainty of the unseparated cross sections could be reduced to a point - to - point uncertainty below 2% , plus a scale uncertainty of less than 3.5% .
the final separated cross sections have a total uncertainty ( statistical plus systematic ) between 8 and 15% .
the longitudinal cross section shows the characteristic fall - off with @xmath8 due to the pion pole , and largely behaves as function of according to @xmath105 .
the transverse cross section depends only little on @xmath3 , but our results indicate a clear dependence on , which , including photoproduction data , can be described as @xmath106 .
this is different from what was concluded from earlier electroproduction results at desy .
the interference term is rather small , while the value of drops fast with increasing .
the separated cross sections were compared with the results of model calculations for the reaction , which use regge trajectories , effective lagrangians , or generalized parton distributions .
they all provide a fair to good description of the longitudinal cross section .
the description of the transverse cross section is much worse , however .
the regge model strongly underpredicts , while the lagrangian model yields good agreement at some values of @xmath2 and , but fails when either @xmath2 or is varied . clearly , more theoretical work has to be done to understand the behavior of ( and also of the interference structure functions and ) of the reaction as function of @xmath2 , and @xmath3
the authors would like to thank drs .
guidal , laget , and vanderhaeghen for stimulating discussions and for modifying their computer program for our needs .
we would also like to thank dr .
obukhovsky for supplying the result of their model calculations and for many informative discussions .
this work was supported in part by the u.s .
department of energy .
the southeastern universities research association ( sura ) operates the thomas jefferson national accelerator facility for the united states department of energy under contract de - ac05 - 84150 .
we acknowledge additional research grants from the u.s
. national science foundation , the natural sciences and engineering research council of canada ( nserc ) , nato , fom ( netherlands ) , and kosef ( south korea ) . v.
e. hart _ et al .
_ , _ mechanical design of the lampf medium resolution spectrometer and cebaf short orbit spectrometer magnets _ , 16 .
international linac conference ; 23 - 28 aug 1992 ; ottawa ( canada ) , published in ottawa 1992 , linear accelerator , vol . 2 737 - 739 .
d. j. abbott _ et al .
_ , _ the coda system and its performance in the first online experiments at cebaf _ , proceedings of the 1995 ieee conference on real - time computer applications in nuclear , particle and plasma physics ( may 1995 ) , pp . 147 - 151 | cross sections for the reaction were measured in hall c at thomas jefferson national accelerator facility ( jlab ) using the cebaf high - intensity , continous electron beam in order to determine the charged pion form factor .
data were taken for central four - momentum transfers ranging from @xmath0=0.60 to 2.45 gev@xmath1 at an invariant mass of the virtual photon - nucleon system of @xmath2=1.95 and 2.22 gev .
the measured cross sections were separated into the four structure functions , , , and .
the various parts of the experimental setup and the analysis steps are described in detail , including the calibrations and systematic studies , which were needed to obtain high precision results .
the different types of systematic uncertainties are also discussed .
the results for the separated cross sections as a function of the mandelstam variable @xmath3 at the different values of are presented .
some global features of the data are discussed , and the data are compared with the results of some model calculations for the reaction . |
aim of this contribution is to present a fully covariant model for describing the electromagnetic ( em ) decay of vector mesons ( vm s ) , both in light and heavy sectors . to this end ,
a simple analytic form for the bethe - salpeter ( bs ) amplitude of vm s is adopted in order to perform without any further approximation the calculations of the decay constants . moreover , with such an ansatz one can easily evaluate the so - called transverse momentum distribution of a constituent inside the vm ( see de melo et al @xcite for the pion case ) , that plays an essential role for fixing the value of the parameters appearing in our approach , and in turn for including some non perturbative inputs in our analytical ansatz . a possible form of the bs amplitude for an interacting @xmath0 system with @xmath1 ,
can be written as follows _ ( k , p)=s(k , m_1 ) _ vm(k , k
- p ) s(k - p , m_2 ) where @xmath2 is the dirac propagator of a constituent with mass @xmath3 , @xmath4 the four - momentum of a vm with mass @xmath5 , @xmath6 its polarization four - vector , @xmath7 the helicity , @xmath8 the dirac structure of the amplitude and @xmath9 the momentum dependence of the bs amplitude .
in particular , the adopted covariant form for the dirac structure is the familiar one ( transverse to @xmath4 ) , viz v^(p ) = m m+m_1+m_2 that in the limit of non interacting system leads to the melosh rotations for a @xmath10 system @xcite , as expected .
for the present preliminary calculations , the momentum dependence has the following simple form with single poles , viz _
vm(k , k - p)= n _ i=1,3 1 [ lambda ] where @xmath11 , @xmath12 are the free parameters of our ansatz ( to be determined as described below ) , @xmath13 the normalization factor , that can be derived by imposing the standard normalization for the bethe - salpeter amplitude , in impulse approximation ( i.e. with free propagators for the constituents ) . the form chosen for @xmath9 allows one both to implement the correct symmetry under the exchange of the quark momenta ( for equal mass constituents ) and to avoid any free propagation of the constituents ( cf the numerator in eq .
( [ lambda ] ) ) . to determine @xmath11 in eq .
( [ lambda ] ) , we first define the constituent transverse momentum distribution inside the vm , @xmath14 , along the same guidelines adopted by de melo et al .
@xcite for the pion , within a light - front hamiltonian dynamics approach . in a frame where @xmath15 , one has n(k_)=n_cp_q|q(2)^3 ^ 2 _ 0 ^ 2 d_k__0 ^ 1 d m^2_0(1- ) @xmath16 and @xmath17 is the valence wave function associated to a given bs amplitude , see , e.g. , huang and karmanov@xcite and frederico et al@xcite . in eq .
( [ nk ] ) , the probability @xmath18 of the valence component reads p_q|q=
n_c(2)^3 ^ 2 _ 0 ^ 1 d ( 1-)d*k*_m^2_0 |(,*k _ * ; m_r_i)|^2 finally , @xmath14 is normalized as : @xmath19 . in spite of the simple form assumed for @xmath20
, one can nicely fit the constituent transverse momentum distributions obtained within 3-d approaches , that i ) retain only the valence component of the vm s and ii ) are able to yield a reasonable description of the spectrum . in this work
we have extracted the parameters @xmath21 in eq .
( [ nk ] ) by fitting @xmath22 to the corresponding quantity obtained from i ) a harmonic oscillator model ( see , e.g. figs . 1 and 2 ) , ii ) the godfrey - isgur model @xcite and iii ) an adapted version of the model by salcedo et al @xcite ( ita model ) .
in order to evaluate the em decays constants , @xmath23 , we adopted a mandelstam - like formula@xcite ( see also de melo et al.@xcite ) ) .
the starting point is the _ macroscopic _ definition of @xmath23 , through the transition matrix element of the em current for a given neutral vm , viz 0|j^(0)| p,= f_v _ ^ [ eq : fv ] the decay constant @xmath23 is related to the em decay width as follows _
e^-= [ eq : gamma ] in our model , the transition matrix element in eq .
( [ eq : fv ] ) can be approximated _ microscopically _ _ la _ mandelstam through 0|j^(0)| p,= f_vm n_cn ( 2)^4d^4 k _ vm(k , k - p , m_1,m_2 ) ( k^2-m^2_1 + ) [ ( p - k)^2-m^2_2 + ] [ _ ( p ) v(p ) ( k -p + m_2)^(k+m_1 ) ] [ eq : mandel ] where @xmath24 with @xmath25 the quark charge . in tabs .
[ tab : hosci ] , [ tab : gi ] , [ tab : qcd - in ] , the preliminary results for both valence probability , @xmath18 , and em decay widths , @xmath26 are shown . even if a more refined evaluations are in progress , some comments are in order : i ) for the harmonic oscillator model the light meson decay widths can be reasonably well described ( in the light sector the confining interaction is quite relevant ) , while the @xmath27 one is largely underestimated ; ii ) for the godfrey - isgur model @xcite , the heavy sector is well reproduced , while the light sector is overestimated , and this appears correlated to the poor estimate of the valence probability ( work in progress suggests that an ansatz for the bs amplitude with a more rich structure substantially improves the comparison ) ; iii ) for the adapted version of the ita model @xcite the same pattern of the harmonic oscillator case has been found , even if more dynamical contents are present in this model .
._preliminary vm em decay widths within the harmonic oscillator model
. adopted quark masses : @xmath28 0.310 gev , @xmath29 0.460 gev , @xmath30 1.749 gev , @xmath31 5.068 gev , _ [ tab : hosci ] [ cols="^,^,^,^,^,^",options="header " , ] and @xmath32 vs the quark transverse momentum .
dashed line : harmonic oscillator model .
solid line : fit by using the analytic ansatz in eq .
( [ lambda])_,title="fig:",width=207 ] ._,title="fig:",width=207 ]
in this contribution we have presented the main ingredients of our model for evaluating both the em decay width of the ground states of vm s and the probability of the valence component of the state . in our fully covariant model ,
a simple , analytic ansatz for the bs amplitude is proposed , and the three parameters , @xmath21 , for each neutral vm , are determined through a fitting procedure , based on the transverse momentum distribution of a constituent inside a given vm , obtained within a light - front hamiltonian dynamics framework . from this first comparisons between our results and the experimental data , one could argue that two different regimes occur in the light ( @xmath33 , @xmath32 ) and in the heavy sector ( @xmath27 ) . from the theoretical side , indeed , the harmonic oscillator and the adapted ita@xcite models seem to better reproduce the light mesons ( cf tab . [
tab : hosci ] , [ tab : gi ] ) while for the heavy sector the godfrey - isgur model@xcite seems to work better ( cf . tab .
[ tab : qcd - in ] ) .
the work in progress will substantially improve the present calculations , in two respect : both introducing a more refined ansatz for the bs amplitude and extending our investigation to the em decay of the @xmath34 .
99 j. p. b. c. de melo , t. frederico , e. pace and g. salm , nucl .
phys . * a 707 * , 399 ( 2002 ) .
w. jaus , phys .
d 44 * , 2851 ( 1991 ) . t. frederico , e. pace , s. pisano and g. salm to be published .
dae sung hwang and v. a. karmanov , nucl .
* b 696 * , 413 ( 2004 ) . s. godfrey and n. isgur , phys . rev . *
d 32 * , 189 ( 1985 ) .
l. a. m. salcedo , j. p. b. c. de melo , d. hadjimichel and t. frederico , eur .
* 27 * 213 ( 2006 ) .
s. mandelstam , proc .
royal soc .
( london ) * a233 * , 248 ( 1956 ) .
j. p. b. c. de melo , t. frederico , e. pace and g. salm , phys .
d 73 * , 074013 ( 2006 ) . | a fully covariant model for describing the electromagnetic decay of vector mesons , both in light and in heavy sectors , is presented .
the main ingredients of our approach are i ) an ansatz for the bethe - salpeter vertex for vector mesons , and ii ) a mandelstam - like formula for the electromagnetic decay constant .
the free parameters of our approach are fixed through a comparison with the transverse momentum distribution obtained within a light - front hamiltonian dynamics framework with constituent quarks .
preliminary results for both the decays constants and the probability of the valence component are shown .
= 11.6pt |
the monster tower , also known as the semple tower , lies in the intersection of differential geometry , non - holonomic mechanics , singularity theory , and algebraic geometry .
cartan ( @xcite studied the diffeomorphism group action on jet spaces , which led to developments in the fields of goursat distributions and sub - riemannian geometry .
jean @xcite , luca and risler @xcite , li and respondek @xcite , pelletier and slayman @xcite , and others have studied models of various kinematic systems ( a car pulling @xmath0 trailers , motion of an articulated arm , @xmath0-bar systems ) .
montgomery and zhitomirskii @xcite studied the relationship with curve singularities ; later , so did we @xcite . and
we discovered in @xcite that algebraic geometers have long studied these objects under different names .
we have begun pursuing these connections @xcite and working with algebraic geometers to consolidate understanding and improve existing terminology and techniques @xcite . here , we study the rvt code for the tower , which is invariant under the action of the diffeomorphism group .
this is related to work on the classification problem studied by mormul @xcite , montgomery and zhitomirskii @xcite , the authors @xcite , and others . in the geometric theory of differential equations , we speculate that there may be some interesting connections between the singularity theory of the monster tower and the general monge problem for underdetermined systems of ordinary differential equations with an arbitrary number of degrees of freedom . in @xcite ,
the authors derive sufficient conditions , in terms of truncated multi - flag systems , for the existence of a monge - cartan parametrization of the general solution of such systems in the regular case . to our knowledge
, no connection has been made with the singular theory of multi - flags presented in this note .
similar undetermined systems of ordinary differential equations are common in geometric control theory when studying flat outputs of nonlinear control systems @xcite .
a detailed account of the geometry of differential equations in jet spaces can be found in @xcite , where symmetry methods from contact and symplectic geometry are used to solve non - trivial nonlinear partial and ordinary differential equations .
it remains to investigate the correspondence between finite jets of spatial curves and normal forms of special multi - flags .
one should explore the depth of the correspondence between arnold s a - d - e classification @xcite and the listing of normal forms of goursat multi - flags .
finally , current work with algebraic geometers @xcite extends and generalizes the results of this paper to the case of an @xmath0-dimensional base .
an interesting open question here concerns the existence of moduli in orbits of the action of the diffeomorphism group of the base space .
thus , it is apparent that this object is of interest to a variety of pure and applied mathematicians , and that it presents a wealth of interesting problems which have potential to shed light in surprising areas .
the subject begins with the study of goursat distributions , which are bracket - generating ( completely non - holonomic ) but slow growing .
cartan @xcite studied the model of the canonical contact distribution on the jet space @xmath1 .
all goursat distributions were believed to be equivalent to cartan s under the action of the diffeomorphism group until giaro , kumpera , and ruiz discovered the first singularity in 1978 @xcite .
jean @xcite studied the kinematic model of a car pulling @xmath2 trailers , a system which is locally universal for goursat distributions of corank @xmath3 .
he developed a geometric stratification given by regions in the configuration space of the model in terms of critical angles .
montgomery and zhitomirskii @xcite introduced the monster tower , a sequence of manifolds with distributions in which every goursat germ occurs , allowing for jean s strata to be recast in terms of positions of members of a canonical subflag of the goursat flag .
mormul @xcite labelled the strata from @xcite by words in the letters gst , which became the rvt code in @xcite . in @xcite ,
montgomery and zhitomirskii showed that goursat germs correspond to finite jets of legendrian curve germs , and that the rvt coding corresponds to several classical invariants in the singularity theory of planar curves .
they also gave complete spelling rules for the rvt code in this case .
these studies were all concerned with the monster tower whose base is @xmath4 . in @xcite , we generalized this to towers with base @xmath5 .
we also discovered that this object was known to algebraic geometers as the semple tower .
we also began the effort to generalize the rvt code , and find spelling rules to describe which words were admissible ( theorem [ thm1 ] ) .
that effort is completed here ( theorem [ thm2 ] ) .
the methods used in the present contribution were first developed in @xcite , in which we also classified points in the first four levels of the tower . here
, we will complete the spelling rules for base @xmath6 .
our techniques generalize to towers with base @xmath5 .
the diffeomorphism group of @xmath6 acts on the monster tower , and the rvt code is an invariant labeling of orbits .
note that the combinatorial data in the rvt code forces a finite number of inequivalent classes at each level of the tower , but there may be moduli within a given class ( see @xcite . in @xcite , we stated the following incomplete spelling rules , which followed from @xcite .
[ thm1 ] in the semple tower with base @xmath6 , every rvt code must begin with @xmath7 , and @xmath8 can not follow @xmath7 . here , we add the missing rules , yielding the complete description of realizable rvt codes .
our alphabet is the set @xmath9 .
note that these seven letters correspond precisely to the seven possibilities found in semple s original work @xcite .
we therefore have the following combinatorial description of the diffeomorphism group orbits . in the semple tower with base @xmath6
, there exists at point @xmath10 with rvt code @xmath11 if and only if the word @xmath11 satisfies :
1 . every word must begin with @xmath7 2 .
@xmath7 must be followed by @xmath7 or @xmath12 3 .
@xmath12 and @xmath8 must be followed by @xmath13 , or @xmath14 4 .
@xmath15 must be followed by @xmath16 , or @xmath17 5 .
@xmath18 , and @xmath17 can be followed by any letter . for example , the word @xmath19 is admissible , but @xmath20 breaks rules ( 1)(4 ) .
the following table [ tab : rules ] summarizes this theorem . .rvt
code spelling rules [ cols="<,^,>",options="header " , ] [ tab : critplanes ] the remainder of this paper will be dedicated to explaining why these possibilities are exhaustive .
cc point ( top left ) , a @xmath12 or @xmath8 point ( top right ) , a @xmath15 point ( bottom left ) , and an @xmath21 point ( bottom right).,title="fig:",scaledwidth=17.7% ] point ( top left ) , a @xmath12 or @xmath8 point ( top right ) , a @xmath15 point ( bottom left ) , and an @xmath21 point ( bottom right).,title="fig:",scaledwidth=25.0% ] + + point ( top left ) , a @xmath12 or @xmath8 point ( top right ) , a @xmath15 point ( bottom left ) , and an @xmath21 point ( bottom right).,title="fig:",scaledwidth=24.0% ] point ( top left ) , a @xmath12 or @xmath8 point ( top right ) , a @xmath15 point ( bottom left ) , and an @xmath21 point ( bottom right).,title="fig:",scaledwidth=25.0% ] we now describe the explicit method from which we derive all our results .
this will be applied to specific examples shortly .
the critical hyperplane method was implicit in parts of @xcite , made explicit in @xcite , exploited for the classification problem in @xcite , and is perfected here .
this gives a blueprint for characterizing all baby monsters and determining all spelling rules for the @xmath5 tower for any @xmath0 .
begin with an rvt code @xmath11 of a point @xmath22 .
we wish to understand which critical letters can be added to the end of the code ( one can always trivially add the letter @xmath7 ) . in order to do so , we must understand which critical planes lie above @xmath10 .
since critical planes live within baby monsters , we must determine which baby monsters are present , and for those which are , we seek to find the levels at which they were born .
we first determine the local kr - coordinate chart containing @xmath10 .
we can then describe the distribution @xmath23 in coordinates .
we then choose a critical plane @xmath24 , or @xmath15 , write it in coordinates as in definition [ letters ] , and trace the coordinate representations backwards , projecting down to lower levels of the tower , one at a time .
if at some level @xmath25 we find that both fiber coordinates @xmath26 and @xmath27 are non - vanishing , then our critical plane must arise as the prolongation of the vertical plane @xmath28 .
our critical plane therefore lives in the baby monster born at level @xmath25 , and is equal to @xmath29 .
this would confirm that the critical plane we chose indeed appears in @xmath23 .
if , however , we reach the base without finding such a baby monster , then the plane we chose can not exist in @xmath23 .
we can shorten the procedure of tracing each plane back to the base by using previously established configuration possibilities and proceeding inductively .
while this is not an algorithm in the strictest sense , it can theoretically determine which configurations are possible above any given point .
as one might suspect , this can at times become extremely tedious , and would not be particularly enlightening for the reader . for this reason , we will focus the remainder of the paper on a few specific examples to demonstrate the efficacy of the method for determining spelling rules , while skipping some of the routine verification that was required to complete our results .
it is obvious that the vertical plane @xmath12 appears above every point
it is just the tangent space to the fiber .
so in the method just described , we need only focus on whether or not @xmath8 and @xmath15 exist ( here , since we are concerned with the @xmath6-tower one immediately sees how this method generalizes to the @xmath5-tower ) . some of our results here ( those needed for the proof of theorem [ thm2 ] ) are summarized near the end of the paper in table [ tab : codes ] .
we will prove some of these relations here the rest are obtained by identical methods .
[ ex : rvl1 ] we continue investigating the case begun in example [ ex1 ] .
suppose @xmath30 has rvt code @xmath31 .
| p1.1 cm | p4.1 cm | c
| p3.1 cm | p3.3 cm | level @xmath25 & coordinates on @xmath32 & @xmath33 coords .
& critical planes in @xmath34 & rvt code of @xmath35 + @xmath36 & @xmath37 & n / a & none & n / a + @xmath38 & @xmath39 & @xmath40 $ ] & @xmath41 & @xmath42 + @xmath43 & @xmath44 & @xmath45 $ ] & @xmath46 & @xmath47 + @xmath48 & @xmath49 & @xmath50 $ ] & @xmath51 & @xmath52 + [ tab : ex2 ] .
the left side shows the birth of @xmath53 as the first prolongation of the vertical plane at level 2 .
the right side shows the birth of @xmath54 as the second prolongation of the vertical plane at level 1 .
these two baby monsters meet in @xmath55 , and their intersection is the distinguished line @xmath56 .
see example [ ex : rvl1].,scaledwidth=45.0% ] begin with the global coframe @xmath57 for @xmath58 .
our chart will be centered at @xmath59 where @xmath60 .
introduce affine fiber coordinates @xmath40 $ ] on @xmath61 . without loss of generality ,
assume @xmath62 . then @xmath40=[1 : \frac{dy}{dx } : \frac{dz}{dx}]$ ] .
now let @xmath63 so that @xmath64 since @xmath65 , we know @xmath66 with @xmath67 not both zero . without loss of generality , assume @xmath68 . then near this point we have @xmath45=[\frac{dx}{du_1 } : 1 : \frac{dv_1}{du_1}]$ ] .
this yields the affine coordinates @xmath69 so that @xmath70 now @xmath71 , so we want coordinate representations of the @xmath12 and @xmath8 planes in @xmath72 . according to definition [ letters ] , @xmath73 is given by @xmath74 and @xmath75 is given by @xmath76 , so we have @xmath77 and @xmath78 .
this forces our coordinates near @xmath79 to have the form @xmath50=[\frac{du_1}{dv_2 } : \frac{du_2}{dv_2 } : 1]$ ] .
this yields the affine coordinates @xmath80 so that @xmath81 this completes the first step of the process , as we have determined the local kr - coordinates around @xmath79 and described the distribution @xmath82 in these coordinates .
note that here @xmath83 .
we now determine which of the critical planes @xmath8 and @xmath15 lie above @xmath79 in @xmath82 , which is coframed with @xmath84 , this abuse of notation is convenient and should not cause confusion . ] by @xmath85 $ ] .
first consider @xmath8 , given by @xmath86 $ ] with @xmath87 .
we assume for now that it exists within some baby monster , and we will either find this baby monster or derive a contradiction . since @xmath85=[a : 0 : b]$ ] here with @xmath87 , we see that @xmath88 is identically zero on the baby monster , while @xmath89 and @xmath90 are not .
now , since @xmath91 is coframed by @xmath50 $ ] near @xmath92 , and since @xmath93 and @xmath94 , this forces the baby monster to have the form @xmath50=[0 : c : d]$ ] .
since this is the form of a vertical plane , we can stop and conclude that @xmath95 exists , and lies inside the baby monster born at level 2 .
that is , the plane @xmath96 , which is the first prolongation of the tangent space to the fiber @xmath97 .
next , we repeat this process for @xmath15 , given by @xmath98 $ ] with @xmath87 .
we assume for now that it exists within some baby monster , and we will either find this baby monster or derive a contradiction . since @xmath85=[a : b : 0]$ ] here with @xmath87 , we see that @xmath90 is identically zero on the baby monster , while @xmath89 and @xmath88 are not .
now , since @xmath91 is coframed by @xmath50 $ ] near @xmath92 , and since @xmath93 and @xmath94 , this forces the baby monster to have the form @xmath50=[c : 0 : d]$ ] .
note that unlike the previous case , this is not vertical , so we must continue searching another level down .
since @xmath99 is coframed by @xmath45 $ ] near @xmath100 , and since @xmath101 and @xmath102 , this forces the baby monster to have the form @xmath45=[0 : e : f]$ ] .
since this is the form of a vertical plane , we can stop and conclude that @xmath103 exists , and lies inside the baby monster born at level 1 .
that is , the plane @xmath104 , which is the second prolongation of the tangent space to the fiber @xmath61 .
we conclude that both planes @xmath8 and @xmath15 occur above a point with rvt code @xmath31 , so that both codes @xmath105 and @xmath106 are admissible and realized ( assuming temporarily that @xmath11 is admissible ) .
compare this result with theorem [ thm2 ] and figure [ fig1 ] . also see figure [ fig2 ] for an illustration of this situation .
we summarize the results of this example in table [ tab : ex2 ] . .
this shows the birth of @xmath107 as the third prolongation of the vertical plane at level 1 .
see example [ ex : rvl1t2 ] . ]
[ ex : rvl1t2 ] we continue the work from the previous example , and consider the case of @xmath108 with rvt code @xmath106 .
this is admissible by the preceding computations , and indeed , all results from that example hold here . as the general techniques were made explicit there , we omit some tiresome details here .
first , one finds affine coordinates @xmath109 and @xmath110 for the fiber @xmath111 .
next , recall that @xmath82 , is coframed by @xmath85 $ ] , and @xmath103 locally satisfies @xmath112 , with @xmath113 non - vanishing and @xmath114 not identically zero .
this implies that @xmath115 , but @xmath116 is non - zero .
( if @xmath117 were zero , then there would be no vertical component , and @xmath118 would lie in a regular direction instead of in @xmath15 . )
second , we show that @xmath15 does occur in @xmath119 .
this computation is nearly identical to those presented in the previous example , so we omit it .
one finds that @xmath120 .
finally , we show that @xmath8 can not occur in @xmath119 .
if it did , it would have the form @xmath121=[a : 0 : b]$ ] with @xmath122 .
but @xmath123 with @xmath124 .
this implies @xmath125 , so @xmath126 , which contradicts the fact that @xmath116 is non - zero in a neighborhood of @xmath127 we have shown that the @xmath15 critical plane occurs , but @xmath8 does not , in @xmath119 for @xmath128 in the class @xmath106 .
we conclude that the code @xmath106 can be amended with letters @xmath16 , and @xmath17 , but not with @xmath129 or @xmath130 .
compare with theorem [ thm2 ] , figure [ fig1 ] , and the second row of table [ tab : codes ] . also see figure [ fig3 ] for an illustration of this situation .
in this section we will outline the proof of theorem @xmath131 from the introduction , which we restate here . 1
. every word must begin with @xmath7 2 .
@xmath7 must be followed by @xmath7 or @xmath12 3 .
@xmath12 and @xmath8 must be followed by @xmath13 , or @xmath14 4 .
@xmath15 must be followed by @xmath16 , or @xmath17 5 .
@xmath18 , and @xmath17 can be followed by any letter .
let us begin with an overview of the method of proof .
the first two rules are well known and appear in @xcite and @xcite . rule ( 3 )
can be checked by direct calculation ; this is tedious but straightforward and we omit the computation here .
the same can be said for the part of rule ( 5 ) concerning the letter @xmath14 .
the technique is illustrated by examples in @xcite and the three examples above .
for example , one finds that for any point @xmath132 , the plane @xmath133 is obtained by prolonging the vertical plane from one level below . in other words , @xmath134 .
similarly , the plane @xmath135 is the prolongation of the @xmath8 plane from one level below .
this is independent of the code @xmath136 . to prove the remaining rules , ( 4 ) and most of ( 5 ) , we proceed by induction on the number of letters @xmath137 , or @xmath17 appearing in the code
this proof is more delicate .
set @xmath138 . for the base case
, we must prove that the spelling rules hold for an rvt code @xmath11 containing only one letter @xmath139 .
for the inductive step , we must prove that the spelling rules hold for an arbitrary code @xmath11 , using the inductive hypothesis that the rules hold for any code containing fewer letters @xmath139 .
in both steps , we assume without loss of generality that the letter @xmath140 appears at the end of the code in question
. unfortunately ( but perhaps unsurprisingly given the examples above ) , this method requires investigating a large number of specific cases , as well as a considerable number of tedious calculations .
we therefore spare the reader details of all cases , and the lengthy but routine computations which are required to prove each spelling rule rigorously .
instead , we will focus in detail on one particular rule : the fourth .
we hope that this approach will yield sufficient detail to introduce the mechanics of the method to the reader , while sparing the reader dozens of pages of nearly identical calculations .
we chose these particular cases as they exhibit generally typical behavior , but with a few of the subtleties which necessitate special care and patience .
we assume rules ( 1 )
( 3 ) have been proved . here
we will provide details for rule ( 4 ) ; the remaining proofs are very similar . to this end , let @xmath11 be an rvt code of length @xmath141 , ending with the letter @xmath15 .
we will show that codes @xmath142 , and @xmath143 do occur at level @xmath144 , while @xmath145 , and @xmath146 are impossible .
we prove this by induction on the number of letters @xmath147 appearing in @xmath11 .
we first prove the base case .
assume @xmath148 , where @xmath136 does not contain any letter from @xmath149 .
we will show that rule ( 4 ) holds for this @xmath11 .
we prove this by considering the potential letters preceding @xmath15 . by rules ( 2 ) and ( 3 ) ,
@xmath15 can not be preceded by @xmath7 or @xmath12 or @xmath8 .
since we have assumed that @xmath136 contains no letters from @xmath149 , we know @xmath15 can not be preceded by @xmath137 , or @xmath17 .
we therefore consider the only remaining possibility : @xmath15 is preceded by @xmath14 .
note that for convenience we will use @xmath136 to denote any sub - code of @xmath11 , regardless of its length .
so we proceed assuming our code has the form @xmath150 , where @xmath136 contains no elements from @xmath149 .
thus , the predecessor of @xmath14 can only be @xmath24 , or @xmath14 .
we have three possible cases .
assume our code has length @xmath141 and is of the form @xmath152 with @xmath153 .
if @xmath154 , then @xmath12 precedes @xmath14 ; if @xmath155 , then @xmath8 does .
the third possibility , where @xmath14 precedes @xmath14 , is treated as a separate case below .
in fact , we can assume without loss of generality that @xmath156 .
this is valid because the plane @xmath157 is the ( possibly multi - step ) prolongation of some vertical plane from a lower level .
that is , @xmath158 for some baby monster , and this subtower could not have been born at a level below the last letter @xmath12 in the rvt code .
now consider @xmath156 .
we have @xmath159 .
we wish to show that the spelling rules hold for @xmath11 .
this is to show that the codes @xmath160 are realized for @xmath161 , but are impossible for @xmath162 .
now there are regular and vertical directions in each distribution plane , so it is clear that @xmath163 or @xmath12 are possible .
recall from definition [ letters ] that @xmath164 , @xmath165 , and @xmath166 .
it is therefore sufficient to simply show that @xmath167 is possible , while @xmath168 is not .
the proof here is nearly identical to that provided in example [ ex : rvl1t2 ] .
in fact , that example gives precisely the case where @xmath154 . recall that in that case , @xmath8 could not appear and @xmath169 . for @xmath155
, we easily verify that , again @xmath8 can not appear , and @xmath170 .
the key observation is the following .
the vertical plane @xmath171 is coframed by @xmath45=[a : b : 0]$ ] with @xmath122 .
the prolongation of this plane is @xmath172 , which is coframed by @xmath173=[a:0:b]$ ] with @xmath68 . for @xmath155
, we continue this process and find that the @xmath174th prolongation of @xmath171 is @xmath175 , which is coframed by @xmath176=[a:0:b]$ ] with @xmath68 .
the rest of the steps are the same as in example [ ex : rvl1t2 ] .
the method here is the same as in case 1 , so we will omit some of the readily checked details .
again suppose the length of @xmath11 is @xmath141
. then @xmath181 is coframed by @xmath182 $ ] , and @xmath157 would have the form @xmath182=[a : b:0]$ ] with @xmath183 and @xmath184 .
its projection in @xmath185 will have the form @xmath186=[a : b:0]$ ] with @xmath183 and @xmath187 .
its projection in @xmath188 will have the form @xmath189=[a:0:b]$ ] with @xmath183 and @xmath190 .
finally , its projection in @xmath191 will have the form @xmath192=[0:a : b]$ ] with @xmath183 . at this point
, we can see that this is the vertical plane @xmath193 , so we find that @xmath157 does indeed exist in @xmath181 , and that it is equal to @xmath194 .
a computation similar to this one and those found in example [ ex : rvl1t2 ] shows that @xmath195 can not exist . in short ,
one repeats this computation beginning with @xmath195 of the form @xmath182=[a:0:b]$ ] with @xmath183 , and at some point a contradiction is obtained in that some coordinate is forced to be both zero and nonzero .
this establishes the base case for the proof of rule ( 4 ) by induction .
we showed that rule ( 4 ) holds for any rvt code containing a single member of @xmath149 ( which , in the context of rule ( 4 ) , must naturally be the letter @xmath15 . )
these three cases comprise the top three rows in table [ tab : codes ] .
the remaining cases are displayed as the lower six rows in table [ tab : codes ] ; their proofs are similar .
we now take @xmath11 to be an arbitrary rvt code of length @xmath141 .
we assume that @xmath11 ends with some letter from @xmath149 , and we will show that the spelling rules hold for @xmath11 .
our inductive hypothesis states that the spelling rules hold for any code which contains fewer letters from @xmath149 than @xmath11 does .
as above , we will focus on rule ( 4 ) , so our code should end with the letter @xmath15 .
so we have @xmath148 and our inductive hypothesis allows the assumption that @xmath136 satisfies the spelling rules .
we wish to show that , at level @xmath144 , the codes @xmath160 are realized for @xmath161 , but are impossible for @xmath162 . now there are regular and vertical directions in each distribution plane , so it is clear that @xmath163 or @xmath12 are possible .
recall from definition [ letters ] that @xmath164 , @xmath165 , and @xmath166 .
it is therefore sufficient to simply show that @xmath167 is possible , while @xmath168 is not .
now since @xmath136 clearly has ( exactly one ) fewer letters from @xmath149 than @xmath11 does , it must obey the spelling rules by assumption .
so @xmath15 must be preceded by either @xmath213 , or @xmath17 .
there are four cases here , but we will give details for just the first and second .
the other two are nearly identical .
suppose @xmath215 has rvt code @xmath214 .
from the discussion above , it suffices to prove that @xmath15 appears in @xmath181 , while @xmath8 does not .
now the distribution @xmath216 is coframed by @xmath217 $ ] .
two levels down , @xmath188 is coframed by @xmath189 $ ] , but since @xmath218 , it must have the form @xmath219 with @xmath220 .
we must therefore have @xmath221 coframed by @xmath222= \left[1 : \frac{du_{k-2}}{df_{k-2 } } : \frac{dv_{k-2}}{df_{k-2}}\right]=[1 : u_{k-1 } : v_{k-1}]\ ] ] where @xmath223 and @xmath224 is not identically zero .
moreover , we see that @xmath225 . since @xmath226 , the same argument shows that @xmath227 is coframed by @xmath228= \left[1 : \frac{du_{k-1}}{df_{k-2 } } : \frac{dv_{k-1}}{df_{k-2}}\right]=[1 : u_{k } : v_{k}]\ ] ] where @xmath229 and @xmath230 is not identically zero .
moreover , we see that @xmath231 .
now as an ansatz , suppose @xmath157 indeed appears in @xmath181 .
then it would have the form @xmath232=[a : b : 0]$ ] with @xmath68 .
its projection one level down would have the form @xmath233=[a : b : 0]$ ] with @xmath68 .
we recognize this as @xmath227 , which we know exists in @xmath185 .
therefore @xmath157 indeed exists as it is the prolongation of @xmath227 , and our ansatz is justified .
finally , assume for sake of contradiction that @xmath195 appears in @xmath181 .
it would have the form @xmath232=[a : 0 : b]$ ] with @xmath68 .
its projection one level down would have the form @xmath234=[a : 0 : b]$ ] with @xmath68 .
this forces @xmath235 .
but we saw above that a local fiber coordinate at @xmath236 is @xmath237 , and @xmath238 is not identically zero .
this contradiction disproves the existence of @xmath195 in @xmath181 .
suppose @xmath215 has rvt code @xmath150 .
from the discussion above , it suffices to prove that @xmath15 appears in @xmath181 , while @xmath8 does not .
now the distribution @xmath216 is coframed by @xmath217 $ ] .
two levels down , @xmath188 is coframed by @xmath189 $ ] , but since @xmath239 , it must have the form @xmath219 with @xmath240 .
we must therefore have @xmath241 coframed by @xmath222= \left [ \frac{df_{k-2}}{dv_{k-2 } } : \frac{du_{k-2}}{dv_{k-2}}:1\right]= [ u_{k-1 } : v_{k-1 } : 1].\ ] ] moreover , we see that @xmath187 . since @xmath226 ,
the we can similarly see that @xmath227 is coframed by @xmath242= \left[1 : \frac{du_{k-1}}{dv_{k-2 } } : \frac{dv_{k-1}}{dv_{k-2}}\right]=[1 : u_{k } : v_{k}]\ ] ] where @xmath229 and @xmath230 is not identically zero .
moreover , we see that @xmath184 .
now as an ansatz , suppose @xmath157 indeed appears in @xmath181 .
then it would have the form @xmath243=[a : b : 0]$ ] with @xmath68 .
its projection one level down would have the form @xmath244=[a : b : 0]$ ] with @xmath68 .
we recognize this as @xmath227 , which we know exists in @xmath185 .
therefore @xmath157 indeed exists as it is the prolongation of @xmath227 , and our ansatz is justified .
finally , assume for sake of contradiction that @xmath195 appears in @xmath181 .
it would have the form @xmath243=[a : 0 : b]$ ] with @xmath68 .
its projection one level down would have the form @xmath245=[a : 0 : b]$ ] with @xmath68 .
this forces @xmath235 .
but we saw above that a local fiber coordinate at @xmath236 is @xmath246 , and @xmath238 is not identically zero .
this contradiction disproves the existence of @xmath195 in @xmath181 .
the authors thank richard montgomery ( santa cruz ) for many useful conversations and remarks , and for introducing the authors to this subject .
warm thanks also to gary kennedy ( ohio state ) and susan colley ( oberlin ) for continuing motivation .
we are also grateful to the referee for helpful comments and suggestions .
a. castro , w. howard , and c. shanbrom , bridges between subriemannian geometry and algebraic geometry , _ proceedings of 10th aims conference on dynamical systems , differential equations , and applications _ , ( 2015 ) .
w. respondek , symmetries and minimal flat outputs of nonlinear control systems , _ new trends in nonlinear dynamics and control and their applications _ , lecture notes in control and information science , * 295 * ( 2004 ) , 6486 . | the monster tower , also known as the semple tower , is a sequence of manifolds with distributions of interest to both differential and algebraic geometers .
each manifold is a projective bundle over the previous .
moreover , each level is a fiber compactified jet bundle equipped with an action of finite jets of the diffeomorphism group .
there is a correspondence between points in the tower and curves in the base manifold .
these points admit a stratification which can be encoded by a word called the rvt code . here
, we derive the spelling rules for these words in the case of a three dimensional base .
that is , we determine precisely which words are realized by points in the tower . to this end
, we study the incidence relations between certain subtowers , called baby monsters , and present a general method for determining the level at which each baby monster is born . here , we focus on the case where the base manifold is three dimensional , but all the methods presented generalize to bases of arbitrary dimension . |
in a direct analogy with classical mechanics , composed systems in quantum optics ( which describe interaction between several subsystems ) can be divided into two classes : 1 ) systems which possess a necessary number of integrals of motion , so that , the classical counterpart is an integrable system ; 2 ) systems that do not admit a sufficient number of integrals of motions , so that , the classical counterpart is a non - integrable system caos . a quantum system can have basically two types of integrals of motion : a ) _ kinematic _ integrals , which do not depend on the kind of interaction between subsystems , as for instance , the total number of atoms ; b ) _ dynamic _ integrals which are related to the particular form of interaction between subsystems , as , for instance , a number of excitations preserved in some transitions between the energy levels of the subsystems .
typical for quantum optical systems dipole - like interactions between two subsystems ( @xmath0 and @xmath1 ) can be described with a generic multichannel hamiltonian of the following form:@xmath2where the two first terms represent the free hamiltonians of the subsystems , so that the frequencies @xmath3 , and the last term describes the interaction between them .
the operators @xmath4 , @xmath5 , @xmath6=0 $ ] , are usually elements of some deformed algebra @xcite , and in particular , satisfy the ladder commutation relations@xmath7 = \pm x_{\pm } ^{j},\hspace{0.1in}\left [ y_{0}^{j},y_{\pm } ^{j}\right ] = \pm y_{\pm } ^{j}. \label{escal}\ ] ] in the interaction hamiltonian there are two kinds of terms : of the form @xmath8 and @xmath9 it is easy to observe that in the rotating frame , that is , applying the following unitary transformation@xmath10 \right)\]]to the hamiltonian ( [ h1 ] ) , the _ counterrotating _ terms @xmath11 oscillate in time with a frequency @xmath12 and the _ rotating _
terms @xmath13 oscillate with a frequency @xmath14 .
it is clear that under the condition @xmath15 the rotating term @xmath8 in ( [ h1 ] ) is approximately time independent ( and thus , can generate transitions with a probability of one between the energy levels of the system ) , meanwhile the counterrotating _ _ _ _ term , @xmath16 always oscillates rapidly in the rotating frame , and its temporal average is zero . by neglecting the counterrotating terms in the hamiltonian ( [ h1 ] ) , which is commonly called the rotating wave approximation ( rwa ) , we arrive at the hamiltonian@xmath17which admits several dynamic integrals of motion @xmath18 , generally not allowed in ( [ h1 ] ) .
this implies that the whole representation space of the system is divided into finite dimensional invariant subspaces , and the mathematical treatment is essentially simplified .
it is worth noting , that the semiclassical models , when some of the subsystems are described by @xmath19-numbers instead of operators , are treated using essentially the same type of hamiltonians as ( [ h1 ] ) . the semiclassical transition in ( [ h1 ] )
can be done by going to the rotating frame of the semiclassical system and then , just substituting the transition operators by some complex numbers .
for instance , in the case of single channel interaction , @xmath20 , @xmath21 , when the system @xmath1 acquires classical features , the hamiltonian ( [ h1 ] ) in the rotating frame corresponding to the system @xmath1 takes the form @xmath22so that the corresponding semiclassical hamiltonian is obtained by substituting @xmath23 giving@xmath24such hamiltonians usually appear when a quantum oscillator and/or a collection of atoms is pumped by an external force @xcite , cohen , @xcite . in the rwa - like systems , described by hamiltonians of the form ( [ h3 ] ) , the resonance conditions@xmath25means that the term @xmath26 , which explicitly appears in the hamiltonian , does not depend on time in an appropriate rotating frame
nevertheless , such _
explicit resonances _ are not the only kind of resonances , which can be found in the hamiltonian ( [ h3 ] ) . usually , the composed systems admit several types of _ implicit resonances _ related to effective transitions , which do not appear in the original hamiltonian , between their energy levels .
such effective interactions play important roles in many physical applications and can be revealed by adiabatic elimination of slow transitions @xcite .
the implicit ( effective ) resonances are characterized by their position , @xmath27 , and strength , i.e. in what order of perturbation expansion they appear for the first time .
although , a generic system can possess a large number of different types of effective transitions , all the possible resonance conditions can be classified only by analyzing the free hamiltonian and the integrals of motion .
because all the invariant subspaces are finite - dimensional , there are always a finite number of different resonances
. we will refer to these kinds of resonances as _
kinematic resonances _ , which include both explicit and implicit resonances . in sec .
ii we show with the example of atom - field interactions , that it is possible to classify all the kinematic resonances in a straightforward way .
the situation is quite different in quantum systems with a lack of integrals of motion , corresponding to classically non - integrable dynamic systems . in such systems
an infinite number of different resonances arise and _ a priori _ it is impossible to determine their position and strength , which essentially depend not only on the type of interaction but also on the algebraic properties of each interacting subsystem @xcite , bombeo , @xcite . in sec.iii
we will discuss such _ dynamic resonances _ analyzing different models of interaction of quantum and classical fields with atomic systems .
an important example of kinematic resonances is the dicke model @xcite , this model describes the interaction of a collection of @xmath28 identical two - level atoms with a single mode of a quantized field under the rotating wave approximation .
the hamiltonian that governs this system is given by @xmath29where @xmath30 is the field frequency , @xmath31 is the atomic frequency , and @xmath32 are the collective atomic operators , they represent the atomic inversion , and the transition between the atomic energy levels respectively , their commutation relations are given by the @xmath33 algebra , @xmath34 = s_{\pm } , \quad \left [ s_{+},s_{-}\right ] = 2s_{z},\quad \label{su2}\]]and @xmath35 are the creation - annihilation field operators , obeying the bosonic commutation relations , @xmath36=1 $ ] .
the hamiltonian ( [ hd ] ) admits two integrals of motion , a kinematic integral of motion , given by the total number of atoms ( that is constant for a closed system ) , @xmath37 , where @xmath38 is the atomic population operator for the @xmath39-th atomic energy level , and the dynamic integral of motion , corresponding to the total number of excitations , has the form @xmath40=0 .
\label{n}\ ] ] this system is the simplest non trivial example of one - channel quantum transitions : an absorption of one photon is accompanied by an excitation of one atomic transition . using the integral of motion ( [ n ] ) , the hamiltonian ( [ hd ] ) can be rewritten as follows,@xmath41where the interaction hamiltonian is@xmath42and @xmath43 , is the detuning between the field and the atomic transition frequencies .
if the only possible resonant condition , @xmath44 is held , the interaction hamiltonian ( [ hd2 ] ) is reduced to , @xmath45which implies that the atomic transition probability ( as a function of time ) oscillates between zero and one .
on the other hand , in the far - off resonant ( dispersive ) limit , @xmath46 , \label{hdeff}\ ] ] so that only the phase of the system evolves due to the appearance of the so - called dynamic stark shift terms @xcite , @xcite in the _ effective _ hamiltonian ( [ hdeff ] ) .
let us consider the interaction of a system of @xmath28 identical @xmath47-level atoms of an arbitrary configuration with a single mode of a quantized field of frequency @xmath30 .
the hamiltonian describing this system has the form @xmath48 where@xmath49is the free hamiltonian , and @xmath50 are the collective atomic population operators corresponding to the @xmath51-th atomic level of energy @xmath52 , and @xmath53 is the interaction atom - field hamiltonian , whose explicit form depends on the atomic configuration , in the dipole approximation , i.e. only one - photon transitions are allowed . in this section
we suppose that the rotating wave approximation is imposed , so that the total number of excitations in the atom - field system is preserved , and thus , the whole representation space of this quantum system is divided into finite - dimensional invariant subspaces .
the hamiltonian ( [ hc1 ] ) admits two integrals of motion : a kinematic integral , given by the total number of atoms:@xmath54and a dynamic integral , corresponding to the total number of excitations in the system:@xmath55where the parameters @xmath56 depend on the atomic configuration .
let us note that a _ generic _ interaction term can be written as follows @xmath57where @xmath58 ( @xmath59 @xmath60 , @xmath61 ) are the atomic transition operators satisfying the @xmath62 commutation relations , @xmath63 = \delta _ { im}s^{kj}-\delta _ { kj}s^{im}$ ] , and negative exponents correspond to the hermitian conjugated operators .
the operational coefficients @xmath64 depend on diagonal atomic operators and the photon number operator .
in what follows we will omit the coefficient @xmath64 , since it leads only to some phase shifts and does not change the distribution of excitations in the system 's energy levels .
it is worth noting that ( [ tg ] ) is not a unique way to represent a generic interaction term .
since the rotating wave approximation is imposed , every interaction term should preserve the total number of excitations , so the condition@xmath65 = 0,\]]is held , and thus , the numbers @xmath66 satisfy the following restriction @xmath67thus , any admissible interaction can be described as@xmath68the interaction ( [ git ] ) becomes resonant when the atomic transition energies and the field frequency satisfy the following condition@xmath69the important point here is that interaction terms ( [ git ] ) can be explicitly presented in ( [ hc1 ] ) , or can describe effective interactions , and thus should be obtained from the original hamiltonian by adiabatic elimination of some far - off resonant transitions .
so that , the condition ( [ resc ] ) describes _ explicit _ resonances if the corresponding interaction term is present in the original hamiltonian or _
implicit _ resonances , if such interaction is effective .
it is clear that the total number of both explicit and implicit resonances is finite , which is a consequence of the restriction ( [ k ] ) imposed by the rotating wave approximation .
the number of possible resonances depends on the number of atomic levels , @xmath47 , and the total number of atoms , @xmath28 .
note that the resonance condition ( resc ) is associated with a vector @xmath70 , in order to not repeat the resonances , we consider only vectors @xmath71with coprime components ( that is @xmath72 do not have a common factor ) . then , different vectors @xmath71 satisfying the following condition * * * * @xmath73define different resonances . on the other hand ,
taking into account the dynamic integral of motion ( i d ) , one can rewrite the free hamiltonian ( [ hdiag ] ) as follows,@xmath74summing the kinematic integral @xmath75 ( [ ic ] ) to the above hamiltonian , where the constants @xmath76 , are chosen such that,@xmath77we
obtain @xmath78it is easy to see that the condition @xmath79 for any fixed @xmath51 and some values of @xmath80 , enumerates all the possible resonances ( [ resc ] ) .
let us take for instance @xmath81 if @xmath82 then@xmath83which coincides with ( [ resc ] ) , when @xmath84 for @xmath85 and @xmath86 .
this means that we can always represent the free hamiltonian in a way that all the possible resonance conditions , corresponding to both explicit and implicit resonances , appear as zeros of coefficients of the atomic population operators @xmath50 in the free hamiltonian , after taking out the integral of motion ( [ i d ] ) corresponding to the total number of excitations .
the above allows us to classify all the possible kinematic resonances : 1 . _
multiphoton resonances : _ transitions which involve absorption and emission of photons : a ) simple @xmath87-photon transitions , described by the terms @xmath88 , with the resonance condition @xmath89 .
the terms with @xmath90 can be present in the original hamiltonian and in such case represent explicit resonances .
multiphoton transitions appear even in a single atom case ; b ) collective atomic transitions , described by the terms @xmath91 ( in general a product of several , up to @xmath28 atomic transition operators can appear ) , with the corresponding resonance condition @xmath92 .
it is clear that such resonances can appear only in multi - atom systems .
note , that if @xmath93 ( or @xmath94 ) then we obtain the same resonance condition as in a ) , describing an effective process of absorption of @xmath87 photons with atomic transition from @xmath95-th to @xmath51-th energy levels .
nevertheless , the corresponding term in the effective hamiltonian would be multiplied by the atomic population operator @xmath96 , which means that more than one atom is need to realize such a process .
2 . _ virtual _ _ photon resonances _ : atomic transitions between independent channels caused by the quantum field fluctuations , and thus existing even when the field is in the vacuum state .
such resonances appear only when the system has more than one atom and are described by terms in the form of products of atomic transitions operators .
for instance , the simplest term of this kind ( typically appearing in the first order perturbation expansion ) is @xmath97 , @xmath98 , @xmath99 , describes atomic transitions @xmath100 , @xmath101 , and the corresponding resonance condition is @xmath102obviously , such transitions can be realized in a system which consists of at least two atoms . note that the term @xmath103 which represents the atomic transition @xmath104 , @xmath105 , satisfies the same resonant condition .
more involved interactions , like@xmath106 , can appear in the highest orders of the perturbation theory .
the strength of virtual photon transitions does not depend on the field intensity in the leading order of the perturbation expansion .
_ photon assisted transitions _ : atomic transitions when every photon emission is accompanied by a simultaneous photon absorption .
these transitions appear only when the atomic configuration contains coherent channels , similar to lambda - like configurations .
the strength of such interactions depends on the number of photons in the field and the populations of some atomic energy levels .
the simplest terms describing the photon assisted transition ( which appear in the lowest order of the perturbation theory ) have the form : @xmath107 , corresponding to the resonance condition @xmath108 , where the transition @xmath109 is not present in the original hamiltonian and @xmath110 is a linear polynomial of the photon number operator and the atomic population operators . in the above classification we do not consider interactions corresponding to powers of terms describing some interactions .
for instance , the @xmath87-photon transition corresponding to the term @xmath111 that describes an absorption of @xmath87 photons by at least @xmath87 atoms in the transition @xmath112 , we include in `` one photon transitions '' .
in the case of interaction of an atomic system with classical field the rotating wave approximation implies that the interaction hamiltonian can always be reduced to a time - independent form . in this case , apart from explicit resonances , several types of implicit ( effective ) resonant transitions can take place .
it is clear that no interactions similar to virtual photon resonances can arise .
nevertheless , transitions similar to multiphoton resonances @xmath113 and photon assisted resonances @xmath114 , where the transition @xmath115 does not exist in the initial hamiltonian , actually appear . as an example of the kinematic resonances we will study the interaction of @xmath116 four level diamond configuration atoms with a single mode of a quantized field under rwa .
the hamiltonian governing the evolution of this system has the form@xmath117this hamiltonian describes four one - photon atomic transitions , ( @xmath118 , @xmath119 , ( @xmath120 , @xmath121 , gathered in two pairs of coherent quantum channels . the corresponding resonance conditions ( explicit resonances ) are @xmath122 , @xmath123 , @xmath124 , and @xmath125 . to find all the other ( implicit ) resonances we will follow the method outlined in previous subection .
the dynamic integral of motion for this system is @xmath126so that the coefficients @xmath56 in ( [ i d ] ) , are @xmath127 , @xmath128 , @xmath129 . substituting these @xmath56 into ( resc )
we obtain all the possible resonance conditions : @xmath130@xmath131 and the corresponding ( effective ) interactions are@xmath132the vector @xmath133 associated with possible resonances should satisfy the condition ( [ pr])@xmath134 for a single atom we have the vectors @xmath135 , @xmath136 , @xmath137 , @xmath138 , @xmath139 , @xmath140 , which correspond to the following resonance conditions ( interactions ) : @xmath141 @xmath142 , @xmath143 @xmath144 , @xmath125 @xmath145 , @xmath146 @xmath147 , @xmath123 @xmath148 , @xmath149 @xmath150 , respectively . for two atoms there
there are 15 other vector apart from 6 shown above . for a better understanding of the nature of the implicit resonances , we find the first order effective hamiltonian considering that all the interactions appearing in the initial hamiltonian ( [ h4lev ] ) are far from resonance . following the method outlined in the appendix
, we apply to the hamiltonian ( [ h4lev ] ) the following sequence of unitary transformations @xmath151where @xmath152 @xmath153and the small parameters @xmath154 are given by@xmath155we obtain , in the first order on small parameters , the following effective hamiltonian@xmath156 \nonumber \\ & & + g_{4}\varepsilon _ { 3}\left [ s_{+}^{23}\left ( s^{44}-a^{\dagger } a\right ) + h.c.\right ] -\left ( g_{3}\varepsilon _ { 1}+g_{4}\varepsilon _ { 2}\right ) \left ( a^{2}s_{+}^{14}+h.c.\right ) \nonumber \\ & & + g_{3}\varepsilon _ { 1}\left ( s_{+}^{24}s_{-}^{12}+h.c.\right ) + g_{3}\varepsilon _ { 2}\left ( s_{+}^{24}s_{-}^{13}+h.c.\right ) \nonumber \\ & & + g_{4}\varepsilon _ { 1}\left ( s_{+}^{12}s_{-}^{34}+h.c.\right ) + g_{4}\varepsilon _ { 2}\left ( s_{+}^{13}s_{-}^{34}+h.c.\right ) .
\label{h4eff}\end{aligned}\]]where@xmath157is the dynamic stark shift @xcite .
let us classify the effective resonances present in the above ( first order)hamiltonian : 1 .
_ multiphoton resonances _ : * * * * the only two resonances of this type are one - photon and two - photon resonances
. there are four one - photon transitions corresponding to explicit resonances , which do not appear in the effective hamiltonian ( [ h4eff ] ) .
the term @xmath158 describes two - photon transitions , with corresponding resonance condition @xmath141 .
virtual photon resonances : _ the last four terms in ( [ h4eff ] ) , @xmath159 , @xmath160 , @xmath161 , @xmath162 , with the resonance conditions @xmath163 , @xmath164 , @xmath165 and @xmath166 respectively .
photon assisted resonances : _ the only resonance of this type is of the form * * * * @xmath167 is produced in the first order of the perturbation theory generating transitions between the middle atomic levels , and the corresponding resonance condition is @xmath168
in this section we study quantum systems corresponding to classically non - integrable systems , mainly focusing on the models describing interaction of atoms with quantized and classical fields without applying the rotation wave approximation @xcite , @xcite ( some different systems not preserving the number of excitations are discussed in @xcite ) .
it is convenient to start with a general analysis of such systems .
the main idea consists in removing the counter - rotating terms from the multi - channel hamiltonian ( [ h1 ] ) using a method of adiabatic elimination . in the limit of weak interaction ,
@xmath169 one can apply , for instance , the lie - transformation method outlined in the appendix .
although , such analysis can be performed for a general system , we will focus on the simplest case of a one channel hamiltonian and show , that even in this case the absence of the integral of motion corresponding to the total number of excitations leads to the appearance of a series of ( dynamic ) resonances , which can be classified according the interaction type .
consider a single - channel hamiltonian ( [ h1])@xmath170where @xmath171 ( @xmath172 ) is the free hamiltonian of the subsystem @xmath0 ( @xmath1 ) and @xmath173 ( @xmath174 ) , @xmath175 ( @xmath176 ) are the up and down operators ( see appendix ) , respectively , which describe transitions between the energy levels in the subsystem @xmath0 ( @xmath1 ) , they hold the @xmath177 @xcite commutation relations ( [ escal ] ) . from now on ,
we do not impose any commutation relation between the transition operators , which are generally some functions of the diagonal operators and integrals of motion @xmath178 = \left [ n_{2},y_{0,\pm } \right ] = 0 $ ] , @xmath179 & = & p_{1}\left ( x_{0}\right ) = \nabla _ { x_{0}}\phi _ { 1}\left ( x_{0},n_{1}\right ) , \\ \left [ y_{+},y_{-}\right ] & = & p_{2}\left ( y_{0}\right ) = \nabla _ { y_{0}}\phi _ { 2}\left ( y_{0},n_{2}\right ) .\end{aligned}\ ] ] where @xmath180 , @xmath181 are given by the structural functions ( appendix ( [ estruc ] ) ) , and , in general , ( we are going to omit the dependence of the integral of motion ) @xmath182 the consequences of the existence of the _ counterrotating _ term @xmath183 in the hamiltonian ( [ h2 ] ) are : a ) the dimension of the whole representation space is the product of the corresponding dimensions of the subsystems : @xmath184 , b ) there are some additional resonances apart from @xmath185 , that in the case of a single - channel hamiltonian ( [ h3 ] ) do not exist , c ) the type of resonances depends on the structure of algebras describing the @xmath0 and @xmath1 systems . the counterrotating _ _ _ _ term rapidly oscillates ( with frequency @xmath186 ) , and thus , can be eliminated by applying the lie - type transformation@xmath187 , \quad \varepsilon = g/\left ( \omega + \omega \right ) \ll 1 , \label{u}\]]to the hamiltonian ( [ h2 ] ) and using the standard perturbative expansion * * , * * see appendix . from now on we suppose that @xmath188 the elimination of the above term leads to the appearance of new elements in the transformed hamiltonian .
all these new terms can be divided into three groups : the first group contains the _ non - resonant terms , _ of the form @xmath189 , that can always be eliminated under the condition @xmath190 by applying some suitable transformations ; the second group consists of _ resonant terms , _ that can not be removed if certain relations between @xmath30 and @xmath191 hold , since the interaction becomes resonant ( the transformation which eliminates a given term from the hamiltonian becomes singular ) .
this group contains terms of the form @xmath192 , which describe transitions between energy levels of the whole system .
the third group includes the diagonal terms ( functions only of @xmath193 ) , that can never be removed .
our strategy consists of keeping in the hamiltonian only diagonal terms and resonant terms , and , we conserve only the leading order coefficients in these terms .
all the counterrotating terms , like @xmath194 @xmath195 can be eliminated by applying the lie - type transformation @xmath196 , \label{umaint}\]]where @xmath197 , and @xmath198 is the product of powers of some small parameters @xmath199 , @xmath200 . in fact
@xmath201 is proportional to the coefficient of the term @xmath202 . applying a sequence of appropriate transformations ( [ umaint ] ) we obtain the effective hamiltonian @xcite @xmath203 @xmath204 , \label{heff}\ ] ] where @xmath205 the term @xmath206 represents the dynamic stark shift ( or bloch - siegert shift ) @xcite and
can be expanded on powers of the small parameter @xmath207 as follows @xmath208the terms of the form@xmath209 describe all the admissible resonant interactions .
in particular the term with @xmath21 , @xmath210 represents the principal resonance .
the coupling constants @xmath211 depend on the algebraic structure ( estruc ) of the operators describing both subsystems ( more precisely , on the degree of the structural polynomials @xmath212 and @xmath213 ) , and in particular , they can become zero for some @xmath95 , @xmath214 @xcite .
the coefficient in the principal term has the form @xmath215and @xmath216}c_{l+[k/2]}^{j}\left ( p_{1}\left ( x_{0}+k\right ) \right ) ^{l+[k/2]-j}\left ( -p_{1}\left ( x_{0}\right )
\right ) ^{j}r_{k}(x_{0},y_{0}-2j ) , \label{theta}\ ] ] @xmath217}c_{[\frac{k-1}{2}% ] } ^{j}\left ( p_{1}\left ( x_{0}+k\right ) \right ) ^{[\frac{k-1}{2}]-j}\left ( -p_{1}\left ( x_{0}\right ) \right ) ^{j}\nabla _ { y_{0}}^{k-1}\phi _ { 2}\left ( y_{0}-2j\right ) \\ & & \times \prod\limits_{i=1 - 2j}^{k-2(j+1)}\phi _ { 2}\left ( y_{0}+i\right ) , \end{aligned}\ ] ] for @xmath218 , where @xmath219 are the binomial coefficients and @xmath220 .
note that the product in the last equation is equal to unity if the upper limit is less than the lower one .
in contrast to the case of kinematic resonances discussed in the previous section , the number of possible resonant interactions appearing in the hamiltonian ( [ heff ] ) is infinite .
these interactions ( _ dynamic resonances _ ) can be classified as follows : a ) principal ( explicit ) resonance , @xmath221 corresponding to the interaction term that is explicitly present in the original hamiltonian @xmath222 ; b ) higher - order resonances : @xmath223 where @xmath224 @xmath225(@xmath226 ) , corresponding to the effective interactions @xmath227 , which can be divided into : b.1 ) odd resonances : when @xmath228 , that is @xmath229 , b.2 ) even resonances : @xmath230 , with the resonance condition @xmath231 , b.3 ) fractional resonances : @xmath232 where @xmath95 and @xmath214 are coprime numbers .
it is worth noting that in the vicinity of each resonance @xmath233 , only the interaction term @xmath234 survives .
this means that if some resonance condition is held , the system is in an approximate invariant subspace and there exists an approximate integral of motion @xmath235 .
this simple example shows that even in one - channel systems an infinite number of resonant interactions may arise , if the total number of excitations is not preserved .
the appearance of such _ dynamic _ resonances in chaotic - like systems is expected from the point of view of classical dynamic systems @xcite .
nevertheless , the quantum nature of interacting subsystems imposes certain restrictions on the possibility of surviving of the dynamic resonances .
we will discuss such restrictions for two examples of the interaction of an atomic system with quantum and classical fields .
let us consider a collection of @xmath28 identical two - level atoms interacting with a single mode of a quantized field ( the dicke model ) without rwa .
the hamiltonian that describes this system is@xmath236and the condition @xmath190 is held .
the following identifications @xmath237so that @xmath238 lead to @xmath239 and the effective hamiltonian ( [ heff ] ) takes the form@xmath240 -g\varepsilon \sum_{m=1}^{\infty } \frac{% \left ( 4\delta \varepsilon \right ) ^{2m}}{(2m)!}\left [ a^{4m}s_{+}^{2}+h.c.% \right ] , \nonumber\end{aligned}\]]where @xmath241 note that the hamiltonian ( [ hdeff ] ) contains only the principal ( explicit ) resonance ( @xmath210 ) , the even and the odd order interactions , but no fractional resonances .
this happens because the structural functions for our subsystems are of a first degree polynomial of the photon number operators @xmath242 and a second degree polynomial of the atomic population operator @xmath243 .
the integrals of motion for the series of odd and even ( exact ) resonances are @xmath244 + o\left ( \delta { } ^{3}\right ) , \end{aligned}\]]where @xmath245 in @xmath246 , @xmath247 and @xmath248 for even and odd resonances correspondingly .
let us recall that the higher resonances appear in the effective hamiltonian ( [ hdeff ] ) only under the condition @xmath249 . in the opposite case , @xmath250 , only the principal resonance survives and in the approximation ( [ heff ] ) the whole effect of the counter - rotating terms reduces to the dynamic stark shift , which has the same form as in the above case .
this does not imply that there are no higher resonances at all , but rather that they are essentially suppressed .
we will proceed with the analysis of the interaction of atomic systems with classical fields without rwa .
a generic interaction hamiltonian has the form@xmath251and admits some explicit resonances , @xmath252 .
the whole set of effective ( implicit ) interactions can be easily obtained in the same way as it was outlined in 3.1 .
to be able to use the effective hamiltonian in the form ( [ heff ] ) we first rewrite the hamiltonian ( [ hc ] ) in the floquet form by making use of the phase operators @xmath253 , which are generators of the euclidean algebra : @xmath254 = -\delta _ { jk}e_{j},\quad \left [ e_{0j},e_{j}^{\dagger } \right ] = \delta _ { jk}e_{j}^{\dagger } , \quad \left [ e_{j},e_{k}^{\dagger } \right ] = 0 . \label{ej}\]]each group of operators ( labeled with the same index @xmath51 ) acts in a hilbert space spanned on the eigenstates of the ( hermitian ) operator @xmath255@xmath256so that in the basis ( [ ejes1 ] ) the operators @xmath257 act as rising - lowering operators : @xmath258the phase states @xmath259 , the eigenstates of @xmath260 @xmath261 , are not normalized .
now , let us consider the following time independent hamiltonian @xmath262it is easy to observe that the average value of the above hamiltonian over the phase states in the rotating frame , @xmath263where @xmath264 are the basis states ( [ ejes1 ] ) , coincides with the hamiltonian ( [ hc1 ] ) due to @xmath265the hamiltonian ( [ hf ] ) is a floquet form of the initial time dependent hamiltonian ( [ hc1 ] ) ( from now on we put all the phases @xmath266 equal to zero ) . in the case of weak driven fields , @xmath267 , the hamiltonian ( [ hf ] ) can be represented in a form of expansion over the principal resonances which can be observed in this system according to ( [ heff ] ) . in the single channel case , the dicke model in the classical field , we have the classical problem , first solved by shirley @xcite ( for a single two level atom ) .
let us consider a collection of @xmath28 two level atoms in a linearly polarized em field .
the corresponding hamiltonian has the form @xmath268where @xmath191 is the classical field frequency .
the floquet form of ( [ hdicke ] ) is @xmath269the following identifications @xmath270so that @xmath271 lead to @xmath272 which immediately leads to the effective hamiltonian ( [ heff ] ) @xmath273averaging the above hamiltonian over the phase states we obtain an effective time - dependent hamiltonian @xmath274which means that only odd resonances , @xmath275 , appear in this system .
this result is different from the atom - quantum field interaction ( and surprisingly can not be obtained from the corresponding quantum hamiltonian ( [ hdeff ] ) by substituting the field operators by @xmath276numbers ) , because in this case the structural function ( [ estruc ] ) for the euclidean algebra ( [ ej ] ) is a constant , that is , a zero - th degree polynomial function .
quantum systems possessing as the integral of motion the operator corresponding to the total number of excitations ( classically integrable systems ) admit a simple classification of possible resonant transitions , which are separated into explicit ( which appear in the original hamiltonians ) and implicit ( effective ) resonances .
the total number of such resonances is always finite and depends on the number of atoms and of atomic levels . by a simple algebraic manipulation of the free hamiltonian and the integrals of motion
, one can obtain some specific conditions for the frequencies of interacting subsystems , which define the implicit resonances .
nevertheless , the form of the interaction terms associated with each implicit resonance condition can be found only after obtaining the effective hamiltonian ( in a perturbative way ) .
quantum systems not preserving the total number of excitations ( classically non - integrable systems ) do not admit a simple classification of admissible resonant transitions .
the explicit form of the effective hamiltonian is needed to specify the allowed resonances , and their number is always infinite .
such non - linear ( on the generators of some lie algebra ) hamiltonians can be represented as a series of operators describing all the possible transitions , which might become resonant under specific relations between frequencies of interacting subsystems .
the structure of the effective hamiltonian essentially depends on the algebraic structure of interacting subsystems ( polynomials @xmath277 ) .
in particular , it is reflected in the types of resonances which are allowed for a given system . in the vicinity of each resonant transition
all of the other transitions can be considered as a perturbation .
thus , in the course of evolution some specific finite - dimensional subspaces in the hilbert space of the whole system are approximately preserved , and for each of these invariant subspaces there is a corresponding integral of motion . this work is partially supported by the grant 45704 of consejonacional de ciencia y tecnologia ( conacyt ) .
the method of lie - type transformation ( small rotations ) @xcite , @xcite provides a _ regular _ procedure for obtaining approximate hamiltonians describing the effective dynamics of nonlinear quantum systems .
the idea of this method is based on the observation that several quantum optical hamiltonians can be written in terms of polynomially deformed algebras @xmath278 @xcite , @xcite , @xcite , bonatsos93 , @xcite , @xcite , @xmath279where the operators @xmath280 and @xmath171 are generators of the deformed algebra and satisfy the following commutation relations ( [ escal ] ) , and @xmath281=p(x_{0 } ) , \label{xpmt}\]]where @xmath282 is a polynomial function of the diagonal operator @xmath171 with coefficients that may depend on some integrals of motion @xmath283 if @xmath284 is a linear function of @xmath171 , the usual @xmath285 or @xmath33 algebras are restored . if for some physical reason ( depending on the particular model under consideration ) @xmath286 is a small parameter , the hamiltonian ( [ hintt ] ) is _ almost _ diagonal in the basis of the eigenstates of @xmath171 and can be approximately diagonalized by applying in a perturbative manner the following unitary transformation ( a _ small nonlinear rotation _ ) @xmath287 . \label{u1}\]]applying the transformation ( [ u1 ] ) to the hamiltonian ( [ hintt ] ) according to the standard expansion
@xmath289 is the adjoint operator defined as @xmath290 $ ] , we obtain @xmath291where @xmath292 and we have taken into account that , due to ( [ xpmt ] ) , @xmath293=-\left ( x_{+}+x_{-}\right ) .\]]the effective hamiltonian acquires the following form @xmath294 , \]]where @xmath295 is a function of the diagonal operator @xmath171 and can be represented as a series on @xmath296 : @xmath297and @xmath298is a structural function , @xmath299 ; the prime ( @xmath300 ) in the above sum means that the term with @xmath301 is taken with the coefficient @xmath302 .
the higher - order contributions always have the form @xmath304 .
this makes the procedure of removing the off - diagonal terms somehow trivial at each step , in the sense that it is always obvious which transformation should be applied . for example , to eliminate the terms of the form @xmath305,\]]it suffices to apply the transformation @xmath306with @xmath307 , since the first commutator of @xmath308 with @xmath309 cancels the corresponding term in the hamiltonian .
a. schenzle , a. , _ nonlinear optical phenomena and fuctuations _ , lecture notes in physics * 155 * ( berlin : springer ) ( 1981 ) p. 103 .
shen , _ the principles of nonlinear optics _ ( new york : john wiley ) ( 1985 ) .
lugiato , p. galatola , and l.m .
narducci 1990 , opt .
commun . , * 76 * , 276 ( 1990 ) .
l. sczaniecki , phys .
a * 28 * , 3493 ( 1983 ) .
m. hillery , and l.d .
mlodinow , 1985 , phys .
a , * 31 * , 797 ( 1985 ) .
klein , d. , j. chem .
phys . , 61 , 786 ( 1974 ) . | we show that quantum optical systems preserving the total number of excitations admit a simple classification of possible resonant transitions ( including effective ) , which can be classified by analizying the free hamiltonian and the corresponding integrals of motion .
quantum systems not preserving the total number of excitations do not admit such a simple classification , so that an explicit form of the effective hamiltonian is needed to specify the allowed resonances .
the structure of the resonant transitions essentially depends on the algebraic propereties of interacting subsystems . |
magnetic reconnection is a process whereby the magnetic field line connectivity @xcite is modified due to the presence of a localized diffusion region .
this gives rise to a change in magnetic field line topology and a release of magnetic energy into kinetic and thermal energy .
reconnection of magnetic field lines is ubiquitous in laboratory , space and astrophysical plasmas , where it is believed to play a key role in many of the most striking and energetic phenomena .
the most notable examples of such phenomena include sawtooth crashes @xcite and major disruptions in tokamak experiments @xcite , solar and stellar flares @xcite , coronal mass ejections @xcite , magnetospheric substorms @xcite , coronal heating @xcite , and high - energy emissions in pulsar wind nebulae , gamma - ray bursts and jets from active galactic nuclei @xcite .
an exhaustive understanding of how magnetic reconnection proceeds in various regimes is therefore essential to shed light on these phenomena . in recent years , for the purpose of organizing the current knowledge of the reconnection dynamics that is expected in a system with given plasma parameters , a particular form of phase diagrams have been developed @xcite .
these diagrams classify what `` phase '' of magnetic reconnection should occur in a particular system , which is identified by two dimensionless plasma parameters , the lundquist number @xmath4 and the macroscopic system size @xmath5 here , @xmath6 indicates the system size in the direction of the reconnecting current sheet , @xmath7 is the alfvn speed based on the reconnecting component of the magnetic field upstream of the diffusion region , @xmath8 is the magnetic diffusivity , and @xmath9 is the relevant kinetic length scale .
this length scale corresponds to ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ) @xmath10 \rho_\tau = c_s/\omega_{ci } & \mbox{for guide - field reconnection}. \end{array } \right.\ ] ] of course , @xmath11 is the ion plasma frequency , @xmath12 is the ion cyclotron frequency , and @xmath13 is the sound speed based on both the electron and ion temperatures .
all the proposed phase diagrams @xcite exhibit a strong similarity and only a few minor differences .
they are useful to summarize some of the current knowledge of the magnetic reconnection dynamics , but they lack fundamental aspects that can greatly affect the reconnection process ( some caveats in the use of these diagrams have been discussed by @xcite ) . for example , they do not take into account the dependence of the reconnection process on the external drive or on the magnetic free energy available in the system .
an attempt to include these effects has been discussed by @xcite , who proposed to incorporate them by adjusting the definition of the lundquist number , eq .
( [ def_s ] ) , but this solution should be viewed only as a rough way to circumnavigate the problem .
a further issue is that these diagrams do not consider the evolution of the reconnection process and predict reconnection rates wich are always fast ( the estimated reconnection inflow is always a significant fraction of @xmath7 ) .
this , however , in not what is commonly observed in laboratory , space , and astrophysical plasmas , where magnetic reconnection exhibits disparate time scales and is often characterized by an impulsive behaviour , i.e. , a sudden increase in the time derivative of the reconnection rate ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ) . here
we propose a different point of view in which we include explicitly the effects of the external drive and the plasma viscosity ( neglected in all previous diagrams ) on the magnetic reconnection process by considering a four - dimensional parameter space .
then , in this four - dimensional diagram we identify specific domains of parameters where the reconnection process exhibits distinct dynamical evolutions .
in other words , in each of these domains the reconnection process goes through diverse phases characterized by different reconnection rates .
this analysis leads us to evaluate in greater detail the dynamical evolution of a forced magnetic reconnection process , while collisionless effects have not been taken into account in the present work .
we introduce the considered model of forced magnetic reconnection in sec . [
sec : taylor_model ] , whereas sec .
[ sec : conditions ] is devoted to the presentation of the possible evolutions of the system and the conditions under which these different evolutions occur . in sec .
[ sec : phase_diagrams ] we construct the parameter space diagrams that show which reconnection evolution is expected in a system with given characteristic parameters .
finally , open issues are discussed in sec .
[ sec : discussion ] .
magnetic reconnection in a given system is conventionally categorized as spontaneous or forced .
spontaneous magnetic reconnection refers to the case in which the reconnection arises by some internal instability of the system or loss of equilibrium , with the most typical example being the tearing mode .
forced magnetic reconnection instead refers to the cases in which the reconnection is driven by some externally imposed flow or magnetic perturbation . in this case ,
one of the most important paradigms is the so - called `` taylor problem '' , which consists in the study of the evolution of the magnetic reconnection process in a tearing - stable slab plasma equilibrium which is subject to a small amplitude boundary perturbation .
this situation is depicted in fig .
[ fig1 ] , where the shared equilibrium magnetic field has the form @xmath14 with @xmath15 , @xmath16 and @xmath17 as constants , and the perfectly conducting walls which bound the plasma are located at @xmath18 .
magnetic reconnection is driven at the resonant surface @xmath19 by a deformation of the conducting walls such that @xmath20 where @xmath21 is the perturbation wave number and @xmath22 is a small ( @xmath23 ) displacement amplitude .
the boundary perturbation is assumed to be set up in a time scale that is long compared to the alfvn time @xmath24 , with @xmath25 , but short compared to any characteristic reconnection time scale .
hence , the plasma can be considered in magnetostatic equilibrium everywhere except near the resonant surface at @xmath19 . the first and probably most important contribution to unveiling
the behaviour of forced magnetic reconnection in taylor s model is due to @xcite , who showed that very small amplitude boundary perturbations cause an initial linear phase in which a current sheet builds up at the resonant surface , and successive phases in which the reconnection process evolves according to a linear resistive regime and a nonlinear rutherford regime @xcite .
the scenario discussed by @xcite , which is characterized by a very slow evolution of the reconnection process , was complemented some years later by @xcite , who showed that larger perturbations may foster reconnection to proceed through the nonlinear regime according to a sweet - parker - like evolution @xcite , which only on the long time scale of resistive diffusion gives way to a rutherford evolution . the scenario outlined by @xcite
is characterized by a reconnection evolution faster than that presented by @xcite , but it could still be slow for very small values of plasma resistivity , since in both the sweet - parker - like @xcite and rutherford @xcite regimes , the reconnection rate is strongly dependent on the resistivity , which is known to be extremely small in many laboratory fusion plasmas and space / astrophysical plasmas . however , recent works @xcite have shown that relatively large boundary perturbations lead to a different reconnection evolution in plasmas with small resistivity and viscosity . in these cases , after a linear inertial phase and an initial nonlinear regime characterized by a gradually evolving current sheet , the reconnection suddenly enters into a fast reconnection regime distinguished by the disruption of the current sheet due to the development of secondary magnetic islands ( usually called plasmoids @xcite ) .
in addition to the works discussed above , which adopt a magnetohydrodynamic ( mhd ) description of the plasma , we emphasize that many other efforts have been devoted to investigate the tayor problem assuming mhd , two - fluid and kinetic descriptions ( see * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . indeed , the taylor problem has important applications besides being interesting from the point of view of basic physics . for instance , in laboratory fusion
plasmas the taylor model represents a convenient way to study magnetic reconnection processes driven by resonant magnetic perturbations , while in astrophysical plasmas this model can be adopted to study magnetic reconnection forced by the motions of photospheric flux tubes .
in this section we review the present understanding of the forced magnetic reconnection dynamics in taylor s model focusing on a visco - resistive plasma with @xmath1 greater than @xmath26 .
as shown by @xcite , this dynamics always starts with a linear inertial phase in which a current sheet builds up at the resonant surface and shrinks inversely in time .
concurrently , the current density at the @xmath27-point increases linearly in time .
the reconnection rate during this phase can be evaluated by recalling that the current density is proportional to the out - of - plane electric field at the @xmath27-point , which is equal to @xmath28 for @xmath29 @xcite . here
, @xmath30 stands for the magnetic flux function of the perturbed magnetic field in the reconnection plane ( @xmath31 ) , @xmath32 and @xmath33 indicate the characteristic time for viscous and resistive diffusion , respectively , while @xmath34 parametrizes the contribution of the external source perturbation to the gradient discontinuity of the magnetic flux function at the resonant surface .
it is important to point out that this phase is characterized by a non - constant-@xmath30 behaviour of the magnetic flux function across the island . depending on whether or not this property persists until the beginning of the nonlinear regime
, different scenarios may occur .
if the boundary perturbation is such that @xcite @xmath35 after the inertial phase the reconnection process evolves trough a visco - resistive phase , which is a linear regime characterized by a constant-@xmath30 behaviour , i.e. , the perturbed magnetic flux function can be treated as a constant in @xmath36 over the width of the reconnection layer . during this phase the reconnection rate
is given by @xmath37 where @xmath38 is the standard tearing stability parameter and @xmath39 is a characteristic time defined as @xcite @xmath40 with @xmath41 indicating the gamma function .
( [ reconnrate_vr ] ) is valid for @xmath42 @xcite and a magnetic island width much smaller than the linear layer width , i.e. , @xmath43 @xcite . if the perturbation is sufficient to drive the magnetic island into the nonlinear regime ( @xmath44 ) , the visco - resistive phase ends up into a rutherford evolution , whose island width growth is governed by the rutherford equation @xmath45 where @xmath46 @xcite .
this is a very slow reconnection evolution in which the reconnection rate can be evaluated analytically in the two limits @xcite @xmath47 @xmath48 where @xmath49 if the boundary perturbation is such that @xcite @xmath50 the non - constant-@xmath30 behaviour characteristic of the inertial phase lingers until the nonlinear regime is entered .
therefore , since in this case the magnetic island grows faster than the current can diffuse out of the reconnecting layer , the evolution of the reconnection process is distinguished by a strong current sheet at the resonant surface @xcite .
the reconnecting current sheet turns out to be stable if the boundary perturbation is such that @xcite @xmath51 where the multiplicative constant @xmath52 depends on the critical inverse aspect ratio of the reconnecting current sheet ( specified later ) . in this case
, the reconnection process follows a sweet - parker evolution ( modified by plasma viscosity @xcite ) , whose reconnection rate in taylor s model is @xcite @xmath53 finally , the sweet - parker type of evolution gives way to a rutherford evolution on the time scale of resistive diffusion .
if the boundary perturbation satisfies @xcite @xmath54 and also the condition @xcite @xmath55 the reconnection process does not reach a stable sweet - parker regime , but a different situation occurs . a gradually thinning current sheet evolves until its aspect ratio reaches the limit that allows the plasmoid instability to develop .
the growth of the plasmoids leads to the disruption of the current sheet , and therefore to a dramatic increase of the reconnection rate .
the reconnection rate during this plasmoid - dominated phase has been evaluated in a statistical steady state as @xcite @xmath56 where @xmath57 is the critical inverse aspect ratio of the reconnecting current sheet .
this quantity , whose value has been found to lie in the range @xmath58 - @xmath59 by means of numerical simulations @xcite , represents the threshold below which the reconnecting current sheet becomes unstable to the plasmoid instability @xcite .
in this section we illustrate the domain of existence of the three different scenarios described before with the help of appropriated parameter space maps . for the sake of clarity we state again the three type of reconnection evolutions
we are referring to : + ( 1 ) hahm - kulsrud scenario @xcite , + ( 2 ) wang - bhattacharjee scenario @xcite , + ( 3 ) our scenario @xcite .
+ since each of these scenarios includes different phases / regimes of reconnection , the concept of `` phase diagrams '' is intended here in a broader sense . due to this fact
, they could also be defined in a more general way as `` scenario diagrams '' .
this kind of diagrams can be constructed from the conditions summarized in the previous section .
therefore , the possible evolutions of the reconnection process may be organized in a four - dimensional parameter space map with @xmath60 , @xmath61 , @xmath62 , and @xmath63 on the four axes . however , due to the difficulty in the visualization of such a four - dimensional diagram , it is convenient to consider two - dimensional slices for fixed values of two of the four parameters
. let us first consider four two - dimensional slices with fixed values of the magnetic prandtl number and perturbation wave number .
assuming that the hahm - kulsrud scenario ( which occurs if @xmath64 ) may hold until @xmath65 , the corresponding diagrams for ( a ) @xmath66 , @xmath67 , ( b ) @xmath66 , @xmath68 , ( c ) @xmath69 , @xmath67 , ( d ) @xmath69 , @xmath68 , are shown in figs . [
fig2](a ) - [ fig2](d ) . from this plots
it is clear that the wang - bhattacharjee scenario is limited to a small range of values of the lundquist number and the source perturbation amplitude .
increasing values of the the magnetic prandtl number and perturbation wave number extend the domain of existence of this possible type of evolution of the system .
however , after a threshold value of the lundquist number ( identified by the intersection of the two black lines representing @xmath70 and @xmath71 ) , the wang - bhattacharjee scenario can not occur because it is not possible to obtain a stable sweet - parker - type evolution . in these cases ,
the hahm - kulsrud scenario is facilitated by very small perturbation amplitudes , whereas larger perturbations lead the system to a fast reconnection regime as described in sec .
[ our_scenario ] .
note that while previously proposed phase diagrams always predict fast reconnection @xcite , in clear contrast to what happens in nature , our diagrams show that reconnection proceeds very slowly ( region ( 1 ) ) if the source perturbation is not sufficiently large .
let us now examine the effect of the plasma viscosity by considering the domain of existence of the different scenarios as a function of the parameters @xmath2 and @xmath1 .
[ fig3](a ) shows the functions @xmath70 and @xmath71 for @xmath72 and @xmath73 . for @xmath74
the threshold for the plasmoids formation coincides with that for the nonlinear evolution characterized by a strong reconnecting current sheet .
therefore , for @xmath74 an increase of the amplitude perturbation @xmath2 drives the system directly from scenario ( 1 ) to scenario ( 3 ) .
this situation is depicted in fig .
[ fig3](b ) , where it is clearly shown that the increase of the magnetic prandtl number has the effect of making possible or extending the domain of existence of scenario ( 2 ) . to clarify the effect of the plasma viscosity we also delineate the boundaries of the diverse evolutions ( 1)-(3 ) in a parameter space map @xmath75 ( as in fig .
[ fig2 ] ) for fixed @xmath73 but different values of @xmath1 .
this is shown in fig .
[ fig4 ] for @xmath76 .
the increase of the magnetic prandtl number extends the domain of existence of the slow reconnection scenario ( 1 ) at the expense of the fast reconnection scenario ( 3 ) .
the area of existence of scenario ( 2 ) remains almost unchanged , but shifted towards higher values of the lundquist number .
we now examine in more detail how the possible evolutions of the forced magnetic reconnection process depend on the wave number of the boundary perturbation . fig .
[ fig5](a ) shows the thresholds @xmath70 and @xmath71 as a function of @xmath3 for fixed values of @xmath72 and @xmath77 . below a critical perturbation wave number @xmath78 ( corresponding to @xmath79 for the fixed parameters used in fig .
[ fig5](a ) ) , every time the non - constant-@xmath30 magnetic island pass into the nonlinear regime , the evolution of the system leads to the plasmoid - dominated phase predicted in scenario ( 3 ) .
the domains of existence of the possible evolutions ( 1)-(3 ) are illustrated in fig .
[ fig5](b ) .
the scenario discussed by wang and bhattacharjee happens only for a small range of ( @xmath80 ) parameters .
not also that scenario ( 3 ) is facilitated for @xmath81 , while scenario ( 1 ) may occur for large amplitude boundary perturbations if @xmath82 .
to better evaluate the effects of @xmath3 on the possible evolutions of the reconnection process , we plot in fig .
[ fig6 ] the boundaries between the scenarios ( 1)-(3 ) in a parameter space map @xmath75 ( as in figs .
[ fig2 ] and [ fig4 ] ) for fixed @xmath77 but different values of @xmath3 .
the maximum area of existence of scenario ( 2 ) occurs for @xmath83 , while for @xmath84 and @xmath85 the scenario ( 2 ) appears for a very limited range of ( @xmath80 ) parameters .
note also that the scenario ( 3 ) is greatly facilitated in the case of very large perturbation wave numbers ( @xmath84 ) while the scenario ( 3 ) is facilitated by relatively large amplitude perturbations with ( @xmath86 ) .
the introduction of a new type of phase / scenario diagrams that include explicitly the effects of the external drive has allowed us to graphically organize in a detailed way the possible evolutions of forced magnetic reconnection processes in collisional plasmas .
in contrast to previous versions of the phase diagrams @xcite , this new representation highlights regions of the parameter space ( @xmath80 , @xmath0 , @xmath1 ) in which reconnection is a slow diffusive process ( sec .
[ sechk ] ) in addition to regions where reconnection can be fast ( secs .
[ secwb ] and [ our_scenario ] ) .
we recall that by fast we mean that the out - of - plane inductive electric field at the @xmath27-point is a significant fraction of the one evaluated upstream of the reconnection layer .
we also emphasize that this kind of diagrams respond to the criticism moved by @xcite concerning the fact that these diagrams are not able of taking into account the dynamical evolution of the reconnection process from a slow to a fast regime inside of a given region of the parameter space . indeed ,
scenarios ( 1)-(3 ) describe the forced magnetic reconnection process from the current sheet formation all the way to their specific nonlinear evolution .
we would like to remark that while the proposed parameter space diagrams represent a valid way to summarize the current knowledge of the forced magnetic reconnection dynamics in a collisional plasma , there are a number of conditions that may significantly affect the reconnection process which have not been addressed in this paper .
for instance , two - fluid / kinetic effects should be considered if the length scale associated with the width of the reconnecting current sheet becomes of the order or smaller than the characteristic length scales of these effects .
in fact , in antiparallel reconnection ( i.e. , in the absence of a guide magnetic field ) , hall effects @xcite are known to enhance the reconnection rate , as well as effects associated to finite electron inertia @xcite , electron pressure @xcite and ion gyration @xcite are known to increase the reconnection rate in the presence of a strong guide field .
we would like to remark also that a common condition in many physical systems is the presence of velocity flows , which are known to suppress the reconnection @xcite or to alter the reconnection rate @xcite . in this case
our analysis should be extended by considering also the effects of a plasma flow on the reconnection dynamics .
similarly , also the effects of turbulence should be considered @xcite in order to obtain a more complete description of the magnetic reconnection dynamics .
finally , it is important to recall that all the presented diagrams of magnetic reconnection are based on two - dimensional models and simulations . at present , the knowledge of how magnetic reconnection evolve in large three - dimensional systems is still far behind our understanding of what happens in two - dimensional systems .
therefore , despite the great progress achieved in recent years @xcite , other work is needed in this direction before we can implement a phase diagram description of three - dimensional magnetic reconnection .
the authors would like to acknowledge fruitful conversations with richard fitzpatrick and enzo lazzaro .
this work was carried out under the contract of association euratom - enea and was also supported by the u.s .
department of energy under contract no .
de - fg02 - 04er-54742 . | recent progress in the understanding of how externally driven magnetic reconnection evolves is organized in terms of parameter space diagrams .
these diagrams are constructed using four pivotal dimensionless parameters : the lundquist number @xmath0 , the magnetic prandtl number @xmath1 , the amplitude of the boundary perturbation @xmath2 , and the perturbation wave number @xmath3 .
this new representation highlights the parameters regions of a given system in which the magnetic reconnection process is expected to be distinguished by a specific evolution .
contrary to previously proposed phase diagrams , the diagrams introduced here take into account the dynamical evolution of the reconnection process and are able to predict slow or fast reconnection regimes for the same values of @xmath0 and @xmath1 , depending on the parameters that characterize the external drive , never considered so far .
these features are important to understand the onset and evolution of magnetic reconnection in diverse physical systems . |
according to the standard quantum mechanics ( sqm ) , the complete description of a system of particles is provided by its wavefunction .
the empirical predictions of sqm follow from a mathematical formalism which makes no use of the assumption that matter consists of particles pursuing definite tracks in space - time .
it follows that the results of the experiments designed to test the predictions of the theory , do not permit us to infer any statement regarding the particle not even its independent existence .
but , in the bohmian quantum mechanics ( bqm ) , the additional element that is introduced apart from the wavefunction is the particle position , conceived in the classical sense as pursuing a definite continuous track in space - time [ 1 - 3 ] .
the detailed predictions made by this causal interpretation explains how the results of quantum experiments come about but it is claimed that they are not tested by them . in fact
when bohm @xcite presented his theory in 1952 , experiments could be done with an almost continuous beam of particles , but not with individual particles .
thus , bohm constructed his theory in such a fashion that it would be impossible to distinguish observable predictions of his theory from sqm .
this can be seen from bell s comment about empirical equivalence of the two theories when he said:_it _ [ the de broglie - bohm version of non - relativistic quantum mechanics ] _ is experimentally equivalent to the usual version insofar as the latter is unambiguous_"@xcite .
however , could it be that a certain class of phenomena might correspond to a well - posed problem in one theory but to none in the other ? or might the additional particles and definite trajectories of bohm s theory lead to a prediction of an observable where sqm would just have no definite prediction to make ? to draw discrepancy from experiments involving the particle track , we have to argue in such a way that the observable predictions of the modified theory are in some way functions of the trajectory assumption . the question raised here
is whether the de broglie - bohm particle law of motion can be made relevant to experiment . at first
, it seems that definition of time spent by a particle within a classically forbidden barrier provides a good evidence for the preference of bqm .
but , there are difficult technical questions , both theoretically and experimentally , that are still unsolved about this tunnelling times @xcite .
a recent work indicates that it is not practically feasible to use tunnelling effect to distinguish between the two theories @xcite . on the other hand , englert _ et al .
_ @xcite and scully @xcite have claimed that in some cases bohm s approach gives results that disagree with those obtained from sqm and , in consequence , with experiment .
again , at first dewdney _
_ @xcite and then hiley _
@xcite showed that the specific objections raised by englert _ et al . _ and scully can not be sustained .
furthermore , hiley believes that no experiment can decide between the standard interpretation and bohm s interpretation .
however , vigier @xcite , in his recent work , has given a brief list of new experiments which suggest that the u(1 ) invariant massless photon assumed properties of light within the standard interpretation , are too restrictive and that the o(3 ) invariant massive photon causal de broglie - bohm interpretation of quantum mechanics , is now supported by experiments .
furthermore , in some of the recent investigation , some feasible experiments have been suggested to distinguish between sqm and bqm @xcite . in one work ,
ghose indicated that although bqm is equivalent to sqm when averages of dynamical variables are taken over a gibbs ensemble of bohmian trajectories , the equivalence breaks down for ensembles built over clearly separated short intervals of time in specially entangled two - bosonic particle systems @xcite .
another one @xcite is an extension of ghose s work to show disagreement between sqm and bqm in a two - particle system with an unentangled wavefunction , particularly at the statistical level .
further discussion of this subject can be found in [ 13 - 15 ] . in that experiment , to obtain a different interference pattern from sqm , we must deviate the source from its geometrically symmetric location . in this investigation , we are offering a new thought experiment which can decide between sqm and bqm . here , the deviation of the source from its geometrical symmetric location is not necessary and we have used a system consisting two correlated particles with an entangled wavefunction . in the following section ,
we have introduced a two double - slit experimental set - up . in section 3 ,
bohm s interpretation is used to find some observable results about our suggested experiment .
predictions of the standard interpretation and their comparison with bohmian predictions is examined in section 4 . in section 5 , we have used selective detection and have compared sqm and bqm with our thought experiment at the ensemble level of particles , and we state our conclusion in section 6 .
and then they pass through slits @xmath0 and @xmath1 or @xmath2 and @xmath3 .
finally , they are detected on @xmath4 and @xmath5 screens , simultaneously .
it is necessary to note that dotted lines are not real trajectories.,width=566,height=302 ]
to distinguish between sqm and bqm we consider the following experimental set - up .
a pair of identical non - relativistic particles with total momentum zero labelled by 1 and 2 , originate from a point source s that is placed exactly in the middle of a two double - slit screens as shown in fig .
we assume that the intensity of the beam is so low that during any individual experiment we have only a single pair of particles passing through the slits and the detectors have the opportunity to relate together for performing selective detection process .
in addition , we assume that the detection screens @xmath4 and @xmath5 register only those pairs of particles that reach the two screens simultaneously .
thus , we are sure that the registration of single particles is eliminated from final interference pattern . the detection process at the screens @xmath4 and @xmath5 may be nontrivial but they play no causal role in the basic phenomenon of the interference of particles waves @xcite
. in the two - dimensional system of coordinates @xmath6 whose origin @xmath7 is shown , the center of slits lie at the points @xmath8 .
the wave incident on the slits will be taken as a plane of the form @xmath9}e^{-iet/\hbar}\ ] ] where @xmath10 is a constant and @xmath11 is the total energy of the system of the two particles .
the plane wave assumption comes from large distance between source @xmath7 and double - slit screens . to avoid the mathematical complexity of fresnel diffraction at a sharp - edge slit
, we suppose the slits have soft edges that generate waves having identical gaussian profiles in the @xmath12-direction while the plane wave in the @xmath13-direction is unaffected @xcite .
the instant at which the packets are formed will be taken as our zero of time . therefore , the four waves emerging from the slits @xmath0 , @xmath2 , @xmath3 and @xmath1 are initially @xmath14}\ ] ] @xmath15}\ ] ] where @xmath16 is the half - width of each slit . at time
@xmath17 the general total wavefunction at a space point @xmath6 of our considered system for bosonic and fermionic particles is given by @xmath18\end{aligned}\ ] ] with @xmath19}\ ] ] @xmath20}\ ] ] where @xmath21 is a reparametrization constant that its value is unimportant in this paper and @xmath22 where @xmath23 and @xmath24 , according to bqm , are initial group velocities corresponding to each particle in the @xmath13- and @xmath12-directions , respectively .
in addition , the upper and lower sings in the total wavefunction refer to symmetric and anti - symmetric wavefunction under exchange of particle 1 to particle 2 , corresponding to bosonic and fermionic property , while in the equations ( 5 ) and ( 6 ) they refer to upper and lower slits , respectively . in the next section ,
we have used bqm to drive some results of this experiment .
in bqm , the complete description of a system is given by specifying the position of the particles in addition to their wavefunction which has the role of guiding the particles according to following guidance condition for @xmath25 particles , with masses @xmath26 @xmath27 where @xmath28 and @xmath29 is a solution of schr@xmath30dinger s wave equation .
thus , instead of sqm with indistinguishable particles , in bqm the path of particles or their individual histories distinguishes them and each one of them can be studied separately @xcite .
in addition , belousek @xcite in his recent work , concluded that the problem of bohmian mechanical particles being statistically ( in)distinguishable is a matter of theory choice underdetermined by logic and experiment , and that such particles are in any case physically distinguishable . for our considered experiment ,
the speed of the particles 1 and 2 in the @xmath12-direction is given , respectively , by @xmath31 @xmath32 with the replacement of @xmath33 from ( 4 ) , ( 5 ) and ( 6 ) , we have @xmath34\psi_{a_{1}}\psi_{b^{'}_{2}}\cr & \pm&[-2(y_{1}+y+u_{y}t)/4\sigma_{0}\sigma_{t}-ik_{y}]\psi_{a_{2}}\psi_{b^{'}_{1}}\cr & + & [ -2(y_{1}+y+u_{y}t)/4\sigma_{0}\sigma_{t}-ik_{y}]\psi_{b_{1}}\psi_{a^{'}_{2}}\cr & \pm&[-2(y_{1}-y - u_{y}t)/4\sigma_{0}\sigma_{t}+ik_{y}]\psi_{b_{2}}\psi_{a^{'}_{1}}]\}\end{aligned}\ ] ] @xmath35\psi_{a_{1}}\psi_{b^{'}_{2}}\cr & \pm&[-2(y_{2}-y - u_{y}t)/4\sigma_{0}\sigma_{t}+ik_{y}]\psi_{a_{2}}\psi_{b^{'}_{1}}\cr & + & [ -2(y_{2}-y - u_{y}t)/4\sigma_{0}\sigma_{t}+ik_{y}]\psi_{b_{1}}\psi_{a^{'}_{2}}\cr & \pm&[-2(y_{2}+y+u_{y}t)/4\sigma_{0}\sigma_{t}-ik_{y}]\psi_{b_{2}}\psi_{a^{'}_{1}}]\}\end{aligned}\ ] ] where , for example , the short notation @xmath36 is used .
furthermore , from ( 5 ) and ( 6 ) it is clear that @xmath37 which indicates the reflection symmetry of @xmath33 with respect to the @xmath13axis . utilizing this symmetry in ( 13 ) and ( 14 ) , we can see that @xmath38 which are valid for both bosonic and fermionic particles .
relations ( 16 ) show that if @xmath39 , then the speed of each particles in the @xmath12-direction is zero along the symmetry axis @xmath13 .
this means that none of the particles can cross the @xmath13-axis nor are they tangent to it , provided both of them are simultaneously on this axis .
similar conclusions can be found in some other works [ for example , 8 , 9 , 11 - 13 ] .
it can be seen that there is the same symmetry of the velocity about the @xmath13-axis as for an ordinary double - slit experiment @xcite .
if we consider @xmath40 to be the vertical coordinate of the center of mass of the two particles , then we can write @xmath41 solving the equation of motion ( 17 ) , we obtain the path of the @xmath12-coordinate of the center of mass @xmath42 if at @xmath43 the center of mass of the two particles is exactly on the @xmath13-axis , then @xmath44 , and the center of mass of the particles will always remain on the @xmath13-axis . thus , according to bqm , the two particles will be detected at points symmetric with respect to the @xmath13-axis , as shown in fig
it seems that calculation of quantum potential can give us another perspective of this experiment . as we know
, to see the connection between the wave and particle , the schr@xmath30dinger equation can be rewritten in the form of a generalized hamilton - jacobi equation that has the form of the classical equation , apart from the extra term @xmath45 where function q has been called the quantum potential @xcite .
but , it is clear that the calculation and analysis of q , by using our total wavefunction ( 4 ) , is not very simple . on the other hand
, we can use the form of newton s second law , in which the particle is subject to a quantum force @xmath46 in addition to the classical force @xmath47 @xcite , namely @xmath48 now , if we utilize the equation of motion of the center of mass @xmath12-coordinate ( 18 ) and the equation ( 20 ) , we shall obtain the quantum potential for the center of mass motion @xmath49 .
thus , we can write @xmath50 @xmath51 where the result of equation ( 21 ) is clearly due to motion of plane wave in the @xmath13-direction .
in addition , we assume that @xmath52 in our experiment . thus , our effective quantum potential is only a function of @xmath12-variable and it has the form @xmath53 if @xmath44 , the quantum potential for the center of mass of the two particles is zero at all times and it remains on the @xmath13-axis .
however , if @xmath54 , then the center of mass can not touch or cross the @xmath13-axis .
these conclusions are consistent with our earlier result ( eq .
so far , we have been studying the results obtained from bqm at the individual level .
but it is well known from sqm that the probability of simultaneous detection of two particles at @xmath55 and @xmath56 , at the screen @xmath4 and @xmath5 is equal to @xmath57 the parameter @xmath58 , which is taken to be small , is a measure of the size of the detectors .
it is clear that the probabilistic prediction of sqm is in disagreement with the symmetrical prediction of bqm , because sqm predicts that probability of asymmetrical detection at the individual level of pair of particles can be different from zero , in opposition to bqm s symmetrical predictions .
in addition , based on sqm s prediction , the probability of finding two particles at one side of the @xmath13-axis can be nonzero while we showed that bqm s prediction forbids such events in our experiment . in other words
, its probability must be exactly zero .
thus , if we provide necessary arrangements to perform this experiment , we must abandon one of the two theories or even both as a complete description of the universe .
now the question arises as to whether this difference persists if we deal with an ensemble of pair of particles ? to answer this question , we consider an ensemble of pair of particles that have arrived at the detection screens @xmath4 and @xmath5 at different times @xmath59 .
it is well known that , in order to ensure the compatibility between sqm and bqm for ensemble of particles , bohm added a further postulate to his three basic and consistent postulates @xcite .
based on this further postulate , the probability that a particle in the ensemble lies between @xmath60 and @xmath61 at time @xmath17 is given by @xmath62 thus , using bqm , the probability of simultaneous detection for all pairs of particles of the ensemble arriving at the two screens at different instant of time @xmath59 , with @xmath44 , is @xmath63 where every term in the sum shows only one pair arriving on the screens @xmath4 and @xmath5 at the point @xmath64 and time @xmath59 , weighted by the corresponding density @xmath65 .
if all times of @xmath59 are taken to be @xmath17 , then the summation on @xmath66 can be changed to an integral over all paths that cross the screens @xmath4 and @xmath5 at that time .
now , we can consider the joint probability of two points @xmath55 and @xmath56 on the two screens at time @xmath17 that are not symmetric about the @xmath13-axis , but we know that they are not detected simultaneously .
then , one can obtain the probability of detecting two particles at two arbitrary points @xmath55 and @xmath56 @xmath67 which is similar to the prediction of sqm ( eq .
( 24 ) ) but obtained in a bohmian way @xcite .
thus , it appears that for a geometrically symmetric arrangement , the possibility of distinguishing the two theories at the statistical level is impossible , as was expected [ 1 - 3 , 9 , 17 ] .
in the previous section , we have shown that sqm and bqm have different predictions for our suggested experiment , at the individual level . since sqm talks about individual events in probabilistic terms , the existence of different predictions by the two theories at the individual level is not a strange result .
on the other hand , we have seen that the two theories , for a geometrically symmetric arrangement , are consistent at the ensemble level . here , one can ask whether the individual level is the only area to distinguish between the two theories and whether the disagreement between them can not appear at the ensemble level . in this section
, we answer this question in negative , and we shall provide conditions under which sqm can be interpreted as a vague theory at the ensemble level . we have seen that the assumption of @xmath44 is not in contradiction with the statistically results of sqm and , in consequence , with experiment .
thus , we can assume that initially each particle in the source is statistically distributed according to the absolute square of the wavefunction , but , this distribution is completely symmetric so that the @xmath12-coordinate of the center of mass is on the @xmath13-axis .
if we can prepare such a special source with two correlated particles , we can try to do our experiment in the following fashion : particles emitted from the source @xmath7 into the right hand side of the experimental set - up can pass through slits @xmath0 or @xmath2 .
since we have assumed that the total momentum of the pair of particles is zero , if one of the particles goes through the slit @xmath0 for instance , the other particle must go through the slit @xmath1 on the left hand side of the experimental set - up .
based on bqm and using equation ( 16 ) , we infer that the particle passing through the slit @xmath0 must be detected on the upper half plane of the @xmath13-axis on the screen @xmath4 .
the same thing must occur for the other particle that passes through @xmath1 , but in the lower half plane of the @xmath13-axis on the @xmath5 screen . using this prediction , we assume that only those particles arriving at @xmath5 are recorded for which there is a simultaneous detection of the other particle at the upper side on @xmath4 .
we called this special detection , in which some of the selected particles are recorded , a selective detection .
thus , based on the prediction power of bqm , we will record two particles symmetric with respect to the @xmath13-axis , for each emitted pairs of particles .
if we wait to record an ensemble of particles , we will see an interference pattern of particles on the lower half plane of the @xmath5 screen . on the other hand , based on sqm , the probability of finding a particle at any point on the @xmath5 screen , even at the upper side , is nonzero and
there is no compulsion to detect pairs of particles symmetrically on the two sides of the @xmath13-axis , as it can be seen from equation ( 24 ) and is depicted in fig .
1 . therefore ,
if we accept that sqm is still efficient and unambiguous for the selective detection , the interference pattern will be seen on the whole screen @xmath5 , particularly at the upper side of it , at the ensemble level .
consequently , we shall have observable results to distinguish the two theories , sqm and bqm .
one can argue that @xmath68 can not have a well - defined position and it must be distributed according to born s principle . however , we will show that this objection can not alter our obtained results .
assume that , @xmath69 but @xmath70 . if we provide conditions in which @xmath71 is very small and @xmath72 , we can still detect particles symmetrically with respect to the @xmath13-axis , with a good approximation . to obtain symmetrical detection about the @xmath13-axis with reasonable approximation , the center of mass variation from the @xmath13-axis must be smaller than the distance between any two neighboring maxima , that is @xmath73 where @xmath74 is the de broglie wavelength .
for conditions @xmath72 , @xmath75 and using equation ( 18 ) , one can obtain @xmath76 therefore , if we use a source with @xmath77 , we shall obtain @xmath78 for each individual observation , and our symmetrical detection can be maintained with a good approximation . it is evident that , if one considers @xmath79 , as was done in @xcite , the incompatibility between the two theories will be disappeared .
but , we believe that , instead of the usual one - particle two - slit experiment with @xmath80 , our correlated two - particle system provides a new situation in which we can adjust @xmath68 independent of @xmath16 , so that @xmath81 although it is obvious that @xmath82 , but position correlation between the two entangled particles cause that they always satisfy equation ( 30 ) .
furthermore , if it is assumed that @xmath68 is statistically distributed , another problem can be raised , which is mentioned by marchildon @xcite .
we have shown that , if both particles are simultaneously on the @xmath13-axis , both velocities in the @xmath12-direction vanish , and neither particle could cross or be tangent to the @xmath13-axis . however , under @xmath70 condition , pairs of particles can not be simultaneously on the @xmath13-axis and we have not the aforementioned constraint on the motion of particles ( relations ( 16 ) ) . however , using our selective detection , we can still obtain our last result , because the center of mass of the two particles are on the @xmath13-axis with reasonable approximation .
it is clear that , under such a condition , one can not claim that the particle detected at the upper ( lower ) side must have passed through upper ( lower ) slit , in spite of @xmath44 condition . to confirm these results ,
it is worth noting that , durr _ et al . _
@xcite argue that the selective detection can alter the statistical predictions of the two theories:_note that by selectively forgetting results we can dramatically alter the statistics of those that we have not forgotten .
this is a striking illustration of the way in which bohmian mechanics does not merely agree with the quantum formalism , but , eliminating ambiguities , clarifies , and sharpens it . _ " . elsewhere
@xcite , we have utilized another kind of selective detection by which we could alter statistical prediction of sqm , using bqm for an interference device that contains two unentangled particles .
in this investigation , we have suggested an experiment to distinguish between sqm and bqm .
in fact , we believe that some particular experiments that for one reason or another have not yet been performed can decide between them . thus , it has been shown that a two double - slit experiment set - up , along with a source of two identical non - relativistic particles with total momentum of zero , emitted at suitable time intervals , has the following characteristics : + 1 ) the suggested experiment will yield different observable predictions for sqm and bqm , at the individual level . + 2 ) the two theories yield the same interference pattern at the ensemble level without using a selective detection , as is expected .
+ 3 ) since in bqm the particles are distinguishable and their past history are known , using selective detection , it has been shown that either the two theories will predict different results at the statistical level , or that bqm has more predictive power than sqm .
it is shown that selective detection can be considered as a tool for arriving to a new realm in which trajectory interpretation is sharply formulated , while the standard interpretation is ambiguous and silent even at the ensemble level .
dillo 83 d. bohm , part i , _ phys .
_ * 85 * 166 ( 1952 ) ; part ii , * 85 * 180 ( 1952 ) .
holland , _ the quantum theory of motion _ , cambridge university press , cambridge , 1993 .
cushing , _ quantum mechanics : historical contingency and the copenhagen hegemony _ , the university of chicago press , ltd . ,
london , 1994 .
bell , _ speakable and unspeakable in quantum mechanics _ , cambridge university press , cambridge , 1987 .
m. abolhasani and m. golshani , _ phys .
rev . a _ * 62 * 12106 ( 2000 )
englert , m.o .
scully , g. sussman and h. walther , _ z. naturf . _ * 47*a 1175 ( 1992 ) .
scully , _ phys .
scripta _ , t*76 * 41 ( 1998 ) . c. dewdney , l. hardy and e.j .
squires , _ phys .
lett . a _ * 184 * 6 ( 1993 ) .
hiley , r.e . callaghan and o.j.e .
maroney , quant - ph/0010020 .
vigier , _ phys . lett .
a _ * 270 * 221 ( 2000 ) .
p. ghose , quant ph/0001024 and quant - ph/0003037 .
m. golshani and o. akhavan , quant - ph/0009040 .
m. golshani and o. akhavan , quant - ph/0103100 .
p. ghose , quant - ph/0102131 .
l. marchildon , quant - ph/0101132 .
belousek , _ found .
phys . _ * 30 * 153 ( 2000 ) .
d. durr , s. goldstein and n. zanghi , _ j. stat .
* 67 * 843 ( 1992 ) . | the significance of proposals that can predict different results for standard and bohmian quantum mechanics have been the subject of many discussions over the years . here , we suggest a particular experiment ( a two double - slit experiment ) and a special detection process , that we call selective detection , to distinguish between the two theories . using our suggested experiment , it is shown that the two theories predict different observable results at the individual level for a geometrically symmetric arrangement . however ,
their predictions are the same at the ensemble level . on the other hand
, we have shown that at the statistical level , if we use our selective detection , then either the predictions of the two theories differ or where standard quantum mechanics is silent or vague , bohmian quantum mechanics makes explicit predictions .
pacs number(s ) : 03.65.bz + |
the role of non - local quantum correlations , more familiarly known as entanglement , in modern quantum theory can not be overstated @xcite . on the one hand , entanglement lies at the heart of quantum information theory @xcite , where it is a crucial ingredient to computation and communication schemes . on the other hand ,
it is intricately related to phenomena such as decoherence @xcite and quantum phase transitions in many - body systems @xcite .
one has come to realize that entanglement is also a resource that can for instance be purified , shared , and possibly irreversibly lost , and should therefore not only be detectable , but also quantifiable @xcite .
one way of doing so is by virtue of entanglement measures @xcite .
these are mathematical functions mapping quantum states to the set of real numbers .
while there is no unique or strict definition for the notion of an entanglement measure , there are a set of properties which are commonly regarded useful , e.g. , that the measure is zero only for separable states and is invariant under local unitary transformations .
another important property which we will assume throughout this work is monotonicity : an entanglement measure must not increase ( on average ) under any protocol involving only local unitary transformations and classical communication . in the following
, we will use the terms ` entanglement measure ' and ` entanglement monotone ' interchangeably . rather understandably , it is difficult to capture all properties of even a pure entangled state with just a single real number , especially in the setting of higher - dimensional and multi - partite systems .
it is thus no surprise that there is quite a number of proposed entanglement monotones of various levels of complexity , generality , and the ability to capture different aspects of entangled states more or less successfully than others .
as indicated previously , most of these entanglement monotones share the fact that they are conveniently defined only for pure states , namely as a function of the amplitudes of the state expressed in a certain standard basis .
the situation becomes more involved in the case of mixed states , where classical and quantum correlations need to be distinguished from one another .
given a density matrix @xmath0 , it is not sufficient to simply calculate the average entanglement of a given decomposition , because this decomposition is not unique .
since there are in general infinitely many ways to write a density matrix as a sum of projectors onto pure states , only the infimum of entanglement over all these possible decompositions can make a reliable statement about the quantum correlations in @xmath0 , because there might be a decomposition of @xmath0 in which all pure states are separable and the total entanglement hence would vanish . taking this infimum of an averaged pure - state entanglement monotone over all decompositions of @xmath0
is called ` convex - roof construction ' or ` convex - roof extension ' of that monotone @xcite .
notably , the thereby obtained measure for mixed states is again an entanglement monotone . calculating the convex - roof for a generic quantum state
is considered extremely difficult @xcite .
in fact , even deciding whether or not a merely bipartite mixed state is separable is a hard problem itself which has no known general solution in hilbert space dimensions larger than six @xcite . in this work ,
we present the computer programs we have written and successfully applied previously to calculate such convex - roof entanglement measures of multi - partite mixed states @xcite .
while we have already described the theory behind our algorithms to some extent in an earlier publication @xcite , we complete this work by publishing here the full source code in the form a user - friendly high - level library called ` libcreme ` .
the package is to a large part written in the matlab programming language @xcite , but great care has been taken to make the library fully compatible with gnu octave @xcite , a free and open - source matlab clone . for the sake of simplicity , we will refer to their common language as m - script .
additionally , functions which have been identified as crucial bottlenecks in terms of execution speed are provided in the form of fast c extensions and have been adapted to be easily callable from matlab and octave through their native c and c++ interfaces , respectively .
while the library already comes with the ability to evaluate a choice of popular entanglement monotones , it is easily extend to calculate user - specified functions .
all that needs to be implemented is the entanglement measure itself and its gradient with respect to the real and imaginary parts of the quantum state vector @xcite .
the library is written in a self - contained and consistent way , making it extremely easy to use in practice and to experiment with different settings , measures , and optimization algorithms .
furthermore , we provide convenience functions hiding most of the steps required to arrive at function handles ready to be optimized .
this essentially simplifies the calculation of a convex - roof entanglement measure to a one - line task .
table [ table : list of functions ] lists each function provided in the library together with a short description of its meaning .
p4.8 cm p11 cm + _ entanglement measures _ & + ` convexsum ` & convex sum parameterized by a stiefel matrix + ` grad_convexsum ` & gradient of convex sum + ` eof2x2 ` & entanglement of formation for 2 qubits ( analytically exact result ) + ` entropyofentanglement ` & entropy of entanglement + ` grad_entropyofentanglement ` & gradient of entropy of entanglement + ` meyer_wallach ` & meyer - wallach measure + ` grad_meyer_wallach ` & gradient of meyer - wallach measure + ` tangle ` & tangle + ` grad_tangle ` & gradient of tangle + + _ optimization routines _ & + ` cg_min ` & conjugate - gradient method + ` bfgs_min ` & bfgs quasi - newton method + ` minimize1d_exp ` & minimization along a geodesic on the stiefel manifold + ` minimize1d_lin ` & minimization along a line in euclidean space + ` get_termination_criteria ` & helper function to check and handle termination criteria for the optimization algorithms + + _ utilities _ & + ` randdensitymatrix ` & random density matrix + ` randstate ` & random pure quantum state + ` randunitarymatrix ` & random stiefel matrix + ` decomposeunitary ` & get angles parameterizing a stiefel matrix + ` dimst ` & dimension of stiefel manifold + ` densityeig ` & get eigendecomposition of a density matrix in the form required by many functions within the library + ` psdecomposition ` & get pure - state decomposition parameterized by a stiefel matrix + ` createconvexfunctions ` & create convex - sum function handles for use with ` cg_min ` + ` createehfunctions ` & create convex - sum function handles for use with ` bfgs_min ` + ` grad_eh_adapt ` & adapter function to calculate the gradient of a convex sum parameterized by stiefel matrix angles
+ ` buildunitary ` & build a complex stiefel matrix from a parameterization vector + ` grad_buildunitary ` & gradient of the above function + ` ptrace ` & partial trace over any subsystems of arbitrary ( finite ) dimensions + ` completegramschmidt ` & helper function for numerical stability used within ` cg_min ` + + _ examples _ & + ` example_eofisotropic ` & main script to run the example from sec .
[ sec : ex_eof ] + ` eofisotropic ` & entanglement of formation of an ` isotropic ' density matrix ( analytically exact result ) + ` example_tangleghzw ` & main script to run the example from sec .
[ sec : ex_tangle ] + ` tangleghzw ` & tangle of ghz / w mixture ( analytically exact result ) + + we would briefly like to mention two other numerical libraries dealing with quantum computing and entanglement .
one is the freely available ` quantum information package ' by t. cubitt @xcite written in m - script as well .
the other one is the cpc library feynman ( catalog identifier adwe_v5_0 ) written by t. radtke and s. fritzsche @xcite for maple .
quantum states obtained from calculations and simulations within these libraries can conveniently be anaylized further using ` libcreme ` s ability to calculate more complex entanglement measures .
the paper is organized as follows : in sec .
[ sec : measures ] , we briefly list and discuss the default entanglement measures coming along with ` libcreme ` .
[ sec : background ] reviews the theory of convex - roof entanglement measures and how to address their calculation on a computer .
[ sec : algorithms ] describes the two central algorithms in ` libcreme ` to solve the optimization problem related to the evaluation of such measures . in sec .
[ sec : examples ] , we discuss two complete examples demonstrating the usage of the library , and sec .
[ sec : conclusions ] concludes the work .
note that the focus of this manuscript lies mainly on the functionality of the library : we have tried to provide short code examples throughout the work for all important functions and concepts in a user - friendly bottom - up way .
these snippets are all valid m - script ( including the line breaks `` ... ` ' which we sometimes use due to spacial restrictions ) and can be copied directly into matlab or octave . finally , we would like to mention that all functions in ` libcreme ` are documented , and more information about them can be inquired by calling `` help ` _ function_name _ ' .
we start the description of our library by listing the entanglement measures currently implemented .
note that pure quantum states , such as the arguments of functions calculating entanglement monotones , are always expected to be represented as column vectors in the standard computational basis . in practice , this means that the @xmath1 orthonormal basis states @xmath2 of each @xmath1-dimensional subsystem ( where @xmath1 may be different for different subsystems ) are always chosen as @xmath3_i = \partial f / \partial { \mathop\mathrm{re}}x_i + \mathfrak{i}\ , \partial f / \partial { \mathop\mathrm{im}}x_i$ ] , where @xmath4 is the imaginary unit .
analytical expressions for all gradients of the measures discussed in this section can be found in [ app : derivatives of measures ] .
the entropy of entanglement @xcite is an entanglement monotone for bipartite quantum systems of arbitrary dimensions .
it is defined as the von neumann - entropy of the reduced density matrix , i.e. , @xmath5,\ ] ] where @xmath6 denotes the partial trace of @xmath7 and [ 0 ; 1 ] .
% % note that in m - script , [ a1 ; a2 ; ... an ] % denotes a column vector , whereas % [ a1 , a2 , ... an ] is a row vector .
p_01 = kron([1 ; 0 ] , [ 0 ; 1 ] ) ; p_10 = kron([0 ; 1 ] , [ 1 ; 0 ] ) ; % define a random superposition of the % above states .
% % rand ( ) yields a random number chosen % uniformly from the interval ( 0 , 1 ) .
r1 = 2*pi*rand ( ) ; r2 = 2*pi*rand ( ) ; psi = sin(r1)*p_01 + exp(1i*r2)*cos(r1)*p_10 ; % dimensions of subsystems sys = [ 2 , 2 ] ; % calulate measure and gradient e = entropyofentanglement(psi , sys ) g = grad_entropyofentanglement(psi ,
sys ) .... note that the entropy of entanglement is of particular importance , because its convex - roof extension is the well - known and widely used ` entanglement of formation ' @xcite . in the special case of a bipartite system composed of two - dimensional subsystems ( qubits )
, there exists an operational solution for the entanglement of formation @xcite , which we have implemented in ` eof2x2 ` .
the three - tangle @xcite is defined specifically for a system of three two - dimensional subsystems .
it reads @xmath8 where @xmath9 and @xmath10 , are the complex amplitudes of the vector @xmath11 in the standard computational basis .
the tangle is implemented in the function ` tangle ` , taking an 8-dimensional vector ` psi ` as its only argument .
here follows a short example : .... % define an 8-dimensional random state psi = randstate(8 ) ; % calculate tangle and its gradient t = tangle(psi ) g = grad_tangle(psi ) .... in the above code , we have introduced the function ` randstate ` , which returns a random pure quantum state of arbitrary specified dimension uniformly distributed according to the haar measure of the unitary group @xcite .
finally , the measure of meyer and wallach @xcite for an arbitrary number @xmath12 of qubits is an entanglement monotone that can be written in the compact form @xcite @xmath13,\ ] ] where @xmath14 is the density matrix obtained by tracing out all but the @xmath15th subsystem out of @xmath16 .
the implementation is given in ` meyer_wallach ` .
this function also makes use of ` ptrace ` .
the usage is analogous to the example given for the three - tangle above .
in this section , we review how to arrive at an optimization problem ( whose solution is the desired value of the convex - roof entanglement measure ) in a form that can be dealt with on a computer .
let @xmath17 be an entanglement monotone for pure states from a hilbert space @xmath18 of finite dimension @xmath19 .
let @xmath0 be a density matrix acting on that space .
our goal is to numerically evaluate the convex roof @xmath20 of @xmath17 , given by @xmath21 where @xmath22 is the set of all pure - state decompositions of @xmath0 . with respect to numerical optimization ,
a convenient parameterization of all subsets of @xmath23 with a constant number of terms @xmath15 ( sometimes referred to as the ` cardinality ' ) is due to the schrdinger - hjw theorem @xcite .
the latter states that ( i ) , every decomposition of a density matrix @xmath0 with @xmath24 into a convex sum of @xmath15 projectors onto pure states can be expressed in terms of a complex @xmath25 matrix @xmath26 obeying @xmath27 and that ( ii ) , conversely , from every such matrix one can obtain a pure - state decomposition of @xmath0 .
the set @xmath28 with @xmath29 is also known as the stiefel manifold .
part ( i ) and ( ii ) together ensure that optimizing over @xmath30 is equivalent to optimizing over the full subset of @xmath23 with fixed cardinality @xmath15 . part ( ii ) also provides an explicit construction of the pure - state decomposition related to an arbitrary given matrix @xmath31 : let @xmath32 , @xmath33 , @xmath34 be the non - zero eigenvalues and corresponding normalized eigenvectors of @xmath0 , i.e. , @xmath35 and @xmath36 . note that we have @xmath37 due to the positive semi - definiteness of @xmath0 .
define the auxiliary states @xmath38 and set @xmath39 in ` libcreme ` , the function ` densityeig ` calculates only the non - zero eigenvalues and corresponding eigenvectors . the eigenvalues are guaranteed to be sorted in decreasing order , which is particularly convenient if one wishes to discard some parts of the density matrix occurring with low probability , such as , e.g. , high - energy sectors in density matrices @xmath40 originating from some hamiltonian @xmath41 .
the function ` psdecomposition ` returns the pure - state decomposition from eqs .
( [ eq : spectral decomposition of rho ] , [ eq : auxiliary states ] , [ eq : p_i of u ] , [ eq : psi_i of u ] ) . as an example , let ` rho ` store a @xmath42 density matrix of rank @xmath43 , and let ` u ` be a matrix from @xmath30 , with arbitrary @xmath29
. then .... % note that in m - script , functions can return % multiple values of arbitrary dimensions . the % syntax to assign several return values to % local variables is % [ a , b , ... ] = foo ( ... ) ; [ chi , lambda ] = densityeig(rho ) ; [ psi , p ] = psdecomposition(u , chi , lambda ) ; .... first yields the eigenvectors of ` rho ` in the columns of the @xmath44 matrix ` chi ` with the corresponding @xmath43 eigenvalues in the vector ` lambda ` . on the second line
then , the pure - state decomposition of ` rho ` ( given in terms of the parameters ` chi ` and ` lambda ` ) corresponding to the parameterization ` u ` is calculated , with the @xmath15 state vectors @xmath2 stored in the columns of the @xmath45 matrix ` psi ` and the @xmath15 corresponding probabilities @xmath46 in the vector ` p ` . by virtue of the schrdinger - hjw theorem
, we can restate the optimization problem eq . as @xmath47 where the dependence on @xmath0 enters implicitly as the dependence of the @xmath46 and @xmath2 on the eigenvalues and eigenvectors of @xmath0 . in practice
, it has turned out to be possible to drop the minimization over @xmath15 completely and set @xmath15 to a constant but large enough value instead .
note that this actually includes all cardinalities @xmath48 with @xmath49 in the search because up to @xmath50 of the @xmath46 are allowed to go to zero without breaking the optimization constraint . in ` libcreme ` , the function ` convexsum ` calculates the value of the expression @xmath51 in eq . , which is , in fact , the objective function of the optimization .
` convexsum ` takes as its first argument a parameterization matrix from the stiefel manifold , as its second a function handle @xcite to the entanglement monotone to be extended , and as its third and fourth arguments the eigendecomposition of the density matrix obtained by ` densityeig ` .
here is a full example : .... % random 8-by-8 density matrix rho = randdensitymatrix(8 ) ; % calculate eigendecomposition of rho for % later use [ chi , lambda ] = densityeig(rho ) ; % random matrix from st(12 , 8) u = randunitarymatrix(12 , 8) ; % evaluate convex sum eq . ( 15 ) with the tangle h = convexsum(u , @tangle , chi , lambda ) .... note that we have introduced the functions ` randdensitymatrix ` and ` randunitarymatrix ` to create random density matrices and random matrices from the stiefel manifold , respectively .
it is important to understand that ` convexsum ` is the key function in the whole library in the sense that it is always this function ( or more specifically , an anonymous function handle @xcite to it , see sec . [
sec : algorithms ] ) , which is ultimately optimized .
as mentioned earlier , the optimization algorithms in ` libcreme ` require the knowledge of the gradient of the objective function , or more precisely , the derivatives of @xmath51 with respect to the real and imaginary parts of the matrix elements of @xmath26 .
these expressions and their derivation can be found in [ app : derivatives of h ] . within the library , this gradient of @xmath52 is implemented in ` grad_convexsum ` .
it requires 5 arguments : the matrix @xmath26 , the entanglement monotone to be extended , the gradient of the latter , and the eigendecomposition of @xmath0 ( eigenvectors and -values ) .
the following code illustrates its application in practice , using the variables ` chi ` , ` lambda ` , and ` u ` from the previous example : .... % evaluate gradient of the convex sum eq .
( 15 ) , % given in eqs .
( b4 - b7 ) with the tangle gh = grad_convexsum(u , @tangle , ...
@grad_tangle , chi , lambda ) ....
we describe in this section two conceptually different optimization algorithms which are both provided in ` libcreme ` .
one is a conjugate gradient method based on the concepts introduced in refs .
it exploits the differential - geometric structure of the nonlinear search space emerging from the optimization constraint @xmath53 .
the other algorithm is a standard broyden - fletcher - goldfarb - shanno ( bfgs ) quasi - newton method employing a transformation of the constrained search space to an unconstrained one .
both algorithms have been discussed in greater detail in a previous work @xcite , where the expressions for the gradients and parameterization of the stiefel manifold given below have been derived .
the interested reader is referred to that earlier work .
here , we just state the final results for the sake of completeness and focus particularly on the implementation and usage within ` libcreme ` .
this algorithm exploits the geometric structure of the unitary group @xmath54 and therefore generally over - parameterizes the true search space @xmath30 , @xmath55 .
this is however not a problem in practice , since we can simply discard the last @xmath56 columns of @xmath26 when calculating the decomposition of the density matrix based on @xmath26 @xcite .
the full algorithm for an input initial guess @xmath57 is given as follows : 1 .
initialization : set @xmath58 .
calculate the gradient @xmath59 according to the formula @xmath60 where the matrices @xmath61 and @xmath62 are given by @xmath63 and the gradient of @xmath51 can be found in [ app : derivatives of h ] .
+ finally , set @xmath64 .
[ enum : step_linmin ] perform the one - dimensional minimization @xmath65 set @xmath66 and compute the new gradient @xmath67 according to eqs .
( [ eq : gradient unitary group ] , [ eq : gradient a ] , [ eq : gradient b ] ) .
3 . define @xmath68 @xmath69 is the gradient @xmath70 parallel - transported to the new point @xmath71 .
4 . calculate the modified polak - ribire parameter @xmath72 where @xmath73 .
5 . set the new search direction to @xmath74 6 .
set @xmath75 .
repeat from step 2 until convergence .
this algorithm is implemented in the function ` cg_min ` .
the minimization in step [ enum : step_linmin ] is performed by the derivative - based method ` minimize1d_exp ` . `
cg_min ` requires the function to be minimized , its gradient , an initial point , and optionally a struct with user - specified termination criteria discussed below . at this point
, we would like to work through a full example demonstrating the use of ` cg_min ` to calculate the convex - roof extended three - tangle of a mixed state .
.... % random 8-by-8 density matrix rho = randdensitymatrix(8 ) ; % calculate eigendecomposition of rho for % later use [ chi , lambda ] = densityeig(rho ) ; % define anonymous function handles [ 24 ] to % the objective function and its gradient f_opt = @(x ) convexsum(x , @tangle , ... chi , lambda ) ; g_opt = @(x ) grad_convexsum(x , @tangle , ...
@grad_tangle , chi , lambda ) ; % choose a random starting point , here for % a decomposition with cardinality 12 u0 = randunitarymatrix(12 , 12 ) ; % perform the optimization [ t , ut , info ] = cg_min(f_opt , g_opt , u0 ) ; .... a few comments about the above code are in order .
first , note that because ` cg_min ` requires the objective and its gradient in the form of one - parameter functions , we need to define the anonymous function handles ` f_opt ` and ` g_opt ` before continuing . in this way , ` f_opt ` is a new function that evaluates ` convexsum ` at a variable unitary input matrix while keeping the constant arguments ` @tangle ` , ` chi ` , and ` lambda ` fixed .
a similar description holds for ` g_opt ` .
second , note that the initial search point ` u0 ` is a unitary and therefore square matrix , although matrices from @xmath76 would be sufficient in the example above to parameterize pure - state decompositions .
the reason is , as mentioned above , that ` cg_min ` is operating on the unitary group instead of the stiefel manifold . however , this is hidden from the user by the fact that both ` convexsum ` and ` grad_convexsum ` can accept larger input than required , and automatically discard any dispensable columns .
third , we would like to draw the reader s attention to the output values of ` cg_min ` . in the above example , ` t ` stores the convex - roof of the entanglement monotone to be evaluated ( in this case the three - tangle ) and ` ut ` ( or more precisely , the first @xmath43 columns of it ) represent the pure - state decomposition of ` rho ` arriving at this value .
the variable ` info ` is a struct that carries useful additional information , namely the criterion that terminated the iteration ( ` info.status ` ) , the function values along the iteration , excluding intermediate values during line searches ( ` info.fvals ` ) , and finally , the traversed points in the search space corresponding to ` fvals ` ( ` info.xvals ` ) . an optional fourth argument containing settings for the termination of the algorithm can be passed to ` cg_min ` .
the following code illustrates the possible struct variables ( the values in this example are also the defaults for any variables not set ) .
.... % create a struct opts = struct ( ) ; % maximum number of iterations opts.maxiter = 1000 ; % tol . on consecutive function values opts.tolfun
= 1e-12 ; % tolerance on norm of difference between % two consecutive gradients opts.tolg = 1e-10 ; % tolerance on norm of difference between % two consecutive iteration points opts.tolx = 1e-10 ; .... the iteration is stopped if either the maximum number of iterations is reached , or one of the checked values is lower than its respective tolerance .
finally , we would like to mention that for the convenience of the user , there is a function called ` createconvexfunctions ` which performs all the necessary steps before the actual optimization in one line : .... % again using the tangle as an example [ f_opt , g_opt ] = createconvexfunctions(rho , ...
@tangle , @grad_tangle ) ; [ t , ut , info ] = cg_min(f_opt , g_opt , u0 ) ; .... as with any other numerical optimization procedure , it is advisable to repeat the computations with different ( random ) initial conditions in order to reach a better approximation to the global minimum .
the second algorithm is a classical bfgs quasi - newton method @xcite that makes use of a transformation which is able to unconstrain the optimization problem eq . from the stiefel manifold to ordinary euclidean space .
this transformation is conceptually identical to the example where one has an optimization problem with the constraint @xmath77 and then sets @xmath78 , @xmath79 and performs the optimization over the new variable @xmath80 .
again , we only state in the following the main results required to implement the algorithm and refer the reader interested in a thorough derivation to ref . @xcite .
the number of independent real parameters required to parameterize @xmath30 is equal to its dimension which is given by @xmath81 . given a tuple of ` angles ' @xmath82 , @xmath83 , we relabel them in the following ( arbitrary but fixed ) way : @xmath84 for @xmath85 $ ] , @xmath86 for @xmath87 + 1 , \ldots , r(2k - r - 1)$ ] , and @xmath88 for @xmath89 .
then , we calculate @xmath90 according to @xmath91 r\ ] ] where @xmath92 , @xmath93 is a @xmath25 matrix with the only non - zero elements being @xmath94 for @xmath95 , and the ` inverse givens matrices ' @xmath96 are defined in terms of their matrix elements as @xmath97_{ij } = \begin{cases } e^{-\mathfrak i\varphi}\cos\vartheta , & \mbox{if } \;\ ; i = j = s , \\
-e^{-\mathfrak i\varphi}\sin\vartheta , & \mbox{if } \;\ ; i = s , j = s + 1,\\ e^{\mathfrak i\varphi}\sin\vartheta , & \mbox{if } \;\ ; i = s + 1 , j = s,\\ e^{\mathfrak i\varphi}\cos\vartheta , & \mbox{if } \;\ ; i = s + 1 , j = s + 1,\\ \delta_{ij } , & \mbox{otherwise}. \end{cases}\ ] ] in ` libcreme ` , calculating a stiefel matrix from a vector of angles ` x ` by eq . is implemented in ` buildunitary ` as a fast c / c++ extension and is demonstrated in the following : .... % dimensions of the stiefel manifold k = 10 ; r = 7 ; % random vector of angles with proper size % ( uses dimst for the dimension of the % stiefel manifold ) .
see footnote [ 30 ] .
% % randn(m , n ) yields an m - by - n matrix of % normally distributed random numbers .
x = 2*pi*randn(1 , dimst(k , r ) ) ; % finally , the stiefel matrix u = buildunitary(x , k , r ) ; .... the derivatives of @xmath98 with respect to the angles @xmath82 are implemented in the function ` grad_buildunitary ` .
the inverse operation of ` buildunitary ` , namely obtaining the parameterizing angles for a given matrix ` u ` can be performed by ` decomposeunitary ` as in .... x = decomposeunitary(u ) ; .... this function is implemented in regular in m - script , because it is called only infrequently and thus is not time critical .
we have now all the tools to describe the full bfgs quasi - newton algorithm starting from an initial vector of angles @xmath99 . 1 . set @xmath58 , @xmath100 , @xmath101 , and @xmath102 .
@xmath103 is the initial guess for the approximate hessian , @xmath52 is the convex sum eq . , and
@xmath104 is the transformation eq . .
2 . perform the line minimization @xmath105 and set @xmath106 3 .
compute the new gradient @xmath107 4 . update the approximate hessian as @xcite @xmath108 where the column vectors @xmath109 and @xmath110 are defined as @xmath111 and @xmath112 .
5 . set the new search direction to @xmath113 6 .
set @xmath75 .
repeat from step 2 until convergence .
the line minimization in step 2 is performed by ` minimize1d_lin ` , a subroutine that is conceptually identical to the function ` minimize1d_exp ` used above in the conjugate - gradient method .
the full algorithm is implemented in ` bfgs_min ` and its input and output parameters are identical to the ones in ` cg_min ` .
hence , the descriptions in the previous section can be adapted analogously to ` bfgs_min ` .
however , the target function ( and its gradient ) look slightly different in the current case and are somewhat more cumbersome in terms of function handles , because the additional intermediate transformation eq .
needs to be incorporated .
the following is a fully working example that should help to clarify this issue .
.... % random 8-by-8 density matrix rho = randdensitymatrix(8 ) ; [ chi , lambda ] = densityeig(rho ) ; % convex - sum function handles for the tangle f_cr = @(x ) convexsum(x , @tangle , ... chi , lambda ) ; g_cr = @(x ) grad_convexsum(x , @tangle , ...
@grad_tangle , chi , lambda ) ; % dimensions of the stiefel manifold r = rank(rho ) ; k = r + 4 ; % objective function and gradient f_opt = @(x ) f_cr(buildunitary(x , k , r ) ) ; g_opt = @(x ) grad_eh_adapt(x , k , r , g_cr ) ; % choose a random starting point x0 = 2*pi*randn(1 , dimst(k , r ) ) ; % perform the optimization [ t , xt , info ] = bfgs_min(f_opt , g_opt , x0 ) ; .... notice the use of the auxiliary function ` grad_eh_adapt ` which calculates the gradient @xmath114 given the derivatives of @xmath51 with respect to the real and imaginary matrix elements of @xmath26 . for the convenience of the user
, there is a function that hides all the above steps just like in the case of the conjugate - gradient algorithm .
it is called ` createehfunctions ` and is analogously called , as exemplified here : .... [ f_opt , g_opt ] = createehfunctions(rho , ... k , r , @tangle , @grad_tangle ) ; [ t , xt , info ] = bfgs_min(f_opt , g_opt , x0 ) ; .... clearly , the same note as in the previous section regarding multiple restarts holds . finally , we would like to make a remark about a detail in our implementation of ` bfgs_min ` .
it has shown to be useful in practice to reset the angles modulo @xmath115 every few iterations .
this improves numerical stability and convergence in the vicinity of a minimum .
it is also advisable to vary the interval size after which this is done as this can improve performance depending on the problem .
if ` bfgs_min ` is to be employed for non - periodic functions , these lines of code must be removed .
before we end this section and look at some more examples , we would like to make a few comments . at this point ,
the reader might wonder why we provide our own implementation of a line search and a bfgs quasi - newton method , instead of resorting to the functions available in matlab and octave .
there are several reasons for that .
first of all , it makes the library independent of the platform , since the standard routines in matlab and octave work differently and hence generally produce unequal results .
furthermore , having a simple implementation at hand allows the user to quickly make modifications and customize the code to specific needs . in octave
this is can only be achieved with quite an effort , whereas in matlab it is generally not possible at all .
additional issues are availability and backward - compatibility .
while matlab s optimization framework is well established , it is only available through the purchase of the ` matlab optimization toolbox ' . on the other hand
, there is a free octave package for non - linear optimization tasks @xcite .
but since this is still under active development , its usage within ` libcreme ` might potentially become incompatible with future releases of the package .
next we would like to address the performance of the algorithms as a function of @xmath116 and @xmath117 .
the dimension of the stiefel manifold ( and hence the problem size ) is given by @xmath118 .
since we must have @xmath29 , we can replace it by @xmath119 , with @xmath120 , yielding @xmath121 .
this shows that the computational cost grows quadratically with the rank of @xmath0 , but only linearly with the ( user - specified ) cardinality .
the algorithms in the library are thus most efficient for low - rank density matrices , whereas experimenting with different cardinalities is not that costly .
actually , already choices for @xmath1 as low as @xmath122 have shown to produce very accurate results in practice ( see also below ) . since
, on the other hand , the scaling with @xmath43 is less favorable , it is advisable to examine whether the rank of @xmath43 can be reduced . particularly in density matrices originating from physical hamiltonians
it is often justified to neglect high - energy states .
as mentioned earlier , reducing the rank of @xmath0 can conveniently be achieved by truncating the output of the function ` densityeig ` .
in this section , we demonstrate the usage of ` libcreme ` by working through two complete examples .
we calculate the entanglement of special states where analytical results are known in order to compare the numerical experiments with theory .
note that we provide initial points for the optimization in separate files instead of generating them randomly , because ( i ) the random number generators in matlab and octave produce different sequences of random numbers and ( ii ) not every initial point leads to the convergence to a global minimum in such high - dimensional spaces . ) , whereas the dashed line does the same for the tangle of a ghz / w mixture ( example in sec .
[ sec : ex_tangle]).,title="fig:"][fig : convergence ] isotropic states are defined as a convex mixture of a maximally entangled state and the maximally mixed state in a system of two @xmath19-dimensional subsystems . the isotropic state with an amount of mixing specified by @xmath123 , where @xmath124 , is given by @xcite @xmath125 where @xmath126 . an analytical solution for the entanglement of formation as a function of @xmath123 and @xmath19 has been found @xcite and is implemented in ` eofisotropic ` .
let us compare now the numerical results with theory .
the full example can be found in the folder ` examples / eofisotropic ` , along with all other related files .
we first choose a dimension for the two subsystems , .... d = 5 ; .... then create the maximally entangled state @xmath127 in these systems and store it in ` psi ` : .... psi = 0 ; for i = 1:d tmp = zeros(d , 1 ) ; tmp(i ) = 1 ; psi = psi + kron(tmp , tmp ) ; end psi = psi / sqrt(d ) ; .... after choosing a value for the mixing parameter ` f ` , .... f = 0.3 ; .... we can construct the isotropic state specified by ` d ` and ` f ` as .... % note that in m - script , a ' is the hermitian % conjugate of a. rho = ( 1 - f)/(d^2 - 1 ) * ( eye(d^2 ) - ... ( psi*psi ' ) ) + f*(psi*psi ' ) ; .... and calculate its eigendecomposition with .... [ chi , lambda ] = densityeig(rho ) ; .... in order to keep this example fully reproducible , we unfortunately have to load and overwrite the eigenvectors ` chi ` from a file at this point .
the reason is that the density matrix is degenerate , yielding different eigendecompositions for the degenerate subspace depending on whether one uses matlab or octave due to the different diagonalization routines employed by these platforms .
clearly , one arrives at comparable results in both cases , but the paths in optimization space are different .
.... chi = load('example_eofisotropic_chi.txt ' ) ; .... after setting an appropriate cardinality .... r = rank(rho ) ; k = 2*r ; .... and defining function handles for the entanglement measure and its gradient .... eoe = @(x ) entropyofentanglement(x , [ d , d ] ) ; grad_eoe = ...
@(x ) grad_entropyofentanglement(x , [ d , d ] ) ; .... we can create the function handles required in the optimization .... f_cr = @(x ) convexsum(x , eoe , chi , lambda ) ; g_cr = @(x ) grad_convexsum(x , eoe , ...
grad_eoe , chi , lambda ) ; .... finally , we choose a random initial value ` u0 ` ( here initialized from a file ) .... u0r = load('example_eofisotropic_u0r.txt ' ) ; u0i = load('example_eofisotropic_u0i.txt ' ) ; u0 = u0r + 1i*u0i ; .... and perform the optimization : .... [ e_res , u_res , info ] = cg_min(f_cr , g_cr , u0 ) ; .... this yields a value of ` e_res ` @xmath128 after 80 iterations .
we can check the convergence and the accuracy of the result by plotting the difference between the function value in each iteration and the theoretical value : .... semilogy(abs(info.fvals - eofisotropic(f , d ) ) ) ; .... this produces the solid line in fig . 1
, showing that the result is exact up to an absolute error of @xmath129 . in this example , we will calculate the three - tangle of a mixture of the two states @xmath130 given by @xcite @xmath131 the example files are in ` examples / tangleghzw ` . in the code
, we define the states .... ghz = [ 1 ; 0 ; 0 ; 0 ; 0 ; 0 ; 0 ; 1]/sqrt(2 ) ; w = [ 0 ; 1 ; 1 ; 0 ; 1 ; 0 ; 0 ; 0]/sqrt(3 ) ; .... choose a particular value for ` p ` .... p = 0.7 ; .... and create the mixed state .... rho = p*ghz*ghz ' + ( 1 - p)*w*w ' ; .... we then specify a value for the cardinality .... r = rank(rho ) ; k = r + 4 ; .... and can create the objective function and gradient handles . note that we use the auxiliary function ` createehfunctions ` to do all the required work : .... [ f_eh , g_eh ] = createehfunctions(rho , ... k , r , @tangle , @grad_tangle ) ; .... after choosing a random initial point ( initialized from a file , as before ) .... x0 = load('example_tangleghzw_x0.dat ' ) ; .... we are ready to perform the optimization : .... [ t_res , x_res , info ] = ... bfgs_min(f_eh , g_eh , x0 ) ; .... the value one obtains in this way after 29 iterations is ` t_res ` @xmath132 .
a comparison with the analytical value @xcite is exact within numerical precision
. we can again plot the error between the function values and the exact result .... semilogy(abs(info.fvals - tangleghzw(0.7 ) ) ) ; .... which yields the dashed line in fig . 1 .
we have presented our library ` libcreme ` which serves to evaluate generic convex - roof entanglement measures .
the library contains all tools required to deal with this problem , including two optimization algorithms working on the space of density matrix decompositions . the first one is based on a conjugate gradient algorithm operating directly on the group of unitary parameterizations , while the second one is a standard bfgs quasi - newton method employed with a transformation from the original search space to unconstrained euclidean space . both implementations accept generic function handles , making it easy to extend their application to user - defined entanglement measures .
all that needs to be done for this is the implementation of the respective pure - state entanglement monotone and the corresponding derivatives with respect to the real and imaginary parts of the input state vector .
we would like to thank stefano chesi for fruitful discussions .
this work was partially supported by the swiss nsf , nccr nanoscience , nccr qsit , solid , and darpa quest .
in the following , we provide the calculations for the derivatives of all entanglement measures included in ` libcreme ` .
let @xmath133 be a state vector from a bipartite system with subsystem dimensions @xmath134 and @xmath135 .
let us rewrite eq . in the form @xmath136 where @xmath137 and @xmath138 .
let @xmath139 be an arbitrary ( complex ) entry of the state vector @xmath133 . then , using the chain rule , we have @xmath140 note that the indices @xmath15 and @xmath141 in the above sum run over the full hilbert space dimension @xmath142 , whereas @xmath143 and @xmath144 only run over the first subsystem with dimension @xmath134 .
we now evaluate each term in the sum separately . for the gradient of @xmath145
we get @xmath146 \\ & = -\nabla_x \bigg\ { \sum_{n = 1}^\infty \frac{(-1)^{n+1}}{n}{\mathop\mathrm{tr}\nolimits}\left[(x - \mathbb{i})^{n+1}\right ] \\ &
\qquad + \sum_{n = 1}^\infty \frac{(-1)^{n+1}}{n}{\mathop\mathrm{tr}\nolimits}\left[(x - \mathbb{i})^n\right]\bigg\ } \\ & = -\bigg\ { \sum_{n = 1}^\infty \frac{(-1)^{n+1}(n+1)}{n}(x - \mathbb{i})^n \\ & \qquad + \sum_{n = 1}^\infty ( -1)^{n+1}(x - \mathbb{i})^{n-1}\bigg\}^t\\ & = -\bigg\{-\sum_{n = 1}^\infty ( \mathbb{i } - x)^n \\ & \qquad + \sum_{n=1}^\infty \frac{(-1)^{n+1}}{n}(x - \mathbb{i})^n + \sum_{n = 0}^\infty ( \mathbb{i } - x)^n\bigg\}^t\\ & = -\log x^t - \mathbb{i},\end{aligned}\ ] ] where we have made use of the formula @xmath147 and the fact that the series expansion of the logarithm is valid because in our case @xmath148 is always a density matrix , thus having real eigenvalues between 0 and 1 .
next we evaluate the derivatives of the partial trace @xmath149 .
we will write coordinate indices of vectors in the full hilbert space as @xmath150 $ ] , where @xmath151 and @xmath152
. then @xmath153\,[j , k]}\ ] ] and thus @xmath154 finally , we have to consider the derivatives of the density matrix itself with respect to the entries of the state vectors .
this is the part where we have to treat @xmath155 and @xmath156 as independent variables because @xmath0 is not analytic in the entries of @xmath133 .
one quickly finds @xmath157 putting all this together we find , after eliminating all kronecker @xmath109-symbols , @xmath158_{n_1 i}\cdot\psi^\ast_{[i , n_2 ] } + [ \nabla s(\rho)]_{i n_1}\cdot\psi_{[i , n_2]}\right\}\ ] ] and analogously for the derivatives with respect to the imaginary parts . exploiting the fact that @xmath159 is hermitian , we arrive at the final expressions @xmath160_{i n_1}\cdot \psi_{[i , n_2]}\right\ } , \\ \frac{\partial e(|\psi\rangle)}{\partial { \mathop\mathrm{im}}\psi_n } & = 2 \sum_{i = 1}^{d_1}{\mathop\mathrm{im}}\left\{[\nabla s({\mathop\mathrm{tr}\nolimits}_b\rho)]_{i n_1}\cdot \psi_{[i , n_2]}\right\}.\end{aligned}\ ] ] defining @xmath161 , where @xmath162 are given in eqs . ( [ eq : tangle_d1 ] , [ eq : tangle_d2 ] , [ eq : tangle_d3 ] ) , it is easy to see that @xmath163 note that the derivatives @xmath164 are well - defined because @xmath19 is an analytic function of the elements @xmath165 of @xmath133 .
we start by calculating the derivative of @xmath166 with respect to an arbitrary complex element @xmath139 of @xmath133 until the point where non - analyticities appear : @xmath167 \\ & = -\frac{4}{n}\sum_{k = 1}^n \sum_{j , l = 1}^2{\mathop\mathrm{re}}\left [ \left(\rho_k^\ast\right)_{jl}\cdot \frac{\partial}{\partial\psi } \left(\rho_k\right)_{jl } \right ] \\ & = -\frac{4}{n}\sum_{k = 1}^n \sum_{j , l = 1}^2{\mathop\mathrm{re}}\left\ { \left(\rho_k^\ast\right)_{jl } \sum_{\alpha , \beta = 1}^{2^n } \left [ \frac{\partial\left(\rho_k\right)_{jl}}{\partial\rho_{\alpha\beta } } \cdot \frac{\partial\rho_{\alpha\beta}}{\partial\psi } \right ] \right\}.\label{eq : d_gamma_final } \end{split}\ ] ] the derivatives of the @xmath168 ( depending non - analytically on @xmath139 ) with respect to the real and imaginary part of @xmath139 have already been stated in eqs .
( [ eq : d_rho_re ] , [ eq : d_rho_im ] ) .
we are thus left to calculate the slightly cumbersome derivatives of multiple partial traces of @xmath0 with respect to the matrix elements @xmath168 . similarly to the calculation in [ sec : gradient_of_eof ] , we will now rewrite indices @xmath169 of the full hilbert space in the binary representation @xmath170 $ ] , where @xmath171 for all @xmath172 ( we will also employ this represention for the indices @xmath173 , and @xmath1 below ) . then , the matrix elements of @xmath14 can be written as@xmath174\ ; [ \nu_1,\ldots,\nu_{k-1 } , j , \nu_{k+1 } , \ldots \nu_n ] } , $ ] where @xmath175 .
hence we have @xmath176 inserting this into eq . and working out all kronecker @xmath109 symbols , we arrive at @xmath177}\cdot\left(\rho_k\right)_{n_k , j}\right ) , \\
\left.\frac{\partial \gamma}{\partial{\mathop\mathrm{im}}\psi_n}\right|_{|\psi\rangle } & = -\frac{8}{n}\sum_{k = 1}^n\sum_{j = 1}^{2 } { \mathop\mathrm{im}}\left(\psi_{[n_1,n_2,\ldots , n_{k-1},j , n_{k+1},\ldots , n_n]}\cdot\left(\rho_k\right)_{n_k , j}\right).\end{aligned}\ ] ]
we will carry out the calculation explicitly only for the derivatives with respect to the real part of @xmath26 , but everything works analogously for the imaginary part . for the sake of readability , we will drop the usual ` ket ' notation and write quantum state vectors as @xmath178 .
we write the @xmath15th element of @xmath179 as @xmath180 . where @xmath19 is the dimension of the total hilbert space .
note that we have specifically emphasized the @xmath26-dependence of the @xmath46 and @xmath179 via eqs .
( [ eq : spectral decomposition of rho ] , [ eq : auxiliary states ] , [ eq : p_i of u ] , [ eq : psi_i of u ] ) .
the first derivative in this expression is given by @xmath183 where we have used in the last step the orthonormality of the @xmath184 and the fact that @xmath185 for any complex number @xmath186 .
we can now insert eqs . and into .
the final result ( including the derivatives with respect to the imaginary part of @xmath26 from an analogous calculation ) reads @xmath188 , \end{split}\ ] ] @xmath189 , \end{split}\ ] ] where @xmath190,\\ \zeta_{\alpha\beta}(u ) & = \left[\mathfrak{i}\sqrt{\frac{\lambda_\beta}{p_\alpha(u)}}\chi_\beta - \frac{\lambda_\beta}{p_\alpha(u)}{\mathop\mathrm{im}}(u_{\alpha\beta})\psi_\alpha(u)\right].\end{aligned}\ ] ] it is easily seen from the cauchy - riemann equations that non - constant functions from @xmath191 to @xmath192 , such as the kind of entanglement measures we are addressing in this work , can not be analytic .
we therefore have to treat the real and imaginary part of complex numbers as independent variables . in m - script
, functions can be passed as arguments using ` function handles ' ( conceptually similar to function pointers in c ) . for a function with name ` foo ` , the corresponding function handle is given by ` @foo ` .
subroutines are then able to call ` foo ` through its function handle , given the number and type of parameters are correct .
a related feature of m - sript used frequently in ` libcreme ` are ad hoc , also called ` anonymous ' , function handles .
these are handles to functions which are neither built - in , nor defined in a script file , but rather created on the spot . as an example , consider the following syntax to create a function of two parameters calculating the sine of the parameter s product : the variable names used to define the function are listed in the parentheses after the `` @ ` ' sign .
directely following is the actual definition of the function .
it can then be called in the usual way , yielding , e.g. , however , in the case where one chooses @xmath193 , it is advantageous to employ an algorithm that works directly on the stiefel manifold instead of the unitary group .
the concepts discussed in ref .
@xcite for the real stiefel manifold can be adapted to the complex case by replacing transpositions with hermitian conjugates and the frobenius inner product @xmath194 by @xmath195 .
we note that this way of generating random angles ( which we have chosen here for simplicity ) is not ideal , since it does not sample uniformly from the haar measure . a better way would be to use ` decomposeunitary(randunitarymatrix(k , r ) ) ` . | we present the software library ` libcreme ` which we have previously used to successfully calculate convex - roof entanglement measures of mixed quantum states appearing in realistic physical systems . evaluating
the amount of entanglement in such states is in general a non - trivial task requiring to solve a highly non - linear complex optimization problem .
the algorithms provided here are able to achieve to do this for a large and important class of entanglement measures .
the library is mostly written in the matlab programming language , but is fully compatible to the free and open - source octave platform .
some inefficient subroutines are written in c / c++ for better performance .
this manuscript discusses the most important theoretical concepts and workings of the algorithms , focussing on the actual implementation and usage within the library .
detailed examples in the end should make it easy for the user to apply ` libcreme ` to specific problems .
entanglement measure , convex roof * program summary * _ manuscript title : _ libcreme : an optimization library for evaluating convex - roof entanglement measures + _ authors : _ beat rthlisberger , jrg lehmann , daniel loss + _ program title : _ libcreme + _ journal reference : _
+ _ catalogue identifier : _
+ _ licensing provisions : _ gnu gpl version 3 + _ programming language : _ matlab / octave and c / c++ + _ operating system : _ all systems running matlab or octave + _ keywords : _ entanglement measure , convex roof + _ classification : _ 4.9 , 4.15 + _ nature of problem : _ + evaluate convex - roof entanglement measures .
this involves solving a non - linear ( unitary ) optimization problem .
+ _ solution method : _ + two algorithms are provided : a conjugate - gradient method using a differential - geometric approach and a quasi - newton method together with a mapping to euclidean space .
+ _ running time : _ + typically seconds to minutes for a density matrix of a few low - dimensional systems and a decent implementation of the pure - state entanglement measure . |
molecular gas is one of the essential components in galaxies because it is closely related to star formation , which is a fundamental process of galaxy evolution .
thus the observational study of molecular gas is indispensable to understand both star formation in galaxies and galaxy evolution .
however , the most abundant constituent in molecular gas , h@xmath15 , can not emit any electro - magnetic wave in cold molecular gas with typical temperature of @xmath16 10 k due to the lack of a permanent dipole moment . instead
, rotational transition lines of @xmath1co , the second abundant molecule , have been used as a tracer of molecular gas .
for example , some extensive @xmath1co surveys of external galaxies , which consist of single pointings toward central regions and some mappings along the major axis , have been reported ( e.g. , @xcite ; @xcite ; @xcite ) .
these studies provided new findings about global properties of galaxies , such as excitation condition of molecular gas in galaxy centers and radial distributions of molecular gas across galaxy disks . in order to understand the relationship between molecular gas and star formation in galaxies further ,
spatially resolved @xmath1co maps covering whole galaxy disks are necessary because star formation rates ( sfrs ) are often different between galaxy centers and disks .
in particular , single - dish observations are essential to measure @xmath17 molecular gas content in the observing beam from dense component to diffuse one avoiding the missing flux ( e.g. , @xcite ) .
so far , two major surveys of wide - area @xmath1co mapping toward nearby galaxies are performed using multi - beam receivers mounted on large single - dish telescopes .
one is the @xmath1co(@xmath0 ) mapping survey of 40 nearby spiral galaxies performed with the nobeyama radio observatory ( nro ) 45-m telescope in the position - switch mode ( @xcite , hereafter k07 ) .
their @xmath1co(@xmath0 ) maps cover most of the optical disks of galaxies at an angular resolution of 15@xmath18 , and clearly show two - dimensional distributions of molecular gas in galaxies .
k07 found that the degree of the central concentration of molecular gas is higher in barred spiral galaxies than in non - barred spiral galaxies .
in addition , they found a correlation between the degree of central concentration and the bar strength adopted from @xcite ; i.e. , galaxies with stronger bar tend to exhibit a higher central concentration .
this correlation suggests that stronger bars accumulate molecular gas toward central regions more efficiently , which may contribute the onset of intense star formation at galaxy centers ( i.e. , higher sfrs than disks ) . using the @xmath1co(@xmath0 ) data , @xcite investigated the physical properties of molecular gas in the barred spiral galaxy maffei 2 .
they found that molecular gas in the bar ridge regions may be gravitationally unbound , which suggests that molecular gas is hard to become dense , and to form stars in the bar .
the other survey is the heterodyne receiver array co line extragalactic survey performed with the iram 30-m telescope @xcite .
they observed @xmath1co(@xmath7 ) emission over the full optical disks of 48 nearby galaxies at an angular resolution of 13@xmath18 , and found that the @xmath1co(@xmath7)/@xmath1co(@xmath0 ) line intensity ratio ( hereafter @xmath8 ) typically ranges from 0.6 to 1.0 with the averaged value of 0.8 .
in addition , @xcite examined a quantitative relationship between surface densities of molecular gas and sfrs for 30 nearby galaxies at a spatial resolution of 1 kpc using the @xmath1co(@xmath7 ) data .
they found a first - order linear correspondence between surface densities of molecular gas and sfrs but also found second - order systematic variations ; i.e. , the apparent molecular gas depletion time , which is defined by the ratio of the surface density of molecular gas to that of sfr , becomes shorter with the decrease in stellar mass , metallicity , and dust - to - gas ratio .
they suggest that this can be explained by a co - to - h@xmath15 conversion factor ( @xmath19 ) that depends on dust shielding .
however , such global co maps of galaxies have raised a new question ; the cause of the spatial variation in star formation efficiencies ( sfes ) defined as sfrs per unit gas mass .
it is reported that sfes differ not only among galaxies ( e.g. , @xcite ) but also within locations / regions in a galaxy ( e.g. , @xcite ) ; i.e. , higher sfes are often observed in galaxy mergers rather than normal spiral galaxies and also observed in the nuclear star forming region rather than in galaxy disks .
some observational studies based on hcn emission , an excellent dense gas tracer , suggest that sfes increase with the increase in molecular gas density ( or dense gas fraction ) in galaxies ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ) , but the cause of the spatial variation in sfes is still an open question because hcn emission in galaxy disks is too weak to obtain its map except for some gas - rich spiral galaxies ( e.g. , m 51 ; @xcite ; @xcite ) . instead , isotopes of co molecule are promising probes of molecular gas density .
in particular , @xmath2co(@xmath0 ) is thought to be optically thin and thus trace denser molecular gas ( @xmath20 ) rather than @xmath1co(@xmath0 ) , which is optically thick and traces relatively diffuse molecular gas ( @xmath21 ) .
therefore , the relative intensity between @xmath2co(@xmath0 ) and @xmath1co(@xmath0 ) is sensitive to physical properties of molecular gas .
for example , spatial variations in @xmath2co(@xmath0)/@xmath1co(@xmath0 ) intensity ratios ( hereafter @xmath9 ) were observed in nearby galaxy disks ( e.g. , @xcite ; @xcite ; @xcite ) .
such variations in @xmath9 , typically ranging from 0.05 to 0.20 , are interpreted as the variation in molecular gas density ; i.e. , @xmath9 increases with the increase in molecular gas density .
however , some observations suggest that @xmath9 in central regions of nearby galaxies are lower than those in disk regions ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ) although central regions of galaxies often show intense star formation activities , suggesting higher molecular gas density .
the cause of the low @xmath9 in central regions is thought to be high temperature of molecular gas due to the heating by uv radiation from a lot of young massive stars .
such a degeneracy between density and temperature of molecular gas in a line ratio can be solved using two ( or more ) molecular line ratios with a theoretical calculation on the excitation of molecular gas such as the large velocity gradient ( lvg ) model @xcite .
for example , the density and kinetic temperature of giant molecular clouds ( gmcs ) were determined using @xmath9 and @xmath1co(@xmath22)/@xmath1co(@xmath0 ) ratio for large magellanic cloud @xcite and m 33 @xcite , and also determined using @xmath9 and @xmath8 for the spiral arm of m 51 @xcite .
this method to determine molecular gas density is useful to investigate the cause of the variation in sfes .
thus the dependence of sfes on molecular gas density should be investigated for various galaxies at high angular resolution based on multiple line ratios including @xmath9 . in this paper
, we investigate the relationship between sfe and molecular gas density within a nearby barred spiral galaxy ngc 2903 using an archival @xmath1co(@xmath7 ) map combined with @xmath1co(@xmath0 ) and @xmath2co(@xmath0 ) maps which are newly obtained by the co multi - line imaging of nearby galaxies ( coming ) project with the nro 45-m telescope .
ngc 2903 is a gas - rich galaxy exhibiting bright nuclear star formation ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ) .
the distance to ngc 2903 is estimated to be 8.9 mpc @xcite ; thus the effective angular resolution of 20@xmath18 for the on - the - fly ( otf ) mapping with the nro 45-m corresponds to 870 pc .
this enables us to resolve major structures within ngc 2903 , such as the center , bar , and spiral arms although its inclination of @xcite is not so small .
in addition , ngc 2903 is rich in archival multi - wavelength data set ; i.e. , not only the @xmath1co(@xmath7 ) map to examine @xmath8 but also h@xmath23 and infrared images to calculate sfrs are available .
thus this galaxy is a preferable target to examine the cause of the variation in sfe in terms of molecular gas density .
basic parameters of ngc 2903 are summarized in table 1 .
the structure of this paper is as follows : we describe the overview of the coming project and explain the detail of the co observations and data reduction for ngc 2903 in section 2 .
then , we show results of observations ; i.e. , spectra and velocity - integrated intensity maps of @xmath1co(@xmath0 ) and @xmath2co(@xmath0 ) emission in section 3 .
we obtain averaged spectra of @xmath1co(@xmath0 ) , @xmath2co(@xmath0 ) , and @xmath1co(@xmath7 ) emission for nine representative regions , and measure averaged @xmath9 and @xmath8 for each region in section 4.1 .
we determine molecular gas density and kinetic temperature for the center , bar , bar - ends , and spiral arms using @xmath9 and @xmath8 based on the lvg approximation in section 4.2 .
finally , we investigate the cause of the variation in sfe by examining the dependence of sfe on molecular gas density and kinetic temperature .
coming is a project to map @xmath0 emission of @xmath1co , @xmath2co , and c@xmath3o molecules simultaneously for 70% area of optical disks of 238 galaxies using the four - beam receiver system on 45-m telescope ( forest ; @xcite ) at nro .
the main purposes of the coming are to characterize properties of molecular gas as sequence of , hubble types , dynamical structures , central concentrations , and star formation activities , as well as surrounding environments of galaxies .
more detailed information on coming project including the current status of the survey will be reported in the forthcoming paper ( sorai et al . in preparation ) .
@xmath1co(@xmath0 ) , @xmath2co(@xmath0 ) , and c@xmath3o(@xmath0 ) emission observations of ngc 2903 were performed using the nro 45-m telescope from april to may , 2015 , employing the otf mapping mode . the observed area is about @xmath4 , which corresponds to @xmath24 kpc at the distance of 8.9 mpc , as indicated in figure 1 .
the total time for the observations was 13 hrs .
we used a new @xmath25 focal - plane dual - polarization sideband - separating sis mixer receiver for the single side band ( ssb ) operation , forest , which provides 8 intermediate frequency ( if ) paths ( i.e. , 4 beam @xmath5 2 polarization ) independently . owing to the wide
if range of 4 to 12 ghz , we could simultaneously observe @xmath1co(@xmath0 ) emission at 115 ghz ( if = 10 ghz ) and @xmath2co(@xmath0 ) and c@xmath3o(@xmath0 ) emission at 110 ghz ( if = 5 ghz ) when the frequency of the local oscillator was set to 105 ghz .
the backend was an fx - type correlator system , sam45 , which consists of 16 arrays with 4096 spectral channels each .
we set the frequency coverage and resolution for each array of 2 ghz and 488.24 khz , which gives velocity coverage and resolution of 5220 km s@xmath26 and 1.27 km s@xmath26 at 115 ghz , and those of 5450 km s@xmath26 and 1.33 km s@xmath26 at 110 ghz .
we assigned 8 of 16 arrays to 115 ghz band ( i.e. , if = 9 11 ghz for @xmath1co(@xmath0 ) emission ) and other 8 arrays to 110 ghz band ( i.e. , if = 4 6 ghz for @xmath2co(@xmath0 ) and c@xmath3o(@xmath0 ) emission ) .
the half - power beam widths of the 45-m with the forest were @xmath27 at 115 ghz and @xmath28 at 110 ghz , respectively .
the system noise temperatures were 300 500 k at 115 ghz and 200 250 k at 110 ghz during the observing run .
we performed the otf mapping along the major and minor axes of the galaxy disk whose position angle was ( k07 ) .
the separation between the scan rows was set to @xmath29 , and the spatial sampling interval was @xmath30 applying a dump time of 0.1 second and a scanning speed of @xmath31 s@xmath26 .
the data sets scanned along two orthogonal axes were co - added by the basket - weave method @xcite to remove any effects of scanning noise . in order to check the absolute pointing accuracy every hour , we observed an sio maser source , w - cnc , using a 43 ghz band receiver .
it was better than @xmath32 ( peak - to - peak ) throughout the observations .
in addition , we observed @xmath1co(@xmath0 ) and @xmath2co(@xmath0 ) emission of w 3 and irc+10216 every day to obtain the scaling factors for converting the observed antenna temperature to the main beam temperature for each if .
note that these scaling factors not only correct the main - beam efficiency ( @xmath33 ) of the 45-m antenna but also compensate the decrease in line intensity due to the incompleteness of the image rejection for ssb receiver ( e.g. , @xcite ) .
the absolute error of the temperature scale for each co spectrum was about @xmath34% , mainly due to variations in @xmath33 and the image rejection ratio of the forest .
data reduction was made using the software package nostar , which comprises tools for otf data analysis , developed by nro @xcite .
we excluded bad spectra , which includes strong baseline undulation and spurious lines , from the raw data .
then , linear baselines were subtracted , and the raw data were regridded to 6@xmath35 per pixel with an effective angular resolution of approximately 20@xmath35 ( or 870 pc ) .
we binned the adjacent spectral channels to a velocity resolution of 10 km s@xmath26 for the spectra .
finally , we created three - dimensional data cubes in @xmath1co(@xmath0 ) , @xmath2co(@xmath0 ) , and c@xmath3o(@xmath0 ) emission .
the resultant r.m.s .
noise levels ( 1 @xmath36 ) were 60 mk , 39 mk , and 40 mk for @xmath1co(@xmath0 ) , @xmath2co(@xmath0 ) , and c@xmath3o(@xmath0 ) , respectively .
figure 2 shows @xmath1co(@xmath0 ) spectra of the whole optical disk in ngc 2903 . as is the case in earlier studies ( e.g. , @xcite , k07 , @xcite ) , strong @xmath1co(@xmath0 ) emission , whose peak temperature was @xmath16 0.6 k , was found at the center and significant @xmath1co(@xmath0 ) emission was detected in the bar ( @xmath16 0.3 k ) , bar - ends ( 0.4 0.5 k ) , and spiral arms ( 0.1 0.3 k ) .
we calculate velocity - integrated @xmath1co(@xmath0 ) intensities ( @xmath37 ) from the spectra . in order to obtain more accurate line intensities ( in other words , to minimalize the effects of the noise and the undulation of baseline for weak line ) , we defined the `` line channels '' , which are successive velocity channels where significant emission exists , in advance for each pixel as described below . in order to define the `` line channels ''
, we utilized @xmath1co(@xmath7 ) data ( @xcite ) , which was regridded and convolved to 20@xmath35 to match our @xmath1co(@xmath0 ) spectra . since the 1 @xmath36 r.m.s . of @xmath1co(@xmath7 ) data of 6 mk at 20@xmath35 and
10 km s@xmath26 resolutions was 10 times better than that of our @xmath1co(@xmath0 ) data , the @xmath1co(@xmath7 ) spectra are appropriate for the decision of `` line channels '' in each pixel .
we first identified a velocity channel exhibiting the peak @xmath1co(@xmath7 ) temperature and defined the channel as the `` co peak channel '' for each pixel .
then , successive channels whose @xmath1co(@xmath7 ) emission consistently exceeds 1 @xmath36 including the `` co peak channel '' are defined as `` line channels '' . finally , we calculated @xmath37 for the specified `` line channels '' in each pixel .
figure 3 shows the @xmath37 map of ngc 2903 .
the strongest @xmath37 of 92 k km s@xmath26 is observed at the center , and the secondary peak of 55 k km s@xmath26 is at the northern bar - end .
the total molecular gas mass for the observed area in ngc 2903 is estimated to ( @xmath38 ) @xmath39 @xmath40 under the assumptions of the constant @xmath19 of @xmath41 @xmath42 ( k km s@xmath26)@xmath26 @xcite over the disk and the uncertainty of 20% in brightness temperature scape of co line .
this value is consistent with @xmath43 @xmath40 obtained by k07 , which is recalculated using the same distance and @xmath19 .
we also compare @xmath37 obtained by coming with those obtained by k07 to confirm the validity of our @xmath1co(@xmath0 ) data .
we examined the pixel - by - pixel comparison for the two @xmath37 maps at the same angular resolution of 20@xmath18 as shown in figure 4 , and confirmed that both @xmath37 are well correlated with each other .
the median and the standard deviation in @xmath37 are 12.9 k km s@xmath26 and 12.4 k km s@xmath26 for coming dataset , and those are 15.2 k km s@xmath26 and 12.6 k km s@xmath26 for k07 dataset .
figure 5(a ) shows the global @xmath2co(@xmath0 ) spectra , which are overlaid by @xmath1co(@xmath0 ) spectra for comparison .
in addition , figure 5(b ) , ( c ) , and ( d ) show the magnified @xmath2co(@xmath0 ) spectra at the northern bar - end , the center , and the southern bar - end , respectively .
we found significant @xmath2co(@xmath0 ) emission at the center and both bar - ends .
however , we could not detect any significant c@xmath3o(@xmath0 ) emission .
we calculated the velocity - integrated @xmath2co(@xmath0 ) intensities ( @xmath44 ) . as is the case of @xmath1co(@xmath0 ) , we utilized the `` line channels '' defined by @xmath1co(@xmath7 ) spectra .
figure 6 shows a spatial distribution of @xmath44 in pseudo - color overlaid by @xmath37 in contour .
the global distribution of @xmath2co(@xmath0 ) is similar to @xmath1co(@xmath0 ) ; several peaks whose @xmath44 exceeds 5 k km s@xmath26 are observed at the center , bar - ends , and in spiral arms .
intensity ratios of two ( or more ) molecular lines provide important clues to estimate physical properties of molecular gas , such as density and temperature .
we examined the spatial variations in line intensity ratios among @xmath1co(@xmath0 ) , @xmath2co(@xmath0 ) , and @xmath1co(@xmath7 ) emission .
figure 7 shows spatial distributions of @xmath8 and @xmath9 over the disk of ngc 2903 . in these maps
, we displayed pixels with each line intensity exceeding 2 @xmath36 ( 5 k km s@xmath26 , 3 k km s@xmath26 , and 0.5 k km s@xmath26 for @xmath1co(@xmath0 ) , @xmath2co(@xmath0 ) , and @xmath1co(@xmath7 ) , respectively ) .
we found some local peaks of @xmath8 ( @xmath16 1.0 ) near the center and at the downstream side of the northern spiral arm , whereas lower @xmath8 ( @xmath16 0.5 0.6 ) was observed in the bar . the spatial distribution of @xmath9 is unclear due to poor signal - to - noise ( s / n ) ratio of @xmath2co(@xmath0 ) emission .
as described in section 3.3 , the spatial distribution of @xmath9 seems noisy and unclear due to the poor s /
n although we could obtain the spatial distribution of @xmath44 . in order to improve the s /
n of weak emission such as @xmath2co(@xmath0 ) , the stacking analysis of co spectra with the velocity - axis alignment seems a promising method .
the stacking technique for co spectra in external galaxies are originally demonstrated by schruba et al .
( 2011 , 2012 ) .
since the observed velocities of each position within a galaxy are different due to its rotation , a simple stacking causes a smearing of spectrum . in order to overcome such difficulty
, @xcite demonstrated the velocity - axis alignment of co spectra in different regions in a galaxy disk according to mean h velocity .
they stacked velocity - axis aligned co spectra , and successfully confirmed very weak @xmath1co(@xmath7 ) emission ( @xmath45 1 k km s@xmath26 ) with high significance in h - dominated outer - disk regions of nearby spiral galaxies . in addition , @xcite applied this stacking technique to perform the sensitive search for weak @xmath1co(@xmath7 ) emission in dwarf galaxies .
furthermore , @xcite applied the stacking technique to @xmath2co(@xmath0 ) emission in the optical disk of the nearby barred spiral galaxy ngc 3627 . by the stacking with velocity - axis alignment based on mean @xmath1co(@xmath0 ) velocity , they obtained high s
/ n @xmath2co(@xmath0 ) spectra which are improved by a factor of up to 3.2 compared to the normal ( without velocity - axis alignment ) stacking analysis . these earlier studies clearly suggest that the stacking analysis is very useful to detect weak molecular line . in this section
, we employ the same stacking technique as @xcite to improve the s / n of @xmath2co(@xmath0 ) emission and to obtain more accurate line ratios . based on our @xmath37 image ( figure 3 ) ,
we have separated ngc 2903 into nine regions according to its major structures ; i.e. , center , northern bar , southern bar , northern bar - end , southern bar - end , northern arm , southern arm , inter - arm , and outer - disk .
the left panel of figure 8 shows the separation of each region overlaid by the grey - scale map of @xmath37 . for each region , we stacked @xmath1co(@xmath0 ) , @xmath2co(@xmath0 ) , and @xmath1co(@xmath7 ) spectra with velocity - axis alignment based on the intensity - weighted mean velocity field calculated from our @xmath1co(@xmath0 ) data ( right panel of figure 8) .
we successfully obtained the stacked co spectra as shown in figure 9 .
the s / n of each co emission is dramatically improved , and thus we could confirm the significant @xmath2co(@xmath0 ) emission for all the regions .
we found the difference in the line shape of stacked co spectra according to regions .
in particular , the stacked @xmath1co spectra in the bar show flat peak over the velocity width of 100 150 km s@xmath26 .
this is presumably due to the rapid velocity change in the bar , which makes the velocity - axis alignment difficult .
we summarize the averaged line intensities and line ratios for each region in table 2 .
we found that the averaged @xmath8 shows the highest value of 0.92 at the center , and a moderate value of 0.7 0.8 at both bar - ends and in the northern arm . a slightly lower @xmath8 of 0.6 0.7
is observed in the bar , southern arm , inter - arm , and outer - disk .
such a variation in @xmath8 ranging from 0.6 to 1.0 in ngc 2903 is quite consistent with those observed in nearby galaxies ( e.g. , @xcite ) .
however , the highest @xmath9 of 0.19 is observed not at the center but in the northern arm .
the central @xmath9 of 0.11 is similar to those in other regions ( 0.08 0.13 ) except for the northern arm and outer - disk ( @xmath16 0.04 ) .
the typical @xmath9 of @xmath16 0.1 is frequently observed in nearby galaxies ( e.g. , @xcite ; @xcite ) , but slightly higher than the averaged @xmath9 in representative regions of ngc 3627 , 0.04 0.09 @xcite .
using @xmath8 and @xmath9 , we derive averaged physical properties of molecular gas , its density ( @xmath10 ) and kinetic temperature ( @xmath11 ) , in seven regions ( center , northern bar , southern bar , northern bar - end , southern bar - end , northern arm , and southern arm ) of ngc 2903 based on the lvg approximation .
some assumptions are required to perform the lvg calculation ; the molecular abundances @xmath46(@xmath1co ) = [ @xmath1co]/[h@xmath15 ] , [ @xmath1co]/[@xmath2co ] , and the velocity gradient @xmath48 .
firstly , we fix the @xmath46(@xmath1co ) of @xmath49 and @xmath48 of 1.0 km s@xmath26 pc@xmath26 ; i.e. , @xmath1co abundance per unit velocity gradient @xmath46(@xmath1co)/(@xmath48 ) was assumed to @xmath49 ( km s@xmath26 pc@xmath26)@xmath26 .
this is the same as the assumed @xmath46(@xmath1co)/(@xmath48 ) for gmcs in m 33 @xcite .
then , we determine the [ @xmath1co]/[@xmath2co ] abundance ratio to be assumed in this study by considering earlier studies .
@xcite found a systematic gradient in the @xmath1c/@xmath2c isotopic ratio across in our galaxy ; from @xmath16 30 in the inner part at 5 kpc to @xmath16 70 at 12 kpc with a galactic center value of 24 . for external galaxies ,
the reported @xmath1c/@xmath2c isotopic ratios in their central regions are 40 for ngc 253 @xcite , 50 for ngc 4945 @xcite , @xmath50 for m 82 and @xmath51 for ic 342 @xcite .
@xcite reported a higher [ @xmath1co]/[@xmath2co ] abundance ratio of 50 75 in the central region of m 82 .
@xcite also reported a higher @xmath1c/@xmath2c isotopic ratios of @xmath52 50 100 in the central regions of m 82 and ngc 253 . in summary , reported @xmath1c/@xmath2c isotopic ( and [ @xmath1co]/[@xmath2co ] abundance ) ratios in nearby galaxy centers ( 30 100 ) are typically higher than that in the inner 5 kpc of our galaxy ( 24 30 ) , but the cause of such discrepancies in @xmath1c/@xmath2c and [ @xmath1co]/[@xmath2co ] between our galaxy and external galaxies is still unresolved . here , we assumed an intermediate [ @xmath1co]/[@xmath2co ] abundance ratio of 50 in ngc 2903 without any gradient across its disk for our lvg calculation .
note that we perform an additional lvg calculation for the center of ngc 2903 assuming the [ @xmath1co]/[@xmath2co ] abundance ratios of 30 and 70 to evaluate the effect of the variation in the assumed [ @xmath1co]/[@xmath2co ] abundance ratio on results of lvg calculation .
figure 10 shows results of lvg calculation for each region in ngc 2903 .
the thin line indicates a curve of constant @xmath8 as functions of @xmath10 and @xmath11 , and the thick line indicates that of constant @xmath9 .
we can determine both @xmath10 and @xmath11 at the point where two curves intersect each other .
under the assumption of [ @xmath1co]/[@xmath2co ] abundance ratio of 50 , the derived @xmath10 ranges from @xmath161000 @xmath13 ( in the disk ; i.e. , bar , bar - ends , and spiral arms ) to 3700 @xmath13 ( at the center ) and the derived @xmath11 ranges from 10 k ( in spiral arms ) to 30 k ( at the center ) .
note that both @xmath10 and @xmath11 vary depending on the assumption of [ @xmath1co]/[@xmath2co ] abundance ratio ; at the center of ngc 2903 , the abundance ratio of 30 yields lower @xmath10 of 1800 @xmath13 and higher @xmath11 of 38 k , whereas the abundance ratio of 70 yields higher @xmath10 of 5900 @xmath13 and intermediate @xmath11 of 29 k. it seems that @xmath10 is proportional to [ @xmath1co]/[@xmath2co ] abundance ratio .
this trend of @xmath10 can be naturally explained if we consider the optical depth of @xmath1co and @xmath2co emission .
@xmath1co is always optically thick and thus its emission emerges from the diffuse envelope of dense gas clouds , while @xmath2co emission emerges from further within these dense gas clouds due to its lower abundance . since the increase in the assumed [ @xmath1co]/[@xmath2co ] abundance ratio means that @xmath2co becomes more optically thin , @xmath2co emission emerged from deeper within the dense gas clouds and thus it probes denser gas . derived physical properties , @xmath10 and @xmath11 , are summarized in table 3 .
we compare the derived @xmath10 and @xmath11 in ngc 2903 with those determined in other external galaxies .
@xcite determined @xmath10 and @xmath11 for gmcs associated with the giant h region ngc 604 in m 33 at a spatial resolution of 100 pc using three molecular lines , @xmath1co(@xmath0 ) , @xmath2co(@xmath0 ) , and @xmath1co(@xmath22 ) , based on the lvg approximation .
the derived @xmath10 and @xmath11 are 800 2500 @xmath13 and 20 30 k , respectively , which are similar to our study for ngc 2903 in spite of the difference in the spatial resolution .
however , @xcite obtained different physical properties for gmcs in spiral arms of m 51 .
they performed the lvg analysis using @xmath9 and @xmath8 at a spatial resolution of 120 180 pc .
for the case of constant @xmath48 = 1.0 km s@xmath26 pc@xmath26 , the derived @xmath11 ranges from 10 to 50 k , which is similar to our study for ngc 2903 , whereas @xmath10 ranges from 100 to 400 @xmath13 , which is 5 10 times lower than that in the disk of ngc 2903 in spite that the values of @xmath8 and @xmath9 in m 51 are not so different from those in ngc 2903 .
this is presumably due to the differences in assumed @xmath46(@xmath1co ) and [ @xmath1co]/[@xmath2co ] abundance ratio .
the authors assumed @xmath46(@xmath1co ) of 8.0 @xmath53 , which is higher than that assumed in our study , and a lower [ @xmath1co]/[@xmath2co ] abundance ratio of 30 .
under the lvg approximation with the assumption of @xmath46(@xmath1co ) of 8.0 @xmath53 , we found that the derived @xmath10 is typically @xmath163 times lower than that with assumption of @xmath46(@xmath1co ) of 1.0 @xmath53 .
physically , high @xmath46(@xmath1co ) means abundant @xmath1co molecules among molecular gas . in this condition ,
the optical depth of @xmath1co line also increases , and thus the photon - trapping effect in molecular clouds becomes effective . since this effect contributes the excitation of @xmath1co molecule , an effective critical density of @xmath1co line decreases .
in other words , since the @xmath1co is easily excited to upper @xmath54 levels even in low molecular gas density , @xmath10 at a given @xmath8 decreases . as a result , lvg analysis with the assumption of @xmath46(@xmath1co ) of 8.0 @xmath53 yields lower @xmath10 .
in addition , the low [ @xmath1co]/[@xmath2co ] abundance ratio of 30 causes the decrease in the derived molecular gas density as described above .
therefore , the difference in the derived @xmath10 between ngc 2903 and m 51 can be explained by the difference in the assumed @xmath46(@xmath1co ) and the [ @xmath1co]/[@xmath2co ] abundance ratio . as described in section 1 ,
sfes often differ between galaxy centers and disks .
since ngc 2903 has a bright star forming region at the center , its sfe is expected to be higher than those in other regions .
here , we calculate sfes for seven regions where averaged physical properties of molecular gas are obtained , and compare sfe with @xmath10 and @xmath11 in each region to examine what parameter controls sfe in galaxies .
sfe is expressed using the surface density of sfr ( @xmath55 ) and that of molecular hydrogen ( @xmath56 ) as follows : @xmath57= \left ( \frac{\sigma_{\rm sfr}}{m_{\odot}\,{\rm yr^{-1}\,pc^{-2 } } } \right ) { \displaystyle \biggl/ } \left ( \frac{\sigma_{\rm h_2}}{m_{\odot}\,{\rm pc^{-2 } } } \right ) \end{aligned}\ ] ] we calculated extinction - corrected sfrs from a linear combination of h@xmath23 and @xmath58/mips 24 @xmath59 luminosities using a following formula @xcite : @xmath60 where @xmath61 and @xmath62 mean h@xmath23 and 24 @xmath59 luminosities , respectively .
@xmath63 is the inclination of for ngc 2903 and @xmath64 is the covered area for each region ( in the unit of pc@xmath65 ) .
we used archival continuum - subtracted h@xmath23 and 24 @xmath59 images of ngc 2903 obtained by @xcite and the local volume legacy survey project @xcite , respectively .
in addition , we calculated @xmath56 using @xmath37 as follows : @xmath66 & = & 2.89 \times { \rm cos } \
i \left ( \frac{i_{\rm 12co(1 - 0)}}{{\rm k\,\,km\,\,s^{-1 } } } \right ) \times \left\ { \frac{x_{\rm co}}{1.8 \times 10^{20}\,{\rm cm}^{-2}\,({\rm k\,\,km\,\,s^{-1}})^{-1 } } \right\ } .\end{aligned}\ ] ] here , we adopted a constant @xmath19 value of @xmath41 @xmath42 ( k km s@xmath26)@xmath26 @xcite .
we found that sfe at the center , @xmath67 yr@xmath26 , is 2 4 times higher than those in other regions .
calculated sfes are listed in table 3 .
we examined the dependence of sfe on @xmath10 and @xmath11 as shown in figure 11 .
we found that sfe positively correlates with both @xmath10 and @xmath11 .
however , the trend of these correlations might change because it is possible that variations in the [ @xmath1co]/[@xmath2co ] abundance ratio and @xmath19 affect the estimate of @xmath10 , @xmath11 , and sfe .
in fact , both the [ @xmath1co]/[@xmath2co ] abundance ratio and @xmath19 often differ between galaxy centers and disks .
therefore , we examine how variations in the [ @xmath1co]/[@xmath2co ] abundance ratio and @xmath19 alter the estimate of @xmath10 , @xmath11 , and sfe at the center of ngc 2903 .
we first consider the effect of the variation in [ @xmath1co]/[@xmath2co ] abundance ratio on the estimate of @xmath10 and @xmath11 .
as described in section 4.2.1 , it is reported that the @xmath1c/@xmath2c abundance ratio in our galaxy increases with the galactocentric radius @xcite .
thus we examine the case of the lower [ @xmath1co]/[@xmath2co ] abundance ratio at the center of ngc 2903 .
if we adopt the [ @xmath1co]/[@xmath2co ] abundance ratio of 30 at the center , @xmath10 and @xmath11 are estimated to be 1800 @xmath13 and 38 k , respectively .
this @xmath10 value is slightly lower than that in the northern arm , whereas the positive correlation between sfe and @xmath10 is not destroyed .
similarly , the @xmath11 of 38 k does not destroy the positive correlation between sfe and @xmath11 .
next , we consider the effect of the variation in @xmath19 on the estimate of sfe . in central regions of disk galaxies , @xmath19 drops ( i.e. , co emission becomes luminous at a given gas mass ) by a factor of 2 3 or more ( e.g. , @xcite ; @xcite ) , including the galactic center ( e.g. , @xcite ; @xcite ) .
such a trend is presumably applicable to ngc 2903 considering the relationship between @xmath19 and metallicity , 12 + log(o / h ) .
in general , @xmath19 decreases with the increase in metallicity because the co abundance should be proportional to the carbon and oxygen abundances ( e.g. , @xcite ; @xcite ) .
in addition , it is reported that metallicity decreases with the galactocentric distance in ngc 2903 ( e.g. , @xcite ; @xcite ) .
these observational facts suggest a smaller @xmath19 by a factor of 1.5 2 at the center than in the disk of ngc 2903 , which yields a smaller gas mass , providing a higher sfe than the present one shown in table 3 and figure 11 .
however , even if a higher sfe by a factor of 2 is adopted for the center , the global trend of the correlations shown in figure 11 does not change so much because the original sfe at the center is already the highest in ngc 2903 .
therefore , we concluded that variations in the [ @xmath1co]/[@xmath2co ] abundance ratio and @xmath19 do @xmath68 affect the correlations of sfe with @xmath10 and @xmath11 in ngc 2903 .
note that the smaller @xmath19 corresponds to the larger @xmath46(@xmath1co ) at the center of ngc 2903 , but the larger @xmath48 is also suggested because the typical velocity width at the center ( 250 300 km s@xmath26 ) is wider than those in other regions ( 150 200 km s@xmath26 ) due to the rapid rotation of molecular gas near the galaxy center .
thus we consider that @xmath69co)/(@xmath48 ) itself does not differ between the center and the disk in ngc 2903 even if the @xmath69co ) at the center is larger than that in the disk .
finally , we examine the correlation coefficient for the least - square power - law fit @xmath14 between sfe and @xmath10 and that between sfe and @xmath11 shown in figure 11 .
we found that the former is 0.50 and the latter is 0.08 .
the significant correlation between sfe and @xmath10 with @xmath14 of 0.50 suggests that molecular gas density governs the spatial variations in sfe .
this speculation is well consistent with earlier studies based on hcn emission ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ) . in order to confirm whether such a relationship between sfe and @xmath10 is applicable to other galaxies or not , we will perform further analysis toward other coming sample galaxies , considering variations in the [ @xmath1co]/[@xmath2co ] abundance ratio , @xmath19 , and @xmath69co)/(@xmath48 ) , in forthcoming papers .
we have performed the simultaneous mappings of @xmath0 emission of @xmath1co , @xmath2co , and c@xmath3o molecules toward the whole disk ( @xmath4 or @xmath24 kpc at the distance of 8.9 mpc ) of the nearby barred spiral galaxy ngc 2903 with the nro 45-m telescope equipped with the forest at an effective angular resolution of @xmath6 ( or 870 pc ) .
a summary of this work is as follows . 1 . we detected @xmath1co(@xmath0 ) emission over the disk of ngc 2903 .
in addition , significant @xmath2co(@xmath0 ) emission was found at the center and bar - ends , whereas we could not detect any significant c@xmath3o(@xmath0 ) emission .
2 . in order to improve the s / n of co emission and to measure @xmath8 and @xmath9 with high significance
, we performed the stacking analysis for our @xmath1co(@xmath0 ) , @xmath2co(@xmath0 ) , and archival @xmath1co(@xmath7 ) spectra with velocity - axis alignment in nine representative regions ( i.e. , center , northern bar , southern bar , northern bar - end , southern bar - end , northern arm , southern arm , inter - arm , and outer - disk ) of ngc 2903 .
we successfully obtained the stacked co spectra with highly improved s / n , and thus we could confirm the significant @xmath2co(@xmath0 ) emission for all the regions .
we examined the averaged @xmath8 and @xmath9 for nine regions , and found that the averaged @xmath8 shows the highest value of 0.92 at the center , and moderate or lower values of 0.6 0.8 are observed in the disk .
however , the highest @xmath9 of 0.19 is observed not at the center but in the northern arm .
the central @xmath9 of 0.11 is similar to those in other regions ( 0.08 0.13 ) except for the northern arm and outer - disk ( @xmath16 0.04 ) .
4 . we determined @xmath10 and @xmath11 of molecular gas using @xmath8 and @xmath9 based on the lvg approximation . under the assumption of [ @xmath1co]/[@xmath2co ] abundance ratio of 50 ,
the derived @xmath10 ranges from @xmath161000 @xmath13 ( in the bar , bar - ends , and spiral arms ) to 3700 @xmath13 ( at the center ) and the derived @xmath11 ranges from 10 k ( in the bar and spiral arms ) to 30 k ( at the center ) .
we examined the dependence of sfe on @xmath10 and @xmath11 of molecular gas , and found the positive correlation between sfe and @xmath10 with the correlation coefficient for the least - square power - law fit @xmath14 of 0.50 .
this suggests that molecular gas density governs the spatial variations in sfe .
we thank the referee for invaluable comments , which significantly improved the manuscript .
we are indebted to the nro staff for the commissioning and operation of the 45-m telescope and their continuous efforts to improve the performance of the instruments .
this work is based on observations at nro , which is a branch of the national astronomical observatory of japan , national institutes of natural sciences .
this research has made use of the nasa / ipac extragalactic database , which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
co(@xmath0 ) emission in ngc 2903 .
the grid spacing was set to 12@xmath18 in order to display the spectrum in each pixel clearly .
the temperature scale of spectra is indicated by the small box inserted in the lower left corner .
, width=302 ] in ngc 2903 obtained by coming and those obtained by k07 .
the vertical and horizontal lines indicate the 2 @xmath36 of @xmath37 for coming and k07 data , respectively .
the diagonal solid line indicates the ratio of @xmath37 by coming to those by k07 of unity , and the dashed lines indicate the ratio of 0.5 and 2.0 .
both @xmath37 are well correlated with each other . ,
width=302 ] co(@xmath0 ) emission in ngc 2903 .
for comparison , spectra of @xmath1co(@xmath0 ) emission multiplied by 0.5 are overlaid in red line . as well as figure 2 , the grid spacing
was set to 12@xmath18 .
the temperature scale of spectra is indicated by the small box inserted in the lower left corner .
( b ) magnified spectra with the grid spacing of 6@xmath18 at the northern bar - end .
( c ) same as ( b ) , but at the center .
( d ) same as ( b ) , but at the southern bar - end .
, width=642 ] ( left ) and @xmath9 ( right ) superposed on the contour map of @xmath37 of ngc 2903 .
the contour levels of @xmath37 are the same as figure 3 .
there are some local peaks of @xmath8 ( @xmath16 1.0 ) near the center and at the downstream side of the northern spiral arm , whereas lower @xmath8 ( @xmath16 0.5 0.6 ) was observed in the bar .
the spatial distribution of @xmath9 is unclear due to poor s / n of @xmath2co(@xmath0 ) emission .
, width=642 ] of ngc 2903 .
the green frame indicates the center , the purple and orange indicate the northern and southern bars , the blue and red indicate the northern and southern bar - ends , the cyan and magenta indicate the northern and southern arms , the yellow indicates the inter - arm , and the grey indicates the outer - disk .
( right ) intensity - weighted mean velocity field calculated from @xmath1co(@xmath0 ) data superposed on the contour map of @xmath37 .
the contour levels of @xmath37 are the same as figure 3 .
, width=642 ] co(@xmath0 ) emission multiplied by 0.5 , the red indicates @xmath1co(@xmath7 ) emission multiplied by 0.5 , and the blue indicates @xmath2co(@xmath0 ) emission .
the s / n of each co emission is dramatically improved , and thus a significant @xmath2co(@xmath0 ) emission is confirmed for all the regions . , width=642 ] ( thin line ) and @xmath9 ( thick line ) as functions of molecular gas density @xmath10 and kinetic temperature @xmath11 .
@xmath1co fractional abundance per unit velocity gradient @xmath69co)/(@xmath48 ) was assumed to be @xmath49 ( km s@xmath26 pc@xmath26)@xmath26 .
the [ @xmath1co]/[@xmath2co ] abundance ratio was assumed to be a fixed value of 50 for the bar , bar - ends , and spiral arms , but three different [ @xmath1co]/[@xmath2co ] abundance ratios of 30 , 50 , and 70 were assumed for the center .
dashed lines indicate @xmath70 error of each line ratio .
, width=642 ] lc + morphological type@xmath71 & sab(rs)bc + map center@xmath72 : & + right ascension ( j2000.0 ) & + declination ( j2000.0 ) & + distance@xmath73 & 8.9 mpc + linear scale & 43 pc arcsec@xmath26 + inclination@xmath74 & + + lccccc + & @xmath37 & @xmath75 & @xmath44 & @xmath8 & @xmath9 + & [ k km s@xmath26 ] & [ k km s@xmath26 ] & [ k km s@xmath26 ] & & + + center & @xmath77 & @xmath78 & @xmath79 & @xmath80 & @xmath81 + northern bar & @xmath82 & @xmath83 & @xmath84 & @xmath85 & @xmath86 + southern bar & @xmath87 & @xmath88 & @xmath89 & @xmath90 & @xmath91 + northern bar - end & @xmath92 & @xmath93 & @xmath94 & @xmath95 & @xmath96 + southern bar - end & @xmath97 & @xmath98 & @xmath99 & @xmath95 & @xmath86 + northern arm & @xmath100 & @xmath101 & @xmath102 & @xmath103 & @xmath104 + southern arm & @xmath105 & @xmath106 & @xmath107 & @xmath108 & @xmath109 + inter - arm & @xmath110 & @xmath111 & @xmath112 & @xmath108 & @xmath96 + outer - disk & @xmath113 & @xmath114 & @xmath115 & @xmath116 & @xmath117 + + lcccc + region & [ @xmath1co]/[@xmath2co ] & @xmath10 & @xmath11 & sfe + & & [ @xmath118 @xmath13 ] & [ k ] & [ @xmath120 yr@xmath26 ] + + center & 50 & @xmath121 & @xmath122 & @xmath123 + northern bar & 50 & @xmath124 & @xmath125 & @xmath126 + southern bar & 50 & @xmath127 & @xmath128 & @xmath129 + northern bar - end & 50 & @xmath130 & @xmath131 & @xmath132 + southern bar - end & 50 & @xmath112 & @xmath133 & @xmath134 + northern arm & 50 & @xmath135 & @xmath136 & @xmath137 + southern arm & 50 & @xmath138 & @xmath139 & @xmath140 + center & 30 & @xmath141 & @xmath142 & @xmath123 + center & 70 & @xmath143 & @xmath144 & @xmath123 + + | we present simultaneous mappings of @xmath0 emission of @xmath1co , @xmath2co , and c@xmath3o molecules toward the whole disk ( @xmath4 or 20.8 kpc @xmath5 13.0 kpc ) of the nearby barred spiral galaxy ngc 2903 with the nobeyama radio observatory 45-m telescope at an effective angular resolution of @xmath6 ( or 870 pc ) .
we detected @xmath1co(@xmath0 ) emission over the disk of ngc 2903 .
in addition , significant @xmath2co(@xmath0 ) emission was found at the center and bar - ends , whereas we could not detect any significant c@xmath3o(@xmath0 ) emission . in order to improve the signal - to - noise ratio of co emission and to obtain accurate line ratios of @xmath1co(@xmath7)/@xmath1co(@xmath0 ) ( @xmath8 ) and @xmath2co(@xmath0)/@xmath1co(@xmath0 ) ( @xmath9 ) , we performed the stacking analysis for our @xmath1co(@xmath0 ) , @xmath2co(@xmath0 ) , and archival @xmath1co(@xmath7 ) spectra with velocity - axis alignment in nine representative regions of ngc 2903 .
we successfully obtained the stacked spectra of the three co lines , and could measure averaged @xmath8 and @xmath9 with high significance for all the regions .
we found that both @xmath8 and @xmath9 differ according to the regions , which reflects the difference in the physical properties of molecular gas ; i.e. , density ( @xmath10 ) and kinetic temperature ( @xmath11 ) .
we determined @xmath10 and @xmath11 using @xmath8 and @xmath9 based on the large velocity gradient approximation .
the derived @xmath10 ranges from @xmath12 @xmath13 ( in the bar , bar - ends , and spiral arms ) to 3700 @xmath13 ( at the center ) and the derived @xmath11 ranges from 10 k ( in the bar and spiral arms ) to 30 k ( at the center ) .
we examined the dependence of star formation efficiencies ( sfes ) on @xmath10 and @xmath11 , and found the positive correlation between sfe and @xmath10 with the correlation coefficient for the least - square power - law fit @xmath14 of 0.50 .
this suggests that molecular gas density governs the spatial variations in sfes . |
when a magnetic impurity is located in the bulk of a metal it undergoes a non - trivial many - body scattering with the electronic states at the fermi energy , @xmath0 , called the kondo effect @xcite . the bulk kondo effect ( to be contrasted with the surface kondo effect studied in this paper ) is one of the best studied problems in condensed matter physics and many different techniques from renormalization group @xcite to bethe ansatz @xcite have been used over the years .
the basic mechanism of the kondo effect is the magnetic screening of the impurity ( the formation of the kondo singlet ) at temperatures @xmath1 below the kondo temperature @xmath2 . above @xmath2
the magnetic impurity behaves paramagnetically but for @xmath3 a resonance is formed close to the fermi surface .
the kondo effect is very important in modern condensed matter theory and appears in many different areas of research , from u and ce intermetallics ( heavy - fermions ) @xcite to quantum dots @xcite .
there is a wide variety of phenomena that can be described within the universality of the kondo problem .
the kondo effect can be observed experimentally in measurements of the temperature dependence of resistivity ( the so - called resistivity minimum ) @xcite , and also in thermodynamic measurements such as specific heat and magnetic susceptibility @xcite of dilute magnetic alloys , due to the enhancement of the density of states close to the fermi energy ( the abrikosov - suhl resonance @xcite ) .
the enhancement of the density of states is related to the change of the characteristic energy scales from @xmath0 to @xmath4 and therefore in the density of states from @xmath5 to @xmath6 since @xmath7 .
the kondo effect was also observed recently in stm studies of magnetic atoms on metallic surfaces @xcite ( see fig.[fig - system ] ) . in a stm experiment electrons from a sharp tip tunnel into the material to be studied , creating a tunneling current @xmath8 due to the application of a potential @xmath9 . roughly speaking a stm experiment measures the local density of states at the fermi energy via the differential tunneling conductance @xmath10 @xcite . when a stm tip is away from the magnetic impurity it measures the substrate density of states , @xmath11 ; however , close to the impurity ( that we call the adatom ) electrons from the tip can tunnel directly to the impurity .
the theory of stm is far from trivial because electrons from the tip not only tunnel to the adatom but also to the bulk and surface states , that is , there are various different channels of tunneling that lead to interference effects that have to be taken into consideration for the proper interpretation of the experimental data .
is the distance of the adatom to the surface , @xmath12 is the distance of the stm tip from the surface , and @xmath13 is the distance , along the surface , between the stm tip and the adatom.,width=226 ] the basic starting point for the study of the kondo effect is the anderson impurity hamiltonian @xcite : @xmath14 where @xmath15 ( @xmath16 ) is the annihilation ( creation ) operator for an electron on a localized atomic state ( impurity state ) with angular momentum labelled by @xmath17 , energy @xmath18 and spin @xmath19 ( @xmath20 ) , @xmath21 ( @xmath22 ) is the annihilation ( creation ) operator for an electron on the conduction band with dispersion @xmath23 and momentum @xmath24 , @xmath25 is the number operator , @xmath26 is the coulomb energy for the double occupancy of the impurity state , and @xmath27 is the hybridization energy between the impurity and conduction states ( the electron operators obey anti - commutation rules : @xmath28 ) .
one of the characteristics of the anderson impurity model is the distinction between substrate and adatom wavefunctions .
although most theoretical works do not question the distinction made _ a priori _ between these quantum states , it turns out that this distinction is not completely natural .
the reason is very simple : when an impurity atom is introduced in a metal , it hybridizes with the metallic states losing its identity .
however , it leaves behind a phase shift @xmath29 in the original metallic states .
thus , the impurity state can not be really distinguished from the host states from the quantum mechanical point of view .
nevertheless , anderson @xcite makes the point that because the d - orbitals are a inner shell , the coulomb energy @xmath26 for double occupancy of those orbitals is large and they must be distinguished from metallic states where the coulomb energy is strongly suppressed by screening effects .
thus , the distinction between these two types of states can only be clearly made when these states are orthogonal to each other so that the impurity does not cause a direct perturbation in the substrate spectrum .
nevertheless , even in metals where the electronic bands are generated out of d - orbitals ( as in the case of cu ) , the strong metallic bonding leads to a large s - wave character of the bands and to very small overlap with the adatom d - orbitals ( that is , these states are `` naturally '' orthogonal @xcite ) .
nevertheless , this orthogonality can only be distinguished _
a posteriori_. in fact , we have found , by direct numerical computation , that this is the case in the surface kondo effect . finally , as pointed out by anderson @xcite ( see also ref.@xcite ) , the orthogonality of these states is not fundamental for the magnetic phenomenon which is essentially a local effect and all the subtleties associated with orthogonality become encapsulated into the hybridization matrix elements @xmath27 which become a phenomenological parameter to be obtained indirectly from the experiment .
however , in trying to understand the stm experiments , and especially the role played by the surface and bulk states , we can not simply take these matrix elements as phenomenological parameters since we would not be able to separate the contributions coming from the bulk and the surface of the cu substrate .
hence , in this paper we follow the scheme proposed by tsvelik and wiegmann @xcite and work with non - orthogonalized states . in this case
the hybridization energies are given by : @xmath30 where @xmath31 is the atomic state of the impurity which is localized on the impurity site ( that is , it decays exponentially away from the impurity ) and @xmath32 is the bloch wavefunction of the crystal .
the hybridization energy depends not only on the amplitude but also on the direction of @xmath24 .
nevertheless , it is usually assumed that the hybridization is constant in momentum space , that is , @xmath33 .
although this assumption maybe warranted for spherical gapless fermi surfaces ( since the momentum dependence averages out " and most of the interesting physics occurs at the fermi surface ) it is not valid for metals where the chemical potential crosses gaps in the electronic spectrum .
the kondo hamiltonian can be obtained from ( [ hamiltonian ] ) in the limit of @xmath34 via the schrieffer - wolf transformation @xcite that maps the hamiltonian ( [ hamiltonian ] ) onto : @xmath35 where @xmath36 is the impurity spin , @xmath37 is the electron spin at the impurity site ( @xmath38 with @xmath39 are pauli matrices ) and @xmath40 is the exchange interaction between local spins and conduction electrons ( the energy is measured relative to the fermi energy ) . in going from ( [ hamiltonian ] ) to ( [ kondo_hamiltonian_1 ] )
we have disregarded a series of physical processes that can not be described as spin exchange interactions , foremost being the variable occupancy of the impurity site .
even simple metals like cu have energy gaps in their electronic spectrum due to the periodic potential generated by the cu ions @xcite .
for instance , it is well - known that for cu along the ( 111 ) direction there is a gap of @xmath41 ev at the fermi energy while there are no gaps along the ( 100 ) direction ( see fig.[fs ] ) . on the one hand
, @xmath27 does not exist for bulk states when @xmath42 points along the ( 111 ) direction .
however , because of the bulk gap , one has a surface state in the cu(111 ) surface @xcite ( a state that decays exponentially in the ( 111 ) direction but it is extended in the directions perpendicular to ( 111 ) ) that can hybridize with the impurity located on the cu(111 ) surface . on the other hand , there are no surface states in the ( 100 ) direction to hybridize with an impurity located on the cu(100 ) surface , although bulk states hybridize with it since their wavefunctions decay exponentially outside of the crystal .
previous studies @xcite on the surface kondo effect either assume a momentum - independent hybridization @xmath43 or sum average over a gapless fermi surface @xcite . by ignoring momentum dependence or the gap structure of the fermi surface , the comparison between cu(111 ) and cu(100 ) surface kondo effects of noble metals is not possible .
in fact , we show that recent stm experiments of atoms located in the cu(111 ) and cu(100 ) surfaces have been interpreted incorrectly because of these differences .
in particular , we show that the surface kondo effect between a magnetic atom located on the cu(111 ) surface and the cu(111 ) surface state is quantitatively similar to the surface kondo effect of a magnetic atom on a cu(100 ) surface with bulk states .
the main reason for this similarity comes from the fact that the bulk state that dominates the kondo scattering is the one with the largest penetration in the work - function barrier , that is , the state with momentum perpendicular to the cu(100 ) surface .
the state with momentum perpendicular to the cu(100 ) surface has essentially the same decay rate out of the crystal as the surface state in the cu(111 ) surface and therefore similar hybridization with the magnetic impurity .
-l is the ( 111 ) direction ( notice the absence of states due to the bulk gap ) ; @xmath44-x is the ( 100 ) direction.,width=226 ] in the presence of the tip new terms have to be added to the hamiltonian ( 1 ) that describe the tunneling processes of tip - to - adatom and tip - to - substrate @xcite .
the tunneling processes of tip - to - adatom is given by : @xmath45 where @xmath46 ( @xmath47 ) annihilates ( creates ) and electron at the tip and @xmath48 is the hybridization energy between tip and adatom .
the tip - to - substrate hybridization is given by : @xmath49 where @xmath50 is the hybridization energy between substrate ( bulk or surface state ) and adatom .
the tip hamiltonian is simply @xmath51 where @xmath52 is the energy of the electron states on the tip .
the total hamiltonian of the tip - substrate - adsorbate system is @xmath53 . as in the case of ( [ hamiltonian ] )
the hybridization energies are given by : @xmath54 with @xmath55 labeling the wavefunctions , @xmath56 , for substrate , adatom and tip , respectively .
notice that the complexity in this problem comes from the interference between the different channels of tunneling . in the present work
, we will consider the case of co on cu surfaces while our theory is applicable to any magnetic adatom on noble metal surfaces .
our main objectives are : ( _ i _ ) the calculation of the hybridization matrix elements between the different wavefunctions in the problem ( bulk , surface , impurity , and stm tip ) with the smallest number of adjustable parameters , and ( _ ii _ ) the description of the surface kondo effect via stm measurements .
while the spectrum of bulk and surface states is readily available in the literature , very little has been published on the actual form of the states . instead of embarking on a complicated calculation of wavefunctions via heavy numerical techniques , we opted for a simpler approach that can provide quantitative results that can be directly compared to the experiment as well as intuitive understanding of the problem . as we are going to show , our simplified models have limitations in explaining some of the available stm data and more detailed work is required .
we hope our work stimulates other works that make use of more sophisticated methods to calculate the matrix elements because , as we are going to show , they are fundamental for the understanding of magnetism in metallic surfaces .
we model the periodic potential inside of the crystal in the nearly - free electron approximation and obtain the bulk gaps as observed in metallic cu .
the surface states are obtained after the modeling of the surface potential in terms of the image charge in the vacuum .
the energies of surface states and resonances in each surface are in very good agreement with photoemission results in cu(111 ) and cu(100 ) surfaces . following the literature in stm theory ,
the impurity is modeled via standard hydrogen - like wave functions .
the paper is organized as follows : in sec .
[ sec - wavefunction ] , we model the ionic potentials that give rise to the wavefunctions of the metallic host , adatom , and the stm tip . using the wavefunctions and their spectrum we discuss the cu band structure , both the bulk and cu(111 ) surface state . in sec .
[ sec - hybridization ] we present the results for the hybridization matrix elements that determine the hybridization energies as a function of the distance from the surface and adatom and also as a function of the momentum .
the surface kondo effect is discussed in sec .
[ sec - surface - kondo ] in the context of the schrieffer - wolf transformation and the slave - boson mean - field theory . using the calculated hybridization energies
the differential conductance is calculated in sec .
[ sec - stm ] and compared with the stm data .
[ sec - conclusions ] contains our conclusions .
in this section , we obtain wavefunctions needed to calculate the tunneling matrix elements ( [ tunneling ] ) .
this is the first step towards the stm differential conductance . as discussed in the introduction we are going to use simple models for the various potentials involved in the problem so that we can make analytical progress in calculating the hybridization energies .
metallic cu is a very good metal with a nearly spherical fermi surface except for the gaps in the ( 111 ) direction .
these gaps have their origin in the periodic potential generated by the ions ( see fig.[fs ] ) .
although the cu states can be obtained from first principle calculations @xcite we adopt the nearly free electron approximation and model the bulk potential with a simple periodic function .
furthermore , in order to study the surface states we consider a semi - infinite crystal where the bulk is located at @xmath57 ( see fig . [ fig - potential ] ) .
we write the potential in the substrate as : @xmath58 where @xmath59 is the bulk potential , @xmath17 is the lattice spacing in the direction of the surface , @xmath60 is the work function measured from the bottom of the conduction band , @xmath61 is the image plane , chosen in such a way that the image potential joins the constant potential @xmath62 continuously , and is related to @xmath63 as @xmath64 in the potential eq .
( [ potential ] ) , @xmath63 is the only free parameter .
the eigenfunctions and eigenenergies for the schrdinger equation associated with the potential ( [ potential ] ) can be calculated exactly and the details are given in appendix a. let us consider first the case of the ( 111 ) direction . in order to fit the experimental band structure@xcite along the ( 111 ) direction in the nearly - free - electron potential we find that @xmath65 ev , @xmath66 ev . these values provide a very good description of the band gap in the ( 111 ) direction and also of the ionization energy of metallic cu .
the cu(111 ) surface has a band of surface states as measured by photoemission , starting @xmath67 mev below the fermi surface@xcite .
by solving the schrdinger problem for the potential ( [ potential ] ) we find the surface band ( see appendix a ) and fix the value of @xmath68 in order to fit the photoemission results
. using this value for @xmath63 , we also obtain the next surface state at @xmath69 mev below the vacuum level .
this result agrees well with photoemission experiments that find the second surface state at @xmath70 mev below the vacuum level@xcite .
thus , we were able to get a consistent description of two independent experiments with a single fitting parameter . in fig .
[ fig - wavefunction ] we show a plot of the surface state wavefunction as a function of the distance away from the surface ( the wavefunction is chosen to be a real function ) .
notice the slow decay of this state inside the crystal in the direction perpendicular to the surface .
ev below the fermi surface .
the vertical axis represents its normalized wavefunction multiplied by the box size @xmath71 .
the horizontal axis represents @xmath72 in @xmath73.,width=302 ] in the 3-dimensional @xmath74-space of cu , a gaped cavity " is centered around the ( 111 ) direction near the fermi level and has the bulk band surrounding this gaped region @xcite . by decomposing the momentum of a bulk state @xmath75 into @xmath76 ( perpendicular to the surface ) and @xmath77 ( parallel to the surface ) , one expects that the larger @xmath78 , the larger wavefunction tail outside the crystal , which makes a significant difference on their contributions to the surface kondo effect . as a result
, it is important to carefully exclude the ( 111 ) gaped region from the fermi sphere when calculating the surface kondo effect .
it is known that a nearly - free - electron band has a parabolic dispersion not too close to the band edge , where band edge " refers to the interface between the bulk band and the gap in the 3d @xmath74-space .
when calculating the stm tunneling conductance , one would have to include the contribution close to the band edge as well .
however , the band - edge contribution is typically much smaller than the one coming from the parabolic band , so we can take merely the contribution from the parabolic band . for the cu(111 ) band ,
we terminate the parabolic band @xmath79 ev below the band edge .
as long as we work in a parabolic band , the bulk - state wavefunction inside the crystal takes a plane - wave form .
we now turn to the case of the ( 100 ) direction .
we find that we get good agreement with band structure calculations @xcite by choosing @xmath80 ev , and @xmath81 ev . in contrast to the ( 111 ) direction , the ( 100 ) direction has no true surface state near the fermi surface .
however , there is a surface resonance at @xmath82 ev above the fermi level @xcite that can be well described by choosing @xmath83 .
when calculating the tunneling conductance , we can safely sum over the entire fermi sphere and adopt a parabolic dispersion , i.e. a plane - wave wavefunction inside the crystal .
it is the d - orbital of the adatom that actually participates in the kondo resonance .
when one of these atoms is placed on a metal surface , its outmost s - wave electrons either get transfered to the metal conduction band or to its own d - orbital .
for example , the co adatom tends to form [ ar]@xmath84 on a cu surface@xcite and therefore has an effective spin @xmath85 . as a result ,
an adatom d - wavefunction can be modeled as a hydrogen atom with an effective charge @xmath86 given by : @xmath87 where @xmath88 are spherical harmonics and @xmath89 are hydrogen radial wavefunctions .
the effective charge @xmath90 can be determined by comparing @xmath91 , defined by , @xmath92 with the experimentally observed d - orbital radius for metallic co , @xmath93 . in our calculation
we orient the axis of quantization for the atomic problem along the direction perpendicular to the surface .
the stm tip wavefunction is modeled following tersoff and hamann@xcite who proposed a wavefunction of the form : @xmath94 where @xmath95 is the radius of curvature of the tip about its center , the decay constant @xmath96 controls the wavefunction tail , and @xmath97 is the tip work function .
notice that the adatom and tip wavefunctions are normalizable , that is , @xmath98 , when integrated over the entire space .
the set of three different wavefunctions ( substrate , adatom , and tip ) are the main elements required for the calculation of the hybridization energies in the anderson impurity hamiltonian .
using the results of eq.([surface - wavefunction ] ) , ( [ co ] ) and ( [ tip ] ) we can evaluate the overlap integrals in ( [ tunneling ] ) that are important for the hybridization energies . let us first remark that the normalization of each of these wavefunctions is different ; for instance , the bulk state is normalized in a box of size @xmath99 , where @xmath71 is one of the dimensions of the box . the surface state , however , since it decays exponentially away from the surface but is extended on the surface , is normalized in an area of size @xmath100 .
finally , the adatom and tip states always decay exponentially and therefore do not require any box normalization .
thus , it is convenient to redefine the wavefunctions so that : @xmath101 where the subscript @xmath102 refers to bulk and @xmath103 to surface .
note that in this case the hybridization energies between substrate ( bulk and surface ) and adatom in ( [ tunneling ] ) can be written as : @xmath104 where @xmath105 for bulk and @xmath106 for surface .
the matrix elements are , therefore , @xmath107 notice that @xmath108 is not only a function of the momentum and the distance @xmath109 between adatom and surface , but also depends on which surface - here labeled by the subscript @xmath110 - the adatom is located . in fig .
[ fig - v111 ] we show the result for @xmath111 , the hybridization matrix element between @xmath112 adatom state and surface state in the cu(111 ) surface , as obtained by numerical integration of ( [ vd ] ) using ( [ surface - wavefunction ] ) and ( [ co ] ) , as a function of @xmath109 ( in units of the distance between the planes in the ( 111 ) direction ) by assuming @xmath24 to be at the fermi surface of the surface band [ since the surface states have essentially a spherical fermi surface there is no dependence of @xmath113 with the direction on the cu(111 ) surface ] .
the first striking result is the strong oscillation of the value of the matrix element with @xmath109 .
this oscillation is the result of the interference between the d - state of the adatom and the surface state .
naturally , the largest overlap occurs at the surface ( where the surface state is maximum , see fig .
[ fig - wavefunction ] ) but we notice that there is substantial overlap between the surface state and the adatom at one lattice spacing from the surface . in principle , we can not tell what is the actual orientation of the adatom relative to the surface but we have checked that the @xmath112 state is the one with biggest overlap as compared with higher angular momentum states , thus , instead of a sum over the angular momentum states in eq.([hamiltonian ] ) we could have kept only the one with @xmath112 . furthermore ,
first principle calculations for adatoms in metallic surfaces indicate that the atomic orbitals are oriented in such a way to generate the largest overlaps @xcite . :
dashed - dotted line , @xmath114-@xmath115 : thick line , and @xmath114-@xmath116 : thin line , as a function of @xmath109.,width=302 ] in fig .
[ fig - v111 ] we also plot the value of @xmath117 , the hybridization matrix element between the bulk states and adatom in the cu(111 ) surface as a function of the distance from the surface for two different bulk states , all of the them at the cu fermi surface , with momenta along the @xmath44-x , and @xmath44-k directions ( see fig.[fs ] ) .
once again we notice the oscillations due to the interference between those states and the adatom states but we also observe that the overlap now is a factor of 3 smaller because these states do not have a large amplitude at the surface . another interesting result of these calculations is the fact that as the momenta moves towards the ( 111 ) direction the overlap grows .
this is due to the fact that away from the ( 111 ) direction the fermi surface of cu is essentially spherical and therefore the states at the fermi surface have essentially the same amplitude , however , the states with large momentum along the direction of the cu(111 ) surface have bigger penetration on the potential barrier generated by the work function and therefore larger hybridization with the adatom .
this effect will be very important later for the interpretation of the experimental data . in fig .
[ fig - v100 ] we show the value of @xmath118 for the overlap between the state of an adatom on the cu(100 ) surface with bulk states with momenta along the @xmath44-x ( this direction is perpendicular to the cu(100 ) surface ) , and @xmath44-k ( since there are no surface states for the cu(100 ) surface we do not need to calculate @xmath119 for this surface ) .
besides the oscillations found in the previous calculations , we find that when the momentum is oriented perpendicular to the surface its penetration is maximum and therefore has largest overlap .
in fact , we find that the overlap between the adatom state on a cu(111 ) surface with the surface state is essentially the same value as the overlap between the state of the adatom on the cu(100 ) surface with a bulk state with momentum oriented in the ( 100 ) direction
. one can easily understand this effect from the nearly free electron picture : in the absence of the periodic potential of the crystal , the fermi surface states in the ( 111 ) and ( 100 ) direction would be essentially identical and their decay in the vacuum of their respective surfaces would be the same .
when the periodic potential of the crystal is added a gap opens in the ( 111 ) direction but not on the ( 100 ) direction .
however , the bulk state that disappeared at the fermi surface in the ( 111 ) direction now becomes the surface state in this same surface .
therefore , its amplitude at the surface and close to it is essentially the same as the state in the ( 100 ) direction ( this is can be readily understood due to the long decay rate of the surface state shown in fig .
[ fig - wavefunction ] ) and the overlap is essentially the same .
thus , one has to be careful when discussing what happens to adatoms in different surfaces because there is very little difference between surface states along gaped directions and bulk states along non - gaped directions in the fermi surface of noble metals .
-@xmath116 : thin line , @xmath120 : thick line , as a function of @xmath109.,width=302 ]
there are two main limits of interest in the kondo problem . the first one is reached when @xmath121 and we can treat the hybridization energy as the smallest scale in the problem .
this is the so - called kondo regime where the average occupation of the impurity state is essentially constant , that is , @xmath122 . in order to study this limit
it is convenient to split the electron states into bulk ( b ) and surface ( s ) so that we write the anderson impurity hamiltonian ( [ hamiltonian ] ) as : @xmath123 where @xmath124 ( @xmath125 ) describes the bulk ( surface ) electron annihilation operator , respectively .
we have a two - band problem where both bulk and surface states hybridize with the impurity at the surface .
notice that this is a single channel problem since only one effective channel couples to the impurity . in the kondo limit
we can use a schrieffer - wolf transformation @xcite and rewrite the anderson impurity hamiltonian as a exchange problem : @xmath126 where @xmath36 is the localized impurity spin and @xmath127 is the exchange energy matrix ( we have defined @xmath128 and @xmath129 as the dimensionalities of bulk and surface states , respectively ) , and @xmath130 notice that , unlike the single band kondo effect ( [ kondo_hamiltonian_1 ] ) , electrons from one band can scatter from the impurity between two different bands with spin flip ( all energies are measured relative to the fermi energy ) .
if we ignore the off - diagonal matrix elements of ( [ exchanges ] ) the problem trivially reduces to two independent kondo effects with characteristic kondo temperatures @xmath131 where @xmath132 is the bandwidth cut - off ( @xmath133 ev , @xmath134 ev @xcite ) and @xmath135 is the effective coupling constant , @xmath136 is the fermi surface density of states ( for spherical fermi surface we have @xmath137 where @xmath138 is the effective mass ) . because of
the difference in the kondo temperatures given in ( [ tk1 ] ) one of the scattering channels dominates the screening process and the singlet state can be either formed with the surface state when @xmath139 or with the bulk states if @xmath140 .
if we apply this argument to the results of the overlap integrals of the previous section we see that for the adatom in the cu(111 ) surface the kondo screening is dominated by the surface state since the coupling constant ( [ ln ] ) is bigger for this state than the bulk states . on the other hand , if the adatom is sitting at the cu(100 ) surface , the kondo effect is dominated by the bulk state with momentum perpendicular to the cu(100 ) surface .
the above discussion is valid as far as the occupation of the impurity level is close to @xmath141 .
however , in the problem at hand the adatom occupation can change substantially with the hybridization since the matrix element is strongly dependent of the distance to the surface and the type of states involved , that is , @xmath27 can be of the order of @xmath142 ( but still much smaller than @xmath26 ) . in this case , the mapping into the kondo problem fails to properly describe the actual occupation on the adatom .
instead , we should start from the anderson impurity hamiltonian ( [ hamiltonian ] ) and compute the kondo temperature directly from it . here
we will do so by adopting the large @xmath143 calculation of newns and read that allows for average fluctuations in the valence of the adatom in the limit of @xmath144 ( in fact , estimates based on lsda@xcite give a value of @xmath145 ev and @xmath146 ev ) .
the idea of the large @xmath143 calculation is to generalize ( [ hamiltonian ] ) to include more degenerate electron states ( besides the @xmath147 states associated with the spin @xmath85 ) and take the limit where the degeneracy diverges .
mathematically this is accomplished by generalizing the spin index in the electron operator to a new index @xmath148 that varies from @xmath141 to @xmath143 .
it is convenient to write anderson impurity hamiltonian ( [ hamiltonian ] ) in the limit of @xmath149 as : @xmath150 where @xmath151 is the projector operator that projects out any states of the adatom with double occupancy . in the adatom we can , in principle ,
have three different states with zero , @xmath152 , single , @xmath153 , and double , @xmath154 , occupancy .
we now expand the hilbert space by introducing a `` phantom '' boson state such that when the state of the adatom is unoccupied by an electron it is occupied by a boson , say @xmath155 , and when it is unoccupied by a boson it is occupied by an electron , @xmath156 .
observe that this construction constrains the number of bosons and electrons to be @xmath141 for all states @xmath157 in the hilbert state : @xmath158 the utility of this representation becomes clear when we consider the action of the operator @xmath159 on the state @xmath160 producing the state @xmath156 .
further application of @xmath159 to @xmath156 annihilates the state while application of @xmath161 to @xmath156 returns the system to the state @xmath160 .
thus , the repetitive application of @xmath162 or @xmath161 never produces the doubly occupied state . in other words
, we can replace @xmath163 in ( [ largen ] ) by a boson operator if we ensure that the constraint ( [ constraint ] ) is obeyed .
this constraint can be seen as an energy cost , that is , states of the hamiltonian that violate the constraint ( [ constraint ] ) are highly excited states
. if we are interested in the low energy physics then we can convert ( [ constraint ] ) into a new energy scale and add it to the hamiltonian as : @xmath164 where @xmath165 is a lagrange multiplier .
notice that in the limit of @xmath166 the constraint ( [ constraint ] ) is automatically obeyed since the energy is minimized when ( [ constraint ] ) is fulfilled . as is , hamiltonian ( [ largen_1 ] ) is still very complicated and at this point we have to use the large @xmath143 approximation that consists in treating ( [ largen_1 ] ) in a mean - field approximation . in the mean - field limit
we consider the average value of the boson fields instead of their actual value , that is , we replace the operators @xmath167 and @xmath168 by @xmath169 notice that by treating the bosons in the mean - field approximation we obtain a quadratic hamiltonian in ( [ largen_1 ] ) that can be diagonalized by standard techniques and the total energy of the system can be minimized with respect to @xmath72 and @xmath165 .
the physical meaning of these parameters becomes clear when we calculate physical observables .
for instance , the average occupation of the impurity is given by : @xmath170 that shows that @xmath72 has to do with the change in the occupancy of the adatom state due to the hybridization with the substrate .
on the other hand we can also show that @xmath165 is the renormalization of the energy of the impurity state , that is , the renormalized impurity state energy is given by : @xmath171 .
the mean - field parameters @xmath172 and @xmath72 are calculated by solving a set of coupled transcendental equations : @xmath173^{2}\right\}^{1/2 } } { d}\right),\end{aligned}\ ] ] where @xmath174 are matrices defined , in general , as : @xmath175 and @xmath176 is bandwidth of the conduction band ( notice that @xmath177 in reference to the adatom and tip , respectively ) .
the kondo temperature can be directly extracted from this formalism as @xmath178 where @xmath179 therefore , using the large @xmath143 account we can deal with the case where the hybridization energies are of the order of the adatom energies .
we use these results in studying the stm differential conductance in the next section . , on the cu(111 )
surface.,width=302 ] we are not aware of any measurement of co - to - cu distance being reported , so this distance @xmath180 will be treated as a parameter of the order of the cu lattice space .
[ fig - kondo - t111 ] and [ fig - kondo - t100 ] shows our calculated kondo temperature as a function of @xmath180 .
an interesting feature is that the kondo temperature decays fast with @xmath180 , which can be understood because of the exponential overlap between the substrate and co wavefunctions . when calculating the kondo temperature for co / cu(111 ) assuming that only the bulk state exists , we find @xmath181 for the @xmath180 s as fig . [ fig - kondo - t111 ] , which clearly means that the surface state dominates the kondo effect .
the domination is not surprising since a significant portion of large-@xmath78 bulk states are gaped around the ( 111 ) direction . comparing the cu fermi surface with a perfectly spherical one
, one finds that the cu surface state contributes to the kondo effect in a way similar to the bulk band of the spherical fermi surface around the ( 111 ) direction .
similarly , the bulk band with respect to the cu(100 ) surface can be considered as a perfectly spherical fermi surface for calculating the kondo effect . as a result
, the kondo effect of the co / cu(111 ) , dominated by the surface state , should have the same order of magnitude as co / cu(100 ) .
the positions that provide the experimentally measured kondo temperatures @xcite ( @xmath182k and @xmath183k ) are @xmath184 and @xmath185 from the ( 111 ) and ( 100 ) surfaces , respectively .
hence the stm conductance will be calculated by fixing co atom at these optimized positions .
we should point out that from an experimental point of view we would expect the co adatom to be located around one lattice spacing away from the surface .
thus , our obtained values of their position seem to be quite small ( especially in the case of the ( 100 ) surface where the adatom has to be located at the surface ) and therefore unrealistic . nevertheless , one has to have in mind that the there are many assumptions made in the calculation that can affect significantly the value of the kondo temperature , namely , the form of the image potential close to the surface , the large @xmath143 approximation , and the limit of @xmath186 .
furthermore , it is clear from ( [ tk2 ] ) , that the kondo temperature is exponentially sensitive on the details of the hybridization and energy scales and therefore is significantly affected by the various assumptions .
thus , the plots in fig.[fig - kondo - t111 ] and fig.[fig - kondo - t100 ] have to taken as the trends for the variation of the kondo temperature with the distance from the surface and not as realistic values .
, on the cu(100 ) surface.,width=302 ]
we assume that the stm tip and the adsorbate - substrate complex are each in local equilibrium , and use the standard green s function formalism for calculation of the differential conductance @xcite : @xmath187^{2}\right\}_{\omega = ev } , \label{current}\ ] ] where @xmath174 is defined in ( [ f - green ] ) and @xmath188 is formally the adsorbate retarded green s function associated with the total hamiltonian @xmath189 .
however , because the effect on the adsorbate from the tip is much smaller than that from the substrate , it is always safe to regard @xmath188 as the one without tip . in this work , @xmath188 ( without tip ) is calculated by the mean - field slave - boson technique as : @xmath190 where the symbols here are given in the previous section . by substituting ( [ ga ] ) into ( [ current ] )
, the stm differential conductance can be written as : @xmath191 \ , , \label{didv}\ ] ] where the differential conductance with a subscript 0 " refers to the background signal , @xmath192 is the amplitude of the stm conductance , @xmath193 @xmath194 is the fano lineshape parameter , @xmath195 and @xmath196 is the dimensionless bias voltage @xmath197 expression ( [ didv ] ) has been frequently used for the fitting of stm data . ,
as a function of the bias @xmath9 in mev for an adatom on the cu(111 ) ( thick line ) and cu(100 ) ( thin line ) surfaces , as obtained from the theoretical calculations.,width=302 ] the shape of the differential conductance versus the bias voltage is determined by the fano line shape parameter @xmath194 . in fig.[didv ] we show the differential conductances obtained from our microscopic theory for the cu(111 ) and cu(100 ) surfaces . from eq .
( [ fano ] ) we see that @xmath194 is related to three microscopic quantities @xmath198 , @xmath199 , and @xmath200 .
it is found that @xmath201 and therefore we can set @xmath202 in ( [ fano ] ) . obtaining the first two is straightforward since they only depend on the physics _ at _ the fermi energy . from ( [ f - green ] ) , we have : @xmath203 notice that the amplitude of the stm signal , ( [ amp ] ) , depends essentially on this quantity and can be calculated reliably . in fig.[amplitude ] we plot @xmath204 for @xmath205 and compare it with the experimental data @xcite . as one can see , the agreement between the theory and experiment and theory is quite good . as compared with the experimental data @xcite.,width=302 ] the other quantity of interest , namely , @xmath200 , requires an integration over the entire cu band . for the cu(111 ) surface ,
the surface band has a free - electron dispersion starting @xmath67 mev below the fermi level up to @xmath206 mev above it , where the upper limit is the intercept of the surface band to the bulk bands .
we calculate @xmath200 for this surface band and determine the tunneling conductances , ( [ didv ] ) , for different values of @xmath207 as shown in shown in fig . [ didv_r ] . .
from bottom to the top : @xmath208 , width=302 ] we notice that for short distances away from the adatom , the theory can describe very well the lineshape of the tunneling conductance .
nevertheless , as the distance of the stm tip to the adatom is increased , the lineshape becomes more symmetric and a positive lorentzian is developed ( see fig.[didv_r ] ) .
the origin of this result can be easily understood : notice that in fig.[amplitude ] the amplitude of the stm conductance becomes very small for large values of @xmath207 .
this smallness can be tracked down to the small value of @xmath209 in ( [ amp ] ) indicating that the interference between tip and adatom is fading away .
however , one sees from ( [ fano ] ) that this indicates that the fano parameter @xmath194 is becoming very large and positive given rise to a positive lorentzian .
this feature is general of any theory of the stm conductance that starts with nearly free electron wavefunctions and can not be avoided since it has its origins on the oscillating nature of the surface states density of states @xcite .
this result is at odds with the experiment when the distance between tip and adatom becomes large . instead of a positive lorentzian
, it is observed a negative lorentzian @xcite . in fig.[fig - q ]
we plot the calculated values of @xmath194 against the values used by the experimentalists in order to fit the data @xcite .
the discrepancy between theory and experiment is clear .
more work will be required in order to check this issue . for co on cu(111 ) as compared with the values obtained by the experimental fit @xcite.,width=302 ] in order to calculate @xmath200 for an adatom located at the cu(100 ) surface we need information of the whole cu bulk band .
the cu bulk bands start at approximately @xmath210 ev below the fermi level and disperse up to the vacuum level that is located at approximately at @xmath211 ev above the fermi energy . at the fermi energy the cu band has dominant s - wave symmetry which gives the quasi - spherical shape of the fermi sea
. however , there are also cu d - bands located from @xmath147 to @xmath212 ev below the fermi level with very large density of states @xcite . using the nearly free electron approximation ( s - wave ) with a cut - off at the vacuum level we have calculated @xmath200 and have found that it yields a value of @xmath213 that changes sign from positive to negative as @xmath207 increases , which is not the case observed in the experiment . in principle the quantity @xmath200 must be calculated including the d - bands . using a simple parabolic dispersion for the d - bands we find their contribution to be less than 0.1% of the total value of @xmath200 and therefore these bands are not able to fix the problem of the sign of @xmath213 .
thus , we have problems with the use of bulk states in our calculation .
we believe that band structure effects are important for the exact calculation of the surface kondo effect due to bulk states .
as far as we know there are no such detailed calculations . as compared with the experimental data @xcite.,width=302 ] in order to remedy this problem we introduce a cut - off in the problem such that by integrating bulk states starting from the bottom of the nearly free electron band we stop the integration when the value of @xmath200 fits the experimental curve at @xmath214 . we find that if we integrate up to @xmath215 ev above the fermi level we can quantitatively explain the data , as seen in fig . [ didv ] . given the limitations of our theory this is essentially equivalent to consider @xmath200 for adatom on the cu(100 ) surface for bulk states as a free parameter of the problem .
more accurate calculations , beyond the ones presented in this paper , are required in order to calculate @xmath200 from a microscopic theory . .
from bottom to the top : @xmath216 , width=302 ] in fig . [ a100 ]
we show the amplitude @xmath204 for co on cu(100 ) for @xmath217 .
there is a qualitative agreement between theory and the experiment . in fig .
[ didv_100 ] we show the evolution of the tunneling conductance as a function of the distance from the adatom .
once again , the lineshape is correct for short distances but for long distances the lineshape becomes a positive lorentzian , in contrast with the experiment .
in fact , in fig . [ q100 ]
we show the dependence of the fano parameter , @xmath194 , as a function of the lateral distance , @xmath207 .
notice that overall result is better than the case of co on cu(111 ) ( fig .
[ fig - q ] ) but the divergence in the value of @xmath194 is very clear . for co on cu(100 ) as compared with the values obtained by the experimental fit @xcite.,width=302 ]
in this paper we have studied the microscopic theory of the single impurity surface kondo resonance starting from an anderson impurity hamiltonian .
we have calculated the hybridization energies and wavefunctions of the model adopting simple models for the crystal and surface potentials in the nearly - free electron approximation .
our model reproduces the main features of the band structure of cu , the energy of the surface states and surface resonances as measured in photoemission experiments .
we have shown that the hybridization energies ( or matrix elements ) , calculated from the solution of the single particle schrdinger equation , are strongly dependent on the direction of the electron momenta , the distance of the adatom from the surface , and the actual surface where the magnetic impurity is located .
the hybridization matrix elements oscillate with distance from the surface due to the interference between the adatom localized state and the substrate states ( surface and bulk ) that oscillate close to the surface .
we have shown that the surface kondo effect occurs preferentially with the surface state ( when it exists ) or with the bulk state at the fermi energy that has the largest component of its momentum perpendicular to the surface .
we have demonstrated that although the surface kondo effect can be obtained from the anderson impurity hamiltonian by a schrieffer - wolf transformation onto the kondo hamiltonian , the hybridization energies can be comparable with the adatom atomic energy requiring a more sophisticated approach that allows for fluctuations in the mean occupation of the adatom ( in other words , the adatom is in what is called a mixed - valence state ) .
we have used the mean - field slave - boson formalism that allows for the calculation of the kondo temperature in the limit when the on - site coulomb energy @xmath26 is much larger than the hybridization energy .
our results produce reasonable values of the kondo temperature for these systems and the distance of the adatom from the surface .
we have calculated the tunneling conductance of a stm tip for different surfaces in the presence of a co impurity and found that for the surface state in the cu(111 ) surface we get quantitative agreement without any further adjustable parameters .
however , for the adatom on the cu(100 ) surface where the kondo effect is dominated by the bulk state with momentum along the ( 100 ) direction , we have to add an extra cut - off energy in order to fit the data .
more detailed band structure calculations are required in order to eliminate the need for this extra parameter .
we have shown that in both cases we can fit very well the data obtained in stm measurements at @xmath214 but there is a discrepancy between the theory and experiments when the distance between the tip and the adatom becomes large . while in the experimental paper where the measurements are reported it is conjectured that the kondo effect in the cu(111 ) and cu(100 ) surfaces are dominated by the bulk states@xcite , we show that in fact , the kondo effect is dominated by the surface state in the cu(111 ) surface and by the bulk state with momentum in the ( 100 ) direction in the ( 100 ) surface .
we show that this occurs because there is very little difference between surface states and bulk states close to the surface in a noble metal and both states have maximum penetration on the work function potential . in summary , we provide a microscopic interpretation for the stm experiments of surface kondo effects .
our results can be used as basis for the interpretation of more complicated situations where , for instance , co adatoms interact via direct or indirect exchange on artificial anderson lattices in metallic surfaces .
* acknowledgments * we would like to thank i. affleck , p. w. anderson , a. balatsky , p. coleman , d. eigler , a. heinrich , p. johnson , c. lutz , v. madhavan , p. simon , and a. tsvelik , for illuminating discussions .
we also would like to acknowledge support under darpa contract no .
daad19 - 01-c-0060 .
for the cu(111 ) surface state , the parallel momentum @xmath218 is a good quantum number , and the wavefunction can be solved from potential eq .
( [ potential ] ) as : @xmath219 where @xmath220 is the standard 2nd solution of confluent hypergeometric equation@xcite , @xmath221 , and @xmath222 . inside the crystal one has a phase shift @xmath223 , a crystal decay factor @xmath224 , and the energy of the surface band bottom @xmath225 related by @xmath226 where @xmath227 is the bulk - band effective mass .
the plane wavefunction has a wavevector @xmath228^{1/2 } \
, , \ ] ] while vacuum decay factor @xmath229 is defined as @xmath230^{1/2 } \ , .\ ] ] the energy @xmath225 is obtained from the continuity conditions at @xmath231 and @xmath232 . as pointed out in ref.@xcite ,
the radius of the d - orbital , @xmath233 , is much smaller than the average interparticle distance , @xmath234 ( where @xmath235 is the fermi momentum ) : @xmath236 .
hence , the corresponding overlap integrals between the adatom and the substrate are small . | we develop a microscopic theory of the single impurity kondo effect on a metallic surface .
we calculate the hybridization energies for the anderson hamiltonian of a magnetic impurity interacting with surface and bulk states and show that , contrary to the kondo effect of an impurity in the bulk , the hybridization matrix elements are strongly dependent on the momentum around the fermi surface . furthermore , by calculating the tunneling conductance of a scanning tunneling microscope ( stm ) , we show that when the magnetic impurity is located at a surface the kondo effect can occur with equal strength between bulk and surface states .
we compare our results with recent experiments of co impurities in cu(111 ) and cu(100 ) surfaces and find good quantitative agreement . |
in this paper , we analyse a stochastic generalisation of the abelian sandpile model ( asm ) .
informally ( we provide a formal definition later ) , the asm operates on a graph where each vertex has a number of ` grains of sand ' on it . at every unit of time
, another grain of sand is added at a random vertex @xmath5 . if this causes the number of grains at @xmath5 to exceed its degree , @xmath5 topples , sending one grain to each of its neighbours .
this may cause other vertices to topple , and we continue until the configuration is stable , i.e. no vertex can topple anymore .
a special vertex , the sink , can absorb any number of grains and never topples .
it is possible to show that eventually this model will be trapped in a set of configurations , called recurrent configurations .
this model arose from work by bak , tang and wiesenfeld @xcite and was named and formalised by dhar @xcite .
it displays a phenomenon known as self - organised criticality @xcite , in which characteristic length or time scales break down in the ` critical ' steady state .
when this happens , the correlation between the number of grains at two vertices obeys a power - law decay , as opposed to an exponential decay often found in models away from criticality .
likewise , the average number of topplings that result from a single grain addition also obeys a power - law distribution . in this sense ,
the model is ` non - local ' , as grains added at a vertex may have an effect on vertices that are far away .
physically , this model ( and self - organised criticality in general ) has been used in applications as wide as forest fires @xcite , earthquakes @xcite and sediment deposits @xcite .
mathematically , the asm has been heavily studied , and we shall not list out all the references here .
we refer interested readers to dhar s papers @xcite and to the excellent review on the subject by redig @xcite . some relevant results to our current work discussed in @xcite include : * the number of recurrent configurations is equal to the number of spanning trees of the graph .
* there is an algorithm ( called the _ burning algorithm _ ) which determines if a given configuration is recurrent or not .
this finds , or establishes the non - existence of , a subgraph not including the sink on which the configuration is stable .
this algorithm constructively establishes a bijection between recurrent configurations and spanning trees . * in the steady state of the model , each recurrent configuration is equally likely . in the asm
, the only randomness occurs in the vertices that we add grains to .
we introduce a variation on this model , where the topplings themselves are also random .
more precisely , we fix a probability @xmath1 $ ] , and when a site is unstable , each neighbour independently has a probability @xmath6 of receiving a grain from the unstable site . in this way
, an unstable site may remain unstable after toppling but , as in the original model , the process continues until the configuration is stable . if @xmath2 , this is identical to the asm .
although this new model appears similar to the asm , a closer inspection reveals some qualitative differences . in particular , the model will again become trapped in a set of recurrent configurations , but this set is not equal to the set of recurrent configurations in the asm .
furthermore , each recurrent configuration is not equally likely , and the steady state measure now depends on @xmath6 .
the aim of this paper is to study the behaviour of this new model , particularly in these respects .
we prove a characterisation of the recurrent configurations in terms of orientations of the graph edges , and provide a tutte polynomial - like " formula which counts these configurations in terms of their numbers of grains of sand .
the stochastic sandpile model ( ssm ) we propose is an appropriate generalisation of the asm , due to the aforementioned relation between the tutte polynomial and a counting of the recurrent configurations .
a similar result has been proved by lpez @xcite for the asm ; in this model , the lacking polynomial we define later has an expression using the `` standard '' tutte polynomial if the number of lacking particles is @xmath7 . ] .
see also cori and le borgne @xcite and bernardi @xcite for combinatorial explanations of this fact .
although several instances of random sandpile models have already been introduced in the literature ( see for example dhar ( * ? ? ? * section 6 ) , kloster _
@xcite , manna @xcite ) , as far as we are aware none of these models have any known link with tutte - like polynomials .
variations on the basic sandpile model also include an asymmetric version @xcite , which devised an analogous algorithm to the burning algorithm called the _ script algorithm_. in section [ sec : model ] , we formally define the asm and introduce our stochastic generalisation of it ( the ssm ) . in section [ sec : result ] , we present our results on the ssm , giving detailed proofs in section [ sec : proof ] .
finally , we offer a brief conclusion in section [ sec : conclusion ] .
in this section , we formally define our model and its associated notation . we start with a brief review of the established abelian sandpile model .
we first define the class of graphs @xmath8 that underly the model .
@xmath0 must be finite , unoriented , connected and loop - free .
it may have multiple edges , and it contains a distinguished vertex @xmath9 that we call the _ sink_. the set of these graphs is denoted @xmath10 .
we use the notation @xmath11 to denote that @xmath12 and @xmath5 are adjacent in @xmath0 , i.e. @xmath13 . a sandpile _ configuration _ on @xmath0 is a vector @xmath14 .
the number @xmath15 represents the number of grains of sand present at the vertex @xmath5 in the configuration @xmath16 .
when this number exceeds a certain threshold , the vertex is said to be _ unstable _ , and will _
topple _ , sending one grain of sand to each of its neighbours . typically and throughout this paper , the threshold
is set to the degree of that vertex .
the sink plays a special role in that it can absorb any number of grains , and as such never topples .
two different configurations play an important role in this paper .
the first one @xmath17 ( or @xmath18 when @xmath0 is clear from the context ) is the configuration ^_g:=(d^g(v ) , v v ) , where @xmath19 is the degree of the vertex @xmath5 in @xmath0 .
the second one is @xmath20 for some vertex @xmath21 in @xmath22 , _
a:=(_a , v , v v ) , where @xmath23 is the kronecker symbol , meaning that all co - ordinates of @xmath20 are 0 except for at position @xmath21 , where it is 1 . a configuration @xmath24 is called _ stable _ if @xmath25 for all @xmath26 .
we write @xmath27 for the set of all stable configurations on @xmath0 .
@xmath17 is clearly the maximum stable configuration in terms of the total number of grains in the configuration .
now define the toppling operator @xmath28 corresponding to a toppling at @xmath29 by [ eq : full - toppling ] t_x()=- d^g(x)_x+_y~x , ys _ y , where configurations are added site - by - site . a toppling @xmath28 is called _ legal _ if @xmath30 .
we now define for any configuration @xmath16 its _ stabilisation _ @xmath31 as follows .
if @xmath16 is stable , then @xmath32 ; otherwise ( ) = t_x_m( ... (t_x_1 ( ) ) ... ) , with the requirements that @xmath33 is a sequence of legal topplings , and that the configuration @xmath31 is stable .
the fact that this operator is well - defined , i.e. the stabilisation of a configuration is independent of the order of the topplings that is used to stabilise it , is not immediately obvious .
we refer the interested reader to proposition 3.7 ( and lemma 3.6 ) in redig @xcite .
the asm has a markov chain structure which is defined as follows .
assume that there are defined some i.i.d .
random variables @xmath34 taking their values in @xmath22 according to a distribution @xmath35 ( where the support of @xmath35 is @xmath22 ) , defined on a common probability space @xmath36 .
the sandpile process starts from any stable configuration @xmath37 .
we define a markov chain @xmath38 with values in @xmath27 . given @xmath39 for any @xmath40 , @xmath41 is obtained as follows : * add a grain at position @xmath42 to the configuration @xmath39 .
let @xmath43 be the obtained configuration .
* let @xmath41 be the stabilisation of @xmath44 , that is @xmath45 ( in some cases no toppling is needed ) .
since a grain can be added anywhere with positive probability , it is immediately seen that from any configuration @xmath16 , the maximal configuration @xmath46 can be reached with positive probability .
it follows that the set of recurrent configurations for the markov chain is the unique recurrent class containing the maximal configuration .
we denote this set by @xmath47 , and as previously noted , it is in bijection with the set of spanning trees of @xmath0 ( see redig @xcite ) . in this model
, we make the topplings random .
a probability @xmath1 $ ] is fixed , and when a site is unstable , each neighbour independently has a probability @xmath6 of receiving a grain from the unstable site . in this way ,
an unstable site may remain unstable after toppling , but as in the original model , the process continues until the configuration is stable .
different topplings are done independently . if @xmath2 , this reduces to the asm ,
so we assume this is not the case .
there are some slight difficulties in establishing that this model is indeed well - defined .
two of them are taken into account in the proof of theorem [ thm : well - defi ] below .
a third one arises by noticing that topplings commute in the asm .
however , in the ssm , when two legal topplings are possible , it is not clear that they commute , since we only talk about what to do in probability " . to solve this
, we describe a probability space where the laws of the topplings are indeed those discussed at the beginning of this section and for which the commutation of legal topplings takes place as needed .
we now give a formal definition of the ssm .
the definition of terms introduced in relation to the asm , apart from the toppling operator , are unchanged .
all the random variables discussed below are defined on a common probability space @xmath36 , and they are all independent . * the variables @xmath48 are distributed according to a distribution @xmath35 with support equal to @xmath22 . as before ,
they represent a sequence of arrival places of grains . * for any @xmath49 such that @xmath50 and any @xmath51 , @xmath52 is a bernoulli random variable with parameter @xmath6 .
it represents the number of particles going from @xmath53 to @xmath54 along @xmath55 due to the @xmath56th toppling of @xmath53 .
we now define @xmath57 as the @xmath56th toppling at @xmath29 ^(i)_x()=- _ e xb_i(x , e ) _ x + _ y s , e=\{x , y } eb_i(x , e)_y
. we will call any realisation of such a toppling a _
stochastic toppling_. note that with positive probability , the @xmath56th toppling at @xmath5 is a _ full toppling , meaning that @xmath58 , in which case @xmath57 coincides with the asm toppling @xmath28 .
_ on the probability space @xmath59 , the ssm is well defined when a starting configuration is specified .
we then use the same definition of a legal toppling , and define the stabilisation of a configuration @xmath16 by ( ) = _
x_n( ... (_x_1 ( ) ) ... ) , where we take @xmath60 if the vertex @xmath61 appears exactly @xmath62 times in the sequence of topplings @xmath63 , and with the same requirements as for the asm , i.e. that @xmath64 is a legal sequence of topplings and @xmath65 is stable .
the notation @xmath66 stands for the `` random stabilisation '' we define .
[ thm : well - defi ] for any graph @xmath0 in @xmath10 , the stabilisation operator @xmath66 is almost surely well - defined .
two things have to be shown .
the first one is that the stabilisation process eventually ends . here , as opposed to the asm
, a loop may appear in the toppling process .
for example ,
if @xmath67 and @xmath21 and @xmath68 are both unstable , and for all @xmath56 we have @xmath69 and @xmath70 , then a repeated loop occurs where @xmath21 only sends one grain to @xmath68 and @xmath68 only sends one grain to @xmath21 , _
ad infinitum_. this example is obviously contrived , and it is easy to see that this occurs with probability 0 , but some more complex loops could occur . to show that this almost surely does not happen , we note that in the asm , stabilisation takes a finite number of topplings regardless of the initial configuration .
let @xmath71 be the maximum number of full topplings at @xmath5 needed ( by the asm ) to stabilise any configuration arising from the addition of 1 grain to a stable configuration ; let @xmath72 .
since there are a finite number of such configurations , @xmath73 is finite .
now for any @xmath56 and @xmath53 , @xmath57 is the `` full toppling '' @xmath28 with positive probability .
a simple renewal argument shows that in an infinite sequence of topplings , the number of full topplings done at every vertex is almost surely larger than @xmath73 . from here
the conclusion follows ( note that this does nt mean that full topplings are needed to stabilise a configuration ) .
the second thing we have to prove is that that the stabilisation operator is well - defined ; that is , starting from an unstable configuration @xmath16 , changing the order of the site stabilisations does not change the final result @xmath65 .
this is the case on @xmath36 since whatever is done elsewhere , the @xmath56th toppling at @xmath5 is @xmath74 .
we can then use the same argument as in lemma 3.6 and proposition 3.7 of redig @xcite , with a simple adjustment to take into account the fact that @xmath75 is not constant .
the ssm has a markov chain structure which is analogous to the asm .
the sandpile process starts from any stable configuration @xmath37 .
we define a markov chain @xmath38 with values in @xmath27 . given @xmath39 for any @xmath40 , we obtain @xmath41 as follows : * add a grain at position @xmath42 to the configuration @xmath39 . let @xmath43 be the resulting configuration .
* let @xmath41 be the stabilisation of @xmath44 , that is @xmath76 ( in some cases no toppling is needed ) .
note that @xmath66 is not defined _ ex nihilo _ as in the asm .
on @xmath36 , when @xmath66 is applied , its action depends on all the previous topplings taken to reach the current state . in view of this
, it is more proper to write @xmath77 or even @xmath78 instead of @xmath66 .
however , for the sake of brevity we write @xmath66 instead . once again , since the support of @xmath35 is @xmath22 , the maximal configuration @xmath46 is recurrent , and the set of recurrent configurations for the markov chain is the unique recurrent class containing @xmath46 .
we denote this set by @xmath79 , and call any element of it _ stochastically recurrent _ ( or _ sr _ ) .
we call the configurations which are recurrent under the asm _ deterministically recurrent _ or _ dr _ in contrast . [ ds ] for any @xmath80 , @xmath81 this is easily seen because @xmath82 always uses a finite number of topplings . when constructing a stochastic stabilisation ,
the deterministic stabilisation must have a positive probability of being reconstructed using only full topplings .
the two sets @xmath47 and @xmath79 are not equal in general .
a simple counter - example is shown in figure [ fig : triangle ] .
for this graph , the configuration which has @xmath83 grains at each vertex is sr .
it can be reached from the dr configuration @xmath84 $ ] ( with 3 grains at the vertex @xmath85 leading to the sink ) , by adding a grain to @xmath85 and toppling it , sending one grain to the sink and one to @xmath86 .
however , it is not dr as it fails the burning algorithm test the graph with the sink removed is a forbidden subconfiguration . to determine @xmath79 ,
it is clear that @xmath16 is sr if and only if there exists a finite sequence of adding of grains and topplings such that @xmath16 is reached from @xmath46 through this sequence . in the rest of the paper ,
we prove a more useful characterisation of the stochastically recurrent states .
in this section , we state the main results of this paper .
our two main results are a characterisation of the stochastically recurrent states in terms of graph orientations and a recurrence for the lacking polynomial ( which we define below ) .
proofs of these results will be given in section [ sec : proof ] .
our first result characterises the sr states in terms of graph orientations . take a graph @xmath87 .
we define an orientation on @xmath0 to be an orientation of each edge of @xmath88 ( when @xmath0 has multiple edges , all of them are oriented independently ) .
we write @xmath89 or @xmath90 to denote that the edge @xmath91 is oriented from @xmath21 to @xmath68 .
[ comp_or ] let @xmath87 .
take a sandpile configuration @xmath16 on @xmath0 .
we define the _ lacking number _ of @xmath16 at @xmath5 as the number of grains at @xmath5 less than its maximum value : @xmath92 now let @xmath93 be an orientation on @xmath0 and let @xmath94 be the number of incoming edges to @xmath5 in @xmath93 .
we say that @xmath16 is _ compatible _ with @xmath93 ( and likewise @xmath93 is compatible with @xmath16 ) if @xmath95 , [ comp_cond ] ^g_o(v ) 1 + l^g_(v ) .
we denote the set of stable configurations that are compatible with @xmath93 as @xmath96 . in situations where it is clear
, we will omit the superscript @xmath0 for brevity .
note that there may be several configurations compatible with a particular orientation .
likewise , there may be several orientations compatible with any given configuration .
for instance , the maximal configuration @xmath46 is compatible with any orientation where each vertex has at least one incoming edge .
[ car_sr ] let @xmath87 . then a ( stable ) configuration @xmath16 is stochastically recurrent if and only if there exists an orientation @xmath93 on @xmath0 such that @xmath97 .
in other words , ( g ) = _ o ( o ) , where the union is taken over all orientations on @xmath0 . furthermore , there is also a way of characterising dr configurations using orientations .
[ th : car_dr ] a ( stable ) configuration @xmath16 is deterministically recurrent if and only if there exists an orientation @xmath93 of @xmath0 with no directed cycles such that @xmath98 .
this theorem is intuitive given the bijection between dr configurations and spanning trees , as any spanning tree can induce a ( not necessarily unique ) orientation with no directed cycles .
our second result uses theorem [ car_sr ] in order to classify sr configurations according to the total number of grains , or equivalently to the number of grains removed from the maximal configuration .
we do this by means of the lacking polynomial , which we now define .
let @xmath87 .
the _ lacking polynomial _
@xmath4 of @xmath0 is the generating function of the stochastically recurrent configurations on @xmath0 , with @xmath53 conjugate to the number of lacking particles in the configuration : @xmath99 where @xmath100 an example of the lacking polynomial is shown in figure [ fig : exa1 ] .
note that we can use the lacking polynomial to count the number of sr configurations , as @xmath101 . before stating the main result of this section ( theorem [ thm : lp ] ) , which gives a tutte - like formula to compute @xmath4 , we state ( without proof ) some propositions concerning some special graphs which serve as an initialisation for the computations ( since @xmath4 will be expressed in terms of the lacking polynomials of some graphs smaller than @xmath0 ) .
related illustrations can be found in figure [ fig : illus ] .
[ pro : init_cond ] let @xmath102 . 1 .
if @xmath103 and there are @xmath104 edges between @xmath12 and @xmath9 , then @xmath105 2 . if @xmath0 is a tree , then @xmath106 .
if we can write @xmath0 as the union of connected graphs @xmath107 , so that the @xmath108 are mutually disjoint , then @xmath109 we say that @xmath0 is the _ product _ of the @xmath110 .
we now state that the pruning of tree branches " of a graph @xmath0 does not change its lacking polynomial .
[ def_tb ] let @xmath102 .
a _ tree branch _ of @xmath0 is a subgraph @xmath111 of @xmath0 which is a tree attached to the rest of @xmath0 at the vertex @xmath112 . in other words
, @xmath113 is a tree , @xmath114 for any @xmath115 , and @xmath116 .
[ pro_tb ] let @xmath102 , and let @xmath111 be a tree branch of @xmath0 .
define @xmath117 as @xmath0 with the tree branch removed .
then @xmath118 note that this lemma implies the second statement of proposition [ pro : init_cond ] .
we now give a formula allowing one to compute the lacking polynomial for a given graph ; this formula is similar to the one used to compute the tutte polynomial .
first we require some definitions of edge deletion and contraction , similar to those used in the tutte polynomial relation ( see e.g. bernardi @xcite and references therein ) .
let @xmath87 , and consider an edge @xmath119 , with @xmath120 . 1 .
* edge deletion*. the graph @xmath121 is the graph @xmath0 with @xmath55 removed , i.e. @xmath122 .
general edge contraction*. define the graph @xmath123 as follows : * if @xmath55 is simple , then @xmath123 is @xmath0 with @xmath55 contracted , i.e. @xmath124 , where edges adjacent in @xmath0 to either @xmath53 or @xmath54 are now connected to @xmath125 instead .
* if @xmath55 has multiplicity @xmath126 , contract one of these edges as above , and replace the other @xmath127 edges with @xmath127 edges @xmath128 .
we illustrate these operations in figure [ fig : edge_manip ] .
[ thm : lp ] let @xmath102 , and let @xmath55 be an edge of @xmath88 which is neither a bridge ( i.e. removing @xmath55 does nt disconnect the graph ) , nor connected to the sink .
then [ eq : master - formula ] l_g(x ) = x l_ge(x)+l_g.e(x ) . by way of contrast ,
the corresponding relation for the tutte polynomial , where @xmath55 is not a loop or bridge , is @xmath129 although here loops are allowed in @xmath0 and so @xmath123 denotes regular edge contraction rather than the version defined above .
one can check that the use of theorem [ thm : lp ] , proposition [ pro : init_cond ] , and lemma [ pro_tb ] allows one to compute @xmath4 for any graph @xmath0 in @xmath10 without referring to the ssm . more specifically ,
theorem [ thm : lp ] expresses @xmath4 in terms of graphs with one less edge
. we can continue to use this theorem , and case 3 in proposition [ pro : init_cond ] , until we express @xmath4 in terms of graphs which only contain edges to the sink and bridges .
the bridges must then form tree branches which are removed by lemma [ pro_tb ] , and case 1 in proposition [ pro : init_cond ] provides the lacking polynomials of the remainders .
it is not immediately obvious that the polynomial so obtained does not depend on the edges that we choose to delete / contract ; however , since the lacking polynomial itself is well - defined , this follows from theorem [ thm : lp ] .
a simple corollary of theorem [ thm : lp ] gives us the degree of the lacking polynomial .
let @xmath130 .
let the _ number of cycles _ @xmath131 be the number of edges it is necessary to remove in order to turn @xmath0 into a tree , i.e. @xmath132 .
then the degree of the lacking polynomial @xmath4 is the number of cycles @xmath131.[cor : degree ] we use induction on @xmath133 . if @xmath134 then @xmath135
. thus @xmath136 , @xmath137 as desired .
now consider a graph @xmath138 , and an edge @xmath55 which is not a bridge or connected to the sink .
if there is no such edge , then after removing tree branches , which does not affect @xmath131 , we can express @xmath0 as the product of graphs of the form described in the first case of proposition [ pro : init_cond ] , which obviously satisfies the corollary . otherwise , the graphs @xmath121 and @xmath123 both have @xmath139 edges .
moreover , @xmath140 and @xmath141 .
the result then immediately follows by induction using .
let @xmath142 be the set of stable configurations on @xmath0 compatible with some orientation .
we wish to show that @xmath143 .
we do this by showing that they are subsets of each other .
let @xmath130 .
we have @xmath144 firstly we show that @xmath145 . for this , consider a spanning tree of @xmath0 and orient all edges of this tree outwards from the sink ; that is , if @xmath91 is an edge of the spanning tree such that @xmath21 is closer to the sink than @xmath68 , orient the edge @xmath146 .
orient all remaining edges in any direction
. then for the resulting orientation every non - sink vertex has at least one incoming edge .
condition now shows that @xmath46 is compatible with this orientation , so @xmath145 .
now let @xmath93 be an orientation compatible with @xmath46 , and let @xmath147 .
we shall construct an orientation @xmath148 which is compatible with @xmath149 . to do this , consider a history of grain additions and stochastic topplings that leads from @xmath46 to @xmath149 .
we construct @xmath148 iteratively from @xmath93 by making the following changes to the orientation at each step of this history : * if a grain topples from a vertex @xmath21 to a neighbour @xmath68 and the edge @xmath91 is oriented @xmath146 , we reverse the orientation of this edge so that it is oriented @xmath150 . * otherwise that is , if the edge @xmath91 is oriented @xmath150 , or if a grain is added do nothing .
these changes are shown in figure [ fig : history ] .
[ fig : history ] we now show that @xmath151 .
let @xmath152 be the configurations constructed at each step of the history , some of which may be unstable .
let @xmath153 be the corresponding orientations .
we show by induction on @xmath56 that for any @xmath154 , @xmath41 is compatible with @xmath155 .
for @xmath156 we have @xmath157 by definition . at any fixed @xmath158 , there are two cases : 1 . @xmath41 is reached from @xmath39 by addition of a grain at some vertex @xmath21 .
+ for any vertex @xmath5 , we have @xmath159 and by construction @xmath160 . since @xmath39 is compatible with @xmath161 , condition is satisfied for @xmath41 and @xmath155 for all vertices in @xmath22 .
thus @xmath41 is compatible with @xmath155 .
@xmath41 is reached from @xmath39 through a ( legal ) toppling of some vertex @xmath21 .
+ let the neighbours of @xmath21 be @xmath162 , and suppose the toppling transfers @xmath163 grains from @xmath21 to each of these vertices respectively . since the toppling is legal , @xmath164 , and so after the toppling , @xmath165
now grains have toppled out from @xmath21 along @xmath166 edges , so by construction all these edges are oriented towards @xmath21 .
therefore @xmath167 and condition is satisfied at @xmath21 .
+ it remains to check that is still satisfied at each @xmath168 . at most @xmath169 incoming edges to @xmath168 in @xmath161 are now outgoing in @xmath155 , so @xmath170 . furthermore ,
@xmath171 , so by induction is satisfied at @xmath168 . +
no other vertices apart from @xmath21 and its neighbours are changed , so @xmath41 is compatible with @xmath155 .
taking @xmath172 , we have shown that @xmath148 is compatible with @xmath149 .
this completes the proof of the lemma .
let @xmath130 .
we have @xmath173 we use induction on the number of vertices @xmath174 .
if @xmath175 , then @xmath0 is simply one vertex @xmath5 , connected to @xmath9 by say @xmath104 edges
. then @xmath176 consists of all configurations @xmath16 where @xmath177 , which are sr by proposition [ pro : init_cond ] .
suppose now that the lemma is true for any graph with @xmath178 .
take a graph @xmath179 with @xmath180 and a configuration @xmath181 .
we wish to show that @xmath16 is sr . to do this ,
take an orientation @xmath93 of @xmath0 such that @xmath97 , and choose a vertex @xmath12 connected to @xmath9 .
we can assume that all edges connecting @xmath12 to @xmath9 are oriented towards @xmath12 ( if some are not , we can reverse them and the resulting orientation remains compatible with @xmath16 ) .
write @xmath182 for the number of such edges .
let @xmath183 be the graph obtained from @xmath0 by identifying @xmath12 with @xmath9 , removing all @xmath184 edges ( denote the new sink by @xmath185 ) . note in particular that the degree of any vertex in @xmath183 ( except @xmath185 ) is equal to its degree in @xmath0 .
we now let @xmath186 be the orientation on @xmath183 coinciding with @xmath93 on all edges still present , except that any edge connected to @xmath185 is oriented away from it . for any vertex @xmath5 in @xmath183 ,
let @xmath187 be the number of edges oriented @xmath188 in @xmath93 ( @xmath189 ) .
we define a configuration @xmath190 on @xmath183 by @xmath191 for @xmath192 in @xmath183 .
now for any such @xmath5 , we have @xmath193 and @xmath194 . since @xmath97
, we deduce that @xmath195 .
hence by induction @xmath196 , and therefore there exists a history of grain additions and legal topplings that leads from @xmath46 to @xmath190 on @xmath183 .
we now start from @xmath46 on @xmath0 and copy this history ( all legal topplings remain legal because the degrees in the two graphs are identical ) .
this results in a configuration @xmath149 on @xmath0 with @xmath197 if @xmath198 .
we then continue the history by either adding grains to @xmath12 or repeatedly toppling grains from @xmath12 to @xmath9 until @xmath12 has @xmath199 grains , i.e. it is minimally unstable .
we now make one final toppling at @xmath12 , sending @xmath187 grains to each of its neighbours @xmath5 , and @xmath200 grains to the sink . in order to do this , we must have at least @xmath201 edges @xmath184 .
however , since @xmath16 and @xmath93 are compatible , we know that @xmath202 , so this is true .
denote by @xmath203 the configuration we finally reach .
we have : * if @xmath5 is neither @xmath12 nor one of its neighbours , then @xmath204 . * if @xmath5 is a neighbour of @xmath12 , then @xmath205 .
* @xmath206 .
together , this shows that @xmath203 is identical to @xmath16 except at @xmath12 , where it may have less grains .
we then merely add the difference in grains to @xmath12 , and have thus created a history of grain additions and legal topplings which leads from @xmath46 to @xmath16 .
therefore @xmath16 is sr and the lemma , and theorem [ car_sr ] , is proved .
let @xmath207 be compatible with an orientation with no directed cycles .
let @xmath93 be such an orientation .
since @xmath93 has no directed cycles , and we can take all edges adjacent to the sink as oriented away from it , we can order the vertices of @xmath0 as @xmath208 so that there exist no edges @xmath209 for @xmath210 .
now we apply the burning algorithm to @xmath16 .
we claim by induction that this can burn the vertices in the order described above .
since @xmath211 , the initial condition is trivial .
now suppose we have burned vertices @xmath212 .
all incoming edges to @xmath213 are burnt , so the number of unburned edges adjacent to @xmath213 is @xmath214 .
but by , @xmath215 therefore @xmath213 can be burnt .
thus the burning algorithm burns all the vertices of the graph , and @xmath16 is deterministically recurrent .
conversely , let @xmath216 .
now apply the burning algorithm to @xmath16 , and every time we burn a vertex @xmath5 , orient all edges from previously burnt vertices to @xmath5 as incoming edges to @xmath5 . since all vertices are burnt , this produces a full orientation @xmath93 on @xmath0 , which obviously has no directed cycles . from the burning condition , we know that for all @xmath5 , @xmath15 is greater than the number of unburnt edges , which is @xmath217 .
this gives @xmath218 since these are integers , this means that is fulfilled for all vertices .
thus @xmath93 is compatible with @xmath16 and the theorem is proved .
we construct a bijection @xmath220 from @xmath221 to @xmath79 such that for any @xmath222 , @xmath223 .
define for @xmath222 and @xmath26 @xmath224 take an orientation @xmath93 on @xmath225 which is compatible with @xmath16 , and extend this orientation to @xmath0 by orienting each edge in @xmath113 away from @xmath112 .
then @xmath226 is compatible with the resulting orientation , so @xmath227 .
moreover , @xmath220 is clearly an injection .
it remains to show that it is surjective . to see this ,
consider a configuration @xmath228 and a compatible orientation @xmath93 .
each vertex in @xmath229 must have at least one incoming edge in @xmath230 .
but since @xmath113 is a tree , @xmath231 .
thus each edge in @xmath230 points to a different vertex in @xmath113 , so @xmath232 for all @xmath115 .
this implies that @xmath233 .
furthermore , all edges in @xmath230 adjacent to @xmath112 point away from it .
now define @xmath149 on @xmath225 according to @xmath234 , and let @xmath148 be the restriction of @xmath93 to @xmath225 .
we have @xmath235 for all @xmath236 , so clearly @xmath149 is compatible with @xmath148 and @xmath237 .
thus @xmath220 is a bijection .
let @xmath102 , and let @xmath55 be an edge of @xmath88 which is neither a bridge nor connected to the sink . write @xmath238 .
for @xmath228 we distinguish the following two cases : a. there exists an orientation @xmath93 on @xmath0 , compatible with @xmath16 , such that @xmath55 is oriented @xmath146 in @xmath93 , and @xmath239 . b. for all orientations @xmath93 compatible with @xmath16 , all @xmath91 edges are oriented @xmath150 , or @xmath240 .
we write @xmath241 if @xmath16 satisfies condition @xmath242 .
obviously @xmath243 .
now we define a function @xmath244 as follows : * if @xmath245 then @xmath246 is a configuration on @xmath121 . *
if @xmath247 then @xmath248 is a configuration on @xmath123 . to simplify further calculations , we note that in the first case , @xmath249 and @xmath250 . in the second case , @xmath251 .
it is easy to see that if @xmath16 is stable , @xmath252 is also stable the lacking number can decrease by at most one , and this occurs only at @xmath68 when @xmath245 , where by definition @xmath253 .
moreover , we have @xmath254 in light of this , it is sufficient to show the following theorem to prove theorem [ thm : lp ] since the two sets @xmath255 and @xmath256 are disjoint , being configurations on different graphs .
[ thm : bij ] let @xmath102 .
the function @xmath257 defined above is a bijection from @xmath79 to @xmath258 .
this theorem is itself a direct consequence of the four following lemmas .
the first lemma shows that for any @xmath228 , the configuration @xmath252 is indeed stochastically recurrent . [ well_def ] for any @xmath259 in @xmath10 , @xmath260 take @xmath228 , and fix an orientation @xmath93 on @xmath0 such that @xmath97 .
there are two cases .
1 . @xmath245 . + by construction
, we may choose @xmath93 such that @xmath55 is oriented @xmath146 .
let @xmath148 be the orientation on @xmath261 which is identical to @xmath93 on all edges of @xmath121 .
then for any vertex @xmath5 , @xmath262 so @xmath263 . thus , by theorem [ car_sr ] , @xmath264 .
2 . @xmath247 .
+ let @xmath104 be the number of @xmath91 edges in @xmath0 .
then in @xmath123 , these will be replaced by @xmath127 edges @xmath265 .
orient these as @xmath266 , and orient all other edges of @xmath123 as they are oriented in @xmath93 .
denote by @xmath148 the resulting orientation on @xmath123 .
then @xmath267 so condition is satisfied at @xmath268 . since @xmath269 for @xmath270 ,
it is clearly also satisfied elsewhere , so @xmath263 , and by theorem [ car_sr ] , @xmath271 .
we write @xmath272 ( resp .
@xmath273 ) for the restriction of @xmath257 to the set @xmath274 ( resp .
@xmath275 ) . we will show that each of these are bijections onto their respective images .
[ bij_a]for any @xmath259 in @xmath10 , the function @xmath272 is a bijection from @xmath274 to @xmath276 .
the fact that @xmath272 is injective follows immediately from the definition of @xmath257 . to show that @xmath272 is surjective onto @xmath276 , let @xmath277 and take an orientation @xmath93 on @xmath121 compatible with @xmath16 .
let @xmath278 , and define @xmath279 as a configuration on @xmath0 .
it is obvious that @xmath149 is stable and @xmath280 .
then for any vertex @xmath5 , @xmath281 thus @xmath282 so @xmath149 is sr .
moreover , @xmath55 is oriented @xmath146 in @xmath148 and @xmath283 , so @xmath284 as desired .
[ surj_b ] for any @xmath259 in @xmath10 , the function @xmath273 is surjective onto @xmath256 . for any @xmath285 ,
let @xmath93 be a compatible orientation on @xmath123 .
let @xmath104 be the multiplicity of the edge @xmath55 in @xmath0 ( as before we may have @xmath286 ) .
we may assume that the @xmath127 edges @xmath287 in @xmath123 are all oriented @xmath266 in @xmath93 .
write @xmath288 ( resp .
@xmath289 ) for the number of edges oriented into @xmath268 in @xmath93 which correspond to edges into @xmath21 ( resp .
@xmath68 ) in @xmath0 , from vertices other than @xmath68 ( resp .
@xmath21 ) .
this is illustrated in figure [ fig : edges ] . since @xmath16 is compatible with @xmath93
, we have [ cond_a.b ] ^g.e_o(a.b ) = i_a + i_b + k-1 1 + l^g.e_(a.b ) .
let @xmath290 be the extension of @xmath273 to all of @xmath79 , i.e. @xmath291 the difference between @xmath292 and @xmath273 lies in the fact that @xmath292 is defined over all @xmath79 , whereas @xmath273 is defined only on @xmath275 .
we will now define a configuration @xmath147 such that @xmath293 .
firstly let @xmath294 if @xmath295 .
likewise let @xmath148 be an orientation on @xmath0 where all edges in @xmath123 are oriented identically to @xmath93 ( the remaining edges are as yet unspecified ) .
obviously is satisfied for @xmath149 and @xmath148 at vertices other than @xmath21 and @xmath68 .
we now assign grains to @xmath21 and @xmath68 , and orientations to the @xmath91 edges , according to 3 cases .
1 . @xmath296 .
+ we set @xmath297 , @xmath298 . in @xmath148 ,
orient one @xmath91 edge as @xmath150 and the remaining as @xmath146 .
clearly , is satisfied at @xmath21 ( which has one incoming edge ) , and @xmath299 from , so is also satisfied at @xmath68 .
2 . @xmath300 .
+ we set @xmath301 , @xmath302 . in @xmath148 , orient all @xmath104 edges @xmath91 as @xmath146 .
then @xmath303 @xmath304 again using , so is satisfied at both @xmath21 and @xmath68 .
3 . @xmath305 .
+ we set @xmath306 , @xmath307 . in @xmath148 ,
orient all @xmath104 edges @xmath91 as @xmath146 .
then we have @xmath308 , and @xmath309 .
thus condition is satisfied at @xmath21 and @xmath68 .
now , in each of these cases we have @xmath310 , so @xmath293 .
likewise , @xmath149 is compatible with @xmath148 , so @xmath147 .
it remains to show that we may choose an @xmath311 so that @xmath280 . to show this ,
define @xmath312 .
since there exists at least one such @xmath149 , this is well - defined .
now take @xmath147 such that @xmath313 and @xmath293 .
we show that @xmath311 . if @xmath314 , this is true by definition .
now suppose that @xmath315 and there exists an orientation @xmath148 on @xmath0 compatible with @xmath149 with an edge oriented @xmath146 .
we define the orientation @xmath316 as @xmath148 with that edge reversed and all other edges oriented as in @xmath93 .
likewise , define the configuration @xmath317 , so that @xmath318 , @xmath319 , and @xmath320 elsewhere .
since @xmath282 , we have @xmath321 by construction .
now we have @xmath322 and @xmath323 , but @xmath324 .
this is a contradiction of the definition of @xmath325 .
therefore no such orientation @xmath148 exists , and @xmath311 .
[ inj_b]for any @xmath259 in @xmath10 , the function @xmath273 is injective .
let @xmath326 such that @xmath327 , that is @xmath328 and @xmath329 if @xmath330 , but suppose @xmath331 .
assume without loss of generality that @xmath332 .
now choose compatible orientations @xmath333 and @xmath334 respectively on @xmath0 .
suppose that the edge @xmath55 has multiplicity @xmath335 in @xmath0 .
since @xmath336 and @xmath337 , these @xmath104 edges must be oriented @xmath150 in @xmath334 .
now , if @xmath338 , then reversing the orientation of @xmath55 to @xmath146 in @xmath334 results in another orientation compatible with @xmath339 , contradicting the fact that @xmath336 .
therefore these quantities are equal and [ eq : lem47 ] ^g_o_1(a ) 1 + l^g__1(a ) > 1 + l^g__2(a ) = ^g_o_2(a ) .
now let @xmath340 be the set of edges in @xmath334 which are oriented differently from @xmath333 .
we define a subgraph @xmath341 of @xmath0 as the union of all directed paths in @xmath334 starting from @xmath21 whose edges are in @xmath340 ( and the induced vertices ) . from ( [ eq : lem47 ] ) , this contains at least one edge adjacent to @xmath21 .
firstly , we claim that @xmath342 .
otherwise , there exists a directed path from @xmath21 to @xmath68 in @xmath334 . starting from @xmath334 , we may reverse @xmath55 and all edges of this path to reach an orientation with the same number of incoming edges at each vertex as @xmath334 , and therefore compatible with @xmath339 , but with @xmath55 oriented @xmath146 .
this contradicts the assumption that @xmath336
. now start from @xmath334 and reverse the orientation of @xmath55 and all edges in @xmath343 .
denote this orientation by @xmath344 .
we show that @xmath344 is compatible with @xmath339 by checking condition at @xmath68 and vertices in @xmath345 ( which include @xmath21 ) : * @xmath346 , since @xmath55 has been reversed and @xmath342 . * for @xmath347 , all incoming edges to @xmath5 in @xmath333 are identically oriented in @xmath344 by construction .
therefore @xmath348 , where the last inequality is strict if @xmath349 and an equality otherwise .
this gives us an orientation compatible with @xmath339 containing an edge @xmath146 .
again , this is a contradiction of the assumption that @xmath336 .
thus there can not exist configurations @xmath331 in @xmath275 such that @xmath327 , and @xmath273 is injective .
in this paper , we have devised a generalisation of the asm in which the topplings are stochastic .
this model behaves qualitatively differently to the established asm of dhar . in particular
, the set of recurrent states of this model contains that of the former model . we have proved a characterisation of these states using graph orientations .
we also define a generating function of these states which counts the number of lacking " grains , and show that this lacking polynomial " satisfies a recurrence relation which resembles that of the tutte polynomial .
there are two directions in which to advance this work .
given the many combinatorial interpretations of the tutte polynomial , it would be of interest to see if the lacking polynomial demonstrates similar interpretations .
in other words , the lacking polynomial may count certain combinatorial objects for given values of its parameter , and we would like to determine what these objects are . alternatively , the lacking polynomial may be related in some way to the tutte polynomial , and the nature of this relation should be determined precisely .
the other topic of interest is to probe further into the behaviour of the markov chain structure , more specifically the steady state . in the asm ,
all recurrent states are equally likely , but this is not the case for the ssm
. it would be interesting to calculate the probabilities for the stochastically recurrent states .
once we have done so , we can analyse the behaviour of the model in the steady state , and see if it displays a similar power - law behaviour to that observed for the classic model . | we introduce a new model of a stochastic sandpile on a graph @xmath0 containing a sink .
when unstable , a site sends one grain to each of its neighbours independently with probability @xmath1 $ ] . for @xmath2 ,
this coincides with the standard abelian sandpile model . in general , for @xmath3 , the set of recurrent configurations of this sandpile model is different from that of the abelian sandpile model .
we give a characterisation of this set in terms of orientations of the graph @xmath0 .
we also define the lacking polynomial @xmath4 as the generating function counting this set according to the number of grains , and show that this polynomial satisfies a recurrence which resembles that of the tutte polynomial . |
the entanglement properties of many - body systems have attracted considerable attention and have become the topic of a large number of studies in the last decade @xcite . in particular
, the study of a suitable measure of entanglement such as the entanglement entropy was triggered by the need to understand ground state properties that lead to the emergence of area laws @xcite .
this keyword refers to a rather generic property of ground states of bipartitioned systems in which the entropy @xmath0 of a subsystem scales as the number of contact points with the environment .
the most notable and well understood counterexamples are represented by one - dimensional critical quantum systems with an underlying conformal symmetry where the area law is violated in the form of universal terms , scaling as the logarithm of the subsystem size @xcite . in the translationally invariant case
this can be written as @xmath1 with @xmath2 being the central charge of the conformal field theory ( cft ) and @xmath3 a non - universal constant .
the presence of impurities has interesting effects on the entanglement entropy @xcite .
an example is given by critical lattice models where single defective links separate the two subsystems .
this can lead to a modified prefactor for the logarithm varying continuously with the defect strength in models with free fermions @xcite , while for interacting electrons the defect either renormalizes to a cut or to the homogeneous value in the @xmath4 scaling limit @xcite . in the more general context of a conformal interface separating two cfts
the effective central charge has been calculated recently and was shown to depend on a single parameter @xcite .
another broad and intensively studied class of impurity problems is related to the kondo effect@xcite . from the viewpoint of block entropies
a spin - chain version of the kondo model has been studied @xcite using density matrix renormalization group ( dmrg ) methods @xcite . here
a different geometry was considered with an impurity spin coupled to one end of a finite chain .
the induced change in the entropy varies between zero and @xmath5 and its qualitative behavior for various system sizes and coupling values can be described in terms of an impurity valence bond picture .
a counterpart of the kondo effect can be found in a single impurity model introduced by anderson @xcite . a similar
, exactly solvable model was studied independently by fano @xcite where the interaction of a discrete state with a continuum of propagation modes leads to scattering resonances .
the exact form of these resonances is controlled by the couplings between the modes .
here we consider a simple geometry where an impurity is side - coupled to a linear chain of electron conduction sites .
this setting can also be realized experimentally with gated semi - conductor heterostructure quantum dots . in accordance with theoretical predictions @xcite ,
fano resonances were detected as dips in the conductance measurements at low temperatures@xcite . despite the large amount of theoretical and experimental work on fano resonances @xcite ,
the question how they affect entanglement properties is still unanswered . in the present paper
we address this question by investigating the simplest model capturing the main features of the fano resonance .
the fano - anderson model @xcite is described by a free fermion hamiltonian , thus standard techniques are available for the study of its entanglement properties @xcite .
we find that the entropy scaling of a block neighboring the impurity is of the form ( [ eq : enthom ] ) with a prefactor @xmath6 which decreases monotonously as the parameters are tuned towards the fano resonance and depends only on a well - defined scattering amplitude .
the numerical analysis leads to the same functional form of @xmath6 that has recently been derived for simpler fermionic models with interface defects @xcite and also seems to be closely related to the one found for conformal interfaces @xcite . if the subsystem contains an open boundary we find a rapidly oscillating subleading term in the entropy that persists even in the large @xmath7 limit . in the following we
first introduce the model and the geometries considered , along with the methods used to calculate the entropy . in section [ sec : inf ] we treat an impurity attached to an infinite hopping model , determine the correlation matrix and present results on the scaling of the entropy as well as the particle number fluctuations .
the effects of an open boundary will be detailed in sec .
[ sec : sinf ] followed by our concluding remarks in sec .
[ sec : concl ] .
the derivation of an asymptotic form of the correlation functions and some of the necessary formulae for the semi - infinite case are presented in two appendices .
we consider a model of non - interacting spinless electrons described by a 1d tight - binding chain that is coupled to an impurity .
the impurity is represented by an additional site from which tunneling events can only occur to a specific site @xmath8 of the chain .
the hamiltonian is @xmath9 where @xmath10 and @xmath11 are the fermion annihilation operators along the chain and at the impurity site , respectively , while @xmath12 and @xmath13 are the corresponding tunneling strengths .
the potential @xmath14 sets the filling of the chain and @xmath15 is the site - energy of the impurity .
we will consider two different geometries : * in the _ infinite geometry _
the @xmath16 sites are indexed as @xmath17 while the impurity is located at the center of the chain ( @xmath18 ) and we take periodic boundary conditions * in the _ semi - infinite geometry _
one has @xmath19 sites with @xmath20 while @xmath8 is arbitrary and we have open boundary conditions the two geometries are depicted in fig .
[ fig : geom ] . in the following we set @xmath21 . after a fourier transformation one has @xmath22 where @xmath23 .
the couplings and the allowed wavenumbers are @xmath24 with @xmath25 for the infinite and @xmath26 with @xmath27 for the semi - infinite case , respectively , where @xmath28 runs over the corresponding site - indices . the fermi - level @xmath29 is set to zero by applying a potential @xmath30 .
the hamiltonian ( [ eq : hamq ] ) is known as the fano - anderson model and is diagonalized by introducing new fermionic operators @xmath31 by the transformation @xcite @xmath32 where @xmath33 @xmath34 with the self - energy term defined as @xmath35 the infinitesimal terms @xmath36 are needed to regularize the sum .
note , that the number of different @xmath37 values equals the number of all sites , including the impurity . in the limit @xmath38
one has a continuum of the unmodified single - particle spectrum @xmath39 of the chain while additional bound states can emerge as real solutions @xmath40 of the equation @xmath41 our aim is to calculate the entanglement entropy @xmath42 of a block of @xmath7 sites neighboring the impurity , where @xmath43 denotes the corresponding reduced density matrix .
since one is dealing with free fermions , this can be obtained through the eigenvalues @xmath44 of the correlation matrix @xmath45 restricted to the block @xmath46 as @xcite @xmath47
we first consider the infinite geometry and calculate the matrix @xmath48 analytically ; this result is in turn used to numerically obtain the entropy .
the particle number fluctuations are also considered where explicit calculations are performed using an asymptotic form of the correlations .
the correlations @xmath49 are calculated by first going over to fourier modes @xmath50 and subsequently applying the transformation ( [ eq : munu ] ) .
the expectation values to be evaluated are then trivial and read @xmath51 where in the ground state @xmath52 for the modes @xmath53 and zero otherwise . moreover , in the infinite geometry it was shown that eq .
( [ eq : bseq ] ) always gives two real solutions with @xmath54 and @xmath55 corresponding to bound states above and below the band @xcite .
the energies @xmath56 are obtained numerically by solving the quartic equation ( [ eq : bseq ] ) with @xmath57 these modes @xmath58 have to be dealt with separately in the transformation ( [ eq : munu ] ) with the factors @xmath59 where @xmath60 is set by the normalization condition @xmath61 . since @xmath62
, one has a contribution @xmath63 to the correlations coming from the lower bound state . setting @xmath64 and taking @xmath65 the sums
are replaced with integrals over the full band @xmath66 $ ] and can be rewritten as contour integrals over the complex plane .
evaluating the residues one has @xmath67 with @xmath68 for @xmath69 while the correlation length is defined by @xmath70 .
there are two contributions from the conduction band .
first we have the translationally invariant terms @xmath71 that give the correlations without the impurity arising from the @xmath72 term in ( [ eq : mu ] ) .
the remaining term gives rise to the correlation contributions @xmath73 ; these can be evaluated by applying the same methods used for determining @xmath74 .
the calculation yields the simple form @xmath75 where the scattering phases for @xmath76 are defined as @xmath77 here @xmath78 is the imaginary part of the retarded self - energy with the real part being zero . collecting all the contributions ,
the correlation matrix is given by @xmath79 the integrals in ( [ eq : corr1inf ] ) must be evaluated numerically and are shown on fig .
[ fig : corrfried ] together with the contribution of the bound state for some fixed values of the parameters @xmath13 and @xmath15 .
the values of @xmath80 are found to oscillate around the exponentially decaying curve of @xmath74 .
since they appear with opposite signs in ( [ eq : corrinf ] ) , the average value will cancel , corresponding to the screening of the impurity induced localized state by the conduction electrons . and from the conduction band @xmath80 for the parameter values @xmath81 and @xmath82 .
the inset shows the difference between these two contributions , compared to the asymptotic form ( [ eq : corrfried ] ) represented by the dashed lines . ]
the overall contribution is shown in the inset of the figure .
a careful analysis shows that the asymptotic behavior has the form of friedel oscillations @xmath83 where only the scattering phase @xmath84 at the fermi - level enters .
@xmath85 is depicted by the dashed lines on the inset of fig .
[ fig : corrfried ] and shows a good agreement with the numerical data .
the derivation of eq .
( [ eq : corrfried ] ) is summarized in appendix a. it relies on an appropriate transformation of the integrals in ( [ eq : corr1inf ] ) and is valid for @xmath86 where the length scale @xmath87 is given in eq .
( [ eq : lnull ] ) .
the elements of the correlation matrix ( [ eq : corrinf ] ) will be used to obtain the entanglement entropy of a block numerically as described in section [ sec : mm ] . in some simple cases
one can already give an answer by looking at the limiting form of the correlations .
taking @xmath88 , the impurity site will either be completely empty or completely occupied and therefore it becomes decoupled from the rest of the chain .
one has @xmath89 while the contributions from the bound state also vanish @xmath90 , hence @xmath91 and the entropy will just be that of a homogeneous hopping model given by ( [ eq : enthom ] ) with @xmath92 .
obviously , the limit of a vanishing coupling yields the same behavior . on the other hand , one could take the limit @xmath93 for very strong coupling .
this corresponds to @xmath94 , that is , one has resonant scattering at every wavelength .
the ground state will correspond to a singlet formed by the impurity and site zero that is otherwise decoupled from the rest of the system .
the bound state correlations contribute only on site zero @xmath95 while for sites @xmath96 one has @xmath97 , which is the form for a semi - infinite chain .
the entropy of this geometry is known to scale logarithmically with a coefficient @xmath98 , which is half of the value that appears in the homogeneous case @xcite . at half filling and for various coupling strengths @xmath13 .
the upper and lower - most dashed lines have slopes @xmath99 and @xmath98 , respectively , acting as guides to the eye . ]
the previous analysis shows that resonant scattering tends to decrease the coefficient of the entropy scaling .
however , in general a fano resonance is concentrated around a single wavenumber .
this is expected to have the largest effect on the asymptotic behavior if the location of the resonance coincides with the fermi - level @xmath100 yielding the condition @xmath101 .
we take this value of the site energy in figure [ fig : entcrossover ] , in which we plot the entropy for a range of values for the coupling @xmath13 .
we see that the entropy crosses over between the two limiting behaviors , giving a slope of @xmath99 for small @xmath7 and @xmath98 for large @xmath7 when plotted against @xmath102 .
therefore , taking the block size large enough the impurity behaves like a cut in the chain .
this effect is reminiscent of the kondo screening mechanism , where conduction electrons form a cloud around a magnetic impurity .
the kondo effect has an associated length scale @xcite that being the distance beyond which the impurity appears to be screened off . in the present case a natural length scale also enters , which marks the asymptotic regime of the correlations and
could be expected to appear in the entropy scaling . from eq .
( [ eq : lnull ] ) one has @xmath103 for @xmath104 and @xmath105 , which seems to be consistent with the location of the crossover moving toward larger @xmath7 in fig .
[ fig : entcrossover ] as the coupling @xmath13 decreases . in order to reliably verify the scaling of the crossover length one would have to consider larger blocks .
unfortunately , the achievable sizes were limited to @xmath106 by the increasing numerical difficulty in evaluating the oscillatory integrands for the matrix elements ( [ eq : corr1inf ] ) .
for @xmath107 the screening cloud is effectively in a singlet state with the impurity electron and cuts the system in two parts .
the coefficient of the entropy therefore renormalizes to the value @xmath98 corresponding to a semi - infinite chain with an open boundary .
note , that a similar renormalization behavior was found for _ interacting _ electrons in the luttinger liquid regime bisected by a hopping defect @xcite .
our further numerical analysis shows that , for arbitrary parameter values and @xmath108 , the entropy can be written in the form @xmath109 where the effective central charge @xmath6 varies continuously between the values @xmath110 and @xmath111 .
the same behavior was found for free fermionic models with hopping defects @xcite where the dependence of @xmath6 on the defect strength was determined by data fits .
recently , sakai and satoh derived the same form as eq .
( [ eq : entropy ] ) for the case of two conformal field theories coupled through a conformal interface . in their case they found @xmath6 in a closed form with dependence only on a single scattering amplitude @xcite .
these conformal interfaces describe a discontinuity in the compactification radii of two bosonic cfts , which correspond to luttinger liquids with unequal interaction parameters on the left and right hand side @xcite .
although the result from @xcite is not directly applicable , a closely related form was recently found to describe the case of free fermions with simple interface defects @xcite .
we therefore attempted to generalize these results to the present model .
motivated by eq .
( [ eq : corrfried ] ) , we argue that the asymptotic behavior of @xmath112 and thus @xmath6 should depend only on the scattering phase at the fermi - level and a suitable scattering amplitude can be defined as @xmath113 .
hence , we propose @xmath114 where @xmath110 comes from the unmodified boundary while the integral is , up to a sign , the same as in @xcite and describes the contribution of the impurity .
figure [ fig : ceffint ] shows the resulting @xmath6 obtained by fitting ( [ eq : entropy ] ) on different @xmath112 data sets together with ( [ eq : ceffint ] ) evaluated as a function of @xmath15 for various fixed values of the coupling @xmath13 and filling @xmath115 .
one has excellent agreement with the conjectured analytical form .
the only visible deviations arise for @xmath116 which can be understood through the crossover phenomenon shown on fig .
[ fig : entcrossover ] . in this parameter regime
the asymptotic behavior sets in only for larger @xmath7 and one has considerable finite - size corrections . and for various values of @xmath13 and @xmath115 .
the lines show the analytical form eq .
( [ eq : ceffint ] ) evaluated for the corresponding values of the parameter @xmath117 . ] the integral ( [ eq : ceffint ] ) depends on the model parameters only through the variable @xmath118 interestingly , the above formula _
exactly _ coincides with the transmission coefficient of the fano - anderson model @xcite with the fermi - energy set equal to zero .
therefore , the effective central charge is directly linked to a simple physical quantity that , in experimental realizations , is obtainable via conductance measurements at very low temperatures .
although the description of fano resonances measured in various nanostructure experiments @xcite typically require more detailed impurity models , the simple relation between @xmath6 and the transmission coefficient might still survive in some parameter regime .
this could open up the possibility of an indirect measurement of the leading term in the entropy .
it should be pointed out that the impurity integral in @xmath6 has been obtained _ analytically _ for a transverse ising chain in a dmrg geometry @xcite .
the calculation relies on a correspondence between transfer matrix spectra of classical 2d ising models and the single - particle eigenvalues of @xmath119 in the reduced density matrix @xmath120 of the quantum chain .
this analogy is well - known in the homogeneous case @xcite and can be extended by considering the transfer matrix of a 2d strip with a defect line . in turn , an alternative representation of the integral in ( [ eq : ceffint ] ) via dilogarithm functions was obtained . in our present model , however , a direct calculation of @xmath6 through the single - particle eigenvalues seems to be a far more challenging problem . of the entropy as a function of @xmath15 for different fillings and @xmath121 .
the homogeneous value @xmath3 has been subtracted for better comparison . ]
we also extracted the subleading contribution @xmath122 in ( [ eq : entropy ] ) from our data as shown in fig .
[ fig : kfits ] for different values of the filling .
its value @xmath3 without the impurity also depends on @xmath115 @xcite and was subtracted in the figure .
one has a peaked structure around @xmath104 , however , unlike @xmath6 the constant @xmath122 is in general not symmetric with respect to @xmath15 .
instead , one has the relation @xmath123 which is a direct consequence of the symmetry property eq .
( [ eq : corrsymm ] ) of the correlations proven in appendix a. in the last part of this section we will investigate the fluctuations in the number of electrons @xmath124 contained in the block next to the impurity .
although being a simpler physical quantity , its scaling properties were shown to be similar to that of the entanglement entropy in a variety of 1d quantum systems @xcite and , in the case of free fermions , also in arbitrary dimensions @xcite .
therefore it is an interesting question whether this connection persists for the present impurity problem .
the particle number fluctuation is a simple quadratic function of the correlation matrix and without the impurity is readily evaluated as @xcite @xmath125 with @xmath126 .
note that here the leading - order logarithmic behavior was seen to emerge from the large distance decay properties of the correlations .
therefore , we will use the approximation in eq .
( [ eq : corrfried ] ) to write @xmath127 . carrying out the traces ,
one finds that the additional term @xmath128 has a logarithmic contribution while @xmath129 evaluates to a constant .
the fluctuations then read @xmath130 with the prefactor given by @xmath131 .
the result is tested by comparing with the fitted values of @xmath132 as shown on figure [ fig : fluctfits ] .
the agreement is very good with the only sizeable deviations appearing in the region @xmath133 where the fano - resonance is close to the fermi - level and the approximation ( [ eq : corrfried ] ) breaks down for the numerically attainable block sizes .
note , however , that in general the small distance deviations from the asymptotic form lead only to a different numerical value of @xmath134 as compared with the one obtained by evaluating the subleading terms in the traces
. finally , one has to point out that , apart from the similar logarithmic scaling , the prefactor @xmath132 of the fluctuations is given by a much simpler function of the scattering amplitude than that governing the entropy . of the logarithmic term in the particle number fluctuation for various values of @xmath13 and @xmath115 , plotted against the scaling variable @xmath135 .
the dotted line is the function @xmath136)/2 .
the dashed line shows @xmath6 as a function of @xmath135 for better comparison . ]
we now turn to the investigation of the semi - infinite model .
our motivation behind studying a system with an open boundary is two - fold . on one hand
, we would like to cross - check our argument used for explaining eq .
( [ eq : ceffint ] ) and show that the contributions in @xmath137 are additive .
in other words , we want to test our expectation @xmath138 for the semi - infinite effective central charge @xmath139 of a subsystem with an open end . on the other hand ,
boundaries were shown to induce interesting subleading behavior of the entropy for pure chains@xcite , therefore it is interesting to extend these investigations to the present impurity problem .
the main difference from the infinite case is that the couplings @xmath140 are now dependent on the wavenumber . in principle
the calculations are very similar , leading to slightly more lengthy expressions that are therefore summarized in appendix b. however , there is an additional feature which leads to the emergence of bound states in the continuum ( bic ) which were studied before @xcite .
we will show how these bic naturally enter the problem at the correlation function level .
our main interest is to determine the entropy of a block with @xmath141 sites located on the left hand side of the impurity .
apart from the usual bound state contribution , the correlations from the conduction band now include @xmath142 with @xmath143 and we have defined the function @xmath144 where @xmath145 and @xmath146 are the real and imaginary parts of the self - energy ( [ eq : sigmaqsi ] ) , respectively , and are related to the parameter @xmath78 of the infinite system as @xmath147 note that the term @xmath148 does not now appear in the correlations .
it is instructive to analyze the main features of the integral in eq .
( [ eq : corr1sil ] ) . taking @xmath149 one has @xmath150 and the integral reproduces the correlations of a pure semi - infinite chain .
for very strong coupling @xmath151 one has @xmath152 and therefore both of the parameters in ( [ eq : reimq ] ) take very large values apart from a discrete set defined by the condition @xmath153 , @xmath154 where @xmath155 .
therefore , the function @xmath156 will become sharply peaked around these @xmath157 values and formally one can substitute @xmath158 where the factor @xmath159 is needed for proper normalization . setting @xmath160 ,
the @xmath157 correspond exactly to the allowed wavenumbers in a _
finite _ chain of length @xmath7 .
therefore the block is effectively decoupled by the impurity and @xmath161 reproduces the correlations in a finite segment . at half filling and @xmath121 .
the dashed lines have corresponding slopes @xmath162 and are shown for comparison . ]
the above argument breaks down for special @xmath157-values fulfilling @xmath163 where @xmath156 becomes zero and the corresponding peak is missing .
however , this is exactly the condition for the existence of the bic .
it has been shown that these states have non - vanishing amplitude only at the impurity site and inside the segment where they reproduce , up to normalization , the eigenstates of a chain of length @xmath141 @xcite . for @xmath151
even the normalization becomes exact and the contribution of the missing peak is therefore reincluded this way .
while for these special values of @xmath15 the bic are exactly located at @xmath157 , the resonances of @xmath164 gradually shift towards the wavenumbers of the decoupled block as the coupling to the impurity becomes larger .
hence , their contribution to the entropy is expected to diminish .
the numerical study of the entropy is now carried out using the exact form of the correlations @xmath165 .
the scaling of the entropy at half filling is shown on fig .
[ fig : entsicomm ] for various @xmath15 values . in agreement with our expectations ,
the slope of the curves coincides well with @xmath138 as illustrated by the dashed lines .
additionally , one observes large amplitude oscillations in the data which seem to persist for large values of @xmath7 .
although such oscillations were already pointed out in case of quantum chains with a boundary , they were found to decay according to a power law and therefore vanish in the limit @xmath4 @xcite . a qualitative argument for this alternation
can be given by considering the main parameters ( [ eq : reimq ] ) entering the integrals .
they contain oscillatory functions of @xmath37 that , for large @xmath8 , are expected to average out upon integration , yielding @xmath166 and @xmath167 .
these are just the values for the infinite case , explaining the average behavior of the entropy . however ,
as seen before the fermi level plays an important role in the asymptotics and at @xmath168 the parameters assume values which differ from the average . for half - filling one
has @xmath169 while @xmath170 for @xmath8 even and @xmath171 for @xmath8 odd , and thus oscillates around the average @xmath172 .
this effect could result in subleading corrections to the entropy that survive the @xmath173 limit . in the entanglement
entropy at half - filling , as obtained from data fits .
the symbols correspond to different values of @xmath15 and @xmath13 and show a nice collapse when plotted against the scaling variable @xmath135 of the infinite geometry .
] to extract this term , we have fitted our data according to @xmath174 .
this form , in general , gives a good description of the alternating part .
the results on @xmath175 are plotted on fig .
[ fig : deltasfit ] against the variable @xmath135 .
this choice is justified through our previous argument , where the only non - vanishing parameter is given by @xmath176 and thus the scattering phase of the infinite case is still expected to be a relevant scaling variable .
the data shows indeed a nice collapse to an interpolated scaling function which is shown by the dotted line .
it has a zero at @xmath177 corresponding to the pure semi - infinite chain where the alternating term is known to vanish asymptotically @xcite .
note , that the symmetry under the exchange @xmath178 is valid also for @xmath112 itself .
the behavior of the entropy becomes more complicated when we choose another value of the filling .
this is illustrated on fig .
[ fig : entsiincomm ] where we have compared @xmath179 , corresponding to one - third filling , with @xmath180 which gives an _ incommensurate _
value for the filling . in the former case
the entropy curve splits into three parts , shifted from each other in a similar way as was observed for half - filling on fig .
[ fig : entsicomm ] .
this corresponds to the three possible values @xmath181 and @xmath182 can take .
however , in the incommensurate case the number of distinct values is infinite , resulting in a highly irregular entropy behavior .
since @xmath183 , the effective central charges for the two cases are almost equal which explains why the averaged slope of the curves is very similar
. nevertheless , for the small block sizes numerically achievable , the amplitude of the oscillations has the same magnitude as the average entropy .
therefore , a quantitative understanding of the subleading term would be clearly desirable and requires further investigation . and an incommensurate @xmath180 value of the filling .
the parameters are @xmath184 and @xmath185 . ]
we have studied the entanglement in the ground state of a single impurity problem described by the fano - anderson model , focusing on the analysis of the effective central charge that appears in the entropy scaling .
we provided strong numerical evidence for the validity of a formula which gives the functional form of @xmath6 for arbitrary parameter values of the model .
in particular , this formula depends only on a single parameter , given by eq .
( [ eq : intpar ] ) , that has a simple physical interpretation , namely it is the transmission coefficient of the impurity .
this , in turn , establishes a connection between the entanglement entropy and the low temperature conductance of the quantum chain that can be measured in experiments with quantum - dot nanostructures @xcite .
therefore , it would be interesting to generalize the study of the block entropy for anderson - type impurities , including on - site coulomb interaction between electrons of different spin , that give a more realistic model of the quantum - dots in experiments and yet are still tractable with dmrg techniques . on the other hand , an exact analytical treatment of the present problem is still lacking .
one could follow the lines of ref.@xcite and try to relate the reduced density matrix to the transfer matrix of an appropriate 2d classical system .
another possible approach would be the direct calculation of the entropy from the correlation matrix , which , in the pure case , requires knowledge of the eigenvalues of a toeplitz matrix @xcite . with an impurity
one has a sum of toeplitz and hankel matrices , which poses a much more difficult problem .
however , the analysis of the correlations has led to a simple asymptotic form , enabling us to calculate the particle number fluctuation that involves , instead of diagonalization , only the traces of the correlation matrix .
the investigation of the semi - infinite geometry has , on one hand , supported our general belief that the leading contributions from both of the interfaces bordering the segment simply add up in the entropy . on the other hand , we found a subleading term which shows an interesting oscillatory behavior .
the amplitude of the oscillations remains finite even asymptotically , an effect that has not been observed for pure chains .
although some arguments were given to describe its main features , a proper understanding could only be achieved through , analogous to the infinite case , a careful investigation of the asymptotic correlations .
finally , it would be worth extending this study to the case where the impurity is located further apart from the boundary of the subsystem .
if this distance grows large , the effective central charge must return to the value of a pure system and an interesting crossover behavior might emerge .
we thank ingo peschel and michael wolf for discussions .
v.e . acknowledges financial support by the danish research council , quantop and the eu projects coquit and quevadis . s.s.g . acknowledges financial support from cqiqc
the expression for the correlation functions involves the integrals in eq .
( [ eq : corr1inf ] ) containing an oscillatory integrand . in the limit @xmath186
one could approximate it by considering the scattering phase to be a constant .
however , this method only works for slowly varying @xmath187 functions which is not fulfilled by ( [ eq : deltaqinf ] ) for @xmath188 and @xmath189 .
it will be shown , that with a suitable transformation of the integral one can extract the resonant contributions which exactly reproduce the term @xmath74 associated with the bound state . to show this , we first rewrite ( [ eq : corr1inf ] ) in the form @xmath190 where we defined @xmath191 note , that the denominator is just the product @xmath192 which is strictly positive along the integration path in ( [ eq : corr1infb ] ) but in general
has four roots on the complex plane@xcite . introducing the variable @xmath193 the integration
is now taken along an arc of the complex unit circle which can be written as the sum of the two contours shown in fig .
[ fig : intcont ] . ) .
the dashed line shows the boundary of the complex unit disk . ] after changing variables and using the symmetry under @xmath194 it takes the following form @xmath195 with the fourth order polynomial @xmath196 it is easy to show , that the polynomial has two real solutions @xmath197 which lie inside the unit disk and @xmath198 ( @xmath199 ) is related to the lower ( upper ) bound state solutions as @xmath200 .
the remaining solutions form a complex - conjugate pair @xmath201 and lie outside the unit disk .
the closed contour @xmath202 therefore only encircles the pole at @xmath198 and the residue is evaluated as @xmath203 where @xmath74 is defined under eq .
( [ eq : corrbinf ] ) . the remaining integral
is taken along the contour @xmath204 which is parametrized as @xmath205 and using the analytic continuation of the scattering phase it reads @xmath206 note , that @xmath207 varies smoothly and vanishes for @xmath208 .
for @xmath186 the integrand is rapidly decaying , thus one can approximate the phase around @xmath209 as @xmath210 . if the derivative is small , that is we are far away from the fano resonance , one can set @xmath211 and the integral ( [ eq : corr1intc2 ] ) can be trivially carried out to yield the result in eq .
( [ eq : corrfried ] ) .
the derivative term is then used to define a length scale @xmath212 which marks the regime @xmath86 where the approximation is valid .
note , that @xmath87 can only grow large if the conditions @xmath213 are satisfied .
the contour integral can also be used to prove an exact symmetry property of the correlations .
we define the particle - hole transformed functions @xmath214 and @xmath215 by performing the changes @xmath216 and @xmath217 .
now , one can show that the sum @xmath218 can be written as the integral ( [ eq : corr1intc1 ] ) over the _ complete _ unit circle which is exactly evaluated as the sum of the residues at @xmath198 and @xmath199 . using the symmetry property @xmath219
, the latter pole contributes @xmath220 which , after reordering , yields the relation @xmath221 .
\label{eq : corrsymm}\ ] ]
we present here some of the formulae which is necessary for the calculations in the semi - infinite geometry .
first , the self - energy function outside the band reads @xmath222 \label{eq : sigmapmsi}\ ] ] and the bound state energies @xmath223 and @xmath224 are obtained numerically as the roots of @xmath225 note , that @xmath226 ( @xmath227 ) exists only for parameter values fulfilling @xmath228 ( @xmath229 ) . on the other hand , using the definition of the inverse correlation length @xmath230 the second term in the parentheses in ( [ eq : sigmapmsi ] ) can be written as @xmath231 . therefore
, if the impurity lies deep in the bulk of the chain with @xmath173 one has @xmath232 and consequently the same solutions for @xmath233 as in the infinite case .
the contribution @xmath234 from the lower bound state can be evaluated as @xmath235 with the normalization factor @xmath236 the limit @xmath173 gives @xmath237 and the right hand side of ( [ eq : corrbsi ] ) becomes @xmath238 with the shifted indices @xmath239 and @xmath240 , thus recovering again the result ( [ eq : corrbinf ] ) for the infinite case . finally , one needs the expression for the retarded self - energy inside the band @xmath241 which is an oscillatory function of the position @xmath8 .
this eventually leads to the emergence of bound states in the continuum , as described in the text .
the scattering phases are defined as @xmath242 with @xmath243 and @xmath244 .
the band contributions also have to be treated separately on either sides of the impurity . on the right hand side , analogously to the infinite case , it can be expressed in the form @xmath245 with @xmath246 while the translationally invariant term @xmath247 is unmodified .
the correlations on the left hand side can not be given in that simple form and are analyzed in the text . | we study the entanglement in the ground state of a chain of free spinless fermions with a single side - coupled impurity . we find a logarithmic scaling for the entanglement entropy of a segment neighboring the impurity .
the prefactor of the logarithm varies continuously and contains an impurity contribution described by a one - parameter function , while the contribution of the unmodified boundary enters additively .
the coefficient is found explicitly by pointing out similarities with other models involving interface defects .
the proposed formula gives excellent agreement with our numerical data .
if the segment has an open boundary , one finds a rapidly oscillating subleading term in the entropy that persists in the limit of large block sizes .
the particle number fluctuation inside the subsystem is also reported .
it is analogous with the expression for the entropy scaling , however with a simpler functional form for the coefficient . |
the celebrated laser cooling processes and magneto - optical traps ( mot ) @xcite have allowed the study of many exciting topics in atomic physics . among them , much interest was devoted to the study of bose - einstein condensates @xcite , which had a profound impart in our understanding of condensed matter .
apart from this , a new trend has recently begun that rekindled the interest on basic properties of mot physics .
this is related to the increasing number of astrophysical phenomena that we can simulate and study using ultra cold atomic clouds . in particular
, we can refer to a new mechanism associated with the laser cooling process , which can lead to the formation of static and oscillating photon bubbles inside the gas @xcite .
photon bubbles have been considered in the astrophysical context @xcite where huge photon densities are required to have any significant impact on high energy particles .
moreover , kaiser _ et al _ @xcite were recently able to achieve random lasing in a cloud of ultra cold atoms under laboratory conditions .
a similar effect was first seen decades ago in stellar clouds @xcite and in some planetary atmosphere @xcite , when random lasing was first proposed to explain why certain specific emission lines in the stellar gas are more intense than theoretically expected @xcite .
finally we refer to a recent work by teras _
et al _ @xcite , where the hydrodynamic equilibrium and normal modes of cold atomic traps are investigated , combining the effects of multiple photon scattering with the thermal fluctuations inside the system , cast in the form of a polytropic equation of state .
this analysis results in a generalized lane - emden equation , describing the equilibrium density profiles of the atomic cloud , similar to that describing astrophysical fluids @xcite .
the process of multiple scattering of light , which typically becames significant for a number of atoms above @xmath0 , is responsible for the rich and complex behaviour of ultra cold atomic vapours .
this mechanism has been described , since the early stages of mots , as the principal limitation for the compressibility of the cloud @xcite . in this regime ,
the atoms in the cloud are strongly correlated due to the presence of a coulomb type long - range interaction @xcite , and the description of the system as a one component plasma becomes feasible and very fruitful @xcite . in the present paper ,
we present a significant extension of the previously mentioned work by teras _ et al _ @xcite , to include the case of a rotating cloud of ultra cold atoms . in section
ii , we begin with the derivation of the equilibrium density profiles of such systems .
numerical analysis predict the existence of a special set of solutions containing satellite rings .
such satellite rings have indeed been observed in rotating clouds since the early nineties @xcite , although their nature was not completely elucidated . in section iii
, we will investigate the stability of such systems , by computing the normal oscillation modes that can be excited in the cloud . finally , in section iv
, we discuss the validity of our model and state some conclusions .
the starting point to compute the equilibrium profiles of the ultra cold cloud of atoms confined in a mot corresponds to the setting of the fluid equations , where the collective force due to multiple scattering of photons by the atoms is included : @xmath1 @xmath2 the fluid description of the system requires some relation between the pressure @xmath3 and the atom density @xmath4 , and for that purpose we assume the existence of a generic polytropic equation of state for the mot , of the form @xmath5 as we are dealing with rotating clouds , the total force acting on an fluid element is @xmath6 , with @xmath7 , and @xmath8 the collective force determined by @xmath9 notice that in the expression for the collective force @xmath8 , the quantity @xmath10 represents the square of an effective atomic charge @xcite , where @xmath11 is the speed of light and @xmath12 the total intensity of the six cooling laser beams .
the terms @xmath13 and @xmath14 represent the emission and absorption cross sections , respectively @xcite .
the force @xmath15 includes the doppler cooling force , with an equivalent damping coefficient @xmath16 , and the trapping force , with an equivalent spring constant @xmath17 .
the difference with respect to the non rotating case is the presence of a new force term @xmath18 .
rotation in the system can easily be achieved by a slight misalignment in four of the six laser beams , and can be described by @xcite @xmath19 for this reason , we will consider from now on a cylindrically symmetric system .
assuming equilibrium conditions , @xmath20 and @xmath21 , and a - dimensioning the system as @xmath22 and @xmath23 with @xmath24 we then get , in a dimensionless form @xmath25 with @xmath26 the effective plasma frequency , and @xmath27 , where @xmath28 is the angular velocity of the fluid element .
the parameter @xmath29 gives the ratio between the rotation angular frequency and the frequency associated with the confinement trap , @xmath30 .
the first is equivalent to an expansion force , and the second one to a contraction force , whereby this constant will be related with the stability of the cloud . in particular
, the system will become unstable for @xmath31 , as it will become clear in section * iii * , when we derive the frequency of the allowed oscillation modes . for non rotating clouds we would simply have @xmath32 .
( [ eq : lane0 ] ) therefore depends on the rotation state of the system .
we can now realize that , by redefining the parameter @xmath33 as @xmath34 and introducing a redefined plasma frequency as @xmath35 we get @xmath36 this new equation becomes independent of rotation parameter , which is incorporated now inside the definitions of @xmath33 and @xmath37 .
it remarkably implies that rotating and non rotating systems share the some kind of solutions for the equilibrium density profiles , differing only by a scale factor @xmath38 .
( [ eq : lane ] ) can be seen a generalized lane - emden equation @xcite , which was first derived for astrophysical fluids with gravitational confinement .
the corresponding density profiles will only depend on the polytropic exponent @xmath39 , and on the constant @xmath37 , which accounts for the three forces in play .
the latter corresponds to the _ ratio _ of the collective force , determined by @xmath40 , and the confining force ( @xmath41 ) , _ subtracted _ by the centripetal force due to rotation ( @xmath42 ) .
the confining force is then directly contour - posed by the centripetal one .
this ratio is sufficient to determine the density profiles . with this interpretation of @xmath37 in mind
, we can easily realize that a stable cloud of atoms can only exist if @xmath43 which implies an overall attractive force greater than the repulsive one ( multiple scattering ) .
since we are interested in associating systems with different rotation states , we realize that degenerate density solutions only exist if @xmath44 , with @xmath45 for a stable solution .
each of these sets of solutions ( for each value of the constant @xmath46 ) differ only by the scale factor @xmath47 .
let us now compute some analytical solutions of eq .
( [ eq : lane ] ) , for some important limiting cases . for traps with a small number of atoms ,
typically @xmath48 multiple scattering effects can be neglected , which allows us to take the limit @xmath49 .
( [ eq : lane ] ) then reduces to @xmath50 which takes the solution @xmath51 $ ] , or equivalently , in dimensional form @xmath52 the isothermal case , which corresponds to @xmath53 , simply yields a gaussian profile @xmath54 = n(0 ) \
, e^{-r^{2 } / 4 r^{2 } } \ , , \ ] ] which results from taking the limit @xmath55 in eq .
( [ eq : gama ] ) . ]
this limit approximately describes a trap in the temperature limited regime , corresponding to a small number of particles , where the thermal effects determine the cloud dynamics and multiple scattering effects are negligible . for a larger number of atoms , as typical of present state - of - the - art traps , this no longer applies . in the opposite case of a large number of trapped atoms , the collective force dominates over temperature effects ,
and we can take the limit @xmath56 in eq .
( [ eq : lane ] ) , giving rise to the solution @xmath57 or , in dimensional form @xmath58 where @xmath59 is the cloud radius and @xmath60 is the heaviside function .
this is usually referred to as the _ water - bag _ profile .
the formation of stable orbital modes in the density profiles is one of the most interesting features observed in rotating clouds of atoms @xcite .
it was widely believed that the existence of such modes was intrinsically related with the rotation of the cloud @xcite .
the investigation presented here indicates that this is not entirely true , since rotating and non rotating clouds share the same density solutions . in fact ,
numerical integration of the generalized lane - emden equation ( [ eq : lane ] ) , reveals a special set of solutions for @xmath61 .
these appear as central cores with rings of atoms around , as depicted in figure [ fig : orbital ] . .
the black line corresponds to @xmath62 , and for the blue and red dashed lines we have @xmath63 and @xmath64 respectively .
the differential equations were integrated using an explicit runge - kutta method , of fourth order . ] in these solutions , there is a central core of atoms surrounded by a ring , typically with higher densities than the central region .
as we increase the number and density of atoms in the system or , equivalently , higher rotation angular velocities - higher @xmath65 - the central core density increases .
this behaviour has been experimentally observed , where a single ring structure can turn into a ring with a central core , for increasing intensity of the laser beams ( capturing more atoms in the trap ) or by increasing the laser misalignments , making the cloud to rotate faster @xcite . in our picture
this corresponds to an increase of @xmath65 .
the novelty of the present analysis is the description of the system in terms of an atom density function , @xmath66 , which is the experimentally accessible quantity , making possible the direct comparison of theory and experiments .
although in references @xcite we can find a theoretical explanation for the rings , it is based on single atom trajectories , not allowing for a detailed comparison between theory and experiments .
since we relate the atom density profiles with the equation of state of the cloud , the results presented in this paper will allow , for the first time , and using already available absorption imaging techniques , to measure the equation of state of the system . in the next section , along with the stability analysis of the solutions , we will also introduce an alternative way of measuring the equation of state , through the normal modes of oscillation in the system .
in fact , the density profiles for @xmath61 present higher order orbital modes , with two or more distinguishable rings around the central core in spatially extended solutions . as the system is limited by the width of the laser beams and strongly depends on the number of trapped atoms , only a central core should be observed , eventually with one or two rings .
experimentally , the transition from one to two rings has been observed @xcite when the number of trapped atoms increases . in figure
[ fig : density ] we show a numerical solution where a second ring of atoms is included . and
two rings are displayed , which should be possible for a larger number of trapped atoms . ]
we now evaluate the nature of localized oscillations , or normal modes of the system , using a perturbative analysis on the above fluid description . it will be useful to notice that the collective force can be derived from a potential , @xmath67 , such that @xmath68 .
this is a coulomb type of potential , due to the existence of an effective electric charge for the neutral atoms , and resulting from the exchange of scattered photons between nearby atoms .
it can be seen as a mean - field potential associated with multiple photon scattering .
we now introduce small perturbations in the equilibrium quantities , labelled with the subscript @xmath69 , @xmath70 with the perturbations varying as @xmath71 .
this prescription allows us to linearize the set of fluid equations , neglecting quadratic terms of the form @xmath72 , which yields @xmath73 \\ + & \frac{1}{m } \boldsymbol{\nabla } \cdot \left [ n_{0}(r ) \boldsymbol{\nabla } \delta \phi_{c } \right ] \
, , \end{split}\ ] ] @xmath74 where we have used @xmath75 , from the polytropic equation of state .
it is worth noticing that this equation for the perturbation @xmath76 is formally identical to that derived for non rotating clouds @xcite .
this is to be expected , since we only consider radial perturbations , of the form @xmath77 , and @xmath78 .
the existence of a finite angular velocity is put in play by the equilibrium density profiles @xmath79 from which the previous equation depends .
recall that , due to the geometry of the problem , we have used cylindrical coordinates . as a last remark ,
notice that the term @xmath80 in the previous equation corresponds to a damping of the modes , which is to be expected since it results from the cooling doppler force , which is a viscous force , @xmath81 .
in fact , along with the mode damping , this term also introduces a small correction in the mode frequency .
but , as it turns out , in typical experimental conditions , the frequency associated with this term is small in comparison with the trapping frequency @xmath30 , and the plasma frequency @xmath82 , for what we can neglect this correction from now on @xcite . in order to combine equations
( [ eq : norm1 ] ) and ( [ eq : llap ] ) , we can now introduce the auxiliary quantity @xmath46 , defined as @xmath83 .
making proper substitutions , the linearized equations can finally be put together , resulting in a single expression @xmath84 \eta ( r ) = -\frac{\gamma c_{\gamma}}{m } \left [ r n_{0 } ^{\gamma - 1 } ( r ) \frac{d}{d r } \left ( \frac{1}{r } \frac{d \eta } { dr } \right ) \right ] \ , , \ ] ] with @xmath85 .
we have then reduced the problem of finding the normal modes of oscillation to an eigenvalue problem . despite the linear form of the differential equation
this is a non - trivial problem , and general solutions involve numerical simulations .
for that reason we will only examine some limiting cases . for small clouds , typically with @xmath48 trapped atoms , the effects of multiple scattering
can be neglected , and we can set @xmath86 , or equivalently @xmath87 in eq .
( [ eq : norm1 ] ) , which simplifies to @xmath88 = 0 \ , .\ ] ] replacing the equilibrium density profile @xmath89 , given by eq .
( [ eq : gama ] ) , into eq .
( [ eq : tlr ] ) and writing the result in terms of the adimensional variable @xmath90 , as defined before , we obtain @xmath91 = 0 \ , , \ ] ] where we performed another change of variables , namely @xmath92 .
for this differential equation we can try out solutions in the form of a power series @xmath93 where we can distinguish between even ( @xmath94 ) and odd ( @xmath95 ) contributions , because they lead to slightly different types of oscillations . for the lowest radial modes , @xmath96 , the solutions correspond to surface excitations @xcite , and the even perturbations solutions , @xmath97 ,
correspond to breathing modes . deriving the ansatz ( [ eq : ansatz ] ) , and replacing in eq .
( [ eq : tlr1 ] ) , we obtain a recursive expression for the coefficients @xmath98 , and the allowed modes of the system are @xmath99 as mentioned before , setting @xmath96 we get the surface modes @xmath100 breathing modes , @xmath97 , are also possible even in small traps , with a spectrum given by @xmath101 similar results are found in non - rotating spherical clouds @xcite . as anticipated in section
* ii * , the cloud becomes unstable for @xmath31 , when the centrifugal force due to rotation is higher than the confining force .
these results , along with those from the previous section , introduce the possibility of _ measuring _ the equation of state for the ultra cold gas , simply determined by the polytropic exponent @xmath39 .
in fact , experimental techniques allow for very precise measurements of oscillation frequencies , and these results relate the polytropic exponent @xmath39 with the modes frequencies and the angular velocity of the cloud .
in contrast with the temperature limited regime for small traps , for a large number of trapped atoms , the system is dominated by the collective effects , and we can neglect the temperature effects by setting @xmath102 in eq .
( [ eq : eq1 ] ) , yielding @xmath103 introducing the water - bag solution for the equilibrium density profile given by eq .
( [ eq : wb ] ) , we obtain @xmath104 again , and as expected , we have an instability for @xmath31 , for the same reasons as before
. remember that the multiple scattering regime corresponds to the limit @xmath105 with @xmath106 , which implies @xmath107 .
we then have a breathing mode in the system at the plasma frequency @xmath82 .
the latter result is formally analogous to that well - known in plasma physics , where uncompressional monopole oscillation of a plasma cloud takes place at the classical plasma frequency .
the same solution is also found for neutral atoms with spherical symmetry @xcite .
these normal modes are oscillations of the cloud , which should not be confused with the hybrid sound waves discussed in ref . @xcite .
in contrast with the present normal modes , which are global oscillations associated with the existence of boundaries , the hybrid sound waves are elementary excitations of the gas , which can also exist in an infinite medium .
the corresponding frequencies are different , and can therefore be identified experimentally .
we shall , however , notice that the solution ( [ eq : plasma_mode ] ) is not unique . a simple manipulation of eq . ( [ eq : norm1 ] ) and ( [ eq : llap ] ) and introducing the equilibrium water - bag profile yields @xmath108 = 0 \ , , \ ] ] with @xmath109 , the effective dielectric function .
this equation holds for the internal region of the cloud , while in the outside we simply have @xmath110 , @xmath111 .
this is quite appropriate for the water - bag density profile .
the breathing mode obtained earlier can be derived from this alternative formulation by setting the particular solution to eq .
( [ eq : eqp ] ) as @xmath112 , which implies @xmath113 as before .
the general solution to eq .
( [ eq : eqp ] ) is given by a simple power series in spherical coordinates . in cylindrical coordinates ,
the solution is more complicate but it can be given in terms of the bessel functions .
we consider solutions valid in the two different regions , the inside region of the cloud , @xmath114 , and the outside region , @xmath115 , where @xmath59 stands for the cloud radius .
the resulting solutions of the above laplace equation then read @xcite @xmath116 and @xmath117 where @xmath118 and @xmath119 are the modified bessel functions , and @xmath120 and @xmath121 are constants .
we have also included the possibility of a perturbation propagating in the @xmath122 direction with wavevector @xmath123 , different from the radial oscillations introduced by eq .
( [ eq : oscillation ] ) .
these solutions have to satisfy regular continuity conditions at the surface of the cloud ( @xmath124 ) , such that @xmath125 @xmath126 the first continuity condition implies @xmath127 , and the second condition yields the allowed modes as @xmath128^{-1/2 } \ , .\ ] ] in figure [ fig : dispersao ] , we represent the dispersion relation given by eq .
( [ eq : omegap ] ) for different values of @xmath95 .
notice that a similar result can be found in the context of a cylindrical pore surface - plasmon modes @xcite . .
] we easily realize that , in the long wavelength limit , @xmath129 , we recover the breathing mode derived earlier at @xmath130 , for @xmath97 . in this same limit ,
all the remaining modes ( @xmath131 ) collapse in a single frequency at @xmath132 , characteristic of a cylindrical system .
it is worth mentioning that for a spherical system , the same problem also yields a breathing mode at @xmath133 and , solving the laplace equation in spherical coordinates @xcite , we would get @xmath134 which collapses into our second breathing mode in the limit @xmath135 , where we get @xmath136 .
notice that by reformulating the problem in terms of an effective dielectric function - see eq .
( [ eq : eqp ] ) - we introduced the possibility for propagating modes in the boundless @xmath122 direction , and determined by radial boundary conditions , in eq .
( [ eq : bound1 ] ) and eq .
( [ eq : bound2 ] ) .
these modes have some similarities with the plasma waves considered in @xcite , which take the form @xmath137 ( @xmath138 is the sound speed ) . in the low wavenumber limit , the dispersion relation of eq .
( [ eq : omegap ] ) , and figure ( [ fig : dispersao ] ) , takes the same form of these plasma waves , @xmath139 , where @xmath140 is some constant for each @xmath95 solution , clarifying plasma like nature of these oscillations .
in contrast , in the high wavenumber limit , the radial boundary conditions would considerably alter the dispersive relation , as indicated above .
such modes could be also be excited in bound extended systems , if the size in the @xmath122 direction is large compared with the typical wavelength @xmath141 . as mentioned before , a cylindrical geometry can experimentally be achieved by lowering the intensity of the laser cooling beams along the rotation axis , allowing for the cloud to expand and to take a cigar - like shape .
kinetic effects are not expected to play an important role here .
in fact , a kinetic description is useful when considering the resonant exchange of energy between atoms moving with velocities near the phase velocity of phonons , @xmath142 .
this is not relevant for normal modes , which can be understood as phonons with infinite wavelength , @xmath143 or @xmath144 .
in this work we have examined the equilibrium profiles and the normal oscillating modes of rotating clouds of ultra cold atoms confined in a mot .
the investigation relies on the assumption that a polytropic equation of state for the ultra cold gas is satisfied , which phenomenologically models a large class of mot s , including the state - of - the - art ones , with a large number of atoms - in fact water - bag density profiles have been experimentally observed since the early stages of large mots @xcite .
an experimental validation of this equation of state , which only depends on the polytropic exponent @xmath39 , can then be performed through the measurement of the density profiles of the atomic cloud , or by determining the spectrum of normal modes . employing direct and absorption imaging techniques ,
this can be done in a very precise manner and would , for the first time , allow for the measurement of an equation of state of the ultra cold matter inside the trap , which constitutes one of the novelties of the present results .
on the other hand , we have shown that this simple polytropic model can account for the observed stable orbital modes , that up so far were thought to be intrinsically related with the cloud rotation @xcite .
the profiles with orbital modes correspond to a special set of solutions with @xmath61 .
the measurement of the normal modes provides a nice way of validating these results .
the model employed here is valid for cylindrically symmetric clouds , which should be a good approximation in the equatorial plane of a spherical one . in order to determine the width of the satellite rings in spherical clouds
, we would however need to develop a spherical model . in the general case , and specially for rotating clouds , the appropriate set of coordinates lies somewhere in between cylindrical and spherical .
for this reason we are now trying to figure out how to employ an elliptical coordinate system that in the non - rotating case would reduce to the spherical geometry ( zero eccentricity ) . this would be a more general and appropriate description of the cloud .
these different aspects of the problem will be addressed in a future work .
also related to these features , and following the annunciated new trend of studying astrophysical related phenomena in the laboratory , the results present here , as well as those in @xcite , allow for an investigation of the equilibrium and dynamical properties of the lane - emden equation , establishing another interesting connection between the communities of ultra - cold matter and astrophysics
. in particular we can refer to the recent observation of a new class of variable stars @xcite .
these rapidly rotating main - sequence stars were not expected to pulsate , or to have any other physical characteristics that would lead to periodic luminosity variations , according to current theories .
one of the hypothesis being considered is that fast rotation could alter the internal conditions of a star , enough to sustain stellar pulsations .
but there is currently no stellar model that can predict whether pulsation can be sustained in very fast rotating stars .
the ability to achieve conditions in laboratory , namely with ultra - cold atomic gases , with similar dynamics to stellar and astrophysical fluids , can prove being important to understand the properties of these objects .
all the similarities between astrophysical and cold atoms systems comes from the concept of radiation pressure , which determines the dynamics of both systems . in astrophysics
its the huge photon densities that leads to this radiation pressure . in cold atoms ,
the low densities are compensated by the high resonances leading to working regimes dictated by the radiation pressure - multiple scattering regime .
in fact , it may be possible to tune the experimental parameters in order to achieve working conditions closer to those found in astrophysical scenarios . | the multiple scattering of light in a gas of ultra cold atoms is responsible for many exciting features observed in magneto - optical traps including the collective behaviour forced by a coulomb like potential .
this field also induces plasma like phenomena in the cloud which allows the treatment of the system as a one component trapped plasma . with a fluid description and casting the thermodynamical behaviour in the form of a polytropic equation of state we investigate the equilibrium profiles of rotating clouds and its dependence on the experiment characteristics .
numerical solutions predict the formation of stable orbital modes both in rotating and non rotating clouds .
we also investigate the normal modes on such rotating systems .
the results obtained allow for the measurement of the equation of state for the cold gas . |
anderson s spin - wave theory ( swt ) provides a good description for the low temperature properties of many two - dimensional ( 2d ) and three - dimensional ( 3d ) quantum antiferromagnetic systems on a bipartite lattice @xcite , including quantum corrections to the classical nel ground state with two alternating sublattices and the doublet transverse spin - wave excitations .
the quantum antiferromagnetic systems on a triangular lattice , however , are the prototypal system with the frustrated spin alignments in the classical picture and have been under intensive study over the last few decades for fundamentally different type of ground states @xcite .
it is now firmly established by various methods @xcite , including a swt based one three - sublattices @xcite , that the ground state of the antierromagnetic heisenberg model has the long - ranged noncollinear order of the @xmath2 magnetic three - sublattice structure with three transverse , gapless spin - wave excitations .
most isotropic antiferromagnets in one - dimension ( 1d ) with low quantum spin numbers do not show nel - like long - ranged order in the ground state due to the strong quantum fluctuations .
the low - lying excitation states are also different from the 2d and 3d counterparts . in particular , by the exact solutions using the bethe ansatz @xcite , the low - lying excitation states of the 1d spin-1/2 heisenberg model have been shown corresponding to the spin-1/2 object ( spinons ) where the spin - wave - like excited states are the triplet states of spinons which always appear in pairs @xcite , contrast to the doublet excitation states by swt ; the excitation states of spin-1 heisenberg model for the linear chain , including the longitudinal one , have an excitation gap above the singlet ground state , first predicted by haldane @xcite .
these theoretical predictions have later been confirmed in the antiferromagnetic compound kcuf@xmath3 for spin-1/2 chains @xcite and csnicl@xmath3 for spin-1 chains @xcite by neutron - scattering experiments .
one interesting remaining question is whether or not there exist longitudinal excitations in quantum antiferromagnetic systems with long - ranged classical order in low temperature , as such modes will represent the oscillations in the magnitude of the long - ranged order parameter .
the answer is affirmative .
there is now ample evidence of the longitudinal excitation states in various quasi-1d structures with the nel - like long - ranged order at low temperature , including the hexagonal @xmath4-type antiferromagnets with both spin quantum number @xmath5 ( csnicl@xmath6 and rbnicl@xmath6 ) @xcite and @xmath7 ( csmni@xmath6 ) @xcite and the tetragonal structure of kcuf@xmath6 with @xmath8 @xcite .
more recently , a longitudinal mode was also observed in the dimerized antiferromagnetic compound tlcucl@xmath6 under pressure with a long - ranged nel order @xcite . to our knowledge ,
no observation of longitudinal modes in any 2d or quasi-2d antiferromagnets has been reported yet .
clearly , such longitudinal modes , which correspond to the oscillations in the magnitude of the magnetic order parameter , are beyond the usual swt which predicts only the transverse spin - wave excitations , usually referred to as quasiparticle magnons in the antiferromagnetic systems .
there are several theoretical investigations in these longitudinal modes using the field theory approach , such as a simplified version of haldane s theory for the spin-1 systems @xcite or the sine - gordon theory for the spin-1/2 systems @xcite , and both treating the inter - chain couplings as perturbations .
a phenomenological field theory approach focusing on the spin frustrations of the hexagonal lattice of the @xmath4-type antiferromagnetic systems has also been made @xcite .
we recently proposed a microscopic many - body theory based on the magnon - density waves for the longitudinal excitations of spin-@xmath9 quantum antiferromagnetic systems , using the original bipartite spin - lattice hamiltonians @xcite .
the basic physics in our analysis follows feynmann s theory on the low - lying excited states of the helium-4 superfluid @xcite : the longitudinal excitation states in a quantum antiferromagnet with a nel - like order are identified as the collective modes of the magnon - density waves , which represent the fluctuations in the long - range order and are supported by the interactions between magnons ; these longitudinal excitation states are constructed by the magnon density operator @xmath10 in contrast to the transverse spin - flip operator @xmath11 of the magnon states @xcite .
our numerical results @xcite for the energy gap values at the magnetic wave vector are in good agreement with the experiments for the energy gap observed in the tetragonal structure of kcuf@xmath6 with @xmath8 @xcite .
we hope that more experimental results for the energy spectra at other wavevectors will be available for comparison . in this article , we extend our microscopic approach to study the longitudinal modes in quantum antiferromagnets on a triangular lattice where the magnetic order is noncollinear hence there are possible more than one longitudinal modes .
we employ the approximate swt ground state in our calculations and find that one of the longitudinal modes is gapless for the isotropic models in the thermodynamic limit , due to the slow divergence of the structure factor , but any finite size effect or small anisotropy will induce a large energy gap at the magnetic wavevectors when compared with the counterpart of the transverse spin - wave spectra , similar to the collinear square lattice model .
we organize this article as follows . for completeness
, we briefly outline the main results of spin - wave theory for the triangular lattice model in sec .
2 , using the one - boson approach .
we then apply our microscopic theory for the longitudinal excitations in sec .
3 , using the approximated ground state from swt .
we find that a large energy gap can be induced by a very tiny anisotropy , similar to the square lattice model .
we also discuss the possible other longitudinal mode of the triangular lattice and the possible extension of our calculations to the more realistic models in the quasi-1d systems such as those hexagonal @xmath4-type antiferromagnets mentioned above in the last section .
the classical ground state of the antiferromagnetic heisenberg model on a triangular lattice consists of three alternating sublattices with spins on each sublattice align at an angle of @xmath12 to the other two sublattices .
the distance between the two nearest - neighbor spins of the same sublattice is @xmath13 of the lattice spacing which is taken as unity in this article .
the spin-@xmath9 heisenberg hamiltonian is given by @xmath14 where @xmath15 is the coupling parameter and the sum on @xmath16 runs over all the nearest - neighbor pairs of the triangular lattice once . following singh and huse @xcite and miyake @xcite
, it is convenient to transform the hamiltonian of eq . by rotating the quantum projection axis of the spins along the classical direction in the @xmath17-plane at the six @xmath18-sublattices surrounding the one @xmath19-sublattice .
this transformation leads to the following rotated spin operators , @xmath20 for all @xmath18 sites , where @xmath21 is the magnetic ordering wavevector at the corner of the hexagonal brillouin zone of the triangular lattice ( see fig .
after the rotation of eq . , the spin - wave theory can be formulated in terms of only one set of bosons , rather than three sets originally employed @xcite .
the hamiltonian operator of eq .
after this transformation is given by @xmath22(s_i^xs_j^x+s_i^zs_j^z)+\delta s_i^ys_j^y\nonumber\\ & + \sin[\mathbf q\cdot(\mathbf r_i-\mathbf r_j ) ] ( s_i^zs_j^x - s_i^xs_j^z)\big],\end{aligned}\ ] ] where we have also introduced an anisotropy parameter @xmath23 along the @xmath24-axis .
we will see later that even a very tiny anisotropy , such as @xmath25 , will induce a large energy gap for the longitudinal excitation spectrum at the magnetic wavevectors @xmath26 when compared with the counterpart of the spin - wave spectra . using the conventional holstein - primakoff transformations , the hamiltonian of eq .
can be expressed in terms of boson operators @xmath27 and @xmath28 in a series in power of @xmath9 .
the classical result for the ground - state energy is given by the @xmath29 term .
the linear terms ( i.e. , linear in @xmath27 and @xmath30 ) cancel each other out .
the quadratic terms are in @xmath31 and are retained in the linear swt . the cubic terms ( in @xmath32 ) and the quartic terms ( in @xmath33 ) have been treated as perturbations for the higher - order corrections to the linear swt @xcite .
therefore , the hamiltonian of the linear swt is given by , keeping only the @xmath29 and @xmath31 terms and after a fourier transofrmation for the boson operators with the fourier component operators @xmath34 and @xmath35 , @xmath36,\ ] ] where @xmath37 and @xmath38 are defined by @xmath39 respectively and @xmath40 is defined as usual by @xmath41 with the summation over the nearest - neighbor index @xmath42 and the coordination number @xmath43 for the triangular lattice .
the hamiltonian @xmath44 is diagonalized by the canonical bogoliubov transformation , @xmath45 , @xmath46 where , @xmath47 , is the spin - wave excitation spectrum with the dimensionless spectrum @xmath48 given by @xmath49 we plot the first brillouin zone ( bz ) of the triangular lattice and the dimensionless energy spectrum @xmath48 of eq . in fig . 1 .
this spin - wave spectrum has three zero modes , one at the center of the zone @xmath50 ) , and the other two at the corners of the bz , the @xmath26 ordering wavevectors .
the magnitude of the spin - wave velocity near these points are different , with values of @xmath51 and @xmath52 respectively .
the sublattice magnetization @xmath53 is given by @xmath54 with the magnon density @xmath55 defined as the ground - state expectation value of boson number operator @xmath56 with value @xmath57 for the isotropic model in the linear swt .
this magnon density of @xmath57 per lattice site is larger than the corresponding number of @xmath58 for the square lattice model , indicating the larger quantum correction in the triangular lattice model due to the frustration .
furthermore , we notice that the three sublattice magnetization are noncollinear with @xmath12 between their directions , hence we may have two longitudinal modes in the triangular lattice model , one corresponding to the @xmath59-component magnon - density fluctuations in our rotated spin basis discussed above , the other corresponding to the perpendicular @xmath60-component magnon - density fluctuations . in the following section , we will discuss the longitudinal excitations in the @xmath59-component magnon - density fluctuations , similar to the collinear square lattice case where there is only one longitudinal mode and leave the discussion of the other possible longitudinal mode to the last section .
as mentioned earlier , the longitudinal excitations in an antiferromagnetic system with a nel - like long - ranged order correspond to the fluctuations in the order parameter . we have proposed a microscopic approach based on the magnon - density waves and have applied to bipartite antiferromagnetic systems where the magnetic order is collinear @xcite . here
we extend our analysis to triangular lattice models using the one - boson approach for the approximation of the ground state and investigate the longitudinal mode of the @xmath59-component magnon - density fluctuations .
briefly , the excitation state of the @xmath59-component magnon - density waves is constructed by applying the magnon - density operator @xmath61 to the ground state @xmath62 , @xmath63 where the density operator @xmath61 , instead of the spin - raising and lowering operators @xmath11 for the transverse spin - wave excitations ( i.e. magnons ) , is given by @xmath10 operator as @xmath64 where the index @xmath65 runs over all lattice sites and the condition @xmath66 ensures that the excited state @xmath67 is orthogonal to the ground state . we notice that in our rotated spin basis of eq .
( 2 ) , the three sublattice magnetizations all point in the same @xmath59-direction . the longitudinal excitation spectrum
is then given by @xmath68 where @xmath69 is given by the ground - state expectation value of a double commutator as @xmath70\rangle_g,\ ] ] and the state normalization integral @xmath71 is the structure factor of the lattice model @xmath72 in eqs . and , the notation @xmath73 indicates the ground - state expectation .
it is straightforward to derive the double commutator for triangular lattice by employing the one - sublattice hamiltonian of eq .
after the rotation to obtain the following general expression , @xmath74 \big[\langle s_l^zs_{l+\rho}^x\rangle_g-\langle s_l^xs_{l+\rho}^z\rangle_g\big]\big].\end{aligned}\ ] ] now we apply the approximation of the linear swt for the ground state in the expectations @xmath73 and obtain @xmath75,\ ] ] where the transverse correlation functions @xmath76 and @xmath77 are defined as @xmath78 both independence of index @xmath65 due to the lattice translational symmetry .
their fourier transformations are obtained as , using the approximation of the linear swt for the ground state in the one - boson approach discussed in sec . 2 , @xmath79 and where @xmath37 and @xmath38 are as given by eqs . .
we obtain the numerical results at the isotropic point @xmath80 as @xmath81 and @xmath82 for all the six nearest neighbors .
as can be seen , @xmath69 is dominated by @xmath83 . within the same approximation of the linear swt for the ground state , the structure factor @xmath71 of eq .
is obtained as @xmath84 where @xmath85 is as given by eq . , @xmath55 is the magnon density as discussed before and @xmath86 we notice that the structure factor of eq .
involves the double magnon spectrum function @xmath87 in the integrals , indicating some effects of the interactions between magnons has been included .
we then obtain the longitudinal excitation spectrum @xmath88 of eq . using the approximations of eqs . and .
by numerical evaluation , we notice that this spectrum of the longitudinal mode is the gapless in the thermodynamic limit , approaching zero @xmath89 at both @xmath90 and @xmath91 .
detailed numerical analysis shows that the gapless spectrum is due to the slow , logarithmic divergence in both the second and third terms in the structure factor @xmath71 of eq . .
more specifically , near @xmath92 and @xmath26 , we find that @xmath93 , and thus the excitation spectrum @xmath94 as @xmath95 or @xmath96 , but with different coefficients for the @xmath92 point from that of the @xmath26 points .
this later feature is reminiscent of the different spin - wave velocities at these points as discussed in the previous section .
the major difference between the longitudinal and transverse modes near @xmath92 and @xmath26 points will become apparent once we introduce the anisotropy as discussed in the followings .
the logarithmic behaviors of the structure factor and of the energy spectrum of triangular lattice model is similar to that of the square lattice model studied earlier @xcite , where we state that the spectrum is quasi - gapped as any finite finite size effect or anisotropy will induce a large energy gap when compared with the counterparts of spin - wave spectrum .
here we consider the effect of anisotropy in the @xmath24-axis as given by eq . .
for a small value of @xmath25 , we obtain energy gap value of @xmath97 at the ordering wavevector @xmath26 , compared with the much smaller gap value of @xmath98 of the corresponding spin - wave spectrum .
more specifically , we find that the longitudinal energy gap value is proportional to @xmath99 $ ] , in contrast to the spin - wave gap which is proportional to @xmath100 , when @xmath101 .
we notice that the spectrum is still gapless at @xmath92 due to the fact that the anisotropy @xmath102 is introduced in the @xmath24-component in the hamiltonian of eq . while the spins are ordered along the rotated @xmath59-axis . in order to make further comparison between the longitudinal mode and the transverse spin - waves mode , we plot both the spectra with @xmath25 in fig . [ 6 ] along the path @xmath103 of the bz .
the different gap values for the longitudinal and transverse mode at both @xmath104 and @xmath105 points can be clearly seen .
both the longitudinal and spin - wave spectra at @xmath0 are still gapless where @xmath106 for any value of @xmath23 .
furthermore , there are slight different peak values at for the longitudinal spectrum but for the spin - wave spectrum the peak values are nearly equal .
this is again similar to the case of the square lattice model discussed earlier @xcite , indicating some effects of the interaction between magnons in the longitudinal mode . of the longitudinal mode together with spin - wave excitation spectrum @xmath107 along ( @xmath108 ) of the bz with an anisotropy @xmath25 .
the gap value at @xmath104 and @xmath105 points is @xmath97 for the longitudinal mode , and @xmath98 for the transverse spin - wave modes.,title="fig : " ] + we now turn our attention to the difference in the excitation spectra between the triangular and square lattice models .
this can be easily demonstrated by the linear swt for both models . due to the frustrations in the triangular lattice ,
there are three sublattice magnetizations with @xmath12 different orientation to one another rather than the two sublattices with complete opposite orientation ( @xmath109 ) in square lattice .
the anisotropy parameter @xmath110 is introduced in the @xmath24 components of the hamiltonian of eq . rather than in the usual the ordered direction of the @xmath59 components for the square lattice model , hence the spin - wave spectra of the triangular anisotropic model are still gapless at the @xmath92 point .
furthermore , the anisotropy parameter @xmath102 inside the square root formula of the spin - wave spectrum of eq . is linear rather than the usual quadratic form for the bipartite models such as the square lattice model .
all these features of the triangular lattice model contribute to the slightly softer nature of the longitudinal model than that of the square lattice model .
in particular , the energy gap values for the longitudinal mode is about @xmath97 at @xmath105 and @xmath111 at @xmath92 for the triangular lattice model with the anisotropy @xmath25 as given before , comparing with the bigger energy gap values of @xmath112 at @xmath113 and @xmath114 at @xmath92 for the square lattice model at the similar value of the anisotropy @xmath115 which enters the hamiltonian in the ordered @xmath59 component @xcite .
more significantly , due to the fact that antiferromagnetic order is noncollinear in triangular lattice , in additional to the above discussed longitudinal model corresponding to the @xmath59-component order parameter fluctuations , we may have another longitudinal mode corresponding to the perpendicular @xmath60-axis fluctuations .
we leave further discussion to the following section .
in this paper , we have applied our microscopic analysis based on the magnon - density waves to study the longitudinal excitations of the @xmath59-component order parameter fluctuations for a triangular lattice antiferromagnets , with the results similar to those of the square lattice model . in particular ,
the longitudinal spectrum for the isotropic model has zero modes at the zone center and the magnetic wavevectors due to the slow , logarithmic divergence of the structure factor . also similar to the square lattice model , any finite size effect or anisotropy
will induce large energy gaps at the magnetic wavevectors when compared with the counterparts of the transverse spin - wave excitations .
furthermore , we find the longitudinal modes of the triangular lattice models are in general softer than that of the square lattice model due to the frustrations in the triangular lattice model . due to the fact that the magnetic order in the triangular lattice is noncollinear
, there may exist another longitudinal mode , corresponding to the order parameter fluctuations in the perpendicular @xmath60 direction . instead of the density operator of eq .
, we use the operator @xmath116 , for the density fluctuations in the x - direction .
we will report the results for this longitudinal mode elsewhere .
one remaining question is the intrinsic lifetime of the longitudinal modes for triangular lattice model . as the results in sec .
3 demonstrate that the energy of the longitudinal mode is always higher than that of the transverse spin - wave excitations , we expect the longitudinal mode in the pure triangular lattice model may be unstable against decaying into two or more magnons , as emphasized in ref
. 28 and in a more recent discussion @xcite .
one way to find the stable longitudinal modes is to consider some frustrated models in triangular lattices near a quantum critical point where the sublattice magnetization is much reduced @xcite .
a more realistic way is to consider some quasi-1d hexagonal systems as mentioned in sec .
i. it is therefore particularly interesting to extend our present analysis to the hexagonal @xmath4-type antiferromagnets with both spin quantum number @xmath5 ( csnicl@xmath6 and rbnicl@xmath6 ) @xcite and @xmath7 ( csmni@xmath6 ) @xcite , where the longitudinal energy gaps were first observed at the magnetic wavevectors .
the spin lattice structure on the basal planes of these systems are triangular as that discussed here .
the field theory approach proposed by afflect based on haldane s theory of spin-1 ( or integer spin quantum number ) chain with five fitting parameters is able to explain many features observed , but there are still some disagreements particularly for the data away from the minimum energy gap at the antiferromagnetic wavevectors @xcite .
we believe that the longitudinal excitations of the quasi-1d antiferromagnets is more general than the extension of haldane s theory of integer spin quantum chains and that our microscopic analysis based on the magnon - density waves as presented here may provide a more general description of such longitudinal modes without any fitting parameter other than those in the model hamiltonian .
j. richter , j. schulenburg , and a. honecker , quantum magnetism in two dimensions : from semi - classical nel order to magnetic disorder , in _ quantum magnetism _ , edited by u. schollwck , j. richter , d. farnell , and r. bishop , volume 645 of _ lecture notes in physics _
, pp . 85153 , springer berlin / heidelberg , 2004 , 10.1007/bfb0119592 . | we study the longitudinal excitations of quantum antiferromagnets on a triangular lattice by a recently proposed microscopic many - body approach based on magnon - density waves .
we calculate the full longitudinal excitation spectra of the antiferromagnetic heisenberg model for a general spin quantum number in the isotropic limit .
similar to the square lattice model , we find that , at the center of the first hexagonal brillouin zone @xmath0 and at the magnetic ordering wavevectors @xmath1 $ ] , the excitation spectra become gapless in the thermodynamic limit , due to the slow , logarithmic divergence of the structure factor .
however , these longitudinal modes on two - dimensional models may be considered as quasi - gapped , as any finite - size effect or small anisotropy will induce a large energy gap , when compared with the counterpart of the transverse spin - wave excitations .
we also discuss a possible second longitudinal mode in the triangular lattice model due to the noncollinear nature of its magnetic order .
+ + pacs numbers : 75.10.jm , 75.30.ds , 75.50.ee |
rrrrrrr & & & & + & & & & & & + 104 & 0.05 & 0.06 & -0.71 & -0.70 & -0.78 & 14.05 @xmath12 0.05 + 288 & 0.03 & 0.04 & -1.40 & -1.07 & -1.14 & 15.40 @xmath12 0.05 + 362 & 0.05 & 0.06 & -1.33 & -1.15 & -1.09 & 15.51 @xmath12 0.05 + 1261 & 0.01 & 0.01 & -1.32 & & -1.08 & 16.68 @xmath12 0.05 + 1851 & 0.02 & 0.03 & -1.23 & & -1.03 & 16.18 @xmath12 0.05 + 1904 & 0.01 & 0.01 & -1.67 & -1.37 & -1.37 & 16.15 @xmath12 0.05 + 3201 & 0.21 & 0.27 & -1.53 & -1.23 & -1.24 & 14.75 @xmath12 0.05 + 4590 & 0.04 & 0.05 & -2.11 & -1.99 & -2.00 & 15.75 @xmath12 0.10 + 4833 & 0.33 & 0.42 & -1.92 & -1.58 & -1.71 & 15.70 @xmath12 0.10 + 5272 & 0.01 & 0.01 & -1.66 & & -1.33 & 15.58 @xmath12 0.05 + 5466 & 0.00 & 0.00 & -2.22 & & -2.13 & 16.60 @xmath12 0.05 + 5897 & 0.08 & 0.10 & -1.93 & -1.59 & -1.73 & 16.30 @xmath12 0.10 + 5904 & 0.03 & 0.04 & -1.38 & -1.11 & -1.12 & 15.00 @xmath12 0.05 + 6093 & 0.18 & 0.23 & -1.75 & & -1.47 & 16.25 @xmath12 0.05 + 6171 & 0.33 & 0.42 & -1.09 & & -0.95 & 15.65 @xmath12 0.05 + 6205 & 0.02 & 0.03 & -1.63 & -1.39 & -1.33 & 14.95 @xmath12 0.10 + 6218 & 0.19 & 0.24 & -1.40 & & -1.14 & 14.70 @xmath12 0.10 + 6254 & 0.28 & 0.36 & -1.55 & -1.41 & -1.25 & 15.05 @xmath12 0.10 + 6341 & 0.02 & 0.03 & -2.24 & & -2.10 & 15.20 @xmath12 0.10 + 6352 & 0.21 & 0.27 & -0.50 & -0.64 & -0.70 & 15.25 @xmath12 0.05 + 6362 & 0.09 & 0.12 & -1.18 & -0.96 & -0.99 & 15.35 @xmath12 0.05 + 6397 & 0.18 & 0.23 & -1.94 & -1.82 & -1.76 & 12.95 @xmath12 0.10 + 6541 & 0.12 & 0.15 & -1.79 & & -1.53 & 15.40 @xmath12 0.10 + 6637 & 0.17 & 0.22 & -0.72 & & -0.78 & 15.95 @xmath12 0.05 + 6656 & 0.34 & 0.44 & -1.75 & & -1.41 & 14.25 @xmath12 0.10 + 6681 & 0.07 & 0.09 & -1.64 & & -1.35 & 15.70 @xmath12 0.05 + 6723 & 0.05 & 0.06 & -1.12 & & -0.96 & 15.45 @xmath12 0.05 + 6752 & 0.04 & 0.05 & -1.54 & -1.42 & -1.24 & 13.80 @xmath12 0.10 + 6779 & 0.20 & 0.26 & -1.94 & & -1.61 & 16.30 @xmath12 0.05 + 6809 & 0.07 & 0.09 & -1.80 & & -1.54 & 14.45 @xmath12 0.10 + 7078 & 0.09 & 0.12 & -2.13 & -2.12 & -2.02 & 15.90 @xmath12 0.05 + in very recent times , new determinations of galactic globular cluster ( ggc ) metallicities have provided us with new homogeneous @xmath13\textrm { } $ ] scales .
in particular , carretta & gratton ( @xcite ; cg ) obtained metallicities from high resolution spectroscopy for 24 ggcs , with an internal uncertainty of 0.06 dex . for an even larger sample of 71 ggcs ,
metallicities have been obtained by rutledge et al .
( @xcite ; rhs97 ) based on spectroscopy of the caii infrared triplet .
the equivalent widths of the caii triplet have been calibrated by rhs97 on both the cg scale and the older zinn & west ( @xcite ; zw ) scale .
the compilation by rhs97 is by far the most homogeneous one which is currently available . in the same period
, we have been building the largest homogeneous @xmath0 photometric sample of galactic globular clusters ( ggc ) based on ccd imaging carried out both with northern ( isaac newton group , ing ) and southern ( eso ) telescopes ( rosenberg et al .
@xcite , @xcite ) .
the main purpose of the project is to establish the relative age ranking of the clusters , based on the methods outlined in saviane et al .
( @xcite , @xcite ; srp97 , srp99 ) and buonanno et al .
( @xcite ; b98 ) .
the results of this investigation are presented in rosenberg et al .
( @xcite ; rspa99 ) . here
suffice it to say that for a set of clusters we obtained @xmath14 vs. @xmath15 color - magnitude diagrams ( cmd ) , which cover a magnitude range that goes from a few mags below the turnoff ( to ) up to the tip of the red giant branch ( rgb ) . at this point
both a spectroscopic and photometric homogeneous databases are available : the purpose of this study is to exploit them to perform a thorough analysis of the morphology of the rgb as a function of the cluster s metallicity . as a first step , we want to obtain a new improved calibration of a few classical photometric metallicity indices .
secondly , we want to provide to the community a self - consistent , * analytic , * family of giant branches , which can be used in the analysis of old stellar populations in external galaxies .
photometric indices have been widely used in the past to estimate the mean metallicities of those stellar systems where direct determinations of their metal content are not feasible . in particular , they are used to obtain @xmath13\textrm { } $ ] values for the farthest globulars and for those resolved galaxies of the local group where a significant pop ii is present ( e.g. the dwarf spheroidal galaxies ) .
the calibration of @xmath0 indices is particularly important , since with comparable exposure times , deeper and more accurate photometry can be obtained for the cool , low - mass stars in these broad bands than in @xmath16 .
moreover , our huge cmd database allows a test of the new cg scale on a large basis : we are able to compare the relations obtained for both the old zw and new scale , and check which one allows to rank ggcs in the most accurate way . indeed , the most recent calibration of the @xmath0 indices ( carretta & bragaglia @xcite ) is based on just 8 clusters .
a reliable metallicity ranking of ggc giant branches also allows studies that go beyond a simple determination of the _ mean _ metallicity of a stellar population . as an illustration
, we may recall the recent investigation of the halo metallicity distribution function ( mdf ) of ngc 5128 ( harris et al .
@xcite ) , which was based on the fiducial gc lines obtained by da costa & armandroff ( @xcite , hereafter da90 ) .
these studies can be made more straightforward by providing a suitable analytic representation of the rgb family of ggcs .
indeed , assuming that most of the ggcs share a common age ( e.g. rosenberg et al .
@xcite ) , one expects that there should exist a `` universal '' function of @xmath17\ } $ ] able to map any @xmath18 $ ] coordinate pair into the corresponding metallicity ( provided that an independent estimate of the distance and extinction of the star are available ) .
we will show here that such relatively simple mono - parametric function can actually be obtained , and that this progress is made possible thanks to the homogeneity of both our data set and analysis . in order to enforce a proper use of our calibrations ,
we must clearly state that , in principle , the present relations are valid only for rigorously old stellar populations ( i.e. for stars as old as the bulk of galactic globulars ) . at fixed abundance ,
giant branches are somewhat bluer for younger ages ( e.g. bertelli et al .
moreover , in real stellar systems agb stars are also present on the blue side of the rgb ( cf .
[ f : pars - partb ] ) .
both effects must be taken into account when dealing with lg galaxies , since they could lead to systematic effects in both the mean abundances and the abundance distributions ( e.g. saviane et al .
@xcite ) . the observational sample , on which this investigation is based , is presented in sect . [ s : sample ] .
[ s : indices ] is devoted to the set of indices which are to be calibrated .
they are defined in sect [ s : defindices ] .
the reliability of our sample is tested in sect .
[ s : checks ] , where we demonstrate that our methodology produces a set of well - correlated indices . in sect .
[ s : newda90 ] we show that , once a distance scale is assumed for the ggcs , our whole set of rgbs can be approximated by a _
single _ analytic function , which depends on the metallicity alone .
this finding allows a new and easier way to determine the distances and mean metallicities of the galaxies of the local group , extending the methods of da costa & armandroff ( @xcite ) , and lee et al .
( @xcite ) .
the metallicity indices are calibrated in sect .
[ s : calibrations ] , where analytic relations are provided both for the zw and for the cg scales . using these indices , we are able to test our analytic rgb family in sect [ s : testfits ] .
our conclusions are in sect .
[ s : conclusioni ] .
clusters have been observed with the eso / dutch 0.9 m telescope at la silla , and at the rgo / jkt 1 m telescope in la palma .
this database comprises @xmath19 of the ggc whose distance modulus is @xmath20 .
the zero - point uncertainties of our calibrations are @xmath21 mag for each band .
three clusters were observed both with the southern and the northern telescopes , thus providing a consistency check of the calibrations : no systematic differences were found , at the level of accuracy of the zero - points .
a detailed description of the observations and reduction procedures will be given in forthcoming papers ( rosenberg et al .
@xcite , @xcite ) presenting the single clusters .
a subsample of this database was used for the present investigation .
we retained those clusters whose cmd satisfied a few criteria : ( a ) the hb level could be well determined ; ( b ) the rgb was not heavily contaminated by foreground / background contamination ; and ( c ) the rgb was well defined up to the tip .
this subsample largely overlaps that used for the age investigation , but a few clusters whose to position could not be measured , are nevertheless useful for the metallicity indices definition .
conversely , in a few cases the lower rgb could be used for the color measurements , while the upper branch was too scarcely defined for a reliable definition of the fiducial line .
two of the cmds that were used are shown in figs .
[ f : pars - partea ] ( ngc 1851 ) and [ f : pars - partb ] ( ngc 104 ) , and they illustrate the good quality of the data . the dataset of clusters used in this paper is listed in table [ t : the - sample ] . from left to right , the columns contain the ngc number , the reddening both in @xmath22 and @xmath15 , the metallicity according to three different scales , and the apparent magnitude of the horizontal branch ( hb ) .
the @xmath23 values were taken from the harris ( @xcite ) on - line table .
the @xmath15 reddenings were obtained by assuming that @xmath24 ( dean et al .
the values of the metallicity were taken from rhs97 : they represent the equivalent widths of the caii infrared triplet , calibrated either onto the zinn & west ( @xcite ) scale ( zw column ) or the carretta & gratton ( @xcite ) scale ( rhs97 column ) .
moreover , the original carretta & gratton metallicities ( cg column ) are also given for the clusters comprised in their sample .
the hb level was found in different ways for clusters of different metallicity .
for the the metal rich and metal intermediate clusters , a magnitude distribution of the hb stars was obtained , and the mode of the distribution was taken .
where the hb was too scarcely populated , a horizontal line was fitted through the data .
the blue tail of the metal poorest clusters does not reach the horizontal part of the branch : in that case , a fiducial hb was fitted to the tail , and the magnitude of the horizontal part was taken as the reference level .
the fiducial branch was defined by taking a cluster having a bimodal hb color distribution ( ngc 1851 , cf .
[ f : pars - partea ] ) and then extending its hb both to the red and to the blue by `` appending '' clusters being more and more metal rich and metal poor , respectively .
the details of this procedure , as well as the errors associated to the @xmath25 in table [ t : the - sample ] , are discussed in rspa99 . for ngc 1851 , @xmath26 was adopted ( dashed line in fig .
[ f : pars - partea ] ) , and this value is just @xmath27 mag brighter than the value found by walker ( @xcite ) and saviane et al .
( @xcite ) . based on this observational sample ,
a set of metallicity indices were measured on the rgbs of the clusters . in the next section ,
the indices are defined and the measurement procedures are described .
consistency checks are also performed .
the metallicity indices calibrated in this study are represented and defined in fig . [
f : pars - partea ] and fig . [ f : pars - partb ] .
the figures represent the cmd of ngc 1851 and ngc 104 in different color - magnitude planes , and the crosses mark the position of the rgb points used in the measurement of the indices . the left panel of fig .
[ f : pars - partea ] shows the apparent colors and magnitudes for ngc 1851 : the inclined line helps to identify the first index , @xmath1 .
this was defined , in the @xmath28 plane , by hartwick ( @xcite ) as the slope of the line connecting two points on the rgb : the first one at the level of the hb , and the second one 2.5 mag brighter .
we use the same definition for the @xmath4 plane here ; however , in order to be able to use our metal richest clusters , we redefined @xmath1 by measuring the second rgb point mag brighter than the hb . since @xmath1 is measured on the apparent cmd , it is independent both from the reddening and the distance modulus .
the right panel of the same figure , shows the apparent @xmath14 magnitude vs. the de - reddened @xmath29 color . in this panel ,
four other indices are identified , i.e. @xmath30 , @xmath2 , @xmath6 , and @xmath3 .
the first one is the rgb color at the level of the hb , and the other three measure the magnitude difference between the hb and the rgb at a fixed color @xmath31 , 1.2 and 1.4 mag .
the former index was originally defined by sandage & smith ( @xcite ) and the latter one by sandage & wallerstein ( @xcite ) , in the @xmath32 plane .
the other two indices , @xmath2 and @xmath6 , are introduced later to measure the metal richest gcs .
these indices require an independent color excess determination . finally , fig .
[ f : pars - partb ] shows the cmd of ngc 104 ( 47 tuc ) in the absolute @xmath33 plane : the adopted distance modulus , @xmath34 , was obtained by correcting the apparent luminosity of the hb according to lee et al .
( @xcite ; cf .
[ s : calibrations ] ) . by comparison ,
harris catalog reports @xmath35 .
two other indices are represented in the figure : @xmath7 and @xmath8 .
they are defined as the rgb color at a fixed absolute @xmath36 magnitude of @xmath37 ( da costa & armandroff @xcite ) or @xmath38 ( lee et al .
the latter index was also discussed by armandroff et al .
( @xcite ) , and a calibration formula was given in caldwell et al .
( @xcite ) .
this is based on the da90 clusters plus m5 and ngc 362 from lloyd evans ( @xcite ) .
since these two indices are defined on the bright part of the rgb , they can be measured even for the farthest objects of the local group ( lg ) . due to the fast luminosity evolution of the stars on the upper rgb ,
this part of the branch was typically under - sampled by the early small - size ccds , so no wide application of these indices has been made for galactic globulars . however ,
this is of no concern for galaxy - size stellar systems .
it will be shown in sect .
[ s : calibrations ] that good accuracies can be obtained even for gcs , provided that the analytic function of eq .
( [ e : iperbole ] ) is used .
lrrrrrrr & & & & + & & & & & & & + 104 & 0.99 & 4.13 & 0.78 & 1.27 & 1.87 & 1.94 & 1.57 + 288 & 0.95 & 6.39 & 1.25 & 1.75 & 2.36 & 1.51 & 1.35 + 362 & 0.90 & 7.28 & 1.67 & 2.09 & 2.57 & 1.45 & 1.28 + 1261 & 0.91 & 7.77 & 1.62 & 2.13 & 2.73 & 1.39 & 1.25 + 1851 & 0.97 & 7.41 & 1.23 & 1.82 & 2.55 & 1.45 & 1.31 + 1904 & 0.94 & 8.56 & 1.58 & 2.14 & 2.83 & 1.35 & 1.24 + 3201 & 0.99 & 8.72 & 1.19 & 1.91 & 2.71 & 1.39 & 1.27 + 4590 & 0.91 & 9.98 & 1.90 & 2.52 & 3.25 & 1.24 & 1.16 + 4833 & 0.92 & 9.25 & 1.80 & 2.36 & 3.12 & 1.28 & 1.19 + 5272 & 0.91 & 7.60 & 1.66 & 2.13 & 2.81 & 1.36 & 1.24 + 5466 & 0.91 & 9.85 & 1.93 & 2.50 & 3.18 & 1.24 & 1.16 + 5897 & 0.97 & 8.73 & 1.34 & 2.00 & 2.79 & 1.35 & 1.25 + 5904 & 0.93 & 6.91 & 1.41 & 1.91 & 2.55 & 1.44 & 1.30 + 6093 & 0.93 & 8.02 & 1.58 & 2.12 & 2.91 & 1.34 & 1.24 + 6171 & 1.07 & 5.66 & 0.31 & 1.09 & 1.93 & 1.67 & 1.49 + 6205 & 0.89 & 7.70 & 1.75 & 2.20 & 2.75 & 1.37 & 1.23 + 6218 & 0.95 & 7.09 & 1.34 & 1.88 & 2.51 & 1.46 & 1.31 + 6254 & 0.90 & 8.25 & 1.75 & 2.29 & 3.17 & 1.30 & 1.21 + 6341 & 0.88 & 9.92 & 2.15 & 2.69 & 3.40 & 1.21 & 1.13 + 6352 & 1.12 & 3.11 & -0.16 & 0.52 & 1.30 & 1.99 & 1.75 + 6362 & 0.93 & 5.84 & 1.31 & 1.76 & 2.32 & 1.55 & 1.37 + 6397 & 0.89 & 9.45 & 1.98 & 2.49 & 3.12 & 1.26 & 1.16 + 6541 & 1.01 & 8.59 & 1.03 & 1.77 & 2.67 & 1.39 & 1.29 + 6637 & 0.96 & 4.39 & 0.96 & 1.41 & 1.97 & 1.82 & 1.53 + 6656 & 0.86 & 10.32 & 2.27 & 2.69 & 2.96 & 1.24 & 1.12 + 6681 & 0.95 & 7.54 & 1.35 & 1.92 & 2.76 & 1.37 & 1.27 + 6723 & 1.01 & 6.02 & 0.76 & 1.38 & 2.18 & 1.55 & 1.41 + 6752 & 0.99 & 7.16 & 1.08 & 1.69 & 2.46 & 1.45 & 1.33 + 6779 & 0.94 & 8.74 & 1.60 & 2.18 & 2.94 & 1.32 & 1.22 + 6809 & 0.93 & 9.38 & 1.72 & 2.29 & 2.87 & 1.32 & 1.20 + 7078 & 0.88 & 9.82 & 2.10 & 2.62 & 3.27 & 1.23 & 1.14 + colors and magnitudes were measured on a fiducial rgb , which has been found by least - square fitting an analytic function to the observed branch .
after some experimenting , it was found that the best solution is to use the following relation : @xmath39 where @xmath40 and @xmath41 represent the color and the magnitude , respectively .
one can see from figs .
[ f : pars - partea ] and [ f : pars - partb ] that the function is indeed able to represent the giant branch over the typical metallicity range of globular clusters .
moreover , it is shown in sect .
[ s : newda90 ] that , when the cmds are corrected for distance and reddening , the four coefficients can be parametrized as a function of [ fe / h ] , so that one is able to reproduce the rgb of each cluster , using just one parameter : the metallicity . at any rate ,
the indices were measured on the original loci , so that an independent check of the goodness of the generalized hyperbolae can be made , by comparison of the measured vs. predicted indices .
all the indices values that have been measured are reported in table [ t : the - indices ] . in this table ,
the cluster ngc number is given in column 1 ; the following columns list , from left to right , @xmath5 , @xmath1 , @xmath2 , @xmath6 , @xmath3 , and finally the rgb color measured at @xmath42 and @xmath43 .
the lee et al .
( 1990 ) distance scale was used to compute the last two indices ( cf .
[ s : calibrations ] ) . before discussing the indices as metallicity indicators
, we checked their internal consistency .
we will show in sect .
[ s : calibrations ] that the index @xmath1 is the most accurate one , as expected , since it does not require reddening and distance corrections .
the rest of the indices are therefore plotted vs. @xmath1 in figs . [ f : intcheck2 ] and [ f : intcheck1 ] , and we expect that most of the scatter will be in the vertical direction .
second order polynomials were fitted to the distributions , and the _ rms _ of the fit was computed for each index . in order to intercompare the different indices , a relative uncertainty has been computed by dividing the _ rms _ by the central value of each parameter ( this value is identified by a dotted line in each figure ) . in this way , the scatter of the metal index @xmath44 is @xmath45 , 0.02 , 0.04 , 0.06 , 0.12 , and 0.26 , for the indices @xmath46 , @xmath8 , @xmath5 , @xmath3 , @xmath6 , and @xmath2 , respectively .
these values confirm the visual impression of the figures , that @xmath47 and @xmath48 are the lowest dispersion indices , followed by @xmath5 and @xmath3 .
the indices will be calibrated in terms of [ fe / h ] in sect .
[ s : calibrations ] ; however , before moving to this section , we want to present a new way to provide `` standard '' ggc branches in the @xmath10 plane , along the lines of the classical da costa & armandroff ( @xcite ) study . using this family of rgb branches ,
we are able to make predictions on the trend of the already defined indices with metallicity ; these trends can thus be compared to the observed ones , and therefore provide a further test of the reliability of our rgb family ( cf .
sect [ s : testfits ] ) .
lcccccc cluster & @xmath49 & @xmath50 & @xmath51 & [ fe / h]@xmath52 & [ fe / h]@xmath53 & [ fe / h]@xmath54 + ngc 104 & 14.05 & 0.050 & 0.064 & @xmath55 & @xmath56 & @xmath57 + ngc 5904 & 15.00 & 0.023 & 0.029 & @xmath58 & @xmath59 & @xmath60 + ngc 288 & 15.40 & 0.036 & 0.046 & @xmath61 & @xmath62 & @xmath63 + ngc 6205 & 14.95 & 0.000 & 0.000 & @xmath64 & @xmath65 & @xmath66 + ngc 5272 & 15.58 & 0.002 & 0.003 & @xmath67 & @xmath65 & + ngc 6341 & 15.20 & 0.010 & 0.013 & @xmath68 & @xmath69 & + rrrrrrrrrrrr & & & & & + @xmath70 & @xmath71 & @xmath70 & @xmath71 & @xmath70 & @xmath71 & @xmath70 & @xmath71 & @xmath70 & @xmath71 & @xmath70 & @xmath71 + 13.782 & 0.978 & 15.359 & 0.914 & 15.492 & 0.852 & 14.725 & 0.926 & 14.645 & 0.867 & 15.060 & 0.852 + 13.604 & 0.994 & 15.107 & 0.939 & 15.151 & 0.874 & 14.457 & 0.942 & 14.322 & 0.890 & 14.720 & 0.872 + 13.443 & 1.008 & 14.849 & 0.960 & 14.789 & 0.892 & 14.221 & 0.961 & 14.033 & 0.909 & 14.395 & 0.894 + 13.317 & 1.021 & 14.593 & 0.984 & 14.597 & 0.910 & 14.040 & 0.978 & 13.788 & 0.929 & 14.079 & 0.916 + 13.075 & 1.045 & 14.342 & 0.999 & 14.359 & 0.929 & 13.878 & 0.994 & 13.595 & 0.944 & 13.789 & 0.937 + 12.862 & 1.070 & 14.109 & 1.018 & 14.143 & 0.955 & 13.700 & 1.009 & 13.381 & 0.966 & 13.533 & 0.953 + 12.619 & 1.101 & 13.881 & 1.036 & 13.796 & 0.990 & 13.456 & 1.032 & 13.170 & 0.984 & 13.303 & 0.974 + 12.346 & 1.136 & 13.649 & 1.062 & 13.517 & 1.021 & 13.190 & 1.061 & 12.984 & 1.005 & 13.082 & 0.994 + 12.035 & 1.185 & 13.376 & 1.090 & 13.265 & 1.046 & 12.916 & 1.091 & 12.832 & 1.019 & 12.850 & 1.020 + 11.761 & 1.231 & 13.058 & 1.132 & 13.005 & 1.076 & 12.655 & 1.122 & 12.631 & 1.045 & 12.611 & 1.039 + 11.461 & 1.281 & 12.766 & 1.173 & 12.759 & 1.110 & 12.419 & 1.154 & 12.363 & 1.077 & 12.351 & 1.067 + 11.101 & 1.362 & 12.534 & 1.210 & 12.519 & 1.148 & 12.231 & 1.183 & 12.118 & 1.111 & 12.075 & 1.102 + 10.696 & 1.459 & 12.380 & 1.233 & 12.302 & 1.187 & 12.073 & 1.212 & 11.945 & 1.138 & 11.771 & 1.148 + 10.330 & 1.600 & 12.163 & 1.268 & 12.109 & 1.227 & 11.868 & 1.254 & 11.844 & 1.156 & 11.492 & 1.195 + 10.062 & 1.720 & 11.928 & 1.317 & 11.878 & 1.275 & 11.615 & 1.305 & 11.707 & 1.178 & 11.284 & 1.233 + 9.877 & 1.856 & 11.617 & 1.411 & 11.741 & 1.310 & 11.335 & 1.371 & 11.571 & 1.204 & 11.154 & 1.265 + 9.706 & 2.019 & 11.427 & 1.483 & 11.575 & 1.344 & 11.116 & 1.422 & 11.395 & 1.252 & 11.008 & 1.295 + 9.602 & 2.148 & & & 11.494 & 1.377 & 10.902 & 1.489 & 11.141 & 1.312 & 10.854 & 1.320 + 9.524 & 2.315 & & & 11.330 & 1.406 & 10.652 & 1.585 & 10.870 & 1.376 & 10.709 & 1.351 + 9.573 & 2.576 & & & 11.240 & 1.447 & 10.457 & 1.680 & 10.643 & 1.444 & & + 9.619 & 2.768 & & & 11.112 & 1.488 & 10.343 & 1.742 & 10.552 & 1.492 & & + & & & & 11.078 & 1.528 & & & & & & + & & & & 11.047 & 1.546 & & & & & & + da costa & armandroff ( @xcite ) presented in tabular form the fiducial ggc branches of 6 globulars , covering the metallicity range @xmath72\leq -0.71 $ ] .
the rgbs were corrected to the absolute @xmath10 plane using the apparent @xmath14 magnitude of the hb , and adopting the lee et al .
( @xcite ) theoretical hb luminosity .
since the da90 study , these branches have been widely used for stellar population studies in the local group .
based on these rgbs , in particular , a method to determine both the distance and mean metallicity of an old stellar population was presented by lee et al .
( @xcite ) . both da90 and lee et al .
( 1993 ) provided a relation between the metallicity [ fe / h ] and the color of the rgb at a fixed absolute @xmath36 magnitude ( @xmath42 and @xmath43 , respectively ) , and recently a new relation for @xmath73 has also been obtained by caldwell et al .
( @xcite ) .
once the distance of the population is known ( e.g. via the luminosity of the rgb tip ) , then an estimate of its _ mean _ metallicity can be obtained using one of the calibrations .
it is assumed that the age of the population is comparable to that of the ggcs , and that the age spread is negligible compared to the metallicity spread ( rspa99 ) . in such case ,
one expects that any rgb star s position in the absolute cmd is determined just by its metallicity , and that a better statistical determination of the population s metal content would be obtained by converting the color of _ each _ star into a [ fe / h ] value .
with this idea in mind , in the following sections we will show that this is indeed possible , at least for the bright / most sensitive part of the giant branch .
we found that a relatively simple _ continuous _ function can be defined in the @xmath9 $ ] space , and that this function can be used to transform the rgb from the @xmath10 plane to the @xmath74},m_{i } $ ] plane . in order to obtain this function , we first selected a subsample of clusters with suitable characteristics , so that a reference rgb grid can be constructed .
the fiducial branches for each cluster were then determined in an objective way , and they were corrected to the absolute @xmath75 plane . in this plane , the analytic function was fitted to the rgb grid .
these operations are described in the following sections .
the clusters that were used for the definition of the fiducial rgbs are listed in table [ t : fiducialgc ] , in order of increasing metallicity .
the table reports the cluster name , and some of the parameters listed in table [ t : the - sample ] are repeated here for ease of use .
the values of the reddening were in some cases changed by a few thousandth magnitudes ( i.e. well within the typical uncertainties on @xmath23 ) , to obtain a sequence of fiducial lines that move from bluer to redder colors as [ fe / h ] increases , and again the corresponding @xmath76 values were obtained assuming that @xmath24 ( dean et al .
@xcite ) . indeed , due to the homogeneity of our sample , we expect that if a monotonic color / metallicity sequence is not obtained , then only the uncertainties on the extinction values must be taken into account . in order to single out these clusters from the total sample ,
some key characteristics were taken into account .
in particular , we considered clusters whose rgbs are all well - defined by a statistically significant number of stars ; they have low reddening values ( @xmath77 ) ; and they cover a metallicity range that includes most of our ggcs ( @xmath78\leq -0.7 $ ] on the zw scale )
. the da90 fiducial clusters were ngc 104 , ngc 1851 , ngc 6752 , ngc 6397 , ngc 7078 and ngc 7089 ( m2 ) .
ngc 104 is the only cluster in common with the previous study , and m2 is not present in our dataset .
the other objects have been excluded from our fiducial sample since they have too large reddening values ( @xmath79 for ngc 6397 and ngc 7078 ) , or their rgbs are too scarcely populated in our cmds ( ngc 1851 and ngc 6752 ) .
nevertheless , the calibrations that we obtain for the @xmath7 and @xmath8 are in fairly good agreement with those obtained by da90 ( for the small discrepancies at the high metallicity end , cf .
[ s : da90indices1 ] and [ s : da90indices2 ] ) , and in particular with the recent caldwell et al .
( @xcite ) calibration for the @xmath73 index .
the ridge lines of our fiducial rgbs were defined according to the following procedure .
the rgb region was selected from the calibrated photometry , by excluding both hb and agb stars .
all stars bluer than the color of the rr lyr gap were removed ; agb stars were also removed by tracing a reference straight line in the cmd , and by excluding all stars blue - side of this line .
this operation was carried out in the @xmath80 plane , where the rgb curvature is less pronounced , and a straight line turns out to be adequate .
the fiducial loci were then extracted from the selected rgb samples .
the @xmath15 and @xmath36 vectors were sorted in magnitude , and bins were created containing a given number of stars . within each bin , the median color of the stars and the mean magnitude were used as estimators of the bin central color and brightness .
the number of stars within the bins was exponentially increased going from brighter to fainter magnitudes . in this way ,
( a ) one can use a small number of stars for the upper rgb , so that the color of the bin is not affected by the rgb slope , and ( b ) it is possible to take advantage of the better statistics of the rgb base .
finally , the brightest two stars of the rgb were not binned , and were left as representatives of the top branch .
after some experimenting , we found that a good rgb sampling can be obtained by taking for each bin a number of stars which is proportional to @xmath81 , where @xmath44 is an integer number .
the resulting fiducial vectors were smoothed using an average filter with a box size of 3 .
the rgb regions of the 6 clusters are shown in fig .
[ f : zooms ] , together with the fiducial lines : it can be seen that in all cases the agbs are easily disentangled from the rgbs .
the values of the fiducial points corresponding to the solid lines in fig .
[ f : zooms ] , are listed in table [ t : fidtable ] .
the fiducial branches defined in sect . [ s : fiducials ] were fitted with a parametrized family of hyperbolae .
first , the rgbs were moved into the absolute @xmath10 plane .
the distance modulus was computed from the apparent magnitude of the hb ( cf .
table [ t : fiducialgc ] ) and by assuming the common law @xmath82+b $ ] ; in order to compare our results with those of da90 , @xmath83 and @xmath84 were used , but we also obtained the same fits using more recent values as in carretta et al .
( @xcite ) , i.e. @xmath85 and @xmath86 .
the rgb was modeled with an hyperbola as in rosenberg et al .
( @xcite ) , but in this case the coefficients were taken as second order polynomials in [ fe / h ] . in other words , we parametrized the whole family of rgbs in the following way : @xmath87\ ] ] where @xmath88^{2}+k_{2}[{\rm fe / h}]+k_{3}\ ] ] @xmath89^{2}+k_{5}[{\rm fe / h}]+k_{6}\ ] ] @xmath90^{2}+k_{8}[{\rm fe / h}]+k_{9}\ ] ] @xmath91 the list of the parameters of the fits in magnitude is reported in table [ t : coeffs ] , together with the _ rms _ of the residuals around the fitting curves .
the table shows that the parameter @xmath92 does not depend on the choice of the distance scale , as expected .
even the other coefficients are little dependent on the distance scale , apart from @xmath93 .
it is affected by the zero - point of the hb luminosity - metallicity relation , and indeed there is the expected @xmath94 mag difference going from the ldz to the c99 distance scale .
one could question the choice of a constant @xmath92 , but after some training on the theoretical isochrones , we found that even allowing for a varying parameter , its value indeed scattered very little around some mean value .
this empirical result is a good one , in the sense that it allows to apply a robust linear least - square fitting method for any choice of @xmath95 , and then to search for the best value of this constant by a simple _ rms _ minimization .
we chose to fit the @xmath96\ } $ ] function , and not the @xmath97 ) $ ] function , since the latter one would be double - valued for the brightest part of the metal rich clusters rgbs .
this choice implies that our fits are not well - constrained for the vertical part of the giant branch , i.e. for magnitudes fainter than @xmath98 .
however , we show in the next section that our analytic function is good enough for the intended purpose , i.e. to obtain the [ fe / h ] of the rgb stars in far local group populations , and thus to analyze how they are distributed in metallicity .
our synthetic rgb families are plotted in figs .
[ f : fitszw ] and [ f : fitscg ] , for the ldz distance scale . in the former figure ,
the zw metallicity scale is used , while the cg scale is used in the latter one .
the figures show that the chosen functional form represents a very good approximation to the true metallicity `` distribution '' of the rg branches .
the _ rms _ values are smaller than the typical uncertainties in the distance moduli within the local group .
we further stress the excellent consistency of the empirical fiducial branches for clusters of similar metallicity .
we have two pairs of clusters whose metallicities differ by at most 0.03 dex ( depending on the scale ) : ngc 288 and ngc 5904 on the one side , and ngc 5272 and ngc 6205 on the other side .
the figures show that the fiducial line of ngc 288 is similar to that of ngc 5904 , and the ngc 5272 fiducial resembles that of ngc 6205 , further demonstrating both the homogeneity of our photometry and the reliability of the procedure that is used in defining the cluster ridge lines .
if the coefficients of the hyperbolae are taken as third order polynomials , the resulting fits are apparently better ( the _ rms _ is @xmath99 mag ) ; however , the trends of the metallicity indices show an unphysical behavior , which is a sign that further clusters , having metallicities not covered by the present set , would be needed in order to robustly constrain the analytic function . in the following section ,
the indices are calibrated in terms of metallicity , so that in sect .
[ s : testfits ] they will be used to check the reliability of our generalized fits .
.the coefficients that define the functions used to interpolate our rgbs ( see text ) ; the top header line identifies the two distance scales used , while the two metallicities are identified in the second line of the header[t : coeffs ] [ cols= " > , > , > , > , > " , ] cccccccc & d.sc . &
metallicity & @xmath100 & @xmath101 & @xmath102 & _ rms _ & fit + & & cg & -0.03 & 0.23 & -1.19 & 0.13 & 2 + & & zw & -0.004 & -0.18 & 0.08 & 0.12 & 2 + & & zw & -0.24 & 0.28 & & 0.12 & 1 + @xmath8 & ldz & cg & 0.00487 & -0.0057 & & 0.13 & @xmath103 + & & zw & -2.12 & 8.81 & -9.75 & 0.13 & 2 + & c99 & cg & 0.0045 & -0.0053 & & 0.15 & @xmath103 + & & zw & -2.05 & 8.57 & -9.61 & 0.12 & 2 + @xmath7 & ldz & cg & 0.0068 & -0.0076 & & 0.15 & @xmath103 + & & zw & -3.34 & 12.37 & -11.91 & 0.14 & 2 + & c99 & cg & 0.0065 & -0.0073 & & 0.15 & @xmath103 + & & zw & -3.233 & 12.23 & -11.96 & 0.14 & 2 + @xmath3 & & cg & -0.34 & 0.93 & -1.37 & 0.16 & 2 + & & zw & -0.063 & -0.56 & 0.41 & 0.16 & 2 + & & zw & -0.87 & 0.77 & & 0.16 & 1 + @xmath6 & & cg & -0.36 & 0.55 & -0.97 & 0.19 & 2 + & & cg & -0.69 & 0.0007 & & 0.22 & 1 + & & zw & -0.13 & -0.38 & -0.28 & 0.20 & 2 + & & zw & -0.82 & 0.06 & & 0.20 & 1 + @xmath2 & & cg & -0.30 & 0.09 & -0.81 & 0.23 & 2 + & & cg & -0.59 & -0.52 & & 0.25 & 1 + & & zw & -0.13 & -0.42 & -0.68 & 0.25 & 2 + & & zw & -0.70 & -0.56 & & 0.25 & 1 + @xmath5 & & cg & 4.25 & -5.37 & & 0.32 & 1 + & & zw & 5.25 & -6.52 & & 0.33 & 1 + in order to obtain analytic relations between the indices and the actual metallicity , our photometric parameters were compared both with the zw and the cg values .
a summary of the resulting equations is given in table [ t : rms ] .
for each index ( first column ) both linear and quadratic fits were tried , of the form : @xmath104=\alpha \cdot index + \beta $ ] and @xmath104=\alpha \cdot index^{2}+\beta \cdot index + \gamma $ ] .
the coefficients of the calibrating relation are given in the columns labelled @xmath105 , @xmath106 , and @xmath107 ; in column 7 , the _ rms _ of the residuals is also given .
in the case of the @xmath7 and @xmath8 indices , neither the linear nor the quadratic fits give satisfactory results , when the cg scale is considered .
instead , a good fit is obtained if a change of variables is performed , setting @xmath108 } $ ] , and linearly interpolating in the index ( i.e. setting @xmath109 ) .
the column 8 of table [ t : rms ] identifies the kind of fitting function that is used for each parameter / metallicity combination : the symbols `` 1 '' , `` 2 '' and `` @xmath103 '' refer to the linear , quadratic , and linear in @xmath103 fits , respectively . relations on both the cg and zw metallicity scales are given , and column 3 flags the [ fe / h ] scale that is used . in order to measure the @xmath46 and @xmath8 indices ( cf .
[ s : indices ] ) a distance scale must be adopted .
the most straightforward way is to use the observed @xmath25 ( cf .
table [ t : the - sample ] ) coupled with a suitable law for the hb absolute magnitude .
it has become customary to parameterize this magnitude as @xmath110+b $ ] , although there is no consensus on the value of the two parameters @xmath111 and @xmath112 .
the current calibrations of these two metallicity indices were obtained by da costa & armandroff ( @xcite ) and lee et al .
( @xcite ) , and they are based on the lee et al .
( @xcite ; ldz ) theoretical luminosities of the hb .
ldz gave a relation @xmath113 + 0.82 $ ] valid for @xmath114 .
as discussed in sect [ s : newda90 ] , since many current determinations of population ii distances within the local group are based on the lee et al .
( @xcite ) distance scale , and for the purpose of comparison with previous studies , we provide a calibration using the latter hb luminosity - metallicity relation .
however , in the last ten years revisions of this relation have been discussed by many authors , so we also calibrated the two indices using @xmath115 + 0.90 $ ] ( carretta et al .
@xcite ) , which is one of the most recent hb - based distance scales .
we must stress that _ metallicities on the zw scale must be used in the @xmath116 vs. _ [ fe / h ] _
relation_. indeed , cg showed that their scale is not linearly correlated to that of zw , so not even the _
@xmath116 vs. _ [ fe / h ] relation will be linear : if one wishes to use the new scale , then _ the absolute magnitude of the hb must be re - calibrated _ in a more complicated way .
the best calibrating relations are shown in figs .
[ f : cals ] to [ f : caldv14 ] . in the following sections , for each index a few remarks on the accuracy of the calibrations and comparisons with past studies
are given .
on the cg scale , the second - order fit has a residual _ rms _ of 0.12 dex in [ fe / h ] . on the zw scale ,
the linear fit is obtained with a _
rms _ of 0.12 dex .
this index can therefore be calibrated on both scales , with a comparable level of accuracy .
a parabolic fit does not improve the relation on the zw scale , since the coefficient of the quadratic term is very small ( -0.004 ) and the _ rms _ is the same .
these relations are shown in fig .
[ f : cals ] as solid lines , where the upper panel is for the zw scale , and the lower panel for the cg scale ( this layout is reproduced in all the following figures ) .
the cluster ngc 6656 ( m22 ) was excluded from the fits , and is plotted as an open circle in fig . [ f : cals ] .
it is well - known that m22 is a cluster that shows a metallicity spread , and indeed it falls outside the general trend in most of the present calibrations . the first definition of the @xmath7 index was given in da costa & armandroff ( @xcite ) , where a calibration in terms of the zw scale was also given : @xmath11=-15.16 + 17.0\ , ( v - i)_{-3}-4.9\ , ( v - i)_{-3}^{2 } $ ] . the same index ( measured on the _ absolute _ rgbs corrected with the ldz hb luminosity - metallicity relation ) is plotted , in fig .
[ f : calvi-3 ] , as a function of the metallicity on both scales , and the solid lines represent our calibrations .
the top panel shows the quadratic relation on the zw scale , whose _ rms _ is 0.14 dex .
the bottom panel of fig .
[ f : calvi-3 ] shows the relation on the cg scale . in this case
, a quadratic fit is not able to reproduce the trend of the observational data .
a better result can be obtained by making a variable change , i.e. using the variable @xmath119 } $ ] ; in this case , a linear relation is found , and its _ rms _ is 0.15 dex .
this measure of the residual scatter has been computed after transforming back to metallicity , so the reliability of the index can be compared to that of the other ones .
again , the index can be calibrated on both scales with a comparable accuracy . the dashed curve in the upper panel of fig .
[ f : calvi-3 ] shows the original relation obtained by da90 : there is a small discrepancy at the high - metallicity end , which can be explained by the different 47 tuc fiducial line that was adopted by da90 ( cf . below the discussion on @xmath8 ) . as already recalled , we checked the effect of adopting another distance scale , by repeating our measurements and fits , and adopting the c99 distance scale . for the zw metallicity scale ,
we obtain the quadratic relation whose coefficients are listed in table [ t : rms ] , and whose _ rms _ is 0.15 dex . the bottom panel of fig .
[ f : calvi-3 ] shows the relation on the cg scale .
again , a quadratic fit is not able to reproduce the trend of the observational data . making the already discussed variable substitution ,
the linear relation in @xmath103 has an _ rms _ of 0.16 dex , so the two metallicity scales yield almost comparable results .
using the same `` standard '' gc branches of da90 , lee et al .
( @xcite ) defined a new index , @xmath8 , to be used for the farthest population ii objects .
it was also calibrated in terms of the zw scale : @xmath11=-12.64 + 12.6\ , ( v - i)_{-3.5}-3.3\ , ( v - i)_{-3.5}^{2 } $ ] .
a new calibration was also given recently in caldwell et al .
( @xcite ) : [ fe / h]@xmath121 , where @xmath122 $ ] .
the index and our calibrations ( solid lines ) are plotted , in fig .
[ f : calvi-3.5 ] , on both metallicity scales .
again , the measurements were made in the absolute cmd , assuming the ldz distance scale .
our quadratic calibration vs. the zw scale has a residual _ rms _ scatter of 0.13 dex , which is the same of the linear relation on the cg metallicity vs. @xmath103 .
the lee et al .
relation ( dashed line ) predicts slightly too larger metallicities on the zw scale , for @xmath104>-1 $ ] .
this can also be interpreted as if the da90 47 tuc branch were @xmath123 mag bluer than ours . indeed , if one looks at fig .
5 of da90 , one can easily see that some weight is given to the brightest rgb star , which is brighter than the trend defined by the previous ones .
the result is a steeper branch , which also justifies the da90 slightly bluer rgb fiducial . since our metal richest point is defined by two clusters , and since the two measured parameters agree very well , we are confident that our calibration is reliable . in any case , the discrepancy between the two scales is no larger than @xmath94 dex .
it is also reassuring that the caldwell et al .
( @xcite ) relation ( pluses ) is closer to the present calibration , since the former is based on a larger set of clusters .
this might be an indication that the lee et al .
relation is actually inaccurate at the metal rich end , due to the small set of calibrating clusters .
as before , we obtained a further calibration also using the c99 @xmath116 vs. [ fe / h ] relation ; the quadratic fit on the zw scale has a residual _ rms _ scatter of 0.13 dex , while the @xmath103 variable can be fitted with a straight line , with an _ rms _ of 0.14 dex . for any @xmath126 index , the quadratic relations vs. the zw metallicity do not improve the _
rms _ and they are not plotted in the figures . the coefficients are listed in table [ t : rms ] .
the best metallicity estimates of the `` @xmath126 family '' are obtained with the @xmath3 index . the errors on @xmath11
$ ] are just slightly larger than the standard uncertainties of the spectroscopic determinations .
the solid lines of fig .
[ f : caldv14 ] show the calibrations that we obtain .
the quadratic equation on the cg scale , and the linear one on the zw scale , are obtained with residual scatters of 0.16 dex .
the rest of the indices in this family , and @xmath127 , lack the precision of the other abundance indicators .
this is due to the fact that the error on any @xmath126 index is proportional to the uncertainty on the color of the rgb ( which depends on the reddening ) , times its local slope where the reference point is measured .
since the rgb slope increases going away from the tip ( i.e. towards bluer colors ) , we expect that the scatter on the @xmath126 indices will also increase as the color of the reference point gets bluer .
indeed , table [ t : rms ] shows that in most cases the _ rms _ uncertainties are @xmath128 dex for these indices .
the residual scatter is largest for the @xmath127 index , which is the most affected by the uncertainties on the reddening .
the @xmath6 and @xmath127 parameters have been earlier calibrated , on the cg scale , by carretta & bragaglia ( @xcite ) . using their quadratic relation for @xmath6 , and both their linear and quadratic relations for @xmath127 , the corresponding _
rms _ of the residuals in metallicity are 0.21 dex and @xmath129 dex , respectively .
our new and the old calibrations are therefore compatible , within the ( albeit large ) uncertainties .
a straightforward test of our new analytic rgbs can be made by generating the same metallicity indices that have been measured on the observed rgbs , and then checking the consistency of the predicted vs. measured quantities . to this aim , for a set of discrete [ fe / h ]
values a @xmath130 vector was generated , and the combination of the two was used to compute the @xmath131 vector of the giant branch , using eqs .
( [ e : general]-[e : d ] ) . then for each branch the metallicity indices were measured as it was done for the clusters fiducials . in figs .
[ f : cals ] to [ f : caldv14 ] , the predicted indices are identified by the small open squares ( spaced by 0.1 dex ) connected by a solid line .
the best predictions are for those indices that rely on the brightest part of the rgb ( i.e. @xmath7 , @xmath8 and @xmath132 ) , while the computations are partially discrepant for those indices that rely on a point that is measured on the faint rgb .
this is easily explained by the nature of our fit : since the best match is searched for along the ordinates ( for the reasons discussed in sect .
[ s : newda90 ] ) , then it is better constrained in the upper part of the rgb , where its curvature becomes more sensitive to metallicity .
we must also stress that the metal richest cluster in the reference grid is 47 tuc ( [ fe / h]@xmath133 on the zw scale ) , whereas ngc 6352 ( [ fe / h]@xmath134 on the same scale ) is the metal richest cluster for which metallicity indices have been measured .
some of the discrepancies that are seen at the highest metallicities are therefore due to the lack of low - reddening clusters that can be used to extend the reference grid to the larger [ fe / h ] values . the mean differences between the predicted and fitted indices are , on the zw scale , around 0.03 dex for the @xmath7 and @xmath8 indices .
they are around 0.08 dex for the @xmath6 , @xmath3 , and @xmath1 indices .
they rise to @xmath94 and @xmath135 dex for the @xmath2 and @xmath5 indices .
a similar trend is seen for the comparison on the cg scale . in this case , the mean differences are @xmath99 dex for @xmath7 , @xmath8 , and @xmath1 ; they are @xmath94 dex for @xmath6 and @xmath3 ; and they are 0.12 and 0.27 for the @xmath2 and @xmath5 indices .
we can therefore conclude that , apart from the @xmath2 and @xmath5 indices , our mono - parametric rgb family gives a satisfactory reproduction of the actual changes of the rgb morphology and location , as a function of metallicity .
it is then expected that , using this approach , one can exploit the brightest @xmath136 mags of the rgb to determine the mean metallicity , and even more important , the metallicity _ distribution _ of the old stellar population of any local group galaxy . in a forthcoming paper , we will demonstrate such possibility by re - analyzing our old photometric studies of the dwarf spheroidal galaxies tucana ( saviane et al .
@xcite ) , phoenix ( held et al .
@xcite ; martnez - delgado et al .
@xcite ) , fornax ( saviane et al .
@xcite ) , lgs 3 ( aparicio et al .
@xcite ) , leo i ( gallart et al . @xcite ; held et al .
@xcite ) and ngc 185 ( martnez - delgado et al .
in this work , we have provided the first calibration of a few metallicity indices in the @xmath137 plane , namely the indices @xmath1 , @xmath2 and @xmath3 .
calibrations on both the zinn & west ( 1984 ) and carretta & gratton ( 1997 ) scales have been obtained . the metallicity indices
@xmath5 , @xmath6 , @xmath7 and @xmath8 have been also calibrated on both scales , and we have shown that our new relations are consistent with existing ones . in the case of the latter two indices ,
we have obtained the first calibration on the cg scale ; for both scales , we have also obtained the first calibration that takes into account new results on the rr lyr distances .
the accuracy of the calibrations is generally better than 0.2 dex , regardless of the metallicity scale that is used .
our results are an improvement over previous calibrations , since a new approach in the definition of the rgb is used , and since our formulae are based on the largest homogeneous photometric database of galactic globular clusters
. the availability of such database also allowed us a progress towards the definition of a standard description of the rgb morphology and location .
we were able to obtain a function in the @xmath9 $ ] space which is able to reproduce the whole set of ggc giant branches in terms of a single parameter ( the metallicity ) .
we suggest that the usage of this function will improve the current determinations of metallicity and distances within the local group , extending the methods of lee et al .
( 1993 ) .
we thank the referee , gary da costa , for helpful suggestions that improved the final presentation of the manuscript .
i.s . acknowledges the financial support of italian and spanish foreign ministries , through an ` azioni integrate / acciones integradas ' grant .
aparicio a. , gallart c. , bertelli g. , 1997 , aj 114 , 680 armandroff t.e . , da costa g.s . , caldwell n. , seitzer p. , 1993
, aj 106 , 986 bertelli g. , bressan a. , chiosi c. , fagotto f. , nasi e. , 1994 , a&as 106 , 275 buonanno r. , corsi c.e . , pulone l. , fusi pecci f. , bellazzini m. , 1998 , a&a 333 , 505 ( b98 ) caldwell n. , armandroff t.e . , da costa g.s . , seitzer p. , 1998
, aj 115 , 535 carretta e. , bragaglia a. , 1998 , a&a 329 , 937 carretta e. , gratton r. , 1997 , a&as 121 , 95 ( cg ) carretta e. , gratton r.g . , clementini g. , fusi pecci f. , 1999 , apj , in press ( c99 ) da costa g.s . , armandroff t.e . , 1990 , aj 100 , 162 ( da90 ) dean j.f . ,
warren p.r .
, cousins a.w.j . , 1978 , mnras 183 , 569 gallart c. , freedman w. , aparicio a. , bertelli g. , chiosi c. , 1999 , aj , , 118 , 2245 harris g.l.h . , harris w.e . , poole g.b . , 1999 , aj 117 , 855 harris w.e
, 1996 , aj 112 , 1487 hartwick f.d.a . , 1968 , apj 154 , 475 held e.v .
, saviane i. , momany y. , 1999a , a&a 345 , 747 held e.v .
, saviane i. , momany y. , carraro g. , 1999b , apj , in press lee m.g .
, freedman w.l . , madore b.f . , 1993 ,
apj 417 , 553 lee y.w . , demarque p. , zinn r. , 1990 , apj 350 , 155 ( ldz ) lloyd evans t. , 1983 , s. afr .
circ . 7 , 86 martnez - delgado d. , aparicio .
a. , gallart c. , 1999a , aj , 118 , 2229 martnez - delgado d. , gallart c. , aparicio a. , 1999b , aj , 118 , 862 rosenberg a. , saviane i. , piotto g. , aparicio a. , 1999a , aj , 118 , 2306 ( rspa99 ) rosenberg a. , piotto g. , saviane i. , aparicio a. , 1999b , a&as , in press rosenberg a. , aparicio a. , saviane i. , piotto g. , 1999c , a&as , submitted rutledge a.g .
, hesser j.e . , stetson p.b . , 1997 ,
pasp 109 , 907 ( rhs97 ) sandage a. , smith l.l .
, 1966 , apj 144 , 886 sandage a. , wallerstein g. , 1960 , apj 131 , 598 saviane i. , held e.v .
, bertelli g. , 1999a , a&a , in press saviane i. , held e.v . , piotto g. , 1996 , a&a 315 , 40 saviane i. , rosenberg a. , piotto g. , 1997 . in :
rood , a.renzini ( eds . ) advances in stellar evolution , cambridge university press , cambridge , p. 65
( srp97 ) saviane i. , piotto g. , fagotto f. , et al . , 1998 , a&a 333 , 479 saviane i. , rosenberg a. , piotto g. , 1999b . in : b. k. gibson , t. s. axelrod , m. e. putman ( eds . ) ,
`` the third stromlo symposium : the galactic halo '' ( srp99 ) walker a. , 1992 , pasp 104 , 1063 zinn r. , west m. , 1984 , apjs 55 , 45 ( zw ) | the purpose of this study is to carry out a thorough investigation of the changes in morphology of the red giant branch ( rgb ) of galactic globular clusters ( ggc ) as a function of metallicity , in the @xmath0 bands . to this aim ,
two key points are developed in the course of the analysis . *
( a ) * using our photometric @xmath0 database for galactic globular clusters ( the largest homogeneous data sample to date ; rosenberg et al .
@xcite ) _ we measure a complete set of metallicity indices _ , based on the morphology and position of the red - giant branch . in particular , we provide here the first calibration of the @xmath1 , @xmath2 and @xmath3 indices in the @xmath4 plane .
we show that our indices are internally consistent , and we calibrate each index in terms of metallicity , both on the zinn & west ( 1984 ) and the carretta & gratton ( 1997 ) scales .
our new calibrations of the @xmath5 , @xmath6 , @xmath7 and @xmath8 indices are consistent with existing relations . *
( b ) * using a grid of selected rgb fiducial points , _ we define a function in the @xmath9 $ ] space which is able to reproduce the whole set of ggc giant branches in terms of a single parameter _ ( the metallicity ) . as a first test , we show that the function is able to predict the correct trend of our observed indices with metallicity
. the usage of this function will improve the current determinations of metallicity and distances within the local group , since it allows to easily map @xmath10 coordinates into @xmath11,m_{i } $ ] ones . to this
aim the `` synthetic '' rgb distribution is generated both for the currently used lee et al .
( 1990 ) distance scale , and for the most recent results on the rr lyr distance scale . |
in the earliest stages of star formation the so - called class0 and class1 phases protostars are still deeply embedded in their parental molecular cloud cores .
this material absorbs almost all of the emitted radiation of the star in the optical and the near - infrared . the spectral energy distribution ( sed ) of class0 protostars peaks at about 100160@xmath2 m , the wavelength of the maximum of a modified blackbody at 3080k .
hence , direct observations of protostars have to be carried out in the far - infrared and in the ( sub-)mm wavelength range .
sub - mm and millimeter observations of some of the sources investigated here have been obtained e.g. by lefloch et al .
@xcite , ladd & hodapp @xcite , chini et al .
@xcite , gueth et al .
@xcite , motte & andr @xcite , and gueth & guilloteau @xcite . the iso satellite ( kessler et al .
@xcite ) with its phot instrument had the capacity to measure the broad - band continuum in the far - infrared .
such observations , covering the peak region of the sed of protostars , help to yield some of the major properties of these objects such as their temperature , the sub - mm slope of their sed , the optical depth and the solid angle under which they emit .
the latter two can not be disentangled due to the limited spatial resolution of the isophot instrument . with higher resolution observations ( e.g. scuba ) , however , we can independently determine the solid angle under which an object is seen and that way infer its optical depth .
these parameters , together with the distance , enable us to calculate the total ( l@xmath0 ) and sub - mm ( l@xmath3 ) luminosities of each object .
we may then decide whether an object really is of class0 or not by determining the l@xmath3/l@xmath0 ratio ( andr et al .
@xcite ) . finally , by placing the inferred values on a temperature bolometric luminosity diagram the equivalent to a hertzsprung - russell diagram for protostars ( myers et al .
@xcite ) we are able for the first time to estimate the ( model dependent ) ages and masses of these sources directly .
bipolar outflows invariably accompany class0 sources : strong inflow and outflow of material are concurrent .
we thus wish to probe how the mass outflow rate is related to the mass accretion rate onto the protostar .
the outflowing material interacts with the ambient medium through radiative shocks .
thus , the luminosity of the outflow may be correlated with some of the source properties ( e.g. the bolometric source luminosity ) , which depend on the mass accretion rate .
therefore , we measured the luminosities of the outflows of 16 class0 sources in the 10s(1 ) line of molecular hydrogen .
this is usually the strongest and easiest line to observe in near - infrared spectra of shocked molecular hydrogen , and due to the short cooling time of h@xmath4 it is a good tracer of the present interaction of the outflow with the surrounding material .
these h@xmath4 luminosities are then compared with various source properties to investigate possible correlations .
modelling of class0 protostars remains in its infancy .
schemes now exist which yield evolutionary tracks , based on relating gas accretion to the dusty envelope ( myers et al.@xcite ) and jet thrust to gas accretion ( bontemps et al .
@xcite , saraceno et al .
@xcite , smith @xcite , and andr et al .
we combine these schemes here in order to test if the simplest assumptions , such as a spherical envelope and a single accreting object , are feasible . in this paper
, we first present our far - infrared iso maps and spectroscopy , and then summarize the data analysis and how we derive temperatures and luminosities ( sect.[datareduction ] ) . in sect.[results ] , we present our results , and comment on individual objects . a discussion of age and mass determination , and the general relationship to the outflows is contained in sect.[discussion ] . a framework within which the data can be interpreted is then put forward ( sect.[evolscheme ] ) .
we used the iso satellite to obtain isophot minimaps of six class0 sources and lws full grating spectra for three of them .
all observations are listed in table[obslog ] .
clcccc observation & object & @xmath5(j2000 ) & @xmath6(j2000 ) & aot & t@xmath7[s ] + number & & & & & + 65903003&rno15fir&03 27 39&+30 13 00&pht22 & 670 + 65903004&rno15fir&03 27 39&+30 13 00&pht22 & 620 + 65903101&hh211 & 03 43 57&+32 00 49&pht22 & 670 + 65903102&hh211 & 03 43 57&+32 00 49&pht22 & 620 + 65201107&hh211 & 03 43 57&+32 00 52&lws01&2268 + 65902801&hh211 west&03 43 57&+32 01 04&lws01&3350 + 66600502&hh211 east&03 43 59&+32 00 36&lws01&2912 + 46601429&l1157 & 20 39 06&+68 02 13&lws01&3390 + 28200120&l1157 & 20 39 06&+68 02 14&lws01&1958 + 52902105&l1157 & 20 39 06&+68 02 14&pht22 & 668 + 52902106&l1157 & 20 39 06&+68 02 14&pht22 & 620 + 54301407&ic1396w & 21 26 06&+57 56 17&pht22 & 668 + 54301408&ic1396w & 21 26 06&+57 56 17&pht22 & 620 + 56300709&l1211 & 22 47 17&+62 01
58&pht22 & 670 + 56300710&l1211 & 22 47 17&+62 01 58&pht22 & 620 + 56600912&cepe south&23 03 13&+61 41 56&lws01&1888 + 56402111&cepe & 23 03 13&+61 42 27&pht22 & 670 + 56402112&cepe & 23 03 13&+61 42 27&pht22 & 620 + 56601113&cepe north&23 03 13&+61 42 59&lws01&1890 + [ tab1 ] minimaps were taken for six objects ( cepe , hh211-mm , ic1396w , l1157 , l1211 , and rno15fir ) with isophot in its pht22 mode by single pointing and moving of the telescope by one ( c100 ) or half ( c200 ) of a detector pixel .
we used four filters ( 60 , 100 , 160 , and 200@xmath2 m ) . for 60 and 100@xmath2
m , the c100 detector ( 3@xmath83 array of ge : ga ) was used to create a 5@xmath83 pixel minimap with a pixel size of 45@xmath8 46 .
the maps thus cover a field of view of 230@xmath8 135 .
for the two longer wavelengths 7@xmath83 mosaics with a pixel size of 45@xmath8 90 were obtained using the c200 detector ( 2@xmath82 array of stressed ge : ga ) , covering thus a field of view of 315@xmath8 270 . for details on the instrument and
the used astronomical observing templates ( aot ) see the iso handbook , volume v : pht the imaging photo polarimeter and lemke et al . @xcite .
the data were reduced with the isophot interactive analysis ( pia v9.1 ) software .
flux measurements in the isophot maps were carried out in two different ways : 1 ) point spread function ( psf ) photometry using psf fractions provided by laureijs @xcite was done for the c100 maps .
we do not provide psf photometry for the c200 detector since the given psf fractions by laureijs @xcite are only for the whole c200 pixel and our maps have a sampling of half a pixel in one direction .
2 ) `` aperture '' photometry was obtained for all filters of both c100 and c200 detectors .
here we attributed each pixel in the maps either to object or to background manually , then summed up both and subtracted background from object to obtain its flux . since at 60 and 100@xmath2
m , i.e. for the c100 data , we were able to do photometry with both methods , we have a means of estimating the consistency of both .
all measured fluxes , including the available iras fluxes of our objects , and the background level in the maps are provided in table[phot_fluxes ] . for three objects (
l1157 , cepe , and hh211 ) we have full grating medium - resolution lws01 scans , which cover a wavelength range from 43 to 196.9@xmath2 m with a resolving power between 150 and 300 .
see the iso handbook , volume iv : lws the long wavelength spectrometer and clegg et al .
@xcite for instruments and aot details .
we reduced the lws data using standard pipeline 7 . for deglitching and flux calibration and defringing of the spectra we employed the iso spectral analyses package ( isap 1.6a ) . for the measurement of the luminosities of the outflows in the 10s(1 ) line of molecular hydrogen at 2.122@xmath2 m near - infrared images
were taken in several observing campaigns and at various telescopes .
the complete list of all observations is provided in table[obslognir ] .
we observed the objects in two filters to distinguish between line and continuum emission .
due to the angular size of the objects , the single images had to be arranged into large mosaics .
all observing campaigns were ( re)-reduced for consistency using own software based on the iraf package dimsum .
the whole procedure includes flatfielding , cosmic ray hit removal and sky subtraction as well as re - centering and mosaicing . for a higher astrometric accuracy we used all available stars in the field for the re - centering .
the photometric calibration was achieved by the observation of faint near - infrared standards with an accuracy of 10% .
for the flux measurements we subtracted the scaled continuum image from the emission line image to measure only the flux in the 10s(1 ) line of h@xmath4 .
our images are being prepared for publication or are already published . since in this paper we will only use the integrated h@xmath4 line luminosities we do not reproduce the images here .
the objects are discussed in the following papers : rno15fir in davis et al .
@xcite and rengel et al .
@xcite ; hh211-mm in eislffel et al .
@xcite ; vla1623 , l1157 in davis and eislffel @xcite ; l1211 in froebrich and eislffel @xcite ; cepe in eislffel et al .
@xcite , smith et al .
@xcite ; l1448 region in eislffel @xcite and froebrich et al .
@xcite ; hh212 , hh24 in froebrich et al .
@xcite and eislffel et al .
@xcite ; ser - firs1 in eislffel and froebrich @xcite .
the observed broad - band continuum fluxes of our sources allow us to fit a sed to the measurements and to infer source properties ( e.g. t@xmath0 and l@xmath0 ) . to fit the sed we used eq.[graybody ] for the flux density @xmath9 of our objects .
@xmath10\,/\,\sigma \omega \ , = \ , \left(1-e^{-\tau}\right ) \ , \cdot \ , b(\lambda , t ) \label{graybody}\ ] ] @xmath11 is the planck function , @xmath12 the solid angle of the source and @xmath13 the optical depth . @xmath13 is set as @xmath14 @xmath15 is in @xmath2 m , the optical depth at 100@xmath2 m ( @xmath16 ) is a free parameter , and @xmath17 is the sub - mm slope of the sed .
the lowest @xmath18 of the fit is obtained when the solid angle of the object is determined by @xmath19 @xmath20 indicates the various used filters , @xmath21 the flux measurements in these filters , and @xmath22 the error of the measurements .
@xmath23 is calculated by @xmath24 for each filter @xmath20 separately using the filter transmission curves @xmath25 . to fit a graybody to the measured seds of each object ,
a grid of graybodys was computed ( see eq.[graybody ] ) in which we varied the three parameters @xmath26 , @xmath16 , and @xmath17 .
we varied the temperature between 15 and 80k , in steps of 0.25k , the optical depth at 100@xmath2 m from 0.09 to 40 , in logarithmic intervals of 1.5 , and the sub - mm slope from 0.0 to 3.0 , in steps of 0.1 .
these graybodys were convolved with the filter curves of the used filter bands ( see eq.[convolve ] ) .
then the solid angle @xmath12 was determined by computing the deviation of the model points from the measurements and minimising this value ( see eq.[sigmaomega ] ) .
finally the @xmath18 of the fit to the measurements was calculated ( see eq.[rms ] ; @xmath27 indicates the number of filters ) and the parameters leading to the minimal @xmath18 were selected .
@xmath28 we find that @xmath16 has almost no influence on the shape of the graybody curve , but only on the absolute flux level , which on the other hand mainly depends on @xmath12 .
thus , the values presented in table[temperatures ] are determined by fixing @xmath16 to unity .
this restriction has no influence in the deduced parameters t@xmath0 and l@xmath0 , but the given source size @xmath12 has no physical meaning . in sect.[results ] we present also graybody fits with @xmath16 as a free parameter , in case that this improves the fit significantly . if sub - mm or millimetre observations yield source sizes , the optical depth at 100@xmath2 m can be constrained .
another way to determine physical meaningful radii for the sources is to follow the assumptions of myers et al .
they adopt an optically thick envelope , a single power law dependence with the frequency of the emissivity and an envelope density proportional to @xmath29 .
taking figure2 in myers et al .
@xcite , we can determine the optical depth using @xmath30 .
we used the bolometric temperatures from table[temperatures ] to derive @xmath16 , and repeated the fit of the sed with this optical depth .
this leads to a new solid angle of the source which is the size of the protostellar envelope where @xmath16 has the correct value , according to the assumptions of myers et al .
@xcite . in first approximation
the solid angle and the optical depth are connected by @xmath31 .
thus , we can determine the radius of the envelope where @xmath16 is unity ( presented as r@xmath32 in table[temperatures ] ) .
this radius is different from the envelope sizes obtained of optically thin emission by sub - mm or millimeter measurements ( e.g. motte & andr @xcite and chini et al .
@xcite ) .
( @xmath26,@xmath12 ) plane for the graybody fit of the cepe photometry .
the fit with the measured fluxes is marked by a cross .
circles indicate fits using fluxes which deviate at most by 0.5@xmath33 , and the small dots represent fits using fluxes with a maximum deviation of 1.0@xmath33 from the measurements . ]
llllllllllllllllll object & c@xmath34 & c@xmath32 & c@xmath35 & c@xmath36 & c@xmath37 & c@xmath38 & b@xmath34 & b@xmath32 & b@xmath35 & b@xmath36 & s@xmath39 & s@xmath40 & i@xmath41 & i@xmath42 & i@xmath43 & i@xmath34 & i@xmath32 + rno15fir & 25.7 & 51.5 & 46.7 & 48.9 & 24.8 & 55.5 & 19 & 25 & 92 & 120 & 9.2@xmath44 & 1.4@xmath44 &
& 0.25@xmath45 & 3.4 & 47.1 & 93.6 + rno15@xmath46 & 3.3 & 11.9 & 7.1&10.8 & 3.7 & 10.5 & 19 & 25 & 92 & 120 & & & & 0.18@xmath45 & 4.2 & 48.8 & 82.2 + hh211-mm & 3.0 & 33.8 & 56.2 & 54.8 & 1.9 & 20.8 & 42 & 138 & 308 & 331 & 16.4@xmath44 & 3.8@xmath44 & 0.9@xmath47 & & & & + hh211firs2@xmath46 & 2.9 & 13.7 & 55.3 & 48.2 & 1.0 & 11.3 & 42 & 138 & 308 & 331 & & & & & & & + l1157 & 6.8 & 37.8 & 42.0 & 38.6 & 5.9 & 35.8 & 11 & 19 & 63 & 72 & 6.0@xmath48 & 0.9@xmath48 & 0.4@xmath48 & 0.25@xmath45&0.25 & 10.9 & 53.5 + ic1396w & 5.7 & 19.9 & 36.6 & 26.6 & 4.6 & 13.8 & 25 & 92 & 210 & 230 & & & & 0.25@xmath45 & 0.6 & 9.7 & 38.3 + ic1396wfirs2@xmath46 & 0.1 & 1.2 & 33.2 & 9.9 & 0.2 & 1.2 & 25 & 92 & 210 & 230 & & & & & & & + ic1396wfirs3@xmath46 & 0.1 & 1.2 & 4.6&17.3 & 0.2 & 2.3 & 25 & 92 & 210 & 230 & & & & & &
+ l1211 & 12.9 & 36.5 & 63.0 & 75.0 & 10.5 & 23.4 & 42 & 157 & 254 & 280 & & & 0.135@xmath49 & 2.7 & 5.7 & 19.8 & 63.6@xmath45 + l1211firs2@xmath46 & 3.4 & 17.0 & 31.6 & 46.2 & 2.2 & 10.9 & 42 & 157 & 254 & 280 & & & 0.345@xmath50 & & & & + cepe & 55.2 & 125.8 & 102.1 & 81.4 & 65.4 & 123.0 & 35 & 99 & 234 & 293 & 43.7@xmath48 & 4.1@xmath48 & 1.0@xmath48 & 0.43 & 5.8 & 61.0&112.0 + upper limit ( 2 ) measured at 1.1 mm ( mccaughrean et al .
@xcite ) ( 3 ) measured at 1.2 mm ( tafalla et al .
@xcite ) , mms4 ( 4 ) measured at 1.2 mm ( tafalla et al .
@xcite ) , superposition of mms1 , mms2 and mms3 ( 5 ) rengel et al .
@xcite ( 6 ) chini et al .
@xcite the above described method to fit the sed was applied to both , the measurements at our four isophot wavelengths and all available data from table[phot_fluxes ] ( except the iras 12 and 25@xmath2 m points ; see below ) .
the inferred object properties from the latter are listed in table[temperatures ] .
we did not find significant changes in the fit parameters between the two methods , except for l1211 and partly for rno15fir .
these differences are discussed in sect.[res_l1211 ] and [ res_rno15fir ] , respectively .
all obtained results are discussed for each object separately in sect.[results ] .
iras 12 and 25@xmath2 m points are excluded from the fit of the sed since recent works show , that these fluxes are usually far above the fit to the sed ( see e.g. chini et al .
@xcite , barsony et al .
barsony et al .
@xcite argue that this excess mid - ir emission is due to ongoing outflow / dust interactions .
another effect , leading to the detection of these extremely red sources at such short wavelength ( and especially at 12@xmath2 m ) could be the tiniest of red leaks of the iras filters .
suppose that there was a red leak of only 0.1% in the 12@xmath2 m filter at the limit of which the filter transmission curve is known over the band pass between 20 and 30@xmath2 m .
the iras detectors were still sensitive in this range .
such leak would increase the measured i@xmath42 flux of cepe , for example , by a factor of 500 .
we also note that the 12 and 25@xmath2 m iras data points , if used , for most deeply embedded sources show huge deviations from the best - fitting seds , compared to the presumed precision of the flux calibration of iras .
less red sources , like class1 or 2 objects , would hardly suffer from such leaks , because their seds are a lot less steep in the 12 and 25@xmath2 m range .
therefore , we decided not to use the iras 12 and 25@xmath2 m data points in our sed fits for our very red objects .
if one uses the measurements for the determination of the bolometric temperature including also the 12 and 25@xmath2 m iras points , nevertheless , we get slightly higher values ( about 4k ) .
since it is not clear how much of this small effect is due to filter leaks , outflow / dust interactions , or envelope emission , we do not list these values in table[temperatures ] .
our best fitting results are given in table[temperatures ] together with the fit errors .
in almost all cases it was not possible or not useful to do a fit for the newly detected objects in our maps , since they are at the edge of the map and so we are missing an unknown part of their flux .
some objects are detected only with the c200 detector since they are outside the slightly smaller maps at the c100 wavelengths .
a determination of the fit errors can not be obtained analytically .
therefore , we varied the measurements within their one sigma error box ( five equidistant values ; @xmath51 ; n = 0 , 1 , 2 ) and computed the best fitting parameters for each of the @xmath52 combinations .
this results in an area of the parameter space into which the error boxes are mapped . as an example
, we show in fig.[error ] this area in the ( @xmath26,@xmath12 ) plane for cepe .
the errors given in table[temperatures ] are read off such diagrams for each of our objects .
this procedure was applied also to the parameters l@xmath0 , l@xmath3 and t@xmath0 .
the determined graybody fits are integrated to obtain the total luminosities of the sources . by integrating only at wavelengths larger than 350@xmath2 m
we obtain the sub - mm luminosities l@xmath3 , which can be compared to the total luminosities l@xmath0 to decide whether an object is a class0 source ( andr et al .
both values , l@xmath0 and l@xmath3/l@xmath0 , are given in table[temperatures ] .
when the ratio l@xmath3/l@xmath0 exceeds 0.005 , then the object is counted as class0 .
this is equivalent to the mass ratio m@xmath53/m@xmath54 being larger than unity ( see andr et al . @xcite and references therein ) .
the given bolometric temperatures t@xmath0 are the temperatures of a blackbody with the same mean frequency as the graybody , where the mean frequency @xmath55 of an sed is determined by @xmath56 lccccccccccccc object & t[k ] & @xmath17 & @xmath12 & r@xmath32[au ] & @xmath18 & r[pc ] & t@xmath0[k ] & l@xmath0[l@xmath57 & l@xmath3/l@xmath0 & class0 + rno15fir & [email protected] & [email protected] & [email protected] & 270&1.0 & 350 & [email protected] & [email protected] & [email protected] & + hh211-mm & [email protected] & [email protected] & [email protected]&520&2.0 & 315 & [email protected] & [email protected] & [email protected] & + l1157 & [email protected] & [email protected] & [email protected] & 440&1.5 & 440 & [email protected] & [email protected] & [email protected] & + ic1396w & [email protected] & 0.3 & [email protected] & 680&1.3 & 750 & [email protected] & [email protected] & [email protected] & ?
+ l1211 & [email protected] & 0.0 & [email protected] & 280&1.4 & 725 & [email protected] & [email protected] & [email protected] & ?
+ l1211firs2@xmath46 & 26.8 & 0.0 & 1.7 & 340&1.6 & 725 & 26.9 & 16.0 & 0.100 & ?
+ cepe & [email protected] & [email protected] & [email protected] & 750&0.9 & 730 & [email protected] & 77.9@xmath5810 & [email protected] & +
our observations of cepe , hh211-mm , ic1396w , l1157 , l1211 , and rno15fir were carried out at their nominal iras positions . in our isophot maps ( shown in figs.[cepe_data][rno15fir_data ] ) we detected more objects than were actually targeted . in four cases other ( partly ) unexpected embedded objects or bright diffuse continuum emission
are found . for l1211
no object was detected at the nominal iras position , but there were two other sources discovered in the maps
. measured fluxes in all filters , including iras fluxes and sub - mm and millimeter points from the literature , are given in table[phot_fluxes ] .
discrepancies of the fluxes between psf and `` aperture '' photometry are for various reasons : first , it is a major problem to determine which pixel contributes to which object when doing `` aperture '' photometry .
a second problem is the determination of the background .
when using the psf fitting method , the background is determined automatically ( provided that the object is a point source and in the centre of a pixel ) , while for `` aperture '' photometry one has to choose background pixels . concerning the absolute calibration errors for the two detectors of 15 and 10% for the c100 and c200 detector , respectively , and an additional error of 20% due to background uncertainties , we find that both flux determination methods lead to consistent results in almost all cases .
most of our investigated objects are of class0 type according to the l@xmath3/l@xmath0 criterion .
we can not decide whether the newly discovered objects in our maps are of class0 , because they are situated at the edges of the isophot maps . due to the different sizes of the maps
we certainly underestimate their fluxes at 60 and 100@xmath2 m , which alters their derived sed in the way that they seem to be proportionally brighter at the longer wavelengths , but to an unknown extent .
the isophot and the iras fluxes at 60 and 100@xmath2 m are consistent within the errors only for cepe .
for all other objects the iras point source catalogue gives values which are a factor of about 1.8 brighter .
apart from the fact that the errors for the iras data are quite large and in some cases only upper limits are given , the main reason for the differences is that the resolution of the iras satellite was not sufficient to resolve close - by sources . only cepe and l1157 seem not to have other young objects in their immediate vicinity , and these are the two objects where the iras and isophot fluxes match the best .
cepe is a known double source ( moro - martn et al .
@xcite ) which can not be resolved by iras nor isophot . for these reasons and the still fairly large errors in the flux measurements , no investigation of the time evolution of the fluxes of these young sources over the 14 year time span
( @xmath59 0.1% of the age of our objects ) between iras and iso is possible .
the psf photometry suggests that all the objects are seen as point sources for the isophot detectors .
when subtracting the fitted psf , no systematic residuals are visible in the difference images .
thus , the angular size of the sources is at maximum 10 , a quarter of the fhwm of the psf .
this leads to an upper limit for the source solid angles of about 100@xmath60 .
this fact is supported by the inferred sizes r@xmath32 on the order of 1@xmath810@xmath61sr ( 4@xmath60 ) , which is less than one percent of the pixel size of the c100 detector .
for the three objects for which we obtained a lws spectrum , we can compare the phot flux with the lws continuum . while the lws continuum is a sum of the continuum of the source and background radiation ,
the phot maps give the true flux of the source .
so , the difference between lws and phot should be the background radiation ( e.g. from cold dust ) .
in all three cases ( cepe , l1157 , and hh211-mm ) we clearly see evidence for such a background emission ( see figs.[cepe_data][hh211_data ] ) . in the following subsections we discuss details of the results for the individual objects .
isophot maps and derived photometry for cepe . stars ( * ) show the lws spectrum and the solid line the best graybody fit ( t=35.3k , @xmath17=1.0 , @xmath62=3.0@xmath810@xmath61sr ) to the data .
@xmath63 is fixed to unity .
error bars for the phot data only include the 15% and 10% errors of the detectors .
they do not include additional uncertainties due to background determination .
the four maps are centred on the central wavelength of the appropriate filter and have the same scale and orientation , given by the scale and the arrow . ]
cepe is the brightest object in our sample .
our isophot maps at the four wavelengths of 60 , 100 , 160 , and 200@xmath2 m are shown in the lower part of fig.[cepe_data ] , all at the same scale and orientation .
photometry from these maps and the lws spectrum , are displayed above the maps .
in addition , we plot the best - fitting graybody to these data as solid line .
the fit was used to deconvolve the measurements and the filter transmission curve for converting the measured fluxes to flux densities at the central wavelengths of the used filter . for cepe ,
the fluxes determined with psf and `` aperture '' photometry were consistent .
deviations of the lws continuum from the isophot data exist for wavelengths shorter than 100 and longer than 150@xmath2 m .
this might be evidence for warm and cold dust .
the psf photometry shows that the object is a point source , perfectly aligned in the middle of our map , and no other embedded object is detected .
nevertheless cepe is at least a double source , separated by 1.4 ( @xmath591000au ) , as shown by the 222ghz observations of moro - martn et al .
@xcite .
cepe was observed by chini et al .
@xcite with scuba ( 450 and 850@xmath2 m ) and the iram 30-m telescope ( 1.3 mm ) .
they measured the fluxes in an aperture with a radius of 40 , comparable to the size of our isophot pixels .
the fluxes are given in table[phot_fluxes ] .
we included these data in the graybody fit .
the deduced temperature is 35.3k for the best graybody fit , which also gives @xmath17=1.0 and @xmath62=3.0@xmath810@xmath61sr .
these parameters are computed fixing @xmath16 to unity .
if we vary the optical depth also , the @xmath18 of the fit is lowered from 0.9 to 0.7 .
the new graybody parameters are then : t=42.8k , @xmath17=1.5 , @xmath12=0.8@xmath810@xmath61sr and @xmath16=25 .
nevertheless , the inferred bolometric luminosities and bolometric temperatures are not affected .
we obtain 77.9 , 79.6l@xmath1 and 45.0 , 45.7k for a fixed and free @xmath16 , respectively .
the l@xmath3/l@xmath0 ratio is 0.017 , a strong hint for the class0 nature of this object , even if we observe the superposition of two sources . using the assumptions of myers et al . @xcite and
t@xmath0 we calculate an optical depth at 100@xmath2 m of 9.4 . from this
we determine a radius of the protostellar envelope where @xmath16 is unity of 750au .
thus , the diameter ( 1500au ) is in agreement with moro - martn et al .
@xcite who found that the double system ( separation of 1000au ) is surrounded by a common envelope .
the bolometric temperature of about 45k is well below the value of 60k given by ladd and hodapp @xcite .
they used the iras data ( 12 , 25 , 60 , and 100@xmath2 m ) and an 800@xmath2 m point to fit the bolometric temperature .
chini et al .
@xcite could fit the 100@xmath2 m iras measurement and the 450 , 850 and 1300@xmath2 m points .
these data still do not cover the emission maximum of the source at about 130@xmath2 m .
for an accurate determination of the temperature , however , the position of the maximum of the sed is needed , which was observed here with isophot .
as fig.[cepe_data ] , but for l1157 .
the best graybody has the parameters t=26.5k , @xmath17=1.4 , and @xmath62=3.3@xmath810@xmath61sr .
@xmath16 was fixed to unity . ]
+ our phot maps of l1157 , and the integrated photometry obtained from these maps , are presented in fig.[l1157_data ] .
this figure also shows our lws spectrum of l1157 , as well as the graybody fit to the photometry .
l1157 was observed with scuba and iram 30-m by chini et al .
they give two different measurements for the fluxes : one for the central source only ( 10 aperture ) and one for the source and the whole envelope ( a 55 by 30 elliptical aperture ) .
the fluxes for the central source are given in table[phot_fluxes ] .
l1157 is a point source with a derived temperature of about 26.5k , @xmath17=1.4 , and @xmath12=4.3@xmath810@xmath61sr , under the assumption that @xmath16 is unity .
varying the optical depth does not improve the fit .
the bolometric temperature and luminosity are 37.8k and 7.6l@xmath1 , while 2.5% of these account for l@xmath3 .
this classifies l1157 as a class0 source . according to the scheme of myers et al
@xcite we determine @xmath64 and r@xmath65au .
if we use the fluxes given by chini et al .
@xcite for the envelope , we get a much worse fit and it seems that we have a second cool component in the sed .
thus , the isophot data reflect the emission of the source itself and not the cold extended envelope .
there are small deviations of the lws continuum from the phot photometry over the whole wavelength range
. this could be due to diffuse emission from warm and cold dust , or reflecting uncertainties in the calibration of isophot or lws .
psf photometry shows that the object is not at the centre of our map , but rather shifted slightly to the east .
this could be a hint for another source nearby or a slight mispointing of the telescope due to the limited accuracy of the iras coordinates .
since nothing is known in the literature about a second source , we attributed all the flux to l1157 .
psf fitting to the c200 maps to confirm this was not possible , due to the unknown psf for our half - pixel sampling in the north - south direction .
also the inferred size of the source ( @xmath12 ) does not support the presence of an additional object . as fig.[cepe_data ] , but for hh211-mm .
the best graybody has the parameters t=21.0k , @xmath17=1.5 , and @xmath62=12.5@xmath810@xmath61sr .
@xmath16 was fixed to unity .
the position of ic348ir is indicated by a + sign , and a circle marks the source ic348mms , found by eislffel et al .
( 2003 ) . ]
our phot maps , derived photometry and a lws spectrum of the hh211 region are displayed in fig.[hh211_data ] .
hh211-mm at the centre of our maps is the dominant source at 60 and 100@xmath2 m .
ic348ir , probably a heavily embedded b - star ( strom et al .
@xcite , mccaughrean et al .
@xcite ) , is visible to its north - east ( marked by a cross ) . at longer wavelengths ,
a very cold source hh211firs2 further north becomes visible and even dominant .
it probably coincides with the object ic348mms ( marked by a circle ) , found by eislffel et al .
@xcite to be the source of a newly detected outflow north of hh211 .
thus , the fluxes of hh211firs2 given in table[phot_fluxes ] are a superposition of two different objects .
the c@xmath34 and c@xmath32 measurements are dominated by ic348ir , while c@xmath35 and c@xmath36 are dominated by ic348mms .
therefore no further investigation of the sed of one of these objects was possible . for hh211-mm , we find large differences of the fluxes at 60 and 100@xmath2 m obtained with psf and `` aperture '' photometry .
these differences are due to the other sources influencing the background determination . in addition
, there is a lot of diffuse background emission present , which can be seen in fig.[hh211_data ] as the difference between the phot photometry and the lws continuum .
so , it is very difficult to determine the background and to state which pixel contributes to the flux of which object .
this is further complicated by the fact that hh211firs2 has a higher surface brightness than our point source hh211-mm at 160 and 200@xmath2 m .
additionally , ic348ir could influence our measured flux for hh211-mm also . due to the difficulties in the determination of the fluxes of hh211-mm we supplemented the isophot data with scuba datapoints at 450 and 850@xmath2 m from rengel et al .
@xcite and jcmt bolometry at 1.1 mm from mccaughrean et al . @xcite .
the fluxes are given in table[phot_fluxes ] .
we find a fit ( @xmath18=2.0 ) with the following object parameters : t=21.0k , @xmath17=1.5 , and @xmath12=12.5@xmath810@xmath61sr . in this case a variable optical depth ( instead of a fixed value of unity ) does not improve the fit .
the inferred bolometric temperature is 31.4k and the bolometric luminosity is 4.5l@xmath1 , while 4.5% account for l@xmath3 .
this verifies the class0 nature of hh211-mm . according to the low bolometric temperature we find a high @xmath16 on the order of 20 and a large radius for the @xmath66 surface of 520au . as fig.[cepe_data ] , but for ic1396w .
we do not have lws data for this source .
the best graybody fit has a temperature of t=30.0k , @xmath17=0.3 , and @xmath62=1.1@xmath810@xmath61sr .
@xmath16 was fixed to unity .
extended cool dust , or a close group of cold sources , are seen northeast of ic1396w , while a very cold bright source appears at 200@xmath2 m to the south - west . ] our phot maps and the derived photometry for ic1396w are shown in fig.[ic1396w_data ] , together with the graybody fit to these data .
in the maps , two additional sources are evident .
ic1396wfirs2 , to the north - east of ic1396w , peaks at 160@xmath2 m , whereas ic1396wfirs3 , to the south - west , is remarkably red , with a flux ratio at c@xmath36/c@xmath35 of 2.6 .
these additional objects were not detected with the c100 detector , because the maps at these wavelengths are slightly smaller and the objects just fall outside .
nevertheless , they influence the flux measurements , especially in the c200 wavelengths range .
we do not have observations at longer ( sub - millimeter or millimeter ) wavelengths to supplement the iso data .
for ic1396w we find a temperature of t=30.0k , @xmath17=0.3 , and @xmath12=1.1@xmath810@xmath61sr for the best graybody fit .
these parameters are obtained fixing @xmath16 to unity .
a variable optical depth does not improve the fit .
the fit is poor ( @xmath18=1.3 ) , however , suggesting errors in the photometry or a source with dust at more than one temperature . with the graybody parameters we determine a bolometric temperature of 32.6k , a luminosity of 16.4l@xmath1 , and a l@xmath3/l@xmath0 ratio of 0.059 .
concerning the difficulties in the flux measurements and the not available observations at longer wavelengths , the classification of ic1396w as a class0 source remains questionable .
at least sub - mm observations are needed to confirm the presence of a spatially extended envelope , to ensure that we do not just see a compact disk edge - on . as fig.[ic1396w_data ] , but for l1211 .
the best fit is a blackbody with t=30.5k , and @xmath62=2.1@xmath810@xmath61sr .
@xmath16 was fixed to unity .
the object appears north of the nominal iras position , and a second cool source is detected south - west of it .
see text for the identification of these sources . ]
our phot maps of l1211 , the derived photometry , and a graybody fit to these data are displayed in fig.[l1211_data ] .
somewhat to our surprise , the l1211 source was not found at its nominal iras position , but is shifted a full pixel , corresponding to 45 to the north .
a second source l1211firs2 , is found in the south - west .
comparing our maps with the work of tafalla et al .
@xcite and anglada & rodrguez @xcite we find that our object l1211 is identical to mms4 or vla5 and the object l1211firs2 seems to be a superposition of mms3 , mms2 , and mms1 , and vla3 and vla1 , respectively .
the best fits to both sources are blackbodies ( @xmath17=0.0 ) and have temperatures of t=30.5k and t=26.8k for l1211 and l1211firs2 , respectively .
we infer a solid angle @xmath12 of the sources of 2.1 and 1.7@xmath810@xmath61sr .
this results in bolometric temperatures of 30.5 and 26.9k and in bolometric luminosities of 33.1 and 16.0l@xmath1 , for l1211 and l1211firs2 , respectively .
even if we get quite high l@xmath3/l@xmath0 ratios ( 0.073 and 0.100 ) , we can not firmly establish the classification of the sources as class0 objects .
this is because our isophot data do not cover the emission maximum and also l1211firs2 is a superposition of the emission from several sources .
tafalla et al .
@xcite classify l1211 as a transitional object between class0 and class1 .
they use the iras fluxes and an additional observation at 1.2 mm ( see table[phot_fluxes ] ) . with these data , as for our iso data , the maximum of the emission could not be determined exactly .
it could only be constrained to lie between 100 and 1200@xmath2 m . with our iso data
, we could corroborate the assumption that l1211 is of class0 , since the maximum of the sed is at @xmath67@xmath2 m . if we use the 1.2 mm datapoint from tafalla et al .
@xcite for our analysis , we get a poor fit . using both
, the isophot and the 1.2 mm point , with plausible values for @xmath17 ( 1.0 .. 2.0 ) , the 1.2 mm flux is always overestimated by about one order of magnitude .
the same applies for l1211firs2 .
accepting values of @xmath17=3.0 or higher , we determine t=36.3 , @xmath62=3.0@xmath810@xmath61sr ( l1211 ) and t=24.8 , @xmath62=5.8@xmath810@xmath61sr ( l1211firs2 ) .
this leads to bolometric temperatures of about 90 and 46k and bolometric luminosities of 77 and 19l@xmath1 for l1211 and l1211firs2 , respectively .
the resulting l@xmath3/l@xmath0 ratios of 0.001 and 0.004 would then classify both objects as class1 ( l1211firs2 is still near the transition phase between class0 and class1 ) . as fig.[ic1396w_data ] , but for rno15fir .
the best fit has temperature of t=34.0k , @xmath17=1.1 , and @xmath62=0.7@xmath810@xmath61sr .
@xmath16 was fixed to unity . south - east of rno15fir
, the warmer source rno15 is detected as well ( marked by a cross ) , especially at the shorter wavelengths .
the positions of two other weak sub - mm sources ( sms1 north , sms2 south ) are indicated by a circle ( rengel at al .
( 2001 , 2002 ) ) . ]
the isophot maps of the rno15fir region are shown in fig.[rno15fir_data ] , together with the derived photometry and the graybody fit .
visible on our maps are rno15fir in the centre , and the source rno15 to the south - east , marked by a cross . since this source is warmer than rno15fir , it is prominent at the shorter wavelengths , but fades considerably relative to rno15fir towards the longer wavelengths ( due to the larger pixel size the two objects also merge ) . from higher spatial resolution sub - mm maps at 450 and 850@xmath2 m taken with scuba ( rengel et al .
@xcite ) we know that two other objects sms1 and sms2 are present to the north and south of rno15fir , but are merged with it at isophot resolution ( marked by a circle in fig.[rno15fir_data ] ) .
they surely influence our flux measurements of rno15fir .
since the isophot measurements show a broad and not well defined maximum of the sed , we supplement these data with scuba measurements of rengel et al .
@xcite to determine more accurate source properties .
the fluxes at 450 and 850@xmath2 m are measured in a 45 by 45 aperture ( rengel priv .
communication ) and listed in table[phot_fluxes ] . using these data in combination with the isophot points , we get t=34.0k , @xmath17=1.1 , and @xmath12=1.7@xmath810@xmath61sr with an optical depth @xmath16 fixed to unity .
the resulting bolometric temperature and luminosity are 44.6k and 8.4l@xmath1 .
1.7% of the luminosity is in the sub - millimeter regime , classifying rno15fir as a class0 source .
if we vary the optical depth at 100@xmath2 m also , the fit is improved ( the @xmath18 is lowered from 1.0 to 0.6 ) , but the parameters t@xmath0 and l@xmath0 do not change .
davis et al .
@xcite suggested that rno15fir might be a double source , due to the observed wiggling in the outflow .
this might also be indicated by the deviation of the data points from the determined sed ( see fig.[scuba_data ] ) .
the bolometric luminosity temperature diagram for the objects analysed here ( thick xs ) , the class0 data from table[lbolvslhtwo ] ( filled diamonds ) , and the class0 data from the review of andr et al .
( 2000 ) ( open diamonds ) .
the superimposed evolutionary tracks are discussed in section[evolscheme ] .
protostars evolve from right to left .
three tracks for final masses of 0.2 , 1 , and 5m@xmath1 are displayed . the model peak accretion rate is reached at 17,000yr , and the power law fall - off is @xmath68t@xmath69 with time t , on a 30,000yr timescale .
the vertical dotted lines on the tracks mark the model ages of 20 , 30 , 40 , 50 , and 75 thousand years .
, width=321 ] do class0 objects develop into class1 and class2 protostars ? to answer this , we wish to determine basic parameters for the class0 protostars , such as age , surrounding mass , present mass and final mass .
these , however , are model dependent quantities . in fig.[classzero ] we plot the locations of our seven ( including l1211firs2 ) sources on the l@xmath0t@xmath0 diagram ( large crosses ) , which is the protostellar equivalent of a hertzsprung - russell diagram ( myers et al .
@xcite ) . also plotted on the diagram are the data for another 37 class0 protostars , as listed by andr et al .
two of these sources possess bolometric luminosities above 1000l@xmath1 , and so fall outside the display . note that class0 protostars possess bolometric temperatures below @xmath70k .
the present sample contains quite powerful and cold class0 members .
two sources lie above the location of the other explored sources .
these are l1211 and cepe . as we demonstrate below , such powerful class0 sources with low bolometric temperature ,
can indeed be included in an evolutionary model through the classes012 .
the large surrounding masses observed restrict the type of model and these objects could go on to produce high - mass stars .
the model tracks plotted represent the evolution of three protostars which end up accumulating masses of 0.2 , 1 , and 5m@xmath1 .
the tracks were derived by combining the unification scheme , as reviewed by smith @xcite , with the framework for protostellar envelopes presented by myers et al .
@xcite , according to the prescription presented below .
we thus determine model ages , present masses of the protostellar nucleus , envelope masses and the final stellar masses ( table[parameters ] ) .
the result is that the more massive class0 protostars possess large envelopes and would become massive stars .
according to the model described here , most of the envelope , however , is not accreted but dispersed , if the majority of protostars here are to form low - mass stars .
note that alternative schemes have been presented by bontemps et al .
@xcite , saraceno et al .
@xcite , and andr et al . @xcite .
the model envelope masses are in agreement with the measurements . just for the two objects where we could not determine a proper bolometric temperature ( l1211 and l1211firs2 )
, there is a significant deviation .
lccccl object & l@xmath0 & t@xmath0 & m@xmath53 & l@xmath71 & ref .
+ & [ l@xmath1 ] & [ k ] & [ m@xmath1 ] & [ 10@xmath73l@xmath1 ] & + rno15fir & 8.4 & 44.6 & 0.9@xmath74 & 0.46 & 1 , 7 + hh211-mm & 4.5 & 31.4 & 1.5 & 3.1 & 2 , 7 + l1157 & 7.6 & 37.8 & 0.5 & 6.1 & 3 , 7 + ic1396w & 16.4 & 32.6 & & 19.0 & 6 + l1211 & 33.1 & 30.5 & 0.8@xmath75 & 10.7 & 7 + cepe & 77.9 & 45.0 & 7.0 & 70.0 & 4 , 7 + l1448n & 11.0 & 70.0 & 2.3 & 2.16 & 7 + l1448irs2 & 6.0 & 70 ? & 0.9 & 2.8 & 7 + l1448c & 9.0 & 60.0 & 1.4 & 5.7 & 7 + iras03282 & 1.5 & 35.0 & 0.6 & 4.46 & 7 + hh212 mm & 14.0 & 70 ?
& 1.2 & 5.3 & 5 , 7 + hh24mms & 5.0 & 20 ? & 4.0 & 1.21 & 1 , 7 + hh25mms & 6.0 & 34.0 & 0.5 & 6.61 & 1 + ngc2264gvla2 & 12.0 & 25.0 & 2.0 & 7.75 & 3 + vla1623 & 1.0 & @xmath7635 & 0.7 & 0.81 & 3 , 7 + ser firs1 & 46.0 & 51.0 & 3.0 & 0.64 & 7 + taken from rengel et al .
@xcite taken from tafalla et al .
@xcite * references : * ( 1 ) davis et al .
@xcite ( 2 ) mccaughrean et al .
@xcite ( 3 ) davis & eislffel @xcite ( 4 ) eislffel et al .
@xcite ( 5 ) zinnecker et al .
@xcite ( 6 ) froebrich & scholz @xcite ( 7 ) own measurements is the luminosity of the outflows from the class0 sources correlated with the properties of the sources like their bolometric luminosity , the temperature , or mass of their envelopes ? to answer this question we measured the luminosity of the outflows from 16 of the class0 sources in andr et al . @xcite and our objects , in the 10s(1 ) line of molecular hydrogen at 2.122@xmath2 m . due to the short cooling times
( some years , smith and brand @xcite ) , h@xmath4 is a good tracer of emission of shocked gas caused by current interactions between outflowing material and the surrounding gas .
the 10s(1 ) line of h@xmath4 is usually the brightest ro - vibrational line in a spectrum of shocked gas and thus most easily detected . in magnetohydrodynamic models of class0 sources ( e.g. shu et al .
@xcite , hirose et al .
@xcite , and ouyed and pudritz @xcite ) the accretion rate onto the protostar is connected to the amount of material injected into the outflowing jet .
this material interacts with the surrounding quiescent gas in shocks .
hence , the luminosity of these shocks may be connected to the mass accretion rate and thus to the source properties . here , we tested for correlations of the 10s(1 ) h@xmath4 luminosity with the source bolometric luminosity , the bolometric temperature and the mass of the surrounding protostellar envelope given by andr et al . @xcite .
the results of these comparisons are shown in figs.[outflows_lbol ] , [ outflows_tbol ] , and [ outflows_menv ] .
we obtained a linear regression for each case and tested if the slope of the regression line differed statistically significantly from a slope value of zero . with a probability of error of 5%
none of the regression lines differs from a constant value .
additionally a kolmogorow - smirnow - test shows , that with a probability of error of 0.1% the data is not consistent with a constant value .
thus , a significant correlation of the outflow luminosity in the 10s(1 ) line of h@xmath4 with any source parameter was not found .
units < 28.571mm,26.087 mm > point at 0 0 x from -0.10 to 2.0 , y from -0.4 to 1.90 left label @xmath77 \right ) $ ] ticks in long numbered from -0 . to 1.8 by 0.4 in short unlabeled from -0.4 to 1.8 by 0.1 / right label l@xmath78[10@xmath73l@xmath1 ] ticks in long logged numbered from 1 to 1 by 1 from 2 to 6 by 2 from 10 to 20 by 10 from 40 to 60 by 20 in short logged unlabeled from 0.5 to 6 by 0.5 from 6 to 10 by 1 from 10 to 20 by 2 from 20 to 75 by 5 / bottom label @xmath79 \right)$ ] ticks in long numbered from 0 to 1.9 by 0.4 in short unlabeled from 0 to 1.9 by 0.1 / top label l@xmath0[l@xmath1 ] ticks in long logged numbered from 1 to 1 by 1 from 2 to 6 by 2 from 10 to 10 by 10 from 20 to 60 by 20 in short logged unlabeled from 1 to 10 by 1 from 10 to 40 by 5 from 40 to 90 by 10 / at 0.9243 -0.33724 at 0.6532 0.49136 at 0.8808 0.78533 at 1.2148 1.27875 at 1.5198 0.90309 at 1.8915 1.83885 at 1.04139 0.33445 at 0.77815 0.44715 at 0.95424 0.75587 at 0.17609 0.6492988 at 1.14613 0.724276 at 0.69897 0.081099 at 0.77815 0.82023 at 0.00000 -0.08951 at 1.07918 0.88899 at 1.66276 -0.19525 units <
-80mm,26.087 mm > point at -1 0 x from 1.9 to 1.15 , y from -0.4 to 1.90 left label @xmath77 \right ) $ ] ticks in long numbered from -0 .
to 1.8 by 0.4 in short unlabeled from -0.4 to 1.8 by 0.1 / right label l@xmath78[10@xmath73l@xmath1 ] ticks in long logged numbered from 1 to 1 by 1 from 2 to 6 by 2 from 10 to 20 by 10 from 40 to 60 by 20 in short logged unlabeled from 0.5 to 6 by 0.5 from 6 to 10 by 1 from 10 to 20 by 2 from 20 to 75 by 5 / bottom label @xmath80 \right)$ ] ticks in long numbered from 1.2 to 1.9 by 0.2 in short unlabeled from 1.2 to 1.9 by 0.1 / top label t@xmath0[k ] ticks in long logged numbered from 20 to 40 by 10 from 60 to 60 by 20 in short logged unlabeled from 15 to 75 by 5 / at 1.6493 -0.33724 at 1.4969 0.49136 at 1.5775 0.78533 at 1.5132 1.27875 at 1.4843 0.90309 at 1.6532 1.83885 at 1.84510 0.33445 at 1.84510 0.44715 at 1.77815 0.75587 at 1.54407 0.6492988 at 1.84510 0.724276 at 1.30103 0.081099 at 1.53148 0.82023 at 1.54407 -0.08951 at 1.39794 0.88899 at 1.70757 -0.19525 units < 46.154mm,26.087 mm > point at 0 0 x from -0.40 to 0.9 , y from -0.4 to 1.9 left label @xmath77 \right ) $ ] ticks in long numbered from -0 . to 1.8 by 0.4 in short unlabeled from -0.4 to 1.8 by 0.1 / right label l@xmath78[10@xmath73l@xmath1 ] ticks in long logged numbered from 1 to 1 by 1 from 2 to 6 by 2 from 10 to 20 by 10 from 40 to 60 by 20 in short logged unlabeled from 0.5 to 6 by 0.5 from 6 to 10 by 1 from 10 to 20 by 2 from 20 to 75 by 5 / bottom label @xmath81 \right)$ ] ticks in long numbered from -0.2 to 0.8 by 0.2 in short unlabeled from -0.4 to 0.9 by 0.1 / top label m@xmath53[m@xmath1 ] ticks in long logged numbered from 2 to 6 by 2 from 0.5 to 0.5 by 0.5 from 1 to 1 by 1 in short logged unlabeled from 1 to 7.5 by 0.5 / at -0.04576 -0.33724 at 0.17609 0.49136 at -0.30103 0.78533 at -0.09691 1.02938 at 0.84510 1.84501 at 0.36173 0.33445 at -0.04576 0.44715 at 0.14613 0.75587 at -0.22185 0.6492988 at 0.07918 0.724276 at 0.60206 0.081099 at -0.30103 0.82023 at -0.15490 -0.08951 at 0.30103 0.88899 at 0.47712 -0.19525 the lack of such correlations may have various explanations .
for example , in each outflow , we observe h@xmath4 emission at various distances from the source and these knots or bow shocks are indicating material which was ejected from the source at different times in the past ( @xmath82 = distance to the source / jet velocity ) .
also , the knot luminosity depends on the local properties of the surrounding gas ( e.g. gas density , atomic fraction ) .
additionally , the extinction gradient in the k - band along the outflow is not known .
it will alter the measured relative and total fluxes in the sense that knots closer to the source appear fainter due to higher extinction . in section [ evolscheme ] , however , we argue that the location on these diagrams depends sensitively on both mass and age , which results in a wide scatter . therefore , the lack of a significant correlation of the present source properties and the outflow luminosity in the h@xmath4 10s(1 ) line may not be surprising . a better tool for comparing the outflows to source properties may be an optically thin line of co ( e.g. the 10@xmath83co line ) , which should give a measurement of the time - integrated power of the outflow without being influenced by local extinction effects .
comparable observations in the same transition and isotope of co are needed for a statistically reasonable sample of objects to study their behaviour . at present , only small samples of class0 sources have been thus analysed ( e.g. bontemps et al .
@xcite , smith @xcite ) .
an evolutionary model for protostars is presented in the appendix .
the outflow scheme has been elaborated by smith @xcite and applied by davis et al .
@xcite , yu et al .
@xcite , and stanke et al .
it is based on a prescribed accretion rate from an envelope .
modelling of outflows has demonstrated that the fraction of mass which escapes through jets must reach a maximum during the class0 stage .
this is required to account for the excess momentum and power of class0 bipolar outflows , as calculated from observations of emission lines of co rotational transitions ( see smith @xcite ) .
we outline in the appendix the fundamental formulae of the evolutionary scheme . according to previous modelling of the envelope ,
three parameters must be introduced to generate plausible models for the bolometric temperature . as shown by myers et al .
@xcite , these are ( 1 ) the envelope s outer temperature ( here t@xmath84=24k ) , ( 2 ) the efficiency of accretion of the envelope into the star - jet system and ( 3 ) the difference in evolutionary timescale between the envelope and the protostar .
the envelope consists of material which will fall onto the central object as well as mass directly lost soon after the class0 stage .
this extra mass component proves necessary to produce a low bolometric temperature , as observed for the class0 sources , yet must be rapidly lost in order to yield ttauri stars within a reasonable time ( myers et al .
@xcite ) . the derived outflow shock luminosity versus the bolometric source luminosity for the class0 sources listed in table[lbolvslhtwo ] , as well as the class0 sample investigated by stanke ( 2000 ) ( symbol : ` x ' ) and the far - infrared line iso luminosities presented by giannini et al . ( 2001 ) ( symbol : * ) .
the model tracks are for the same three models presented in fig[classzero ] and the straight line divides model class0 and model class1 protostars , as determined by the protostar possessing half of its final mass .
, width=321 ] the derived outflow shock luminosity versus the bolometric source luminosity for class1 sources from the sample of stanke ( 2000 ) ( symbol : ` x ' ) and the far - infrared line iso luminosities presented by giannini et al .
( 2001 ) ( symbol : * ) .
the model tracks are for the same three models presented in fig[classzero ] and the straight line divides model class0 and model class1 protostars , as determined by the protostar possessing half of its final mass . , width=321 ] previously , we modelled the envelope evolution by assuming mass conservation . here , we find two significant adjustments are necessary in order to model the new data and maintain plausible time scales .
first , the initial mass in the envelope which will eventually fall inwards is reduced to 87% of the total required mass to form the star and excavate the bipolar outflow .
the other 13% is presumed to initially lie within a flattened disk .
this yields the values in the column infall mass in table[parameters ] . in fig.[classzero ]
, we plot the sample summarised in table[lbolvslhtwo ] for class0 sources for which a correlation of the source properties with the outflow luminosity in the 10s(1 ) line of h@xmath4 has been investigated .
note that this sample includes warmer and less luminous protostars than in the isophot sample investigated above .
according to the tracks , this corresponds to a wide range in final stellar masses .
the lowest mass star forming here is found to be vla1623 ( assuming t@xmath0=35k ) which will reach just 0.07m@xmath1 , owing to its low bolometric luminosity of only 1l@xmath1 ( andr et al .
@xcite ) . the low final mass is a result of this version of the evolutionary scheme employed , for which we maintain the same accretion timescale but alter the accretion rate to generate the tracks .
this implies that the final mass is nearly proportional to the peak accretion luminosity .
future statistical studies will lead to revisions of this first model .
the simplest form of the unifying model , assumed here , is that a fraction of the jet power is instantaneously dissipated in shock waves , while the bipolar outflow is a time - averaged recording of the momentum outflow . to model the outflow ,
we have previously employed the h@xmath4 luminosity , l(h@xmath4 ) , which we estimate to be ten times the 10s(1 ) luminosity .
this is consistent with expectations from shock physics and allows a comparison with previous diagrams presented by stanke @xcite and smith @xcite . here
, however , we shall use the jet power itself as the comparison parameter . for the comparison ,
we assume that the observed emission is produced in the warm shocks where the jets dissipate their energy , l@xmath85 .
the fraction of the jet power dissipated in molecular hydrogen lines is taken to be 2% .
this is consistent with numerical simulations and bow shock modelling which predict , typically , 10% of the infrared radiation from shocks in dense clouds to be in the form of h@xmath4 lines .
we also assume that 80% of the jet energy is hidden by just under two magnitudes of k - band extinction .
the shock power has also been estimated from the far - infrared lines of co , oi , oh , and h@xmath4o , measured by iso ( giannini et al .
@xcite ) . here
, we shall assume that these lines in total , within the iso - lws beam , also represent 2% of the jet power , l@xmath85 .
we thus increase estimated h@xmath4 and sub - mm luminosities by 50 to yield the displayed values .
while these approximations are far from ideal , they permit us to determine if the evolutionary scheme is plausible .
figure[h2-lbol0 ] demonstrates that the class0 protostars possess almost exclusively high ratios of l(@xmath85)/l@xmath0 .
two objects , however , rno15fir and ser - firs1 , lie well within the model class1 regime .
environmental factors could cause a downward shift of the data points , including higher extinction or radiative shocks which are less efficient in h@xmath4 vibrational excitation .
furthermore , these data are consistent with the same model tracks fitted to the bolometric luminosity temperature data .
the lack of a correlation in the data is thus put down to the combination of the distributions in both mass and age .
lcccccrr object & age & mass & final & min .
& infall & env .
& m@xmath53 + & & & mass & mass & mass & mass & + rno15fir & 29.7 & 0.10 & 0.5 & 0.05 & 0.32 & 1.0 & 0.9 + hh211-mm & 24.0 & 0.06 & 0.3 & 0.19 & 0.26 & 1.4 & 1.5 + l1157 & 27.0 & 0.10 & 0.5 & 0.10 & 0.35 & 1.3 & 0.5 + ic1396w & 24.9 & 0.21 & 1.2 & 0.50 & 0.88 & 4.2 &
+ l1211 & 24.0 & 0.43 & 2.6 & 1.53 & 1.91 & 10.3 & 0.8 + l1211firs2&22.5 & 0.21 & 1.4 & 1.42 & 1.06 & 6.9 & 2.1 + cepe & 30.6 & 0.97 & 4.2 & 0.37 & 2.83 & 8.1 & 7.0 + in addition , previously measured class1 outflows almost all lie below the predicted class border line , as shown in fig.[h2-lbol1 ] . note , however , that for many of the h@xmath4 flows in orion detected by stanke @xcite only upper limits for the bolometric luminosity are available .
nevertheless , the division of the two classes with the model straight line where the protostar has acquired half the final stellar mass , is evident .
a similar difference in outflow luminosity between class0 and class1 sources has been found by bontemps et al .
@xcite .
the derived outflow shock luminosity versus the bolometric source temperature for the class0 sources listed in table[lbolvslhtwo ] .
the model tracks are for the same three models presented in fig[classzero].,width=321 ] figures[lshock - tbol ] demonstrates that the model is also consistent with the envelope properties .
the main exceptions apparent from this diagram are a group of low luminosity h@xmath4 objects .
this suggests that the extinction for these sources may far exceed the fiducial two magnitudes .
we have observed the spectral energy distributions for the six deeply embedded objects cepe , l1211 , ic1396w , l1157 , hh211-mm , and rno15fir in the far - infrared with iso .
the inferred temperatures and l@xmath3/l@xmath0 ratios confirm the class0 nature of four of these sources , within the errors .
employing an evolutionary scheme , we are able to estimate the age , surrounding mass and the current and final mass of these sources .
these estimates are , however , model dependent .
two sources , cepe and l1211 appear to develop into intermediate mass stars , while the others will become solar mass stars or lower mass objects .
the comparison of the isophot and lws observations for three of the sources reveals the existense of emission from cold dust in the immediate vicinity of the objects . a comparison of the luminosity in the 10s(1 ) line of h@xmath4 of the related outflows for 16 class0 sources , with the source bolometric luminosity , bolometric temperature , and envelope mass was done .
we found no statistically significant correlation of the outflow luminosity with each of these source parameters .
this could be due to the h@xmath4 luminosity mainly depending on the local properties of the surrounding gas .
the unifying scheme , however , explains the lack of correlations as due to evolutionary effects .
furthermore , the scheme which involves a redistribution of mass between envelope , disk , protostar , jets and outflow , accounts for the differences in source properties according to the class .
we thank manfred stickel from mpia for his help with the data reduction of the phot data .
we thank alex rosen for a critical reading of the manuscript and the department of culture , arts and leisure , northern ireland for financial support .
jochen eislffel and dirk froebrich received financial support from the dlr through verbundforschung grant 50or99049 .
+ the isophot data presented in this paper were reduced using pia , which is a joint development by the esa astrophysics division and the isophot consortium with the collaboration of the infrared processing and analysis center ( ipac ) . contributing isophot consortium institutes are dias , ral , aip , mpik and mpia .
+ the iso spectral analysis package ( isap ) is a joint development by the lws and sws instrument teams and data centers .
contributing institutes are cesr , ias , ipac , mpe , ral and sron .
we present and test a model based on the transfer of gas between components .
we take a spherical envelope of gas and dust , and prescribe an accretion rate from the inner edge of the envelope onto a disk . note
that we assume a centrifugal barrier at 30au , which defines the inner envelope outer disc transition .
the accretion disk processes most of the mass onto the protostar and a fraction into twin jets .
the speed of the jets is assumed to be a fixed fraction of the escape speed from the protostellar surface . the accretion rate from the envelope
is taken to increase exponentially for a short period before decreasing as a power law through the class0 , 1 and 2 phases .
the zero point of time is thus defined as the moment when accretion starts and , simultaneously , a central hydrostatic object forms .
the accretion rate is @xmath86 energy release through accretion and contraction are included . in the models
shown , the peak accretion rate is reached at t@xmath87yr , and the power law index is @xmath88 , on a t@xmath84 = 30,000yr timescale . the accreted mass is predominantly accrued by the growing protostars .
the fraction @xmath89 which escapes through twin jets reaches a maximum of @xmath90 at the peak accretion time : @xmath91^\zeta \label{eqn - hm}\ ] ] where @xmath92 is found to be appropriate .
hence the mass left over , which accretes onto the core to form the star is @xmath93 to form a star like the sun , this model will provide an early accretion peak in which @xmath94 for 10@xmath95 years , and eventually becoming @xmath96 for 10@xmath97years , corresponding to class0 and class2 or classical ttauri stars , respectively .
the power - law has substantial observational support ( calvet et al .
@xcite ) .
we previously modelled the envelope evolution by assuming mass conservation .
here , we make two significant adjustments in order to model the new data .
first , the initial mass in the envelope which will eventually fall inwards is reduced to 87% of the total required mass to form the star and excavate the bipolar outflow .
the other 13% is presumed to initially lie within a flattened disk .
this yields the values in the column infall mass in table[parameters ] .
the total mass can be written analytically in terms of an incomplete gamma function on integrating eqn.[eqnpower ] : @xmath98.\ ] ] secondly , we find that the low bolometric temperatures of class0 protostars can only be attained by introducing an additional mass component to the envelope . in confirmation of the results of myers
@xcite , we find that this extra mass is lost on a shorter timescale than the protostellar accretion timescale . the bolometric temperature is calculated according to the optically thick case of myers et al .
we thus extend the opacity law approximation employed up to 60 to 120@xmath2 m with the same form and take the optically thick envelope throughout the early evolutionary stages .
we have thus found here that an envelope mass @xmath99 where @xmath100 provides bolometric temperatures , timescales and masses consistent with the observed samples .
the envelope mass provides a testable prediction .
this mass is not strongly dependent on the evolutionary path but is necessary to provide the optical depth out to a sufficiently large radius to permit the measured low bolometric temperature .
the total mass is dominated by the outer regions of the envelope , while the total optical depth is controlled by the inner region ( for all plausible density distributions such as @xmath101 , as assumed here ) .
hence , the mass is sensitive to the extent of the envelope .
for this reason , we present three determinations of the envelope mass in table[parameters ] .
masses derived from submillimetre observations yield quite low extended masses ( andr et al .
@xcite ) , consistent with the absence of more mass than necessary to form the star and feed the jets ( smith @xcite ) .
it is clear that both the observationally derived mass and model mass are sensitive to chosen physical parameters and both will need refining .
andr , p. , ward - thompson , d. , barsony , m. 2000 , in protostars and planets iv , 59 anglada , g. , rodrguez , l.f . , 2002 , revista mexicana de astronoma y astrofsica , 38 , 12 barsony , m. , ward - thompson , d. , andr , p. , o'linger , j. , 1998 , aj , 509 , 733 bontemps , s. , andr , p. , terebey , s. , cabrit , s. , 1996 , a&a , 311 , 858 calvet , n. , hartmann , l.w .
, strom , s.e . , 2000 , in protostars & planets iv , ed . v. mannings et al . , ( tucson , u. of arizona press ) , 377 chini , r. , ward - thompson , d. , kirk , j.m . , nielbock , m. , reipurth , b. , sievers , a. , 2001 , a&a , 369 , 155 clegg , p.e . ,
ade , p.a.r .
, armand , c. , et al . , 1996 , a&a , 315 , l38 davis , c.j . ,
eislffel , j. , 1995 , a&a , 300 , 851 davis c.j . ,
smith m.d . ,
moriarty - schieven g.h . , 1998 , mnras , 299 , 825 davis , c.j . , ray , t.p .
, eislffel , j. , corcoran , d. , 1997 , a&a , 324 , 263 eislffel , j. , a&a , 2000 , 354 , 236 eislffel , j. , froebrich , d. , 2004 , in prep .
eislffel , j. , froebrich , d. , stanke , t. , mccaughrean , m.j . , 2003 ,
apj , in press eislffel , j. , smith , m.d . ,
davis , c.j . , ray , t.p .
, 1996 , aj , 112 , 2086 eislffel , j. , ziener , r. , froebrich , d. , 2004 , in prep .
froebrich , d. , eislffel , j. , 2004 , in prep .
froebrich , d. , scholz , a. , 2003 , a&a , in press froebrich , d. , smith , m.d . ,
eislffel , j. , 2002 , a&a , 385 , 239 froebrich , d. , ziener , r. , eislffel , j. , ag abstr .
ser . , 18 , 25 gabriel , c. , et al . , 1997 , in asp conf
125 , astronomical data analysis software and systems ( adass ) vi , ed .
g. hunt & h.e .
payne , ( san francisco asp ) , 108 giannini , t. , nisini , b. , lorenzetti , d. , 2001 , apj , 555 , 40 gueth , f. , guilloteau , s. , 1999 , a&a , 343 , 571 gueth , f. , guilloteau , s. , dutrey , a. , bachiller , r. , 1997 , a&a , 323 , 943 hirose , s. , uchida , y. , shibata , k. , matsumoto , r. , 1997 , pasj , 49 , 193 kessler , m.f . ,
steinz , j.a . ,
anderegg , m.e .
, et al . , 1996 , a&a , 315 , l27 ladd , e.f .
, hodapp , k .- w . , 1997 , apj , 475 , 749 laureijs , r.j . , 1999 ,
point spread function fractions related to the isophot c100 and c200 arrays , iso - data centre , astrophysics division , esa , villafranca spain , http://www.iso.vilspa.esa.es/users/expl_lib/pht/c200fpsf02.ps.gz lefloch , b. , eislffel , j. , lazareff , b. , 1996 , a&a , 313 , l17 lemke , d. , klaas , u. , abolins , j. , et al .
, 1996 , a&a , 315 , l64 mccaughrean , m.j . , rayner , j.t . ,
zinnecker , h. , 1994 , apj , 436 , l189 moro - martn , a. , noriega - crespo , a. , molinari , s. , testi , l. , chernicharo , j. , sargent , a. 2001 , a&a , 555 , 146 motte , f. , andr , p. , 2001
, a&a , 365 , 440 myers , p.c . ,
adams , f.c . , chen , h. , schaff , e. , 1998 , apj , 492 , 703 ouyed , r. , pudritz , r.e .
, 1997 , apj , 482 , 712 rengel , m. , eislffel , j. , hodapp , k .- w . , 2003 , in prep .
rengel , m. , froebrich , d. , eislffel , j. , hodapp , k. , 2001 , ag abstr . ser .
18 , 66 rengel , m. , froebrich , d. , hodapp , k. , eislffel , j. , 2002 , in : the origins of stars and planets : the vlt view , joo alves & mark mccaughrean ( ed . ) saraceno p. , andr p. , ceccarelli c. , griffin m. , molinari s. , 1996 , a&a 309 827 shu , f. , najita , j. , ostriker , e. , wilkin , f. , ruden , s. , lizano , s. , 1994 , apj , 429 , 781 smith m.d .
, 1998 , ap&ss , 261 , 169 smith , m.d . , 2000 , iraj , 27 , 25 smith , m.d . , 2002 , in : the origins of stars and planets : the vlt view , joo alves & mark mccaughrean ( ed . ) smith , m.d . ,
brand , p.w.j.l . , 1990 , mnras , 245 , 108 smith , m.d . , froebrich , d. , eislffel , j. , 2003 , apj , 592 , 245 stanke , t. , 2000 , ph.d .
thesis ( aip , potsdam ) strom , s.e . ,
strom , k.a . ,
carrasco , l. , 1974 , pasp , 86 , 798 tafalla , m. , myers , p.c . , mardones , d. , bachiller , r. , 1999 , a&a , 348 , 479 yu , k.c . , billawala , y. , smith , m.d . , bally , j. , butner , h. , 2000 , aj , 120 , 1974 | we present far - infrared maps and spectroscopy for a number of deeply embedded protostellar objects ( cepe , hh211-mm , ic1396w , l1157 , l1211 , and rno15fir ) from data that we acquired with the iso instruments phot and lws .
several previously undetected deeply embedded sources are found in the vicinity of our targets .
we determine temperatures and luminosities of seven objects and locate them on a l@xmath0-t@xmath0 diagram the equivalent to a hertzsprung - russell diagram for protostars .
their masses and ages , according to their location on tracks taken from our evolutionary model , are derived .
l1211 and cepe appear to be intermediate mass objects which will reach final masses of about 3m@xmath1 , while the other sources are in or below the solar mass range .
the derived ages of 15000 to 30000yr are consistent with their current class0 state . a comparison of the luminosity of the associated outflows in the 10s(1 ) line of molecular hydrogen with the source properties ( bolometric luminosity , bolometric temperature , and envelope mass ) of 16 class0 sources shows no statistically significant correlations .
nevertheless , the data are consistent with a scheme in which the outflow strength and protostar evolve simultaneously .
we show that the relationship is partially disguised , however , by the local properties of the surrounding material , the extinction , and short - term flux variability .
[ firstpage ] stars : evolution stars : formation infrared : stars |
experiments on mesoscopic quantum systems , and sophisticated numerical techniques for quantum many body theory , promise progress on deep questions about the relationship between microphysics and thermodynamics @xcite . we still require analytical theory , however , to clarify the questions .
for example , it should be well known that heat consists microscopically of energy held by degrees of freedom whose evolution is too rapid to perceive or control on macroscopic time scales @xcite .
this simple formulation is nonetheless surprising to many physicists , because heat is rarely discussed explicitly except in terms of ensembles .
since the ensembles themselves are justified fundamentally as proxies for time averaging over rapid microscopic evolution , the role of time scale separation in defining heat is well established as an implicit assumption ; but it is underdeveloped as an explicit theory . investigating
the mesoscopic onset of thermodynamics will require fully developing this theory , because the applicability of ensembles is not obvious in small systems .
microscopic energy that remains adiabatically frozen in fast evolution may of course simply be ignored on longer time scales .
what the first law of thermodynamics means from a microscopic point of view , however , is that energy can be exchanged slowly between slow and fast degrees of freedom : heat and work may transform into each other . to investigate the mesoscopic onset of thermodynamics , therefore , a simple dynamical model that provides a minimal representation of heat exchange offers a useful paradigm . in this work we present a simple but non - trivial model for two quantum subsystems in thermal contact , and show how heat exchange between them may be described in familiar quantum mechanical terms .
each subsystem is a pair of bosonic modes coupled so as to represent an idealized josephson junction a two - mode bose - hubbard system .
we specify that the subsystem characteristic timescale the josephson frequency is to be the shortest one in the entire system , so that observation and control of the system can address only much longer time scales .
each subsystem then has , in isolation , exactly two conserved quantum numbers : particle number and energy . for a fixed number of particles ,
the energy may still be varied independently by exciting the josephson mode ; if the rapid josephson oscillations are averaged over , then states of any energy still count as equilibrium .
thus , the internal josephson oscillations of each subsystem are to be regarded as heat .
this identification is bold enough that it may need some defense .
the 2-mode bose - hubbard model , which has been very extensively studied from many points of view , is certainly a rather minimalist model for a thermal system .
it is technically ergodic , inasmuch as the corresponding classical system explores the entire energy surface ( for fixed particle number ) ; but only because that surface is one - dimensional , so that time evolution has no choice but to explore it all .
the system is integrable , not chaotic ; and chaos is often invoked as a justification for statistical mechanics .
but the precise relevance of chaos to macroscopic behavior , as opposed to time scale separation alone , is far from fully understood .
according to the explicit logic normally used to justify ensemble treatments of dynamical systems , and so relate thermodynamics to mechanics , the fact that time averaging reproduces the phase space average over the energy surface should allow us to consider the n - particle excited states of 2-mode bose - hubbard as equilibrium states , as far as any measurements or control operations performed on much longer time scales are concerned . and
if the standard appeal to time averaging allows us to claim our n - particle excited josephson states as equilibria , then it is at least qualitatively reasonable to say that the energy of those excitations is heat . for a macroscopic sample of ordinary gas held in a fixed external potential , with a fixed number of atoms ,
there is a one - parameter family of equilibrium states , distinguished by total energy ( usually expressed as temperature ) .
if we raise this energy without changing the external potential or the particle number , we say that we have added heat to the gas .
indeed this is almost the definition of heat , practically speaking .
if heat means anything , it must mean this .
we therefore argue that the excitations of a two - mode bose - hubbard system can indeed provide a simple model for heat , if they are involved in a dynamical process that occurs slowly compared to their josephson frequency .
we arrange for this by placing two such two - mode subsystems in thermal contact , through an additional josephson tunneling between them , with a much smaller rate coefficient than that of intra - subsystem tunneling .
the two subsystems can thereby exchange particles , on the slow time scale ; but they can also exchange heat , as the amplitudes of their fast josephson oscillations beat slowly back and forth .
our combined four - mode bose - hubbard system is thus a candidate minimalist model for heat exchange .
the implications of this model for mesoscopic and quantum thermodynamics depend simply on whether our josephson heat really behaves as heat
. if it does not , because our integrable system is in some way too trivial , then this will be a useful insight into the precise role of chaos .
even a reductio ad absurdum is still a reductio .
but in fact we will show that the beating josephson excitations do behave very similarly to heat in a recognized , though unusual , form : second sound , which is normally described as a temperature oscillation .
we identify a nonlinear collective effect that makes josephson oscillations of heat
second josephson oscillations the slowest collective excitation of the 4-mode system about an excited equilibrium state .
our analytical derivation of this effect , from a simple quantum mechanical hamiltonian , indicates how thermodynamics may emerge from quantum dynamics in small systems .
our complete system is thus a rather symmetric four - mode bose - hubbard model , with two sharply differing tunneling rates .
such systems may be realized to a good approximation in current cold atom laboratories @xcite , either using four closely neighboring microtrap potentials , or with a double well trap holding atoms with two internal states coupled by a rabi laser @xcite .
our system can probably also be realized at least classically with superconducting josephson junctions , but with the cold atom context in mind , we will refer to the particles as atoms . by exploiting the small ratio between the slow and fast tunneling rates , we are able to analyze collective excitations in this model beyond the linear bogoliubov approximation , and
show that the low - frequency behavior of the joint system consists of two josephson oscillations : the elementary bogoliubov excitation whereby the two subsystems exchange atoms with each other , and a nonlinear beat mode by which they exchange heat .
insofar as a josephson oscillation represents a two - mode analog of bogoliubov zero sound in a dilute bose gas , we identify our second josephson mode as an analog of second sound ( see for example @xcite ) .
we offer it as a simple example of a typical phenomenon of mesoscopic thermodynamics .
our system s four - mode bose - hubbard ( bh ) hamiltonian with onsite interaction @xmath0 and index @xmath1 is @xmath2 where @xmath3 is the high tunneling rate within each subsystem pair , and @xmath4 the low rate between the two subsystems .
@xmath5 conserves the total atom number @xmath6 , and in the present paper we assume for simplicity an eigenstate of @xmath7 with large eigenvalue @xmath8 . the global phase zero mode corresponding to @xmath7
is therefore eliminated , and we are left with the three nontrivial tunneling modes illustrated in fig . 2 . in the linear bogoliubov approximation
the frequencies of these elementary excitations above the @xmath8-atom ground state are @xmath9 these three modes are josephson oscillations , and @xmath10 , @xmath11 , @xmath12 are the three josephson frequencies for the three independent josephson oscillations shown in fig .
as is well known , these collective mode frequencies are shifted away from the pure tunneling ( or for internal transitions , rabi ) frequencies @xmath4 , @xmath3 by the nonlinear interactions .
repulsive interactions ( @xmath13 ) raise the josephson frequencies , while attractive interactions ( @xmath14 ) lower them . in the @xmath15 case we consider ,
only one of these three elementary modes is slow , and the others are fast . since excitations of the bogoliubov modes provide a complete basis for the @xmath8-particle many - body hilbert space , it may seem puzzling that there are in fact ( as we shall see ) two distinct collective modes in this system with low frequencies ( of order @xmath4 rather than @xmath3 ) .
this situation is actually common in nonlinear quantum many - body systems , however .
the ` type ii ' excitations in the one - dimensional lieb - lieniger gas @xcite , second sound in superfluid helium @xcite , and even ordinary sound in air , are all examples of low - frequency collective excitations that are not to be found among a system s low - frequency elementary excitations , despite the fact that these are a complete set .
the resolution of this paradox is that the frequency of beats between our system s two high frequency modes , @xmath16 is of order @xmath4 , and it is the interference beat between the two nearly degenerate high - frequency modes that provides the extra low - frequency collective mode . what we now show
, by calculating to leading order in the small frequency ratio @xmath17 , but nonlinearly in excitation amplitude , is that within the low - frequency regime this beat mode is an independent excitation in its own right , whose frequency is shifted away from @xmath18 by nonlinear effects .
to go accurately beyond the linear approximation within the standard symmetry - breaking bogoliubov theory requires careful treatment of zero modes , but it is straightforward if we follow a number - conserving approach @xcite . applying the holstein - primakoff transformation ( hpt ) @xcite , we re - write @xmath19 in terms of the total atom number on each side @xmath20 , and atom - moving operators @xmath21 that commute with @xmath22 : @xmath23 it is easy to show that @xmath24=1 $ ] and @xmath25=0 $ ] . in these terms
the hamiltonians of the subsystems read @xmath26 we can then implement the bogoliubov formalism by defining @xmath27 and choosing @xmath28 to diagonalize the quadratic terms in @xmath29 while maintaining @xmath30=1 $ ] .
since we are interested in dynamics slow compared to @xmath3 , and in josephson oscillations of finite but not extreme amplitude , we will apply the rotating wave approximation ( rwa ) by dropping from @xmath29 all terms that do not commute with the leading term proportional to @xmath31 . assuming the regime of large occupations for simplicity , we omit corrections of order @xmath32 , and obtain for @xmath33 @xmath34 in the quadratic term of ( [ halpha1 ] ) we note the familiar josephson frequency of the excitations created by @xmath35 , which we have been unable to resist naming ` josons ' .
the co - efficient of the term quartic in @xmath36 is in general of the same order as @xmath0 , but opposite sign ; this reflects the well - known fact that the period of nonlinear josephson oscillations lengthens with amplitude @xcite . in quasiparticle terms , we may say that where atoms repel each other , josons attract ; and _ vice versa_. to simplify the description of the slow tunneling between the @xmath37 subsystems , we now perform a similar hpt for our two atom numbers , while assuming that our system is in an eigenstate of the conserved total @xmath38 , with eigenvalue @xmath8 .
this defines a number - conserving operator @xmath39 , which satisfies @xmath40=1 $ ] ; it transfers atoms between the @xmath41 and @xmath42 subsystems . in the large @xmath8 limit that we assume for simplicity , this provides @xmath43 \end{aligned}\ ] ] up to terms of @xmath44 .
we can therefore re - write our full hamiltonian ( [ hamiltonian ] ) in terms of @xmath45 , @xmath39 , and @xmath8 . since @xmath11 is on the order of the problem
s high frequency @xmath3 , we see that to leading ( zeroth ) order in @xmath17 , our entire hamiltonian is simply @xmath46 , for @xmath47 in classical terms , the ` total joson number ' @xmath48 is an adiabatic invariant . by means of an rwa
we can now compute low frequency dynamics nonlinearly in excitation amplitude , but to leading order in small frequency ratios , by dropping all terms in @xmath5 that do not commute with @xmath48 .
since @xmath31 are not separately adiabatic invariant , we must retain terms such as @xmath49 .
our result is @xmath50 in the first two lines we can recognize two different forms of the standard single josephson junction hamiltonian .
the first line provides josephson oscillations of atoms between @xmath41 and @xmath42 subsystems , with josephson frequency @xmath51 , while the second implies _
josephson oscillations of josons_. the two josephson modes are exactly analogous , except that @xmath52 and @xmath0 are of opposite sign , as anticipated above following ( 5 ) .
the third line couples the atom and joson modes , and arises because the frequency of fast josephson oscillations in each @xmath37 subsystem depends on the respective atom number .
we can now perform a final hpt to express @xmath36 in terms of @xmath48 and a @xmath53-conserving operator @xmath54 that transfers josons between subsystems . as an adiabatic invariant , @xmath48 is just as good a conserved quantity within the low - frequency regime as @xmath7 itself , and so we may let @xmath55 in the low - frequency theory . to identify collective excitations , we then linearize ( [ hfull ] ) in @xmath56 to obtain for fixed @xmath57 @xmath58 in terms of the atom josephson frequency @xmath59 and the new josephson frequency for josons @xmath60 the decoupled collective mode frequencies in ( [ hfinal ] ) are @xmath61^{1/2}\!\!\!\!\!\!\!\!.\ ] ] we observe that in the low - amplitude regime @xmath62 , either @xmath63 so that @xmath64 ; or else @xmath65 and thus @xmath66 so that the third line in ( [ hfinal ] ) is much smaller than the first two .
thus for @xmath62 the low frequency atom josephson frequency @xmath67 is always essentially unchanged from the standard bogoliubov theory indicated in fig .
( 2 ) ; but it is joined by a second josephson oscillation whose frequency @xmath68 can be significantly lower ( for @xmath13 ) , because @xmath52 is negative . this is our main result .
note that the linearization that yields ( [ hfinal ] ) is different from the standard bogoliubov approximation of retaining only quadratic terms in fluctuations of the original @xmath69 about the mean - field ground state .
it is a resummed bogoliubov formalism , in which certain dominant nonlinear interactions among atoms have been included in the linearized interactions among josons , which are nonlinearly related to atoms through their definition via the hpt .
one may also say that we do pursue the standard ( conserving ) bogoliubov approach , but applied to a two - mode system of interacting conserved atoms and josons . in this sense
the extra low - frequency mode is a non - elementary collective mode in the full theory , but an elementary excitation within the adiabatic effective theory for low frequency dynamics .
although our derivation is post - bogoliubov in the sense of being nonlinear , it is not essentially non - classical , because we have assumed large @xmath8 , and because all our hpt and rwa steps are exactly paralleled in classical adiabatic theory .
we can therefore test our theory of second josephson oscillations numerically by solving the four - mode gross - pitaevskii nonlinear schrdinger equation for the classical limit of our system .
we excite intra - subsystem josephson excitations with an oscillating drive , as follows .
we relax to an @xmath8-particle mean field ground state , then excite josons by adding a time - dependent potential tilt to each subsystem , which varies periodically with the josephson frequency @xmath11 .
we excite the second josephson mode by giving the tilt drives slightly different amplitudes for the @xmath41 and @xmath42 pairs , so that slightly more josons are excited on one side than on the other . the resulting evolution , after the drive is turned off , is shown in fig . 3 for a typical case . and
@xmath70 in the mean - field system for the parameters @xmath71 , @xmath72 , @xmath73 , @xmath74 and @xmath75 .
red solid line : numerical mean - field @xmath76 .
blue solid line : oscillation at second josephson frequency @xmath77 computed in the text , fitted for phase .
blue dotted line : oscillation at beat - frequency @xmath78 .
black dashed line : numerical @xmath70 . ]
the agreement with our linearized theory is excellent , and the nonlinear shifting of the second josephson frequency well below the trivial bogoliubov beat is clear .
some atom population oscillation between @xmath41 and @xmath42 subsystems occurs at frequency @xmath79 , this mode having been slightly excited by our driving procedure , but the main atom oscillation is exactly opposite in phase to the joson oscillation .
we note that while in incompressible superfluids second sound includes no include significant particle density change , the coupling of density and heat oscillations is expected for second sound in compressible superfluids @xcite .
we interpret the coupling between the slow atom and joson oscillations as a simple example of the interconversion of heat and work .
the question whether our proposed oscillation of heat carries oscillating entropy is as much about the microscopic definition of entropy , as about the validity of our identification of heat
. our josons do carry at least one physically meaningful form of entropy : the von neumann entropy of the single - particle reduced density matrices @xmath80 , projected onto either half of our system .
these are simply two - by - two matrices , diagonal in the basis of even and odd modes excited by @xmath81 : @xmath82\!\ ! , \label{palpha}\end{aligned}\ ] ] for @xmath83 .
the local single - particle entropies @xmath84 on each side , @xmath85\end{aligned}\ ] ] are genuine entropies inasmuch as they characterize the first order coherence , or equivalently the condensate depletion , within the two - mode bose - hubbard subsystems .
they are direct analogs , for their respective two - mode subsystems , of the single - particle entropy which has been used since boltzmann to represent thermodynamic entropy in statistical mechanics .
since @xmath86 is a simple function of @xmath87 , it is clear that our josons carry entropy at least in this single - particle sense .
one could even consider letting @xmath88 and @xmath89 in ( [ palpha ] ) and ( [ entropy ] ) , in order to define hermitian entropy operators @xmath90 which would be ordinary quantum mechanical observables .
one of the basic questions in quantum thermodynamics is whether quantum entropy should be an observable operator just like quantum energy , with state - independent eigenvalues that appear as measurements with state - dependent probabilities , or whether it should remain a fundamentally statistical quantity like the full von neumann entropy @xmath91 , which is a nonlinear function of the density operator @xmath92 that represents the quantum state of an open system .
our simple model thus addresses this question by suggesting a form of entropy operator , whose performance in the role of thermodynamic entropy may be assessed experimentally . the general implication of our work for mesoscopic quantum thermodynamics would seem to be that it is rather easy to identify first law phenomenology , whereby parametric oscillations among high - frequency internal degrees of freedom provide those extra contributions to the low - frequency energy budget which in thermodynamics are called heat .
inasmuch as this heat oscillates in our case , rather than diffusing , we must recognize that second law phenomenology , with irreversible spontaneous processes , may be more elusive in small quantum systems .
on the other hand , the full range of quantum dynamics in even the four - mode bose - hubbard model , in parameter ranges not considered here , is rich beyond the scope of this paper .
it includes quantum chaos @xcite , whose role in mesoscopic irreversibility amply deserves further study .
support from the deutsche forschungsgemeinschaft via the graduiertenkolleg `` nichtlineare optik und ultrakurzzeitphysik '' is gratefully acknowledged . | as a model for mesoscopic quantum systems in thermal contact , we consider a four - mode bose - hubbard model with two greatly differing tunneling rates . by a series of holstein - primakoff transformations we show that the low - frequency dynamics of this system consists in general of two slow josephson oscillations , rather than the single slow mode predicted by linear bogoliubov theory .
we identify the second slow josephson oscillation as a heat exchange mode analogous to second sound . |
nuclear magnetic resonance ( nmr ) is based on the resonant absorption of radio frequency ( rf ) electromagnetic radiation by nuclear spin - split states by a magnetic field .
the absorption of rf field is limited by small differences of the populations of the nuclear - spin states in thermodynamic equilibrium .
therefore nmr requires probing of a macroscopically large number of nuclei ( @xmath0 ) @xcite .
the nmr detection sensitivity can be greatly improved by preparing a nuclear - spin state with degree of polarization higher than the thermodynamic equilibrium value at a given magnetic field @xcite .
additional improvement is achieved by using optical methods to detect the state of the nuclear - spin system @xcite .
the efficiency of optically detected nmr was demonstrated for bulk semiconductors @xcite , quantum wells @xcite , and even quantum dots ( qds ) @xcite , containing @xmath1 nuclear spins only , thanks to the high values of dynamic nuclear - spin polarization ( dnp ) , of the order of tens percent , that can be achieved in qds by optical pumping @xcite . for self - assembled qds
the observation of nmr is complicated by the large spread of parameters in a qd ensemble @xcite .
usually , single qd spectroscopy is used to diminish this spread @xcite . however
, a considerable broadening of the resonances remains even at the single qd level @xcite .
the main origin of this broadening in self - assembled qds is the inhomogeneous quadrupole splitting of the nuclear - spin states by the strain - induced gradient of crystal field within a dot @xcite .
the broadening results in overlapping resonances from the same and from different isotopes , making their identification difficult . here ,
we show that such an identification is nevertheless possible .
we investigate two samples with self - assembled ( in , ga)as / gaas qds grown by the stranski - krastanov method and annealed at different temperatures .
the annealing gives rise to a decrease of the deformation of the crystal lattice , caused by the mismatch of the qd and barrier lattices and , hence , to a decrease of the inhomogeneous quadrupole splitting of nuclear states .
we make use of polarization modulation of the optical pumping of the dots .
such modulation has been the common basis for the all - optical nmr technique @xcite . as the hyperfine interaction of nuclei with oriented electron spins
can be described as an effective magnetic field ( the knight field ) acting on the nuclei @xcite , the modulation of this field caused by the modulated optical pumping is similar to the action of an rf field @xcite .
optical detection of nmr under light - polarization modulation brings about a resonance cooling of nuclear spins or , in other words , a cooling of nuclear spin system in the rotating frame @xcite .
the effect was observed in a magnetic field oriented perpendicular to the excitation light beam ( voigt geometry ) .
resonance cooling results in appearance of overhauser field oriented parallel or antiparallel to the static magnetic field @xcite . in our study
we exploit another resonant effect connected with the modulation of the polarization of the optical pumping as described in ref . .
if the modulation frequency coincides with that of a nuclear - spin precession about the magnetic field , considerable nuclear - spin polarization perpendicular to the magnetic field appears in the system .
this dnp component favorably increases the electron - spin polarization monitored in the experiment . as a result , additional maxima associated with the nuclear resonances
appear in the magnetic field dependence of the electron spin polarization , i.e. , in the hanle curve . to accurately determine the positions of these nuclear resonances ,
we have measured the hanle curves at different modulation frequencies varied from units of kilohertz up to of about @xmath2 mhz . an rf field synchronized with the optical modulation
was applied to magnify the resonant effects .
identification of the different resonances was done by a theoretical analysis based on modeling the quadrupole - zeeman splitting of the nuclear - spin states .
comparison of the calculated results with the experimental data allows us to identify several resonances corresponding to transitions between the states characterized by spin projection @xmath3 for all the nuclei contained in the qds and by @xmath4 for the in and ga nuclei . besides , resonances with transitions between the higher spin - projection states , @xmath5 , for the in nuclei and between the states having different projections , @xmath6 , for the in and ga nuclei were identified .
the large broadening of resonances is confirmed to be caused by the spread of strain within the qds .
we study a heterostructure containing 20 layers of self - assembled ( in , ga)as qds sandwiched between gaas barriers .
the barriers are @xmath7-doped by donors with a concentration , which supplies every dot with , on average , a single resident electron after the donor ionization .
the structure was grown by molecular - beam epitaxy on a ( 001)-oriented gaas substrate .
rapid thermal post - growth annealing of the structure provides inter - diffusion of indium and gallium atoms between the qds and gaas barriers so that the nominally pure inas qds are partially enriched by the in - diffusion of ga atoms which increases the band gap , resulting in a blue shift of the ground qd optical transition @xcite .
the annealing also causes a decrease of built - in strain in the qds that is an important source of the quadrupole splitting of the nuclear - spin states .
we study two pieces of this structure annealed at @xmath8 c ( sample 1 ) and at @xmath9 c ( sample 2 ) .
the samples were placed in a cryostat with a superconducting magnet such that the magnetic field could be applied perpendicular to the structure growth axis ( voigt geometry ) coinciding with the @xmath10 $ ] crystallographic axis .
the experiments were performed at a sample temperature @xmath11 k. to detect dnp , we look over the circular polarization of the pl from the qds . to excite the qds we use a continuous - wave ti : sapphire laser , which photon energy is tuned to the optical transition of the wetting - layer exciton .
the ground - state pl is dispersed by a half - metre spectrometer and is detected with a silicon avalanche photodiode .
the degree of circular polarization of the pl , @xmath12 , is measured using a photo elastic modulator operated at a frequency of @xmath13 khz and a two - channel photon - counting system . here
, @xmath14 is the pl intensity for co- ( cross- ) polarization relative to that of excitation . in the maximum of pl band of the qds ,
the polarization is negative ( fig .
[ fig : pl ] ) and reflects the mean spin polarization of the resident electrons along the optical axis ( @xmath15 axis ) , as has been extensively discussed earlier @xcite . due to the interaction of the resident electrons with the qd nuclei , the negative circular polarization ( ncp )
can be used as a sensitive tool to monitor the nuclear spin state @xcite . in our experiments
, the helicity of the optical excitation was periodically modulated between @xmath16 and @xmath17 by an electro - optical modulator followed by a quarter - wave plate .
in addition , an rf field was applied by means of a small coil oriented parallel to the optical @xmath15 axis .
the rf - field frequency coinciding with the frequency of optical modulation was fixed , while the hanle curves were measured by scanning the magnetic field .
the pl of the qds at low excitation density shows a broad emission contributed mostly by the ground state optical transition but also the first excited state transition shows up on the high - energy flank , see fig .
[ fig : pl ] .
we focus here on the behavior observed for the lowest transition .
the corresponding emission line has the half - width at half - maximum of about 10 mev , centered at 1.418 ev for sample 1 and 1.337 ev for sample 2 .
the ncp is observed across the whole emission line with a peak amplitude reaching 30% at optimal excitation density , which is about 30 w/@xmath18 for sample 1 and three times larger for sample 2 .
the analysis shows @xcite that the upper limit of ncp is determined by the fraction of qds containing a single resident electron .
optimization of the excitation density allows one to fully polarize the electrons .
note that the ncp is found to depend very sensitively on any residual magnetic field adding to the external field applied in the voigt geometry .
therefore the residual magnetic fields including the earth magnetic field were carefully compensated in our experiment by means of helmholtz coils .
magnetic - field dependencies of ncp were measured at the pl maximum for different frequencies of polarization modulation of the excitation .
all hanle curves presented in this paper were obtained by scanning the magnetic field in one direction , defined here as from negative to positive values . as an example , the hanle curves for sample 1 are shown in fig .
[ fig : hanle ] .
the hanle curve measured with nonmodulated polarization of the excitation ( black line ) contains the so - called w - structure , which consists of a very narrow central peak and two maxima separated from the peak by dips .
the appearance of the w - structure results from polarization of this nuclear spins along the magnetic field @xcite .
when the polarization of excitation is modulated , the hanle curve becomes strongly modified . as seen from fig .
[ fig : hanle ] , the central peak drastically broaden and additional intense maxima appear .
they move to higher magnetic fields with increasing modulation frequency .
the frequency dependence indicates the resonant nature of the maxima .
the rapid decrease of their amplitude at higher frequencies does not allow , however , to observe them in wide frequency range , which aggravates their identification . to increase the amplitude of the resonant peaks in the hanle curves and to extend the range of transverse magnetic fields , in which the peaks can be observed , we applied to the sample an alternating rf magnetic field of small amplitude along the optical axis .
the frequency of the alternating field was identical to that of the modulation of the optical excitation .
the phase difference , @xmath19 , between the polarization and rf modulations was varied .
the effect of the rf field on the hanle curve for sample 1 is shown in fig .
[ fig : phase ] .
obviously , the effect strongly depends on @xmath19 @xcite . at some optimal @xmath19 ,
the wings of the hanle curve are greatly increased and the overall width of the curve is strongly enlarged @xcite .
phase inversion of the rf field relative to the optimal case results in the opposite effect
the additional wings almost disappear and only the central part of the hanle curve survives . for intermediate @xmath19 , an additional structure appears in the hanle curve , indicating asynchronous magnification or suppression of the different resonances .
this non - synchronicity facilitates the separation of resonances .
similar experiments were done for sample 2 where the hanle curves are considerably wider due to the stronger quadrupole interaction .
the effect of the rf field for this sample is similar to the one for sample 1 . as seen from fig .
[ fig : phase ] , the wings of the hanle curve in most cases are rather smooth due to the large resonance broadening , whose origin we will discuss below . therefore , to uniquely identify the position of the nmr resonances for different nuclei isotopes , we measured a set of hanle curves at different modulation frequencies @xmath20 and phase differences @xmath19 .
the amplitude of the rf field has been chosen to be relatively small ( fraction of millitesla ) and has been optimized for the best separation of the resonances .
the experimental data were compared then with the results of theoretical modeling of the zeeman splittings of the nuclear spin states in the qds in presence of strain .
the analysis of the whole set of hanle curves measured at different modulation frequencies allowed us to identify the observed resonances , to retrieve the nuclear zeeman splittings , and to determine the nuclear quadrupole splittings for the studied samples .
the results of this analysis are discussed below .
an estimation of the crystal lattice deformation in qds annealed at different temperatures was given in ref . .
here , we give insight into the modeling of the strain - induced electric field gradient in the annealed qds .
the results are summarized in fig .
[ fig : efg ] . the annealing of the ( in , ga)as / gaas heterostructure initiates an interdiffusion of in and ga atoms followed by an increase of the qd size and a decrease of the in dot content .
we model this process by solving a diffusion equation for the spatially variable indium fraction , @xmath21 , in the in@xmath22ga@xmath23as compound ( see ref . for details ) .
figure [ fig : efg](a ) shows the cross - section of the resulting distribution of @xmath21 for the qd annealed at a temperature above @xmath24 c. it should be noted that the electron localization volume increases with annealing due to both , the increase of effective size of the dot and the decrease of the potential well depth for the conduction - band electron @xcite .
additionally , the strain in the qd crystal lattice , arising from the lattice mismatch of the qd and barrier materials , also changes with annealing .
for modeling the quadrupole - zeeman splittings of the nuclear sublevels in the qds , it is important to account for the component of biaxial strain , defined as @xmath25 @xcite .
this component relaxes from the average value of @xmath26 in the unannealed qd down to @xmath27 ( sample 1 ) and @xmath28 ( sample 2 ) in the annealed qds , as calculations show .
these values agree well with our experimental data as demonstrated below .
the calculations of @xmath29 were made for a hat - shaped qd , which we take as characteristic for the dots in the ensemble .
taking advantage of the cylindrical symmetry of our model , the biaxial component of the strain can be rewritten as @xmath30 .
as we are interested in ensemble - averaged values of the observables , the strain - tensor components are calculated using continuum elasticity theory using the transverse - isotropic - media approximation @xcite , which is valid in our case . to obtain microscopic insight into an individual qd
, a more complicated atomistic approach should be taken @xcite .
the crystal lattice deformation is enhanced within the qd and causes there the appearance of an electric field gradient ( efg ) at nuclear sites .
the main origin of the efg is the strain due to lattice mismatch .
there is also a crystal field gradient due to the statistical population of anion sub - lattice sites with in and ga atoms .
this statistics concerns a fraction of the as nuclei and gives rise to such a strong quadrupole splitting that the corresponding resonances can be observed only in very high magnetic fields , which are available in our experiments .
the efg relates to the elastic strain by a fourth - order tensor @xcite , i.e. , @xmath31 where summation over iterated subscripts is assumed . for @xmath32 compounds ,
there are only three nonzero components of the @xmath33-tensor and only two of them are independent @xcite , @xmath34 , @xmath35 , and @xmath36 , written in the voigt notation .
the goal of our modeling is to obtain the principle components of the strain - induced efg , @xmath37 , @xmath38 , and @xmath39 . in the first step , we calculate all nonzero components of @xmath40 using eq . and , in the second step , we diagonalize the obtained matrix @xmath41 . by doing this , we find the spatial dependencies of the principal component of the efg , @xmath39 , and the asymmetry parameter , @xmath42 , for all the nuclei in the dot .
the results of these calculations are shown in figs .
[ fig : efg](b)[fig : efg](d ) for the @xmath43ga , @xmath44 in , and @xmath45as nuclei , respectively . as one can see from the figures , the principal component of the efg tensor
is directed along the @xmath15-growth axis of the structure for all the nuclei .
additionally , the efg direction at the as nuclei is opposite to that for the other nuclei because of the different signs of @xmath34 and @xmath36 for anions and cations ( see ref . for details ) .
large variations of the parameter @xmath46 are found for the as nuclei located in the barrier to the left and right of the qd . inside the dot where the ground - state electron density is large ,
this variation is negligibly small . to illustrate that
, we show in figs .
[ fig : efg](b)[fig : efg](d ) the isosurfaces of the electron density distribution , calculated in the effective - mass approximation @xcite . at zero magnetic field ,
the nuclear - spin states are split by the quadrupole interaction into doublets @xmath47 for nuclei with spin @xmath48 , possessing a non - zero quadrupole moment @xmath49 .
the magnitude of the splitting , expressed in terms of the frequency @xmath50 , is determined by @xcite @xmath51 since the principal component of efg is aligned along the @xmath15 axis , we set @xmath52 as the asymmetry parameter @xmath46 of the quadrupole interaction is neglected , the hamiltonian describing the splitting of the nuclear - spin states is @xcite @xmath53 , \label{eq : ham}\ ] ] where @xmath54 is the gyromagnetic ratio .
the first term of the hamiltonian describes the zeeman interaction and the second term is the quadrupole interaction .
the matrix of this hamiltonian using the basis states @xmath55 is diagonalized to obtain the energies of the nuclear spins under the influence of the zeeman and quadrupole interactions . for the calculations ,
we use the values of @xmath34 , @xmath49 , and @xmath54 collected in table [ tab : nucparam ] .
.summary of nuclear parameters .
the values of @xmath34 , @xmath49 , and @xmath56 are taken from refs .
, , and , respectively .
[ cols="<,^,^,^,^",options="header " , ] the spin splittings calculated for the ga , as , and in nuclei in the qds under study are shown in fig .
[ fig : fitting ] .
one can see that the resonances for transitions between the split - off states @xmath57 are observed at larger magnetic fields than those for the @xmath58 states . besides
, the resonances at a given frequency for the split - off states are shifted to larger magnetic fields in sample 2 relative to those in sample 1 due to the larger @xmath50 .
+ + the shape of the resonances contributing to the hanle curves is determined by their physics origin . under action of optical pumping with non modulated polarization , dnp appears aligned parallel to the external magnetic field .
the alternating magnetic field from the radio - frequency excitation close to the resonant frequency of nuclear spin transitions , applied along the optical axis , tilts this nuclear - spin magnetization . when scanning the external field around the resonance , the tilting angle changes from @xmath59 to @xmath60 taking on the value @xmath61 at the resonance .
consequently , the sign of the projection of the nuclear - spin magnetization onto the axis of external magnetic field changes while scanning through the resonance .
this results in a dispersion like singularity in the hanle curve .
such resonant profiles were previously observed in different systems and discussed in literature @xcite .
in contrast to non modulated excitation , optical pumping with circular polarization modulated at a frequency close to a resonance in a qd ensemble leads to a nuclear - spin magnetization aligned _ perpendicular _ to the external magnetic field .
this dnp component is created by nuclear spins , which coherently precess in the external magnetic field .
synchronization of the precession of the dnp field with the modulated knight field maintains the electron - spin polarization .
the latter results in positive peaks on the hanle curve .
these resonances can be observed outside the electronic peak ( @xmath62 peak ) in the hanle curve given by pure electronic polarization @xcite .
analysis of our experimental data shows that all the resonances must be modeled by positive peaks rather than the dispersion like curves for which we used gaussians .
the resonances corresponding to the @xmath63 transitions overlap so that their phenomenological analysis does not allow us to derive definite conclusions about their nature . to simplify the fit of experimental data
, we also used gaussians for their modeling @xcite .
so , it is assumed that each hanle curve is a superposition of gaussian peaks centered near the nmr positions calculated for each type of nuclei .
besides , the hanle curve is contributed by an electronic peak ( @xmath62 peak ) observed in absence of dnp .
the shape of this peak is determined experimentally by measuring the hanle curve using an amplitude modulation of the excitation with a large ratio of dark to bright intervals . as shown in ref .
, dnp is strongly suppressed under such experimental conditions .
the calculation of splittings described above allows us to determine the magnetic - field positions of nuclear - spin resonances at each particular modulation frequency and , thus , to fit the hanle curves .
the fitting parameters are the width and the amplitude of the gaussians .
the magnetic - field position of the resonances is taken , at the first step , from the modeling ( see fig . [
fig : fitting ] ) and then slightly varied to obtain the best fit to the experiment . for more precise determination of the resonance positions ,
a set of hanle curves measured with different phase shifts @xmath19 between the polarization modulation and the rf field is analyzed .
the resonance positions are found to be close to those obtained from the modeling .
however , the widths of the resonances are much larger than those typically observed in standard nmr . the origin of this broadening of the resonances will be discussed in the next section .
examples of deconvolutions of the hanle curves into gaussian - like resonances are given in fig .
[ fig : hanlefit ] .
the figure shows the hanle curves measured at moderate modulation frequencies when only the transitions @xmath63 and @xmath64 contribute to the curves .
as one can see , the strongly broadened resonances overlap .
the central part of the hanle curve for sample 1 can be well modeled by the sole contribution of the transitions @xmath63 of @xmath43ga isotope . in the hanle curve of sample 2 , contributions of transitions in the in nuclei can be also resolved .
the wide part of hanle curves is given by the transitions @xmath64 of the in and ga nuclei .
the relative amplitude of these resonances is found to be very sensitive to the experimental conditions , in particular , to the phase shift @xmath19 and does not reflect the content of in and ga nuclei in the qds .
for example , although the in content is larger in sample 2 , the respective resonance @xmath64 of the in nuclei gives rise to a stronger peak in the hanle curve of sample 1 for the experimental conditions used for recording the data in fig .
[ fig : hanlefit ] .
a comparison of the resonance positions obtained from the experiment with those obtained from the calculation of the zeeman splittings of the nuclear - spin states is shown in fig .
[ fig : resonancepositions ] . figures [ fig : resonancepositions](a)[fig : resonancepositions](d ) demonstrate the data for the observed transitions @xmath63 , @xmath64 , and @xmath65 .
variations of the phase shift @xmath19 between the polarization modulation and the rf field slightly change the resonance positions .
this is reflected in some horizontal spread of the experimental data in the figures .
figures [ fig : resonancepositions](e ) and [ fig : resonancepositions](f ) show similar data for the transitions @xmath66 and @xmath67 .
the resonance frequencies for these transitions can be determined only with less accuracy .
however , the hanle curve can not be successfully described if these resonances are neglected .
the results obtained show that , in spite of the large broadening of the resonances , most of them can be identified and their behavior can be described by a simple model containing only one free parameter , @xmath29 .
the values of this parameter obtained from experiment for the samples under study are very close to those calculated ( see sec . [
sec : efgmodeling ] ) . , experiencing some spread due to the inhomogeneity of the biaxial strain ( solid and dotted lines ) .
because of this spread , the resonances @xmath68 broaden and the broadening increases with @xmath20 ( for details see the text ) .
( b ) relation between the biaxial strain and the magnetic - field position of the resonance at different frequencies of the transitions @xmath64 in the @xmath43ga nucleus . , title="fig : " ] + , experiencing some spread due to the inhomogeneity of the biaxial strain ( solid and dotted lines ) .
because of this spread , the resonances @xmath68 broaden and the broadening increases with @xmath20 ( for details see the text ) .
( b ) relation between the biaxial strain and the magnetic - field position of the resonance at different frequencies of the transitions @xmath64 in the @xmath43ga nucleus . , title="fig : " ] to @xmath69 khz for sample 1 .
all resonances are normalized to the same integral area .
the resonances corresponding to the transitions @xmath64 for @xmath43ga are shown in panel ( a ) and for @xmath44 in in panel ( b ) .
the thick solid lines are the averaged profiles and the dashed lines are gaussian fits to them : @xmath70 $ ] .
the fitting parameters are @xmath71 and @xmath72 for @xmath43ga , @xmath73 and @xmath74 for @xmath44 in .
+ , title="fig : " ] + to @xmath69 khz for sample 1 .
all resonances are normalized to the same integral area .
the resonances corresponding to the transitions @xmath64 for @xmath43ga are shown in panel ( a ) and for @xmath44 in in panel ( b ) .
the thick solid lines are the averaged profiles and the dashed lines are gaussian fits to them : @xmath70 $ ] .
the fitting parameters are @xmath71 and @xmath72 for @xmath43ga , @xmath73 and @xmath74 for @xmath44 in .
+ , title="fig : " ] the fitting of the hanle curves by sums of resonant peaks ( see figs . [
fig : hanlefit ] and [ fig : resonancepositions ] ) has revealed the large widths of the resonances , which are considerably broader than those observed for quantum wells @xcite . particularly , large widths are observed for the peaks corresponding to the transitions @xmath64 and @xmath65 .
these widths are about an order of magnitude larger than those caused by random dipole - dipole fields ( fraction of mt ) or by the knight field ( of about @xmath2 mt in sample 1 , see ref .
variation of these fields has been considered in ref .
[ ] as the origin of broadening the nmr resonances in single qds .
we assume that a more broader effect on the resonance broadening originates from the spread of strain causing a spread of the quadrupole splittings of the nuclear - spin states .
the spread can be present in each individual qd [ @xcite ] and in the qd ensemble as a whole .
the zeeman splitting of the nuclear - spin states strongly depends on the biaxial strain .
as one can see from figs .
[ fig : resonancepositions](a)[fig : resonancepositions](e ) , the splitting in sample 2 with large strain is smaller than that in sample 1 for the same magnetic field . at a fixed resonance frequency , each particular magnitude of @xmath29 corresponds to a unique magnitude of magnetic field , @xmath75 , at which the resonance is observed .
correspondingly , the spread of strain gives rise to a spread of resonant magnetic fields as illustrated in fig .
[ fig:3dstatistics](a ) .
there examples of resonance profiles , @xmath68 , are shown , which describe the probability of nmr as a function of magnetic field at a given frequency . as seen ,
the profiles are shifted to higher magnetic fields and become broader with increasing frequency which qualitatively explains the experimental observations .
the shape of the resonance profiles @xmath68 should be unambiguously related to the function @xmath76 describing the spread of biaxial strain in the structure . to determine the relation between the functions @xmath68 and @xmath77 )
, we have calculated the magnetic - field dependence of the splitting of the nuclear - spin states @xmath78 for the ga and in nuclei at different values of @xmath29 .
results of these calculations for the @xmath43ga isotope are shown in fig .
[ fig:3dstatistics](b ) .
as one can see , there is a monotonic , almost linear , dependence , @xmath79 , for modulation frequencies @xmath20 fixed in the range used in experiment .
this allows one to interconnect the two probability distributions : @xmath80 where the derivative @xmath81 can be easily calculated .
figure [ fig:3dstatistics](b ) shows that the derivative weakly depends on @xmath82 and is a function of @xmath20 .
the resonance peaks , @xmath83 , obtained from the experiments and modeled by gaussians , can be treated except for some scaling factor , as the probability distribution @xmath68 of nmr at a given frequency . using
the equation given above one can transform @xmath83 into @xmath76 . because the spread of strain does not depend on modulation frequency
, the resonant peaks measured at different frequencies and transformed into functions of @xmath29 should be very similar to each other , if the broadening is actually due to this spread .
the results of such a processing of our experimental data for the transitions @xmath84 of the @xmath43ga and @xmath44 in nuclei are shown in fig .
[ fig : statistics ] .
as seen , variation of the modulation frequency over a wide range from @xmath85 to @xmath69 khz weakly affects the positions and the widths of the resonance profiles @xmath76 .
note , that the width of @xmath83 increases several times with the modulation - frequency the relatively small variations of the parameters of the experimentally determined resonance profiles shown in fig .
[ fig : statistics ] are probably caused by experimental errors and some uncertainty in the fitting procedure .
there is also some difference in the averaged positions and widths of the @xmath76 for the nuclei @xmath43ga and @xmath44 in . in the @xmath43ga ( thick blue line ) and @xmath44 in ( thin red line ) nuclei calculated in the framework of the model described in sec .
[ sec : efgmodeling ] for the sample 1 annealed at 980 c. dashed lines indicate the mean values @xmath86 for @xmath43ga and @xmath87 for @xmath44 in . ] to model the function @xmath76 in a single qd , we perform a simple calculation based on the model described in sec . [
sec : efgmodeling ] . using the distribution of the in fraction , @xmath21 , and the electron density distribution , @xmath88
, one can estimate the distribution of the principal component of the efg tensor by @xmath89 d\mathbf{r}. \label{eq : rho_th}\ ] ] to obtain equivalent distributions for the other isotopes one should replace @xmath21 by @xmath90 ( for ga ) and by @xmath2 ( for as ) in eq . .
the resulting distributions @xmath76 for the in and ga nuclei are shown in fig .
[ fig : rho_th ] .
as one can see , the distributions are relatively wide , which supports the assumption of strain non - homogeneity within the qd as primary source of the resonance broadening shown in fig .
[ fig : statistics ] .
there is a strong asymmetry of the distribution with an abrupt right edge , which corresponds to the maximal value of @xmath29 in the middle of the qd . the experimentally obtained distribution shown in fig .
[ fig : statistics ] is more symmetric , which is possibly due to smoothing of the distribution in the qd ensemble .
there is also some difference in the values @xmath91 for @xmath43ga and @xmath44 in averaged over the corresponding statistics .
the in content is maximal in the qd middle where the efg is also maximal so that the distribution for the in nuclei is shifted to larger values of @xmath29 relative to those for the ga nuclei .
this result correlates well with the experimental observations , see fig .
[ fig : statistics ] .
the broadening of the resonance peaks for the transitions @xmath63 requires some further consideration .
this broadening is considerably smaller than that for the transitions @xmath64 .
however , it is still much larger than the one typically observed in solid state nmr @xcite . to understand the possible origin of this broadening
we should consider the zeeman splitting of the states @xmath58 in presence of quadrupole splitting .
when the magnetic field is small and mostly perpendicular to the principal axis of the efg , the splitting is given by @xcite @xmath92,\ ] ] where @xmath93 is the zeeman splitting of the states @xmath58 in absence of strain and @xmath94 is the quadrupole splitting of the nuclear - spin states into doublets corresponding to @xmath58 and @xmath95 . clearly , when @xmath96 , the splitting @xmath97 is approximately twice larger than the zeeman splitting and almost independent of the deformation .
this means that the spread of @xmath29 can not be responsible for the observed broadening .
we have to assume that the broadening is caused by a deviation of the principle axis of the efg from the direction orthogonal to the magnetic - field .
one possible reason for that is an asymmetry of the qds in the ensemble due to the statistical nature of the assembly of atoms in the qds during the growth process .
another reason might be the inclination of the deformation axis at peripheral parts of the qds @xcite . note that even a relatively small inclination of the axis may cause a remarkable shift of the resonance position .
indeed , when the efg axis is parallel to the magnetic field ( inclination is @xmath61 ) , the splitting of the @xmath58 states becomes twice smaller and corresponds to the ordinary zeeman splitting , @xmath98 .
therefore the spread of inclinations of the efg axis in the qd ensemble can be responsible for the observed broadening of the resonances @xmath99 .
to conclude , the obtained results demonstrate that resonant optical pumping of the electron - nuclear spin system in qds subject to transverse magnetic fields is an efficient tool for studying transition between nuclear spin states split by a magnetic field . using this method , we managed to experimentally detect a number of resonances of the in , ga , and as nuclei in an inhomogeneous qd ensemble , namely those corresponding to the transitions between the @xmath100 , @xmath84 , and @xmath101 states as well as to the transitions between the @xmath102 and @xmath103 states .
the comparison of experimental data recorded on ( in , ga)as / gaas qd samples annealed at different temperatures show that the resonant frequencies strongly depend on the strain of the crystal lattice in the qds .
the strain - induced efg at the nuclear sites splits the nuclear spin states with non - zero quadrupole moment resulting in a remarkable modification of the nmr spectrum .
our model considering zeeman splitting of the nuclear states in presence of quadrupolar interaction has been developed .
it allows us to identify all the observed resonances .
essentially , the model having only one free parameter , @xmath29 , allows one to satisfactorily describe the magnetic field dependences of the resonant frequencies observed .
the obtained values , @xmath104 for sample 1 annealed at @xmath105 c and @xmath106 for sample 2 annealed at @xmath107 c , are in good agreement with those obtained theoretically from the qd structure modeling .
we have also found that the resonances are broadened in qds much stronger than in bulk crystals and quantum wells . in particular , this is valid for the split - off resonances @xmath108 , and @xmath109 .
our analysis shows that the main origin of the resonance broadening is related to the spread of the biaxial strain , @xmath29 .
the value of the spread is of the order of @xmath29 itself .
this work was supported by the deutsche forschungsgemeinschaft , the bmbf quahl - rep 16bq1035 , the eu fet - program spangl4q , and by the russian ministry of education and science ( contract no . 11.g34.31.0067 with spbsu and leading scientist a. v. kavokin ) .
a. i. tartakovskii , t. wright , a. russell , v. i. falko , a. b. vankov , j. skiba - szymanska , i. drouzas , r. s. kolodka , m s. skolnick , p. w. fry , a. tahraoui , h .- y .
liu , and m. hopkinson * 98 * , 026806 ( 2007 ) . the asymmetry of the hanle curves , which is seen in fig .
[ fig : phase ] , is a result of the well - known effect of hysteresis observed for the electron - nuclear spin system in bulk crystals @xcite as well as in qds @xcite .
the hamiltonian of the quadrupole interaction has a simple form in the intrinsic coordinate system of the efg tensor . since the principal axis of the efg coincides with the structural @xmath15 axis , the quadrupole part of the total hamiltonian written in the @xmath110 coordinate frame has the same form as in the intrinsic coordinate system .
i. a. yugova ( private communication ) .
a theoretical analysis shows that , under strong enough optical pumping , the shape of the resonances for the transitions @xmath100 deviates from the dispersion like curve and resembles a gaussian peak . | the photoluminescence polarizations of ( in , ga)as / gaas quantum dots annealed at different temperatures are studied as a function of external magnetic field ( hanle curves ) . in these dependencies ,
remarkable resonant features appear due to all - optical nuclear magnetic resonances ( nmr ) for optical excitation with modulated circular polarization .
application of an additional radio - frequency field synchronously with the polarization modulation strongly modifies the nmr features .
the resonances can be related to transitions between different nuclear spin states split by the strain - induced gradient of the crystal field and by the externally applied magnetic field .
a theoretical model is developed to simulate quadrupole and zeeman splittings of the nuclear spins in a strained quantum dot .
comparison with the experiment allows us to uniquely identify the observed resonances .
the large broadening of the nmr resonances is attributed to variations of the quadrupole splitting within the quantum dot volume , which is well described by the model .
2aexperimentelle physik 2 , technische universitt dortmund , d-44221 dortmund , germany |
quantizing gravity and/or the quantum nature of spacetime have been one of the central themes in theoretical physics in the last century .
the two pillars of modern physics , i.e. general relativity and quantum theory are reluctant to be reconciled despite many efforts . in spite of many theoretical advancements the field of quantum gravity ( qg ) still lacks contact with observations and experiments .
however , in recent years the field of quantum gravity phenomenology has started to attract the interest of the community . since to date
we do not have a fully working quantum gravity model , but a plethora of models each with nice features and drawbacks , we should bet on common features shared by different models .
one of the recurring features in models of qg is the idea that spacetime may be fundamentally discrete .
this idea is common to various approaches even if it presents itself in different ways .
@xcite however , when dealing with discrete spacetime , lorentz invariance violating ( liv ) effects are lurking .
indeed studies of liv are a good example of qg phenomenology that has flourished in the last 20 years or so .
we nowadays have stringent constraints on liv coming from experiments ranging from astrophysical observations to earth based high energy experiments .
together with theoretical arguments , these constraints tightly restrict the number of viable liv theories and suggest that lorentz invariance ( li ) may actually be an exact symmetry of nature telling us something deep about spacetime . from these latter considerations
follows the idea that inspired our investigation : to consider discrete spacetime models that preserve lorentz invariance and study their phenomenology .
in this work we consider a model of discrete spacetime , causal set theory ( cs ) , which is li and its associated phenomenology . we will also propose a way to test non - local features , results of the discreteness itself , with state of the art ( or near future ) quantum experiments .
the paper is organized as follows . in section [ sec2 ]
we briefly introduce the basic principles of causal set theory . in section [ sec3 ]
we explore huygens principle violations in flat 4 dimensional spacetime and possible dimensional reduction in causal set theory . in section [ sec4 ]
we propose a way to test non - local effects using opto - mechanical quantum oscillators .
we conclude in section [ sec5 ] with a discussion and future directions .
causal set theory is an approach to quantum gravity that stems from the assumptions that the causal structure is fundamental and spacetime is discrete . in particular
a causal set is defined as a partially ordered , locally finite set .
@xcite already in general relativity ( gr ) spacetime is a partially ordered set , where the partial order represents the causal structure , but not a locally finite one .
moreover , it is striking that the knowledge of solely the causal structure of spacetime can be proved to be enough , under some technical causality conditions , to completely determine the conformal geometry ( see ref . and reference therein ) .
these results in differential geometry shine a light on the primitive importance of causal structure , which is assumed to be fundamental in cs theory .
finally , the locally finite requirement translates into rigorous terms the physical expectation that spacetime is discrete at scales comparable with the planck scale as suggested by different theoretical hints .
@xcite in what follows we will be interested in the kinematic structure of causal sets which offers interesting phenomenological possibilities for what concerns the study of the effects of the spacetime discreteness .
indeed , different phenomenological models and predictions have been produced in the cs literature , ranging from the prediction of the cosmological constant to swerves model .
@xcite in particular , we will consider causal sets that well approximate flat spacetime in different dimensions .
these causal sets are the ones that can arise with relatively high probability from the poisson process called _ sprinkling_. this process consists in selecting points uniformly at random in minkowski with density @xmath0 ( being @xmath1 the fundamental discreteness scale and @xmath2 the spacetime dimension ) so that the expected number of points sprinkled in a region of spacetime volume @xmath3 is @xmath4 .
the sprinkling process produce a random lattice that preserve lorentz invariance in a well defined sense , in this way we have a kinematical randomness that preserves lorentz invariance while working with a discrete structure .
@xcite however , there is a price to pay : a fundamental non - locality of causal sets .
indeed , consider the nearest neighbours to a given point in a causal set well - approximated by minkowski spacetime .
these will lie roughly on the hyperboloid lying one planck unit of proper time away from that point and therefore will be infinite in number .
this non - locality manifests itself also in the definition of the non - local dalembertian for a scalar field on the causal set .
@xcite the non - local dalembertian is a discrete operator that reduces in the continuum limit to the standard ( local ) wave operator .
the precise form of this correspondence is given by performing an average of the causal set dalembertian over all sprinklings of minkowski , giving rise to a non - local , retarded , lorentz invariant linear operator in the continuum , @xmath5 , whose non - locality is parametrised by a scale @xmath6 .
locality is restored in the limit @xmath7 in which @xmath8 .
we will consider the non - locality scale @xmath6 as a free parameter of the theory on which we would like to cast phenomenological bounds .
the general expression for the non - local dalembertians in flat spacetime of dimension @xmath2 was introduced in ref . and is given by @xmath9 where @xmath10 is a dimension dependent positive integer , @xmath11 and the coefficients @xmath12 , @xmath13 and @xmath14 can be found in equations ( 12)-(15 ) of ref . .
in ref . the canonical quantization of the non - local , free scalar field theory was performed .
a different quantization scheme was investigated , considering non - local field theories that share some common features with the ones deriving from causal sets , with similar results .
in the following we consider two features that emerge from the study of eq . , i.e. huygens principle violations in 4 dimensional flat spacetime and dimensional reduction . where the first can be conceived as pointing towards interesting phenomenological scenario the second has intrinsic theoretical importance since dimensional reduction is a common features of different qg approaches .
huygens principle ( hp ) states that in spacetime dimension @xmath15 , @xmath16 , the green functions of the wave equation have support on the light - cone , while in dimensions d = 2n + 1 they also have support inside the light - cone .
this translates into the fact that the field perturbations can propagate only on the light cone when hp is satisfied .
the fact that hp violations are present in 4-dimensional flat spacetime for the non - local causal set dalembertian was observed in ref . by looking at the support of the retarded green function ,
see fig .
[ gr ] . it has been recently shown that , whenever hp is violated an information channel opens that permits information transmission without energy exchange .
@xcite it is then interesting to observe that , the causal set non - local dalembertian shows hp violations in flat 4-dimensional spacetime . indeed ,
if it is possible to extend such results to the electromagnetic field this will open to the possibility of testing hp violations in the lab using high precision experiment to cast bounds on non - local effects deriving from cs discreteness .
spectral dimension in cs theory was first computed in ref . using a random walker on the graph .
the result was that the spectral dimension , computed in this way , diverges in the limit of short diffusion times .
this behaviour was explained using the non - local nature of causal sets discussed previously .
however , we computed the spectral dimension associated to the non - local scalar field dalembertian finding that , for every dimension , in the uv there is a dimensional reduction to 2 of the spectral dimension , see fig .
@xcite as a function of diffusion time @xmath17 for @xmath18 in dimension @xmath19 and 4 respectively .
the dashed line represents the value of the hausdorff dimension ( @xmath2 ) .
the spectral dimension interpolate between @xmath20 at short scales and @xmath2 at large scales .
, title="fig : " ] as a function of diffusion time @xmath17 for @xmath18 in dimension @xmath19 and 4 respectively .
the dashed line represents the value of the hausdorff dimension ( @xmath2 ) .
the spectral dimension interpolate between @xmath20 at short scales and @xmath2 at large scales .
, title="fig : " ] as a function of diffusion time @xmath17 for @xmath18 in dimension @xmath19 and 4 respectively .
the dashed line represents the value of the hausdorff dimension ( @xmath2 ) .
the spectral dimension interpolate between @xmath20 at short scales and @xmath2 at large scales .
, title="fig : " ] to date the precise link between our result and the one of ref . has not been clarified .
it is possible that the two quantities carry different physical meaning given that they are obtained with different methods and in different regimes of the theory ( albeit having the same name ) , or that the spectral dimension computed in ref .
ceases to be valid at scales when the continuum approximation is no longer reliable .
note moreover that additional hints towards dimensional reduction in causal set theory were recently presented .
in this last section we consider a recently proposed way to test models that present non - locality with opto - mechanical experiments .
@xcite the rationale behind this proposal rests in the fact that quantum mechanical experiments are reaching regimes in which it is possible to test some general ideas stemming from qg , see e.g. ref .
we studied the effect of non - locality in the evolution of coherent states of a quantum harmonic oscillator , a system that is central in opto - mechanical experiments .
in particular we derived , starting from a non - local klein - gordon equation of the type @xmath21 with @xmath22 an analytic function , the corresponding non - local schrdinger equation .
we then solved this equation perturbatively around a coherent state solution of the local schrdinger equation , the perturbation parameter being @xmath23 ( where @xmath24 is the mass of the system and @xmath25 its natural frequency ) .
what we found is that the first order correction to the coherent state evolution produce a _ spontaneous , time periodic squeezing _ of the state and , indeed , the state remain of minimum uncertainty throughout its evolution , see fig .
[ nls ] .
we then used current data coming from opto - mechanical experiments to forecast the expected bounds on the non - locality scale that could be achieved with ongoing experiments .
the forecast shows how very stringent bounds will possibly be achieved , of the order of @xmath26 m . even tough
six order of magnitude from the planck scale these kind of bounds can explore a very interesting and uncharted territory , far beyond current bounds coming from lhc data @xmath27 m @xcite .
quantum gravity phenomenology is attracting more and more interest in recent years . in this work
we have reviewed the phenomenological possibilities related to non - local dalembertians in causal set theory as well as a new proposal for testing non - local models with opto - mechanical experiments .
principle violations seems to have the potential for be used to constrain non - local models stemming from cs theory , work in this direction is in progress . on the other hand , tests of ( quantum ) gravity inspired effects with ``
macroscopic '' quantum objects are being studied in different contexts and promise to shed new light on our ides about space , time and gravity .
the author would like to thank dionigi benincasa for comments and suggestions on early drafts and stefano liberati for illuminating discussions and support .
we wish to acknowledge the john templeton foundation for the supporting grant # 51876 .
0 l. j. garay , int .
j. mod .
phys . a * 10 * ( 1995 ) 145 s. surya , arxiv:1103.6272 [ gr - qc ] .
r. d. sorkin , int .
j. theor .
phys . * 36 * ( 1997 ) 2759 r. d. sorkin , print-91 - 0287 ( chicago ) . f. dowker , j. henson and r. d. sorkin , mod . phys .
a * 19 * ( 2004 ) 1829 s. aslanbeigi , m. saravani and r. d. sorkin , jhep * 1406 * ( 2014 ) 024 l. bombelli , j. henson and r. d. sorkin , mod .
a * 24 * ( 2009 ) 2579 r. d. sorkin , in * oriti , d. ( ed . ) : approaches to quantum gravity * 26 - 43 [ gr - qc/0703099 [ gr - qc ] ] . f. dowker and l. glaser , class .
* 30 * ( 2013 ) 195016 a. belenchia , d. m. t. benincasa and s. liberati , jhep * 1503 * ( 2015 ) 036 m. saravani and s. aslanbeigi , phys .
d * 92 * ( 2015 ) 10 , 103504 s. p. johnston , arxiv:1010.5514 [ hep - th ] .
a. eichhorn and s. mizera , class .
* 31 * ( 2014 ) 125007 a. belenchia , d. m. t. benincasa , a. marciano and l. modesto , arxiv:1507.00330 [ gr - qc ] . s. carlip , class .
* 32 * ( 2015 ) 23 , 232001 a. belenchia , d. m. t. benincasa , s. liberati , f. marin , f. marino and a. ortolan , arxiv:1512.02083 [ gr - qc ] . m. bawaj _ et al .
_ , nature communications 6 , 7503 ( 2015 ) t. biswas and n. okada , nucl .
b * 898 * , 113 ( 2015 ) . | this proceeding is based on a talk prepared for the xiv marcel grossmann meeting .
we review some results on causal set inspired non - local theories as well as work in progress concerning their phenomenology . |
0.3 cm in a study of `` discovery potential '' and data analysis of present high energy experiments the problems of precise predictions including , simultaneously , signal signatures of the new ( or studied ) physics , backgrounds , as well as all detector related effects should be analysed .
it is generally believed that a monte carlo simulation of the full chain from the beam collision to detector response is the most convenient technique to address such question .
in general it is indispensable to divide monte carlo simulation into separate blocks : physics event generation and detector response .
later event generation can be divided further into parts , describing for example production and decay of the intermediate states . in the present paper
we will concentrate on the particular class of the processes involving polarised @xmath0 leptons .
the two main goals of the present paper are : @xmath12 presentation of the algorithm for matching @xmath0 lepton decay and its production , with some control over spin effects ; in particular in case of @xmath13 production mechanism , @xmath14 discussion of physical observables sensitive to the spin correlations in the @xmath0 pair production .
spin correlations in the decay of @xmath0 leptons not only can help to suppress irreducible background to the possible resonant @xmath0 pair production at lhc , such as the mssm higgs bosons decays , but also help to determine the spin nature of this resonanse .
in the papers @xcite tauola monte carlo package for simulation of @xmath0 lepton decay was described .
recently , in ref .
@xcite , technical details convenient for using the code in multi - purpose environment were collected , and universal interface for combining the simulation of @xmath0 lepton decay , with different packages for generation of physics event was proposed .
scheme of ref .
@xcite relies on the information stored in the hepevt common block @xcite only , and not on the details specific for the production generator , such as pythia @xcite ( used in our examples ) .
in fact , such an interface can be considered as a separate software project , to some degree independent both from the specific problem of @xmath0 production and its decay .
our paper is organized as follows : in the next section we will describe new algorithm for extracting elementary @xmath15 body reaction for @xmath16 , which is necessary for properly introducing spin correlations into generation chain . in sec .
3 we analyze spin content of such an elementary function . sec . 4
is dedicated to the discussion of their consequences for the distributions of physics interest . in sec .
5 we discuss few observables where spin effects can improve separation of the higgs boson signature , in case of the 14 tev @xmath17 collisions .
summary closes the paper . in appendix
, we explain the basic scheme of the spin treatment used in our code .
it completes the program manual given in ref . @xcite
the exact way of calculating spin state of any final state is with the help of the matrix element and the rigorous density matrix treatment .
this is however not always possible or necessary .
often , like in the case of the production and decay of particles in the ultra - relativistic limit a simplified approach can be sufficient .
such an approach was developed for koralz monte carlo program @xcite and its limitations were studied with the help of matrix element calculations of the order @xmath19 @xcite . in the following , we study the question whwter the approach can be generalised , and the approximate spin correlation calculated from the information stored in the hepevt common block filled by `` any '' @xmath0 production program .
the approximation consists of reconstructing information of the elementary @xmath20 body process @xmath21 , buried inside multi - body production process .
let us stress that such a procedure can never be fully controlled , as its functioning depends on the way the production program fills the hepevt common block .
it will be always responsibility of the user to check if in the particular case the implemented algorithm is applicable .
nonetheless our aim is _ not _ to replace the matrix element calculations , but rather to provide a method of calculating / estimating spin effects in cases when spin effects would not be taken care of , at all .
needless to say such an approach is limited ( for the spin treatment ) to the approximation not better than leading - log , and to the longitudinal spin degrees only .
the principle of calculating kinematic variables is simple .
the 4-momenta of the @xmath20 body process have to be found .
the 4-momenta of the outcoming @xmath0 s are used directly .
initial state momenta are constructed from the incoming and outcoming momenta of the particles ( or fields ) accompanying production of the @xmath22 state state does not need to be explicitly coded in the hepevt common block . note that if available , information from the history part of the event , where the 4-momenta of gluons quarks etc .
are stored , will be used . ] .
we group them accordingly to fermion number flow , and ambiguous additional particles are grouped ( summed ) into effective quarks to minimise their virtualities .
such an approach is internally consistent in the case of emission of photons or gluons within the leading log approximation .
longitudinal polarisation of @xmath0 leptons @xmath23 depends on the spin quantum number of the @xmath0 mother .
it is randomly generated as specified in table [ t : probability ] .
._probability for the configurations of the longitudinal polarisation of the pair of @xmath0 leptons from different origins .
[ t : probability ] _ [ cols="^,^,^,^",options="header " , ] in the first step of our discussion the @xmath23 is shown as a function of @xmath24 , for several centre of mass energies and initial state flavours .
the angle @xmath25 denotes @xmath26 scattering angle in the @xmath27 rest - frame .
it is calculated with respect to the @xmath28 , @xmath29 or @xmath30 effective beam .
the @xmath31 mass @xmath32 = 91.1882 gev , was taken from ref .
@xcite , as well as effective @xmath33 and @xmath34 = 2.49 gev . in fig .
[ rozk2 ] , the angular dependence of the @xmath0 polarisation for the @xmath35 process at the peak of @xmath31 resonanse and for the three different values of the centre - of - mass ( cms ) energies above it ( upper plot ) .
the @xmath0 s are strongly polarised in the forward direction , where polarisation approaches the value approximately twice as big as the average one ( the polarisation of z due to @xmath36 coupling sums with the one due to @xmath37 coupling ) .
the polarisation is very small in the backward regions .
for @xmath38 it is equal to zero , independently of the centre of mass energy .
the reason is the universality of both the @xmath31 and @xmath39 couplings to all leptons .
as can be observed , the @xmath0 polarisation changes significantly ( especially in the forward region ) with the centre of mass energy . at cms energies above the @xmath31-peak
, @xmath0 polarisation is smaller , due to significant contribution from the @xmath40-channel @xmath39-exchange production mechanism not contributing directly to the polarisation .
( 1400,1400 ) ( 800,1300)(0,0)[b ] @xmath41 ( 450 , 600)(0,0)[b ] @xmath42 ( 1200 , 600)(0,0)[b ] @xmath43 ( 330 , 480)(0,0)[lb ] ( -20 , -220)(0,0)[lb ] ( 700 , -220)(0,0)[lb ] [ rozk2 ] let us now turn to the production from quarks . in fig .
[ rozk2 ] we show the @xmath0 polarisation for the @xmath44 and @xmath45 elementary processes and the same cms energies as in the previous case . for the cms energy close to the @xmath31-peak
, the @xmath0 s produced in the forward as well as in the backward directions are strongly polarised .
the polarisation seem to be almost zero for @xmath46 , but in reality is equal ( nearly ) to the same value as in previous case of production from electrons at @xmath46 . also , the average @xmath0 polarisation is close to the case of production from electrons .
it is , as it should be , independent from the initial state flavour , the @xmath31 was produced from .
the initial state couplings of the @xmath31 , affect the angular dependence of the @xmath0 polarisation only , and give sizable angular asymmetry in polarisation .
obviously , the larger the initial state vector couplings to @xmath31 the larger angular dependence of the polarisation .
the above arguments hold , because the contribution from the @xmath39 exchange is small .
this is the case , in the region of the @xmath31 peak only .
quite different polarisation pattern can be observed for quarks and cms energies far above the @xmath31 peak ( see fig . [ rozk2 ] ) .
contribution from the @xmath39 exchange can not be neglected .
the @xmath47 interference complicates the pattern even more .
let us comment briefly on the numerical results . in the case of @xmath29 and @xmath30 initial state
, polarisation is negative and nearly constant over the forward hemisphere ; also the average polarisation is negative .
this is because of rather large and positive forward - backward asymmetry , and polarisation forward - backward asymmetry .
the average polarisation increases above the z - peak and approaches , respectively , -35% and -61% in the case of @xmath48 and @xmath49 annihilation pair production . ] .
let us point , that in case of the @xmath0 production from the @xmath30 quarks in the ultra high - energy limit , polarisation of the @xmath0 leptons in the backward directions approaches zero ( independently of the numerical value of @xmath50 ) .
once we have understood the pattern of the @xmath0 polarisation , let us turn to the question of its measurement .
we will follow the reasoning similar to the one of ref .
the @xmath0 polarisation @xmath23 can be measured from the energy distribution of its decay products . to simplify the discussion of spin effects ,
the decays @xmath51 were used only sensitivity is smaller and depends on energy of leptons ; in case of @xmath52 to exploit sensitivity in full , the reconstruction of @xmath53 is necessary . ] .
we will compare the cases of vector @xmath54 and scalar neutral higgs boson , produced in hadron collisions and decaying into pair of @xmath0 leptons .
the @xmath0 rest - frame can not be accessed experimentally or higgs , can be reconstructed from total bilans of the observed transverse energies for all tracks in the event . ] , the spin effects can be seen , however , through the effects on @xmath55 @xmath56 energy distributions and correlations , observed / defined for the laboratory frame .
( 1600,800 ) ( 375,750)(0,0)[b ] ( 1225,750)(0,0)[b ] ( -20 , -1)(0,0)[lb ] ( 700 , -1)(0,0)[lb ] let us start with the following example , where for the production of the @xmath0 lepton pairs monte carlo program pythia was used , and for the decay monte carlo program tauola , and our interface .
it was assured , that the invariant mass of the pair of two incoming quarks was @xmath57 .
for the time being we discuss energies defined in the @xmath58 pair rest - frame . with the help of variables
@xmath59 , the spin effects are visualized . in fig .
[ pion ] we observe the slope ( as expected ) of @xmath60 energy spectrum due to @xmath0 polarisation .
the slope of the distribution is simply proportional to the polarisation ( small dip at @xmath61 is due to kinematical effect of the @xmath60 mass ) .
@xmath62 in the case of the plot on the right - hand side of fig .
[ pion ] , _ i.e. _ decay of the @xmath31 , polarisation was about -14.7% . in the case of the plot on the left - hand side
the spectrum is flat , as would be in the case of scalar neutral higgs boson ( or pure @xmath39 ) where there is no polarisation .
if polarisation was @xmath63 = @xmath64 100% , then the distribution slope would be maximally negative and touch zero at @xmath65 . for @xmath63
= 100% , slope would be reversed and distribution would touch zero at @xmath66 .
these are the cases of charged higgs boson and charged w boson decays into @xmath67 .
respective distributions are shown in fig .
[ pion - charged ] .
( 1600,800 ) ( 375,750)(0,0)[b ] ( 1225,750)(0,0)[b ] ( -20 , -1)(0,0)[lb ] ( 700 , -1)(0,0)[lb ] let us now turn to the question of the spin correlations .
as we can see from table [ t : probability ] , the @xmath0 pairs are produced with the well defined spin configurations ( @xmath68 or @xmath69 for vector bosons ; @xmath70 or @xmath71 for neutral higgs boson ) , thus , the spin effects are to become visible on the two - dimensional distribution build on @xmath72 and @xmath73 variables . indeed , for the vector bosons we expect more events when both @xmath55 and @xmath56 are on the upper side ( case i ) ( @xmath74 , @xmath75 ) or lower side ( case ii ) ( @xmath76 , @xmath77 ) , with respect to the mixed ones : @xmath74 , @xmath77 ( case iii ) and @xmath76 , @xmath75 ( case iv ) .
the appropriate asymmetry is : @xmath78 this can be compared to the higgs boson case , where asymmetry @xmath79 . for the higgs boson case
the @xmath80 is insensitive on the choice of cms energy , as there is no interference effects at all .
the sign difference is due to the the vector boson nature of @xmath31 , as opposed to scalar nature of higgs boson . in order to better visualize the spin correlation effect we have introduced variable @xmath81 defined as signed part of the following phase space part : surface in @xmath73 , @xmath82 variables between lines @xmath83 and @xmath84 ( the sign of the @xmath85 should be taken ) . in fig .
[ tomek ] respective plots are given for the @xmath86 and @xmath31 decays .
the dashed lines ( which in both @xmath31 and @xmath86 cases are flat ) correspond to the case when spin correlations are switched off .
as we expect in the @xmath87 decays , due to spin effects a fast ( slow ) @xmath88 is most likely associated with a fast ( slow ) @xmath89 .
the solid line has maximum at @xmath90 and approaches zero for @xmath91 .
precisely the opposite is true in the case of @xmath92 .
the maximum is now at @xmath91 and minimum for @xmath90 .
( 1600,800 ) ( 375,750)(0,0)[b ] ( 1225,750)(0,0)[b ] ( -20 , -1)(0,0)[lb ] ( 700 , -1)(0,0)[lb ] ( 1600,800 ) ( 375,750)(0,0)[b ] ( 1225,750)(0,0)[b ] ( 400 , -1)(0,0)[lb ] ( 1600,800 ) ( 375,750)(0,0)[b ] ( 1225,750)(0,0)[b ] ( -20 , -1)(0,0)[lb ] ( 700 , -1)(0,0)[lb ] let us now turn our attention to the quantities which ( at least in principle ) can be measured experimentally .
[ przod - tyl ] shows the invariant mass distribution for the case of pure @xmath93 and @xmath94 configurations of the @xmath0 polarisation and the mass of the resonance equal to the @xmath31 mass .
the spin correlations enhance fraction of the fast - fast and slow - slow configurations which are localised mostly at the shoulders of the @xmath95 invariant mass distributions , the fast - slow configurations would be localised in the centre of the distributions . in fig .
[ invariantmass90 ] we show @xmath95 invariant mass distribution for the higgs and @xmath31 cases .
continuous line with spin effects included , dotted line with spin effects switched off .
left - hand side plot corresponds to the higgs boson case , right - hand side to the @xmath31 . in the case of higgs boson ,
the mass distribution is peaked centrally , whereas in the case of @xmath54 shoulders of the distributions are more profound , especially the one at lower invariant mass additionally enhanced by the polarisation . for the vector case
the spin correlations enhance fraction of the fast - fast and slow - slow configurations which are localised mostly at the shoulders of the @xmath95 invariant mass distributions .
the relative height of the shoulders is sensitive to the average polarisation . for the scalar case , the enhancement in the fraction of the slow - fast configurations effects in the enhancement of the events localised in the middle of the @xmath95 invariant mass distribution , leading to the slightly narrower shape . if all polarisation effects are switched off ( dashed lines ) the distributions in the two cases are identical . that observable , well defined distribution of invariant mass built from the visible decay products of the @xmath0 s , can be helpful in separating higgs boson signal from @xmath27 background .
( 1600,800 ) ( 375,750)(0,0)[b ] ( 1225,750)(0,0)[b ] ( -20 , -1)(0,0)[lb ] ( 700 , -1)(0,0)[lb ] the same distribution have been also studied for the off - peak production of @xmath54 , _
i.e. _ for the larger cms energies . in these cases
the average polarisation is large and negative , also distinct for the @xmath48 and @xmath49 annihilations . as illustrated in fig .
[ invariantmass300 ] , the effect on the @xmath96 invariant mass distribution is noticeable .
the shape of the distribution might give the insight to the structure functions of the colliding protons , once the invariant mass of the @xmath54 state can be reconstructed .
0.3 cm in the search for new phenomena in accelerator experiments , important parameter to estimate discovery chances is the signal sensitivity , equal to the ratio of number of expected events from the new physics divided by the square root of the expected background events .
any possible increase of such a ratio can improve chances of discovery ( or improve limit of the exclusion ) .
use of more sophisticated cuts can be of a great help in a case when it can be combined with the physical properties of signal and/or background distributions .
equally important is the possibility for the verification of the nature of observed new physics , _
e.g. _ the quantum numbers of the new resonances .
the @xmath0 leptons are considered as a very promising signature for the searches of the higgs bosons in the minimal supersymmetric standard model ( mssm ) at lhc collider @xcite .
below , we will briefly discuss possible applications of the discussed spin correlations in the cases of the neutral higgs bosons @xmath86 and @xmath97 decay into @xmath98 pair and the charged higgs boson @xmath99 decays into @xmath100 pair .
the neutral higgs bosons h and a decays into @xmath98 pair , are enhanced for the large values of @xmath101 ( @xmath101 denotes the ratio of the vacuum expectation values of the higgs doublets in the mssm model ) , with the branching ratio of about 10% for most of the range of the interesting higgs boson mass values ( 150 - 1000 gev ) .
accessibility of the hadronic decay mode of the @xmath98 pair has been studied recently by the cms collaboration @xcite .
the triggering on such events , signal extraction from the irreducible background @xmath3 and the reducible backgrounds qcd jets , @xmath102 and @xmath103 , and reconstruction of the resonance peak in the @xmath98 mass distributions seems feasible for the higgs boson masses roughly above 300 gev . in that study the @xmath0 identification
is based on the presence of a single hard isolated charged hadron in the jet using tracker information .
the two hard tracks from @xmath104 and @xmath105 in the signal events have an opposite sign while no strong charged correlation is expected for the qcd jets or @xmath103 events .
the sensitivity of 5@xmath106 can be reached for the large fraction of the mssm parameter space after three years of data collecting at low luminosity .
the expected signal - to - background ratio is very large , being of the order of one for 5@xmath106 sensitivity , with the background dominated by the continuum @xmath3 production .
the resolution for the reconstruction of the gaussian @xmath107 peak is @xmath108 10% of the mass of the higgs boson .
the above performance seems very promising but hopefully can be still improved by exploring the spin correlations and polarisations effects .
the possible improvements may come from the additional suppression of the background .
but it is also important that exploring effect of the spin correlation on the invariant mass of the hadronic decay products might allow to verify hypothesis of the scalar _ versus _ vector nature of the observed resonance peak in the reconstructed invariant mass of the @xmath2 pair . as we have seen in the previous sections , the main difference between production mechanisms due to higgs boson _ versus _
@xmath27 consists of the correlation in energies of the @xmath0 hadronic decay products .
( 1600,800 ) ( 375,750)(0,0)[b ] ( 1225,750)(0,0)[b ] ( -20 , -1)(0,0)[lb ] ( 700 , -1)(0,0)[lb ] ( 1600,800 ) ( 375,750)(0,0)[b ] ( 1225,750)(0,0)[b ] ( -20 , -1)(0,0)[lb ] ( 700 , -1)(0,0)[lb ] let us concentrate on the case of @xmath0 decays to @xmath109 most sensitive to the spin correlations . in fig .
[ cut1 ] the @xmath60 energy spectrum is shown in the laboratory frame for the @xmath86 ( left - hand side ) and @xmath54 ( right - hand side ) decays .
the selection , roughly consistent with what is foreseen in the experimental analysis @xcite , was applied .
the minimal transverse momenta of the @xmath110 were required to be above 15 gev and the pseudorapidity @xmath111 .
solid line shows results with included spin effects , dashed line with the effects switched off .
spin correlations lead to the softer spectrum of @xmath60 in @xmath54 decays , which may slightly suppress the identification efficiency of the @xmath0 with respect to those produced from the higgs boson decay .
the slope in the distribution for the @xmath54 ( sensitive to the large negative average polarisation ) , depends also on the relative fraction of the @xmath48 and @xmath49 production processes simulated in the proton - proton collision , hence the parametrisation of the structure functions .
events were generated with the monte carlo pythia 5.7 @xcite and cteq2l structure functions ; the higgs boson mass of the 300 gev and the width below 1 gev ( as for @xmath112 ) was assumed .
for the continuum @xmath54 , the cms energy of the produced @xmath2 pair was taken in the range 300 @xmath113 10 gev .
[ cut4 ] shows the effect of the @xmath114 energy - energy correlations in the cms frame as discussed in the previous section .
although the effect seems strong enough to discriminate between scalar and vector cases , exploring this effect would require good experimental reconstruction of the effective cms frame which might be very difficult . the invariant mass distribution of the @xmath114 system , see fig .
[ cut5 ] , shows the visible effect of including spin - correlations . as discussed in the previous section , in the vector case
these correlations lead to the more profound shoulders in the distribution , and the relative height of them is sensitive to the average polarisation of the produced resonance . in the case of the scalar higgs boson the correlations lead to the narrower distribution , enhancing fraction of events localised in its central part . with the expected signal - to - background ratio being of one or higher
, this effect seems promising for determining spin property of the studied resonance .
( 1600,800 ) ( 375,750)(0,0)[b ] ( 1225,750)(0,0)[b ] ( -20 , -1)(0,0)[lb ] ( 700 , -1)(0,0)[lb ] for completeness let us now turn to the case of the charged higgs boson .
the decay into @xmath100 pair is a dominant mode below the kinematical threshold of the @xmath115 channel , and accounts for nearly 100% of cases , almost independently of @xmath101 .
this branching ratio is decreasing rather rapidly above @xmath115 thereshold , but is still of the order of 10% for the higgs boson mass of 500 gev and large @xmath101 .
the lhc detectors will be thus sensitive to signal @xcite , with at least @xmath116 significance , for higgs boson mass up to 400 - 500 gev .
this sensitivity is almost independent on @xmath101 below @xmath115 threshold . for the higgs boson mass above top - quark mass
this sensitivity is expected only for large @xmath101 . in both cases of higgs boson masses below and above top - quark mass ,
the main background comes from @xmath117 decay .
harder pions are expected from the @xmath99 decay than from the @xmath118 decays .
the effect of the spin correlations has been already studied , theoretically in ref .
@xcite and experimentally in ref .
@xcite . for completeness
we show in fig .
[ cut6 ] the @xmath60 energy distribution in the laboratory frame for the decays of @xmath99 and @xmath119 spin effects switched on and off .
the higgs boson mass of 130 gev was taken .
( 1600,800 ) ( 375,750)(0,0)[b ] ( 1225,750)(0,0)[b ] ( -20 , -1)(0,0)[lb ] ( 700 , -1)(0,0)[lb ] 0.4 cm
0.3 cm we have discussed the spin effects in the @xmath0 pair production at lhc .
a few distributions presented here , sensitive to the @xmath58 spin correlations , can be possibly used for the mssm higgs boson searches scenarios at lhc to enhance sensitivity of the signal or to verify the hypothesis of the spin zero nature of the higgs boson .
we have extended the algorithm for the interfacing the @xmath0 lepton decay package tauola with `` any '' production generator to include effects due to spin in elementary @xmath120 process .
the interface is based exclusively on the information stored in the hepevt common block .
this code is publicly available _
e.g. _ from the url address @xcite .
let us recall first some details of the universal interface for tauola and `` any '' @xmath0 production generator as described in ref .
the interface uses as an input , the hepevt common block and operates on its content only . as a demonstration example
the interface is combined with the jetset generator , however it should work in the same manner with the pythia , herwig or isajet generators as well . *
the @xmath0 lepton should be forced to be stable in the package performing generation of the @xmath0 production . *
the content of the hepevt common block is searched for all @xmath0 leptons and @xmath0 neutrinos . *
it is checked if there are @xmath0 flavour pairs ( two @xmath0 leptons or @xmath0 lepton and @xmath0 neutrino ) originating from the same mother . *
the decays of the @xmath0 flavour pairs are performed with the subroutine tauola .
longitudinal spin correlations are generated in the case of the @xmath0 produced from decay of : @xmath121 , @xmath122 , the neutral higgs boson @xmath123 , and the charged higgs boson @xmath124 .
parallel or anti - parallel spin configurations are generated , before calling on the @xmath0 decay , and then the decays of 100 % polarised @xmath0 s are executed . * in the case of the higgs boson ( for the spin correlations to be generated ) the identifier of the @xmath0 mother must be that of the higgs boson . here , the particle code convention as that used by the pythia 5.7 monte carlo is adopted . * in case of the @xmath119 and @xmath22 it is not necessary .
if from the same mother as that of the @xmath0 , a @xmath125 is also produced , the @xmath119 is assumed as the mother of the @xmath0 . similarly , if from the same mother another @xmath0 with opposite charge is produced , the @xmath22 is assumed to be the mother of the @xmath0 pair . * photon radiation in the decay
is performed with photos package @xcite .
* let us note that the calculation of the @xmath0 polarisation created from the @xmath31 and/or virtual @xmath39 ( as function of the direction ) represents a rather non - trivial extension .
generally , the dedicated study of the production matrix elements of the host generator is necessary in every individual case . for some of the technical details on how to use the interface we address the reader to ref .
for general documentation of tauola to ref .
for the use of pythia reference to @xcite and references therein will be the best . for every event generated by the production generator , all @xmath0 leptons will be decayed with the single call tauola(0,keyspin ) , ( keyspin=1/0 denotes spin effects switched on / off ) .
then , all @xmath0 leptons , will be first localised , their positions stored in internal common block taupos , and the information necessary for calculation of @xmath0 spin state will be read in , from hepevt common block .
later spin state for the given @xmath0 ( or @xmath0 pair ) will be generated , and finally decay of polarised @xmath0 will be performed with the standard tauola action .
in particular , the decay products of @xmath0 will be boosted to the laboratory frame and added to the complete event configuration stored in hepevt common block .
first , we look in hepevt for the position of @xmath0 s mothers , and store them , in matrix imother(20 ) .
each mother giving @xmath0 lepton(s ) is stored only once , independently of the number of produced @xmath0 s .
later , for every imother(i ) we execute the following steps : 1 .
the daughters which are either @xmath0 leptons , or @xmath125 are searched for . 2 .
daughters are combined in pairs , case of more than 1 pair is not expected to be important and ad hoc pairing is then performed .
3 . the two main cases are thus ( @xmath126 @xmath126 ) or ( @xmath0 @xmath127 ) .
4 . the default choices are , respectively , @xmath22 or @xmath119 , unless the identifier of imother(i ) is explicitely that of neutral ( or charged ) higgs boson . 5 .
calculation of the spin parameters is kinematics independent and straightforward in all cases except @xmath22 ( see section 2 for details on physics ) .
6 . for @xmath22 ,
the @xmath128 is calculated with the help of the function plzapx(hope , im0,np1,np2 ) .
the hope is the logical parameter defined in subroutine tauola placed in file tauface- + jetset.f .
it tells whether spin effects can be calculated or not .
it is set to .false .
, if available information is incomplete , then , plzapx(hope , im0,np1,np2 ) returns @xmath129 .
the im0 denotes position of the @xmath0 mother in hepevt common block np1 position of @xmath105 and np2 of @xmath26 .
7 . to calculate reduced @xmath15 body kinematical variables @xmath40 and @xmath130 subroutine angulu(pd1,pd2,q1,q2,costhe ) is used .
4-momenta of the incoming effective beams and outgoing @xmath131 and @xmath26 are denoted by @xmath132 , respectively . after the series of events is generated the optional call tauola(1,keyspin )
can be executed .
the information on the whole sample , such as number of the generated @xmath0 decays , branching ratios calculated from matrix elements _ etc .
_ , will be printed .
our main program demo.f is stored in subdirectory demo - jetset .
it reads in the file init.dat which includes some input parameters for the particular run , such as number of events to be generated ( by jetset / pythia ) , or the type of the interaction it should use to produce @xmath0 s , _ etc_. we address the reader directly to the code for more details .
it is self - explanatory .
99 s. jadach , j.h .
khn , z. was , _ comput .
* 64 * ( 1990 ) 275 .
m. jeabek , z. was , s. jadach , j.h . khn _ comput
commun . _ * 70 * ( 1992 ) 69 .
r. decker , s. jadach , j.h .
khn , z. was , _ comput .
* 76 * ( 1993 ) 361 .
p. golonka , e. richter - was , z. was , hep - ph/0009302 .
particle date group , c. caso _ et al .
_ , _ eur .
phys . j. _ * c3 * ( 1998 ) 1 .
t. sjstrand , _ comput .
commun . _ * 82 * ( 1994 ) 74 .
s. jadach , b.f.l .
ward , z. was , _ comput .
commun . _ * 79 * ( 1994 ) 503 .
z. was , _ acta phys .
* b18 * ( 1987 ) 1099 .
et al . _ ,
proceedings of the workshop on @xmath31 physics at lep , + edited by g. altarelli , r. kleiss and v. verzegnassi , cern-89 - 08 v. 1 - 3 , + switzerland , geneva ( 1989 ) .
particle date group , d.e .
_ , _ eur .
c15 * ( 2000 ) 1 . s. jadach , z. was , _ acta phys . pol . _ * b15 * ( 1984 ) 1151 ; + erratum _ acta phys . pol . _ * b16 * ( 1985 ) 483 .
aleph collaboration , cern - lhcc/99 - 15 .
cleo collaboration , cern - lhcc/94 - 44 .
r. kinnunen , d. denegri , cms note 1999/037 .
t. sjstrand _
et al . _ , hep - ph/0010017 .
roy , _ phys .
* b459 * ( 1999 ) 607 . .
e. barberio , b. van eijk , z. was , _ comput .
* 66 * ( 1991 ) 115 .
e. barberio , z. was , _ comput .
* 79 * ( 1994 ) 291 . | the proper incorporation of spin effects in @xmath0 lepton decays is often of importance . in present work the case of the @xmath1 production mechanism is studied in detail . as an example , the effects due to the spin correlations on the potential for the minimal supersymmetric standard model ( mssm ) higgs boson(s ) searches in the @xmath2 decay channel at the large hadron collider ( lhc ) are discussed . for these processes ,
the standard model @xmath3-pair production is a dominant background .
the spin effects in high energy physics reactions , can be implemented up to certain approximation , independently of the algorithm and matrix elements used by the production program .
information stored on every generated event can be sufficient .
the algorithm based on such approximation is documented .
question of the theoretical uncertaintity is partly discussed . *
tpju 2/2001 * + * t. pierzchaa@xmath4 , e. richter - was@xmath5 , z.was@xmath6 and m. worek@xmath4 * @xmath7_institute of physics , university of silesia + uniwersytecka 4 , 40 - 007 katowice , poland
_ + @xmath8 _ institute of computer science , jagellonian university + nawojki 11 , 30 - 072 cracow , poland _ + @xmath9_cern , ppe , 1211 geneva 23 , switzerland
_ + @xmath10_institute of nuclear physics + kawiory 26a , 30 - 055 cracow , poland _
+ @xmath11_cern , theory division , 1211 geneva 23 , switzerland
_ _ acta physica polonica _ * b 32 * ( 2001 ) 1277 * january 2001 * |
we are investigating whether or not interacting but not yet merging galaxies have heightened star formation properties . in our spitzer spirals , bridges , and tails interacting galaxy study ( @xcite ) ,
we have compiled a sample of interacting galaxies selected from the arp atlas of peculiar galaxies @xcite .
we have previously presented a detailed study of one of these galaxies , arp 107 , in @xcite . in the current proceeding
we investigate a second system , the interacting pair arp 82 ( ngc 2535/6 ) @xcite .
we have obtained uv , visible , and ir images of arp 82 from galex , sara , and spitzer telescopes respectively .
figure 1 displays various images of arp 82 .
the top left is a galex far - uv image with the 26 clumps identified .
the northern galaxy is ngc 2535 and southern galaxy is ngc 2536 .
the top right image in figure 1 is arp 82 in the spitzer irac 3.6 @xmath0 m band with h@xmath1 contours from the sara telescope .
the bottom left image in figure 1 is arp 82 in the spitzer irac 8 @xmath0 m band with sara h@xmath1 contours .
the tail is more prominent in the uv than in the ir while the center is much less prominent .
note that there are fuv and 8 @xmath0 m clumps in the tail region that are not seen in h@xmath1 .
the star forming regions at 8 @xmath0 m and in the fuv are more prominent than at 3.6 @xmath0 m .
the bottom right image in figure 1 is a snapshot of a smooth particle hydrodynamics model of the gas in red and old stars in blue .
this image shows arp 82 about 1 gyr after the initial closest approach .
the dotted curve shows the companion s passage .
the orbit is nearly planar .
the long duration is needed to allow particles to propagate out to the large distances observed .
four individual plots are seen in figure 2 .
the top left plots the galex fuv / nuv distribution .
a starburst99 @xcite stellar population synthesis model reddened with e(b - v)=0.0(blue ) , 0.2(green ) , and 0.6(red ) mag according to the @xcite reddening law is shown at the top of the histogram .
selected ages are marked .
the top axis is in magnitudes .
most of the clumps have an e(b - v ) between 0.2 and 0.6 mag and ages @xmath2 myr while a few clumps may be @xmath3 myr .
we have determined the star formation rate ( sfr ) for the clumps using two independent methods .
first , we estimated them from the l(ir ) using the calibration in @xcite .
the clumps have a total sfr@xmath4 of @xmath5 m@xmath6 yr@xmath7 .
second , we estimated the sfr from the l@xmath8(fuv ) using the uv sfr calibration in @xcite .
the total sfr@xmath9 of the clumps is @xmath10 m@xmath6 yr@xmath7 , in good agreement with the sfr@xmath4 .
no reddening correction has been applied to the l@xmath8(fuv ) .
the top right plot in figure 2 shows the star formation rate ( sfr ) versus distance from ngc 2536 .
the red symbols are the sfr determined from l(ir ) and the blue symbols are the sfr determined from l@xmath8(fuv ) .
the open boxes represent clumps in ngc 2536 , x s represent clumps in the bridge region , stars represent clumps in the spiral ( ngc 2535 ) region , and filled boxes represent clumps in the tail region .
the top axis is in arcseconds . from this figure
it can be seen that the sfr is greatest in the spiral region of ngc 2535 and in ngc 2536 , with much less star formation in the bridge and tail regions .
it can also be seen that the sfr s of clumps in the bridge and tail regions have much better agreement than do the sfr s of clumps in ngc 2536 and the spiral regions . in most cases
the sfr@xmath9 is greater than the sfr@xmath4 .
if an extinction correction were applied to the l@xmath8(fuv ) the sfr@xmath9 would be greater and the agreement with sfr@xmath4 would be worse .
the clumps in the bridge and tail regions account for about 7% of the total clump sfr@xmath4 , while the 2 clumps in the small companion , ngc 2536 , and the 2 largest clumps in the spiral region of ngc 2535 ( # 13 and # 16 ) make up about 42% of the total clump sfr@xmath4 .
the sfr@xmath4 of the entire arp 82 system is 1.2 m@xmath6 yr@xmath7 , while the entire system sfr@xmath9 is 2.7 m@xmath6 yr@xmath7 .
the total clump sfr@xmath4 accounts for about 36% of the entire system sfr@xmath4 .
the bottom left graph in figure 2 plots the irac [ 4.5]@xmath11[5.8 ] vs [ 3.6]@xmath11[4.5 ] colors of the clumps .
the data symbols are the same as above .
also included in this figure are the predicted irac colors for interstellar dust @xcite , the sloan digitized sky survey quasars in the spitzer wide - area infrared extragalactic survey ( swire ) elais n1 field @xcite , and the colors of m iii stars from m. cohen ( 2005 , private communication ) and field stars from @xcite .
the quasars have redshifts between 0.5 and 3.65 ; since their spectral energy distributions are power laws , their infrared colors do not vary much with redshift . from these figures
, it can be seen that clumps # 23 and # 26 have colors consistent with those of quasars and field stars respectively and may not be part of arp 82 .
most of the clumps have [ 4.5]@xmath11[5.8 ] colors between those of the ism and stars , indicating contributions from both to this color .
clumps # 24 and # 25 , which are in the northern tail , have colors similar to those of ism ( but with large uncertainties ) .
thus these appear to be very young star formation regions with little underlying old stellar population .
the bottom right graph in figure 2 plots the irac [ 4.5]@xmath11[5.8 ] color vs distance from ngc 2536 . the data symbols and horizontal axis
are the same as above . from this plot
it can be seen that clumps in the bridge and tail regions seem to have different relative ages than those in the spiral region .
the [ 4.5]@xmath11[5.8 ] colors are generally very red ( i.e. , very ` starbursty ' ) , except for the two low s / n clumps in the tail , # 23 and # 26 .
this work is based in part on observations made with the spitzer space telescope , which is operated by the jet propulsion laboratory , california institute of technology under contract with nasa .
galex is a nasa small explorer mission , developed in cooperation with the centre national detudes spatiales of france and the korean ministry of science and technology .
this research was supported by nasa spitzer grant 1263924 , nsf grant ast 00 - 97616 , nasa ltsa grant nag5 - 13079 , and galex grant galexgi04 - 0000 - 0026 .
this work 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 nasa . | to help understand the effects of galaxy interactions on star formation , we analyze spitzer infrared and galex ultraviolet images of the interacting galaxy pair arp 82 ( ngc 2535/6 ) , and compare to a numerical simulation of the interaction .
we investigate the uv and ir properties of several star forming regions ( clumps ) . using the fuv / nuv colors of the clumps
we constrain the ages .
the 8 @xmath0 m and 24 @xmath0 m luminosities are used to estimate the far - infrared luminosities and the star formation rates of the clumps .
we investigate possible gradients in the uv and ir colors .
see smith et al .
( 2006a , b ) for global results on our entire interacting sample . |
the superfluid ( sf ) to mott - insulator ( mi ) transition of interacting bosonic particles on a lattice is a paradigmatic example of a quantum phase transition , introduced by fisher @xcite .
the realization in a seminal paper by jaksch @xcite that the bose - hubbard ( bh ) model can describe the dynamics of a system of ultra - cold atoms trapped in an optical lattice sparked interest in the experimental realization of this model which culminated in the successful break - through experiment by the munich group @xcite .
rapid experimental progress in the experimental atomic and molecular physics community is leading to the successful simulation of various interacting models of both boson and fermion many - body systems using trapped atoms and molecules @xcite . in step with these
, various extensions of standard bh model have been vigorously studied theoretically @xcite .
a particularly interesting class of such models take into account not only two- but also three - body local interactions .
the phase properties of these models have recently been widely discussed in the literature for one dimensional chains using exact diagonalization @xcite and density matrix renormalization group methods @xcite .
the so - called perturbative single site mean - field approach was used to obtain the qualitative phase boundaries in @xcite for various parameter regimes . on the experimental side , effects of higher than two - body interactions
have also been observed already @xcite .
below , we consider the effect of three - body interactions at a mean - field level on repulsively interacting bosons undergoing the mi - to - sf quantum phase transition .
the studied quantum many body system is described by the following hamiltonian in grand canonical ensemble form : @xmath0 [ ham0 ] where the sum @xmath1 is over nearest neighbours in the optical lattice .
the operator @xmath2 annihilates a boson at site @xmath3 and @xmath4 is the local density operator .
the chemical potential @xmath5 controles the average number of particles in the system .
the operator @xmath6 represents the on - site interaction between bosons .
in contrast to the standard bh model where two - body interactions ( @xmath7 ) are solely considered , the local interaction consists here also of three - body terms .
the latter are represented in the last term in and describe the energy cost of forming a local triple of bosons at a given lattice site given by the parameter @xmath8 . in this article
, we study the stability properties of the zero temperature insulating phases in the system described by ( 1 ) for repulsive two body interaction ( @xmath9 ) within mean - field approaches .
we recall here that the mean - field approach should be a good approximation for higher dimensional lattices , becoming exact in the limit of infinite lattice dimensionality . in this article
, we compare results obtained by two mean - field approaches the perturbative mean - field and the gutzwiller variational ansatz for the phase boundaries , and further use the latter approach to obtain the full phase diagram properties .
for the considered problem , the superfluid state spontaneously breaking the @xmath10 symmetry of the model ( 1 ) can be described by the local order parameter @xmath11 . in the mean - field approximation ,
one considers the annihilation operator to be described by its average value plus the operator describing fluctuations about this average @xmath12 and neglecting second order fluctuations @xmath13 . in accordance with translational symmetry of the model ( 1 ) , the order parameter is assumed uniform throughout the system and the hopping part of the hamiltonian can be decoupled as @xmath14 . in the studied case , without loss of generality , the order parameter can be set to be real .
the resultant mean - field hamiltonian is a sum of local terms @xmath15 - \mu\hat{n}_i , \label{hmf}\ ] ] where @xmath16 is coordination number of the lattice . due to the fact that in this description all sites are completely independent
, one just considers a chosen site omitting the site index @xmath3 .
the order parameter should in principle be determined self - consistently from the condition @xmath17 in the ground state of the single - site mean field hamiltonian . in particular ,
the boundary between the sf and mi insulator phases is determined by the vanishing of the sf order parameter ( assuming the phase transition is continuous ) .
the boundary can however be determined analytically by the following perturbative argument .
the mott phase ( in mean - field description ) corresponds to a fock state corresponding to some integer filling . moving across the quantum phase transition , the sf order parameter attains a small non - zero value and it contributes a perturbative term to the hamiltonian of the system . treating this perturbation to the lowest non - trivial order , one can ask when this term is energetically favorable , i.e. the sf ground state energy becomes lower than that of the mott state , to obtain the phase boundary . for any @xmath5 ,
the mott state with integer filling @xmath18 has ground - state energy @xmath19 .
to second order in @xmath20 , the perturbed ground - state energy is then given by @xmath21,\end{aligned}\ ] ] where @xmath22 is the off - diagonal part of the on - site perturbation . to determine the phase boundary , it is convenient to rewrite the energy as @xmath23,\ ] ] where @xmath24
now it is clear that whenever @xmath25 the energy is minimized when the order parameter is zero and the system remains in the insulating phase . in contrast , for @xmath26 the energy is minimized for non vanishing order parameter @xmath20 and therefore the sf phase prevails .
the phase boundary is thus obtained from the limiting condition @xmath27 where @xmath28 is the critical value of tunneling .
for the first insulating lobe , i.e. when @xmath29 , the critical tunnelling expressed as a function of chemical potential @xmath5 is given by @xmath30 and therefore , in particular , the boundary is completely insensitive to three - body interactions .
this observation is in contrast to results obtained with more accurate methods .
for example , it was shown that , the tip of the mott lobe shifts under strong three - body interactions in @xmath31 dimensions @xcite .
the second insulating lobe ( @xmath32 ) is constrained by @xmath33 in the limit of vanishing tunnelling @xmath34 the second insulating lobe is bounded by two values of chemical potential @xmath35 and @xmath36 . moreover , its area defined as @xmath37 scales as @xmath38 .
similar analysis can be extended to higher fillings in straightforward manner .
the width of the insulating lobe for a given @xmath18 for zero tunnelling is equal to @xmath39 and its area scales as @xmath40 ^ 2 $ ] .
thus the three - body interactions have an increasingly important role for higher fillings in substantially stabilizing the mott phase in comparison with the standard bh model .
in this section , the background necessary for obtaining the properties of the studied system using the gutzwiller variational ansatz is laid out @xcite . in order to capture both the mott phase with fixed number of bosons per site , and the ( mean field ) sf phase which involves local boson number fluctuations ,
a product state ansatz over individual sites is assumed as an approximation to the true ground state : @xmath41 .
the local states @xmath42 can be expanded in the fock basis , which for numerical purposes is truncated to some sufficiently large maximal number of particles @xmath43 .
thus the variational ansatz is of the form @xmath44 where @xmath45 is the probability amplitude of finding @xmath46 bosons on site @xmath3 .
note that on assuming translational invariance , @xmath47 , one is lead to the simple mean - field theory described in section 2 .
the more general form of the product ansatz corresponds to the so - called unrestricted mean - field approximation .
the basic advantage of the gutzwiller ansatz , however , lies in the possibility of explicitly and directly obtaining the ground state instead of just chosen ground state expectation values .
it is obvious that a product state can not posses any information about non - local correlations .
nevertheless , the sf long - range coherence is captured at this level by non - trivial superpositions of local fock states , i.e. the local order parameter @xmath48 is non - zero only when the local quantum state is a superposition of states with different occupations .
the optimal state describing the state at a given parameter point of the studied hamiltonian is determined variationally through energy minimization . in practice
the minimization procedure is implemented by using standard methods like the conjugate gradient method or imaginary - time evolution .
the latter is based on the assumption that the dynamics of the system is determined by a mean - field hamiltonian where the order parameter is not necessarily a priori uniform over all sites .
the evolution in imaginary time in the infinite time limit then , in principle , projects an arbitrary initial state of the form onto the ground state .
the set of mean - field coupled equations of motion in imaginary time @xmath49 to be solved are @xmath50\varphi_i\nonumber \\ & + \left[\frac{u}{2}n(n-1 ) + \frac{w}{6 } n(n-1)(n-2 ) -\mu n \right ] \alpha_i(n),\end{aligned}\ ] ] where @xmath51 and the sum of the order parameters is over the sites coupled to site @xmath3 in the lattice .
the @xmath52 have to be determined self - consistently .
mean - field results are obtained for the studied system using the gutzwiller method based on the imaginary time evolution technique and compared with the perturbative mean - field approach .
the numerical calculations are performed on an @xmath53 square lattice , with @xmath54 under periodic boundary conditions . the cut - off @xmath55 .
we note here that the gutzwiller method does not assume translational invariance over the system and is in this sense unbiased .
1 shows the phase diagrams obtained by the gutzwiller method for three values of three - body interaction strengths @xmath8 .
the ground state parametrization obtained using this method is translationally invariant as assumed in the perturbative mean field approach .
the phase boundary is clearly identified by inspecting where the sf order parameter @xmath20 becomes negligibly small which is also accompanied by commensurate filling of lattice sites . the phase boundary obtained by the perturbative mean - field
is also plotted for comparison .
the gutzwiller results are generally in good agreement with the latter , although the mott area around the lobe tips seem to be slightly underestimated using the number of simulation steps and cut - off chosen for the considered system size .
as expected from perturbative mean field results , the gutzwiller approach yields a diagram where the first lobe is unaffected by three - body interactions .
indeed , the particle superfluid formed above the upper half of the mott lobe for unit filling has negligible contributions from states where two or more bosons occupy sites .
the non - trivial effect of the three - body interaction is thus on the mott lobes corresponding to higher lobes where three or more particles on site contribute to the ground state . on increasing ( the repulsive )
@xmath8 , the primary effect with respect to the standard bose - hubbard model is the increase of mott lobe widths and increased stability with respect to the hopping strength .
furthermore , the stability of the mott lobe increases with increasing of the filling of the lattice .
this is also in accordance with recent exact results obtained for the one - dimensional chain in the limit of dominant three - body interactions ( @xmath56 ) @xcite where exactly the same effect was observed .
finally note that at the mean - field level , for general @xmath57 , there is no change in character of mi - sf quantum phase transition which is still of second order .
again , this is in accordance with more exact results , e.g. in a one - dimensional studied using ed and dmrg showing no change of universality class , which there is the kosterlitz - thouless type @xcite .
while the above discussion focussed on the repulsive case @xmath58 , a similar simple analysis can be carried out for attractive interactions . there
the lobes shrink in size with increasing filling number and for every value of @xmath8 there is a critical number of bosons above which translational invariance is lost and the bosons collapse on to one site resulting in a trivial phase separation on the lattice .
this research was supported by the ( polish ) national science center grants no .
dec-2011/01/d / st2/02019 ( t.s . ) and dec-2011/03/b / st2/01903 ( r.w.c . ) .
r.w.c . acknowledges a mobility plus fellowship from the polish ministry of science and higher education .
99 fisher m p a 1989 _ phys . rev .
b _ * 40 * 546 jaksch d 1998 _ phys . rev . lett . _ * 81 * 3108 m. greiner 2002 _ nature ( london ) _ * 415 * 39 bloch i 2008 _ rev . mod .
phys . _ * 80 * 885 lewenstein m 2007 _ adv .
phys . _ * 56 * 243 lahaye t 2009 _ rep .
phys . _ * 72 * 126401 mering a and fleischhauer m 2011 _ phys .
rev . a _ * 83 * 063630 sowiski 115301 lewenstein m 2012 _ ultracold atoms in optical lattices : simulating quantum many - body systems _
( oxford university press , oxford ) sowiski t 2012 _ phys . rev . a _ * 85 * 065601 silva - valencia j and souza a 2011 _ phys .
a _ * 84 * 065601 singh m 2012 _ phys .
a _ * 85 * 051604(r ) silva - valencia j and souza a 2012 _ eur .
phys . j. b _ * 85 * 161 sowiski t 2013 _ preprint _
arxiv:1307.6852 chen b l. 2008 _ phys .
rev . a _ * 78 * 043603 zhou k 2010 _ phys .
rev . a _ * 82 * 013634 will s 2010 _ nature ( london ) _ * 465 * 197 rokhsar d and kivelson b 1991 , _ phys .
b _ * 44 * 10328 | the zero temperature properties of the generalized bose - hubbard model including three - body interactions are studied on a mean - field level .
we obtain analytical results using the so - called perturbative mean - field method and more detailed numerical results using the gutzwiller product state variational ansatz .
these two approaches yield equivalent results which compare well on a qualitative level with recent exact results obtained in the literature . |
the equivalence principle ( ep ) , as one of the fundamental hypotheses of einstein s general relativity , has been tested by many experiments , including torsion balance scheme @xcite and free - fall one r5,r6,r7 .
lunar laser ranging from the earth to the moon has provided up to now the most accurate test of the ep to 5@xmath210@xmath3r8 .
recently , some different tests of ep for gravitational self - energy and spin - polarized macroscopic objects have been reported @xcite . however , in all of the experiments including the satellite test of the equivalence principle ( step ) and the galileo galilei ( gg ) space projects as well as the microscope space mission @xcite , the test masses are all non - rotating .
it is well known that spin - interactions of elementary particles , spin - orbit coupling and spin - spin coupling , have been studied in both theory and experiment .
furthermore , the spin - gravitational couplings , i.e. the spin - gravitoelectric coupling and the spin - gravitomagnetic coupling , and the spin - rotation coupling between intrinsic spins have been also investigated for a long time @xcite . over the last few years
there has been a growing interest in effects of gravitational fields on gyroscopes @xcite . from the experimental point of view
, the nasa / stanford relativity mission gravity probe b ( gp - b ) experiment will provide two extremely precise tests of general relativity based on observations of four identical superconducting gyroscopes in a satellite in a 400 miles polar orbit around the earth r25 .
these gyroscopes are carefully isolated from nearly all sources of newtonian torques , and their residual drift is presented less than 0.020 marc - second / year for a gyroscope in a fully inertial orbit @xcite .
general relativity predicts that , though isolated from external torques , the spin axes of these gyroscopes will precess with respect to a distant inertial reference frame at a rate of 6.6 arc - second / year for the geodetic effect , and 0.042 arc - second / year due to frame dragging .
cerdonio et al .
have proposed a novel detector ( gyromagnetic electron gyroscope ) to locally detect the frame dragging due to the terrestrial rotation @xcite .
recently , zhang et al . also developed a phenomenological model for the spin - spin interaction between rotating extended bodies , which predicts the effect of the spin - spin coupling on the orbital acceleration of the gyroscope free - falling in gravitational field rather than the spin procession of the gyroscope@xcite . in the mode ,
a dimensionless parameter representing the strength of violation of ep can be defined as follows @xmath4where _ g _ is the newtonian gravitational constant , _
m_@xmath5 , _ m_@xmath6 and _ m_@xmath7 are the masses of the two gyroscopes and the earth , respectively , and @xmath8 , @xmath9 , and @xmath10 are their spin angular momentums , _
r_@xmath11 and _ r_@xmath6 are the distances between the centers of the two gyroscopes and the earth , respectively , and the parameter @xmath12 represents the universal coupling factor for the spin - spin interaction for rotating extended bodies . as pointed out in refs .
28 and 29 , the phenomenological model developed by zhang et al . is to investigate the effect of the spin - spin coupling on the orbital acceleration of the rotating gyroscope free falling in gravitational field , which is different from the aim of the gp - b .
a preliminary double free - fall ( dff ) experiment to test the ep for rotating extended bodies , in which two gyroscopes with differing rotating senses drop freely , has been performed , and the results show that the ep is still valid for rotating extended bodies at the level of 2@xmath210@xmath13 r29 .
a main limit of preliminary experimental precision has been proved to come from the pump outgassing effect due to the asymmetrical outgassing for the two tubes . in the initial experimental setup ,
the vacuum pump system is set in the top part of the tube . in this case , when the test masses fall through the tee part of the tube ( @xmath14 0.3 s free - fall ) , a force due to the pump outgassing will deflect the test masses or lead them to a more complex motion @xcite .
an abrupt acceleration change of about 20 mgal is observed at this height . to avoid it ,
the pump system is moved down to the bottom part of the tube , and meantime the vacuum level is improved from initial 50 mpa to 2 mpa . at the same time , the vibration excited by the operating pump is effectively isolated by a rubber - gas - steel isolator , and the isolation ratio is measured about -30 db to -60 db in the range of above 3 hz . and
then , the effective falling height is prolonged from initial 20 cm ( 0.2 s ) to 9 m ( 1.0 s ) . in this article , error sources of our dff experiment will be carefully discussed and a new upper limit of the ep for rotating extended bodies will be presented .
a michelson - type interferometer including a frequency - stabilized he - ne laser beam with a relative length standard of 1.3@xmath1510@xmath16 is used to monitor the differential vertical displacement between two gyroscopes , in which one is rotating and another is non - rotating , and then the interference fringes are sampled by means of a 10 mhz 12-bit ad card combined with an external rubidium atomic clock , and then stored in a computer .
the diameter of the laser beam is collimated about 3.0 mm so that the beam wavefront effect can be neglected .
an aligned verticality is kept within 50@xmath17 for each laser beam , and the maximum uncertain differential acceleration due to the aligned verticality is below 20 @xmath18gal .
each of the two test masses consists of a steel gyroscope with a mass of [email protected] g , a diameter of about 55 mm and a height of about 32 mm , which together with a corner - cube - retroreflector ( ccr ) of [email protected] g is sealed in an aluminum frame of [email protected] g. tinned copper wires with a diameter of 0.25 mm are used to suspend the test masses , and the initial suspending differential height between them is kept within 1 mm , which implies the vertical gravity gradient correction is about 0.3 @xmath18gal .
the test mass with a non - rotation rotor is released about 3 ms before the rotating one , which sets a systematic error of about 0.3 @xmath20gal due to the finite speed of light .
the rotating gyroscope is driven by a dc motor and its rotating speed is kept at ( 17000@xmath19200 ) rpm .
an uncertain acceleration due to the residual gas drag effect is less than 0.2 @xmath18gal at @xmath21 = 2 mpa and @xmath22 = 300 k @xcite .
in addition , the outgassing effect on the dropped objects should be carefully considered because of the continuous operation of a turbo - molecular pump with a full rated pumping speed @xmath23 of 1500 l / s . the acceleration contribution for a single dropped object
can be estimated as follows r31 @xmath24where @xmath25 represents the ratio of the surface of the dropped object and the inner surface of the vacuum tube , and is about 0.04 here , @xmath26 and @xmath27 are the residual gas density and mean gas particle speed , respectively .
then , the acceleration for a single dropped object is about 100 @xmath20gal at @xmath21 = 2 mpa .
fortunately , the dff scheme can reduce the common mode effect , and a differential acceleration for the two test masses due to the outgassing effect depends on both the outgassing difference and the gas density difference between the two tubes .
it is very difficult to calculate the real difference due to the complex flux motion exactly .
however , it can be roughly estimated based on the pressure difference between the two tubes .
when the turbo - molecular pump runs normally , the pressure of the bottom part of the tube ( close to the pump ) is measured about 1.7 mpa , and at the same time , the top parts of the two tubes are measured about 8.7 mpa and 5.6 mpa , respectively .
this means that the outgassing speeds in the two tubes are about 0.61 @xmath23 and 0.39 @xmath23 if the pressure distribution in both tubes is the same .
so in this assumption , the acceleration difference for the two test masses due to the outgassing effect is estimated less than 22 @xmath20gal .
a possible lifting force for a rotating rotor due to the residual gas flow s circulation can be calculated based on the zhukovskii s theorem , and this effect can be neglected here @xcite .
a possible horizontal velocity difference @xmath28 is estimated smaller than 4.2 mm / s according to the change of the interference pattern intensity during free - fall of test masses , and an acceleration difference due to the coriolis s effect is less than 54 @xmath20gal @xcite .
it means that the horizontal velocity difference would have to be monitored in the further experiment with a higher precision .
the silent amplitude spectrum of the seismic noise in our laboratory contributes an uncertainty of about 1 @xmath18gal to the final experiment result @xcite .
but the mechanical vibration modes of the optical measurement system are dominant due to excitation of the vacuum pumps .
figure 1 is a typical residual differential displacement of the dff experiment between both non - rotating test masses , in which the linear term has been subtracted .
three main modes have been observed about 16.8 hz , 36.6 hz , and 96.1 hz , and their amplitudes comes to about 0.05 @xmath20 m , which contribute an uncertain acceleration of 8 @xmath20gal based on the following equation @xcite@xmath29where @xmath30 , @xmath31 , and @xmath32 represent the amplitude , the angular frequency , and the phase of the high - frequency vibration , respectively , and @xmath22 is the effective time length .
parabolic curve fitting result shows that the differential acceleration @xmath33 is -58 @xmath20gal , which is consistent with the total systematic uncertainty , as listed in table 1 , of 64 @xmath20gal .
figure 2(a ) and 2(b ) are , respectively , typical residual differential displacements of the dff experiment with non- and left - rotating ( spin vector pointing upward ) gyroscopes , and non- and right - rotating ( spin vector pointing downward ) ones . from both figures , it is very clear that there are a dominant slow frequency motion but not a parabolic term .
this slow frequency ( [email protected] hz ) motion is confirmed to be resulted from the friction coupling between the rotating rotor and the aluminum frame by observing the motion of the reflected beam .
since effective free - fall duration in our dff experiment is about 1 s and the correlative coefficient between a parabolic term ( @xmath34 @xmath14 100 @xmath20gal ) and a harmonic term of 1.6 hz ( @xmath30 @xmath14 0.06 @xmath20 m ) comes to 0.3 @xmath14 0.4 , the slow frequency fluctuation is very difficult to be subtracted by fitting .
nevertheless , a maximum uncertain acceleration due to the slow motion can be estimated based on eq .
( 3 ) , and its effect comes to about 130 @xmath20gal ( @xmath35 0.06 @xmath20 m ) for the non- and left - rotating gyroscopes and 150 @xmath20gal ( @xmath35 0.07 @xmath36 m ) for the non- and right - rotating gyroscopes , respectively .
the fitting results of the both residual curves show that the differential acceleration between the non- and left - rotating gyroscopes @xmath37 is 0.48 @xmath20gal , and that between the non- and right - rotating gyroscopes @xmath38 is -110 @xmath20gal , which are consisted with the systematic uncertainty of 160 @xmath20gal as listed in table 1 .
in addition , a high - frequency mechanical vibration of the ccr at the frequency of the rotating rotor , caused by the friction coupling , has also been observed and its modulation amplitude is in the order of 0.1 @xmath20 m , which contributes an uncertainty of about 5 @xmath20gal .
fortunately , the limit can be suppressed by a factor of sin(@xmath39)/(@xmath40 ) by means of a time - domain data - smoothing - process , here @xmath41 is the time length of the smoothing - process .
based on above statement , the ep is still valid at the level of 1.6@xmath210@xmath1 for rotating extended bodies , which is improved by over one order for preliminary experiment result @xcite , and the spin - spin interaction between the rotating extended bodies has not been observed at this level . according to the eq .
( 1 ) and the approximately uniform sphere mode of the earth , it can be concluded that the coupling factor @xmath421.6@xmath210@xmath43kg@xmath44 , which sets a new upper limit for the spin - spin interaction between a rotating extended body and the earth .
our experiment precision is mainly limited by the mechanical friction of the gyroscope , which could be improved by choosing better gyroscope or extending the free - fall duration . as pointed out in the introduction section ,
gp - b mission measures the spin axes precession of spinning fused - quartz gyroscopes with respect to an inertial reference frame , partly due to the geodetic effect , and partly due to frame dragging .
in general relativity the both effects have small , non - zero values even without violation of the ep .
however , based on the model developed by zhang et al . , a perigean precession of the gyroscope in the gp - b experiment is not greater than 100 arc - second / year based on current experimental result @xmath451.6@xmath010@xmath43kg@xmath44 .
nevertheless , the gp - b experiment measures the spin axis precession rather than the orbital motion of the gyroscope .
as proposed in ref .
28 , a possible scheme is to put a non - spinning shell surrounding a spinning gyroscope in a satellite , and the motion of the spinning gyroscope with respect to the non - spinning reference frame could be monitored using a squid or an inertial sensor @xcite .
if the gap between the gyroscope and the reference shell can be measured in the level of 1 nm / year , and the coupling factor @xmath12 between the spin - spin interaction between the rotating extended bodies could be tested in the level of 10@xmath46 kg@xmath44 , which is improved by about 12 orders . *
acknowledgments * we are grateful to prof .
w. r. hu and a. ruediger for their discussion and useful suggestion .
this work was supported by the ministry of science and technology of china under grant no : 95-yu-34 and 19990754 and by the national natural science foundation of china under grant no : 19835040 , 10175070 , and 10047004 . | improved terrestrial experiment to test the equivalence principle for rotating extended bodies is presented , and a new upper limit for the violation of the equivalence principle is obtained at the level of 1.6@xmath010@xmath1 , which is limited by the friction of the rotating gyroscope .
it means the spin - gravity interaction between the extended bodies has not been observed at this level . |
the study of bianchi type v cosmological models create more interest as these models contain isotropic special cases and permit arbitrary small anisotropy levels at some instant of cosmic time .
this property makes them suitable as model of our universe . the homogeneous and isotropic friedman - robertson - walker ( frw ) cosmological models , which are used to describe standard cosmological models , are particular case of bianchi type i , v and ix universes , according to whether the constant curvature of the physical three - space , @xmath1 = constant , is zero , negative or positive .
these models will be interesting to construct cosmological models of the types which are of class one .
present cosmology is based on the frw model which is completely homogeneous and isotropic .
this is in agreement with observational data about the large scale structure of the universe .
however , although homogeneous but anisotropic models are more restricted than the inhomogeneous models , they explain a number of observed phenomena quite satisfactorily .
this stimulates the research for obtaining exact anisotropic solution for einstein s field equations ( efes ) as a cosmologically accepted physical models for the universe ( at least in the early stages ) .
roy and prasad [ 1 ] have investigated bianchi type v universes which are locally rotationally symmetric and are of embedding class one filled with perfect fluid with heat conduction and radiation .
bianchi type v cosmological models have been studied by other researchers ( farnsworth [ 2 ] , maartens and nel [ 3 ] , wainwright _
[ 4 ] , collins [ 5 ] , meena and bali [ 6 ] , pradhan _ et al . _ [ 7 , 8 ] ) in different context .
models with a relic cosmological constant @xmath0 have received considerable attention recently among researchers for various reasons ( see refs . [
9 - 13 ] and references therein ) .
some of the recent discussions on the cosmological constant `` problem '' and on cosmology with a time - varying cosmological constant by ratra and peebles [ 14 ] , dolgov [ 15 - 17 ] and sahni and starobinsky [ 18 ] point out that in the absence of any interaction with matter or radiation , the cosmological constant remains a `` constant '' , however , in the presence of interactions with matter or radiation , a solution of einstein equations and the assumed equation of covariant conservation of stress - energy with a time - varying @xmath0 can be found .
for these solutions , conservation of energy requires decrease in the energy density of the vacuum component to be compensated by a corresponding increase in the energy density of matter or radiation .
earlier researchers on this topic , are contained in zeldovich [ 19 ] , weinberg [ 10 ] and carroll , press and turner [ 20 ] .
recent cosmological observations by high - z supernova team and supernova cosmological project ( garnavich _ et al . _
[ 21 ] , perlmutter _
[ 22 ] , riess _
[ 23 ] , schmidt _
[ 24 ] ) strongly favour a significant and positive @xmath0 with the magnitude @xmath2 .
their finding arise from the study of more than @xmath3 type ia supernovae with redshifts in the range @xmath4 and suggest friedman models with negative pressure matter such as a cosmological constant , domain walls or cosmic strings ( vilenkin [ 25 ] , garnavich _ et al . _
the main conclusion of these observations on magnitude and red - shift of type ia supernova suggest that the expansion of the universe may be an accelerating one with a large function of cosmological density in the form of the cosmological @xmath0-term .
several ans@xmath5tz have been proposed in which the @xmath0 term decays with time ( see refs .
gasperini [ 26 ] , freese _ et al . _
[ 27 ] , @xmath6zer and taha [ 13 ] , peebles and ratra [ 28 ] , chen and hu [ 29 ] , abdussattar and viswakarma [ 30 ] , gariel and le denmat [ 31 ] , pradhan _ et al . _
of the special interest is the ansatz @xmath7 ( where @xmath8 is the scale factor of the robertson - walker metric ) by chen and wu [ 29 ] , which has been considered / modified by several authors ( abdel - rahaman [ 33 ] , carvalho _ et al . _
[ 13 ] , waga [ 34 ] , silveira and waga [ 35 ] , vishwakarma [ 36 ] ) . in recent years
, several authors ( hajj - boutros [ 37 ] , hajj - boutros and sfeila [ 38 ] , ram [ 39 ] , mazumder [ 40 ] and pradhan and kumar [ 41 ] ) have investigated the solutions of efes for homogeneous but anisotropic models by using some different generation techniques .
bianchi spaces @xmath9 are useful tools in constructing models of spatially homogeneous cosmologies ( ellis and maccallum [ 42 ] , ryan and shepley [ 43 ] ) . from these models ,
homogeneous bianchi type v universes are the natural generalization of the open frw model which eventually isotropize .
recently camci _ et al . _
[ 44 ] derived a new technique for generating exact solutions of efes with perfect fluid for bianchi type v spacetime . in this paper , in what follows , we will discuss bianchi type v cosmological models obtained by augmenting the energy - momentum tensor of a perfect fluid by a term that represents a variable cosmological constant varying with time , and later generalize the solutions of refs .
[ 39 , 44 ] .
this paper is organized as follows : the field equations and the generation technique are presented in section @xmath10 .
we relate three of the metric variables by solving the off - diagonal component of efes , and find a second integral which is used to relate the remaining two metric variables . in section 3 , for the particular form of each metric variables , some solutions are presented separately and solutions of camci _ et al .
_ [ 44 ] and ram [ 39 ] are shown to be particular cases of our solutions .
kinematical and dynamical properties of all solutions are also studied in this section . in section @xmath11 ,
we give the concluding remarks .
in this section , we review the solutions obtained by camci _ et al . _ [ 44 ] .
the usual energy - momentum tensor is modified by addition of a term @xmath12 where @xmath13 is the cosmological term and @xmath14 is the metric tensor .
thus the new energy - momentum tensor is @xmath15 where @xmath16 and @xmath17 , respectively , the energy and pressure of cosmic fluid , and @xmath18 is the fluid four - velocity such that @xmath19 .
+ we consider the spacetime metric of the spatially homogeneous bianchi type v of the form @xmath20,\ ] ] where @xmath21 is a constant . for the energy momentum tensor ( [ eq2 ] ) and bianchi type v spacetime ( [ eq3 ] ) , einstein s field equations @xmath22 yield the following five independent equations @xmath23
@xmath24 @xmath25 @xmath26 @xmath27 here and in what follows the suffix @xmath11 by the symbols @xmath28 , @xmath29 , @xmath30 and @xmath16 denote differentiation with respect to @xmath1 .
the bianchi identity ( @xmath31 ) takes the form @xmath32 for complete determinacy of the system , we consider a perfect - gas equation of state @xmath33 it is worth noting here that our approach suffers from a lack of lagrangian approach .
there is no known way to present a consistent lagrangian model satisfying the necessary conditions discussed in this paper .
+ the physical quantities expansion scalar @xmath34 and shear scalar @xmath35 have the following expressions : @xmath36 @xmath37.\ ] ] integrating eq .
( [ eq9 ] ) and absorbing the integration constant into @xmath29 or @xmath30 , we obtain @xmath38 without any loss of generality .
thus , elimination of @xmath17 from eqs .
( [ eq5 ] ) - ( [ eq7 ] ) gives the condition of isotropy of pressures @xmath39 which on integration yields @xmath40 where @xmath41 is a constant of integration .
hence for the metric function @xmath29 or @xmath30 from the above first order differential eq .
( [ eq16 ] ) , some scale transformations permit us to obtain new metric function @xmath29 or @xmath30 .
+ firstly , under the scale transformation @xmath42 , eq .
( [ eq16 ] ) takes the form @xmath43 where subscript represents derivative with respect to @xmath44 .
considering eq .
( [ eq17 ] ) as a linear differential equation for @xmath29 , where @xmath30 is an arbitrary function , we obtain @xmath45 where @xmath46 is an integrating constant .
similarly , using the transformations @xmath47 , @xmath48 , and @xmath49 in eq .
( [ eq16 ] ) , after some algebra we obtain respectively @xmath50 @xmath51 and @xmath52 where @xmath53 , @xmath54 and @xmath55 are constants of integration . thus choosing any given function @xmath29 or @xmath30 in cases ( i ) , ( ii ) , ( iii ) and ( iv ) , one can obtain @xmath29 or @xmath30 and hence @xmath28 from ( [ eq14 ] ) .
we consider the following four cases : in this case , eq .
( [ eq18 ] ) gives @xmath59 and then from ( [ eq14 ] ) , we obtain @xmath60 hence the metric ( [ eq3 ] ) reduces to the new form @xmath61 - e^{2\alpha x}\left[\left(k_{1}\tau^{n } + 2\ell \tau^{\ell_{1}}\right)^ { 2 } dy^{2 } + \tau^{2n } dz^{2}\right],\ ] ] where @xmath62 for this derived model ( [ eq24 ] ) , the pressure , energy density and cosmological constant are given by @xmath63 @xmath64 + \alpha^{2}\tau^{-n } \left(k_{1 } \tau^{n } + 2\ell \tau^{\ell_{1}}\right)^{-1},\ ] ] @xmath65 @xmath66 - 3\alpha^{2}\tau^{-n } \left(k_{1 } \tau^{n } + 2\ell \tau^{\ell_{1}}\right)^{-1}.\ ] ] eq .
( [ eq25 ] ) , with the use of ( [ eq11 ] ) and ( [ eq26 ] ) , reduces to @xmath67\ ] ] @xmath68 eliminating @xmath69 between ( [ eq26 ] ) and ( [ eq27 ] ) , we obtain @xmath70 @xmath71\ ] ] @xmath72 from eq .
( [ eq28 ] ) , we observe that at the time of early universe the cosmological constant ( @xmath0 ) increases rapidly during a very short time and then it decreases as time increases ( see figure 1 ) .
we also observe that the value of @xmath0 is small and positive at late times which is supported by recent type ia supernovae observations [ 22 - 24 ] .
+ the metric ( [ eq24 ] ) is a four - parameter family of solutions to efes with a perfect fluid . using the scale transformation @xmath73 in eqs .
( [ eq12 ] ) and ( [ eq13 ] ) for this case , the scalar expansion @xmath34 and the shear @xmath74 have the expressions : @xmath75 \left(k_{1}\tau^{n } + 2\ell \tau^{\ell_{1}}\right)^{-3/2}\ ] ] @xmath76 eqs .
( [ eq29 ] ) and ( [ eq30 ] ) lead to @xmath77^{-1}\ ] ] now , we consider four subcases for the parameters @xmath0 , @xmath57 , @xmath41 , @xmath46 whether zero or not .
+ in subcase @xmath78 , metric ( [ eq24 ] ) with expressions @xmath17 , @xmath16 , @xmath34 and @xmath74 for this model are same as that of solution ( [ eq18 ] ) of camci _ et al .
_ [ 44 ] . + in subcase @xmath78 , @xmath79 , after a suitable inverse time transformation , we find that @xmath80,\ ] ] where @xmath81 is a constant of integration and @xmath82 .
the expressions @xmath17 , @xmath16 , @xmath34 and @xmath74 for this model are not given here , since it is observed that the physical properties of this one are same as that of the solution ( [ eq24 ] ) of ram [ 39 ] .
+ in subcase @xmath78 , @xmath83 , after inverse time transformation and rescaling , the metric ( [ eq24 ] ) reduces to @xmath84,\ ] ] where @xmath85 is a constant of integration and @xmath86 . for this solution ,
when @xmath87 and @xmath88 , we obtain einstein and de sitter [ 45 ] dust filled universe . for @xmath89 , @xmath90 and @xmath91 , where @xmath92 is a parameter in ram s paper [ 39 ] , the solution ( [ eq33 ] ) reduces to the metric ( [ eq14 ] ) of ram [ 39 ] . in later case ,
if also @xmath88 , then we get the minkowski spacetime .
+ now , in subcase @xmath78 , @xmath93 , after some algebra the metric ( [ eq24 ] ) takes the form @xmath94,\ ] ] where @xmath95 is a constant , @xmath96 and @xmath97^{2/3}$ ] .
+ for @xmath98 , @xmath99 and @xmath79 from ( [ eq34 ] ) , we obtain that the solution ( [ eq24 ] ) of ram [ 39 ] . +
* some physical aspects of model :* + the model ( [ eq24 ] ) has barrel singularity at @xmath100 given by @xmath101^{\frac{2}{(2 - 5n)}},\ ] ] which corresponds to @xmath102 . for @xmath103 from ( [ eq24 ] ) , it is observed that at the singularity state @xmath100 , @xmath17 , @xmath16 , @xmath0 , @xmath34 and @xmath74 are infinitely large . at @xmath104 , which corresponds to @xmath105 for @xmath106 and @xmath107 , or @xmath108 for @xmath109 and @xmath110 , @xmath17 , @xmath16 , @xmath0 , @xmath34 and @xmath74 vanish .
therefore , for @xmath103 , the solution ( [ eq24 ] ) represents an anisotropic universe exploding from @xmath100 , i.e. @xmath102 , which expands for @xmath111 .
we also find that the ratio @xmath112 tends to a finite limit as @xmath104 , which means that the shear scalar does not tend to zero faster than the expansion .
hence the model does not approach isotropy for large values of @xmath1 .
+ in subcase @xmath78 , @xmath113 , the ratio ( [ eq31 ] ) tends to zero , then the model approaches isotropy i.e. shear scalar @xmath74 goes to zero . for the model ( [ eq33 ] ) , @xmath17 and @xmath16 tends to zero as @xmath104 ; the model would give an essentially empty universe at large time .
the dominant energy condition given by hawking and ellis [ 46 ] requires that @xmath114 thus , we find for the model ( [ eq33 ] ) that @xmath115 .
hence for the values @xmath116 , the universe ( [ eq33 ] ) satisfies the strong energy condition i.e. @xmath117 .
also this model is sheer - free and expanding .
+ in subcase @xmath78 , @xmath93 , for @xmath118 , @xmath10 , it is observed from relations ( [ eq25 ] ) - ( [ eq30 ] ) that @xmath17 , @xmath16 , @xmath34 and @xmath74 are infinitely large at the singularity state @xmath119 . when @xmath104 , these quantities vanish .
we also find that the ratio @xmath120 is a constant .
this shows that the cosmological model ( [ eq34 ] ) does not approach isotropy for large value of @xmath1 . in this model
the dominant energy conditions ( [ eq35 ] ) are then verified for @xmath121 . since @xmath103 , the model ( [ eq34 ] )
satisfies the strong energy condition for @xmath122 .
+ in each of subcases , all the obtained solutions ( [ eq32 ] ) , ( [ eq33 ] ) and ( [ eq34 ] ) satisfy the bianchi identity given in eq .
( [ eq10 ] ) . in this case
( [ eq19 ] ) gives @xmath125 and from ( [ eq14 ] ) , we obtain @xmath126 where @xmath127 .
hence the metric ( [ eq3 ] ) reduces to the form @xmath128 @xmath129,\ ] ] where the constant @xmath53 is taken , without any loss of generality , equal to @xmath130 .
this metric is a three - parameter family of solutions to efes with a perfect fluid .
+ for the above model , the distribution of matter and nonzero kinematical parameters are obtained as @xmath131 @xmath132 @xmath133 @xmath134 with the use of eq .
( [ eq11 ] ) in ( [ eq39 ] ) and ( [ eq40 ] ) , we obtain @xmath135 @xmath136 eliminating @xmath69 between ( [ eq40 ] ) and ( [ eq41 ] ) , we get @xmath137 @xmath138 from eq .
( [ eq42 ] ) , we observe that the cosmological constant ( @xmath0 ) in the model ( [ eq38 ] ) is a decreasing function of time ( see figure 2 ) .
we also observe that the value of @xmath0 approaches a small and positive value at late times which is supported by recent type ia supernovae observations [ 21 - 24 ] .
+ the scalar of expansion @xmath34 and the shear @xmath74 are obtained as @xmath139,\ ] ] @xmath140 from eqs . ( [ eq43 ] ) and ( [ eq44 ] )
, we have @xmath141 in subcase @xmath78 , metric ( [ eq38 ] ) with expressions @xmath17 , @xmath16 , @xmath34 and @xmath74 for this model are same as that of solution ( [ eq27 ] ) of camci _ et al . _
+ in sub - case @xmath78 , @xmath142 ( i.e. @xmath79 ) , we find a similar solution to ( [ eq32 ] ) , and hence this subclass is omitted .
for @xmath83 , the ratio ( [ eq45 ] ) is zero and hence there is no anisotropy . +
after a suitable coordinate transformation , the metric ( [ eq38 ] ) can be written as @xmath143,\ ] ] where @xmath144 is a constant and @xmath145^{2m_{1}}$ ] , @xmath146 , where @xmath147 is a new parameter .
when @xmath148 and @xmath149 , from ( [ eq46 ] ) , we get the solution ( [ eq12 ] ) of ram [ 39 ] . +
* some physical aspects of model : * + the model has singularity at @xmath150 for @xmath151 or @xmath152 for @xmath153 , which corresponds to @xmath154 .
it is a point type singularity for @xmath151 whereas it is a cigar or a barrel singularity according as @xmath153 . at @xmath104 , which correspond to @xmath155 for @xmath151 or @xmath152 for @xmath153 , from eqs .
( [ eq39 ] ) - ( [ eq45 ] ) , we obtain that for @xmath151 , @xmath156 , and @xmath157 ( @xmath158 ) , - @xmath159 ( @xmath110 ) ; for @xmath153 , similar the above ones .
then , clearly , for a realistic universe , it must be fulfill as @xmath150 , @xmath57 and @xmath41 are positive and @xmath160 is an odd positive number ; as @xmath152 , @xmath41 is positive , and @xmath160 an even negative number .
also , since @xmath161 therefore , this model does not approach isotropy for large values of @xmath1 .
+ in subcase @xmath113 for the metric ( [ eq46 ] ) , the pressure , density and cosmological constant are given by @xmath162 @xmath163 eq .
( [ eq47 ] ) , with the use of ( [ eq11 ] ) and ( [ eq48 ] ) , reduces to @xmath164 eliminating @xmath69 between ( [ eq48 ] ) and ( [ eq49 ] ) , we obtain @xmath165}{(t + t_{3})^{2 } } + \frac{\alpha^{2}(1 + 3\gamma)}{k_{4}(t + t_{3})^{2\ell_{1}}}.\ ] ] when @xmath78 , the pressure and energy density are same as that of given in eq .
( [ eq44 ] ) of paper camci _
[ 44 ] . in this case , the weak and strong energy conditions ( [ eq35 ] ) for this solution are identically satisfied when @xmath166 i.e. @xmath167 .
this model is shear - free and expanding with @xmath168 . in this case
( [ eq20 ] ) gives @xmath170 and then from ( [ eq14 ] ) , we obtain @xmath171 hence the metric ( [ eq3 ] ) takes the new form @xmath172 -\ ] ] @xmath173\ ] ] for four - parameters family of solution ( [ eq53 ] ) , the physical and kinematical quantities are given by @xmath174 @xmath175\left(k_{3}t^{n } - 2\ell t^{\ell_{1}}\right)^{-3 } + \alpha^{2 } t^{-n}\left(k_{3}t^{n } - 2\ell t^{\ell_{1}}\right)^{-1},\ ] ] @xmath176 @xmath177\left(k_{3}t^{n } - 2\ell t^{\ell_{1}}\right)^{-3 } - 3\alpha^{2 } t^{-n}\left(k_{3}t^{n } - 2\ell t^{\ell_{1}}\right)^{-1}.\ ] ] eq .
( [ eq54 ] ) , with the use of ( [ eq11 ] ) and ( [ eq55 ] ) , reduces to @xmath178 @xmath179\left(k_{3 } t^{n } - 2\ell t^{\ell_{1}}\right)^{-3 } - 2\alpha^{2}t^{-n}\left(k_{3}t^{n } - 2\ell t^{\ell_{1}}\right)^{-1}.\ ] ] eliminating @xmath69 between ( [ eq55 ] ) and ( [ eq56 ] ) , we have @xmath180 @xmath181 \left(k_{3 } t^{n } - 2\ell t^{\ell_{1}}\right)^{-3}\ ] ] @xmath182 from eq .
( [ eq57 ] ) , we observe that the cosmological constant ( @xmath0 ) in the model ( [ eq53 ] ) decreases as time increases .
we also observe that the value of @xmath0 approaches a small and positive at late times which is supported by recent type ia supernovae observations ( [ eq21 ] ) @xmath183 ( [ eq24 ] ) .
figure @xmath184 clearly shows this behaviour of @xmath0 as decreasing function of time . + the scale of expansion and the shear are obtained as @xmath185 \left(k_{3 } t^{n } - 2\ell t^{\ell_{1}}\right)^{-3/2},\ ] ] @xmath186 from ( [ eq58 ] ) and ( [ eq59 ] ) , we get @xmath187^{-1}.\ ] ] in subcase @xmath78 , metric ( [ eq53 ] ) with expressions @xmath17 , @xmath16 , @xmath34 and @xmath74 for this model are same as that of solution ( [ eq34 ] ) of camci _ et al . _ [ 44 ] . + in subcase @xmath78 , @xmath79 , after an inverse transformation , metric ( [ eq53 ] ) reduces to the form @xmath188,\ ] ] where @xmath189 is an integrating constant .
this model is different from the model ( [ eq32 ] ) by a change of scale .
+ in subcase @xmath78 , @xmath113 , same model as ( [ eq33 ] ) is obtained .
+ further in subcase @xmath78 , @xmath190 , we see that the metric ( [ eq53 ] ) takes the form @xmath191,\ ] ] where @xmath192 is a constant , @xmath193 and @xmath194^{2/3}$ ] .
this metric is only different from ( [ eq33 ] ) by a change of sign .
also , in each of subcase the physical and kinematical properties of obtained metric are same as that of case(i ) .
therefore , we do not consider here them .
+ in this case eq . ( [ eq21 ] ) gives @xmath196 and then from ( [ eq14 ] ) , we obtain @xmath197 hence the metric ( [ eq3 ] ) reduces to @xmath198\ ] ] @xmath199,\ ] ] where , without any loss of generality , the constant @xmath55 is taken equal to @xmath130 .
expressions for physical and kinematical parameters for the model ( [ eq65 ] ) are not given here , but it is observed that the properties of the metric ( [ eq65 ] ) are same as that of the solution ( [ eq38 ] ) , i.e. the case ( ii ) .
in this paper we have described a new exact solutions of efes for bianchi type v spacetime with a perfect fluid as the source of matter and cosmological term @xmath0 varying with time . using a generation technique followed by camci _
[ 44 ] , it is shown that the einstein s field equations are solvable for any arbitrary cosmic scale function . starting from particular cosmic functions , new classes of spatially homogeneous and anisotropic cosmological models
have been investigated for which the fluids are acceleration and rotation free but they do have expansion and shear . for @xmath88 in the metric ( [ eq1 ] ) , we obtained metrics as lrs bianchi type i model ( hajj - boutros [ 37 ] , hajj - boutros and sfeila [ 38 ] , ram [ 39 ] , mazumder [ 40 ] and pradhan and kumar [ 41 ] ) .
it is also seen that the solutions obtained by camci _
[ 44 ] and ram [ 39 ] are particular of our solutions .
+ the cosmological constants in all models given in section @xmath184 are decreasing functions of time and they all approach a small positive value at late times which are supported by the results from recent supernova observations recently obtained by the high - z supernova team and supernova cosmological project ( garnavich _ et al . _
[ 21 ] , perlmutter _
[ 22 ] , riess _
[ 23 ] , schmidt _
et al . _ [ 24 ] ) .
roy and a. prasad : gen .
* 26 * ( 1994 ) 939 .
farnsworth : j. math .
* 8 * ( 1967 ) 2315 .
r. maartens and s.d .
nel : comm .
* 59 * ( 1978 ) 273 j. wainwright , w.c.w .
ince and b.j .
marshman : gen .
* 10 * ( 1979 ) 259 .
+ c. g. hewitt and j. wainwright , phys .
d * 46 * ( 1992 ) 4242 .
collins : comm .
* 39 * ( 1974 ) 131 .
meena and r. bali : pramana - journal of phys .
* 62 * ( 2004 ) 1007 ) .
a. pradhan and a. rai : ast .
* 291 * ( 2004 ) 149 .
a. pradhan , l. yadav and a.k .
yadav : czech . j. phys .
* 54 * ( 2004 ) 487 .
s. weinberg : rev .
* 61 * ( 1989 ) 1 .
s. weinberg : _ gravitation and cosmology _ , wiley , new york , 1972 .
frieman and i. waga : phys .
d * 57 * ( 1998 ) 4642 .
r. carlberg _
et al_. : astrophys . j. * 462 * ( 1996 ) 32 .
m. @xmath6zer and m.o .
taha : nucl .
b * 287 * ( 1987 ) 776 ; + k. freese , f.c .
adams , j.a .
frieman and e. motta : ibid .
b * 287 * ( 1987 ) 1797 ; + j.c .
carvalho , j.a.s .
lima and i. waga : phys .
d * 46 * ( 1992 ) 2404 ; + v. silviera and i. waga : ibid .
d * 50 * ( 1994 ) 4890 .
b. ratra and p.j.e .
peebles : phys .
d * 37 * ( 1988 ) 3406 .
dolgov : in _ the very early universe _ , eds .
gibbons , s.w . hawking and s.t.c .
siklos , cambridge university press , 1983 .
dolgov , m.v .
sazhin and ya.b .
zeldovich : _ basics of modern cosmology _ , editions frontiers , 1990 .
dolgov : phys .
d * 55 * ( 1997 ) 5881 .
v. sahni and a. starobinsky : int .
d * 9 * ( 2000 ) 373 .
zeldovich : sov .
uspekhi * 11 * ( 1968 ) 381 .
carroll , w.h . press and e.l .
turner : ann .
astron . astrophys . *
30 * ( 1992 ) 499 .
m. garnavich _
et al . _ : astrophys
. j. * 493 * ( 1998a ) l53 ( astro - ph/9710123 ) ; astrophys .
j. * 509 * ( 1998b ) 74 ( astro - ph/9806396 ) .
s. perlmutter _
_ : astrophys
. j. * 483 * ( 1997 ) 565 ( astro - ph/9608192 ) ; nature * 391 * ( 1998 ) 51 ( astro - ph/9712212 ) ; astrophys .
j. * 517 * ( 1999 ) 565 .
a. g. riess _
et al . _ : astron .
j. * 116 * ( 1998 ) 1009 ( astro - ph/98052201 ) .
b. p. schmidt _
et al . _ : astrophys . j. * 507 * ( 1998 ) 46 ( astro - ph/98052200 ) .
a. vilenkin : phys .
* 121 * ( 1985 ) 265 .
m. gasperini : phys .
b * 194 * ( 1987 ) 347 ; class .
. grav . * 5 * ( 1988 ) 521 .
k. freese , f.c .
adams , j.a .
friemann and e. mottolla : nucl . phys .
b * 287 * ( 1987 ) 797 .
peebles and b. ratra : astrophys .
j. * 325 * ( 1988 ) l17 .
w. chen and y.s .
wu : phys .
d * 41 * ( 1990 ) 695 .
abdussattar and r.g .
vishwakarma : pramana - j. phys .
* 47 * ( 1996 ) 41 .
j. gariel and g.le denmat : class .
* 16 * ( 1999 ) 149 .
a. pradhan and v.k .
yadav : int . j. mod
d * 11 * ( 2002 ) 893 ; + a. pradhan and i. aotemshi : int .
d * 11 * ( 2002 ) 1419 . a .-
m. abdel - rahaman : gen .
* 22 * ( 1990 ) 655 ; phys .
d * 45 * ( 1992 ) 3492 .
i. waga : astrophys .
j. * 414 * ( 1993 ) 436 . v. silveira and i. waga : phys .
d * 50 * ( 1994 ) 4890 .
r. g. vishwakarma : class .
* 17 * ( 2000 ) 3833 .
j. hajj - boutros : _ lecture notes in physics , gravitation , geometry and relativistic physics _ vol .
212 , p. 51 , springer , berlin ( 1984 ) ; j. math .
* 26 * ( 1985 ) 2297 ; class .
. grav . * 3 * ( 1986 ) 311 .
j. hajj - boutros and j. sfeila : int . j. theor
* 26 * ( 1987 ) 97 .
s. ram : int .
* 29 * ( 1990 ) 901 ; gen .
* 21 * ( 1989 ) 697 .
a. mazumder : gen .
* 26 * ( 1994 ) 307 .
a. pradhan and a. kumar : int
d * 10 * ( 2001 ) 291 .
ellis and m.a.h .
maccallum : comm .
* 12 * ( 1969 ) 108 .
ryan and l.c .
shepley : _ homogeneous relativistic cosmology _ , princeton university press , princeton ( 1975 ) .
u. camci , i. yavuz , h. baysal , i. tarhan and i. yilmaz : ast .
* 275 * ( 2001 ) 391 .
a. einstein and w.de sitter : _ proceeding of the national academy of sciences of the u.s.a . _
* 18 * ( 1932 ) 213 .
s.w . hawking and g.f.r .
ellis , _ the large scale structure of space - time _ , cambridge university press , cambridge ( 1973 ) . | bianchi type v perfect fluid cosmological models are investigated with cosmological term @xmath0 varying with time . using a generation technique ( camci _ et al . _ , 2001 ) , it is shown that the einstein s field equations are solvable for any arbitrary cosmic scale function .
solutions for particular forms of cosmic scale functions are also obtained .
the cosmological constant is found to be decreasing function of time , which is supported by results from recent type ia supernovae observations .
some physical aspects of the models are also discussed . |
recent discovery of exotic baryons @xcite revived the interest in hadron spectroscopy , especially that of exotic hadrons . among several models for exotic baryons ,
promising one is the diquark model @xcite , which assumes a strong correlation of two quarks in the color antitriplet channel . in the jaffe - wilczek ( jw )
diquark model @xcite , the exotic baryons are bound states of two scalar diquarks and one antiquark .
the salient features of jw model are the occurrence of low - dimensional multiplets and small mass difference among the members in the multiplet , which were soon supported by the na49 experiment @xcite .
furthermore , the extremely narrow decay width of exotic baryons was naturally explained in the diquark picture of exotic baryons , where the decay process is suppressed by tunnelling among two diquarks @xcite . as the existence of hadrons containing quarks
more than three is logically possible , its absence in the low - lying hadronic spectroscopy has been a puzzle since the advent of the quark model @xcite . in late seventies , jaffe suggested that the positive - parity scalar mesons of mass less than @xmath0 might be bound states of multi quarks , forming a @xmath1 nonet @xcite . in his scheme
, the scalar nonet is postulated to be the non - exotic components of 4-quark states ( @xmath2 ) .
however , the quark content of light scalar mesons has been controversial and still is @xcite .
recently , it was argued that @xmath3 , the lightest member in the scalar nonet , is not a pure @xmath4 state but mostly a four - quark state , using the unitarity argument in the large @xmath5 analysis @xcite .
( see also a recent paper for the diquark picture of scalar mesons @xcite . ) in this paper , we assume that the scalar nonet is a bound sate of a diquark and an anti - diquark and calculate the decay width of scalar nonet to compare with the experimental data @xcite , using the diquark chiral effective theory @xcite .
the success of quark model in hadron spectroscopy is understood in the framework of the chiral quark effective theory ( @xmath6qet ) , derived by georgi and manohar @xcite .
the @xmath6qet is an effective theory of quantum chromodynamics ( qcd ) below the chiral symmetry breaking scale @xmath7 but above the confinement scale , @xmath8 .
the relevant degrees of freedom of the effective theory therefore contains quarks , gluons and the nambu - goldstone ( ng ) bosons of spontaneously broken @xmath9 chiral symmetry .
the chiral quark effective lagrangian is given as @xmath10 where @xmath11 is the qcd covariant derivative , @xmath12 and @xmath13 is the pion decay constant , @xmath14 the nambu - goldstone fields @xmath15 with @xmath16 . in @xmath6qet
the higher - order terms or explicit chiral symmetry breaking terms are suppressed by powers of @xmath17 .
the typical size of symmetry breaking terms is @xmath18 therefore , they make only small contributions to the naive quark model .
furthermore , by matching with qcd , one finds the axial coupling and the strong coupling to be small as well , @xmath19 and @xmath20 so that the perturbation is reliable .
this is why the naive quark model is so successful in describing non - exotic hadrons @xcite .
similarly , the success of diquark picture for exotic hadrons may be explained by the diquark chiral effective theory ( d@xmath6et ) , proposed recently by sohn , zahed , and one of the authors @xcite .
since the diquarks have mass around @xmath21 @xcite , they fit naturally in the range where the @xmath6qet is applicable . hence ,
in addition to the relevant degrees of freedom of @xmath6qet , the diquark chiral effective theory involves one more degree of freedom , namely a diquark field , defined as @xmath22 where @xmath23 is a mass scale for the diquark field , @xmath24 is the anomalous dimension of the diquark correlator and the greek indices denote colors , while the latin indices @xmath25 denote flavors .
@xmath26 is a charge conjugated field of a quark field , @xmath27 , and @xmath28 .
( here we consider only scalar diquarks , since tensor diquarks @xcite do not contribute to light scalar mesons . )
the diquark chiral effective lagrangian is described by @xmath29 where higher - order interactions of diquarks and quarks is denoted as @xmath30 , which may contain the ( chirally ) covariant derivatives . in this paper
, we will take @xmath31 from the random instanton model @xcite and @xmath32 , the best value fitting the scalar nonet mass @xcite .
the yukawa coupling splits diquark masses , @xmath33 and @xmath34 . and the binding energy between a diquark and an antidiquark in scalar nonet is @xmath35 .
we list these parameters in table [ previous ] .
.the best - fit parameters in @xcite .
they are obtained from the random instanton model and the scalar nonet masses . [
cols="^,^,^,^,^,^",options="header " , ] once we determine the axial couplings @xmath36 and @xmath37 , we calculate in the next section the decay width of @xmath38 in its subdominant mode to see if d@xmath6et works .
since the mass of two kaons is larger than @xmath38 mass , we can not obtain the decay width with eq .
( [ naive2 ] ) .
the broadening of @xmath38 makes the decay into two kaons possible . so the broadening should be considered in order to calculate the decay mode into two kaons . keeping the basic picture
, we introduce the breit - wigner form to revise eq .
( [ naive2 ] ) ; @xmath39 with scalar mass @xmath40 and total decay width @xmath41 .
@xmath42 is @xmath43 and @xmath44 is chosen to be the larger one between the threshold value for the decay mode and @xmath45 .
so @xmath46 for @xmath47 mode .
the wavefunction at the origin and the final momentum are modified as @xmath48 and @xmath49 in ( [ res ] ) .
the normalization constant @xmath50 is adjusted by @xmath51 since @xmath52 is dominant among @xmath38 decay modes , we will use @xmath53 in table [ width1 ] for @xmath54 in eq .
( [ res ] ) .
then we obtain 75 mev with set 1 and 59 mev with set 2 for @xmath55 .
experimentally @xmath56 is 0.14 @xmath57 0.32 and the full decay width of @xmath38 is @xmath58 @xcite . we find our result is a few times larger than the experimental value , @xmath59 . considering the experimental uncertainty and the crude approximation we made for the bound state wavefunction , our result is not too bad .
we calculate the decay width of light scalar mesons in the framework of the diquark chiral effective theory , which was introduced recently to describe exotic hadrons like pentaquarks . using the parameters fixed previously by the mass of scalar mesons , we determine the axial couplings of diquarks with pions and kaons from the experimental data for the decay width of scalars in the dominant decay modes .
the axial couplings turned out to be small , showing that the naive diquark model of jaffe and wilczek is reliable .
then , we calculate the decay width of scalar mesons for the subdominant channel and find that our result is not far off from the experimental data . to conclude , in the diquark chiral effective theory , based on the spontaneously broken chiral symmetry and the diquark picture of exotic hadrons , we find the decay width of scalar mesons in good agreement of experimental data .
it shows that the diquark chiral effective theory captures correct physics of exotic hadrons .
we thank p. ko , f. sannino and y. j. sohn for useful comments .
the work of d.k.h .
is supported by korea research foundation grant ( krf-2003 - 041-c00073 ) .
the work of c.s . was financially supported by pusan national university in the program , post - doc .
2004 .
t. nakano _ et al .
_ [ leps collaboration ] , phys . rev .
* 91 * ( 2003 ) 012002 [ arxiv : hep - ex/0301020 ] ; v. v. barmin _ et al . _
[ diana collaboration ] , phys . atom .
* 66 * ( 2003 ) 1715 [ yad
. fiz .
* 66 * ( 2003 ) 1763 ] [ arxiv : hep - ex/0304040 ] ; s. stepanyan _ et al . _
[ clas collaboration ] , phys .
* 91 * ( 2003 ) 252001 [ arxiv : hep - ex/0307018 ] .
r. l. jaffe and f. wilczek , phys .
* 91 * ( 2003 ) 232003 [ arxiv : hep - ph/0307341 ] .
e. shuryak and i. zahed , arxiv : hep - ph/0310270 . c. alt _ et al . _
[ na49 collaboration ] , phys .
lett . * 92 * , 042003 ( 2004 ) [ arxiv : hep - ex/0310014 ] .
d. k. hong , y. j. sohn and i. zahed , arxiv : hep - ph/0403205 .
r. l. jaffe , phys .
d * 15 * , 267 ( 1977 ) ; _ ibid .
_ 281 ( 1977 ) .
e. van beveren , t. a. rijken , k. metzger , c. dullemond , g. rupp and j. e. ribeiro , z. phys .
c * 30 * , 615 ( 1986 ) ; n. a. tornqvist and m. roos , phys . rev .
lett . * 76 * , 1575 ( 1996 ) [ arxiv : hep - ph/9511210 ] ; j. d. weinstein and n. isgur , phys .
d * 41 * , 2236 ( 1990 ) ; f. e. close , n. isgur and s. kumano , nucl .
b * 389 * , 513 ( 1993 ) [ arxiv : hep - ph/9301253 ] ; f. e. close and a. kirk , phys .
b * 489 * , 24 ( 2000 ) [ arxiv : hep - ph/0008066 ] . f. sannino and j. schechter , phys .
d * 52 * , 96 ( 1995 ) [ arxiv : hep - ph/9501417 ] ; m. harada , f. sannino and j. schechter , phys .
d * 54 * , 1991 ( 1996 ) [ arxiv : hep - ph/9511335 ] ; _ ibid . _ * 69 * , 034005 ( 2004 ) [ arxiv : hep - ph/0309206 ] .
l. maiani , a. polosa , f. piccinini and v. riquer , arxiv : hep - ph/0407017 . m. n. achasov _ et al .
_ , phys .
b * 479 * , 53 ( 2000 ) [ arxiv : hep - ex/0003031 ] ; r. r. akhmetshin _ et al .
_ [ cmd-2 collaboration ] , phys .
b * 462 * , 380 ( 1999 ) [ arxiv : hep - ex/9907006 ] ; a. aloisio _ et al .
_ [ kloe collaboration ] , arxiv : hep - ex/0107024 .
k. hagiwara _ et al .
_ [ particle data group collaboration ] , phys . rev .
d * 66 * , 010001 ( 2002 ) .
a. manohar and h. georgi , nucl .
b * 234 * ( 1984 ) 189 .
t. schafer , e. v. shuryak and j. j. m. verbaarschot , nucl .
b * 412 * , 143 ( 1994 ) [ arxiv : hep - ph/9306220 ] .
s. weinberg , phys .
lett . * 67 * , 3473 ( 1991 ) . | we calculate the decay rate of light scalar mesons , using a diquark chiral effective theory , recently proposed to describe exotic hadrons . in the effective theory
the light scalar mesons are postulated to be bound states of diquark and anti - diquark .
we find our results are in good agreement with experimental data .
the axial couplings of diquarks with pions and kaons are found to be quite small and the perturbation is reliable .
it shows that the diquark picture captures the correct physics of light scalar mesons . |
phase transitions and critical phenomena of rigorously solvable interacting many - particle systems are much sought after in the modern equilibrium statistical mechanics as they offer valuable insight into a cooperative nature of phase changes @xcite . beside this
, the usefulness of mathematically tractable models can also be viewed in providing guidance on a reliability of various approximative techniques , which are often needed for treating more complicated models that preclude exact analytical treatment . _
decorated planar ising models _ , which can be constructed by adding one or more spins on bonds of some original lattice , belong to the simplest mathematically tractable lattice - statistical models ( see ref .
@xcite and references cited therein ) .
the main advantage of decorated ising models consists in a relative simple way of obtaining their exact solutions . as a matter of fact
, several decorated planar ising models can straightforwardly be solved by employing the generalized decoration - iteration transformation @xcite that relates their exact solution to that one of the simple spin-1/2 ising model on a corresponding undecorated lattice , which is generally known for many planar lattices of different topologies @xcite .
quite recently , the decorated ising models consisting of mixed spins have attracted a great deal of attention on account of much richer critical behaviour in comparison with their single - spin counterparts .
exact solutions of the mixed - spin ising models on several decorated planar lattices have furnished a deeper insight into diverse attractive issues of statistical mechanics such as multiply reentrant phase transitions @xcite , multicompensation phenomenon @xcite , annealed disorder @xcite , as well as , the effect of non - zero external magnetic field @xcite .
in addition , the mixed - spin ising models on some decorated planar lattices can also be viewed as useful model systems for some ferromagnetic , ferrimagnetic , and metamagnetic molecular - based magnetic materials ( see refs .
@xcite for excellent recent reviews ) . among the most convenient properties of the generalized decoration - iteration transformation one could mention its general validity , which means that this mapping transformation holds independently of the lattice spatial dimension to be considered
unfortunately , the application of decoration - iteration mapping was until lately basically restricted to one- and two - dimensional decorated lattices due to the lack of the exact solution of the spin-1/2 ising model on three - dimensional ( 3d ) lattices .
the majority of studies concerned with the mixed - spin ising models on 3d decorated lattices were therefore based on approximative analytical methods such as mean - field and effective - field theories @xcite . on the other hand ,
essentially exact results were recently reported by oitmaa and zheng @xcite for phase diagrams of the mixed - spin ising model on the decorated cubic lattice by adopting the decoration - iteration transformation and the critical temperature of the corresponding spin-1/2 ising model on the simple cubic lattice , which is known with a high numerical precision from the high - temperature series expansion @xcite .
another possibility of how rather accurate results can be obtained for the mixed - spin ising model on 3d decorated lattices is to perform extensive monte carlo simulation as recently done by boughrara and kerouad for the decorated ising film @xcite . in the present work
, the mixed spin-1/2 and spin-@xmath0 ising model on the layered 3d decorated lattice will be studied by applying the decoration - iteration transformation , which establishes a precise mapping relationship with the spin-1/2 ising model on the tetragonal lattice .
the reasonable results for the mixed - spin ising model on the 3d decorated lattice can be consequently extracted from the corresponding results of much simpler spin-1/2 ising model on the tetragonal lattice .
two alternative approaches are subsequently used for a theoretical analysis of the latter model : the first analytical approach is based on the zhang s conjectured solution for the spin-1/2 ising model on the orthorhombic lattice @xcite , while the second numerical approach exploits monte carlo simulations . even though there are serious doubts @xcite about a rigour of the conjectured solution for the spin-1/2 ising model on the 3d orthorhombic lattice @xcite , it is quite tempting to utilize it for a theoretical treatment of highly anisotropic spin systems because the zhang s results @xcite correctly reproduce the onsager s exact solution for the spin-1/2 ising model on the 2d rectangular lattice @xcite . from this point of view
, one should expect only small numerical error when treating highly anisotropic quasi-1d or quasi-2d spin systems even if the conjectured solution does not represent the true exact solution and moreover , the correctness of obtained results can easily be checked by the alternative numerical method based on the monte carlo simulations .
the main advantage of the combination of the generalized decoration - iteration transformation with the zhang s conjectured solution is that it preserves the analytic form of the solution to be obtained for the layered ising model of mixed spins .
this advantage is naturally lost in the case of combining the decoration - iteration transformation with monte carlo simulations .
the outline of this paper is as follows . in section [ model ] , the detailed description of the layered mixed - spin ising model is presented at first .
then , some details of the decoration - iteration mapping are clarified together with two alternative ways of how the magnetization and critical temperature can be calculated .
the most interesting results are presented and detailed discussed in section [ result ] .
finally , some concluding remarks are mentioned in section [ conc ] .
let us define the mixed spin-1/2 and spin-@xmath0 ( @xmath2 ) ising model on the 3d layered decorated lattice as it is diagrammatically depicted in fig .
[ fig1 ] . in this figure ,
the solid circles denote lattice positions of the spin-1/2 ising atoms that reside sites of the simple cubic lattice and the empty ones represent lattice positions of the decorating spin-@xmath0 ising atoms lying on the horizontal bonds of the simple cubic lattice . ising model on the layered 3d decorated lattice and its decoration - iteration transformation towards the simple spin-1/2 ising model on the tetragonal lattice .
solid ( empty ) circles denote lattice positions of the spin-1/2 ( spin-@xmath0 ) atoms , while solid and broken lines represent intra- and inter - layer interactions for both mixed - spin as well as effective spin-1/2 ising model , respectively.,scaledwidth=85.0% ] let us further denote the total number of layers by the symbol @xmath3 , the total number of the spin-1/2 atoms within each layer as @xmath4 and the total number of the spin-1/2 atoms as @xmath5 .
the model under investigation can be then defined through the hamiltonian @xmath6 where @xmath7 and @xmath8 are two different kinds of ising spins located in the @xmath9th layer at @xmath10th and @xmath11th lattice position , respectively , and periodic boundary conditions are imposed for simplicity . the parameter @xmath12 denotes the intra - layer interaction between the nearest - neighbour spin-1/2 and spin-@xmath0 atoms , the parameter @xmath13 labels the inter - layer interaction between the nearest - neighbour spin-1/2 atoms from two adjacent layers and the parameter @xmath14 stands for axial zero - field splitting ( azfs ) parameter that acts on the decorating spin-@xmath0 atoms only @xcite .
the partition function of the layered mixed - spin ising model , which is defined through the hamiltonian ( [ eq1 ] ) , can be written after straightforward re - arrangement of some terms in the form @xmath15 , \label{eq2}\end{aligned}\ ] ] where @xmath16 , @xmath17 is boltzmann s constant , @xmath18 is the absolute temperature and the symbol @xmath19 stands for a summation over all possible spin configurations of the spin-1/2 atoms .
it can be readily seen from the structure of the relation ( [ eq2 ] ) that the summation over spin degrees of freedom of the decorating spin-@xmath0 atoms can be performed independently of each other ( there is no direct interaction between the decorating spins ) and before summing over all possible spin configurations of the spin-1/2 atoms .
both these facts enable us to introduce the generalized decoration - iteration transformation @xcite @xmath20 = a \exp ( \beta j_{\rm intra } \sigma_{l , i1 } \sigma_{l , i2 } ) , \label{eq3}\end{aligned}\ ] ] which effectively replaces all the interaction terms associated with the decorating spin @xmath21 and substitutes them by the equivalent expression that depends solely on its two nearest - neighbour vertex spins @xmath22 and @xmath23 .
of course , the decoration - iteration transformation must retain its validity regardless of possible spin states of both the nearest - neighbour vertex spins @xmath22 and @xmath23 and this `` self - consistency '' condition unambiguously determines until now not specified transformation parameters @xmath24 and @xmath25 @xmath26 \left [ \sum_{n =- s}^s \exp(\beta d n^2 ) \right ] \right \}^{1/2 } , \label{eq4a } \\ \beta j_{\rm intra } & = & 2 \ln \left [ \sum_{n =- s}^s \exp(\beta d n^2 ) \cosh(\beta j n ) \right ] - 2 \ln \left [ \sum_{n =- s}^s \exp(\beta d n^2 ) \right ] .
\label{eq4b}\end{aligned}\ ] ] at this stage , the substitution of the decoration - iteration transformation ( [ eq3 ] ) into eq .
( [ eq2 ] ) yields , after straightforward re - arrangement of few terms , the following mapping relationship for the partition function @xmath27 it is quite obvious that the mapping relation ( [ eq5 ] ) relates the partition function of the layered ising model on 3d decorated lattice to that one of the corresponding spin-1/2 ising model on the tetragonal lattice ( see fig . [ fig1 ] ) .
notice furthermore that the effective intra - layer interaction @xmath25 of the corresponding spin-1/2 ising model on the tetragonal lattice is temperature dependent parameter satisfying the self - consistency condition ( [ eq4b ] ) , while the effective inter - layer interaction @xmath28 is temperature independent parameter that is directly equal to the interaction parameter @xmath13 . a calculation of the spontaneous magnetization and other thermodynamic quantities can be now accomplished in an easy and rather straightforward way . adopting the mapping theorems developed by barry _
et al_. @xcite , the sublattice magnetization @xmath29 relevant to the spin-1/2 atoms of the mixed - spin ising model on 3d decorated lattice directly equals to the magnetization of the corresponding spin-1/2 ising model on the tetragonal lattice @xmath30 above , the symbols @xmath31 and @xmath32 denote canonical ensemble averaging performed within the mixed - spin ising model on the 3d decorated lattice and its corresponding spin-1/2 ising model on the tetragonal lattice , respectively . on the other hand , the sublattice magnetization @xmath33 of the spin-@xmath0 atoms
can easily be calculated by combining the exact callen - suzuki spin identity @xcite with the differential operator technique @xcite .
it is noteworthy that this kind of mathematical treatment essentially follows kaneyoshi s procedure @xcite originally developed for the decorated planar ising models , which connects the sublattice magnetization of the spin-@xmath0 atoms with that one of the spin-1/2 atoms through the relation @xmath34 if both sublattice magnetization are known , the total magnetization of the mixed - spin ising model on the 3d decorated lattice is given by the definition @xmath35 .
it is quite obvious from eqs .
( [ eq6 ] ) and ( [ eq7 ] ) that it is now sufficient to find the spontaneous magnetization of the corresponding spin-1/2 ising model on the tetragonal lattice in order to complete our calculation of both sublattice magnetizations . for this purpose
, we will utilize two alternative approaches : the first method adopts the conjectured solution for the spin-1/2 ising model on the orthorhombic lattice @xcite , while the second method takes advantage of numerical monte carlo simulations .
the former analytic procedure employs an explicit expression for the spontaneous magnetization of the spin-1/2 ising model on the tetragonal lattice , which can be easily descended from the zhang s results for the spin-1/2 ising model on the orthorhombic lattice @xcite @xmath36^{3/8 } , \label{eq8}\end{aligned}\ ] ] where @xmath37 and @xmath38 . within the framework of this analytic method , it is also easy to obtain the critical condition that thoroughly determines a critical point of the order - disorder phase transition of the layered ising model on the 3d decorated lattice .
namely , both sublattice magnetization @xmath29 and @xmath33 tend necessarily to zero if the spontaneous magnetization @xmath39 of the corresponding spin-1/2 ising model on the tetragonal lattice vanishes as well .
accordingly , the critical condition that enables to locate the order - disorder phase transition of the mixed - spin ising model on 3d decorated lattice can readily be found from the zhang s critical condition for the spin-1/2 ising model on the orthorhombic lattice @xcite , which contains as a particular case the following critical condition for the spin-1/2 ising model on the tetragonal lattice @xmath40 where @xmath41 and @xmath42 denotes the critical temperature .
it should be nevertheless mentioned that the above critical condition thoroughly determines a critical behaviour of the layered ising model of mixed spins on assumption that the effective intra - layer interaction @xmath25 satisfies the mapping relation ( [ eq4b ] ) and the effective inter - layer interaction is equal to @xmath43 . to avoid a danger of over - interpretation of the obtained results ,
the monte carlo simulations @xcite were further used as the other alternative approach with the aim to provide an independent calculation of the spontaneous magnetization of the corresponding spin-1/2 ising model on the tetragonal lattice defined via the effective interactions @xmath25 and @xmath28 . the main advantage of a combination of the decoration - iteration transformation with the monte carlo method consists in a drastic reduction of the total hilbert space , because the total number of available spin configurations reduces from @xmath44^{n_t}$ ] to @xmath45 after performing the decoration - iteration transformation .
apparently , this drastic reduction of the total hilbert space makes from the monte carlo simulations much more efficient tool for obtaining meaningful results . to be more specific
, we have performed the monte carlo simulations for the spin-1/2 ising model on the tetragonal lattice with the linear size @xmath46 , @xmath47 , @xmath48 , and @xmath49 .
note furthermore that periodic boundary conditions were imposed and all initial spin states were randomly assigned .
the last spin configuration at any temperature was used as an input for maintained simulation at lower temperature .
spin configurations were generated by random passing through the tetragonal lattice and making single spin - flip attemps , which were accepted or rejected according to the standard metropolis algorithm @xcite .
finally , canonical ensemble averages were calculated using @xmath50 monte carlo steps per site after discarding the initial @xmath51 monte carlo steps per site .
the magnetization per site was calculated from the definition @xmath52 , where the symbol @xmath53 denotes the ensemble average performed within the spin-1/2 ising model on the tetragonal lattice with the linear size @xmath3 .
it is worthy to remind that the magnetization @xmath39 then directly equals to the sublattice magnetization @xmath29 relevant to the spin-1/2 atoms of the mixed - spin ising model on 3d decorated lattice .
the other sublattice magnetization @xmath33 of the spin-@xmath0 atoms can easily be enumerated from the relation ( [ eq7 ] ) .
for better accuracy , the critical temperature was determined with the help of fourth - order binder cumulants @xmath54 $ ] @xcite , which intersect each other for different lattice sizes @xmath3 at a critical point according to the finite - size scaling theory @xcite .
in this part , let us proceed to a discussion of the most interesting results obtained for the layered ising model on 3d decorated lattice . before doing this
, it is worthy to mention that all analytical results presented in the preceding section are rather general as they hold for arbitrary quantum spin number @xmath0 of the decorating spins and also independently of whether ferromagnetic or antiferromagnetic interactions @xmath12 and @xmath13 are assumed . in what follows , we will restrict ourselves for simplicity just to an analysis of the particular case with both ferromagnetic interaction constants @xmath55 and @xmath56 .
it should be mentioned , however , that the presented zero - field phase diagrams should remain valid also for layered ising models with the antiferromagnetic interaction(s ) @xmath12 and/or @xmath13 due to an invariance of ising spin systems with respect to the transformations @xmath57 and/or @xmath58 , which merely cause a rather trivial change of the ferromagnetic ( @xmath59 ) alignment to the metamagnetic ( @xmath60 ) , the ferrimagnetic ( @xmath61 ) , or the antiferromagnetic ( @xmath62 ) one .
first , let us take a closer look at finite - temperature phase diagrams , which are shown in fig .
[ fig2 ] in the form of the critical temperature vs. the azfs parameter dependences for several values of the decorating spins @xmath0 and the selected ratio @xmath63 . for several values of the decorating
spins @xmath0 when the ratio between the inter- and intra - layer interactions is fixed to @xmath63 .
solid lines depict the critical temperatures calculated from eq .
( [ eq9 ] ) of our analytical procedure , whereas symbols connected by dotted lines show the corresponding numerical results obtained by using monte carlo simulations.,scaledwidth=95.0% ] in this figure , the solid lines depict analytical results obtained from the critical condition ( [ eq9 ] ) , while symbols connected by dotted lines show the corresponding numerical results acquired by the use of monte carlo simulations .
it is quite obvious from fig .
[ fig2 ] that the phase diagrams obtained from both independent theoretical approaches are in a good qualitative agreement , the numerical data for critical temperatures stemming from monte carlo simulations are in fact just slightly above the respective analytical results .
moreover , it can be also clearly seen from fig .
[ fig2 ] that the overall critical behaviour basically depends merely on whether the decorating spins are half - odd - integer or integer ones . the critical temperature for the spin systems with half - odd - integer decorating spins ( fig .
[ fig2]a ) monotonically decreases upon decrease of the azfs parameter until it asymptotically reaches the critical temperature of the special case with @xmath64 that is of course independent of the azfs parameter .
this rather trivial finding can be straightforwardly attributed to a consecutive lowering of the spin state of the half - odd - integer decorating spins , which generally takes place at sufficiently strong negative values of the azfs parameters @xmath65 on assumption that the relevant spin state changes from @xmath66 to @xmath67 ( @xmath68 ) .
similarly , the critical temperature for the spin systems with integer decorating spins ( fig .
[ fig2]b ) monotonically decreases upon decrease of the azfs parameter until it tends towards zero temperature at some boundary value of the azfs parameter .
the monotonous decrease of the critical temperature can be again explained in terms of a gradual decline of the spin state of integer decorating spins , which takes place at the following values of the azfs parameter @xmath69 provided that the spin state changes from @xmath70 to @xmath71 ( @xmath72 ) .
contrary to our expectations , the critical temperatures of the spin systems with integer decorating spins do not vanish at the boundary value of the azfs parameter , @xmath73 , below which all integer decorating spins tend towards their non - magnetic spin state @xmath74 .
this is the most remarkable finding of our study and we will henceforth explore this striking critical behaviour , which represents a general feature of the spin systems with integer decorating spins , on the simplest model with the integer decorating spins @xmath75 .
let us consider first possible spin arrangements to emerge in the ground state of this particular model system .
it turns out that three different phases may appear in total at zero temperature in dependence on a relative strength of the intra - layer interaction @xmath12 , the inter - layer interaction @xmath13 , and the azfs parameter @xmath14 .
the azfs term @xmath14 plays the role of the anisotropy parameter that forces all decorating spins @xmath75 towards their non - magnetic spin state @xmath74 provided that this parameter is a sufficiently large negative number .
the usual ferromagnetic phase ( fp ) , which can be characterized through the following spin states of the decorating and vertex spins @xmath76 , consequently represents the lowest - energy state just if @xmath77 . note that this finding is consistent with the relevant results of our analytical approach as well as the numerical monte carlo simulations .
on the other hand , it directly follows from the critical condition ( [ eq9 ] ) that the striking quasi-1d ferromagnetic phase ( qfp ) constitutes the ground state in a range of intermediate strong anisotropy parameters @xmath78 , where it exhibits an outstanding spontaneous long - range order unambiguously determined through the spin states @xmath79 .
it should be stressed that the qualitatively same behaviour is also predicted by monte carlo simulations even although it becomes rather hard to estimate accurately the lower boundary of qfp within this numerical technique ( see for details the subsequent part ) .
the absence of any spontaneous long - range order can finally be detected in the disordered phase ( dp ) , which becomes the lowest - energy state on assumption that @xmath80 when the critical condition ( [ eq9 ] ) is taken into account . in this particular case , the sufficiently strong ( negative )
azfs parameter energetically favours the non - magnetic spin state @xmath74 of the decorating spins and hence , there appears the spin state @xmath81 with a complete randomness in the states of the vertex spins ( the vertex spins from the same layer do not effectively feel each other ) .
the most surprising finding resulting from our study of the ground state is a pure existence of qfp , which exhibits a remarkable spontaneous long - range order in spite of the non - magnetic nature of the decorating spins and the effectively quasi-1d character of the spin system . to provide a deeper insight into the mechanism that drives the spin system into one of those three available spin states
, it might be useful to take a closer look at the effective coupling parameters @xmath82 and @xmath83 of the corresponding spin-1/2 ising model on the tetragonal lattice , which were used both in our analytical approach as well as monte carlo simulations .
the effective inter - layer coupling @xmath84 is evidently monotonously decreasing function of the temperature , which diverges as @xmath85 when approaching the zero temperature .
by contrast , the effective intra - layer coupling @xmath82 exhibits much more complex thermal variations , which are for better illustration depicted in fig . [
fig3]a ) for several values of the azfs parameter @xmath86 .
it can be directly proved from the definition ( [ eq4b ] ) that @xmath82 diverges as @xmath85 when reaching the zero temperature either according to the law @xmath87 valid for @xmath88 , or according to the formula @xmath89 valid for @xmath90 .
furthermore , the effective intra - layer coupling tends towards the constant value @xmath91 when approaching zero temperature for the special case @xmath92 , while it exponentially goes to zero by following the law @xmath93 $ ] in the region @xmath94 . notice that all aforedescribed features can also be clearly seen in the dependences shown in fig .
[ fig3]a ) .
are shown in fig .
[ fig3]a ) for several values of the azfs parameter @xmath86 . note that @xmath82 is given by the mapping relation ( [ eq4b ] ) and it does not depend on a strength of the inter - layer interaction @xmath13 .
[ fig3]b ) displays in a semi - logarithmic scale a graphical solution of the critical condition ( [ eq9 ] ) . solid ( broken ) lines depict temperature dependences of the left - hand - side ( right - hand - side ) of the critical condition ( [ eq9 ] ) for several values of the azfs parameter @xmath86 and the ratio @xmath63 .
the points of intersection between broken and solid lines ( full circles ) determine critical points.,scaledwidth=95.0% ] this comprehensive analysis of the effective intra - layer coupling demonstrates that there does not exist ( at least at zero temperature ) any effective intra - layer interaction between the spin-1/2 atoms if @xmath94 and thus , the spin-1/2 atoms from the same layer should become completely independent of each other under this condition . this reasoning would have a simple physical explanation , since the relative strength of azfs parameter @xmath92 is just as strong as to make energy balance between the non - magnetic ( @xmath74 ) and magnetic ( @xmath95 ) spin state of the decorating spins and accordingly , all vertex spins should be effectively separated by the non - magnetic decorating spins @xmath74 whenever @xmath94 . bearing all this in mind , one would intuitively expect that the layered ising model on 3d decorated lattice must be disordered at any finite temperature when @xmath94 . under this assumption ,
the only non - zero term at the zero temperature is the effective inter - layer interaction @xmath96 and the layered ising model on 3d decorated lattice should therefore break into a set of the independent spin-1/2 ising chains ( running perpendicular to the layers ) that do not possess a finite critical temperature .
however , the mathematical structure of the critical condition ( [ eq9 ] ) as well as the numerical results from monte carlo simulations indicate a more involved situation .
our analytical approach implies that the spin system is spontaneously ordered ( disordered ) if the product on the left - hand - side of the critical condition ( [ eq9 ] ) is greater ( less ) than unity .
thus , there exists a possibility that the product on the left - hand - side of the critical condition ( [ eq9 ] ) might be greater than unity despite the zero value of the effective intra - layer coupling , for instance , if a divergence of the effective inter - layer coupling @xmath83 overwhelms the asymptotic vanishing of the intra - layer coupling @xmath82 .
one actually finds in the zero temperature limit ( @xmath97 or equivalently @xmath98 ) that @xmath99 = \biggl \
{ \begin{array}{ll } \infty & \ , { \rm if } \ , \ , \ , \frac{d}{j } > - 1 - \frac{j'}{j } \\ \ : 0 & \ , { \rm if } \ , \ , \ , \frac{d}{j } < - 1 - \frac{j'}{j } \end{array } , \end{aligned}\ ] ] which means that the spontaneous order disappears only at @xmath100 notwithstanding the simple intuitive expectations given above . among other matters ,
this argument might serve in evidence of the outstanding spontaneous long - range ordering qfp that emerges in a range of the intermediate strong anisotropy parameters @xmath78 despite the non - magnetic nature of all decorating spins .
for better illustration , fig .
[ fig3]b ) shows in a graphical form several temperature dependences of the left - hand - side of the critical condition ( [ eq9 ] ) for one particular value of the ratio @xmath101 , which confirm a correctness of the aforedescribed analysis .
it is noteworthy that this figure can also be regarded as a graphical solution of the critical condition ( [ eq9 ] ) that determines a critical point of the layered ising model on 3d decorated lattice as an intersection of both sides of the eq .
( [ eq9 ] ) .
finally , it is worth noticing that the above mentioned analysis is also consistent with the numerical results of monte carlo simulations , which predict the spontaneous order for intermediate values of the azfs parameter @xmath102 as well . for comparison ,
we depict in fig .
[ fig4 ] the critical temperature as a function of the azfs parameter for the particular spin case @xmath75 and two different values of the interaction ratio @xmath103 and @xmath104 .
the critical temperatures , which are displayed in fig .
[ fig4 ] as solid and dashed lines , were obtained by numerically solving the critical condition ( [ eq9 ] ) .
the symbols connected by dotted lines depict the relevant numerical data obtained by using monte carlo simulations . for the particular spin case @xmath75 and two different values of the ratio @xmath103 and @xmath104 between the inter- and intra - layer interactions .
the solid and dashed lines without any symbol show the critical lines obtained from eq .
( [ eq9 ] ) .
the symbols connected by dotted lines display the relevant critical points acquired from monte carlo simulations.,scaledwidth=60.0% ] it is worthwhile to remark that the critical line displayed in fig .
[ fig4 ] for the special case @xmath105 is fully consistent with the formerly published exact results @xcite . in this particular case
, the critical line actually ends up at the expected ground - state boundary @xmath92 at which the spin state @xmath95 changes to the non - magnetic one @xmath74 and there does not appear a striking spontaneous order inherent to qfp .
this is a direct consequence of the fact that the critical condition ( [ eq9 ] ) extracted from the zhang s solution for the spin-1/2 ising model on the orthorhombic lattice @xcite essentially reduces to the famous onsager s solution for the spin-1/2 ising model on the square lattice @xcite .
it should be pointed out , moreover , that the relevant numerical data from monte carlo simulations are lying on this critical line , which confirms accuracy of our monte carlo simulations .
the critical temperature monotonically decreases with a decrease of the azfs parameter also for any non - zero inter - layer interaction @xmath106 until it tends to zero at some stronger ( more negative ) values of the azfs parameter ( see the curve for the particular case @xmath63 ) .
this surprising finding is evident both from our analytical results as well as monte carlo simulations .
however , the decorating spins reside the spin state @xmath95 just if @xmath107 , while they reside the non - magnetic spin state @xmath74 whenever @xmath108 .
from this perspective , the boundary value of the azfs parameter @xmath73 divides the critical line into two different region : the part where @xmath107 corresponds to the critical points of the fp , while the part where @xmath108 corresponds to the critical points of the qfp .
it should be also mentioned that the lower boundary allocating a presence of the spontaneously ordered qfp is @xmath109 according to the critical condition ( [ eq9 ] ) , while it becomes rather hard to locate precisely the lower boundary with the help of monte carlo simulations .
namely , the more and more extensive monte carlo simulations are needed at sufficiently low temperatures in order to overcome finite - size effects that become very important in the parameter space @xmath110 , because the relevant spin system effectivelly splits into a set of weakly interacting spin-1/2 ising chains running perpendicular to the layers . to provide an independent check of a presence of
spontaneously ordered qfp , it might be quite useful to take a look at thermal dependences of the total and sublattice spontaneous magnetizations .
for this purpose , some temperature variations of the total and sublattice magnetizations are displayed in fig .
[ fig5 ] for the particular value of the interaction ratio @xmath101 and several values of the anisotropy parameter @xmath86 .
it is noteworthy that the results obtained from our analytical procedure are in a good qualitative accordance with the numerical estimates of monte carlo simulations .
however , there appears just a small deviation between the relevant results at relatively high temperatures close to a critical point , because our analytical procedure slightly underestimates the critical temperature in comparison with the monte carlo predictions .
[ fig5]a ) shows thermal dependences of the total and sublattice magnetizations , which are typical for @xmath111 and which lead to the most common q - type temperature dependence of the total magnetization .
( solid lines ) and the sublattice magnetization @xmath29 ( dashed lines ) , @xmath33 ( dotted lines ) for the spin case @xmath75 , the fixed value of the ratio @xmath63 and several values of the azfs parameter @xmath86 .
the lines ending at lower critical temperature show the relevant results of our analytical approach , while the lines having a high - temperature tail come from the monte carlo simulations of a lattice with the linear size @xmath112.,scaledwidth=95.0% ] on the other hand , the s - type temperature dependence of the total magnetization can be observed on assumption that the azfs parameter is slightly greater than the boundary value @xmath92 [ see fig .
[ fig5]b ) for @xmath113 .
the stair - like s - shaped dependence with a rapid initial decrease of the total magnetization obviously appears owing to preferred thermal excitations of the decorating spins to the non - magnetic spin state @xmath74 .
namely , these thermal excitations are also reflected in the temperature dependence of the sublattice magnetization @xmath33 and the non - magnetic spin state @xmath74 is close enough in energy to the spin state @xmath95 to emerge in the ground state under this condition .
interestingly , the standard thermal dependences of q - type are recovered for the total and both sublattice magnetizations by selecting the boundary value @xmath92 ( see fig . [
it is worthwhile to remark , nevertheless , that the sublattice magnetization @xmath33 pertinent to the decorating spins starts in this particular case from one half of its saturation value on behalf of the energetic equivalence between the spin states @xmath74 and @xmath95 , which are populated with the same probability . as a result , both sublattice magnetization exhibit the qualitatively same dependences that can not be distinguished within the displayed scale .
last but not least , the interesting l - type dependence of the total magnetization can be found for the azfs parameters @xmath94 as depicted in fig .
[ fig5]d ) for the particular case @xmath114 . as one can see from this figure , the sublattice magnetization @xmath33 of the decorating spins starts from zero and this might be regarded as another convincing evidence of the existence qfp . besides , the temperature - induced increase of the total magnetization evidently comes from the relevant thermal excitations of the decorating spins , which are clearly reflected in the thermal behaviour of the sublattice magnetization @xmath33 . in agreement with this suggestion ,
the observed temperature - induced increase of the magnetization is the more robust , the closer is the azfs parameter to the boundary value @xmath92 , i.e. the closer in energy is the excited magnetic spin state @xmath95 to the non - magnetic spin state @xmath74 emerging at @xmath115 .
in the present work , the critical behaviour and magnetic properties of the layered ising model of mixed spins on 3d decorated lattice are investigated by the use of generalized decoration - iteration transformation , which establishes a precise mapping relationship between the investigated model system and the corresponding spin-1/2 ising model on the tetragonal lattice .
this exact mapping method was subsequently combined either with the conjectured solution for the spin-1/2 ising model on the orthorhombic lattice @xcite or numerical monte carlo simulations with the aim to obtain the meaningful results for the mixed - spin ising model on the layered 3d decorated lattice .
the main advantage of the former procedure is that it preserves essentially analytical form of the results obtained for critical and thermodynamic properties of the layered mixed - spin ising model , while the main advantage of the latter procedure rest in a drastic reduction of the total hilbert space that makes monte carlo simulations very efficient . in the spirit of both these techniques , the ground - state and finite temperature phase diagrams
have been studied along with possible temperature dependences of the total and sublattice magnetizations .
the most interesting finding presented in this work surely represents a theoretical prediction of the striking spontaneous long - range ordering qfp , which appears in spite of the non - magnetic nature of all decorating spins and the effectively quasi-1d character of the spin system .
it should be pointed out , however , that the analogous spontaneous long - range order of the effectively quasi-1d spin system have already been exactly confirmed in the mixed - spin ising model on a decorated square lattice with two different kinds of decorating spins on the horizontal and vertical bonds @xcite .
this noticeable and rather surprising coincidence can readily be understood from the mathematical structure of the critical condition ( [ eq9 ] ) .
indeed , the proposed critical condition ( [ eq9 ] ) for the spin-1/2 ising model on the tetragonal lattice formally coincides with the onsager s critical condition @xcite derived for the spin-1/2 ising model on the anisotropic square ( rectangular ) lattice to which the mixed - spin ising model on anisotropically decorated square lattice is effectively mapped @xcite .
finally , it is worthwhile to remark that the presented solution can be rather straightforwardly extended to account for several additional interaction terms not included in the hamiltonian ( [ eq1 ] ) such as the biaxial zero - field splitting parameter acting on the decorating spins , the next - nearest - neighbour interaction between the vertex spins , the multispin interaction between the decorating spin and its two nearest - neighbour vertex spins and so on . | phase transitions of the mixed spin-1/2 and spin-@xmath0 ( @xmath1 ) ising model on a three - dimensional ( 3d ) decorated lattice with a layered magnetic structure are investigated within the framework of a precise mapping relationship to the simple spin-1/2 ising model on the tetragonal lattice .
this mapping correspondence yields for the layered ising model of mixed spins plausible results either by adopting the conjectured solution for the spin-1/2 ising model on the orthorhombic lattice [ z .- d .
zhang , philos . mag .
87 ( 2007 ) 5309 - 5419 ] or by performing extensive monte carlo simulations for the corresponding spin-1/2 ising model on the tetragonal lattice .
it is shown that the critical behaviour markedly depends on a relative strength of axial zero - field splitting parameter , inter- and intra - layer interactions .
the striking spontaneous order captured to the quasi-1d spin system is found in a restricted region of interaction parameters , where the zero - field splitting parameter forces all integer - valued decorating spins towards their non - magnetic spin state . , , and ising model , decoration - iteration transformation , monte carlo simulations , phase transitions 05.50.+q , 05.70.jk , 64.60.cn , 75.10.hk , 75.10.-b , 75.30.kz , 75.40.cx |
among the so - called ` secondary ' physical mechanisms that re - process the cmb anisotropy , such as gravitational lensing , time delay , and the sunyaev - zeldovich effect , time delay by foreground inhomogeneities in the matter distribution appears to be investigated least . in the recent period a detailed treatment of the problem
was provided by hu & cooray ( 2001 , hereafter hc01 ) , although the general framework for calculating gravitational perturbation effects on the cmb as published in the review of cmb lensing by lewis & challinor ( 2006 ) could also be employed to carry the study further .
in hc01 the authors found an infra - red logarithmic divergence in the variance of the _ absolute _ ( or total ) time delay along a randomly chosen direction to the lss .
they ` renormalized ' this infinity by subtracting the contribution from the ` monopole ' term of the matter power spectrum , corresponding to the removal of a constant uniformly across the sky .
nevertheless , the remaining ( finite ) quantity still carries the divergence trend .
more precisely the variance is dominated by long wavelength fluctuations , prompting hc01 to consider it as an effect of very large coherence length , @xmath9 2 which does not affect our interpretation of the cmb anisotropy , because the net outcome is simply a gentle and arbitrary distortion of the spherical shape of the lss , completely negligible over the size of one ( or a few ) cycles of cmb acoustic oscillations . the coherence length inferred by hc01
should be viewed with some caution , however , because the large delay excursion of @xmath10 1 mpc / c calculated there , which involved only the _ zeroth _ order term , the path integral of the perturbing potential itself , stems from the part of the matter power spectrum which carries the scale - invariant harrison - zeldovich dependence @xmath11 , i.e. there is a danger that the large coherence length may simply be due to the infra - red divergence of the variance rather than any genuine physical scale in the matter spectrum . to find the coherence length of relevance to the question of cmb anisotropy distortion ,
it is necessary to pursue the perturbation expansion to the next two orders .
not only are both results free from divergences , but also it is only through a comparison of these two terms that the true coherence scale for variations in the _ relative _ delay between two light paths separated by a small angle @xmath12 would become transparent . we shall find that this scale is defined by a physically significant parameter , viz .
the characteristic wavenumber at which departures of @xmath4 from the harrison - zeldovich behavior occurs for the first time .
the consequence is that appreciable distortion of the lss radius , with both amplitudes and wavelengths on par with the dimension of the primary acoustic oscillations at the time of last scattering , can exist in principle .
although in the @xmath13 0 limit the matter power spectrum has the form @xmath11 , the behavior of @xmath4 at large @xmath14 is more complicated than an exponential cutoff .
it is possible , however , to break down _ any _ general @xmath4 into constituent terms , each of the form @xmath15 , and sum up the time delay fluctuation contributions from all the terms , because @xmath4 has the meaning of a variance , i.e. it too is additive .
we may therefore write @xmath16 this empirical representation of @xmath4 is not the same as the more commonly used ones ( e.g. efstathiou , bond , and white 1992 ) but , as shall be seen in section 3 , the exponential form reveals coherent length scales of foreground effects in a transparent way ; in any case , provided our formula for @xmath4 fits the observational data ( see below ) the detailed structure of the terms used to model the spectrum is of no significance . the resulting value of @xmath17 in eq .
( 1 ) that we shall obtain is @xmath18 and leads , by eq .
( a-6 ) , to @xmath19 for a @xmath20 0.3 , @xmath21 0.7 , and @xmath22 0.7 cosmology ( bennett et al 2003 , spergel et al 2007 ) .
this agrees well with the cmb temperature modulation of @xmath23 at small @xmath14 as measured by wmap ( see e.g. bennett et al 2003 ) , as it ought to , because from eq .
( a-5 ) @xmath24 where the final step is explained in the material around eq .
( 18.14 ) of peacock ( 1999 ) .
the consistency between @xmath8 as derived from our @xmath25 0 matter spectrum and the large scale cmb anisotropy re - assures us that any corrections we ignored , such as the effect of vacuum domination at @xmath26 0.3 , are indeed minor .
if @xmath4 has the simple form involving only the first term of eq . ( 1 ) with @xmath27 1 , we may work out from eq .
( a-5 ) the correlation function for the perturbing newtonian potential @xmath28 that arises from the linear growth of primordial density contrasts , as @xmath29 where @xmath28 is assumed to be time - independent and gaussian distributed , and in the final step use was made of eqs .
( a-4 ) and ( a-6 ) . if @xmath4 is given by the full eq .
( 1 ) instead , then the rightmost side of eq .
( 5 ) will be a sum of similar terms , each carrying the exponent @xmath30 and with @xmath31 replaced by @xmath32 .
implementing now the observed power spectrum , the most up to date data are from the 2dfgrs galaxy survey ( cole et al 2005 ) after they are deconvolved and aligned with the wmap1 normalization by setting the @xmath33 parameter to @xmath34 0.74 ( sanchez et al 2006 ) .
we found that the resulting dataset can adequately be fitted with a function for @xmath4 of the form given by eq .
( 1 ) and involving three exponential terms , with the value of @xmath17 as already quoted in eq . ( 2 ) and the values of @xmath35 and @xmath30 ( @xmath36 1,2,3 ) as shown in table 1 .
this best - fit spectrum , which closely follows that of wmap1 s @xmath6cdm model ( spergel et al 2003 ) is plotted in fig .
1 . model of the wmap1 normalized 2dfgrs data as given by eq .
( 1 ) with @xmath37 and the remaining parameters as shown in table 1 .
the hubble constant assumed is @xmath22 0.7.,width=384 ]
we proceed towards calculating the excursion in time delay along two light paths of equal lengths .
our starting point is a flat universe having its average density at the critical value , as required by the wmap observations ( bennett et al 2003 , spergel et al 2007 ) .
we employ cartesian comoving coordinates and the conformal time coordinate @xmath38 , so that in a perturbed flat frw space light propagates along the null geodesics of the metric @xmath39 as though the expansion factor @xmath40 plays no role .
let light signals arrive at the observer s origin from a source at @xmath41 .
although the signal is coming towards us , we can , by optical reciprocity , solve the equation starting at @xmath42 and following its path backwards . .parameters ( with 1-@xmath43 errors ) for the best @xmath4 model of the wmap1 normalized 2dfgrs data ( @xmath22 0.7 )
. the form of the model is given by eq .
three exponentials were needed to fit the data . [ cols="<,^,^,>",options="header " , ] suppose a signal arrives from a direction making a small angle @xmath44 w.r.t . the @xmath45 axis .
this may correspond to some ` off - axis ' point on the lss in the case of the cmb .
the conformal time of travel for our light signal is , from eq .
( 6 ) , @xmath46dx',\ ] ] with an ensuing time delay of @xmath47 where @xmath45 and @xmath48 can be present day physical distances if we set @xmath49 ( here we ignored the geometric time delay , with is comparatively negligible as pointed out by hc01 ) . moreover , if we use the current value of @xmath28 , as in the previous section , to calculate @xmath50 , then @xmath51 .
the ` absolute ' variance of the time delay along a random direction @xmath44 to the lss @xmath52 is logarithmically divergent because the quantity @xmath53 is , by eq .
( 5 ) , of the form @xmath54 at small @xmath14 . as stated in the beginning of the paper , hc01 renormalized @xmath55 by subtracting an infinite constant from it . of more relevance to understanding the cmb acoustic peaks
is the relative delay in the arrival time between the above signal and another light signal emitted simultaneously from the same distance , but along the ` on - axis ' direction @xmath56 , i.e. @xmath57 where @xmath58 is the gradient operator transverse to the vector @xmath59 .
assuming for the time being that the form of @xmath4 is given by eq .
( a-4 ) , we can construct the correlation function with the help of eq .
( 5 ) , as @xmath60 + { \frac}{r_ir_j}{r^2}\left[{\frac}{3b}{r}\arctan\left({\frac}{r}{b}\right)-2 -{\frac}{b^2}{r^2+b^2}\right]\right\},\nonumber\\\end{aligned}\ ] ] where @xmath61 and the indices @xmath62 denote _ two _ orthogonal components in directions transverse to @xmath45 .
this enables us to derive the lowest order term for the variance in the relative time delay @xmath63 ^ 2 = { \langle}[{\tau}({\mbox{\boldmath{$\theta$}}})-{\tau}({\mbox{\boldmath{$0$}}})]^2{\rangle}$ ] , viz . the first term on the right side of eq .
( 10 ) , as @xmath64 ^ 2 = ( { \delta}{\phi})^2{\mbox{\boldmath{$\theta$}}}^2\int_0^x\,2x'dx'\int_0^x\,2x''dx '' \left[{\frac}{1}{r^2}-{\frac}{b}{r^3}\arctan\left({\frac}{r}{b}\right)\right],\ ] ] where @xmath65 , consistent with eq .
( 12 ) .
transforming now to the new variables @xmath66 and @xmath67 , the resulting integrand is symmetric in @xmath68 , so we can restrict the range of @xmath68 to positive values , introducing an extra factor of 2 .
thus @xmath69 ^ 2 & = & 2({\delta}{\phi})^2{\mbox{\boldmath{$\theta$}}}^2 \int_0^x d\tilde x\int_{\frac{\tilde x}{2}}^{x-\frac{\tilde x}{2}}d\bar x { \frac}{4\bar x^2-\tilde x^2}{\tilde x^2 } \left[1-{\frac}{b}{\tilde x}\arctan\left({\frac}{\tilde x}{b}\right)\right]\nonumber\\ & = & { \frac}{4}{3}({\delta}{\phi})^2{\mbox{\boldmath{$\theta$}}}^2\int_0^x d\tilde x { \frac}{2x^3 - 3x^2\tilde x+\tilde x^3}{\tilde x^2 } \left[1-{\frac}{b}{\tilde x}\arctan\left({\frac}{\tilde x}{b}\right)\right ] .
\end{aligned}\ ] ] by restricting ourselves to the limit @xmath70 ( ( appropriate to emission distances @xmath71 0.1 mpc - see table 1 - the cmb lss clearly satisfies this criterion ) , it becomes straightforward to complete the calculation , because only one term stands out .
the result is @xmath64 ^ 2={\frac}{4}{3}({\delta}{\phi})^2{\mbox{\boldmath{$\theta$}}}^2 \frac{x^3}{b}\arctan\left({\frac}{x}{b}\right ) = { \frac}{2\pi}{3}({\delta}{\phi})^2{\mbox{\boldmath{$\theta$}}}^2 { \frac}{x^3}{b}.\ ] ] note that according to eq .
( 14 ) the variation in the time delay difference between two points a and b on the lss subtending an angle @xmath44 at the observer o is @xmath72 .
this behavior indicates that we are in the regime of _ coherent delay _ , i.e. provided @xmath12 is sufficiently small the two rays sampled a primordial matter potential gradient which may be regarded as constant .
thus , if in figure 2a a third point s@xmath73 on the lss is collinear with two other points s@xmath74 and s@xmath75 , then a constant gradient would imply equality between the ` s@xmath74 to s@xmath75 ' and ` s@xmath75 to s@xmath73 ' relative delays , i.e. both are @xmath76 .
consequently the net s@xmath74 to s@xmath73 delay will be @xmath77 .
since the angular separation between s@xmath74 and s@xmath73 is @xmath78 at o , we have @xmath79 , fully consistent with the @xmath72 dependence .
is small enough that only the first spatial derivative of the potential is responsible for any delay , then the signals from s@xmath74 and s@xmath75 will e.g. have reached a and b when that from s@xmath73 is at c@xmath80c , where aa @xmath80 2bb. the result is a net displacement of the angular position of all three sources by the same amount to the right .
_ 2b _ : if we reverse the light paths by letting the observer be a single source and the former ( s@xmath74 , s@xmath75 , s@xmath73 ) be replaced by ( b , o , a ) , then the source will appear shifted in the opposite direction ( i.e. to the left ) w.r.t .
the observer whose telescope aperture is ab , as it should do.,title="fig:",width=384 ] is small enough that only the first spatial derivative of the potential is responsible for any delay , then the signals from s@xmath74 and s@xmath75 will e.g. have reached a and b when that from s@xmath73 is at c@xmath80c , where aa @xmath80 2bb. the result is a net displacement of the angular position of all three sources by the same amount to the right . _
2b _ : if we reverse the light paths by letting the observer be a single source and the former ( s@xmath74 , s@xmath75 , s@xmath73 ) be replaced by ( b , o , a ) , then the source will appear shifted in the opposite direction ( i.e. to the left ) w.r.t .
the observer whose telescope aperture is ab , as it should do.,title="fig:",width=384 ] turning to the @xmath81 term , which takes into account the difference in the potential gradient between the two rays , its exact mathematical form can be derived directly from the last integral of eq .
( 10 ) , but in this paper we will present instead another method of calculation which also provides unusual insights to the relationship between time delay and gravitational lensing . such material is to be found in sections 4 and 5 .
for now we simply give the result , i.e. an expansion of @xmath63 ^ 2 $ ] to include the @xmath81 term .
it is @xmath82 ^ 2 \approx { \frac}{2\pi}{3}({\delta}{\phi})^2\theta^2 { \frac}{x^3}{b } \left(1 - \frac{1}{20 } \frac{x^2 \theta^2}{b^2 } \right)~{\rm for}~\theta < \theta_m;\ ] ] or , if @xmath4 has the more general form of eq .
( 1 ) , the contribution from the @xmath83th term to the variance ( note the label @xmath83 here has a _ different _ meaning from that in eq .
( 11 ) ) will be @xmath84 ^ 2 \approx { \frac}{2\pi}{3}({\delta}{\phi})^2\theta^2 x^3 \frac{a_i}{b_i } \left(1 - \frac{1}{20}\frac{x^2 \theta^2}{b_i^2}\right ) ~{\rm for}~\theta < \theta_m^i,\ ] ] where in eqs .
( 15 ) and ( 16 ) the parameters @xmath85 and @xmath86 are yet to be defined below - see eqs . (
17 ) and ( 20 ) . from eq .
( 15 ) we see that the higher order ( @xmath87 ) term , once it assumes importance , will halt the linear rise of @xmath76 with @xmath12 . eventually , when @xmath12 becomes large enough , all the higher order terms of eq .
( 10 ) will take their place to ensure that @xmath63 ^ 2 $ ] reaches constancy because of the logarithmic divergence problem mentioned in section 1 .
this has no significant impact on any of the results presented in this paper , however . ] , as it must do , because two widely separated rays are _
uncorrelated_. thus , once the @xmath87 term is no longer negligible , _ incoherence _ takes over .
a reasonable ( and conservative ) way of estimating the maximum ( or ` plateau ' ) value of @xmath63 ^ 2 $ ] is to find the angle at which the first @xmath12-derivative of the right side of eq .
( 15 ) vanishes .
this occurs when @xmath88 at which point @xmath89 reaches the constant ( saturation ) value of @xmath82 ^ 2 = [ \delta\tau ( \theta_m)]^2 = \frac{10\pi}{3 } ( \delta\phi)^2 xb~{\rm for}~\theta \geq \theta_m\ ] ] here @xmath85 may be defined as the _ coherence angle _ for time delay .
it is also closely related to the expansion parameter of the perturbation expansion of eq .
( 15 ) , because the ratio of the second order to the first order term of eq .
( 15 ) is @xmath90 .
when a more general form of @xmath4 , like eq .
( 1 ) , is considered , we have @xmath84 ^ 2 = [ \delta\tau_i ( \theta_m^i)]^2 = \frac{10\pi}{3 } ( \delta\phi)^2 x a_i b_i~{\rm for}~\theta \geq \theta_m^i,\ ] ] where @xmath91 for each component @xmath83 of @xmath4 in eq . ( 1 ) .
the total variance of time delay is then the sum over @xmath83 of all the individual variances @xmath92 ^ 2 $ ] as given by eqs .
( 16 ) and ( 19 ) .
obviously , for rays separated by angles @xmath93 the time delay is always completely incoherent ( or random ) .
as will be demonstrated in the next section , the significance of the transition angle @xmath85 is that for separation @xmath94 , the variance for the coherent ( or systematic ) and incoherent delay is given by the first and second order terms of eq .
( 15 ) respectively .
relative time delay @xmath95 between two rays from the lss due to the various terms @xmath83 of primordial density perturbation of eq .
( 1 ) . in the limit @xmath96 where @xmath86 given by eq .
( 20 ) , @xmath95 is given by eq .
( 16 ) . when @xmath97 , @xmath95 saturates to the value in eq .
long dashes correspond to the first term @xmath98 , and the dotted line to the second term @xmath99 ( the @xmath100 term was ignored due to its insignificance ) .
values of @xmath35 and @xmath30 are shown in table 1.,width=384 ] how does time delay affect cmb observations ? when light signals emitted during the same redshift @xmath101 from two points on the lss which subtend the angle @xmath12 at the observer are to arrive simultaneously _ and yet _ the times of flight are delayed w.r.t .
each other , the signals would necessarily have covered different distances - the slower one must have undertaken a shorter journey .
as illustrated in figure 4 , the consequence is a ` tilt ' of the lss in that vicinity .
hence , an angle @xmath12 which normally corresponds to an anisotropy on the lss spanning the comoving distance @xmath102 would now be randomly mapped to a _ larger _
distance @xmath103 where , from figure 4 , @xmath104 ^ 2}{2x^2\theta^2 } = \frac{1}{2x^2\theta^2 } \sum_i [ \delta\tau_i ( \theta)]^2,\ ] ] with the contribution to @xmath105 from the @xmath83th term of eq .
( 1 ) given by eqs . ( 16 ) and ( 19 ) .
the net outcome is a _
systematic _ shift of the cmb acoustic peaks , which are features with fixed comoving scales , towards smaller angles or higher harmonic numbers @xmath0 .
it should be mentioned that another possible test of cosmological time delay concerns two light paths propagating through different parts of the universe , but connecting the same source with the observer , i.e. a strong gravitational lensing scenario , under which multiple images occur .
if this source undergoes flaring behavior , a time lag between the light curves of a pair of multiple images might result from perturbations in the distribution of primordial matter . in practice , however , this is a sensitive test only if the lowest order ( @xmath106 ) term contributes to the delay .
as it turns out , this term does _ not _ play a role because the potential time delay we have hitherto been considering is cancelled by another effect - the geometric delay - which _ is _ important under this scenario ( see bar - kana 1996 and seljak 1994 ) . since the higher order terms are too feeble to be measured , no meaningful constraints can be provided here . , the quadrature sum ( over all @xmath83 ) of the @xmath107 contributions as given by eqs .
( 16 ) and ( 19 ) , because it is the relative delay between the light signals emitted at a and c , with the path ao suffering from more delay than bo in this case .
since any lss plane samples the same distribution of hot and cold spots as any other , the tilting of the lss by time delay invariably means an observer charted slightly larger physical distance on the lss for a certain angular separation between two temperature sensors .
this _ systematically _ causes all acoustic peaks to move towards smaller @xmath12 , or larger @xmath0 , in the manner of eq .
( 21).,width=288 ]
a significant gain in our general understanding of foreground effects is afforded by examining also the question of cmb lensing .
in addition to the @xmath45-axis of section 3 , which points along a radial direction from the observer s origin back to the lss , we now introduce two more cartesian ( comoving ) coordinates @xmath108 , both measuring distances transverse to the light path . tracing the path towards the lss as before ,
the equation of motion for a light signal is @xmath109 let a signal arrive from some direction which makes a ( small ) angle @xmath44 with the @xmath45 axis .
then the solution to ( 28 ) is @xmath110 with appropriate initial conditions . from eq .
( 30 ) emerges that the correlation between the values of the deviation angle @xmath111 for different values of @xmath44 , @xmath112 by means of eqs .
( 24 ) and ( 11 ) , one can expand the ensuing correlation function @xmath113 as a taylor series ( see appendix b for details on intermediate steps ) , i.e. @xmath114 where the sum represented by a repeated index is over the two transverse dimensions .
an interesting connection exists between sections 3 and 4 , in that the key formulae derived in each section , viz .
( 15 ) and ( 25 ) , are closely related to each other because of a physical reason - the reciprocity of light propagation .
we first consider the scenario of a source s emitting light signals that enter an observer s circular telescope aperture at the extremeties a and b such that the angle asb is @xmath115 , see fig 2b .
let us assume that when the universe is homogeneous , the wavefronts are parallel to the aperture as they arrive , i.e. the signals reach points a and b simultaneously .
we further suppose that the same is true for two points c and d on the extremeties of another aperture diameter perpendicular to the ab line - the source is ` on - axis ' . in the presence of primordial matter density perturbation ,
let the signal propagating in the sb direction reach the point b when the sa signal has already arrived at a , i.e. the sa signal suffered from a smaller delay .
the outcome is a tilting of the wavefronts as they enter the aperture - the source is now seen to have moved ` off - axis ' along the ab line to a new position leftwards of the center o. the angle of tilt is @xmath116 in the limit of coherent time delay ( i.e. small @xmath12 ) where @xmath117 is the distance bb and @xmath118 is the diameter ab . repeating our above argument to the points c and d , and noting that the relative delay between the sc and sd directions is _ independent _ of that between sa and sb , we realize that the variance @xmath119 ^ 2}{d^2}\ ] ] applies to the overall shift in the position of the source on the two dimensional sky .
of particular interest is the fact that for a given source distance @xmath45 the quantity @xmath120 thus , in order for the positional shift to be the same amount irrespective of aperture size @xmath118 we must have @xmath121 where @xmath122 . from section 3
we saw that this condition holds only when @xmath123 , hence @xmath118 , is small .
to link sections 3 and 4 , however , we have to consider a second scenario , under which the distribution of primordial matter remains the same as before , but a , o , and b are now three simultaneously emitting sources and s is the observer ( equipped with a small telescope ) .
this reversal of the light paths converts figure 2b into figure 2a . by the reciprocity of light propagation
, the signals must arrive at s in such a way that the change in the positions of the three sources as perceived by s after the ` turning on ' of the matter perturbation involves an angular shift by the same amount @xmath124 for each source . yet according to section 4 , this shift has a variance given precisely by the correlation function @xmath113 , where @xmath44 is the angle the two sources subtend at the observer s , viz .
@xmath125 with @xmath113 being given by eq .
( 25 ) . if we now take the limit @xmath126 . by eqs ( 28 ) and ( 22 ) , the latter in section 4 , we have @xmath127 which is a constant shift ( of o w.r.t .
a , and b w.r.t .
o ) independent of @xmath118 and @xmath123 .
such a behavior is also consistent with the requirement stated at the end of the previous paragraph when we considered our first scenario .
thus , from eqs .
( 26 ) , ( 27 ) , and ( 29 ) we deduce that @xmath64 ^ 2= { \frac}{2\pi}{3}({\delta}{\phi})^2{\mbox{\boldmath{$\theta$}}}^2 { \frac}{x^3}{b},\ ] ] which is the low @xmath12 limit of the variance in the relative time delay between two light paths separated by angle @xmath44 , as derived in eq .
( 15 ) of section 3 .
thus , it is now clear that sections 3 and 4 can be unified by the principle of light reciprocity . of even more interest , however , is the regime of larger @xmath123 , where incoherence between the two rays becomes important .
here we already saw from eq .
( 25 ) of section 4 that @xmath113 , hence @xmath128 by eq .
( 28 ) , is no longer a constant , but decreases away from constancy as @xmath123 increases .
this decrease is expected , because when @xmath123 becomes sufficiently large the two light paths are independently perturbed by _
different _ primordial density fluctuations , i.e. the correlation function @xmath129 , hence @xmath128 by eq .
( 28 ) , assumes relatively small values .
the two sources shift their positions randomly w.r.t .
each other , leaving behind little net systematic drift of their centroid across the sky ( according to s ) .
in fact , when we compare eq .
( 25 ) in full with eqs .
( 27 ) and ( 28 ) , we obtain an expression for @xmath63 ^ 2 $ ] in complete agreement with eq .
this consistency provided an important cross - check of the robustness of our analysis , and explains why the @xmath130 term in the variance represents incoherence effects - it is precisely this term that decorrelates @xmath113 . to labor upon this point even further , we observe that the variance of the _ random relative deflection _ between the two rays : @xmath131 ^ 2{\rangle}=2[c(0)-c({\mbox{\boldmath{$\theta$}}})],\ ] ] has the @xmath130 coefficient in is leading term . thus , it is completely clear that while the @xmath132 coefficient concerns absolute deflection and coherent delay , the physics of @xmath130 is relative deflection and incoherent delay . a word of caution , however , before we leave this section .
although the phenomena of time delay and deflection are two sides of the same coin , the ` flipping of the coin ' involves reversing light path arrows , i.e. for a given set of arrows we should _ not _ conclude that delay presents no new physical effects than those ensuing from lensing .
in particular , for the cmb which is emitted from a three dimensional distribution of sources , a varying time delay on simultaneously observed signals from different directions allows _ depth _ to play a role in the problem - as is already explained towards the end of section 3 .
this depth effect can not be reproduced in any way by lensing .
we finally return to the original subject of this paper , viz .
the degree to which time delay by foreground primordial matter re - processes the cmb primary anisotropy . in eq .
( 1 ) and table 1 , we consider only the first two terms in our breakdown of density fluctuations , viz .
@xmath133 of eq .
. their coherence lengths are , by eq .
( 20 ) , @xmath134 0.0289 or @xmath135 108.8 , and @xmath136 0.00886 or @xmath137 354.4 . using eq .
( 21 ) , then , one derives the gaussian smoothing width of cmb anisotropy power , due to the @xmath98 perturbation only , as @xmath138 likewise one also derives the width due to the @xmath99 term as @xmath139 the total width @xmath140 at a given @xmath12 is then the quadrature sum of the contributions ( at that value of @xmath12 ) from both the @xmath98 and @xmath99 terms , as in eq .
( 21 ) . a graph of @xmath141 versus @xmath142 for the @xmath98 , @xmath99 components and their total is shown in figure 5 .
we explained in the end of section 3 that the smoothing kernel for @xmath140 is a one - sided gaussian function in the case of time delay perturbations , i.e. structures on the lss are invariably biased towards having smaller perceived angular sizes irrespective of how time delay may tilt the lss .
specifically the kernel @xmath143 that enters the integrand for the lensed cmb temperature correlation function , viz .
( a5 ) of seljak ( 1996 ) is @xmath144 where @xmath145 with @xmath146 given by eqs ( 31 ) and ( 32 ) added in quadrature , and the factor of 0.6 is due to the fact that the standard deviation of a one - sided gaussian having the usual form for its exponent has standard deviation 0.6@xmath43 .
thus , after applying the ( normal ) simplifying procedures of limber approximation and retention of only the isotropic term @xmath147 in seljak s ( a5 ) , one arrives at the lensed correlation function due to time delay @xmath148 which is to be contrasted with eq .
( a7 ) of seljak ( 1996 ) . owing to the smallness of the effect as dictated by the numerical coefficients in eqs .
( 31 ) and ( 32 ) , however , we did not convolve the acoustic peaks according to the recipe of eq . ( 34 ) , because the distortion is completely unnoticeable to be at least 10 % at the center of the first peak . ] .
we can therefore corroborate hc01 , after this comparatively more detailed and specific investigation ( see section 1 ) , by concluding that relative time delay by foreground primordial density fluctuations does not lead to any appreciable changes in the prediction of the standard cosmological model on the cmb acoustic anisotropy .
due to time delay differences along the lss , i.e. eqs .
( 31 ) and ( 32 ) , is plotted as a function of @xmath0 for contribution from the mode @xmath98 ( long dashes ) and @xmath99 ( short dashes ) of primordial density fluctuation in eq .
( 1 ) and table 1 .
the total @xmath149 curve , obtained by adding the two dashed lines in quadrature , is shown as the solid line .
note that the @xmath100 mode is ignored because of the smallness of its effect.,width=384 ]
in spite of the feeble acoustic distortion induced by the time delay perturbation of the primordial matter distribution , its effect of biasing the acoustic peaks towards higher values of @xmath0 ( as compared with the values under the scenario of a homogeneous universe ) remains in principle an interesting phenomenon . to illustrate , let us repeat the acoustic smearing computation in the previous section using a kernel @xmath143 as determined by assuming a larger normalization of the matter power spectrum ( than the wmap value of eq .
( 3 ) ) , viz . @xmath150
the consequence , shown in figure 6 , is no longer negligible as before .
the systematic shift of the acoustic peaks towards higher @xmath0 ( or smaller @xmath12 ) is now evident , and tends to force the best - fit model parameters to assume new values - that involve @xmath151 1 in particular .
thus clumping does not just introduce local fluctuations in the metric tensor ( as first noted in section 3 of einstein 1917 ) , they _ can _ mimic space curvature , by altering the mean geometry of space as revealed by the statistical behavior of remotely emitted light along many and varied directions .
the magnitude of this influence is by no means necessarily small : even the value of @xmath8 in eq .
( 35 ) lies well within the perturbation criterion of @xmath152 1 , yet it led to a finite cmb @xmath140 which , for the @xmath98 term of eq .
( 1 ) , is _ not _ @xmath153 1 because it maxmizes at 32.5 % when eq .
( 31 ) is rescaled on the right side according to @xmath154 . in the light of the above development ,
let us return to the question of why global space as measured by wmap has zero curvature .
it has widely been accepted that inflation solved the flatness problem , but does not answer why @xmath8 is as small as the value in eq .
( 3 ) . which implies an unusually weak scalar - field coupling
this view is still correct , so long as the ` flatness ' is understood as an intrinsic rather than observed property of space , because indeed @xmath8 , no matter what values it assumes , does not alter the mean intrinsic curvature .
but since , as demonstrated in this work , even within the perturbation regime a larger @xmath8 can make an intrinsically flat geometry appear to wmap as , on average , curved , then one must acknowledge that inflation has not explained why the _ observed _ mean curvature is zero .
for the primordial density perturbation , i.e. with eq .
( 35 ) replacing eq .
( 3 ) , the _ rescaled _ ( hence _ mock _ ) wmap3 data and rescaled wmap3 standard model prediction of the acoustic peaks are plotted , with the latter as a dotted line .
this model is then re - processed by the lensed correlation function due to time delay , viz .
( 34 ) of section 5 , again with the new @xmath8 value of eq .
( 35).,width=384 ] the authors are grateful to shaun cole , tom kibble , chris kochanek , and an anonymous referee for their most helpful advice .
rl thanks esra bulbul for the preparation of figures 2 , 4 , and 7 .
* the basic relationship between gravitational potential perturbation at various length scales and the matter power spectrum * during some epoch @xmath155 , or redshift @xmath156 , let the typical random excursion of the gravitational potential ( from its mean value in a given cosmology ) be @xmath157 at wavenumber @xmath14 or lengthscale @xmath158 . by the poisson equation , @xmath157 is expressible in terms of the matter density variation @xmath159 as @xmath160 , where @xmath161 is the expansion factor .
squaring both sides , we obtain the variance @xmath162 , the additive nature of which enables us to write the equation in differential form , as @xmath163 ^ 2 a^4 \frac{d\delta_k^2}{k^4}.\ ] ] under a matter dominated scenario , @xmath164 , and since @xmath165 , we see that @xmath166 is independent of epoch .
thus it may be written in terms of the present day quantities @xmath167 ^ 2 \frac{d\delta_k^2 ( 0)}{k^4},\ ] ] where @xmath168 relates to the @xmath169 matter power spectrum @xmath4 as inferred from wmap s standard model parameters ( or , better still , the wmap1/2dfgrs external correlation , see below ) by @xmath170 it should be noted , however , that if vaccum energy ( or dark energy ) rather than matter dominates the low @xmath156 universe , eq .
( a-2 ) would have slightly overestimated the value of @xmath157 at an earlier epoch . before proceeding further , therefore
, we must estimate a strict limit on the inaccuracy of eq .
( a-2 ) . under a @xmath6-dominated scenario ,
the rapid expansion freezes the density distribution - the masses can not move fast enough to counter it . as a result , @xmath171 tends to a constant independent of @xmath172 ( i.e. no further growth of density contrasts in the linear regime ) , and @xmath157 decreases as @xmath173 . in a @xmath20 0.3 , @xmath21 0.7 , and @xmath22 0.7 cosmology ( bennett et al 2003 , spergel et al 2007 ) , vacuum domination occurs at about @xmath174 0.32 ( when @xmath175 ) .
thus , between @xmath174 0.32 and @xmath25 0 , @xmath176 decreased from its hitherto constant value by the fraction @xmath177 30 % ( where @xmath172 refers to the expansion factor at @xmath178 0.32 ) .
it is clear then that any lightpath integrations of the potential from @xmath178 0 back to some remote past ( or vice versa ) will certainly not lead to an overestimation of @xmath157 by more than 20 % if ones uses eq .
( a-2 ) for @xmath157 .
let us _ for the time being _ suppose that the @xmath169 matter power spectrum has the simple form @xmath179 so that at small @xmath14 the power spectrum takes the harrison - zeldovich form @xmath11 . then , assembling eqs .
( a-2 ) and ( a-3 ) , we have @xmath180 where @xmath181 with @xmath182 being the hubble constant and @xmath183 1 here - and - after . at sufficiently large @xmath14
, corresponding to wavelengths smaller than the size of the universe ( in today s distance scale ) during matter - radiation equipartition , @xmath4 cuts off because the modes could not grow .
thus , in the simple manner by which eq .
( a-4 ) depicted the matter spectrum , @xmath184 33 mpc , the equipartition horizon for a flat universe with @xmath22 0.7 . the advantage of eq .
( a-4 ) , as can be seen in sections 3 and 4 , is that it affords us an analytical treatment of cmb foreground re - processing by primordial matter - time delay and lensing in particular - by a method of successive approximation from low to high orders which reveals unamiguously the coherent spatial scales of these effects . the correlation function @xmath113 as defined by the first equality of eq .
( 25 ) , section 4 , may be computed using eqs .
( 24 ) and ( 11 ) .
we find @xmath185 , \end{array}\ ] ] where @xmath186 , @xmath187 as in section 3 , and @xmath188 because with the new @xmath45-variables we have , from eq .
( 12 ) , @xmath189 . by setting @xmath190
one arrives at the variance of the deflection angle for a single light ray , @xmath191\nonumber\\ & = & { \frac}{8({\delta}{\phi})^2}{3x^2}\int_0^x d\tilde x { \frac}{2x^3 - 3x^2\tilde x+\tilde x^3}{\tilde x^2 } \left[1-{\frac}{b}{\tilde x}\arctan\left({\frac}{\tilde x}{b}\right)\right ] \end{aligned}\ ] ] evaluating the integral yields @xmath192 in the limit @xmath193 b. in general , the cross - correlation function @xmath194 of eq . ( b-1 ) may be expanded as a taylor series : @xmath195 where @xmath196_{{\theta}=0}\nonumber\\ & = & { \frac}{32{\delta}{\phi}^2}{x^2}\int_0^xd\tilde x \int_{\frac{\tilde x}{2}}^{x-\frac{\tilde x}{2 } } d\bar x{\frac}{[(x-\bar x)^2-{\frac}{1}{4}\tilde x^2]\bar x^2}{\tilde x^4 } \left[3{\frac}{b}{\tilde x}\arctan{\frac}{b}{\tilde x } -2-{\frac}{b^2}{\tilde x^2+b^2}\right]\nonumber\\ & = & { \frac}{4{\delta}{\phi}^2}{15x^2}\int_0^xd\tilde x { \frac}{4x^5 - 10x^3\tilde x^2 + 5x^2\tilde x^3+\tilde x^5}{\tilde x^4 } \left[3{\frac}{b}{\tilde x}\arctan{\frac}{b}{\tilde x } -2-{\frac}{b^2}{\tilde x^2+b^2}\right ] .
\end{aligned}\ ] ] again as before , in the limit @xmath193 b only the leading term needs to be kept .
the result is @xmath197 we finally arrive at the expansion @xmath198 after substituting eqs .
( b-4 ) and ( b-6 ) into eq . | in this paper we specifically studied one aspect of foreground primordial matter density perturbations : the relative gravitational time delay between a pair of light paths converging towards an observer and originating from two points on the last scattering surface separated by the physical scale of an acoustic oscillation .
it is found that time delay biases the size of acoustic oscillations _ systematically _ towards smaller angles , or larger harmonic numbers @xmath0 , i.e. the mean geometry as revealed by cmb light becomes that of an open universe if @xmath1 . since the effect is second order , its standard deviation @xmath2 where @xmath3 is the normalization of the primordial matter spectrum @xmath4 , the consequence is too numerically feeble to warrant a re - interpretion of wmap data .
if , however , this normalization were increased to @xmath5 which is still well within the perturbation limit , the shift in the positions of the acoustic peaks would have been substantial enough to implicate inflationary @xmath6cdm cosmology .
thus @xmath7 is not the only parameter ( and by deduction inflation can not be the only mechanism ) of relevance to the understanding of _ observed _ large scale geometry .
the physics that explains why @xmath8 is so small also plays a crucial role , but since this is a separate issue independent of inflation , might it be less artificial to look for an alternative solution to the flatness problem altogether ? |
the dynamics of interacting species has attracted a lot of attention since the pioneering works of lotka @xcite and volterra @xcite . in their independent studies
, they showed that simple prey - predator models may exhibit limit cycles during which the populations of both species have periodic oscillations in time .
however , this behavior depends strongly on the initial state , and it is not robust to the addition of more general non - linearities or to the presence of more than two interacting species @xcite . in many cases the system reaches a simple steady - state .
a better understanding of the properties of such oscillations is clearly desirable , as such population cycles are often observed in ecological systems and the underlying causes remain a long - standing open question @xcite .
one of the best documented example concerns the canadian lynx population .
this population was monitored for more than hundred years ( starting in 1820 ) from different regions of canada .
it was observed that the population oscillates with a period of approximately 10 years and that this synchronization was spatially extended over areas of several millions of square kilometers @xcite .
several attempts were made to explain these facts ( climatic effects , relations with the food - web , influence if the solar cycle ) without success .
more recently , blasius et al .
@xcite introduced a deterministic three level vertical food - chain model .
the three coupled nonlinear differential equations defining the model contain eight free parameters and two unknown nonlinear functions .
the authors showed that an ad - hoc choice of the free parameters and nonlinear functions explains the experimental data for the canadian lynx . in such mean - field type models
, it is assumed that the populations evolve homogeneously , which is obviously an oversimplification .
an important question consists in understanding the role played by the local environment on the dynamics @xcite .
there are many examples in equilibrium and nonequilibrium statistical physics showing that , in low enough dimensions , the local aspects ( fluctuations ) play a crucial role and have some dramatic effects on the dynamics of the system .
accordingly , a lot of activities have been devoted during the past years to the study of extended prey - predator models .
the simplest spatial generalization are the so called two patches models , where the species follow the conventional prey predator rules within each patches , and can migrate from one patch to the other @xcite .
other works have found that the introduction of stochastic dynamics plays an important role @xcite , as well as the use of discreet variables , which prevent the population to become vanishingly small .
these ingredients are included in the so called individual based lattice models , for which each lattice site can be empty or occupied by one individual of a given species or two individuals belonging to different species .
it was recognized that these models give a better description of the oscillatory behavior than the usual lotka - volterra ( l - v ) equations .
indeed , the oscillations in these lattice models are stable against small perturbations of the prey and predator densities , and they do not depend on the initial state .
it was also found ( in two dimensional systems ) that the amplitude of the oscillations of global quantities decreases with increasing system size , while the oscillations persist on local level .
it was argued that coherent periodic oscillations are absent in large systems ( although , @xcite do not discard this possibility ) . in @xcite lipowski
state that this is only possible above a spatial dimension of 3 . in @xcite provata
emphasize that the frequency of the oscillations are stabilized by the lattice structure and that it depends on the lattice geometry . in some papers ,
the stationary phase diagram was also derived , and different phases were observed as a function of the model parameters , such as an empty phase , a pure prey phase , and an oscillatory region of coexisting preys and predators . in @xcite , a coexistence region without oscillations and a domain of the control parameter space for which the stationary states depend strongly upon the initial condition , were found .
however , in all the above works no systematic finite size studies have been performed , allowing to draw firm conclusions on the phase diagram of the models as a function of their sizes .
it is known @xcite , that in ecological problems the fact that a system has a finite size is more relevant than in most of the cases encountered in statistical physics , for which one concentrates on the thermodynamic limit .
particularly , the size dependence of the amplitude of the oscillations , as well as a detailed description of the critical behavior near the phase transitions have not been investigated .
another relevant question is how much the stationary phase diagrams of these prey - predator models have some generic properties or how much they depend upon the details of the models .
the goal of this paper is to study a simple models of prey - predators on a two - dimensional lattice for which some of the above questions could be answered .
our model is based on a coarse - grained description in the sense that a given cell models a rather large part of a territory and thus can contain many preys or predators .
moreover , predators can not leave without preys in a given cell .
those are the main differences between our model and satulovsky and tom ( st ) model @xcite .
nevertheless , it turns out that the stationary state phase diagram of the two models are quite different .
our model is defined in sec . [
sec : model ] .
although governed by only two control parameters , this model exhibits a rich phase diagram .
two different phases are observed : a pure prey phase , and a coexistence phase of preys and predators in which an oscillatory and a non - oscillatory region can be distinguished . in some limiting cases
the model can be mapped onto another well known nonequilibrium model : the _ contact process _
( cp ) @xcite . in sec .
[ sec : mean - field ] the properties of our model are analyzed in dynamical one and two - points mean - field approximations and no undamped oscillatory behavior is found . in sec .
[ sec : mc ] , extensive monte - carlo simulations are performed .
it is shown that , as a function of the values of the control parameters , two types of continuous nonequilibrium phase transitions towards a prey absorbing state are present .
the system size dependence of the amplitude of the oscillations is studied and several scaling relations between the amplitude of the oscillations and the correlation length are obtained . in sec .
[ sec : discuss ] an underlying mechanism responsible for the spatial oscillations is proposed , which leads to a qualitative explanation of the properties of the phase diagram . in particular , we show that the spatially extended aspect of the problem is crucial to have an oscillatory region . finally , conclusions are drawn in sec .
[ sec : conc ] .
our system models preys and predators living together in a two dimensional territory .
this territory is divided into square cells , and each of them can contain several preys and predators . in this coarse - grained description in which each cell represents a rather large territory , one can assume that each cell containing some predator will also contain some preys .
hence , a three state representation is made .
each cell of the two - dimensional square lattice ( of size @xmath0 , with periodic boundary condition ) , labeled by the index @xmath1 , can be at time @xmath2 , in one of the three following states : @xmath3 . a cell in state 0
, 1 or 2 corresponds respectively to a cell which is empty , occupied by preys or simultaneously occupied by preys and predators .
the dynamics of the system is defined as a continuous time markov process .
the transition rates for site @xmath1 are * @xmath4 at rate @xmath5 , * @xmath6 at rate @xmath7 , * @xmath8 at rate 1 , where @xmath9 denotes the number of nearest neighbor sites of @xmath1 which are in the state @xmath10 .
4 is the coordination number of this two dimensional square lattice .
the first two processes model the spreading of preys and predators .
the two control parameters , @xmath11 and @xmath12 , characterize a particular prey - predator system .
the reason for considering the sum , @xmath13 , in the first rule is simply that all the neighboring cells of @xmath1 containing some prey , ( hence @xmath14 or @xmath15 ) , will contribute to the prey repopulation of cell @xmath1 .
the third process represents the local depopulation of a cell due to too greedy predators
. it can be interpreted as the local extinction of the species or as the moving of them to neighboring occupied sites .
spontaneous disappearance of a prey state ( @xmath16 : @xmath17 ) or that of the predators alone ( @xmath16 : @xmath18 ) is forbidden .
these assumptions are reasonable because the occurrence of these processes is improbable .
the rate of the third process is chosen to be 1 , which sets the time scale . as a consequence , @xmath11 and @xmath12 are dimensionless quantities .
the above dynamical rules are an extension of the contact process model ( cp ) @xcite introduced as a description of epidemic spreading .
the cp is a 2-state model , @xmath19 ; the status 0 and 1 represent respectively the healthy and the infected individuals .
the cp dynamical rules are * @xmath4 at rate @xmath20 , * @xmath17 at rate 1 .
an epidemic survives for @xmath21 = @xmath22 @xcite and disappears for @xmath23 .
the transition towards this absorbing state is of second order and belongs to the directed percolation ( dp ) universality class @xcite .
our model differs from most of the lattice models previously investigated by the fact that on each site , each species may be represented by several individuals rather than just one . in the previously investigated models the spreading rate of the preys is simply proportional to @xmath24 . under this assumption ,
our model reduces essentially to the st model , in which the control parameters are defined as @xmath25 and @xmath26 .
it is worth discussing first the behavior of our model in two limiting cases . in the @xmath27
limit the proportion of empty cells is negligible since the empty cells are reoccupied by preys instantly after their extinction .
hence , the lattice is completely covered by preys and the @xmath28 sites behave as the infected species in the cp .
namely , when decreasing @xmath12 the predator density is decreasing continuously and vanishes at the cp critical value @xmath29 = @xmath30 .
one can think of the @xmath31 limit in similar terms . in this case
, the proportion of the prey cells ( @xmath32 ) should be negligible since the high productivity of the predators , while the prey - predator cells should behave as the infected species in the cp .
this is indeed the case if @xmath33 , but when @xmath11 gets smaller than @xmath30 , the prey density increases again instead of being zero , as we shall see later .
although apparently simple , there is no way to solve analytically the model defined above .
however , analytic solutions can be obtained by making some approximations .
the simplest one is the one - point mean - field approximation in which all spatial fluctuations are neglected .
thus , the system is characterized by the densities of prey , @xmath34 , and predator , @xmath35 , sites a=1l^2 _ i ( _ _ i , 1 + _ _ i , 2 ) , b=1l^2 _ i _ _ i , 2 , which values satisfy the @xmath36 conditions by definition . in terms of these densities ,
the mean - field dynamical equations read : = _ a a(1-a)-b [ mfa ] and = _
b b(a - b)-b [ mfb ] note , that for @xmath37 ( @xmath38 ) initial condition the predator ( and prey ) densities remains 0 .
the ( [ mfa],[mfb ] ) equations clearly differ from the usual l - v ones .
the main difference lies in the interaction terms as , although a larger prey density increases the predator growth rate , the rate of the predated preys only depends on the predator density .
this is a simple consequence of the fact that there are no pure predator sites without preys in this model .
thinking of a real prey - predator system it makes sense , as a predator has to consume a certain amount of preys in a given time to survive , independently of the number of preys around it .
the @xmath39 term in the first equation plays the role of a simple verhlaust factor which assures an upper limit for the prey density @xmath40 , and similarly the @xmath41 term in the second equation do not let the density of predators exceed that of the preys .
the stationary states are obtained by setting the left hand sides of eqs .
( [ mfa ] , [ mfb ] ) to zero .
contrary to the simplest l - v equations , qualitatively different stationary states are obtained varying the parameters , @xmath11 and @xmath12 , as illustrated on fig .
[ fig : phase_mf ] . for @xmath42 and @xmath43 ,
the stationary state is a pure prey absorbing state @xmath44 . for @xmath45 the stationary state
is also a prey state , @xmath46 , however , the value of @xmath47 depends upon the initial state . in the rest of the plane @xmath48 ,
the stationary solution is a^s= and b^s = a^s- , which describes a coexistence of preys and predators ( coexistence phase ) .
for @xmath49 the @xmath34 and @xmath35 densities are approximately the same a^s = b^s + o ( ) = \ { 1 - + o ( ) & _ a > 1 + o ( ) & _ a = 1 + o ( ) & _ a < 1 + . and as a function of @xmath11 they show a mean field cp behavior as it is expected from the argument given in sec . [ sec : model ] . in the @xmath50 limit ( and for @xmath51 )
the system is full of preys
, namely a^s = 1 - ( 1- ) + o ( ) [ aapp ] and the predator density reads b^s = ( 1-)(1- ) + o ( ) [ bapp ] and , as expected , its @xmath12 dependence agrees with the prediction of the mean - field approximation for cp .
this approximation predicts a second order phase transition along the whole @xmath52 line , as in the @xmath53 limit @xmath34 and @xmath35 approach linearly the values @xmath54 and @xmath55 respectively a^s & = & 1 - + o((_b-1)^2)b^s & = & _ a + o((_b-1)^2 ) .
the behavior of the densities is rather surprising at the @xmath45 boundary of the coexistence phase . for @xmath56 and for @xmath51 a^s = + _ a ( + ) + o(_a^2 ) , [ a : la<<1 ] while the stationary solution , @xmath47 , for @xmath57 depends on the initial state .
thus the mean field approximation predicts a discontinuity of the prey density along this boundary .
however , the density @xmath58 is proportional to @xmath11 and continuous in @xmath57 .
important quantities are the fluctuations of the prey and the predator densities ( mean square deviations ) , which are normalized to be size independent for large systems _ = l^2 ( -)^2 , = a b , [ defchi ] and @xmath59 means the time average in the stationary state . for @xmath60
the majority of the sites are in state @xmath15 , with a few holes in it , hence one can suppose that the holes are independent .
consequently , the number of the holes follows a poisson distribution , from which the average hole number equals to the mean square deviation .
there are @xmath61 holes made of sites in the state @xmath62 and @xmath63 holes made of sites in the states @xmath62 or @xmath54 . thus _ a 1-a^s _ b 1-b^s , [ appchi ] which is in good agreement with the simulations in a region ( non - oscillatory part ) of the coexistence phase ( see fig . [
fig : chi ] ) .
the stability of the stationary state can be analyzed by linear stability .
one has to investigate the eigenvalues , @xmath64 , of the jacobian matrix related to the mean field equations ( [ mfa ] , [ mfb ] ) at the stationary densities .
( _ a a & _ b a + _ a b & _ b b + ) |_s= ( _ a(1 - 2a^s ) & -1 + _ b a^s -1 & 1 - _ b a^s ) .
it turns out that the real parts of the eigenvalues are always negative , assuring the stability of the solutions .
this mean field approximation do not predict limit cycles , which would correspond to having an eigenvalue , @xmath65 , with a zero real part .
however , in some part of the coexistence phase the imaginary part is nonzero , so the stationary solution is approached in spirals ( poles ) , instead of straight lines ( nodes ) ( see fig . [
fig : phase_mf ] and [ fig : flow ] ) , as it was also observed in the st model @xcite .
note , that an unexpected node region appears for @xmath66 .
one can consider the presence of poles as a hint for the appearance of oscillations beyond the mean - field approximation .
notice , that in this pole case , the damped oscillations are strong along the @xmath12 axes ( i.e. for @xmath67 )
. the strength of them can be characterized by the ratio of the imaginary and real part of the eigenvalues , which has a singularity in the @xmath68 limit .
using ( [ a : la<<1 ] ) , we obtain || = 4 _ a^-1/2 + o ( _ a^1/2 ) for @xmath67 and @xmath51 . in this limit one can derive an expression also for the frequency , @xmath69 , of the damped oscillations = | ( ) | = 2 _ a^1/2 + o ( _ a^3/2 ) . the mean field results can be interpreted in the following way .
the approximation predicts two distinct phases : the pure prey phase and the coexistence one .
it also gives some hints for a possible presence of oscillations in some parts of the coexistence phase .
the phase boundaries of the two phases are described by two lines : the @xmath52 and the @xmath45 .
several quantities show a power - law behavior close to these boundaries , like @xmath35 and @xmath70 at the @xmath52 boundary , and @xmath35 , @xmath69 and the strength of the damped oscillations at the @xmath45 boundary .
this implies that the transitions are of second order , and the predator density , @xmath35 , seems to be a good candidate for the order parameter .
the order parameter goes to zero at the phase boundaries as @xmath71 and @xmath72 with a mean - field exponent @xmath73 .
we performed also a pair approximation , in which the nearest neighbor correlations are also considered as parameters .
it turns out that the results differ only quantitatively from that of the one point approximation .
contrary to @xcite , our system does not show limit cycle behavior on the pair approximation level either .
on general grounds , one expects that the fluctuations will play an important role in low dimensions .
our model is supposed to describe a two dimensional world and accordingly , we have performed extensive monte - carlo ( mc ) simulations for systems of sizes @xmath0 , @xmath74 varying between 100 and 1000 .
we used periodic boundary conditions .
although our model is formulated as a continuous - time process , an equivalent ( at least for not very short times ) discrete time formulation is more suitable for numerical simulations .
in one elementary time step one lattice site is chosen randomly and its state evolves according to the rules defined is sec .
[ sec : model ] using rescaled rates ( all less than 1 ) as transition probabilities .
one mc step is defined as the time needed such that all the sites have been , on the average , visited once . in this paper
we always use the original time units defined by the model , which can be obtained simply by rescaling the time measured in mc steps . for sufficiently large system , the stationary state does not depend on the initial conditions .
usually we filled up the lattice completely with preys as an initial state and put a few predators on it . to obtain the stationary phase diagram and the stationary values of the quantities of interest the number of mc steps performed varied in the range @xmath75 to @xmath76 for systems of linear size @xmath77 to 1000 respectively .
the corresponding phase diagram is depicted on fig .
[ fig : phase ] .
two different phases are present as a function of the two control parameters @xmath11 and @xmath12 : a pure prey phase , a prey and predator coexistence phase with an oscillatory and a non - oscillatory region . in the oscillatory region ,
oscillations with a well defined frequency were observed in the prey and the predator densities ( see fig . [
fig : dens_time ] ) .
although theoretically possible , we never observed an empty lattice absorbing state .
the reason for that is simply that even one surviving prey fill up the system with preys in the absence of predators . as fig .
[ fig : phase_t ] shows , the locations of the different regions of the phase space differ essentially from those obtained for the st model .
the phase boundaries of the prey phase ( see fig . [
fig : phase ] ) were obtained in the following way .
simulations were started at parameter values for which the coexistence is maintained practically forever ( up to the maximal number of mc steps investigated ) , and we decreased one of the parameter values by @xmath78 . if the predators were still alive after a given time , @xmath79 , we decreased the parameter further .
the extinction of the predators defines the phase boundary .
@xmath78 was chosen to be in the range 0.005 to 0.04 , while @xmath80 mc steps .
the result was very similar with @xmath81 and @xmath82 mc steps .
the definition of the boundary between the oscillatory and the non - oscillatory region of the coexistence phase will be described later . on fig .
[ fig : phase ] , the boundary of the prey phase is displayed for different system sizes ( @xmath83 ) . apparently , in the @xmath84 regime the size dependence is negligible , but relevant for @xmath85 .
note that this strong size dependence of the boundary coincides with the presence of oscillations .
decreasing @xmath12 at any fixed value of @xmath11 , a second order phase transition takes place between the coexistence and the prey absorbing phases along a transition line @xmath86 . as for the mean field case ,
the predator density is considered to be the order parameter . as @xmath87
, the order parameter , @xmath35 , and @xmath70 go to zero as b ~1-a ~(_b - _
b^*(_a))^_1 .
[ dp_op ] as seen on fig .
[ fig : phase ] , the values of @xmath86 obtained by fitting the data with eq .
[ dp_op ] are in very good agreement with the phase boundary obtained previously .
fitting the data leads @xmath88 ( with satisfactory precision for @xmath89 ; see fig .
[ fig : dens_dp ] ) . in the same limit
the fluctuations of the predator density also follow a power law behavior , @xmath90 .
the exponent has been determined to a good precision as @xmath91 for several values of @xmath11 between 1 and 50 .
the same behavior has been obtained ( only for @xmath92 and 3 ) for the prey fluctuations , @xmath93 , with the exponent @xmath94 .
the critical behavior seems to be the same when the transition line is crossed while decreasing @xmath11 at fixed values of @xmath12 .
the two exponents , @xmath95 and @xmath96 , are compatible with those obtained for dp in @xmath97 dimension @xcite .
thus we conclude , that this absorbing state phase transition belongs to the dp universality class , as expected on general grounds @xcite .
note that dp type phase transition in a similar lattice prey - predator system has already been observed in 1 dimension @xcite . for @xmath68
, the transition line , @xmath86 , ends in a special point , ( @xmath98 ) , where all the three phases meet . for @xmath99 ,
the mc results for the finite size dependent phase boundary suggest us that the transition happens at @xmath100 in the @xmath101 limit . approaching this transition line , @xmath68 ,
the prey density , @xmath34 , does not go to 1 but to a finite value depending on @xmath12 ( see fig .
[ fig : dens_a ] ) .
however , according to the results depicted on fig .
[ fig : dens_b ] , the predator density , @xmath35 , always goes to zero in this limit as a power of @xmath11 with an exponent @xmath102 .
surprisingly , this second order phase transition to the prey absorbing phase does not belong to the dp universality class .
the presence of power law behavior , however , confirms that for an infinite system the transition occurs at @xmath100 for @xmath103 .
this means that for this range of @xmath12 , and for any arbitrary small @xmath11 , the coexistence of the species is possible providing that the system is large enough .
for @xmath99 the fluctuations of the two densities , @xmath93 and @xmath104 , behaves similarly .
for a given @xmath12 , there is a clear crossover at @xmath105 from a mean field like behavior to a regime where the correlations are more important . for @xmath106
the behavior of @xmath93 and @xmath104 agrees with that predicted by mean - field theory , reflecting the fact that in this range of @xmath11 the dominant behavior comes from the noise . as the @xmath107 condition coincides with the presence of oscillations , the crossover point , @xmath105 ,
is taken as the definition of the border between the oscillatory and the non - oscillatory region .
after a proper normalization , the relative fluctuations collapse on a single curve for @xmath108 ( see fig . [
fig : chi ] ) .
namely , k_1(_b ) , where the numerical factor , @xmath109 , depends only on @xmath12 . however , the precision of the simulation results was not satisfactory enough to obtain the functional form of @xmath109 ( and of the forthcoming @xmath110 for @xmath111 , 3 and 4 either ) .
the simulation showed that @xmath112 ( @xmath113 or @xmath35 ) is size independent as it was expected from its definition ( [ defchi ] ) .
as a consequence , the deviation from the average density , @xmath114 @xcite , is smaller for larger systems and evidently scales with @xmath115 .
certainly , this deviation increases with the intensity of the oscillations .
the above finite size behavior is in agreement with the results of earlier simulations which claimed that the oscillations in the global densities disappear with increasing system size @xcite .
our simulations predict more pronounced oscillations for smaller @xmath11 and for larger @xmath12 .
the oscillations have to show up also in the correlation functions , c_a(i , ) & = & ( 1 - _ _
j(t ) , 0 ) ( 1 - _ _
j+i(t+ ) , 0 ) , c_b(i , ) & = & _ _
j(t ) , 2 _ _
j+i(t+ ) , 2 , where @xmath116 labels a lattice site distant of @xmath1 lattice spacing from the site @xmath117 . @xmath118 ( @xmath113 or @xmath35 ) depends only on @xmath1 and @xmath119 due to the homogeneity of the system in space and time . for @xmath120 the correlation function , @xmath121 , obtained numerically
could be fitted by an exponential @xmath122 . in the oscillatory region
the correlation lengths of preys and predators differ only through a @xmath12 dependent factor , @xmath123 , and they turned out to be proportional to the fluctuations of the prey density , @xmath124 .
it means that a more correlated system displays stronger oscillations .
the reason for that is simply that the dynamics of the different sites shows some synchronization within a correlation length , which results in larger oscillations ( see sec .
[ sec : discuss ] for more details ) . in order to determine the characteristic frequency , @xmath125 , and the amplitude , @xmath126 ( @xmath113 or @xmath35 ) , of the oscillations
, we measured the fourier spectrum of the time dependent densities s _ ( ) = _ t 1 t | _ t=1^t ( t ) ( it ) |^2 .
[ defps ] the presence of oscillations is reflected as a peak at nonzero frequency in the fourier spectrum . extracting this peak from a background noise ,
enable us to define @xmath127 and @xmath128 as the zeroth and the first momentum of this distribution .
this analysis shows clearly that the frequency of the oscillations is independent of the system size ( see fig .
[ fig : freq ] ) , and is the same for preys and predators .
moreover , for a wide range of the parameters in the oscillatory phase the frequency , @xmath129 , is well approximated by @xmath130 .
this linear behavior differs from the mean field prediction . in the oscillatory region
the oscillations are present for arbitrary large systems , however , their amplitude decreases with increasing system size , as @xmath131 . at this point
it is important to emphasize that this fact does not imply that only small oscillations are present in large systems .
indeed , for a large system the amplitude of the oscillations can be made larger by decreasing @xmath11 .
on the other hand , when increasing @xmath11 the amplitude goes to zero as a power law which makes difficult to define a phase boundary for the oscillations in this way . however , there is a simple scaling relation between the amplitude and the correlation length for the preys in the oscillatory region _
a^2 2 _ a l^2 a_a , [ xca ] as it can be observed on fig .
[ fig : scale ] .
the analogous expression for the predators is slightly more complicated _
b^2 ( ) ^2 k_3(_b)_b k_4(_a ) l^2 a_b , [ xcb ] with appropriate @xmath132 and @xmath133 values .
another quantity which characterizes the oscillations is the time dependent local correlations , @xmath134 .
a similar investigation was made in @xcite with time dependent correlations of the average local densities . in the oscillatory region
@xmath135 displays damped , size independent oscillations . more precisely , the time correlations are size independent for any @xmath136 , while for any @xmath137 the system evolves to the prey absorbing state .
clearly , this critical system size is proportional to the correlation length , @xmath138 .
the size independence of @xmath139 is a simple consequence of the fact that areas which are further than @xmath140 apart are uncorrelated .
the investigation of the time dependent correlations , however , provides a rather ambiguous way to define the boundary of the oscillatory region .
indeed , one can observe local oscillations everywhere in the coexistence phase simply because , due to the cyclic dominance nature of the model , each site has to evolve in a loop ( @xmath141 ) .
thus , according to the value of the damping factor , it is somehow arbitrary to decide if the state is oscillatory or not .
it is worth noting , that at some particular values of @xmath12 ( @xmath142 or 20 ) and for small @xmath11 values ( @xmath143 or 0.4 respectively ) , where the correlation length is comparable to the system size ( @xmath144 ) , the system can evolve to a stripe like state . in this state
3 stripes of size @xmath74 , made of predator , prey and empty cells , are drifting through the system .
however , for given @xmath11 and @xmath12 values , this behavior disappears when increasing the size of the system .
the comparison of the mc results with the mean - field prediction shows that the later gives a qualitatively correct description of the phase diagram ( see fig .
[ fig : phase_mf ] ) , as well as of the discontinuity in the prey density , @xmath34 , along the @xmath45 boundary .
a qualitative understanding of the phase diagram is possible .
if the birth rates are much larger than the death rate ( @xmath145 and @xmath146 ) the system is full of preys and predators ( @xmath147 ) , while for small values of @xmath12 the system evidently reaches the pure prey absorbing state . as already discussed in sec .
[ sec : model ] , in the @xmath27 the system is full of preys ( @xmath148 ) and the predators behave like the infected species in the cp .
it means that they could survive only for @xmath149 , where a dp like second order transition occurs .
this is in agreement with the mean field results and with the simulation for @xmath150 ( see fig . [
fig : dens_dp ] ) .
one can also derive an approximate formula for the position of the phase boundary between the non - oscillatory and the prey phase @xmath86 .
for @xmath151 , the system is almost full of preys ( @xmath152 ) and , in some sense , the dynamics of the predators is close to that of the cp .
the predators die at rate 1 and spread at rate @xmath12 , but they can not enter into the empty sites .
one can introduce an effective @xmath153 and describe the process as a cp , namely , the predators can enter any neighboring site at this rate . as the number of empty sites is proportional to leading order to @xmath154 , the effective parameter should be @xmath155 , where @xmath156 is a fitting parameter .
as this cp displays a phase transition at @xmath157 , in terms of the original parameter the transition occurs at @xmath158 .
this conjecture is in excellent agreement with the simulation data for @xmath159 with @xmath160 ( see fig . [
fig : phase ] ) .
for @xmath146 the new prey sites are usually immediately occupied by predators as well . however , with a small but finite probability , a predator site can disappear before the predators spread to the new born prey site , and in this way , a prey site can be left alone and grow ( similarly to the eden model @xcite ) .
this rare event is negligible when the predator density is large enough and a prey island can not grow for long periods of time .
practically this is the case for @xmath33 . in this case , the number of prey sites is negligible small , and the predator sites behave as the infected species in the cp .
one can see on the fig .
[ fig : dens_a ] and [ fig : dens_b ] , that for @xmath161 the two densities ( @xmath162 ) are equal to that of the cp if @xmath33
. however , in the vicinity of @xmath30 the densities are low , which allows an isolated prey island to grow for a long time .
if @xmath163 the predator islands are shrinking and , if @xmath11 and @xmath74 are not too small , they can survive until a growing prey island reaches one of them .
at this moment , the predators invade very quickly the prey territory and increase their population size ( see fig . [
fig : conf100 ] ) .
these new predator sites start to die out leaving a few prey sites alone , and the whole procedure starts again .
this mechanism insures the survival of the predators much bellow the cp critical density and results in oscillations in the population sizes . for @xmath103 , but not too large ,
the qualitative picture is slightly different . as one can observe on fig .
[ fig : conf5 ] , groups of predators are wandering through the system towards prey - dense areas .
if two fronts of predators meet they usually stop moving and the local population of predators starts shrinking .
the oscillations are maintained in a somewhat similar way than for the @xmath146 case : these predators can only survive if the preys become dense around them .
this is more probable for larger values of @xmath11 , and it is also clear that the predators have a better chance to survive in larger systems . according to the above statements , the key point in the underlying mechanism of oscillations is the existence of blocked predator islands which are located and trapped in sparse prey areas .
indeed , blocked predators in sparse prey areas result in growing prey populations ; however , the resulting dense prey population allows predators to move and predate again .
this mechanism drives back the system to the beginning of the loop .
clearly , predators can only be trapped in sparse prey areas if @xmath11 is smaller or of the order of the death rate , 1 .
this explains the location of the oscillatory region .
note , that the above argument is based on the spatial nature of the system , suggesting that the spatially extended character is fundamental for the existence of such prey - predator type of oscillations .
this mechanism also provides a qualitative understanding of the key properties of the system .
the trapped predators can invade the prey area only when the preys are dense enough again , which takes a time proportional to @xmath154 , and leads to @xmath164 . according to the simulations
the correlation length , @xmath140 , increases with decreasing @xmath11 .
indeed , as @xmath11 decreases , the trapped predators have to wait longer to escape , hence fewer groups of predators survive .
this increases the distance between the groups , resulting in larger prey islands , which average size is proportional to @xmath165 .
when the correlation length is of the order of the system size , there are islands of preys of typical size l , extruding the predators out of the system .
hence , the condition @xmath166 characterizes the phase boundary between the oscillatory and the prey phase . on the other hand , a correlation length of order one ( @xmath167 ) ,
means that the noise dominates the system .
thus , @xmath167 characterizes the boundary between the oscillatory and the non - oscillatory region of the coexistence phase . as shown by the study of the time dependent correlations , domains separated by a distance larger than @xmath165
oscillate asynchronously around a constant value with the same frequency , @xmath168 . according to this picture
, one can derive a more quantitative description for the oscillatory region .
let us assume that for @xmath169 , the global densities of each species can be written as the sum of local coarse - grained densities at a typical length scale @xmath165 .
moreover , we assume that all these local densities oscillate with the same frequency but a different phase , @xmath170 .
in general , the amplitude of the local oscillations should depend on the parameters @xmath11 and @xmath12 .
however , as one can observe on fig .
[ fig : conf100 ] and [ fig : conf5 ] , the predators can only enter an almost fully dense prey area and the predator fronts leave an almost empty field behind them .
hence , as suggested by the numerical simulations , everywhere in the oscillatory region , the local amplitude for the prey density can be considered as a constant , @xmath171 .
= a^s + d ( ) ^2 _ l=1^()^2 ( t + _ l ) .
supposing that the @xmath170 values change much more slowly than @xmath69 , @xmath172 shows a simple @xmath173 behavior for long periods of time ( see fig .
[ fig : dens_time ] ) .
thus , for @xmath172 one can derive the value of the density fluctuations , @xmath93 , and the amplitude of these oscillations , @xmath174 , using ( [ defchi ] ) and ( [ defps ] ) , and take the average over all the possible @xmath170 configuration taken from a flat distribution .
this procedure reproduces the result of eq .
( [ xca ] ) up to a multiplicative factor in front of the correlation length .
we have studied a two dimensional prey - predator model , ( size @xmath0 ) , which exhibits a rich stationary state phase diagram .
a particular attention has been payed to the study of finite size effects , and we were able to draw clear cut conclusions concerning the behavior of the model both for @xmath74 finite as well as for the limit @xmath175 .
three kinds of stationary states can be reached according to the values of the control parameters : a pure prey state , and two coexisting prey - predator ones with and without oscillations .
two different kinds of second order transitions were found when going into the prey absorbing phase .
the transition between oscillatory and non - oscillatory coexistence phase is , in fact , a cross - over between two asymptotic regimes characterized by a very small and a large correlation length respectively . in the oscillatory regime
, scaling relations were established between several physical quantities .
a qualitative explanation for the existence of such oscillatory regime is given , pointing out the crucial role of the spatial extension of the system .
indeed , the frequency of the oscillations is determined locally due to the dynamics related to blocked predator islands in sparse prey areas .
regions of linear size @xmath165 oscillate with the same frequency but with different phases .
this explains the decreasing amplitude of oscillations with increasing system size . on the other hand , slowly changing phases result periodic oscillations in the overall prey density for long periods of time .
moreover , for suitable choices of the control parameters one can have synchronized oscillations with finite amplitude across arbitrary large systems .
thus we think that our simple model could offer a qualitative explanation for the behavior of the lynx population problem described in sec .
[ sec : intro ] . | a coarse grained description of a two - dimensional prey - predator system is given in terms of a 3-state lattice model containing two control parameters : the spreading rates of preys and predators .
the properties of the model are investigated by dynamical mean - field approximations and extensive numerical simulations .
it is shown that the stationary state phase diagram is divided into two phases : a pure prey phase and a coexistence phase of preys and predators in which temporal and spatial oscillations can be present .
the different type of phase transitions occuring at the boundary of the prey absorbing phase , as well as the crossover phenomena occuring between the oscillatory and non - oscillatory domains of the coexistence phase are studied .
the importance of finite size effects are discussed and scaling relations between different quantities are established .
finally , physical arguments , based on the spatial structure of the model , are given to explain the underlying mechanism leading to oscillations .
= 10000 2 |
simulation of hadron production at low energies ( @xmath1 gev ) is relevant for various physical problems . as the cross - section in this region
is considered to be saturated with intermediate resonances , it is possible to study the properties of these resonances through mass , momentum and angular distributions .
simulation of different channels allows one to better account for selection rules and interference effects .
in particular , the problem of measuring r ( total cross - section of @xmath0 annihilation into hadrons ) is a crucial information for contemporary high energy physics , especially for the @xmath2 problem @xcite . at low energies this value
can only be obtained from an experiment .
simulation allows to calculate the ratios between different channels and estimate the accuracy of measurements
. of other related problems we can highlight the studies of @xmath3 at @xmath4 and @xmath5 decays allowing to check the vector current conservation hypothesis in the standard model @xcite , studies of the @xmath6 decay mode as well as @xmath7- and @xmath8-meson decays .
at this moment there is no theory , describing the strong interactions reliably at low energies .
we used a common phenomenological approach , assuming the production of a final state in consequent resonance decays : + @xmath9 only tree - level diagrams including 2- and 3-body decays , are considered ; + @xmath9 there may be several interfering mechanisms ( e.g. @xmath10 ) which may contribute to different final states ( e.g. @xmath11,@xmath12 ) in accordance with charge and strong isospin conservation ; + @xmath9 possible permutations of final particles are taken into account .
+ to simplify adding new matrix elements , we introduced the following requirements : + @xmath9 the tensor structures are chosen according to the field transformation properties ( spin , isospin , parities ) ; + @xmath9 expressions for matrix elements are written in gauge - invariant form with the help of specialized tensor library ( e.g. , @xmath13 ) ; + @xmath9 indices contraction and particles permutations are performed by the software . + allowed charge states and relative phases of permutations are based on the strong isospin part of the matrix element .
scalar parts of propagators may have arbitrary forms , including form - factors and dependence of width on virtuality . in case of the vector intial state ( @xmath14 )
, the absolute value of the transverse part of hadronic current is used as the value of matrix element .
several interfering matrix elements with arbitrary complex relative coefficients may be used .
by now the package has been used to calculate 4 and 6-pion production cross - sections in @xmath0-collisions .
some distributions of simulated data vs. experimental results @xcite are given in fig .
[ 2picomp ] ( parameters not fitted ) .
comparison with analytical calculations given in ref .
@xcite show a reasonable agreement .
+ at the vepp-2000 collider ( the major upgrade of vepp-2 m , binp , novosibirsk ) @xmath15 production will become possible . despite the lack of absolute normalization factor in matrix elements , the ratio of cross - sections of different final charge states with the same production mechanism
can be tested for @xmath15 data . in the channel @xmath16
our simulation ( left ) and experimental data ( right).,title="fig:",width=260 ] in the channel @xmath16 our simulation ( left ) and experimental data ( right).,title="fig:",width=260 ]
the package for simulation of multipion production has been developed .
it includes generators for @xmath17 @xmath18 , @xmath19 , @xmath20 , and @xmath21 production mechanisms
. this tool may be useful for many problems involving production of multipion final states .
we are thankful to o.yu .
dashevskij , i.f .
ginzburg , p.p .
krokovny , a.s .
kuzmin , a.i . milstein and n.i .
root for useful discussions and helpful contributions .
d.a . thanks dfg foundation for support of his participation in pic2003 .
the work is supported by rfbr grants 02 - 02 - 17884-a and 03 - 02 - 06651-mac .
brown et al . , phys .
* 86 * ( 2001 ) 2227 .
eidelman and v.n .
ivanchenko , nucl . phys .
( proc . suppl . )
* 55 c * ( 1999 ) 181 ; r. sobie , z. phys . *
69 c * ( 1995 ) 99
. t. bergfeld et al .
* 79 * ( 1997 ) 2406 ; a.
anastassov et al .
* 86 * ( 2001 ) 4467 .
akhmetshin et al .
( cmd-2 collaboration ) , phys . lett . *
b 466 * ( 1999 ) 392 .
kopylov , the basics of resonance kinematics , nauka , moscow , 1970 . | software package for monte - carlo simulation of @xmath0 exclusive annihilation channels written in the c++ language for linux / solaris platforms has been developed .
it incorporates matrix elements for several mechanisms of multipion production in a model of consequent two and three - body resonance decays .
possible charge states of intermediate and final particles are accounted automatically under the assumption of isospin conservation .
interference effects can be taken into acccount .
package structure allows adding new matrix elements written in a gauge - invariant form .
= 14.5pt |
the detection of contacting interfaces which can slide or open under applied load is a difficult task but one that is relevant in many technologies .
examples include closed cracks or defects , glued bondings , and partially delaminated material interfaces .
ultrasonic methods based on linear wave scattering are efficient for detecting defects and characterizing material elasticity , but are less sensitive to closed cracks and contacting interfaces . using the nonlinear behavior of these defects , nonlinear ultrasonic techniques such as nonlinear resonance @xcite , sub- and higher - harmonic generation @xcite , and frequency - modulation @xcite have been shown to be sensitive to micro - cracks or closed cracks . for an overview of nonlinear acoustics applications see @xcite .
when an ultrasonic wave with large enough amplitude is incident on a contact with a frictional interface ( e.g. closed cracks ) , higher harmonics appear in the frequency spectrum of transmitted and reflected waves .
this effect , called contact acoustic nonlinearity ( can ) @xcite , is of increasing interest for characterization of closed cracks or imperfectly bonded interfaces @xcite .
analysis of model systems has helped to understand the nonlinear interaction of waves and frictional cracks .
in particular , @xcite analysed the partition of energy resulting from a time harmonic obliquely incident plane sh wave reflected and refracted by a frictional contact interface that slips . both longitudinal and shear wave propagation through a rough surface
were investigated in @xcite using an interface contact model based on hertz theory , again for time harmonic incidence .
the partial contact model was subsequently applied to model scattering from surface breaking cracks @xcite , and numerical simulations indicated efficient production of second harmonics .
time domain studies have concentrated on numerical implementations , such as bem formulation of sh slip motion on an arbitrary interface @xcite . using a generalization of this method to include in - plane motion , it was shown in @xcite that the amplitudes of the higher harmonics of the scattered far - fields can be useful in determining both the pre - stress and the frictional coefficient .
measurements of second harmonic generation for normal incidence of longitudinal waves on a contacting interface between aluminum blocks have been reported in @xcite .
these experiments indicate that the amplitude of the second harmonic decreases rapidly with applied normal contact pressure initially , and then falls off in magnitude at a lesser rate .
these findings are in agreement with experimental measurements on contacting adhesive bonds @xcite .
a simple but instructive model was proposed in @xcite for the nonlinear interaction of an sh wave normally incident on a frictional contact interface .
in contrast to the other studies mentioned , this friction model predicts that only odd harmonics are generated for time harmonic incident wave motion .
the objective here is to investigate the effect of friction on shear wave scattering from a interface .
physically the friction leads to nonlinearity due to switching between the sticking and sliding states of the interfacial contact .
analytical and numerical methods are combined to understand the role of energy loss and the generation of higher harmonics and their dependence on the system parameters , such as applied normal load , incident amplitude , and the friction coefficient .
we consider a shear wave incident normally on a flat frictional interface .
the friction behavior of the interface is described by two friction laws : coulomb s law with a constant friction coefficient and a slip - weakening friction law including static / dynamic friction coefficients .
the outline of the paper is as follows .
the analysis in section [ sec2 ] develops the basic model of @xcite to consider both time harmonic and transient pulses .
particular attention is given to estimating the energy lost in frictional sliding , and how it effects the scattered waves .
a closed form expression is obtained for the energy dissipation under time harmonic incidence .
the time domain numerical model is presented in section [ sec3 ] .
we note that the numerical simulation contains more of the actual physical effects encountered in practice , such as finite contact interface and cylindrical wave spreading .
simulation results are discussed in section [ sec4 ] in light of the simpler analytical model , and several conclusions are drawn concerning the relation between the total energy dissipated and the amplitudes of the harmonics in terms of non - dimensional system parameters .
we consider uni - dimensional sh motion in a solid with shear modulus @xmath0 , density @xmath1 , impedance @xmath2 and wave speed @xmath3 . a plane sh wave is normally incident from @xmath4 on the interface @xmath5 , such that the total shear stress @xmath6 comprises the incident and reflected waves above the interface , and a transmitted wave below : @xmath7 instantaneous particle velocity in the @xmath8 direction is @xmath9 and equal to @xmath10 for @xmath4 and @xmath11 for @xmath12 .
shear stress is continuous across the interface , regardless of the state of contact , with stress @xmath13 , where the latter follows from as @xmath14 when the interface is in contact the wave is completely transmitted with zero reflection : @xmath15 , @xmath16 , and there is no energy dissipation .
when slipping occurs , the velocity jump across the interface follows from - as @xmath17 \equiv v(t,0+)- v(t,0- ) = -2 z^{-1}g(t ) .\ ] ] slipping incurs irreversible loss of mechanical energy into heat , which can be quantified in two equivalent ways .
first , the work done per unit interface length by the slippage over a time increment @xmath18 is ( force@xmath19distance =) @xmath20\delta t$ ] .
the total amount of energy dissipated ( per unit interface length ) then follows from eqs . and
as @xmath21 \ , \text{d }
t = -2 z^{-1 } \int g ( t)h(t)\ , \text{d } t .\ ] ] alternatively , the rate of flux of incident energy per unit length ( power flow ) is ( stress@xmath19velocity =) @xmath22 .
the fluxes of the reflected and transmitted waves are @xmath23 and @xmath24 , respectively . the loss in energy ( per unit length and time ) is therefore @xmath25 .
the three waves are assumed to be plane waves , and hence the fluxes are independent of @xmath26 , with @xmath27 for the incident and reflected waves , and @xmath28 for the transmitted wave .
taking @xmath29 and using to express the stresses , the total energy loss per unit length can be written @xmath30 substituting from eq .
shows that the lost wave energy is the same as the dissipated interfacial energy of eq . .
before the incident wave strikes the interface it is assumed to be in a state of constant normal stress @xmath31 ( compression ) . slipping commences at the first instant that @xmath32 depends on the friction model . for simple
_ coulomb friction _
: @xmath33 , where @xmath34 is the frictional coefficient .
slipping continues for as long as @xmath35 and @xmath36 while it occurs . by considering the two possibilities : @xmath37 , @xmath38
, it follows that the reflected wave during slipping is @xmath39 where @xmath40 is the heaviside step function . the transmitted wave amplitude @xmath41 is given by eq . and the dissipated energy is therefore @xmath42 the slip condition is equivalent to the model proposed in @xcite .
it predicts that the reflected and transmitted waves are clipped " versions of the incident pulse at the positive and negative values equal in magnitude to the critical yield stress , as illustrated in figure [ fig2 ] .
the simple expression for the dissipated energy is new and will be central to subsequent developments . as an example , consider the time harmonic incident wave @xmath43 , @xmath44 . during a single period ,
@xmath45 the slip condition @xmath46 prevails for @xmath47 and @xmath48 where @xmath49 , with the same amount of energy lost in each sub - interval .
it follows that the reflected stress function is @xmath50 the energy dissipated in one cycle then follows from eq . as @xmath51,\ ] ] where the nondimensional function @xmath52 is shown in figure [ fig3 ] .
nonlinear effects occur only for values of the non - dimensional stress parameter @xmath53 less than unity . if @xmath54 no energy is dissipated as the wave transmits unaltered .
we will also find the scaling factor @xmath55 critical for interpreting the numerical simulations ( see [ 3.3 ] ) .
the reflected and transmitted stress waves together sum to equal the wave shape of the incident wave .
this means that the the higher harmonic amplitudes of the reflected and transmitted waves equal the content of the incident wave .
if the latter is a single frequency time harmonic wave , then higher harmonic amplitudes of the reflected and transmitted waves are equal in magnitude and of opposite sign .
we consider this case first .
the incident wave is @xmath43 , @xmath44 .
the clipped nature of the reflected wave implies that its fourier series contains only odd harmonics @xcite , i.e. @xmath56 where @xmath53 is defined in .
the first harmonic amplitudes and the magnitudes of the third and fifth harmonics are plotted in figure [ fig3 ] . note that by definition the non - dimensional amplitudes are normalized such that @xmath57 where @xmath58 is the amplitude of the first harmonic of the transmitted wave . where @xmath59 is the amplitude of the incident wave in figure [ fig2]b .
] a narrow band pulse is shown in figure [ fig2]b .
note that the pulse is antisymmetric about the center @xmath60 because it represents the incident shear stress , which is proportional to the velocity .
the incident displacement amplitude is symmetric about @xmath60 .
explicit expressions for the dissipation and for the harmonic content are not available , but are easily found by numerical quadrature .
thus , the dissipation is computed from , and the harmonic amplitudes of the reflected wave are given by analogy with @xmath61 as @xmath62 note that all of these quantities ( @xmath63 , @xmath64 ) are proportional to fourier components of the reflected wave , either at zero frequency for @xmath63 , or finite values for the harmonics .
figure [ fig5 ] shows the dissipation ( normalized by the scaling factor @xmath55 of eq . ) and the magnitudes of the several harmonics . comparing figures [ fig3]a and [ fig5 ]
we note that the harmonics @xmath65 and @xmath66 for the pulse are larger in magnitude as compared with the time harmonic case , whereas the nondimensional dissipation @xmath52 is smaller .
the overall magnitudes are still comparable with those in the time harmonic case because , even though the integral is over all @xmath67 rather than a single period , the majority of the wave clipping and dissipation occurs within one cycle .
note that @xmath68 but it is small in comparison with @xmath65 .
we next consider a more realistic model using numerical simulation .
the system modeled is shown in figure [ fam1 ] .
it is composed of two identical isotropic elastic media p@xmath69 and p@xmath70 separated by an interface governed by unilateral contact with a coulomb friction law .
the material used in the simulations is aluminium with the young modulus , poisson s ratio and density equal to 69gpa , 0.33 and 2700kg / m@xmath71 respectively .
the problem is considered in terms of a material ( lagrangian ) description . in the first step of the simulation a normal contact stress between the two solids is applied by numerically imposing appropriate boundary conditions on the external surfaces .
the lower surface of the body p@xmath70 is fixed ( @xmath72 ) and the upper surface of the body p@xmath69 is shifted by the appropriate amount @xmath73 in the @xmath26 direction ( @xmath74 ) .
the normal contact stress profile is shown in figure [ fam2]a .
the normal contact stress @xmath75 is quasi - constant on a part of interface ( in figure [ fam2]a , @xmath76=-0.3mpa ) . in the second step ,
a non - planar shear wave is generated by imposing a dynamic displacement along the @xmath8 axis on the upper surface of p@xmath69 . during this stage of the simulation
the lower surface of body p@xmath70 is fixed ( @xmath72 ) and the displacement @xmath77 imposed on the upper surface of p@xmath69 is expressed as a product of a weight function @xmath78 and a function of time @xmath79 shown in figure [ fam2]b .
the interface is supposed to remain in contact during the transmission / reflection of the shear wave .
this assumption has been verified for all simulations .
plast2 , an explicit dynamic finite element code in 2d , is used to simulate the behaviour of the system ( p@xmath69 and p@xmath70 ) during frictional contact .
plast2 is designed for large deformations and non - linear material behaviour @xcite .
it uses a forward lagrange multiplier method for the contact between deformable bodies . for this dynamic study ,
the formulation is discretized spatially using the finite element method and discretized temporally by the @xmath80 method .
the contact algorithm uses slave nodes ( situated on the contact surface of p@xmath69 ) and target surfaces ( on the contact surface of p@xmath70 ) described by a four node quadrilateral element with 2@xmath192 gauss quadrature rule .
the elementary target segments are described by two nodes and approximated by bicubic splines @xcite .
the forward lagrange multiplier method is formulated for the equations of motion at time @xmath81 with the displacement condition imposed on the slave node at time @xmath82 : @xmath83 where @xmath84 , @xmath85 and @xmath86 are symmetric and positive definite matrices of mass , rayleigh s proportional damping @xmath87 and stiffness of the system , respectively .
@xmath88 , @xmath89 , @xmath90 are respectively the vectors of nodal displacements , nodal velocities and nodal accelerations .
@xmath91 is the vector of nodal external forces .
the vector @xmath92^t$ ] comprises normal ( n ) and tangential ( t ) forces at contact points @xmath93 .
@xmath94 $ ] is the global matrix of the displacement conditions ensuring non - penetration and the contact law of the bodies in contact .
the equations of motion are discretized in time using an explicit newmark scheme .
the vectors are expressed at each time step using a time scheme of type @xmath80 , @xmath95 : @xmath96 the displacements @xmath97 of the nodes situated on the contact surface ( p@xmath69 and p@xmath70 ) are first computed with @xmath98 equal to zero . for simplicity @xmath80
is fixed at 0.5 and the nodal displacements @xmath99 at time @xmath82 are obtained according to @xmath100 a constraint matrix @xmath101 is formulated for the slave nodes if they have penetrated through a target segment .
calculation of the contact forces @xmath98 and of the correct nodal displacements @xmath102 at time @xmath82 are then performed : @xmath103 equations are solved using the gauss - seidel method .
the contact conditions solved during each internal iteration of this method are expressed as follows : @xmath104 where @xmath105 is the relative tangential velocity for a slave node related to the surface , @xmath75 and @xmath106 are normal and tangential stress respectively , subscript @xmath107 and @xmath108 referring to the normal and tangential vectors defining the contact .
@xmath109 is an apparent friction coefficient .
two friction laws are considered in the simulations : a standard coulomb friction law with a constant friction coefficient ( @xmath109=@xmath110 , see figure [ fam4]a ) and a slip - weakening ( sw ) friction law ( figure [ fam4]b ) for which @xmath109 is defined in @xcite as : @xmath111 where @xmath112 and @xmath113 are static and dynamic friction coefficient , @xmath114 is the slip and @xmath115 the critical slip distance .
once the dynamic regime is reached , the friction corresponds to a classical coulomb friction law with a dynamic friction coefficient @xmath116 .
this physically motivated model of slip weakening friction ( i.e. the decrease of the friction force with slip ) , which was first introduced in the geophysical context @xcite , can be associated to a phenomenon of adhesion caused by long time stationary contact @xcite . for more details on rate and state - variable friction law ( including the slip - weakening one )
see for example @xcite .
figure [ fam5 ] shows the velocity on the @xmath8-axis versus time and @xmath26-position ( depth ) for three different friction coefficients . at time @xmath117
, the shear wave is generated on the upper face of p@xmath69(see fig .
[ fam1 ] ) .
the incident wave then propagates through p@xmath69 and the contact interface gives rise to reflected and transmitted waves .
it can be seen from figure [ fam5 ] that the magnitude of the transmitted wave is increasing with the friction coefficient .
consequently , the magnitude of the reflected wave is decreasing .
points a ( @xmath8=0,@xmath26 = 20 mm ) and b ( @xmath8=0,@xmath26=-5 mm ) are chosen for fast fourier transform ( fft ) processing of the reflected and transmitted waves , respectively .
figure [ fam5]c and [ fam5]d shows the fft of velocity along the @xmath8-axis for the reflected and transmitted wave for @xmath110=0 , @xmath110=0.4 and @xmath110=1 .
nonlinear harmonics should appear in the spectrum as soon as slipping zones occur at the contact interface .
for friction coefficient equal to 0 or 1 , no harmonics are observed in the reflected and transmitted wave spectrum .
for the case @xmath110= 0 , the interface is in a sliding state during transmission of shear waves but the system is perfectly linear under the previously made assumption of no separation .
when the friction coefficient is equal to 1 , the contact law is nonlinear , but the friction coefficient is too large to permit enough sliding such that would produce a noticeable effect on the transmitted / reflected wave spectrum . in other words , during transmission / reflection of the shear wave the contact nodes remain almost always in the sticking state of the coulomb friction law . for friction coefficient equal to 0.4 , the third harmonic ( 1.5mhz ) is observed in the reflected and transmitted wave spectra due to the nonlinearity of the friction law and the occurrence of sliding zones during the transmission / reflection stage .
as expected , the magnitude of the wave decreases along the propagation direction because of diffraction effects . in order to compare the numerical results with the analytical ones presented in
[ sec2 ] , the non - dimensional parameter @xmath118 used in the analytical model ( where @xmath59 denotes the maximal stress magnitude for the plane wave ) needs to be modified .
let @xmath119 correspond to the case in which the wave magnitude is not large enough to initiate slip at the interface . from this definition a critical friction coefficient for slip , @xmath120 ,
is determined numerically , and the modified @xmath121 is defined as follows : @xmath122 the harmonic magnitudes @xmath123 are extracted from the fft of the computed velocity ( figure [ fam5]c and d ) .
energy dissipated during transmission of the shear wave @xmath124 is numerically calculated along the interface from its definition @xmath125 . since the incident wave is non - plane , @xmath124 is not constant along the interface , and accordingly a mean value @xmath63 on the zone of constant normal stress ( figure [ fam2]a ) is chosen for the results .
non - dimensional measures of the harmonic magnitudes and energy dissipated per unit cycle during transmission of the shear wave are defined respectively as @xmath126 where @xmath127 is the magnitude of the first harmonic of the transmitted wave velocity for @xmath119 at the probe point b and the factor @xmath128 is the number of cycles in the incident signal ( see figure [ fam2]b for @xmath79 ) .
figure [ fam7]a displays the evolution of the non - dimensional third and fifth harmonic magnitudes and non - dimensional dissipated energy versus @xmath121 for two different values of initial normal stress : 0.3mpa and 0.6mpa .
it can be seen that the initial normal contact stress has no influence on the non - dimensional third , fifth harmonics and dissipated energy . comparing figures
[ fam7]a and
[ fig5 ] shows a good qualitative agreement between the numerical and analytical results for @xmath129 and @xmath130 .
the fact that the values obtained numerically are about half of the analytical ones is because the sliding is not uniform for a cylindrical shape of wave front used in the numerical model , while the analytical model assumes a plane wave .
the appearance of odd harmonics in the spectrum is due to distortion of the incident wave during the transmission / reflection as the interface partially slides under friction .
the evolution of the odd harmonics , directly related to this distortion , depends on two competing effects .
first , for a fixed value of the sliding distance , more energy is dissipated and the wave is more highly distorted as the friction coefficient increases . on the other hand , since the incident wave is the same for the different simulations , the cumulative sliding distance is actually diminished as the friction coefficient increases , as shown in figure [ fam7]b .
these two antagonist aspects of nonlinearity of a frictional interface explain the existence of a distinguished value of the normal stress that maximizes energy dissipation : @xmath131 0.4 in figure [ fam7]a .
the numerical simulations allow us to explore dependence on parameter values describing the contact interface ( displacement , normal and tangential stresses , sliding distance , ... ) .
figure [ fam7]b shows the tangential stress and displacement of the contact node situated on the interface at @xmath8 = 0 mm versus time .
the displacement corresponds to the tangential displacement of the slave node ( of p@xmath69 ) in a local frame of p@xmath70 whose origin is situated at the initial contact position of this node ( of p@xmath69 ) on the p@xmath70 interface before the arrival of the incident shear wave . as the friction coefficient increases the relative displacement decreases , because the resistance to sliding is increasing . for @xmath121
= 0.1 , the sliding of the slave node is important but the sinusoidal form is weakly affected , whereas for @xmath121 = 0.8 the sliding of the slave node is weak , but the evolution of the displacement of the contact node is strongly distorted from the initial waveform . in this simulation
the numerical model takes into account a slip - weakening friction law that includes @xcite the combined effects of static and dynamic friction , with coefficients @xmath112 and @xmath113 respectively .
a third parameter @xmath115 describes the slope of the decreasing friction coefficient , see figure [ fam4]b . for all simulations ,
static friction coefficient and incident shear wave magnitude remain the same ( @xmath112=0.3 and @xmath59=0.33mpa ) .
the simulations were performed for different values of the initial normal stress @xmath132 ( from 0.15 mpa to 1.2 mpa ) .
note that in these conditions no slip is allowed when @xmath132 is greater than 1.1mpa .
figure [ fam9 ] shows the results for two dynamic friction coefficients in comparison with those for the constant friction law .
figure [ fam9]a confirms that for this model the third harmonic again follows the behavior of the dissipated energy .
maximum values of the third harmonic @xmath133 and dissipated energy per unit surface @xmath63 are decreasing when the dynamic friction coefficient is decreasing .
the maximum of both is obtained for a larger value of @xmath132 .
figure [ fam9]b shows the tangential stress @xmath134 and the tangential slip @xmath135 of the contact node situated at @xmath8 = 0 on the interface versus time .
the interface is sticking because @xmath136 is less than @xmath137 .
when @xmath136 is equal to @xmath137 , the system begin to slide .
@xmath136 remains the same for constant coulomb friction ( @xmath138 ) while , for the sw law , @xmath139 after the slip distance becomes greater than @xmath115 .
hereafter the interface behavior corresponds to a classical coulomb friction law with @xmath140 . as expected , since the dynamic friction coefficient is less than the static one , the relative displacement at the interface is greater than for the constant friction coefficient case ( @xmath138 )
. also , the relative displacement increases when the dynamic friction coefficient decreases .
the resistance to motion due to friction is less important and so the dissipated energy is reduced .
this explains the offset of the position of the maximum dissipated energy @xmath63 ( figure [ fam9]a ) .
the form of the curve @xmath141 is directly related to the interface friction law in effect .
figure [ fam10 ] highlights an important effect of the parameter @xmath115 on the tangential stress and slip at the interface , and consequently on dissipated energy and third harmonic magnitude . for a low value of @xmath115 ( @xmath142 m ) ,
the system behavior is similar to the constant friction law with @xmath143 .
the sliding is delayed due to the static friction coefficient . for low value of @xmath144
this delay has little influence on the interface behavior . on the contrary ,
if @xmath145 0.6mpa , the delay is growing and the interface slip decreases reaching zero for @xmath146=1.1mpa .
this limit value of @xmath132 depends on the parameter @xmath115 namely , it decreases as @xmath115 increases . as a result , for @xmath147 m
, the limit value of @xmath146 is very low and hence the dissipated energy and the third harmonic curves tends to those for constant friction coefficient @xmath138 .
this study presents a combined numerical and analytical analysis of the propagation of a shear wave through a frictional interface .
a simple analytical model is developed that describes the case of a classical coulomb friction law and a plane wave .
the model pinpoints the pivotal non - dimensional parameters of the problem and yields an explicit dependence of the measurable quantities on these parameters .
good agreement between the analytical and numerical results for this case is demonstrated .
the numerical analysis is also applied to a more complex case which involves a non - planar wave and a slip - weakening ( sw ) friction law with static and dynamic coefficients @xmath112 / @xmath113 . for the both cases
, it is shown that the amplitude of the third harmonic is directly related to the friction - induced dissipated energy .
the evolution of the third harmonic as well as of the dissipated energy depends on two aspects : on the cumulative sliding distance during the transmission of the shear wave and on the friction coefficient . when a sw friction law is taken into account ,
the maximum values of third harmonic magnitude and of the dissipated energy are achieved for a larger value of the applied normal stress @xmath132 .
this is explained by the fact that the static effect in the sw law delays the interface sliding in comparison with the case @xmath148 .
less sliding and @xmath149 both imply lower values of dissipated energy and of third harmonic .
the main conclusions of this study are as follows : + ( i ) there exists a specific value of the normal stress that maximizes energy dissipation .
this value results from the competition of two antagonist aspects of a frictional interface : sliding ( cumulative sliding distance ) and resistance to motion ( tangential stress ) .
+ ( ii ) the sw friction law affects the form of the curves for the dissipated energy , @xmath150 , and third harmonic , @xmath151 .
the maximal value is obtained for a larger value of @xmath132 .
each of the three parameters of the sw friction model has a different impact on interface sliding and dissipated energy : the dynamic friction coefficient @xmath113 acts on the steady state sliding , the static friction coefficient @xmath152 delays the sliding in comparison with the constant friction law @xmath148 , and @xmath115 acts directly on this delay . + ( iii ) for the form of shear incident wave used in this study , the third harmonic evolution follows the variation of the dissipated energy at the interface .
thus the measurement of transmitted or reflected wave signal can provide information on dissipated energy at the interface , and the form of the @xmath151 curve can reveal the type of friction law of the interface .
s. biwa , s. yamaji and e. matsumotoa . quantitative evaluation of harmonic generation at contacting interface . in b. o. enflo , c. m. hedberg , and l. kari , editors , _ nonlinear acoustics - fundamentals and applications _ , cp1022 , pages 505508 .
aip , new york , 2008 . | analytical and numerical modelling of the nonlinear interaction of shear wave with a frictional interface is presented .
the system studied is composed of two homogeneous and isotropic elastic solids , brought into frictional contact by remote normal compression . a shear wave , either time harmonic or a narrow band pulse , is incident normal to the interface and propagates through the contact .
two friction laws are considered and their influence on interface behavior is investigated : coulomb s law with a constant friction coefficient and a slip - weakening friction law which involves static and dynamic friction coefficients . the relationship between the nonlinear harmonics and the dissipated energy , and their dependence on the contact dynamics ( friction law , sliding and tangential stress ) and on the normal contact stress
are examined in detail .
the analytical and numerical results indicate universal type laws for the amplitude of the higher harmonics and for the dissipated energy , properly non - dimensionalized in terms of the pre - stress , the friction coefficient and the incident amplitude .
the results suggest that measurements of higher harmonics can be used to quantify friction and dissipation effects of a sliding interface . |
large - scale , large - amplitude disturbances propagating in the solar atmosphere were first recorded by @xcite and @xcite in the chromospheric h@xmath3 spectral line ; therefore called `` moreton waves '' .
typical velocities are in the range @xmath4 km s@xmath1 , and the angular extents are @xmath5 ( e.g. * ? ? ?
* ; * ? ? ?
since there is no chromospheric wave mode which can propagate at such high speeds , moreton waves were interpreted as the intersection line of an expanding , coronal fast - mode shock wave and the chromosphere which is compressed and pushed downward by the increased pressure behind the coronal shockfront @xcite .
they occur in association with major flare / cme events and type ii bursts @xcite , the latter being a direct signature of coronal shock waves .
typically , the first wavefront appears at distances of @xmath6100 mm from the source site and shows a circular curvature .
the fronts are seen in emission in the center and in the blue wing of the h@xmath7 spectral line , whereas in the red wing they appear in absorption .
this behavior was interpreted as a downward motion of the chromospheric plasma with typical velocities of @xmath8 km s@xmath1 @xcite .
sometimes the trailing segment of the wave shows the upward relaxation of the material , i.e. , the chromosphere executes a down - up swing @xcite . in the beginning of the wave propagation ,
the leading edge is rather sharp and intense .
as the disturbance propagates , the perturbation becomes more irregular and diffuse and its profile broadens @xcite .
thus , the perturbation amplitude decreases and the wavefronts
get fainter until they can no longer be traced at @xmath9 mm from the source active region @xcite . a decade ago
, large - scale waves were for the first time directly imaged in the corona by eit ( extreme - ultraviolet imaging telescope ; * ? ? ? * ) aboard the soho ( solar and heliospheric observatory ) spacecraft , so - called `` eit waves '' @xcite .
similarities in the propagation characteristics led to the assumption that in at least a fraction of the events are the eit waves the coronal counterpart of the chromospheric moreton waves ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
basic questions regarding the nature of moreton and eit waves are whether they are caused by the same or by different disturbances , and whether they are initiated by the associated flare or the cme . for recent reviews ,
we refer to @xcite , @xcite , @xcite and @xcite . here , we study the fast and globally propagating moreton wave that occurred in association with the powerful x17.2/4b flare and fast cme event from the noaa ar10486 ( s16@xmath10 , e08@xmath10 ) on october 28 , 2003 . due to its extreme powerfulness and geo - effectiveness ,
diverse aspects of this flare / cme event have been analyzed in a number of studies ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the relationship of the moreton wave to radio observations has already been studied in @xcite . in this paper
, we focus on the kinematical analysis of the moreton wave , its relationship to the flare , the cme , coronal dimmings and type ii radio bursts , in order to get insight into the wave characteristics and its initiating agent .
we analyzed the moreton wave of 2003 october 28 and associated phenomena ( flare , eit wave , coronal dimmings , type ii radio burst ) using the following data sets . 1 . the moreton wave is studied in h@xmath7 filtergrams recorded by the meudon heliograph ( france ) which provides simultaneous observations of the full sun at three different wavelengths in the h@xmath3 spectral lines ( h@xmath3 line center , h@xmath3 @xmath11 0.5 @xmath12 and h@xmath3 @xmath13 0.5 @xmath12 ) with an imaging cadence of @xmath141 min .
2 . the eit wave and the coronal dimmings are studied in full - disk euv images by eit / soho @xcite .
the analysis is conducted in the 195 @xmath12 bandpass which has a time cadence of @xmath15 min .
the coronal dimmings are also analyzed with trace ( transition region and coronal explorer , * ? ? ?
high - resolution 195 @xmath12 filtergrams with a field - of - view ( fov ) of 380@xmath16340@xmath17 around the flare site are available with a time cadence of @xmath18 8 sec .
the associated cme was observed by soho / lasco @xcite .
we use the information from the lasco cme catalog at ` http://cdaw.gsfc.nasa.gov/cme_list/ ` @xcite .
the associated flare is studied in soft x - rays ( sxrs ) by the goes10 satellite and in hard x - rays ( hxrs ) @xmath19150 kev provided by the spectrometer spi onboard integral @xcite .
the associated type ii radio burst is analyzed utilizing the dynamic radio spectrum recorded by the radio spectrographs of the astrophysikalisches institut potsdam ( aip ; * ? ? ?
positions of the radio source are derived from observations of the nanay radioheliograph ( nrh ; * ? ? ?
6 . a full - disk magnetogram recorded by soho / mdi ( at 11:00:03 ut ; pixel size @xmath18 2@xmath17 )
is used to study the magnetic context of the event @xcite .
the moreton wave under study was launched during a powerful flare / cme event which occurred in noaa ar10486 ( s16@xmath10 , e08@xmath10 ) on october 28 , 2003 .
noaa ar10486 had a complex magnetic configuration of @xmath20 and was surrounded by several other large and complex ars ( e.g. ar10484 , ar10488 ; see fig .
[ img : plot_magnetogramar ] ) . the time range between 19-oct-2003 and 4-nov-2003
was characterized by an extremely high level of solar activity during which 12 x - class flares occurred .
on 28-oct-2003 ar10486 produced a x17.2/4b two - ribbon flare .
the goes10 sxr flux showed the flare onset in the 1@xmath138 @xmath12 channel at @xmath1811:01 ut reaching peak at 11:10 ut . the integral hxr observations cover the total flare impulsive phase which lasted roughly 15 min .
the increase of the hxr flux ( @xmath19150 kev ) started at 11:02:00 ut and peaked for the first time at 11:02:40 ut .
the associated fast halo cme had a mean plane of sky velocity @xmath21 2500 km s@xmath1 and its linear back - extrapolated launch time is 11:01 ut ( lasco cme catalogue , @xcite ) .
this goes along with a filament activation in the southeastern quadrant at 11:01 ut .
+ the moreton wave can be identified during the time interval 11:0211:13 ut .
the kinematics of the moreton wave is analyzed applying two different methods .
the first method relies on the visual determination of wavefronts in a series of difference images .
the second one is based on the analysis of intensity profiles , derived along the wave fronts , so - called perturbation profiles .
we subtracted red and blue wing filtergrams recorded simultaneously at h@xmath22 and h@xmath23 with the meudon heliograph .
these images contain information on the plasma line - of - sight velocity ( `` dopplergrams '' ) . in order to maximize the wave contrast , we produced difference images by further subtracting a blue - minus - red pre - event image ( 11:00 ut ) . figure [ img : meudon_wavedifffullsun ] shows this dopplergram sequence .
the moreton wave appears as a bright , arc - like transient disturbance , propagating away from the flare site in different directions , spanning almost over the complete solar disk .
we studied the wave kinematics separately for four different directions , in which the wave could be distinctly followed .
the four sectors 1@xmath133 , 4@xmath136 , 7@xmath138 , 9@xmath1311 ( denoted by hours on the clock - face ) and the visually identified wavefronts are indicated in fig . [ img : meudon_wavedifffullsun ] .
we note that the wavefronts in direction 9@xmath1311 were affected by refraction and reflection at the strong magnetic fields around ar 10488 , which resulted in the wave propagation towards north after 11:06 ut ( denoted as sector 12 in fig .
[ img : meudon_wavedifffullsun ] ) . the wave behavior in sector 12
is not considered in the following analysis .
the moreton wave `` radiant point '' ( source center ) was estimated by applying circular fits to the earliest observed wavefronts for each propagation direction ( see fig .
[ img : waveellorzoominblackred ] ) , whereby the projection effect due to the spherical solar surface was taken into account ( following * ? ? ?
* ; * ? ? ?
* ) . in order to estimate the error of the determined wave center , each of the earliest wavefronts was drawn for six times ( first appearance ) or four times ( second appearance )
the determined wave centers for the four directions are : + @xmath24\approx[-253\arcsec\pm31\arcsec,-307\arcsec\pm6\arcsec]$ ] for direction 78 , + @xmath24\approx[-190\arcsec\pm11\arcsec,-315\arcsec\pm7\arcsec]$ ] for direction 911 , + @xmath24\approx[-51\arcsec\pm16\arcsec,-376\arcsec\pm8\arcsec]$ ] for direction 13 , and + @xmath24\approx[+69\arcsec\pm~9\arcsec,-381\arcsec\pm14\arcsec]$ ] for direction 46 .
+ the wave centers derived for the four propagation directions can be grouped into two significantly different radiant points at the mean coordinates @xmath24\approx[-221\arcsec\pm39\arcsec,-311\arcsec\pm8\arcsec]$ ] and @xmath24\approx[9\arcsec\pm61\arcsec,-379\arcsec\pm11\arcsec]$ ] , on the east and west borders of the source ar ( see fig . [
img : waveellorzoominblackred ] ) .
this suggests that the wave propagation to the east ( 78 and 911 ) and to the west ( 13 and 46 ) can be attributed to two different source sites .
accordingly , the wave kinematics are derived using two different source site coordinates . to derive the wave kinematics , we calculated for each point of the wavefronts its distance from the respective radiant point along great circles on the solar surface , and then averaged over all points of the wavefront . figure [ img : velocityallresult ]
shows the distance - time diagrams for the four different propagation directions .
the wave was first visible at a distance of @xmath1890 mm from the wave s radiant point in direction 1@xmath133 , and can be followed up to distances of 500600 mm in the different directions . in direction 4@xmath136 , the winking of a filament can be observed after the wave passage ( @xmath1811:04 ut ) , which makes the wavefront determination difficult for this phase .
linear and quadratic functions are fitted to the derived kinematical curves of the wave ( see fig . [ img : velocityallresult ] ) , from which we extrapolated the start times , ranging from 10:59:40 to 11:01:40 ut .
the mean velocities derived from the linear fits are 940@xmath2510 km s@xmath1 , 1050@xmath2510 km s@xmath1 , 880@xmath2530 km s@xmath1 , and 1020@xmath2520 km s@xmath1 for the orientations 1@xmath133 , 4@xmath136 , 7@xmath138 , and 9@xmath1311 , respectively .
the quadratic fits indicate wave deceleration for directions 1@xmath133 and 9@xmath1311 with 1420@xmath25360 m s@xmath26 and 2960@xmath25280 m s@xmath26 , respectively . in the second approach ,
the moreton wave propagation was followed by analyzing the perturbation intensity - profiles @xmath27 .
it has to be noted that the perturbation profile analysis can only be carried out if the disturbance is strong enough ( see examples in * ? ? ?
moreton waves are in most cases too faint and irregular to be analyzed by this method .
we also note that there are simplifying assumptions involved in the profile method : it is supposed that the ambient medium is homogeneous and the wavefronts expand circularly .
+ first of all , a specific sector is chosen for which the intensity evolution is calculated .
then the intensity values are averaged over annuli with a chosen angular width and a thickness of 5 mm .
this results in a mean intensity as a function of distance @xmath28 ( measured along the solar chromosphere ) from the wave s radiant point .
the procedure is repeated for each filtergram , until the moreton wave fades away .
since we investigate intensity changes @xmath29 , we use the blue - minus - red wing base - difference images , where the wave contrast is best . the method was carried out for the directions 9@xmath1311 and 4@xmath136 because of their high intensities and their rather constant sector spans .
however , the wavefronts become more and more irregular and show noticeable changes in their shape and propagation direction . thus , with the profile method we can track the wave over shorter distances than with the visual method .
the advantage of this method is that it provides us with more information on the perturbation characteristics .
figure [ img : rb9 - 11rb4 - 6 ] displays the perturbation profiles of the propagating wavefronts . from these profiles ,
the locations of the leading edge @xmath30 , the trailing edge @xmath31 and the peak amplitude @xmath32 are extracted ( see arrows in fig . [
img : rb9 - 11rb4 - 6 ] ) .
the difference @xmath33 gives the width of the propagating disturbance . from the calculated perturbation profiles ,
we derived the wave kinematics . figure [ img : intvelocitylast ] shows the time - distance plots of the moreton wave derived from the profile method , in comparison with the curves derived from the visual method .
the extrapolated start times lie in the range of 11:00:4011:01:50 ut , and correspond well to the starting times derived from the visual method .
for the direction 9@xmath1311 , the mean velocities derived from the linear fits are 1320 km s@xmath1 , 1270 km s@xmath1 and 1230 km s@xmath1 for @xmath30 , @xmath32 and @xmath31 , respectively .
the values for orientation 4@xmath136 are 1140 km s@xmath1 , 1040 km s@xmath1 and 850 km s@xmath1 .
the velocities derived by the profile method differ from those derived from the visual - method . however , this difference is not a result from the different methods , but merely results from the fact that with the profile method the late - phase evolution of the wave , when it has already decelerated , can not be extracted . using an identical time range for the fitting procedure for both methods ( orientation 911 : 11:02 ut
11:06 ut ; orientation 46 : 11:04 ut 11:09 ut ) , we obtain from the visual method mean velocities of 1280 km s@xmath1 and 1070 km s@xmath1 , which is consistent with the velocities obtained from the profile method ( peak amplitude ) for the two sectors ( cf.fig .
[ img : intvelocitylast ] ) .
figure [ img : intvelocitylast ] also indicates that the distance between the leading edge @xmath30 , the peak amplitude @xmath32 and the trailing edge @xmath31 increases with time , which is especially prominent in direction 46 .
the peak amplitude of the perturbation increases in the beginning ( see fig . [
img : rb9 - 11rb4 - 6 ] , panels g and h ) , and after achieving maximum ( panel i ) , subsequently decreases ( panels j to l ) . in direction 911 ,
the wavefront steepens from panel a to c. we emphasize that panel c ( 11:03:14 ut ) shows the situation close to the first appearance of the type ii burst source ( see sect .
[ seq : type2 ] ) , which reveals the formation of the coronal shock .
the dip in the profile ( at the trailing part of the wave ) is a signature of the upward relaxation of the compressed chromospheric plasma .
it is not visible in the beginning but becomes more and more remarkable as the wave evolves .
eit images provide information on large - scale coronal changes associated with flare / cme events .
figure [ img : eit_basediff924blackwave ] shows a base - difference eit 195 @xmath12 image at 11:11:50 ut ( reference image is 9:24:50 ut ) . due to the 12 min time cadence of the eit instrument
, the transient eit wave ( bright diffuse arcs toward north and northwest of ar 10488 ) is observable in only one single frame at 11:11:50 ut in two sectors .
it has already traveled across a significant portion of the solar disk . since the eit wave can only be traced in one single image ,
no kinematics can be derived .
however , the position of the eit wavefront can be compared to the moreton wavefronts observed in h@xmath7 ( fig .
[ img : plot_hxr ] ) .
figure [ img : plot_hxr ] shows a combined plot of the moreton wave kinematics ( direction 13 ; visual method ) , the flare hxr flux measured by integral , the back - extrapolated cme propagation curve ( from the lasco catalog ) and the location of the eit wavefront ( direction 13 ) . extrapolating the 2nd order fit to the moreton wave distance - time diagram in this direction
, we find that the eit wave is slightly ahead of the moreton wave but basically on the same kinematical curve .
the fact , that the eit wavefront is located some 10 mm ahead of the corresponding moreton wave was also reported from other studies ( e.g. * ? ? ?
* ; * ? ? ?
assuming that both are related phenomena , it can be explained by the different heights of the moreton wave ( chromosphere ) and eit wave ( lower corona ) observations and the inclination of the expanding coronal wave dome as envisaged in uchida s theory .
bipolar coronal dimmings are believed to be the footprints of large - scale flux - rope ejections ( e.g. * ? ? ?
* ; * ? ? ?
* ) . in the event under study , coronal dimmings are detectable in eit 195 @xmath12 ( see fig . [
img : eit_basediff924blackwave ] ) as well as trace 195 @xmath12 ( see fig .
[ img : dimming_overview_8panel ] ) .
eit observations are available as full - disk images with a time cadence of @xmath18 12 min .
trace observations are available for a fov of 380@xmath34340@xmath17 around the flare site with a high time cadence ( @xmath18 8 sec ) during 10:14 ut and 11:45 ut . figure [ img : eit_basediff924blackwave ] shows an eit 195 base - difference image .
it displays the global coronal dimmings , which are expanding over the whole hemisphere indicative of a complete restructuring of the corona ( e.g. * ? ? ?
in addition to these global or `` secondary '' dimmings @xcite , localized bipolar dimmings in the near vicinity of the flare site are also observed .
these `` core dimmings '' , as quoted by @xcite are difficult to observe in eit but are well captured with trace . here
, we focus on these small , bipolar dimming regions and their evolution .
figure [ img : dimming_overview_8panel ] shows a sequence of trace 195 @xmath12 base - difference images .
a pre - event image recorded at 10:59:26 ut was subtracted from the whole image series . at 11:00:29 ut
we can observe the first small brightenings along the neutral line in ar10486 .
the first signature of the bipolar dimmings , which are observed as localized dark regions on each side of the flare site , can be detected at 11:01:00 ut ( see fig .
[ img : dimming_overview_8panel ] ) .
they lie on opposite east - west edges of the source ar and probably reflect the two cme expanding flanks .
we also note that they are roughly located on the same axis as the two radiant centers of the wave .
the time of first appearance of these bipolar dimming corresponds to the cme onset , which is determined from the linear back - extrapolation of the cme time - distance curve to be at @xmath1811:01 ut . in the high cadence trace euv images
, we can study the evolution of the bipolar dimmings .
since the western dimming region was only partially covered by the trace fov ( cf.fig .
[ img : dimming_overview_8panel ] ) , we concentrate here only on the eastern dimming region . to obtain information on the area of the bipolar dimming regions , we used specific intensity levels and summed up the enclosed area ( given in percentage of the entire solar disk ; see fig [ img : dimming_evolution_left]a ) .
the expansion velocity of the dimming regions ( presumably related to the cme flanks ) is derived by measuring the propagation of the outer edge of the dimming region , indicated by grey lines in fig .
[ img : dimming_overview_8panel ] .
figure [ img : dimming_evolution_left]b shows the distance - time diagram for the kinematics of the eastern dimming region . due to the distinctive change in its propagation velocity , we applied two separate linear fits to the kinematical curve .
the dimming shows a fast expansion ( @xmath35 km s@xmath1 ) until 11:06 ut , thereafter it drastically slows down by almost an order of magnitude to @xmath36 km s@xmath1 .
we note that fig .
[ img : dimming_evolution_left]a and fig .
[ img : dimming_evolution_left]b show a similar evolution with fast growth up to @xmath1811:06 ut and much slower changes thereafter .
at the time of the fast growth , the cme is in its initial phase . according to @xcite ,
moreton waves appear at the intersection line of an expanding , coronal fast - mode shock and the chromosphere .
another shock signature are type ii radio bursts @xcite , excited by electrons accelerated at the shock front .
type ii bursts usually show the fundamental and harmonic emission band , both frequently being split in two parallel lanes , so - called band - split @xcite .
the interpretation of the band split as the plasma emission from the upstream and downstream shock regions , was affirmed by @xcite .
thus the band - split can be used to obtain an estimate of the density jump at the shock front @xcite . from that , one can infer the pressure jump , which governs the downward compression of the chromosphere after the shock - front passage .
we note that several aspects of the radio observations of this event were already analyzed by @xcite , @xcite and @xcite . the aip dynamic radio spectrum ( fig .
[ img : osra_spect ] ) shows complex and intense radio emission , consisting of a group of type iii bursts , a type ii bursts @xcite , and a complex type iv burst . for a detailed dynamic spectrum covering frequency range 20 - 800 mhz
, we refer to fig . 4 in @xcite . in the present study we focus on the type ii burst that started around 11:03 ut , at a frequency of 420 mhz .
the back - extrapolation of the emission lanes point to 11:02:30 ut , i.e. , close to the first peak of the hxr burst , and the intense type iii bursts which indicate the impulsive phase of the flare .
this is also close to the steepening phase of the moreton wavefront of direction 911 ( see fig .
[ img : rb9 - 11rb4 - 6 ] , panel 3 ) .
nanay radioheliograph measurements at 411 mhz at 11:03:01 ut and 11:03:11 ut show that the type ii burst source was located at @xmath18 @xmath37=@xmath38 . comparing the type ii source location with the moreton wave front
, we find that the position of the radio source is cospatial with the first moreton wave front in the direction 4@xmath136 . in fig .
[ img : intvelocitylast ] the position of the radio source is marked by a thick vertical bar , with the length of the bar corresponding to the measurement error .
the back - extrapolation of the type ii burst to 11:02:30 ut indicates that the disturbance , which developed into a shock , was probably not launched before that time .
on the other hand , we found the back - extrapolated start time of the moreton wave to the western radiant point at 11:01:35 ut ( see fig . [
img : velocityallresult ] ) .
this indicates that the moreton wave was not launched from a point - like source but rather from an extended source . from fig .
[ img : intvelocitylast ] we find that at 11:02:30 ut the back - extrapolation of the moreton wave trajectory was at a distance of @xmath39 6080 mm , which provides us with a lower limit for the source region size .
the upper limit for the source region is given by the first appearance of the moreton wave , which was at @xmath18 200 mm .
due to the complexity and overlapping of various radio emissions it is difficult to analyze in detail the high - frequency part of the type ii burst .
nevertheless , in the period 11:03:10 to 11:03:40 ut , we were able to recognize a band - split pattern , with a relative bandwidth of @xmath400.25 ( see fig .
[ img : osra_spect ] ) .
the relative bandwidth @xmath41 is determined by the density jump at the shock front @xmath42 , where @xmath43 and @xmath44 are the densities upstream and downstream of the shock front ( for details see * ? ? ?
since @xmath45 , we find @xmath461.56 , where the subscript c " stands for corona " .
we found two separate initiation centers for the moreton wavefronts on opposite east - west edges of the source ar 10486 .
this implies that actually two waves were launched simultaneously .
this argues against an initiation by the flare pressure pulse : first , a simultaneous launch of two flare pressure pulses is very unlikely ; second , at the time of the first flare hard x - ray peak ( 11:02:40 ut ) , indicating the first episode of powerful flare energy release , we already observed the moreton wave , this means that if initiated by the flare there would have been no time for the wave to reach a large amplitude to be observable .
nevertheless , we note that an initiation scenario caused by the flare can not be completely excluded ( for instance , assuming that the wave was launched already _ before _ the time of the first peak in the flare energy release ) .
the most likely interpretation of the observations is that the moreton wave is caused by the expanding flanks of the cme , evidenced by the observed bipolar dimmings on east - west edges of the flare site .
this is further strengthened by the observations of two moreton wave radiant points , which lie roughly on the same axis as the bipolar dimmings .
we also note that the initial acceleration of the wave coincides with the fast expanding phase of the dimming evolution with @xmath21 100 km s@xmath1 , which lasts until 11:06 ut ( cf .
[ img : dimming_evolution_left ] ) .
this corresponds also well with the end of the steepening phase ( presumably indicating the shock formation ) of the moreton wavefront profiles ( cf .
[ img : rb9 - 11rb4 - 6 ] ) .
we also emphasize that the moreton wave is not launched from two localized points but from two extended sources .
this result is derived from the type ii burst appearance revealing the shock formation , which is first detectable at @xmath1811:03 ut at a certain distance ( @xmath1860 mm ) from the moreton wave radiant point . at this time
, the perturbation profiles of the moreton wave are steepening ( cf .
[ img : rb9 - 11rb4 - 6 ] ) .
the surface of these extended source sites is located at a distance @xmath47 mm from the moreton wave radiant points .
the first value is determined from the type ii occurrence , while the second value is given by the first moreton wavefront appearance in this direction . a strong relation between the cme lateral expansion and the occurrence of coronal shock waves
was already suggested by radio observations .
@xcite reported about a halo cme event , accompanied by a moreton and an eit wave , in which the appearance of new radio sources were attributed to the arc expansion of the cme magnetic field .
finally , the encounter of the propagating wave with magnetic field structures caused the opening of field lines .
@xcite related the systematic occurrence of radio sources during the early evolution of a cme to the successive destabilization or interaction of loop systems which resulted in large - scale coronal restructering and finally cme formation .
similar conclusions are drawn by @xcite comparing the moreton wave observations of january , 17 2005 and the related type ii burst resulting in an outcoming analytical mhd model .
it can be presumed that at low heights the coronal shock is a quasi - perpendicular fast - mode mhd shock @xcite .
considering jump relations at the perpendicular shock front ( e.g. , * ? ? ?
* ) , the shock magnetosonic mach number @xmath48 ( where @xmath49 is the coronal shock velocity and @xmath50 the magnetosonic speed ) can be expressed as @xmath51 where @xmath52 is the ambient coronal plasma - to - magnetic pressure ratio and @xmath53 the density jump at the shock front . for the specific - heat ratio ( the polytropic index ) we have taken @xmath54 . using @xmath461.56 for the density jump , as derived from the type ii band - split observations ( cf.sect .
3.6 ) , we find @xmath551.42 for @xmath56 , and @xmath571.45 for @xmath58 . the pressure jump at the shock front , @xmath59 , can be written as @xmath60 where @xmath61 represents the sound mach number of the coronal shock , which for @xmath54 reads : @xmath62 using @xmath461.56 , we find @xmath639.34 for @xmath58 , and @xmath642.22 for @xmath56 .
thus , the type ii burst observations indicate that the pressure jump ranges between @xmath182 and 10 .
the pressure jump at the shock front causes an abrupt pressure increase at the top of a given chromospheric element after the shock sweeps over it .
consequently , also a pressure discontinuity forms between the perturbed corona and the yet undisturbed chromosphere .
the appearance of the pressure discontinuity at the top of the chromosphere implies that a downward - traveling shock should be formed . behind the shock front
the plasma is compressed and moves downwards ( including the corona / chromosphere `` interface '' ) . as the shock penetrates deeper , a successively larger portion of the chromosphere is compressed and set into motion , increasing the h@xmath7 blue - minus - red doppler signal at the leading part of the moreton wave .
@xcite obtained from a cme case study that the speed and density of the cme front and flanks , derived from coronagraphic white light images , were consistent with the existence of a shock .
comparison of the observations with mhd simulations confirmed that the observed wave feature is a density enhancement from a fast - mode mhd shock ( for further similar white light observations of coronal shocks see also @xcite ) .
@xcite found from radio imaging observations that the emitting sources of type iii bursts were located at the compression region between the cme flanks and the neighboring open field lines .
the compression region was revealed as a narrow feature in white light , and was interpreted as a coronal shock driven by the cme lateral expansion .
it is occasionally also observed as a type iv - like burst propagating colaterally with the moreton wave @xcite . especially important
are observations of broadening and intensity changes of uv emission lines in front of cmes , associated with type ii bursts , since such measurements provide an insight into the physical state of plasma in the compression region @xcite .
occasionally , the plasma parameters can be inferred also from the soft x - ray observations @xcite .
we analyzed the intense morton wave of october 28 , 2003 together with its associated phenomena : the x17.2/4b flare event , the fast halo cme ( @xmath21 2500 km s@xmath1 ) , the eit wave , the radio type ii burst , and coronal dimmings , in order to study the wave characteristics and its driver .
the main findings of the study and our interpretation are summarized in the following : 1 .
the moreton wave propagates in almost all directions over the whole solar disk , and can be observed for a period as long as 11 min ( during 11:0211:13 ut ) .
these are both unusual properties of a moreton wave .
but we note that a wave with similar characteristics ( global propagation , @xmath21 1100@xmath131200 km s@xmath1 ) was launched from the same ar on 2003 october 29 in assocation with an x10 flare / cme event @xcite .
we studied the wave kinematics in detail for four different propagation directions finding mean velocities in the range 9001100 km s@xmath1 .
typical values of the sound speed and the alfv@xmath65n speed in the solar corona are 180 km s@xmath1 and 300500 km s@xmath1 ( mann et al .
1999 ) , respectively , which corresponds to a magnetosonic mach number of @xmath1823 of the moreton wave under study , if interpreted as a propagating fast - mode shock .
the band - split of the associated type ii burst was determined to 0.150.25 , from which we inferred magnetosonic mach number @xmath66 .
we find two radiant points ( source centers ) for the moreton wave on opposite east - west edges of the source active region ( ar 10486 ) .
this has not been reported before and indicates that either two independent waves were launched simultaneously , or that the wave was caused by the expansion of an extended source with a preferred axis .
the associated coronal wave ( `` eit wave '' ) was observed by the eit instrument in one frame .
the determined eit wavefront lies on the extrapolated kinematical curve of the moreton wave , identifying both the eit wave and the h@xmath7 moreton wave as different signatures of the same disturbance , according to uchida s theory .
4 . in two propagation directions ,
a deceleration of @xmath67 m s@xmath26 and @xmath68 m s@xmath26 was observed in the distance - time diagrams of the h@xmath7 moreton wavefronts . in the initial phase ,
the perturbation profiles show amplitude growth and steepening of the disturbance .
later on , they display a broadening , amplitude decrease , and signatures of chromospheric relaxation appear at the trailing segment of the disturbance .
this indicates that the coronal perturbation is a freely propagating large - amplitude simple - wave .
the width of the disturbance is in the range of 50150 mm .
the extrapolated start time of the wave fits roughly with the extrapolated start time of the cme observed in lasco ( but we note that the uncertainty of the cme start time is of the order of 10 min ) but occurs slightly before ( 2 min ) the first hxr peak of the flare impulsive phase ( cf . fig .
[ img : plot_hxr ] ) . 6 .
the associated bipolar coronal dimmings , which are generally interpreted as footprints of the expanding cme and associated coronal mass depletion , appear on the same opposite east - west edges of ar 10486 as the moreton wave ignition centers do .
the bipolar dimmings show a fast increase during the impulsive phase , followed by a much slower evolution , which indicates a characteristic restricted growth .
the expansion velocity of the dimming region is @xmath18100 km s@xmath1 at the time , where the moreton wave is launched .
the position of the type ii burst source is co - spatial with the moreton wavefront in direction 4@xmath136 .
the back - extrapolated start time for the type ii burst is @xmath1811:02:30 ut , i.e. , somewhat later than the start time derived for the moreton wave with respect to its radiant point .
this suggests that the perturbation was not launched from a point - like source but from an extended source whose surface was at a distance @xmath6960 mm from the back - extrapolated radiant point .
the launch time of the type ii burst , which indicates the shock front formation , and the moreton wave perturbation profile steepening coincide within the measurement errors . from several case studies ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) it was derived that radio as well as white light signatures in coronagraphs provide evidence that the lateral expansion of a cme is able to produce shocks in the corona .
the results of the study , presented in this paper , support a scenario in which the launch of the observed moreton and eit wave is initiated by the lateral expansion of the cme .
in such a case we expect that the perturbation is driven only over a short time / distance , as the cme lateral - expansion range is limited ( for the theoretical background see , e.g. * ? ? ?
* ; * ? ? ?
thus , the disturbance would evolve into a freely propagating wave , which is consistent with the observed perturbation characteristics .
the behavior of cme expansion ( radial and lateral direction ) at low coronal heights might be an important condition for the formation of coronal shock waves .
we thank monique pick and jean - marie malherbe for providing the h@xmath7 data from the meudon observatory .
we also thank the soho , lasco , trace , integral , aip and nanay teams for their open data policy . n.m .
acknowledges support by the austrian research program moel - plus . m.t .
acknowledges project apart 11262 of the austrian academy of sciences .
acknowledges the austrian fonds zur frderung der wissenschaftlichen forschung under project fwf p20867-n16 . ,
g. , klassen , a. , estel , c. , & thompson , b. j. 1999 , in esa special publication , vol .
446 , 8th soho workshop : plasma dynamics and diagnostics in the solar transition region and corona , ed .
vial & b. kaldeich - sch , 477 | we study the well pronounced moreton wave that occurred in association with the x17.2 flare / cme event of october 28 , 2003 .
this moreton wave is striking for its global propagation and two separate wave centers , which implies that two waves were launched simultaneously .
the mean velocity of the moreton wave , tracked within different sectors of propagation direction , lies in the range of @xmath0 km s@xmath1 with two sectors showing wave deceleration .
the perturbation profile analysis of the wave indicates amplitude growth followed by amplitude weakening and broadening of the perturbation profile , which is consistent with a disturbance first driven and then evolving into a freely propagating wave .
the eit wavefront is found to lie on the same kinematical curve as the moreton wavefronts indicating that both are different signatures of the same physical process .
bipolar coronal dimmings are observed on the same opposite east - west edges of the active region as the moreton wave ignition centers .
the radio type ii source , which is co - spatially located with the first wave front , indicates that the wave was launched from an extended source region ( @xmath2 mm ) .
these findings suggest that the moreton wave is initiated by the cme expanding flanks . |
the relationship between mergers , starbursts , and active galactic nuclei ( agn ) has long since been the subject of vigourous research .
it is clear that at least some of the ingredients necessary to trigger starbursts are also necessary to trigger agn : both phenomena require fuel , and both require that this fuel be somehow displaced from one point to another , whether to replenish the material in the accretion disk , or to compress it resulting in enhanced star formation .
there is also evidence that mergers are important to both phenomena , although the precise role that they play is still debatable .
the high incidence of mergers in starburst galaxies ( * ? ? ? * and references therein ) might indicate that mergers are indeed necessary to trigger massive starbursts , but they are not sufficient , since mergers often result in only moderately enhanced star formation ( e.g. , * ? ? ? * ) .
mergers have , on the other hand , only been shown to be required to trigger nuclear activity in far - infrared ( fir ) loud qsos @xcite , and are certainly not sufficient , as can be seen from the relatively low incidence of agn in any catalog of interacting galaxies .
what are the additional ingredients necessary to trigger starbursts and agn ?
are these ingredients common to both phenomena ?
we already know that starbursts can exist in the absence of agn , as has been found in the case of many ultraluminous infrared galaxies ( at least at the lower luminosity end ; e.g. , * ? ? ?
* ) . in these proceedings ,
we address the question of whether the converse is true , i.e. , whether it is possible to trigger a high luminosity agn ( specifically a qso ) without triggering also a starburst . in
the following sections , we review our current knowledge of star formation in different kinds of qsos . for the purposes of this discussion , have grouped qsos in four classes according to some of their observational characteristics rather than their intrinsic properties .
one hypothesis that clearly ties in merger - induced star formation with qso activity is that of @xcite , who suggest that ultraluminous infrared galaxies ( ulirgs ) play a dominant role in the formation of all qsos . according to this hypothesis ,
ulirgs are the result of strong interactions or mergers which funnel gaseous material into the central regions of galaxies , thus fueling intense star formation and the qso activity .
ulirgs are then dust - enshrouded qsos which , after blowing away the dust , become classical qsos .
we have tested this hypothesis through a systematic imaging and spectroscopic study of host galaxies of low - redshift qsos found in a region of the far - infrared ( fir ) two - color diagram between the region where most qsos lie and the region occupied by ulirgs ( * ? ? ? * and references therein ) .
these objects are presumably in some transition stage between the ulirg and `` normal '' qso phases .
spectra were obtained of the host galaxies and/or strongly interacting companions for all qsos in the sample with the keck low - resolution imaging spectrometer ( lris ) .
we obtained ages for the starburst or post - starburst component in different regions of each host galaxy using the procedure illustrated by fig .
[ 3c48spec ] . using these data along with @xmath2 and ground - based images , we constructed detailed star - formation and interaction histories for each of the objects in the sample .
2 e of the quasar ( heavy solid line ) .
this spectrum has been fitted with a bruzual - charlot model ( dotted line ) comprising a starburst component of weighted average age 5 myr ( upper light solid line ) and an older component with exponentially decreasing star formation ( e - folding time 5 gyr ; lower light solid line ) .
the lower balmer absorption lines are contaminated with emission and are not used in the fitting procedure ; these and other emission lines are due , not to hii regions , but rather to extended emission photoionized by the quasar . the exposure totaled 3600 s with lris on keck ii through a 1 slit .
( from * ? ? ?
[ 3c48spec],width=297 ] we found that every `` transition '' qso is undergoing a strong tidal interaction , and most are major mergers where at least one of the parent galaxies was a disk galaxy .
the spectra are characteristic of e+a galaxies , and are successfully modeled by an underlying old population ( the stellar component present in the parent galaxies prior to interaction ) plus a superposed instantaneous burst population ( presumably resulting from the interaction ) .
all of the hosts have very young starburst ages , ranging from ongoing star formation to @xmath3 myr . by modeling spectra from many discrete regions across the hosts , we created velocity fields and age maps from the stellar populations . by comparing the starburst ages of the central stellar components with those of the more extended emission
, we determined the relative ages between stellar populations in various regions of the host galaxies .
these estimates , along with dynamical ages , place constraints on the timescale for concentrating material in the nucleus .
the concentration of material is likely to have triggered the central strong starbursts and the qso activity roughly simultaneously .
the age of the peak starburst is , therefore , representative of the age of the qso activity .
to summarize , our study showed that the qso and ulirg phenomena are physically related in these transition objects , and firmly established that at least some qsos can be traced back to a merger and a starburst phase .
star formation in these qsos is then , not only prominent , but clearly linked to the triggering of the qso activity .
however , these fir qsos may not be telling the story of the qso population as a whole . by comparing the distribution of the ratio @xmath4 for pg qsos @xcite and for `` transition '' qsos
, we estimate that the transition sample is representative of @xmath5 of the optically selected pg qso sample . while there are many uncertainties in this estimate ( especially since the ratio may be more indicative of reddening than of intrinsic characteristics in the objects )
, it clearly shows that these objects are not representative of the majority of ( at least optical ) qsos , but rather form a relatively small fraction of the population .
the rare class of post - starburst quasars shows simultaneously an agn and a massive luminous starburst of a few @xmath6100 myr old . because of their composite spectra displaying broad emission lines as well as the balmer jump and strong balmer absorption lines characteristic of type - a stars , these qsos are sometimes called `` q+a '' objects .
a striking case is un j1025@xmath70040 @xcite .
the strong starburst component has an age of @xmath8 myr , and a bolometric luminosity of @xmath9l@xmath10 , equal to that of the quasar .
a younger un j1025@xmath70040 ( tens of myr after the starburst ) would have a more luminous stellar population and would likely be dust enshrouded , placing it in the ulirg class .
un j1025@xmath70040 has a nearby companion galaxy also in a post - starburst phase @xcite .
the companion appears to be interacting with the quasar host galaxy , and this interaction may have triggered both the starburst and the quasar activity in un j1025@xmath70040 .
@xmath2 wfpc2 imaging @xcite shows that the starburst is nuclear , the host resembles a merger remnant , and a less massive young starburst is also present .
acs snapshot of the post - starburst qso sdss231055@xmath7090107 and its interacting companion galaxy .
[ qaimg],width=288 ] the new large qso surveys 2df and sdss are revealing post - starburst qsos in significant numbers for the first time .
we have spectroscopically selected post - starburst quasars in the sdss data release 2 ( dr2 ) using an automated algorithm based on one that @xcite used to select post - starburst galaxies .
we have found @xmath11250 post - starburst quasars roughly 5% of @xmath12 qsos that show clear evidence for post - starburst stellar populations of significant luminosity / mass ( brotherton et al , in preparation ) .
sdss images show that about 40% of our objects appear to be close interacting doubles or have companions within 10 . even most single sources show some extended fuzz from the host galaxy , and often features like tidal tails indicating merger activity .
an @xmath2 advanced camera for surveys ( acs ) snapshot survey of these objects is currently underway and is already revealing a much higher rate of tidal interaction than can be inferred from the sdss images . in figs .
[ qaimg ] and [ qa ] we show , respectively , the image and spectra of one of the sdss qsos with its companion , which appears to have a similar , though much less luminous , post - starburst population to that of the host . 090107 an its companion , showing that both have undergone dramatic episodes of star formation .
[ qa],width=316 ] post - starburst qsos represent the most dramatic cases of star formation in qso host galaxies . however , they are only a small fraction of the total qso population , and their relationship to other qsos as well as to ulirgs and e+a galaxies is currently unknown .
we are conducting a comprehensive study including @xmath2 acs ( fig . [ qaimg ] ) , keck lris ( fig . [ qa ] ) , and spitzer irs observations that will help us determine the nature of this relationship .
another class of qsos that is being uncovered with the advent of large area surveys at longer wavelengths , such as 2mass and first , is that of the so - called `` red '' qsos . in general , these objects have some of the same characteristics as blue qsos , such as ( at least some of the ) strong broad emission lines and high bolometric luminosities , but with much redder continua . at present
, there is not a clear definition for red qsos , so that the different objects that are catalogued as red qsos do not form a a homogeneous class .
for example , in the sample of red qsos drawn from the parkes half - jansky flat - radio - spectrum sample ( phfs ; * ? ? ?
* ) , the reddening is most likely due to red synchrotron emission .
the majority of red qsos discovered to date , however , appear to be reddened by dust , and thus they are considered the dust - obscured equivalent of the blue qso population ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
even then , there are cases where the dust is intervening rather than intrinsic to the qso ( e.g. , * ? ? ?
here we focus on those objects where the dust is presumably near the nucleus .
the nature of red qsos remains uncertain .
there are two popular theories to explain why some quasars appear reddened by dust .
the first uses an analogy with the seyfert galaxy and radio galaxy / quasar orientation - based unified schemes in which whether the observer sees a type-1 ( seyfert-1 , normal quasar ) object or a type-2 ( seyfert-2 , quasar-2 ) object depends on the angle at which the observer views the agn . in this model , dust - reddened quasars
are predicted to be objects viewed close to the transition angle between type-1 and type-2 which just scrapes the surface of a torus of gas and dust surrounding the agn .
the radio - loud red quasar 3c68.1 is thought to be at an angle to the line of sight between those of normal radio - loud lobe - dominated quasars and their fully obscured counterparts , radio galaxies @xcite .
the second theory is that the dust which reddens the qso is produced in young objects as a result of the starburst which follows the merger of two galaxies and which also triggers the qso , once again as in the scenario of @xcite .
thus a red qso phase might be expected at some point early in the life of most quasars . in either case , the heavy extinction of the qso nuclei at optical wavelengths makes red qsos excellent candidates to study stellar populations in their host galaxies , yet this fact has remained largely unexploited .
we have obtained keck esi spectra of 15 red qsos from the @xcite 2mass sample ; spectra of two of these objects are shown in fig .
[ redspec ] .
the spectra , which clearly emission characteristic of qsos , the spectra of the host galaxies suffer little contamination from the qso at shorter wavelengths .
the spectra shown here have not been corrected for scattered qso light to illustrate this point .
the spectrum of 2m015721 + 1712 shows a post - starburst spectrum characteristic of post - starburst qsos while that of 2m022150 + 1327 shows a somewhat older population.[redspec],width=297 ] show stellar absorption lines , have not been corrected for scattered qso light , demonstrating that they suffer little contamination from the nucleus at wavelengths shorter than h@xmath13 .
roughly half of the sample show strong post - starburst populations similar to those of post - starbursts qsos . of the remaining objects , most , if not all , show evidence for relatively recent star formation .
thus it appears that starbursts also play an important role in these objects .
how common are red qsos ?
@xcite estimate that heavily dust - reddened qsos comprise 20% of the total qso population for @xmath14 .
@xcite and @xcite , among others , suggest that so far we have only detected the `` tip of the red quasar iceberg '' , which could comprise up to several times the blue qso population .
as more surveys are published and selection methods are refined , estimates for the fraction of red qsos are likely to become more accurate .
this will have important consequences in areas ranging from galaxy evolution to cosmology and , for the current discussion , it will allow us to have better statistics on the number of qso hosts that show young post - starburst populations .
we use the term `` classical '' here to describe any luminous optical qso not covered by the first three classes of qsos discussed above .
recent imaging studies have shown that the majority of low - redshift classical qsos , whether radio - loud or radio - quiet , reside in the centers of galaxies that have relaxed light distributions like those of elliptical galaxies ( e.g. , * ? ? ? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . unlike the case of red qsos , the study of stellar populations in classical qso hosts is hampered by the difficulty of observing absorption line spectra in galaxies that tend to be overpowered by their bright nuclei .
for this reason , there have been few spectroscopic studies that deal with star formation in these objects .
* and references therein ) carried out imaging and spectroscopy of a statistically - matched sample of radio galaxies , quasars , and radio - quiet qsos at @xmath15 .
they find that essentially all of the host galaxies having nuclei in the quasar luminosity range are bulges that have properties `` indistinguishable from those of quiescent , evolved , low - redshift ellipticals of comparable mass . '' in particular , they suggest that these galaxies have truly old stellar populations with no episodes of massive star formation in the recent past .
several other surveys of agn host galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) also indicate that galaxies hosting the most luminous agn are most often bulge - dominated . however , all of these surveys find that the host colors are significantly bluer than those of inactive elliptical galaxies and they are consistent with the presence of intermediate age starbursts . based on their position in the @xmath16(4000)/h@xmath17 plane ,
@xcite suggest that these agn hosts have had significant bursts of star formation in the past @xmath18 gyr .
the objects studied in these surveys are on average @xmath19 magnitude fainter than those in the dunlop et al . sample .
however , it is possible that the same signs of intermediate age starbursts may be present in the hosts of higher luminosity objects , but that they are simply more elusive due to the technical difficulties mentioned above . to investigate this possibility ,
we have obtained and modeled very deep ( @xmath20 hr exposures ) keck lris spectra of 14 @xmath15 classical qso host galaxies in the dunlop et al .
sample that are dominated by spheroids .
the surprising results are that we have found a starburst component in all but one of the host galaxies , and in the majority of objects we find traces of major starburst episodes involving @xmath21 of the mass of the stellar component , with ages ranging from 0.6 gyr to 2.2 gyr ( see fig . [ oldsb ] ) . while there are uncertainties in the precise starburst age estimates due to metallicity , initial mass function , and the models themselves , the high signal - to - noise ratio of the spectra clearly show , in addition to an underlying very old component , a younger stellar component that is significantly older than a few @xmath22 yr .
this component can not be mimicked either by a small amount of constant star formation or by one less dramatic and substantially younger starburst phase . in any case ,
our results show conclusively that the host galaxies of classical qsos are @xmath23 made up almost purely of ancient stellar populations .
fit of a two component model to the data ( dotted line ) .
the model is the sum of a 12 gyr old population and a 2.2 gyr instantaneous starburst model which comprises 40% of the mass of the total population .
the bottom panel shows residuals of the fit.[oldsb],width=326 ] are these major starbursts episodes connected to a merger event ?
elliptical hosts formed through mergers would be expected to show fine structure indicative of past tidal interactions .
this tell - tale structure , which includes shells , ripples , and boxy isophotes , is commonly seen in nearby ellipticals and indicates that these early type galaxies were either formed or structurally modified by mergers in the not so distant past .
fine structure tends to persist even after the more overt signs of interaction have faded , and it is thought to be sensitive to the dynamical age of the merger ( e.g. , * ? ? ?
* ; * ? ? ?
these studies indicate that fine structure can in general be detected even a few gyr after the last major merger event .
we are obtaining very deep @xmath2 images of five of the host galaxies for which we have keck spectra , using acs with the f606w broad - v filter .
the subsample was chosen to include those objects for which we have the most robust age determinations and that do not show emission lines in the host galaxy to avoid confusion in the interpretation of the imaging when using a broad filter .
three objects have been observed so far , and all three show signs of tidal interactions : one appears to be in a system of three or four interacting galaxies , while the other two show traces of past major merger events . figure [ shells ] shows a dramatic example of shell structure in one of the qso hosts .
acs image of the @xmath15 classical qso mc2 1635 + 119 showing spectacular shell structure .
the second panel shows unsharp - like masking to highlight the sharp features that make up the shells.[shells],width=288 ] acs image of the @xmath15 classical qso mc2 1635 + 119 showing spectacular shell structure .
the second panel shows unsharp - like masking to highlight the sharp features that make up the shells.[shells],width=288 ] the results from both our keck spectroscopic and @xmath2 imaging observations point toward the same conclusion : classical qsos reside in host galaxies that have had a major tidal interaction and starburst episode within the last two or so billion years .
but the question remains : is there a connection between this catastrophic event in the past and the current nuclear activity ?
an oversimplistic interpretation assuming that the gyr old merger triggered the qso activity at that time implies a qso duty cycle of more than a billion years , a result in stark contrast to current theoretical estimates ( e.g. , @xmath24 yr by * ? ? ? * ) .
the scenario may be more complicated , however .
qso activity may require , for example , a two step process : first , collection of gas by a major merger , and then , triggering of the nuclear activity by other mechanisms such as a minor merger .
however , before we worry about these complex possibilities , a more fundamental question needs to be considered : are the host galaxies of classical qsos truly distinct from typical inactive elliptical galaxies ?
in other words , if we were to study a sample of inactive ellipticals at @xmath15 in the same depth , might we find similar shell structure and traces to major starburst episodes ?
a matching study of a control sample of elliptical galaxies in environments similar to those of qsos and with similar effective radii is clearly necessary in order to answer these questions .
such a study would allow us to interpret our results in the proper context and determine once and for all whether classical qsos reside in normal elliptical galaxies or not .
although we are not yet able to answer the question of whether the triggering of a qso is necessarily accompanied by the triggering of a starburst , we have strong hints that this may well be the case .
star formation plays a prominent role in three of the qso classes we have discussed : fir qsos , q+a s , and red qsos .
most of these objects , as well as a small fraction of classical qso hosts , show relatively young post - starburst populations of a few @xmath610@xmath25 yr .
however , is still unclear what fraction of all qsos is represented by these objects .
a conservative estimate would indicate that they form at least one fourth of the total qso population if we add 7.5% for fir qsos , 5% for q+a s , and 20% for red qsos , and allow for some overlap between these classifications .
however , the fraction could be as high as 80% or more if the population of red qsos proves to be as numerous as currently suspected @xcite .
the rest of the qso hosts appear to have traces of massive starbursts with ages @xmath2610@xmath27 yr . whether these starburst episodes are linked to the event that triggered the nuclear activity remains to be established .
we thank nicola bennert for her assistance in the reduction of q+a spectra , and zhaohui shang and rajib ganguly for providing the original version of fig .
[ qaimg ] .
this work was supported in part under proposal go-10421 by nasa through a grant from the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, under contract nas5 - 26555 .
additional support was provided by the national science foundation , under grant number ast 0507450 .
some of the data presented herein were obtained at the w.m .
keck observatory , which is operated as a scientific partnership among the california institute of technology , the university of california and the national aeronautics and space administration .
the observatory was made possible by the generous financial support of the w.m .
keck foundation .
bahcall , j. , kirhakos , s. , saxe , d. h. , schneider , d. p. 1997
, apj , 479 , 642 bergvall , n. , laurikainen , e. , aalto , s. 2003 , a & a , 405 , 31 brotherton , m. s. , wills , b. j. , dey , a. , van breugel , w. , antonucci , r. 1998 , apj , 501 , 110 brotherton , m. s. , et al . 1999 ,
apj , 520 , l87 brotherton , m. s. , et al .
2002 , pasp , 114 , 593 canalizo , g. , stockton , a. , brotherton , m. s. , & van breugel , w. 2000 , aj , 119 , 59 canalizo , g. , & stockton , a. 2000 , apj , 528 , 201 canalizo , g. , stockton , a. 2001 , apj , 555 , 719 cutri , r. m. , nelson , b. o. , francis , p. j. , smith , p. s. 2002 , in agn surveys , asp conference proceedings , 284 , 127 disney , m. j. , et al .
1995 , nature , 376 , 150 dunlop j. s. , mclure , r. j. , kukula , m. j. , baum , s. a. , odea , c. p. , hughes , d. h. 2003 , mnras , 340 , 1095 floyd , d. j. e.,kukula , m. j. , dunlop , j. s. , mclure , r. j. , miller , l. , percival , w. j. , baum , s. a. , odea , c. p. 2004 , mnras , 355 , 196 francis , p. j. , drake , c. l. , whiting , m. j. , drinkwater , m. j. , webster , r. l. 2001 , pasa , 2001 , 18 , 221 glikman , e. , gregg , m. d. , lacy , m. , helfand , d. j. , becker , r. h. , white , r. l. 2004 , apj , 607 , 60 gregg , m. d. , lacy , m. , white , r. l. , glikman , e. , helfand , d. , becker , r. h. , brotherton , m. s. 2002 , apj , 564 , 133 hall , p. b. , et al .
2002 , apjs , 141 , 267 jahnke , k. , kuhlbrodt , b. , wisotzki , l. 2004 , mnras , 352 , 399 kauffmann , g. , et al .
2003 , mnras , 346 , 1055 lacy , m. , gregg , m. , becker , r. h. , white , r. l. , glikman , e. , helfand , d. , winn , j. n. 2002 , aj , 123 , 2925 marble , a. r. , dean , h. c. , schmidt , g. d. , smith , p. s. 2003 , apj , 590 , 707 sanchez , s. f. , et al .
2004 , apj , 614 , 586 sanders , d. b. , soifer , b. t. , elias , j. h. , madore , b. f. , matthews , k. , et al .
apj , 325 , 74 sanders , d. b. , phinney , e. s. , neugebauer , g. , soifer , b. t. , & matthews , k. 1989 , apj , 347 , 29 sanders , d. b. & mirabel , i. f. 1996 , ara & a , 34 , 749 sansom , a. e. , hibbard , j. e. , schweizer , f. 2000 , aj , 120 , 1946 schweizer , f. , seitzer , p. , faber , s. m. , burstein , d. , dalle ore , c. m. , gonzalez , j. j. 1990 , apj , 364 , l33 white , r. l. , helfand , d. j. , et al . 2003 , 126 , 706 yu , q. & tremaine , s. 2002 , mnras , 335 , 96 zabludoff , a. i. , zaritsky , d. , lin , h. , tucker , d. , hashimoto , y. , shectman , s. a. , oemler , a. , kirshner , r. p. 1996 apj , 466 , 104 | many of the conditions that are necessary for starbursts appear to be important in the triggering of qsos . however , it is still debatable whether starbursts are ubiquitously present in galaxies harboring qsos . in this paper
we review our current knowledge from observations of the role of starbursts in different types of qsos .
post - starburst stellar populations are potentially present in the majority of qso hosts .
qsos with far - infrared colors similar to those of ultraluminous infrared galaxies invariably reside in merging galaxies that have interaction - induced starbursts of a few hundred myr or less .
similar , but dramatically more luminous post - starburst populations are found in the recently discovered class of qsos known as post - starburst qsos , or q+a s . both of these classes , however , comprise only a small fraction ( @xmath0% ) of the total qso population .
the so - called `` red '' qsos generally suffer from strong extinction at optical wavelengths , making them ideal candidates for the study of hosts .
their stellar populations typically show a post - starburst component as well , though with a larger range of ages .
finally , optical `` classical '' qso hosts show traces of major star formation episodes ( typically involving @xmath110% of the mass of the stellar component ) in the more distant past ( 1 - 2 gyr ) . these starbursts appear to be linked to past merger events .
it remains to be determined whether these mergers were also responsible for triggering the qso activity that we observe today .
quasars : general , galaxies : interactions , galaxies : starburst 98.54.aj , 98.54.ep , 98.62.lv , 98.65.fz |
type ia supernovae ( sne ) are the end - products of white dwarfs with a mass approaching the chandrasekhar limit , which results in a thermonuclear explosion of the star .
in addition to their use as cosmological distance indicators ( e.g. , * ? ? ? * ; * ? ? ?
* ) , type ia sne ( henceforth sne ia ) are a major contributor to the chemical evolution of galaxies .
it is therefore unfortunate that we do not yet know what makes a sn ia .
this lack of knowledge makes it difficult to gain a physical understanding of the explosions , so that we can model possible evolution , which compromises their use as distance indicators .
it also means we do not fully understand the timescale over which sne ia turn on , adding a large uncertainty to our understanding of the chemical evolution of galaxies .
unveiling the progenitor scenario for sne ia is difficult because white dwarfs ( wds ) can , theoretically , reach their fatal chandrasekhar mass in many ways , and disentangling which is the correct one ( if there is just one ) , is challenging from an observational point of view . nonetheless , there are two basic families of models leading to a sn ia , the single - degenerate model ( sd ) and the double - degenerate model ( dd ) . in the sd scenario , a wd accretes mass from a hydrogen - rich companion star before reaching a mass close to the chandrasekhar mass and going off as supernova . in the dd scenario ,
two wds merge , with the more - massive wd being thought to tidally disrupt and accrete the lower - mass wd ( see , e.g. , * ? ? ? * and references therein ) .
observations can potentially discriminate between the progenitor models of sne ia .
for example , in all scenarios with mass transfer from a companion , a significant amount of circumstellar gas is expected ( see , e.g. , * ? ? ?
* ) , and therefore a shock is bound to form when the supernova ejecta are expelled
. the situation would then be very similar to circumstellar interaction in core - collapse sne , where the interaction of the blast wave from the supernova with its circumstellar medium results in strong radio and x - ray emission @xcite .
on the other hand , the dd scenario will not give rise to any circumstellar medium close to the progenitor system , and hence essentially no radio emission is expected .
radio and x - ray observations of sn 2011fe have provided the most sensitive constraints on possible circumstellar material @xcite around a normal sn ia .
the claimed limits on mass loss rate from the progenitor system are @xmath5 and @xmath6 from radio @xcite and x - rays @xcite , respectively , assuming a wind velocity of 100 km s@xmath7 .
radio ( e.g. , * ? ? ?
* ; * ? ? ?
* ) and x - ray ( e.g. , * ? ? ?
* ; * ? ? ?
* ) observations of other , more distant sne ia , have resulted in less constraining upper limits on wind density .
the non - detections of radio and x - ray emission from sne ia have added to a growing consensus that a large fraction of sne ia may not be the result of sd scenarios ( e.g. , * ? ? ?
despite the non - detection of radio and x - ray emission , there is evidence of possible circumstellar material in the form of time - varying absorption features in the optical na i d line for a few sne ia @xcite , supposed to arise in circumstellar shells .
the exact location of the absorbing gas is still debated ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , and probably varies from case to case .
the number of sne ia showing indications of circumstellar shells could be significant , although the uncertainty is still large ( ( 18@xmath8)% ; @xcite ) . just as with the radio and x - rays , no optical circumstellar emission lines from normal sne
ia have yet been detected ( e.g. , * ? ? ?
* ) , although there are a few cases with strong emission ( see , e.g. , * ? ? ?
* for an overview ) .
those sne ia with strong circumstellar interaction constitute a very small fraction of all sne ia , probably only @xmath9% @xcite .
recently , @xcite serendipitously discovered in the nearby galaxy m 82 ( d=3.5 mpc ) .
@xcite classified sn 2014j as a sn ia , which makes it the closest sn ia since sn 1986 g in cen a , almost three decades ago .
the supernova exploded between ut 14.56 jan 2014 and 15.57 jan 2014 according to the imaging obtained by @xcite , and its j2000.0 coordinates are ra=09:55:42.121 , dec=+69:40:25.88 @xcite . for a further discussion on the discovery and early rise of the optical / ir emission
, we refer to @xcite and @xcite .
the vicinity of makes it a unique case for probing its prompt radio emission , and thus constrain its progenitor system .
we observed sn2014j with the electronic multi element radio interferometric network ( emerlin ) at 1.55 and 6.17 ghz , and with the electronic european very long baseline interferometry network ( evn ) at a frequency of 1.66 ghz .
we show in table 1 the summary for our observations , along with radio data obtained by others .
we observed with emerlin on 28 january 2014 , at a frequency of 1.55 ghz , and on 29 - 30 january 2014 , at a frequency of 6.17 ghz .
our observing array included , at both frequencies , all emerlin stations ( lovell , jodrell mk2 , darham , pickmere , cambridge , defford , knockin ) .
given the expected faintness of we used a phase - reference observing scheme , with @xmath108 minutes spent on the sn , and @xmath102 minutes on the nearby , bright phase - calibrator j0955 + 6903 ( ra=09:55:33.1731 ; dec=69:03:55.061 ) .
we used 3c286 as our absolute flux density calibrator , and oq208 as bandpass calibrator .
we observed in dual - polarization mode at both frequencies .
the bandwidth at 1.55 ( 6.17 ) ghz was of 512 ( 1024 ) mhz .
each of those frequency bands was split into 4 ( 8) spectral windows ( spw ) of 128 mhz each .
each spw was in turn split into 512 channels / polarisation .
[ tab : radiolog ] we loaded the data into the nrao astronomical image processing system ( _ aips _ ) of the national radio astronomy observatory ( nrao , usa ) , after having averaged them to 128 channels per spw ( i.e. , channel width of 500 khz ) .
we used aips for calibration , data inspection , and flagging , using standard procedures .
we lost @xmath1015% of the data due to radio frequency interference ( rfi ) .
we imaged the @xmath1013(@xmath104 ) field of view of our 1.55 ( 6.17 ) ghz observations , including m82 , which has a strong and spatially complex radio structure , using a robust=0 uv - weighting scheme .
we used those in - beam radio sources to refine the target field calibration via several rounds of phase and amplitude self - calibration . following these rounds of self - calibration , we reweighted the target visibilities to account for difference in the sensitivity of the individual e - merlin antennas .
our final imaging yielded 13.6 and 12.4 @xmath11jy / bm r.m.s .
noise levels at the location of sn2014j at 1.55 and 1.67 ghz , respectively .
we observed our target source , on 3 - 4 february 2014 and 19 february 2014 , using the eevn at 1.66ghz .
we used a sustained data recording rate of 1024 mbit s@xmath7 , in dual - polarisation mode and with 2-bit sampling .
each frequency band was split into 8 intermediate subbands of 16 mhz bandwidth each , for a total synthesized bandwidth of 128 mhz .
each subband was in turn split into 128 ( 64 ) spectral channels of 125 ( 250 ) khz bandwidth each for the observations on 3 - 4 february ( 19 february ) 2014 .
our observations on 3 february included the following six antennas of the evn : effelsberg , westerbork ( phased array ) , jodrell bank ( mk 2 ) , medicina , onsala , and torun .
in addition to these antennas , our observing run on 19 february included also the antennas of noto and sheshan .
we observed our target source , sn 2014j , phase - referenced to the core of the nearby galaxy m81 , known to be very compact at vlbi scales , with a typical duty cycle of 5 minutes .
we used the strong source da193 as fringe finder and bandpass calibrator .
all the data were correlated at the evn mkiv data processor of the joint institute for vlbi in europe ( jive , the netherlands ) , using an averaging time of 1 s. we used _
aips _ for calibration , data inspection , and flagging of our eevn data , using standard procedures .
those steps included a - priori gain calibration ( using the measured gains and system temperatures of each antenna ) , parallactic angle correction and correction for ionosphere effects .
we then aligned the visibility phases in the different subbands , i.e. , `` fringe - fitted '' the data , solved for the residual delays and delay rates , and interpolated the resulting gains into the scans of .
we then imaged a field of view of 3@xmath123centered at the position given by @xcite , and applied standard imaging procedures using _ aips _ , without averaging the data either in time , or frequency , to prevent time- and bandwidth smearing of the images .
we used natural uv - weighting to maximize the signal to noise ratio in our final images .
the radio and x - ray non - detections of sne ia , in conjunction with indications of circumstellar shells around some sne ia ( see [ sec : intro ] ) , is a conundrum that yet has to find a solution .
the nearby northern hemisphere sne 2011fe and 2014j offer a possibility to use the most sensitive radio facilities present to probe circumstellar emission .
in particular , we now interpret the upper limits on radio emission from sn 2014j in [ sec : obs ] within the framework of circumstellar interaction . indeed , when the supernova shock - wave ploughs through the circumstellar gas , a high - energy density shell forms . within this shell
, electrons are accelerated to relativistic speeds and significant magnetic fields are generated , especially if the circumstellar gas is pre - ionized .
for the low wind densities discussed in this paper , pre - ionization is likely to occur @xcite .
the relativistic electrons radiate synchrotron ( radio ) emission ( e.g. , * ? ? ?
a proper modeling of the radio emission from sne requires , in principle , taking into account coulomb , synchrotron , and ( inverse ) compton losses of the relativistic electrons .
however , since we only have upper limits for the radio emission from sn 2014j , we will discuss the radio emission from sne ia within a scenario of type ib / c sne ( see , e.g. , * ? ? ?
* ) , neglecting energy losses for the relativistic electrons ( c.f . * ? ? ?
* ; * ? ? ?
* for a more general treatment ) .
the spectrum of the radio emission from those sne follows the `` synchrotron self - absorption '' ( ssa ) form , i.e. , a rising power law with @xmath13 ( low - frequency , optically thick regime ) , and a declining power law , @xmath14 ( high frequency , optically thin regime ) , where @xmath15 is assumed to be constant . for most
well studied sne , @xmath16 @xcite .
we assume that electrons are accelerated to relativistic energies , with a power law distribution , @xmath17 ; where @xmath18 is the energy of the electrons and @xmath19 is the lorentz factor . for synchrotron emission , @xmath20 , which indicates that @xmath21 should be used . here , we study both the case of a circumstellar structure created by a wind , as well as the case with constant density circumstellar gas . for the wind case
, we make the standard assumption that the sn progenitor has been losing matter at a constant rate , @xmath22 , so that the circumstellar density has a radial profile : @xmath23 , where @xmath24 is the wind velocity , @xmath25 is the radial distance from the star , @xmath26 is the particle density and @xmath11 is the mean atomic weight of the circumstellar matter . to calculate the shock expansion , we use the thin - shell approximation @xcite , with the extensions of @xcite .
we assume that the innermost ejecta has a density slope of @xmath27 , which at some velocity of the ejecta rolls over to a steeper density profile , @xmath28 ( @xmath29 ) .
we assume @xmath30 , motivated by the explosion models of @xcite , and use @xmath31 , which is a good approximation to the outer density profile of a supernova that stems from a radiative star @xcite .
assuming an ejecta mass of @xmath32 and a kinetic energy of the explosion of 10@xmath33 erg , the break in power - law index in our model occurs at @xmath34 , which agrees with the angle - averaged results of @xcite .
the supernova expansion can be well approximated by a power law , @xmath35 , so that the shock speed , @xmath36 @xcite .
here , @xmath37 , and @xmath38 is the density slope of the circumstellar gas , which for the steady wind case is @xmath39 .
( in [ sec : ism ] we also discuss the case @xmath40 . )
the shock speed at 10 days in this model is @xmath41 for @xmath42 and @xmath43 ; @xmath44 and @xmath45 both scale as @xmath46 . for any sensible pre - supernova wind speed ,
the supernova shock is strong . assuming a polytropic gas with @xmath47 , the compression of the gas across the shock is @xmath48 , and the post - shock thermal energy density is @xmath49 , where @xmath50 is the pre - shock density .
following @xcite , we denote @xmath51 , where @xmath52 is the ( post - shock ) magnetic energy density ; and @xmath53 , where @xmath54 is the energy density of the relativistic particles , assumed for simplicity to be electrons .
we assume that the power law index of the relativistic electron population stays constant with time at @xmath55 , although we have also studied cases with @xmath56 ( see [ sec : sens - res ] ) .
the most uncertain parameters refer to the microphysics of the shocked gas , namely @xmath57 and , to a greater extent , @xmath58 .
indeed , it seems that @xmath59 with some small dispersion around this value @xcite , whereas @xmath58 appears to vary more among supernovae , and is hence largely unknown .
therefore , we fix @xmath60 , and take @xmath58 as a free parameter
. we can easily find @xmath61 by integrating the relativistic electron distribution between @xmath62 and infinity , which yields @xmath63 .
we estimate the minimum lorentz factor of the relativistic electrons , @xmath64 , assuming that all postshock electrons go into the power - law distribution with energy index @xmath65 ( cf . * ? ? ?
this means that @xmath66 $ ] ( see also * ? ? ?
* ) . here , @xmath67 is the electron density of the pre - shocked gas when it is fully ionized .
we assume a mix of h and he with an abundance ratio 10:1 , which together with @xmath60 and @xmath48 means that @xmath68 ^ 2 $ ] .
following @xcite , we add the constraint that @xmath69 .
to calculate the synchrotron spectrum , we follow the method by @xcite , i.e. , we use the observational evidence that the brightness temperature , @xmath70 , is expected to be somewhat below 10@xmath71 k ( cf .
* ; * ? ? ?
* ) . while we defer a more complete discussion about this to a future paper ( c .- i .
bjrnsson , in preparation ) , we have chosen a likely value of @xmath72 k , which should be correct to within a factor of @xmath73 .
the intensity at the frequency of the peak of the synchrotron spectrum , @xmath74 , is then defined as @xmath75 , whereas the intensity at any frequency is @xmath76 $ ] . here
@xmath77 is the source function and @xmath78 the synchrotron optical depth .
the latter is just @xmath79 , where @xmath80 is the path length through the emitting region along the line of sight , and @xmath81 . like @xcite
, we make the simplification that @xmath82 , where @xmath83 is the particle pitch angle . the constant @xmath84 can be found in , e.g. , @xcite .
the path length @xmath80 depends on the thickness of the synchrotron emitting region , @xmath85 . at the center , @xmath86 ( assuming the supernova ejecta to be transparent to radio emission ) , but can become significantly larger than unity toward the limb .
@xmath87 is the normalized impact parameter , so that @xmath88 .
we assume constant properties of the plasma within @xmath85 .
for @xmath39 , we have assumed a thickness of @xmath89 , which corresponds to that of the shocked circumstellar gas for @xmath90 and @xmath91 in the similarity solutions of @xcite , namely , @xmath92 for @xmath93 . for @xmath94 ,
the similarity solutions give @xmath95 for the same range in @xmath96 , and we have chosen @xmath97 for this @xmath38-value .
@xmath98 is therefore just due to the geometrical increase of the path length as @xmath87 increases . for convenience , we introduce , in addition to @xmath74 , also the frequency @xmath99 , defined as @xmath100 . in general , @xmath101 . for @xmath102 ,
we denote @xmath103 , @xmath104 and @xmath105 .
we can then derive the intensity for any impact parameter as @xmath106 , \label{eq : intensity}\ ] ] where @xmath107 depends on @xmath65 such that
@xmath108 \label{eq : fx}\ ] ] ( see also * ? ? ?
* ) . for @xmath109 , @xmath110 and @xmath111 .
to obtain the luminosity , one integrates over @xmath87 , so that @xmath112
. the longer path length toward the limb makes @xmath113 larger for the optically thin part of the spectrum than just assuming @xmath114 . for @xmath115 ,
the factor @xmath116 in the optically thin part is a weak function of @xmath85 , being @xmath117 for @xmath118 . for the optically thick part , @xmath119 .
this makes the _ observed _ spectrum peak at a somewhat higher frequency than @xmath74 .
( solid lines ) , and for @xmath120(dashed lines ) .
for the former we used @xmath121 and for the latter @xmath122 . the data points ( cf .
table [ tab : radiolog ] ) with 3@xmath123 upper limits for are in the right part of the figure . shown in the figure
is also the earliest 5.9 ghz 3@xmath123 upper limit for sn 2011fe @xcite , scaled to its distance of 6.4 mpc , together with a dotted line marking the predicted evolution for @xmath124(for @xmath121 ) .
common parameters in all models are @xmath60 , @xmath115 and @xmath125 .
see text for further details .
, width=340 ]
the radio emission from the supernova is subject to both ssa and possible external free - free absorption in the ambient medium .
while we include ssa in our model , we do not include free - free absorption since , as we show below , it is negligible . in previous analyses of sne ia @xcite ,
free - free absorption was assumed to be the most important factor to derive wind densities . for sne 2011fe and 2014j , x - ray non - detections
@xcite have put limits on @xmath126 of order @xmath127 for @xmath43 . from equation 6 in @xcite
it follows that the free - free optical depth , @xmath128 , for a fully ionized wind at @xmath129 k and moving at @xmath43 , is @xmath130 , where @xmath131 is in cm .
for such a low wind density , the shock radius is @xmath132 cm already at 2 days , which means that @xmath133 at 5.5 ghz at such an early epoch .
free - free absorption is thus insignificant and can be dismissed from our analysis .
we also note that @xcite used a similar argument to dismiss free - free absorption in their analysis of radio emission from sn 2011fe . in what follows ,
we therefore only consider ssa .
we now compare the radio data for sn 2014j in [ sec : obs ] with the predictions of the model presented in
[ sec : model ] .
if the supernova happens in an sd scenario , the accreting wd is expected to have lost some of the accreted material from the donor star through a wind .
this sets up a @xmath4 circumstellar structure ( cf .
[ sec : model ] ) .
as we show below , for the epochs of the radio observations , sn 2014j was clearly in its optically thin phase .
this simplifies the expressions above , so that the luminosity becomes @xmath134 where @xmath135 from this , together with expressions in [ sec : model ] and assuming @xmath115 , one gets @xmath136 if @xmath64 is not fixed .
if it is fixed @xmath137 at early epochs , when the shock velocity is high , @xmath64 is always larger than unity , and decreases as time goes on .
therefore , equation [ eq : lum1 ] applies . at later epochs ,
when the shock velocity is such that it would formally imply @xmath138 , our constraint on @xmath64 takes effect , and equation [ eq : lum2 ] applies . as stated in
[ sec : model ] , we fixed @xmath57 at 0.1 , and allowed @xmath58 to vary .
equations [ eq : lum1 ] and [ eq : lum2 ] can be used to scale @xmath57 even if only @xmath58 is varied . in figure
[ fig : lcurves ] we show models for @xmath139 and 0.1 .
an almost perfect overlap between modeled light curves occurs for the combination @xmath140 and @xmath121 , and @xmath141and @xmath122 . only at very early epochs ( @xmath142 days after explosion ) ,
does ssa play a role for the lowest frequencies .
the overlap is not surprising , since equation [ eq : lum1 ] shows that @xmath143 for fixed luminosity at early epochs in the optically thin part in our model .
we note that @xcite obtain a slightly different power - law index , @xmath144 , in their model . for all models in figure [ fig : lcurves ] , @xmath145 for the time span shown .
this means that equation [ eq : lum1 ] describes all light curves well , except for the lowest frequencies at @xmath146 days .
the values of @xmath147 for in figure [ fig : lcurves ] are chosen so that the 5.50 ghz light curves go through the jvla 3@xmath123 upper limit on day 8.2 .
the light curves for other frequencies lie below their corresponding upper limits .
the second most constraining limit is from our 1.55 ghz emerlin observation on day 13.8 , yielding @xmath148 for @xmath149 and @xmath150 .
we show in table [ tab : radiolog ] upper limits for all data points .
figure [ fig : lcurves ] also includes the most constraining upper limit for sn 2011fe @xcite , together with a 5.9 ghz light curve using @xmath151 , @xmath150 and @xmath152 .
the limit on mass - loss rate is somewhat below that of @xcite , who obtained @xmath153 .
the difference in those values probably stems from the difference in shell thickness of the emitting region , where we have adopted @xmath154 vs. @xmath155 @xcite , and our fixed @xmath70 . in any case
, the difference in the upper limit on @xmath147 is much smaller than that due to the uncertainty in @xmath58 . in principle , radio non - detections could also be due to ssa during the observed epochs . in this case , the observed frequency @xmath156 , and from equation [ eq : intensity ] we find that the observed flux at @xmath157 is @xmath158 .
the combination of @xmath159 is a weak function of @xmath147 , and to make ssa important for the observations discussed here would require values of @xmath147 much larger than those at which free - free absorption becomes important .
we can therefore fully dismiss ssa as a cause for the radio non - detections of sn 2014j . , but for circumstellar gas with constant density ( @xmath94 ) , i.e. , the insterstellar medium ( ism ) .
solid lines are for @xmath160 , and dashed line is for @xmath161 , assuming @xmath121 and @xmath122 , respectively .
the data upper limits for are the same as in figure [ fig : lcurves ] .
the constraint on @xmath162 is set by the stacked 1.66 ghz eevn data from days 20 and 35 .
the change in spectral slope around 25 days for the @xmath122 model is due to the @xmath69 constraint then coming into effect . for a comparison
we also show the 3@xmath123 upper limit from the stacked 5.9 ghz sample for sn 2011fe between days @xmath163 @xcite , together with part of the light curve for a model ( dotted red line ) assuming @xmath164 and @xmath121 .
, width=340 ] , but for different values of @xmath65 and @xmath96 .
the values of @xmath147 were chosen for the 5.50 ghz light curve to cross the upper jvla limit from day 8.2 .
note the change in slopes for the dashed lines around day 9 , when the condition @xmath165 comes into effect .
see text for further details.,width=340 ] if the progenitor of followed the double - degenerate channel , then the exploding wd is expected to be surrounded by the interstellar medium ( ism ) , which has a constant density .
@xcite discussed this scenario for sn 2011fe , and obtained a limit for the density of @xmath166 ( @xmath121 ) .
the general behavior of the radio light curves for the constant density ism case is different from the @xmath167 wind case in [
sec : wind ] . for the constant density case and @xmath115 ,
the radio luminosity increases with time ( see also * ? ? ?
* ) according to @xmath168 if @xmath64 is not fixed .
if it is fixed @xmath169 here , we substituted @xmath170 with @xmath162 to highlight the likely origin of the gas in the @xmath40 case . figure [ fig : lcurves_constant ] shows models with densities @xmath160 ( @xmath121 ) and @xmath161 ( @xmath122 ) . scaling according to @xmath171 ( cf . equation [ eq : lum3 ] ) makes the light curves for these models overlap fully , except for @xmath172 days , when the condition @xmath69 becomes important for the @xmath161 model .
the most constraining data are our eevn 1.66 ghz data , stacked together , and the model parameters were chosen for the modeled radio luminosity to match those data . however , due to the @xmath69 constraint , the 35 day data alone are almost as constraining as the stacked data . in figure [
fig : lcurves_constant ] , we also show the stacked 5.9 ghz data for sn 2011fe @xcite , together with a model characterized by @xmath164 and @xmath121 .
we are thus close to @xcite regarding the limit on @xmath162 for sn 2011fe .
for @xmath40 , we used @xmath173 , which is close to the value 0.1 used by @xcite .
the limit on @xmath162 we find for is @xmath174 times lower than for sn 2011fe , and is therefore clearly the lowest limit on density for the constant density case in any sn ia . as for the @xmath39 case ,
ssa is unimportant for the @xmath40 case . to be efficient enough to mute the radio emission to be consistent with the observed upper limit
, @xmath162 would have to be @xmath175 ( for @xmath176 ) , which is fully ruled out from x - ray limits , as well the normal optical behaviour of the supernova .
while usually not acknowledged in the sn ia literature , the obtained radio and x - ray upper limits on the circumstellar density are model dependent . in particular , equations [ eq : lum1 ] through [ eq : lum4 ] show how most parameters influence the results .
we already mentioned the uncertainty in @xmath57 and , especially , @xmath58 , whereas @xmath70 is observationally constrained by other similar radio sources .
the thickness of the radio - emitting region is yet another source of uncertainty , but probably small in comparison to other uncertainties .
there is also an uncertainty in the upper limit on @xmath147 due to values chosen for @xmath96 and @xmath65 .
figure [ fig : lcurves_params ] shows two models , both with @xmath121 and @xmath60 , but where we have also assumed @xmath177 and @xmath178 ( solid lines ) , and @xmath115 and @xmath179 ( dashed lines ) . in both models we have used the earliest @xmath180 jvla point for to constrain @xmath147 . for the @xmath177 model , @xmath181 , i.e. , a factor @xmath182 lower than the @xmath115 model in figure [ fig : lcurves ] .
the choice of @xmath115 therefore gives a conservative limit on @xmath147 , unless @xmath183 .
judging from type ib / ic sne , the expected deviation from @xmath115 is not large @xcite , so we do not consider the uncertainty in @xmath65 being a major source of uncertainty for @xmath147 . for the @xmath179 model , @xmath184 , i.e. , a factor @xmath185
higher than the @xmath178 model in figure [ fig : lcurves ] . the choice of @xmath178 may give too low a limit on @xmath147 , unless a shallower density profile than @xmath178 is expected .
indeed , @xcite argue that an exponential density profile of the outer ejecta fits early explosion models better than a power law , and could indicate steeper profiles than at least @xmath186 for the outermost ejecta
. we have also run models for @xmath186 , but the value for @xmath147 then becomes so low ( @xmath187 ) that the model breaks down , producing shock velocities in excess of @xmath188 .
for such low wind densities , a relativistic treatment of the shock interaction is needed , similar to for gamma - ray bursts . from the models of @xcite
it seems reasonable to assume that @xmath189 for the outermost ejecta , which in our model would imply an upper limit of @xmath190 .
the span in upper limit on @xmath147 by a factor of @xmath191 between @xmath192 shows that the unknown density profile for the outermost ejecta is an important source of uncertainty , and that accurate models for the outermost ejecta are needed .
this is even more evident for the constant density case @xmath40 , for which we find that our limit on @xmath162 ranges from @xmath193 for @xmath194 to @xmath195 for @xmath179 , assuming @xmath121 , @xmath60 and @xmath115 .
the solution for @xmath194 is , however , unphysical due to too large velocities for the shock during the first @xmath196 days , calling for a relativistic treatment of the dynamics .
the sd progenitor systems involve only one wd and include , in decreasing order of mass - loss rate from the supernova progenitor , symbiotic systems , wds with steady nuclear burning , and recurrent novae . in a symbiotic system ,
the wd accretes mass from a giant star @xcite .
the wd loses this accreted matter at rates of @xmath197and @xmath198 .
the radio emission from those systems should have been detected by our deep sensitive observations .
thus , our radio non - detection rules out a symbiotic system as the progenitor of sn 2014j ( red region in figure [ fig : mdot - vwind ] ) .
another possible sd scenario is one where a main sequence , subgiant , helium , or giant star undergoes roche lobe overflow onto the wd at rates of @xmath199 @xcite . at those accretion rates , the wd experiences steady nuclear burning @xcite . for an assumed fraction @xmath200 of the transferred mass to be lost from the system
, the mass - loss rate is constrained to @xmath201 and typical speeds of 100 @xmath202 3000 , where the low speeds apply for steady nuclear burning , while the high speeds apply to the systems with the highest accretion rates . at the lower end of @xmath203 , the mass loss through the outer lagrangian points of the system
proceeds at speeds up to @xmath10600 .
most of the parameter space for the low - accretion rate scenario is ruled out by our radio observations , if @xmath204 ( blue region in fig .
[ fig : mdot - vwind ] ) . at the upper end of @xmath203
the winds become optically thick , limiting the accretion rate to @xmath205 and wind speeds of a few @xmath206 1000 @xcite .
our data essentially rule out completely the high - accretion rate scenario of a wd with steady nuclear burning ( cyan region in fig .
[ fig : mdot - vwind ] ) . finally , another possible sd channel is that of recurrent novae , which lie at the lowest accretion rate regime among popular sd scenarios . here , a wd accreting at a rate @xmath207 , ejects shells of material at speeds of a few @xmath208 , with typical recurrence times of a few years .
the radio observations in table [ tab : radiolog ] probe a radius of @xmath209 cm ( for @xmath39 and @xmath210 ) , which constrains the presence of shells with recurrence times of @xmath211 yr . models of recurrent novae seem to indicate that as much as @xmath1015% of the accreted material over the recurrence time is ejected @xcite . for the typical accretion rates above
, this implies an ejected shell mass of @xmath212 , which should have been detected by our sensitive observations ( see gold region in fig .
[ fig : mdot - vwind ] ) .
unfortunately , the short duration of the nova radio burst , a few days at most , may have prevented its detection , so we can not rule out completely the possibility of a nova shell ejection . during the quiescent phase between nova shell ejections , the wd accretes at a rate of @xmath213 , so that the mass - loss wind parameter is @xmath214 . if @xmath121 , our observations rule out almost completely the scenario with wd accretion during the quiescent phase of the star , whereas the case with @xmath122 can not be excluded completely ( green region in fig . [
fig : mdot - vwind ] ) . in summary ,
our observations exclude completely symbiotic systems and the majority of the parameter space associated with stable nuclear burning wds , as viable progenitor systems for sn 2014j .
recurrent novae with main sequence or subgiant donors can not be ruled out completely , yet most of their parameter space is also excluded by our observations .
the alternative to a sd scenario is the dd channel , which involves two wds in a binary system . in this case
, the progenitor star is expected to have exploded in a constant ambient density medium .
we can estimate the density in the region surrounding sn 2014j from the column density of neutral hydrogen toward the supernova position , @xmath215 @xmath216 @xcite .
assuming a path length , @xmath217 100 pc , and solar abundance ( @xmath218 ) , the particle number density at the supernova location is @xmath219 0.9 @xmath220 .
our stacked eevn limits imply @xmath221 @xmath220 for @xmath222 ( 0.01 ) , and are thus consistent with the sn directly expanding into the interstellar medium
. therefore , our radio non - detections are consistent with the dd channel for sne ia .
we note that the limit imposed on @xmath162 by the microphysical parameter @xmath121 is formally similar to the likely value of @xmath162 at the sn location . yet , the uncertainties involved in this estimate are such that both values are in agreement . at any rate ,
the sensitivity of @xmath58 on @xmath162 demonstrates the usefulness of late - time radio observations to constrain this relevant microphysical parameter in sne ia .
for example , a non - detection of one year after explosion with the same observational limit as from our stacked 1.66 ghz observations would , according to our model , constrain @xmath162 to @xmath223 , assuming @xmath224 , @xmath225 , @xmath178 , @xmath115 and @xmath226 .
the supernova shock wave will at this point be located at @xmath227 cm
. a non - detection at such late epochs and such low flux levels , will certainly be very useful in constraining @xmath58 .
the recent and nearby sne 2011fe and 2014j have offered a remarkable possibility to learn about the origin of sne ia .
we have shown here that deep radio observations can be used to rule out several progenitor models for and a similar discussion was made for sn 2011fe by @xcite .
in addition to this , @xcite provided deep limits for sn 2011fe from x - rays ( cf .
[ sec : intro ] ) , which do not depend on @xmath58 , but where the limit on circumstellar density has a stronger dependence on @xmath57 and @xmath64 than for radio emission . while @xmath228 has been used by us and others ,
it must be cautioned that , e.g. , sn 1993j had a much lower value @xcite , although it is clear that this supernova bears little resemblance with type ib / ic sne , which have been used as templates to model sne ia . for sn 2011fe , the non - detections in radio and x - rays were accompanied with no circumstellar line absorption @xcite and a non - detection of late nebular emission from gas ablated off an sd companion @xcite .
this , together with other evidence for sn 2011fe ( see * ? ? ?
* ) , has been used to argue for an increased likelihood of sne ia being the endpoint of a dd scenario rather than sd scenarios .
our non - detections of radio emission from in principle add to this evidence .
however , @xcite suggested that a sd scenario , with a spun - up / spun - down super - chandrasekhar wd ( see also * ? ? ?
* ; * ? ? ?
* ) , can still be possible if the donor star shrinks far inside its roche lobe prior to the explosion .
this would make the sd companion smaller and more tightly bound , and only very dilute circumstellar gas would be expected in the immediate vicinity of the wd .
@xcite argue that density could be of the same order as typical interstellar densities .
continued radio observations of both sne 2011fe and 2014j could be useful to test the presence of such low - density gas ( cf .
[ sec : dd ] ) . for a typical time - scale of @xmath229 years between last roche - lobe overflow and explosion @xcite , and a wind speed of @xmath1 , the last traces of substantial circumstellar gas could in this scenario
be at a distance of @xmath230 cm , and could explain the presumed shell around , e.g. , sn 2006x @xcite . if the supernova ejecta would start to interact with such a shell , radio emission would increase .
radio observations of sn 2006x @xcite two years after explosion , however , failed to detect any emission . @xcite
estimated a shell radius of @xmath231 cm , which according to the estimate in
[ sec : sd ] , was most likely overtaken by the supernova ejecta by @xmath232 years .
continued monitoring of would be useful to trace these putative shells .
we emphasize that shells around supernovae do not have to lie along the line of sight to be detected in radio , as opposed to the narrow absorption line features .
this should increase the possibility to detect such shells in the radio , especially if shells are as common as suggested ( see [ sec : obs ] ) .
interaction with shells are also better observed in the radio than in x - rays , as no inverse compton scattering is expected at late times when the supernova has faded and the distance from the line emitting sn ejecta and the shell is large .
@xcite caution that the lack of signatures from an sd companion could be a problem for the model of @xcite , as only @xmath233 of ablated mass from the companion can be accommodated by the observations of @xcite , in combination with an extrapolation of the models presented in @xcite and @xcite , before giving rise to detectable h@xmath15 emission in the nebular phase ; all models calculated by @xcite , @xcite and @xcite predict more than @xmath234 of ablated mass . a way to avoid h@xmath15
emission is , of course , if the donor is he - rich .
the models of @xcite show that of order @xmath235 of he - rich gas would then reside in the centre of the supernova in the nebular phase .
however , even if the donor is h - rich , the caution by @xcite should not be over interpreted .
the opacity in the nebular emission models of @xcite and @xcite does not contain as many spectral lines as more recent models by , e.g. , @xcite .
scattering in the spectral region around h@xmath15 could be more severe than previously anticipated , and the constraint from lack of nebular h@xmath15 less important .
nebular lines further out in the red should suffer from less scattering , and the models discussed in @xcite show that narrow ( @xmath236 ) [ ca ii ] lines are present in the red , and could thus be more constraining than h@xmath15 .
traces of these lines could be useful to test scenarios with both h- and he - rich donors .
further detailed modeling of the nebular phase is indeed needed , as well as modeling of supernova ejecta colliding with compact companions such as those in the models of @xcite .
deep nebular spectra of sn 2014j are warranted to test those impact models , due to the sn proximity .
a potential problem is , however , the extinction toward sn 2014j and the complicated interstellar imprint on the supernova spectrum @xcite .
the latter may , in particular , make the search for narrow - line variations more cumbersome than for , e.g. , sn 2006x . at the moment ,
our deepest radio limits on circumstellar gas are for sne 2011fe and 2014j . with the advent of the square kilometre array ( ska )
, we will be able to obtain significantly deeper radio limits ( or , potentially , a detection ) for sne ia exploding at the distance of m 82 . for more distant supernovae
, we will obtain similar limits to those obtained for sne 2011fe and 2014j , which will allow us to build a picture from a larger statistical sample .
the first phase of ska considers three different components .
one of them , ska1-mid , promises to yield 1@xmath123 sensitivities of @xmath237jy / b in one hour at a fiducial frequency of 1.7 ghz .
this figure is five times better than currently provided by the most sensitive array , the jvla .
therefore , ska1-mid should be able to either detect the putative radio emission of sn 2014j - like objects up to distances @xmath238 mpc in less than one hour , or put significantly better constraints on some of the parameter space of sd scenarios for the next sn ia that explodes in m 82 , some of which could not be completely ruled out even by our very deep radio observations .
however , the expected number of sn ia per year in such a volume of the local universe is small . indeed
, since the volumetric sn ia rate is @xmath239 sn / yr / mpc@xmath240 @xcite , we should expect on average one sn ia every @xmath1015 yr within a distance of @xmath2418 mpc , which is a small value . to obtain a statistically significant sample of sne ia observed in radio , with similar upper limits to those obtained by us for sn 2014j
, we need to sample significantly larger volumes and need much more sensitive radio observations .
for example , by sampling out to a distance of 25 mpc , we can expect @xmath102 sne ia per year within the sampled volume , which in 10 years would result in a total of @xmath1020 sne ia , enough to extract statistical results . at this maximum distance
, we need a sensitivity of @xmath1050 times better than obtained by the observations discussed here , to be as constraining , or @xmath242 njy / b .
when ska is completed , the fiducial 1@xmath123 sensitivity should be 10 times better than for ska1-mid , or about @xmath243 njy / b in one hour , and such statistical studies will be perfectly possible in short amounts of time . at this level of sensitivity , a non - detection would be essentially as meaningful as a direct detection , since the former would imply that only the dd scenario is viable , while the latter would tell us which of the sd channels result in sne ia .
we report deep eevn and emerlin radio observations of the type ia sn 2014j in the nearby galaxy m 82 , along with a detailed modeling of its radio emission .
our observations result in non - detections of the radio emission from sn 2014j . yet , radio data and modeling allow us to place a tight constraint on the mass loss rate from the progenitor system of sn 2014j .
namely , if the exploding wd was surrounded by a wind with a density profile @xmath4 , as expected for a sd scenario , then our upper limit to the mass - loss rate is @xmath244 , for a wind speed of @xmath1 .
if , on the contrary , the circumstellar gas has a constant density , as expected to be the case for the dd scenario ( but also in a small region of the parameter space of sd scenarios ) , then our modeling yields an upper limit on the gas density , such that @xmath2 . our stringent upper limits to the circumstellar density around sn 2014j
allow us to exclude completely symbiotic systems and the majority of the parameter space associated with stable nuclear burning wds , as viable progenitor systems for sn 2014j .
for the case of recurrent novae with main sequence or subgiant donors , we can not rule out them completely , yet most of their parameter space is also excluded by our observations for the standard assumption of @xmath121 , where @xmath58 is the ratio of magnetic energy density to post - shock thermal energy density .
we have also reassessed the radio limits on wind density for sn 2011fe , and for @xmath121 we obtain @xmath245 ( for a wind speed of @xmath1 ) and @xmath246 .
these limits are close to those calculated by @xcite .
our limit on @xmath247 for is thus similar to that for sn 2011fe , whereas for the constant density case we obtain a much lower limit than for sn 2011fe , and hence the lowest limit for a constant density ambient around a sn ia .
the combined radio limits on circumstellar gas around sne 2011fe and 2014j add to evidence from mainly non - detections of x - rays from sn 2011fe and 2014j@xcite and no detection of h@xmath15 in the nebular phase of sn 2011fe @xcite , that sne ia are very likely to stem from the dd scenario , rather than sd scenarios . finally , we highlight future observations with the square kilometre array ( ska ) . when fully completed , the ska is likely to yield limits on circumstellar gas for future sne ia similar to the limits reported here for the nearby sne 2011fe and 2014j , but for distances well beyond the virgo cluster . for nearby sne
ia , ska limits are likely to be fully conclusive regarding the origin of the progenitor systems of sne ia .
we are grateful to carles badenes for useful comments on the manuscript , to rubn herrero - illana for pythonic advice to produce figure 5 , and to vicent peris and oscar brevi from the observatorio astronmico de aras ( valencia , spain ) for the optical image of m82 used in figure 1 .
we acknowledge the emerlin and evn programme committees , and the directors of the evn stations , for supporting the radio observations of sn2014j .
the european vlbi network ( evn ) is a joint facility of european , chinese , south african , and other radio astronomy institutes funded by their national research councils .
the electronic multi - element radio linked interferometer network ( emerlin ) is the uk s facility for high resolution radio astronomy observations , operated by the university of manchester for the science and technology facilities council ( stfc ) .
the research leading to these results has received funding from the european commission seventh framework programme ( fp/2007 - 2013 ) under grant agreement no 283393 ( radionet3 ) .
aa , jcg , jmm , mapt , er , and imv acknowledge support from the spanish micinn through grants aya2012 - 38491-c02 - 01 and aya2012 - 38491-c02 - 02
. p.l . acknowledges support from the swedish research council .
the research leading to these results has received funding from the european commission seventh framework programme ( fp/2007 - 2013 ) under grant agreement no 283393 ( radionet3 ) .
bjrnsson , c .- i .
, & lundqvist , p. 2014
, , 787 , 143 branch , d. , livio , m. , yungelson , l. r. , boffi , f. r. , & baron , e. 1995 , , 107 , 1019 cao , y. , kasliwal , m. m. , mckay , a. , & bradley , a. 2014 , atel , 5786 , 1 chandler , c. j. , & marvil , j. 2014 , atel , 5812 chandra , p. , chevalier , r. , & patat , p. 2008
, atel , 1391 , 1 chevalier , r. a. 1982a , , 258 , 790 chevalier , r. a. 1982b , , 259 , 302 chevalier , r. a. 1998 , , 499 , 810 chevalier , r. a. , & fransson 2006 , , 651 , 381 chomiuk , l. , soderberg , a. m. , moe , m. , et al .
2012 , , 750 , 164 chugai , n. n. 2008 , astronomy letters , 34 , 389 chugai , n. n. , chevalier , r. a. , & lundqvist , p. 2004
, , 355 , 627 cumming , r. j. , lundqvist , p. , smith , l. j. , pettini , m. , & king , d. l. 1996 , , 283 , 1355 dilday , b. , smith , m. , bassett , b. , et al .
2010 , , 713 , 1026 dilday , b. , howell , d. a. , cenko , s. b. , et al .
2012 , science , 337 , 942 di stefano , r. , voss , r. , & clayes , j. s. w. 2011 , , 738 , l1 dwarkadas , v. v. , & chevalier , r. a. 1998 , , 497 , 807 fink , m. , kromer , m. , seitenzahl , i. r. , et al .
2014 , mnras , 438 , 1762 fossey , j. , cooke , b. , pollack , g. , wilde , m. , & wright , t. 2014 , cbet , 3792 , 2 fransson , c. , & bjrnsson , c .- i .
1998 , , 509 , 861 goobar , a. , johansson , j. , amanullah , r. , et al .
2014 , , 784 , l12 hachisu , i. , kato , m. , & nomoto , k. 1999 , , 522 , 487 hachisu , i. , kato , m. , & nomoto , k. 2008 , , 679 , 1390 hachisu , i. , kato , m. , saio , h. , & nomoto , k. 2012 , apj , 744 , 69 hancock , p. j. , gaensler , b. m. , & murphy , t. 2011 , , 735 , 35 horesh , a. , kulkarni , s. r. , fox , d. b. , et al .
2012 , , 746 , 21 hughes , j. p. , chugai , n. , chevalier , r. , lundqvist , p. , & schlegel , e. 2007 , , 670 , 1260 itagaki , k. , kaneda , h. , yamaoka , h. , et al .
2014 , cbet , 3792 , 1 jerkstrand , a. , fransson , c. , & kozma , c. 2011 , , 535 , 45 justham , s. 2011 , , 730 , l34 liu , z. w. , pakmor , r. , rpke , f. k. , et al .
2012 , , 548 , a2 liu , z. w. , pakmor , r. , seitenzahl , i. r. , et al .
2013 , , 774 , 37 lundqvist , p. , & fransson , c. 1988 , , 192 , 221 lundqvist , p. , mattila , s. , sollerman , j. , et al
2013 , , 435 , 329 maoz , d. , mannucci , f. , & nelemans , g. 2014 , , in press ; arxiv1312.628 m margutti , r. , parrent , j. , kamble , a. , et al .
2014 , , submitted , arxiv:1405.1488 margutti , r. , soderberg , a. m. , chomiuk , l. , et al .
2012 , , 751 , 134 marietta , e. , burrows , a. , & fryxell , b. 2000 , apjs , 128 , 615 mart - vidal , i. , marcaide , j. m. , alberdi , a. , et al .
2011 , , 526 , a143 mattila , s. , lundqvist , p. , sollerman , j. , et al .
2005 , , 443 , 649 matzner , c.d . , &
mckee , c.f .
, , 510 , 379 nomoto , k. , saio , h. , kato , m. , & hachisu , i. 2007 , , 663 , 1269 pan , k .- c . ,
ricker , p. m. , & taam , r. e. 2012 , , 750 , 151 panagia , n. , van dyk , s. d. , weiler , k. w. , et al .
2006 , , 646 , 369 patat , n. , chandra , p. , chevalier , r. , et al .
2007 , science , 317 , 924 patat , n. , cordiner , m. a. , cox , n. l. j. , et al .
2013 , , 549 , 62 prez - torres , m. a. , alberdi , a. , & marcaide , j. m. 2001 , , 374 , 997 perlmutter , s. , aldering , g. , goldhaber , g. , et al .
1999 , , 517 , 565 readhead , a. c. s. 1994 , , 426 , 51 riess , a. , filippenko , a. v. , challis , p. , et al . , 1998 , , 116 , 1009 rybicki , g. b. , & lightman ,
, radiative processes in astrophysics ( new york : wiley ) russell , b. r. , & immler , s. 2012 , , 748 , l29 shappee , b. j. , kochanek , c. s. , & stanek , k. z. 2013a , , 765 , 150 shappee , b. j. , stanek , k. z. , pogge , r. w. , & garnavich , p. m. 2013b , , 762 , l5 shen , k. j. , & bildsten , l. 2007 , , 660 , 1444 shen , k. j. , & bildsten , l. 2009 , , 692 , 324 simon , j. d. , gal - yam , a. , gnat , o. , et al .
2009 , , 702 , 1157 smith , l. j. , strolger , l. , mutchler , m. , ubeda , l. , & levay , k. 2014 , atel , 5821 , 1 soaker , n. 2014 , arxiv:1405.0173 sternberg , a. , gal - yam , a. , simon , j. d. , et al .
2014 , , submitted , arxiv:1311.3645s truelove , j. k. , & mckee , c. e. , 120 , 299 welty , d. e. , ritchey , a. m. , dahlstrom , j. a. , & york , d. g. 2014 , , submitted , arxiv:1404.2639 yaron , o. , prialnik , d. , shara , m. m. , & kovetz , a. , 623 , 398 zheng , w. , shivvers , i. , filippenko , a. v. , et al .
2014 , , 783 , l24 zwaan , m. , walter , f. , ryan - weber , e. , et al .
2008 , , 136 , 2886 | we report deep evn and emerlin observations of the type ia sn 2014j in the nearby galaxy m 82 .
our observations represent , together with jvla observations of sne 2011fe and 2014j , the most sensitive radio studies of type ia sne ever . by combining data and a proper modeling of the radio emission , we constrain the mass - loss rate from the progenitor system of sn 2014j to @xmath0 ( for a wind speed of @xmath1 ) . if the medium around the supernova is uniform , then @xmath2 , which is the most stringent limit for the ( uniform ) density around a type ia sn .
our deep upper limits favor a double - degenerate ( dd ) scenario involving two wd stars for the progenitor system of sn 2014j , as such systems have less circumstellar gas than our upper limits .
by contrast , most single - degenerate ( sd ) scenarios , i.e. , the wide family of progenitor systems where a red giant , main - sequence , or sub - giant star donates mass to a exploding wd , are ruled out by our observations ( for a wind speed of @xmath1 ) and @xmath3 , for the @xmath4 wind and constant density cases , respectively .
as these limits are less constraining than ours , the findings by @xcite do not alter our conclusions .
the x - ray results are , however , important to rule out free - free and synchrotron self - absorption as a reason for the radio non - detections . ] .
our estimates on the limits to the gas density surrounding sn2011fe , using the flux density limits from @xcite , agree well with their results .
although we discuss possibilities for a sd scenario to pass observational tests , as well as uncertainties in the modeling of the radio emission , the evidence from sne 2011fe and 2014j points in the direction of a dd scenario for both . |
despite many years of dedicated research , the progenitor systems and explosion mechanisms of sne ia remain unclear .
it is generally accepted that sne ia originate from thermonuclear explosions of massive carbon - oxygen white dwarfs in binary systems . depending on the nature of the companion star , two different progenitor systems have been proposed . in the _ single degenerate _
scenario @xcite a carbon - oxygen white dwarf accretes from a non - degenerate companion star until it reaches the chandrasekhar - mass and explodes ( but note that an explosion before reaching the chandrasekhar mass may also be possible , e.g. * ? ? ?
* ; * ? ? ?
in contrast , in the _ double degenerate _
scenario @xcite a merger of two carbon - oxygen white dwarfs causes a thermonuclear explosion of the merged system .
many recent findings including sn rates from population synthesis studies @xcite , studies of the delay time distribution of observed sne ia ( e.g. * ? ? ?
* ) , the lack of radio emission of sne ia ( e.g. * ? ? ?
* ) , the lack of hydrogen emission in nebular spectra of sne ia @xcite , the lack of x - ray emission in elliptical galaxies @xcite and studies of sn ia remnants ( e.g. * ? ? ?
* ) seem to favor the double degenerate scenario .
in addition , first studies of sn 2011fe seem to disfavor a single degenerate progenitor ( see , e.g. * ? ? ? * ) .
there is , however , no unambiguous proof for any progenitor scenario yet . for a detailed discussion about constraints on the progenitor scenarios see @xcite . to date , most theoretical work on sne ia has concentrated on the single degenerate scenario , as mergers of white dwarfs were thought not to lead to thermonuclear explosions .
this was mainly based on the picture that mergers would leave behind the more massive white dwarf with a hot envelope made up of the material of the less massive white dwarf .
fast accretion from the envelope will then turn the carbon - oxygen white dwarf into an oxygen - neon white dwarf @xcite which collapses to a neutron star as it approaches the chandrasekhar mass @xcite .
ways to avoid the transformation of the carbon - oxygen white dwarf into an oxygen - neon white dwarf have been proposed ( e.g. * ? ? ?
* ) , but without conclusive results . recently , however , we demonstrated that violent mergers of two carbon - oxygen white dwarfs could directly lead to a thermonuclear explosion while the merger is still ongoing @xcite .
we also showed that the observables for such an explosion with two white dwarfs of @xmath4 show good agreement with the observed properties of subluminous 1991bg - like sne ia .
furthermore , we found that for a primary mass of @xmath4 a mass ratio of at least about @xmath5 is required for the scenario to work @xcite .
lately , @xcite showed that using exact initial conditions can change the properties of the merger .
in particular , this leads to a significantly longer inspiral in their simulations .
however , @xcite were only able to run the simulation with a resolution of @xmath6 particles ( for comparison , the violent merger calculations by @xcite used @xmath7 particles ) . in this work ,
we combine high resolution merger simulations with exact initial conditions .
we present the results of a simulation of the massive merger of a @xmath8 and a @xmath4 carbon - oxygen white dwarf .
we follow the evolution of the binary system through the merger phase , thermonuclear explosion and nucleosynthesis .
finally we use three - dimensional radiative transfer simulations to obtain synthetic lightcurves and spectra .
the inspiral and merger is modeled using a modified version of the gadget code @xcite .
modifications include the helmholtz equation of state @xcite and a 13 isotope nuclear reaction network that contains all @xmath9-elements from @xmath10 to @xmath3 .
radiative cooling effects are not included in our simulation .
a detailed description of the modifications will be given in a forthcoming paper .
in addition the maximum smoothing length of a particle was restricted to @xmath11 .
this affects only particles ejected from the binary system during the merger but leads to a significant speedup of the code . since these particles are at very low densities and contain only less than one percent of the total mass , they have no noticable influence on the explosion dynamics and observables .
the initial binary system consists of a @xmath8 and a @xmath4 carbon - oxygen white dwarf constructed from a total of @xmath12 equal - mass particles .
both white dwarfs are set up in isolation and relaxed with an additional friction force for @xmath13 .
we then apply the method described in @xcite to slowly move the two white dwarfs close together .
when the first particle of the less massive white dwarf crosses the inner lagrange - point we stop and start the actual simulation . at this time , the binary system has an orbital period of about @xmath14 . after the explosion of a slice in the x - z plane . the mass fraction is shown for carbon , oxygen , silicon , stable iron and radioactive @xmath3 . ]
the evolution of the binary system is shown in fig .
[ fig : merger ] .
the mass transfer is stable for more than @xmath15 orbits .
after about @xmath16 the secondary white dwarf becomes dynamically unstable and is disrupted on a timescale of one orbit .
as the material of the secondary is accreted violently onto the primary , material is compressed and heated up on the surface of the primary white dwarf . as a consequence hotspots form in which carbon burning is ignited .
when the first hotspot reaches a temperature larger than @xmath17 at a density of about @xmath18 we assume that a detonation forms @xcite .
note that despite the high resolution we use , we still tend to underestimate the temperature in the hotspot compared to even higher resolution simulations @xcite .
only future detailed investigations might be able to decide whether or not a detonation really forms but the conditions in our sph simulations suggest that it is plausible . at this time we map the whole state of the simulation on a uniform cartesian grid with a resolution of @xmath19 grid cells and a box size of @xmath20 .
about @xmath21 of material is lost by the mapping as it is outside the box . since this material makes up less than two percent of the total mass and has a density too low to contribute significantly to nuclear burning ( i.e. it stays unburned ) it does not affect the dynamics of the ejecta or the synthetic observables derived from the model . with these initial conditions we use the leafs code @xcite that applies the level - set technique to model detonation
flames @xcite .
we ignite the detonation at the cell with the highest temperature and follow its propagation through the merged object until most of the material is burned .
[ fig : merger ] shows that the primary white dwarf and most parts of the secondary are already burned @xmath22 after the detonation formed .
the energy release from nuclear burning unbinds the object . using an expanding grid @xcite we follow the dynamic ejecta until they reach homologous expansion about @xmath13 after the detonation was ignited .
they have an asymptotic kinetic energy of @xmath23 .
in order to obtain detailed isotopic abundances of the ejecta we add @xmath24 tracer particles to the simulation of the explosion that record their local temperature and density . in a post - processing step
we run a detailed nuclear network containing @xmath25 isotopes on these trajectories @xcite . to mimic the effect of solar metallicity we choose the initial composition of the tracer particles as @xmath26 , @xmath27 and @xmath28 by mass . in the explosion ,
a total of @xmath29 of iron group elements are synthesized .
they consist predominantly of radioactive @xmath30 ( @xmath31 ) with a small fraction of stable @xmath32 ( @xmath33 ) and stable @xmath34 ( @xmath35 ) .
in addition , @xmath36 of intermediate mass elements are produced in the explosion .
the ejecta contain about @xmath37 of oxygen and about @xmath38 of unburned carbon .
the spatial composition and density structure of the ejecta in homologous expansion are shown in fig .
[ fig : comp ] .
its inherently three - dimensional structure is a result of the state of the merged object at the time the detonation forms .
since the detonation propagates faster at higher densities , the primary white dwarf is burned first and its ashes expand .
as they expand they sweep around the material of the partially intact secondary white dwarf that is still being burned .
therefore , the ashes of the primary have already expanded significantly before burning of the secondary is completed ( roughly one second later ) , meaning that the ashes of the secondary completely dominate the center of the ejecta .
hence , when the ejecta reach the homologous expansion phase several ten seconds later , the very inner parts of the ejecta do not contain material from the primary white dwarf and therefore no iron - group elements .
lcccccc @xmath39 & 18.6 & 18.0 & 20.8 & 23.2 & 21.6 + @xmath40 & 0.74 & 1.30 & 0.95 & 0.67 & 0.42 + @xmath41 & -19.2 & -19.6 & -19.0 & -19.2 & -19.2 + @xmath42 & -19.1 & -19.5 & -19.0 & -19.2 & -19.2 + using the three - dimensional density structure from explosion modeling and the corresponding spatial abundance distribution from the tracer particles we apply the monte carlo radiative transfer code artis @xcite to calculate synthetic spectra and lightcurves for our model .
[ fig : lightcurves ] shows angle - averaged and line - of - sight dependent lightcurves of our model compared with several well - observed normal sne ia .
the angle - averaged lightcurves show good agreement with observations in the optical bands and reproduce general properties of normal sne ia having a b - band rise time of @xmath43 days , a peak brightness of @xmath44 , @xmath45 , and @xmath46 in the u , b , and v band , respectively , and a light curve decline rate in the b band of @xmath47 ) .
these values are well within the range of normal sne ia @xcite .
more detailed values for selected bands of the angle - averaged light curves are given in table [ tab : lc_paras ] .
in particular , the angle - averaged lightcurves in the @xmath48 and @xmath49 band agree very well with those observed for normal sne ia .
the angle - averaged synthetic @xmath50-band lightcurve is about @xmath51 magnitudes fainter at maximum than the three supernovae we compare with and slightly brighter from @xmath52 days after maximum onwards .
the angle - averaged @xmath53-band lightcurve of our model agrees well with the data up to about @xmath54 days after maximum , but declines slightly more slowly afterwards . due to the asymmetric ejecta structure ,
our model lightcurves show a non - negligible sensitivity to line - of - sight effects . in the @xmath50 band , for example
, we find peak magnitudes between -19.5 and -18.7 . at the same time @xmath55
varies between 0.5 and 1.4 , roughly along the phillips relation @xcite . for the @xmath48 band ,
the scatter is even larger while it decreases in the redder bands . for lower photon energies ,
the asymmetry of the ejecta is less important , since the optical depths are smaller and photons probe a larger fraction of the total ejecta , thereby making the observables less sensitive to line - of - sight effects ( compare @xcite ) .
the angle - averaged lightcurve of our model in the @xmath56 band agrees well with the observations up to maximum light .
afterwards it is between @xmath57 and @xmath58 magnitudes brighter than our comparison sne ia .
this offset ( in the @xmath56 band after maximum ) can be attributed to a flux excess in the caii nir triplet in the synthetic spectra .
this could point to an over - abundance of calcium in the model .
however , it is more likely to be a shortcoming of our radiative transfer treatment , which uses the simple van regemorter approximation @xcite to treat collisional excitation .
thus we likely underestimate the effectiveness of cooling by forbidden transitions of heavy elements and therefore predict that too much cooling occurs via the strong dipole transitions of caii .
there is good agreement between model and data in the nir bands ( @xmath59,@xmath60,@xmath61 ) . in particular ,
our model reproduces time and brightness of the second peaks and even the light curve decline at more than @xmath62 days after maximum quite accurately .
however , high precision modeling of the nir lightcurves requires an extensive atomic data set to properly simulate flux redistribution by fluorescence @xcite . here , however , we have restricted ourselves for computational reasons to a simplified atomic data set ( cd23_gf-5 of @xcite ) with only @xmath63 lines .
this has been shown to give reliable results in the optical bands but slightly underpredicts the nir brightness @xcite between the primary and secondary peaks .
[ fig : spectrum ] shows the angle - averaged spectrum of our model one day after b - band maximum . for comparison
we overplot an observed spectrum of the normal sn 2003du @xcite at the same epoch .
our model shows most of the characteristic features of sne ia , particularly the defining siii doublet at @xmath646347,6371 but also the weaker siii features at @xmath645958,5979 and @xmath654128,4131 , the mgii triplet at @xmath664481 and the caii h and k absorption .
the sii w - feature at @xmath675400 ( which is a blend of several lines ) is also visible though it is weaker than in sn 2003du . in the red tail of the spectrum
the caii nir triplet is clearly visible as well as some indication of the oi triplet @xmath647772,7774,7775 .
moreover , the overall spectral shape and the velocity - shifts of most of the line features in the observed spectrum are well reproduced , indicating that the velocity structure of our model is a good representation of that in real sne ia .
the dominant parameter that determines the brightness of the explosion in our model is the mass of the primary white dwarf . in the merger
the primary remains mostly unaffected whereas the secondary white dwarf is destroyed .
therefore , because the density profile of the primary white dwarf to first order only depends on its mass and the density stays high enough to burn to iron - group elements ( @xmath3 ) only in the primary white dwarf , the brightness of the supernova directly correlates with the mass of the primary white dwarf .
physical parameters of secondary importance are the mass of the secondary white dwarf and the composition of both white dwarfs ( c / o - ratio and metallicity ) .
the material of the secondary is burned only to intermediate - mass elements and oxygen and therefore only affects the brightness of the explosion indirectly .
the metallicity of the two white dwarfs has only a small impact on the final composition of the ejecta by changing the mixture of iron - group elements synthesized @xcite .
the moment at which the detonation forms is an artificial parameter of our model .
physically it will be determined by the properties of the binary system discussed above .
since we can not resolve the formation of the detonation microscopically , however , we have to infer the most likely place and time for it to form from macroscopic properties .
once we assume that a detonation forms on the surface of the primary its exact place and time are only secondary effects : the choice for the moment of detonation affects the densities at which the material of the secondary white dwarf is burned ( since the detonation occurs while the secondary is being destroyed ) but it does not strongly affect the burning of the primary because it is not dynamically affected by the merger .
compared to the merger of two @xmath68 white dwarfs @xcite the merger we present here mainly differs in the mass of the primary white dwarf .
since it is considerably more massive , its central density is higher and its material is burned predominantly to iron group elements .
in addition , the mass ratio of the binary system we model here is well below unity .
therefore , the more compact primary is not subject to noticeable tidal forces from the secondary .
in addition , although exact initial conditions may weaken the merger , we show that it still leads to the formation of hot spots on the surface of the primary , in which densities and temperatures are sufficiently high that a detonation is plausible .
this is consistent with the properties of hot spots seen in previous studies with less relaxed initial conditions @xcite and also with very recent simulations by @xcite who found similar hot spots in calculations with relaxed initial conditions and only slightly smaller resolution than our simulation .
note that @xcite do not find hotspots in similar systems with exact initial conditions but more than a factor of ten fewer particles .
our model can be interpreted as a pure detonation of an isolated sub - chandrasekhar mass white dwarf as described in @xcite with an additional component from the remains of the secondary white dwarf surrounding the system .
quantitatively , however , the color lightcurves of our model agree significantly better with observations than the toy models of @xcite .
in particular , our model also shows good agreement in the nir - bands and reproduces the position of the secondary peaks in these bands very well .
this difference to the toy models is likely to be associated with a different density profile and composition structure in the inner parts of the ejecta . as discussed in section [ sec : nucleosynthesis ] this is a result of the presence of the secondary white dwarf and its interaction with ashes of the primary when they expand . also , the asymptotic kinetic energy per mass of the ejecta is smaller for our merger model than for detonations of isolated sub - chandrasekhar mass white dwarfs , because the burning of the secondary white dwarf is less complete and hence the average energy release per mass is smaller .
overall , our explosion reproduces observational data of normal sne ia of the same brightness reasonably well . since the time span from the onset of mass transfer between the two white dwarfs until the ejecta reach homologous expansion is very short ( only of the order of a few minutes ) there is no extended outer envelope the explosion ejecta interact with later than a few minutes after the explosion , in contrast to the model by @xcite .
observational constraints @xcite on the number of double white dwarf binaries significantly more massive than a chandrasekhar mass are still not very restrictive .
if violent mergers contribute noticeably to the total sne ia rate , our calculations suggest that it will be hard to distinguish the merger events from the bulk of normal sne ia via optical / nir lightcurves and optical spectra .
further detailed work is therefore needed to investigate whether there are other characteristic properties of sn ia explosions from violent white dwarf mergers and to explore how the observational display changes with the properties of their progenitor system .
this work was supported by the deutsche forschungsgemeinschaft via the transregional collaborative research center trr 33 `` the dark universe '' , the excellence cluster exc153 `` origin and structure of the universe '' and the emmy noether program ( ro 3676/1 - 1 ) .
parts of the simulations were carried out at the john von neumann institute for computing ( nic ) in jlich , germany ( project hmu14 ) .
r. p. gratefully acknowledges financial support of the klaus tschira foundation . | one of the most important questions regarding the progenitor systems of type ia supernovae ( sne ia ) is whether mergers of two white dwarfs can lead to explosions that reproduce observations of normal events . here
we present a fully three - dimensional simulation of a violent merger of two carbon - oxygen white dwarfs with masses of @xmath0 and @xmath1 combining very high resolution and exact initial conditions .
a well - tested combination of codes is used to study the system .
we start with the dynamical inspiral phase and follow the subsequent thermonuclear explosion under the plausible assumption that a detonation forms in the process of merging .
we then perform detailed nucleosynthesis calculations and radiative transfer simulations to predict synthetic observables from the homologously expanding supernova ejecta .
we find that synthetic color lightcurves of our merger , which produces about @xmath2 of @xmath3 , show good agreement with those observed for normal sne ia in all wave bands from u to k. line velocities in synthetic spectra around maximum light also agree well with observations .
we conclude , that violent mergers of massive white dwarfs can closely resemble normal sne ia .
therefore , depending on the number of such massive systems available these mergers may contribute at least a small fraction to the observed population of normal sne ia . |
cold gases of rydberg atoms are currently receiving a growing attention in the communities of quantum optics @xcite , quantum information @xcite , and many - body physics @xcite .
this is rooted in the fact that they offer strong and long - ranged interactions and at the same time grant long coherent lifetimes
. currently , considerable efforts are devoted to developing all - optical quantum information protocols @xcite with the rydberg - atom - mediated interaction between individual photons @xcite .
fundamentally important optical devices that operate on the single photon level , such as phase shifters @xcite , switches @xcite and transistors @xcite , have been demonstrated experimentally in rydberg gases .
single photon switchs might form a central building block of an all - optical quantum information processor @xcite .
the prime function of such switches is to control the transmission of an incoming photon through a single gate photon .
one promising way to realize this is to store the gate photon in form of a gate ( rydberg ) atom immersed in an atomic gas which is in a delocalized spinwave state @xcite .
the gate atom then prevents transmission of incident photons through the gas , while ideally the coherence of the rydberg spinwave state is preserved @xcite .
the latter property would permit the subsequent coherent conversion of the rydberg spinwave into a photon which would pave the way for gating the switch with superposition states that can also be subsequently retrieved .
currently , there is only a basic understanding of how the coherence of the rydberg spinwave might be affected by the scattering of incoming photons and no systematic study of this important question exists . ,
excited state @xmath0 ( decay rate @xmath1 ) and rydberg state @xmath2 are resonantly coupled by a single photon field @xmath3 ( with collective coupling strength @xmath4 ) and a classical field of rabi frequency @xmath5 .
initially a gate photon is stored as a spinwave in the rydberg state @xmath6 ( indicated by the green circle ) .
( b , c ) polarization profiles @xmath7 for a spinwave consisting of two possible gate atom positions @xmath8 ( @xmath9 ) and their dependence on the blockade radius @xmath10 and the system length @xmath11 .
( b ) for @xmath12 and @xmath13 the polarization profiles associated with the two gate atom positions are distinguishable .
( c ) when @xmath14 the polarization profile is independent of the gate atom position which leads to enhanced coherence of the stored spinwave . ] in this work we address this outstanding issue within a simple model system .
we study the propagation of a single photon under conditions of electromagnetically induced transparency ( eit ) in a cold atomic gas in which a gate photon is stored as a rydberg spinwave .
an incident photon subsequently experiences a rydberg mediated van der waals ( vdw ) interaction with this stored gate atom which lifts the eit condition and renders the atomic medium opaque . in this case
the incident photon is scattered incoherently off the rydberg spinwave .
we study the photon propagation and explore the dependence of rydberg spinwave coherence on the interaction strength ( parameterized by the blockade radius @xmath15 ) , the system length @xmath11 and bandwidth of the incident photon pulse .
our findings confirm that strong absorption , i.e. high gain , can be achieved already for large systems ( @xmath12 ) while coherence of the spinwave is preserved only for sufficiently strong interactions , i.e. @xmath14 . intuitively ,
this can be understood by regarding the scattering of the incoming photon as a measurement of the position of the gate atom . when @xmath14 this measurement is not able to resolve the position of the excitation and hence coherence of the rydberg spinwave is maintained .
our study goes beyond this simple consideration by taking into account propagation effects , a realistic interaction potential and a finite photon band width .
the results can therefore be considered as upper bounds for the fidelity with which a rydberg spinwave can be preserved and re - converted into a photon in an experimental realization of a coherent cold atom photon switch .
the paper is organized as follows . in section
ii , we introduce a one - dimensional model system to study the propagation dynamics of single source photons in the atomic gas prepared in a rydberg spinwave state . in sec .
iii , the model system is solved numerically with realistic parameters .
we identify the working regime for a single photon switch where the source photon is scattered completely . in sec .
iv , we numerically study the fidelity between the initial spinwave state and the final state after the source photon is scattered .
our calculation shows that the coherence of the spinwave is preserved when @xmath16 while the final state becomes a mixed state when @xmath17 . in sec .
v , we provide analytical results for a coherent single photon switch ( @xmath18 ) .
we reveal that the transmission and switch fidelity depend nontrvially on the optical depth and bandwidth of the source photon field .
we summarize in sec . vi .
our model system is a one - dimensional , homogeneous gas consisting of @xmath19 atoms , whose electronic levels are given in fig .
[ fig : illustration]a . the photon field @xmath20 and the eit control laser ( rabi frequency @xmath5 ) resonantly couple the groundstate @xmath21 with the excited state @xmath0 and @xmath0 with the rydberg state @xmath2 .
following ref .
@xcite , we use polarization operators @xmath22 and @xmath23 to describe the slowly varying and continuum coherence of the atomic medium @xmath24 and @xmath25 , respectively .
all the operators @xmath26 are bosons and satisfy the equal time commutation relation , @xmath27=\delta(z - z')$ ] . initially , the atoms are prepared in a delocalized spinwave state with a single gate atom in the rydberg state @xmath6 , @xmath28 is the wavenumber of the spinwave and @xmath29 abbreviates many - body basis with the gated atom located at position @xmath30 and the rest in the groundstate .
the rydberg spinwave state is created routinely in experiments @xcite .
when interacting with the incoming single photon , the general many - body state of this one - dimensional system is expanded as @xcite @xmath31|\psi_n(0)\rangle , \label{eq : state}\end{aligned}\ ] ] where @xmath32 is probability amplitude of the initial spinwave state . in the weak field approximation
, we will assume @xmath33 at any moment .
we have defined @xmath34 , i.e. the expectation value of the operator @xmath35 .
specifically one finds that @xmath36 is the probability amplitude in the one photon state , @xmath37 and @xmath38 are the amplitude of one atom in the @xmath0 and @xmath2 state , respectively . in order to develop a first intuition for the physics at work we first consider a spinwave that is delocalized merely over two atoms embedded in the atomic cloud ( see fig .
[ fig : illustration]b , c ) .
we assume furthermore that the interaction between atoms in state @xmath2 and the gate atom is infinite for distances smaller than the so - called blockade radius @xmath10 and zero otherwise . outside the blockade region , the photon propagates ( along the @xmath39 direction ) as a dark - state polariton by virtue of eit @xcite . inside the blockade region
the medium behaves like an ensemble of two - level system . here
the incoming photon is building up a non - zero polarization @xmath37 , whose modulus square is the probability density distribution for finding an atom in the decaying state @xmath0 according to eq .
( [ eq : state ] ) @xcite .
eventually , this leads to the loss of the incoming photon and makes the medium opaque .
in order to understand how such photon scattering affects the coherence of the properties of the spinwave one needs to analyze the shape of the polarization profile . as shown in fig . [
fig : illustration]b this in general depends on the position of the gate atom when the system length is larger than the blockade radius @xmath40 . here , since @xmath41 and @xmath13
, it is possible to distinguish the profiles @xmath7 which are associated with the two possible positions of the gate atom .
conversely , the polarization @xmath7 becomes independent of the gate atom position when @xmath14 ( see fig .
[ fig : illustration]c ) . in this case
as discussed in detail later the coherence of the spinwave will be preserved as one can not distinguish gate atoms from the scattered photon .
let us now consider the actual photon propagation together with a realistic interaction potential .
the dynamics of the system follows the master equation @xcite @xmath42+\gamma\int_0^l dz\hat{p}(z , t)\hat{\rho}(t)\hat{p}^{\dagger}(z , t),\end{aligned}\ ] ] where the first term on the right - hand side ( rhs ) is the evolution of @xmath43 under the effective hamiltonian @xmath44 , and the spontaneous decay ( with rate @xmath1 ) from the state @xmath0 is governed by the second term . in the effective hamiltonian ,
the photon propagation in the medium is governed by the hamiltonian @xmath45 with the vacuum light speed @xmath46 .
the atom - photon coupling is described by @xmath47 , \end{aligned}\ ] ] where @xmath48 with @xmath49 being the single atom - photon coupling strength .
the vdw interaction between an atom in the state @xmath2 and the gate atom at position @xmath30 is @xmath50 the interaction potential depends on the gate atom position , @xmath51 where @xmath52 gives the vdw interaction with @xmath53 being the dispersion coefficient . for the case of a single incoming photon which we consider here
the solution of the master equation ( [ eq : masterequation ] ) is @xcite @xmath54 where @xmath55 and @xmath56 .
the first term on the rhs describes the unhindered photon propagation through the medium , while the second term accounts for the photon scattering , i.e. photon - loss from the medium .
to calculate ( [ eq : densityevolution ] ) we first treat the dynamics under the effective hamiltonian in the heisenberg picture . to this end
we obtain the equation of motion for the expectation values @xmath57 from the corresponding operator heisenberg equation @xcite .
note , that due to the linearity of the equations we can moreover calculate the expectation value for each component @xmath58 of the rydberg spinwave , i.e. each of the possible positions of the gate atom , separately .
this yields the set of equations [ eq : me ] @xmath59 where the index @xmath60 labels the respective spinwave component .
alternatively , these equations can be obtained from a heisenberg - langevin approach @xcite .
we solve the coupled equations ( [ eq : me ] ) through a fourier transform yielding the formal solution for the polarization @xmath61 here we have abbreviated @xmath62 and introduced the electric susceptibility @xmath63(\omega+i\gamma/2)}.\end{aligned}\ ] ] from @xmath64 one can actually extract the blockade radius as the critical distance at which the vdw interaction and the control laser are equally strong .
this yields @xmath65 @xcite .
the polarization ( [ eq : pfft ] ) depends on the fourier transform @xmath66 of the photon field at position @xmath67 . to be specific
we take the photon pulse to be a gaussian at @xmath68 which is normalized in space , @xmath69.\ ] ] here @xmath70 is the temporal duration of the pulse and @xmath71 is the initial central position ( @xmath72 ) . the band width of the pulse is then given by @xmath73 .
note , that it is generally not possible to evaluate the formal solution ( [ eq : pfft ] ) analytically .
moreover , numerical calculations are challenging since the involved time and length scales span several orders of magnitude @xcite .
let us now calculate the photon transmission as a function of the pulse duration @xmath70 , which to our knowledge has not been examined previously .
we define the transmission of the photon pulse as @xmath74 . in fig .
[ fig : transmission]a , we show @xmath75 as a function of the pulse width for two values of the atom - photon coupling strength @xmath4 . for fixed pulse length @xmath70
, we find that stronger couplings generally are accompanied by a lower transmission .
furthermore , we observe that the transmission increases with decreasing pulse duration @xmath70 .
this is due to the fact that the pulse contains increasingly more weight on frequency components , which are outside the absorption window of the medium . for the purpose of complete photon scattering ,
one thus has to utilize narrow frequency band pulses . as a function of the pulse duration @xmath70 for @xmath76 ( squares ) and @xmath77 ( circles ) .
the medium becomes transparent when @xmath78 , i.e. the band width @xmath79 of the pulse is large .
( b ) photon transmission as a function of the coupling constant @xmath4 for @xmath80 ( squares ) and @xmath81 ( circles ) .
the solid curve is the analytical result obtained from eq .
( [ eq : transmission ] ) .
the dashed curve is plotted as a guide to the eye .
note , that @xmath82 when @xmath83 for both @xmath80 and @xmath81 .
the data is calculated for rubidium atoms with the parameters , @xmath84 m , @xmath85 , and @xmath86 mhz .
the blockade radius can be changed through selecting different rydberg states .
for example , @xmath87 m when @xmath88 and @xmath89 , where @xmath90 . ]
next , we briefly discuss the dependence of the transmission @xmath75 on the strength of the atom - photon coupling @xmath4 .
[ fig : transmission]b shows data for two choices of the blockade radius , @xmath80 and @xmath81 . as expected , @xmath75 decreases with increasing @xmath4 .
however , for the system parameters chosen here , there is virtually no dependence of @xmath75 on the value of the blockade radius when @xmath76 , where @xmath82 .
these findings indicate that one reaches the strong scattering regime when @xmath91 and @xmath92 .
this is the working regime for the single photon switch where the medium becomes opaque for the incident photon .
focusing on this regime , our next task is to investigate how the photon scattering influences the rydberg spinwave .
we quantify the difference between the initial rydberg spinwave @xmath93 and the final state @xmath94 by the fidelity @xcite @xmath95 ^ 2.\ ] ] as the initial spinwave is a pure state , this simplifies to @xmath96 , where @xmath97 .
this shows that a high fidelity can be obtained only if the polarization profiles @xmath98 for each spinwave component are essentially equal : only when @xmath99 and thus @xmath100 the fidelity is close to one .
this is the formal version of the intuitive statement that we made earlier in conjunction with the discussion of fig .
[ fig : illustration]b , c . for completeness
we provide a numerical example for which we choose @xmath81 and select only two components of the spinwave , where the gate atom is located at either @xmath101 or @xmath102 . the resulting polarization profile @xmath103 is shown in fig . [ fig : psi2]a , b . for @xmath101 ,
non - vanishing polarization is built up within the blockade region as long as the photon is inside the medium ( fig .
[ fig : psi2]a ) . integrating over time
we obtain the intensity @xmath104 which clearly shows a decay to zero within a blockade distance @xmath10 ( see fig . [
fig : psi2]c ) .
in contrast , for @xmath102 , appreciable polarization is built up also outside @xmath10 and the profile is peaked at approximately @xmath105 ( fig .
[ fig : psi2]b , d ) . clearly , both polarization profiles are strikingly different which in turn causes a loss of fidelity when the blockade radius is smaller than the system length .
we verify this by numerically calculating the fidelity as a function of the blockade radius . the data is displayed in fig .
[ fig : fidelity]a , together with the corresponding transmission @xmath75 .
as anticipated , the fidelity decreases significantly below unity when @xmath15 is decreased with respect to @xmath11 .
note , that the transmission is close to zero throughout . and
( c , d ) the time - integrated intensity @xmath106 for @xmath81 and @xmath76 for two different positions of the gate atom . the gate atom ( green circle )
is located at @xmath101 in panels ( a ) and ( b ) and at @xmath102 in panels ( c ) and ( d ) .
the dashed line marks the blockade radius with respect to the gate atom position . ]
a fidelity smaller than unity directly indicates the formation of a mixed state after the photon scattering . the final state density matrix is @xmath107 } |z_j\rangle\langle z_k|.\ ] ] the final state can only be pure when @xmath108 and hence @xmath109 .
the formation of a mixed state is a consequence of the actual measurement of the gate atom position @xcite which is performed by the photon scattering : when @xmath40 one in principle gains information on the position of the gate atom since the spatial uncertainty of its wave function is reduced from @xmath11 to the blockade region .
the final state is then a mixture of all states compatible with this additional information .
in the remainder of the paper we will focus on the case of a coherent photon switch , i.e. @xmath110 . here the expression for the susceptibility of the medium simplifies to that of an ensemble of two - level atoms , @xmath111 which permits the derivation of analytical results . for a narrow band width pulse we can derive explicit solutions to eq .
( [ eq : me ] ) that have no dependence on the position of the gate atom @xcite .
for example , the polarization @xmath37 is given by @xmath112\nonumber\\ & & \times \text{ec}\left[\frac{c\gamma^3\tau^2 - 2\gamma^2(c\mathcal{t}+z_0)-8g^2z}{2\sqrt{2}c\tau\gamma^2 } \right],\end{aligned}\ ] ] where @xmath113 is the complementary error function .
the corresponding time - integrated profile @xmath106 agrees perfectly with the numerical result from eq .
( [ eq : me ] ) ( see fig . [
fig : fidelity]b ) .
the transmission @xmath75 is given by @xmath114^{-1}\left[1+\text{er}\left(\frac{l - z_0}{c\tau}-\frac{\alpha}{\gamma\tau}\right)\right],\ ] ] where @xmath115 is the error function and @xmath116 is the optical depth of a resonant two - level medium . the excellent agreement between the analytical and numerical calculation is shown in fig .
[ fig : transmission]b . neglecting the finite band width of the photon pulse ,
i.e. when all the frequency components are in the absorption window , eq . ( [ eq : transmission ] ) reduces to the well - known form @xmath117 @xcite . with @xmath76
the transmission is already negligible when @xmath118 , but the fidelity approaches unity only when @xmath119 .
( b ) intensity @xmath106 in the strong blockade regime ( @xmath80 ) for @xmath76 ( square ) and @xmath120 ( circle ) .
the squares and circles are numerical data .
the solid and dotted curves are the analytical results obtained from eq .
( [ eq : polarization_response ] ) . ]
finally , the fidelity can be expressed as a function of the optical depth and pulse band width @xmath121.\ ] ] this shows that indeed a small band width is a requirement for reaching a large fidelity .
for example , the transmission is negligible ( @xmath122 ) when @xmath123 and @xmath76 according to the data in fig .
[ fig : transmission]a . however , the fidelity is below unity ( @xmath124 ) due to non - negligible contributions from the terms accounting for the finite band width . in the limit of very long pulses
one finds @xmath125 and thus the fidelity is solely determined by the transmission .
in summary , we have studied the coherence of a rydberg spinwave in the operation of a signle photon switch .
the current study is limited to a single gate atom and an incoming single - photon pulse , which permits the description of multi - photon scattering , however , only if the photons enter the switch sequentially .
addressing this limitation and extending the discussion to correlated and entangled photon pulses that fall in the operation regime of single photon transistors will be subject to future studies . _
acknowledgements. _ we acknowledge helpful discussions with d. viscor , b. olmos and m. marcuzzi .
the research leading to these results has received funding from the european research council under the european union s seventh framework programme ( fp/2007 - 2013 ) / erc grant agreement no .
335266 ( escquma ) , the eu - fet grants no .
295293 ( quilmi ) and no .
512862 ( hairs ) , as well as the h2020-fetproact-2014 grant no .
640378 ( rysq ) .
wl is supported through the nottingham research fellowship by the university of nottingham .
here we will show how to obtain the analytical solution to the coupled equations ( 3 ) in the main text in the strong blockade regime ( @xmath126 ) and for narrow band pulses . in the frequency domain ,
the solution to equations ( 3 ) is given by , @xmath127\tilde{\mathcal{e}}_0(\omega),\\ \label{eq : p } \tilde{p}(z,\omega)&=&\chi\tilde{\mathcal{e}}(z,\omega),\\ \label{eq : s } \tilde{s}(z,\omega)&=&-\frac{\omega}{\omega - v(z)}\chi\tilde{\mathcal{e}}(z,\omega).\end{aligned}\ ] ] due to the strong blockade condition , we have removed the dependence of @xmath128 , @xmath129 , and @xmath130 on the gate atom index @xmath60 and replaced the susceptibility by the one corresponding to two - level atoms , @xmath131 .
moreover , we set @xmath132 , which is a good approximation as @xmath133\approx \omega / v(z)\approx 0 $ ] in eq .
( [ eq : s ] ) .
our aim is to obtain analytical expressions of @xmath3 and @xmath37 . applying the inverse fourier
transform on the both sides of eqns .
( [ eq : e ] ) and ( [ eq : p ] ) , we obtain the formal solution for @xmath3 and @xmath37 in time domain , @xmath134 the integration is in general difficult to carry out analytically due to the complicated form of the susceptibility .
we overcome this difficulty by expanding the susceptibility in powers of @xmath135 , @xmath136.\ ] ] first let us calculate the approximate solution for @xmath3 . to carry out analytical calculations and at the same time
take into account contributions due to the finite band width , we will keep terms up to the second order of @xmath135 in eq .
( [ eq : s_expansion ] ) .
this yields the solution for @xmath3 @xmath137,\end{aligned}\ ] ] with @xmath138 . for the current problem
, we always have @xmath139 as the photon travelling time through the medium is the shortest time scale .
for example , @xmath140 second for @xmath141 m . with the solution for @xmath3
, we can calculate the transmission @xmath74 .
we need to carry out the respective two integrals over time at @xmath67 and @xmath142 .
this can be done analytically , @xmath143 and @xmath144\approx \frac{e^{-\alpha}}{2c\xi(l)}\left[1+\text{er}\left(-\frac{z_0}{c\tau\xi(l)}-\frac{\alpha}{\gamma\tau\xi(l)}\right)\right].\ ] ] this leads to the analytical form of the transmission ( 7 ) in the main text . with the analytical solution for @xmath3 at hand ,
there are two ways to calculate @xmath37 .
we can directly calculate @xmath37 from eq .
( 3b ) by inserting the solution ( [ eq : efield ] ) and @xmath145 .
this yields the linear response of the medium to the photon electric field , @xmath146 the integration over time can be carried out analytically , which gives @xmath147 \left\{1+\text{ec}\left[\frac{24g^2z+2\gamma^2(c\mathcal{t}+z_0)-c\gamma^3\tau^2}{2\sqrt{2\xi}c\tau\gamma^2 } \right]\right\}.\\ p(z , t ) & \approx & ig\sqrt{\frac{\sqrt{\pi}\tau}{2c}}\exp\left[\frac{c\gamma^2\tau^2 - 4\gamma(c\mathcal{t}+ z_0)}{8c}-\frac{6g^2z}{c\gamma}\right ] \left[1+\text{er}\left(\frac{1 - 3\xi^2(z)}{4\sqrt{2}\xi(z)}\gamma\tau + \frac{c\mathcal{t}+z_0}{\sqrt{2}\xi(z)c\tau } \right)\right].\end{aligned}\ ] ] however it is difficult to calculate the fidelity from eq .
( [ eq : p_response ] ) due to the presence of the error function .
we thus calculate @xmath37 alternatively using the fourier transform method .
we note that the susceptibility @xmath148 appears at two places in eq .
( [ eq : psi2ifft ] ) : one in front of @xmath149 and another one in the exponential function .
in order to obtain an analytical result , we will expand the former susceptibility up to the second order of @xmath135 while the latter up to the linear order .
after performing the inverse fourier transform , we obtain the expression for @xmath37 , @xmath150 ^ 2\right\}.\end{aligned}\ ] ] 43 ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1 '' '' @noop [ 0]secondoftwo sanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12 @startlink[1 ] @endlink[0 ]
@bib@innerbibempty in @noop _ _ , vol .
( , ) p. link:\doibase 10.1038/nature11361 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.110.153601 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.82.2313 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.101.250601 [ * * , ( ) ] link:\doibase 10.1103/physreva.80.033418 [ * * , ( ) ] link:\doibase 10.1038/nphys1614 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.104.043002 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.106.025301 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.107.060406 [ * * , ( ) ] link:\doibase 10.1103/physreva.86.033422 [ * * , ( ) ] link:\doibase 10.1103/physreva.87.053414 [ * * , ( ) ] link:\doibase 10.1103/physreva.88.053627 [ * * , ( ) ] link:\doibase 10.1038/35051009 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.112.040501 [ * * , ( ) ]
link:\doibase 10.1103/physrevlett.100.170501 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.107.133602 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.112.133606 [ * * , ( ) ] link:\doibase 10.1038/nature12512 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.112.073901 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.113.053602 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.113.053601 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1038/nphoton.2010.94 [ * * , ( ) ] link:\doibase 10.1038/nphoton.2012.181 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.133602 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.113.083601 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.114.043602 [ * * , ( ) ] link:\doibase 10.1103/physreva.66.023818 [ * * , ( ) ] link:\doibase 10.1103/physreva.78.053816 [ * * , ( ) ] link:\doibase 10.1103/physreva.85.033811 [ * * , ( ) ] link:\doibase 10.1103/physreva.87.023821 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.112.243601 [ * * , ( ) ] link:\doibase 10.1103/physreva.65.022314 [ * * , ( ) ] link:\doibase 10.1126/science.1217901 [ * * , ( ) ] link:\doibase 10.1038/nature12227 [ * * , ( ) ] @noop link:\doibase 10.1103/revmodphys.70.101 [ * * , ( ) ] link:\doibase 10.1103/physreva.76.033805 [ * * , ( ) ] @noop link:\doibase 10.1016/0034 - 4877(76)90060 - 4 [ * * , ( ) ] @noop _ _ ( , ) @noop | we study coherence in a cold atom single photon switch where the gate photon is stored in a rydberg spinwave . with a combined field
theoretical and quantum jump approach and by employing a simple model description we investigate systematically how the coherence of the rydberg spinwave is affected by scattering of incoming photons . with large - scale numerical calculations
we show how coherence becomes increasingly protected with growing interatomic interaction strength . for the strongly interacting limit
we derive analytical expressions for the spinwave fidelity as a function of the optical depth and bandwidth of the incoming photon . |
a model where the participants vote for the projects of redistribution of their own property was analyzed by a.v .
malishevskii in the late 1960 s [ 1 , pp .
9395 ] . in this model
, voting is greatly manipulatable by the organizers , as it is the case with the participants whose `` ideals '' are the points in the multidimensional space of programs [ 2 ] ( see also [ 3 ] ) .
the monograph [ 4 ] is devoted to the spatial voting models .
the problems of relations between selfishness , altruism , and rationality were considered in [ 57 ] , and voting as a method of making decisions about redistribution of social benefits by means of taxation and social programs was discussed in [ 811 ] . for a stochastic environment oriented to the analysis of `` effectiveness '' of the voters collectivist and selfish attitudes in the conditions where new programs are randomly generated by the `` environment , '' rather than developed by the organizers or participants of voting , a model of voting
was suggested by the present author and analyzed in [ 12 ] .
consideration was given to the case of neutral environment , the rest of the cases being analyzed mostly in qualitative terms .
emphasis was made on the time dependencies of the participant `` capitals '' under various values of the model parameters . in what follows
, the present author obtained explicit expressions of these dependencies , the exact formulas of expectations and their normal approximations . these formulas
enable one to determine the nature of social dynamics under any values of the model parameters .
expressions for the expectations of the capital increments were obtained in sec .
[ sec_main ] , the necessary lemmas being proved in sec . [ sec_lem ] .
we give a thumbnail of the simplest variant of the model .
the `` society '' is assumed to consist of @xmath0 `` participants '' of which @xmath1 are `` egoists '' and @xmath2 are the `` group '' members . at any time instant ,
the participant is characterized by its `` capital '' expressed by a real number and interpreted in the most general sense ( like _ utility _ ) .
each participant has some starting capital .
the `` environment proposal '' is the vector @xmath3 of the algebraic increments of the capitals of all participants .
these increments are independent - in - aggregate random variables with identical distribution @xmath4 , where @xmath5 and @xmath6 are the model parameters . at each step
, one random `` environment proposal '' is put to the vote .
each `` egoist '' votes `` for '' if and only if the proposal brings it a positive capital increment .
each member of the group votes `` for '' if and only if the group gains from the realization of this proposal .
the `` gain '' may be understood differently .
the basic model considers two fundamental principles of decision voting .
* principle a*. _ the group votes `` for '' the proposal @xmath7 if and only if as the result of its approval the number of group members getting a positive capital increment exceeds that of the group members getting a negative increment . _ * principle b*. _ the group votes `` for '' the proposal @xmath7 if and only if the sum of increments of group members is positive@xmath8 @xmath9 , where the sum is taken over the subscripts of the group participants .
_ the results of voting are summarized using the `` @xmath10-majority '' procedure : the proposal is accepted if and only if more than @xmath11,@xmath12 , participants vote for it . if the proposal is accepted , then the capitals of all participants get the corresponding increments @xmath13 , or remain the same , otherwise .
capital dynamics of the participants is analyzed in terms of their social roles ( `` egoist '' or `` group member '' ) and model parameters .
it is implied , in particular , that a scenario is plausible where the egoists join the group and the `` group egoism '' resembles more and more the decision making in the interests of the entire society . by another hypothetical scenario ,
an `` ineffective '' group dissolves , and its members either become `` egoists '' or make other groups .
it is planned also to consider the case of a socially oriented group supporting its poorest members and preventing their ruin , a variant of the model where the capital increments depend on their current values , the impact on the social dynamics of the mechanisms of taxation and collection of the `` party dues , '' and so on . at the same time
, it is not planned to consider purely economical mechanisms of capital reproduction and loss because our aim lies in analyzing the social , rather than the economic phenomena .
consideration is given not only to the traditional decision threshold @xmath14 corresponding to the `` simple majority , '' but to all thresholds ranging from @xmath15 to @xmath16 .
that is due to the fact that the most important for example , `` constitutional''decisions are accepted by the a `` qualified parliamentary majority '' with a threshold greater than @xmath17 . on the other hand ,
there are `` initiative '' decisions such as forming new deputy groups in parliament , putting question on agenda , sending requests to other state authorities , initiating referendums , and so on which can be approved by a certain number of votes smaller than one half .
this model is discussed in more detail in [ 12 ] .
the relation between the numbers of egoists and group members is defined by the parameter @xmath18 , half of the portion of egoists among all participants .
we denote by @xmath19 and @xmath20 , respectively , the one - dimensional density and the cumulative distribution function corresponding to the distribution @xmath4;@xmath21 and @xmath22 stand for density and the distribution function of the normal distribution with center at @xmath15 and variance @xmath16 ; @xmath23 and @xmath24 are , respectively , the expectation and the deviation of any of the random variable @xmath25 at hand .
each @xmath26th egoist participant votes for a proposal if and only if @xmath27 . the probability of this event is as follows : @xmath28 the probability of voting `` against '' the proposal is as follows : @xmath29 in what follows , we also need the notation @xmath30 according to the model , the probability that a participant abstains from voting is zero because the normal distribution is continuous .
therefore , the voting of an egoist participant is the bernoulli test with the parameter @xmath31 .
then , since the values @xmath32 are independent , the number of egoists voting `` for '' is distributed _ binomially _ with the parameters @xmath1 and @xmath31 .
the mean value and the variance of this distribution are , respectively , @xmath33 and @xmath34 .
in this section we prove lemmas that underlie the following calculations . by the `` _ normal voting sample _ '' of size @xmath1 with the parameters @xmath35 and voting threshold @xmath36 is meant the totality of random variables @xmath37 , where @xmath38 is a sample from the distribution @xmath4 , @xmath39 and @xmath40 the following lemma holds .
[ l1 ] let @xmath41 be a normal voting sample with the parameters @xmath35 and the voting threshold @xmath36 .
then , for any @xmath42 @xmath43 + 1}^{\ell}\left(\mu+\frac{\sigma f}{q}\left(\frac{x}{p\ell}-1\right)\right ) \left(\begin{array}{c } \ell \\ x \end{array}\right ) p^xq^{\ell - x},\end{gathered}\ ] ] where @xmath44 , and @xmath45 are defined in .
the proofs are given in the appendix .
to calculate ( [ mvot ] ) , one may use the well - known relation between the binomial distribution and the beta - distribution : @xmath46 the right - hand side contains the cumulative function of the distribution of beta - distribution with @xmath47 and @xmath48 degrees of freedom .
however , the normal approximation of the binomial probability can be used even for a comparatively small @xmath1 : @xmath49 where @xmath50 and @xmath51 are the same parameters as for the binomial distribution : @xmath52 , @xmath53 .
summation in ( [ appro0 ] ) provides an approximation to the binomial distribution function : @xmath54 the normal approximation is advisable for @xmath55 . for a fixed @xmath34 ,
its accuracy is maximal for @xmath56 and decreases with @xmath31 approaching @xmath15 or @xmath16 .
therefore , the normal approximation is often recommendable for @xmath57 already if @xmath58 . for the values of @xmath31 that are very close to @xmath15 or @xmath16 ,
usually the condition @xmath59 is imposed .
[ l2 ] let @xmath41 be a normal voting sample with the parameters @xmath35 and the voting threshold @xmath36 .
then substitution of the standard normal approximation of the random variable @xmath60 for its binomial distribution provides the following approximation of @xmath61 for any @xmath62 @xmath63 where @xmath64 + 0{.}5-p\ell)/\sqrt{pq\ell}$ ] .
this approximation is equivalent to the replacement of a finite number of the participants of voting by a continuous `` voting field . ''
we present for completeness sake the following lemma where @xmath65 differs from @xmath66 ( [ vartheta ] ) in the sign of inequality in the definition .
its proof is similar to that of lemmas [ l1 ] and [ l2 ] .
[ l3 ] let @xmath67 , where @xmath68 , be a sample from the distribution @xmath4 and @xmath69 then , for any @xmath42 @xmath70}\left(\mu+\frac{\sigma f}{q}\left(\frac{x}{p\ell}-1\right)\right ) \left(\begin{array}{c } \ell \\ x \end{array}\right)p^xq^{\ell - x},\end{gathered}\ ] ] and the normal approximation provides the approximation @xmath71 where @xmath64 + 0{.}5-p\ell)/\sqrt{pq\ell}$ ] . to sum up the auxiliary results , we denote by @xmath72 the expectation of the normal voting sample of size @xmath1 with the parameters @xmath35 and the voting threshold @xmath36 , and by @xmath73 , the expectation of the set of random variables as defined in lemma [ l3 ] .
then , the exact and approximate values of @xmath74 obey the formulas ( [ mvot ] ) and ( [ mvoti ] ) , and the exact and approximate values of @xmath73 obey the formulas ( [ mivot ] ) and ( [ mvoti ] ) .
let us calculate the expectation @xmath75 of the egoist s capital increment in one step .
we recall that a step lies is considering one proposal of the environment by the participants , independently of whether it is accepted or not .
let the event @xmath76 be supported by the group , and @xmath77 be the opposite .
then , @xmath78 we denote @xmath79 and @xmath80 , respectively , by @xmath81 and @xmath82 .
support of a proposal by the group brings it @xmath83 votes .
consequently , for approval of a proposal on condition that it is supported by the group , it is necessary and sufficient that more than @xmath84 egoists vote for it .
if the group does not support the proposal , then it is necessary and sufficient that more than @xmath11 egoists vote for it .
therefore , @xmath85 where @xmath86 and @xmath87 $ ] is the integer part of @xmath47 .
hence , @xmath88 by substituting expression ( [ mvot ] ) for @xmath89 in ( [ mde ] ) and using the notation @xmath90 we obtain @xmath91 + 1}^{\ell}\left(\mu+\frac{\sigma f}{q}\left(\frac{x}{p\ell}-1\right)\right)b(x\mid\ell)%+ \\ % \hphantom{m(\widetilde{d}_{\mathcal{e } } ) } & + & q_g \sum_{x=[\alpha n]+1}^ { \ell } \left(\mu+\frac{\sigma f}{q}\left(\frac{x}{p\ell}-1\right ) \right)b(x\mid\ell)%= \nonumber \\
\nonumber % \hphantom{m(\widetilde{d}_{\mathcal{e } } ) } & = & p_g \sum_{x=[\gamma n]+1}^{[\alpha
n ] } \left(\mu+\frac{\sigma f}{q } \left(\frac{x}{p\ell}-1\right ) \right)b(x\mid\ell)%+ \nonumber \\
% \hphantom{m(\widetilde{d}_{\mathcal{e } } ) } & + & \sum_{x=[\alpha n]+1}^ { \ell } \left(\mu+\frac{\sigma f}{q}\left(\frac{x}{p\ell}-1\right ) \right)b(x\mid\ell).\nonumber\end{aligned}\ ] ] if the group follows principle a , then @xmath92 + 1}^gb(x\mid g)\quad\text{and}\quad q_g^a=1-p_g^a = \sum_{x=0}^{[{g}/{2}]}b(x\mid g)\end{gathered}\ ] ] should be substituted for @xmath81 and @xmath82 , respectively , and in the case of principle b , @xmath93 because for an arbitrary proposal the mean capital increment of the group has the distribution @xmath94 .
now , by substituting in ( [ mde ] ) the normal approximation ( [ mvoti ] ) for @xmath95 and using the notation @xmath96 + 0{.}5-p\ell}{\sqrt{pq\ell}}\right),\quad f_{\theta}=f\left ( \frac{[\theta n]+0{.}5-p\ell}{\sqrt{pq\ell}}\right),\end{gathered}\ ] ] we get @xmath97\left(\mu \left[\begin{array}{c } f_{\gamma}\\ f_{\alpha } \end{array } \right ] + \frac{\sigma f}{\sqrt { pq\ell}}\left[\begin{array}{c } f_{\gamma}\\ f_{\alpha } \end{array } \right]\right)\end{aligned}\ ] ] where matrix notation is used . in the case of principle
a , the approximate values @xmath98 + 0{.}5-pg}{\sqrt{pqg}}\right ) \approx p_g^a , \\ & 1-f\left(-\frac{\left[\tfrac{g}{2}\right]+0{.}5-pg}{\sqrt{pqg}}\right)= f\left ( \frac{\left[\tfrac{g}{2}\right]+0{.}5-pg}{\sqrt{pqg}}\right ) \approx q_g^a \end{split}\end{gathered}\ ] ] or exact values ( [ pga ] ) must be substituted in this formula ; in the case of principle b , the exact values of ( [ pgb ] ) .
now we derive formulas similar to ( [ mde ] ) and ( [ mdeap ] ) for the expectation @xmath99 of the capital increment of a group member in one step .
let @xmath100 be an event of egoists giving more than @xmath101 votes to the proposal ; @xmath102 , otherwise . if @xmath103 is satisfied , then the proposal is accepted independently of the group voting ; if @xmath104 takes place , then acceptance / rejection of the proposal is defined by its support by the group . if @xmath105 is realized , then the proposal can not be accepted . therefore , @xmath106 if @xmath103 is satisfied , then the proposal is always accepted
therefore , @xmath107 by virtue of independence of the capital increments . if @xmath108 is satisfied , then acceptance of the proposal is defined by the group voting , and generally @xmath109 .
we make use of the notation @xmath110 + 1}^{\ell}b(x\mid\ell ) \approx f_{\theta}=f\left(-\frac{[\theta n]+0{.}5-p\ell}{\sqrt{pq\ell}}\right),\end{gathered}\ ] ] to rearrange ( [ mdg ] ) in @xmath111 if the group follows principle a , then @xmath112 , the substitution of ( [ mvot ] ) in @xmath113 provides @xmath114 + 1}^{g}\left(\mu+\frac{\sigma f}{q}\left(\frac{x}{pg}-1\right ) \right)b(x\mid g)(p_{\gamma}-p_{\alpha})%= \\ \nonumber & = & \mu \sum_{x=[\alpha n]+1}^{\ell}b(x\mid\ell ) +
\sum_{x=[g/2]+1}^{g}\left(\mu+\frac{\sigma f}{q}\left(\frac{x}{pg}-1\right ) \right)b(x\mid g ) \sum_{x=[\gamma n]+1}^{[\alpha n]}b(x\mid\ell).\end{aligned}\ ] ] in order to obtain the normal approximation of this value , we replace @xmath115 in ( [ mdga ] ) by the approximation ( [ mvoti ] ) , and @xmath116 and @xmath117 , by the approximations @xmath118 and @xmath119 ( see ( [ pthe ] ) ) . by using the notation @xmath120 + 0{.}5-pg}{\sqrt{pqg}}\right),\end{gathered}\ ] ] we obtain @xmath121\left(\mu\left [ \begin{array}{c } p_g^a \\[1 mm ] q_g^a \end{array } \right ] + \frac{\sigma ff_g^a}{\sqrt{pqg}}\left [ \begin{array}{r } 1 \\ -1 \end{array } \right]\right).\end{aligned}\ ] ] if the group follows principle b , then @xmath122 , and ( [ mdg ] ) is reshaped in @xmath123 to determine @xmath124 , one may substitute the corresponding arguments in ( [ mvot ] ) or , which is even simpler , multiply ( a.4 ) by @xmath31 : @xmath125 by using the notation ( [ pgb ] ) and @xmath126 we obtain @xmath127\left(\mu\left[\begin{array}{c } p_g^b \\[1 mm ] q_g^b \end{array } \right ] + \frac{\sigma f_g^b}{\sqrt{g}}\left [ \begin{array}{r } 1 \\
-1 \end{array}\right]\right ) , \nonumber\end{aligned}\ ] ] where either the exact values or normal approximations ( see ( [ pthe ] ) ) can be substituted for @xmath117 and @xmath116 .
the expected values of the capital increments of the egoists and members of the group in a series of @xmath128 steps obey , respectively , @xmath129 and @xmath130 .
in addition to principles a and b , the group can make use of principle a@xmath131 [ 12 ] enabling it sometimes to minimize and sometimes to eliminate the advantage of the egoists in those zones of variation of the threshold @xmath10 where they have this advantage .
this is reached at the expense of reduced group capital increment in these zones : relative gain of the group leads to absolute losses .
we formulate below this voting principle .
the threshold @xmath133 vs. @xmath10 and @xmath134 is depicted in fig . 1 .
vs. @xmath10 . ] * principle a@xmath131*. _ the group votes `` for '' the proposal of the environment if and only if as the result of accepting it the part of its members getting a positive capital increment exceeds the threshold _
@xmath135 \tfrac{1}{2}+\tfrac{\delta}{2\beta } , & \text{where\ } \delta=\alpha-(1-\beta ) , & \text{for\ } \alpha>1-\beta;\\[3 mm ] \tfrac{1}{2 } & & \text{for\ } \alpha\in[\beta,1-\beta ] \end{array}% \right.\hspace*{4 mm } % = \\ & = & \left\{% \begin{array}{ll } \tfrac{\alpha}{2\beta } & \text{for\ } \alpha<\beta;\\[1 mm ] 1-\tfrac{1-\alpha}{2\beta } & \text{for\ } \alpha>1-\beta;\\[1 mm ] \tfrac{1}{2 } & \text{for\ } \alpha\in[\beta,1-\beta ] . \end{array}% \right.\nonumber\end{aligned}\ ] ] vs. the decision threshold @xmath10 .
values of the parameters : @xmath136 ; @xmath137 ; @xmath138 ; @xmath139 ; @xmath140 . ] vs. the decision threshold @xmath10 .
values of the parameters : @xmath136 ; @xmath137 ; @xmath138 ; @xmath139 ; @xmath140 . ]
since principle a@xmath131 differs from principle a only in the internal threshold of group voting , for calculation of the expected capital increments it suffices to replace @xmath141 $ ] by @xmath142 $ ] in ( [ pga ] ) , ( [ pgaqga ] ) , ( [ mdg ] ) , and ( [ fga ] ) .
thus , in ( [ mde ] ) , ( [ mdeap ] ) , and ( [ mdgap ] ) one has to substitute @xmath143 + 1}^gb(x\mid g ) \\ & \text{or the approximation}\quad f\left(-\frac{[\alpha'g]+0{.}5-pg}{\sqrt{pqg}}\right)\approx p_g^{a ' } , \end{split}\end{gathered}\ ] ] for @xmath81 and @xmath144 , @xmath145}b(x\mid g ) \\ & \text{or the approximation } \quad f\left ( \frac{[\alpha'g]+0{.}5-pg}{\sqrt{pqg}}\right)\approx q_g^{a ' } , \end{split}\end{gathered}\ ] ] for @xmath82 and @xmath146 , and @xmath147 + 0{.}5-pg}{\sqrt{pqg}}\right)\end{gathered}\ ] ] for @xmath148 .
we present some examples of applying the above formulas . for @xmath138 , @xmath139 , @xmath149 , @xmath137 participants , and @xmath140 steps , fig .
2 shows the expected capital increments of the egoists and group members vs. the decision threshold , provided that the group makes use of different voting principles . vs. the decision threshold @xmath10 .
values of the parameters : @xmath150 ; @xmath151 ; @xmath138 ; @xmath139 ; @xmath140 . ] vs. the decision threshold @xmath10 .
values of the parameters : @xmath152 ; @xmath137 ; @xmath153 ; @xmath139 ; @xmath140 . ] the functions in fig . 2
are stepwise because the results of voting are defined by the ratio of the number of participant voting `` for '' and the _ integer part _ of the threshold @xmath11 .
the capital increment of a group member in the case of principle a@xmath131 is shown by the white line which for @xmath154 lies slightly below the bold black line corresponding to the capital increment of a group member in the case of principle a. at the same time , at the passage to a@xmath131 the egoists lose more : for @xmath154 , the thin black line that represents the variation of their mean capital passes much lower the bold white line corresponding to principle a. the capital of a group member grows to the limit for @xmath155 by 410 units , whereas in the case of accepting all proposals the reduction in capital is 300 units .
the expected capital increment of the egoist is positive for @xmath156 and maximal for @xmath157 .
the group can use a more radical approach to the choice of the internal threshold assuming that it is equal to the external one . * principle a@xmath132*. _ the group votes `` for '' a proposal of the environment if and only if as the result of accepting it the part of its members getting a positive capital increment exceeds @xmath10_. in this case , the capital increments are calculated as in the case of principle @xmath131 with @xmath158 $ ] in place of @xmath142 $ ] .
an example of using principle a@xmath132 is depicted in fig . 3 , all parameters being the same as in fig . 2 .
the passage from principle a@xmath131 to principle a@xmath132 in the main worsens the group activities and , additionally , changes completely the dynamics of the egoists capital . in the example of fig . 4 ,
practically all proposals of the unfavorable environment are accepted for @xmath159 , and practically no proposal is accepted for @xmath160 . for @xmath161 ,
the expected capital increment of a group member is maximal : @xmath162 ; the capital increment of the egoist does not become positive under any @xmath10 . the main distinction of principle a@xmath132 lies in that it leads to actual disappearance of the maxima of capital increments of the egoists , provided that their fraction is excessive . in the case of principles a and b , these peaks are especially high in zones 4 and 5 namely for a small number of egoists ( see fig . 4 ) , and the passage to principle a@xmath132 eliminates them . for 300 participants ,
the height of these peaks is negligible if the egoists make up less than two thirds of the participants ; low peaks for 80% of egoists ( low `` strongly plowed hills '' ) can be seen in fig .
if an optimality criterion is defined , then one may raise the question of the optimal choice of the intragroup voting threshold .
similarly , it is possible to vary the threshold of proposal support for principles like b. in the case of favorable environment , in particular , the requirements on the mean one - step capital increment of a group member can be set at a higher level than in an unfavorable environment .
all examples show that to a certain extent the group can `` dictate its will . ''
it deserves noting that for 300 participants , neutral environment , and simple - majority procedure , the ratio of the expected capital increment of the member of a three - participant group to the capital increment of an egoist is approximately @xmath163 in the case of principle a and @xmath164 in the case of principle b.
exact and approximate ( based on the normal approximation of the binomial distribution ) formulas for the expectations of the capital increment of the `` egoists '' and the `` group members '' were obtained for the considered model of voting in an environment with certain random parameters . these formulas
enable one to determine the form of the model trajectory for any values of the parameters .
the distinctions of different voting principles were considered using several examples .
let @xmath165 .
we denote @xmath60 by @xmath166 for brevity . using the formula of total probability for expectation , we get @xmath167 + 1}^{\ell}m(\eta_k\mid n^+=x)\,p\{n^+=x\}\end{gathered}\ ] ] for any @xmath42 .
then @xmath168 it follows from independence of the sample elements that @xmath169 this expression is the mean normal random variable , if positive , which is determined by integration and is as follows : @xmath170 where @xmath171 .
similarly , @xmath172 we note that @xmath173 and @xmath174 , substitute the determined values in ( a.2 ) @xmath175 and ( a.1 ) @xmath176 + 1}^{\ell}\left(\mu-\frac{\sigma f}{q}+\frac{x\sigma f}{pq\ell}\right ) \left(\begin{array}{c } \ell \\ x \end{array}\right)p^xq^{\ell - x},\end{gathered}\ ] ] which completes the proof of lemma .
@xmath177 we rearrange ( [ mvot ] ) in @xmath178 + 1}^{\ell } \left(\begin{array}{c } \ell \\ x \end{array}\right)p^xq^{\ell - x } \left(\mu-\frac{\sigma f } { q}+ \frac{\sigma f}{pq\ell}\cdot\frac{\sum\limits_{x=[\ell_0]+1}^{\ell}x \left(\begin{array}{c }
\ell \\ x \end{array}\right)p^xq^{\ell - x } } { \sum\limits_{x=[\ell_0]+1}^{\ell } \left(\begin{array}{c } \ell \\ x \end{array}\right)p^xq^{\ell - x}}\right).\end{gathered}\ ] ] the first sum in the right - hand side is @xmath179 ; it is approximated by @xmath180 .
the ratio of the two remaining sums is @xmath181 . for @xmath182 , @xmath183 ( this formula stems from ( a.4 ) ) .
we use it to approximate @xmath184 and obtain @xmath185 substitution of these approximations in ( a.8 ) provides @xmath186 @xmath177 borzenko , v.i . , lezina , z.m .
, loginov , a.k . ,
tsodikova , ya.yu . , and chebotarev , p.yu . , strategies of voting in stochastic environment :
egoism and collectivism , _ automat .
remote control _ , 2006 ,
67 , no . 2 , pp . | in the simplest variant of the model of collective decision making in the stochastic environment , the participants were segregated into egoists and a group of collectivists .
`` proposal of the environment '' is the stochastically generated vector of algebraic increments of capitals .
the social dynamics was defined by the sequence of proposals accepted by threshold - majority voting .
analytical expressions of the expected values of the capitals of participants , collectivists and egoists were obtained .
distinctions of some principles of group voting were discussed . |
speculations that instantons could induce diquark condensation in low temperature but high density quark matter @xcite have revived the interest in the diquark clustering in the nucleon .
it is sometimes also argued that diquark condensation may occur even at moderate densities , for example in heavy nuclei .
this problem is strongly related to the question of instanton induced diquark clustering in the nucleon .
indeed , the instanton - induced t hooft interaction is strongly attractive for a quark - quark pair with quantum numbers @xmath2 ( scalar diquark ) .
this raises expectations that it binds a scalar diquark and is responsible for the scalar diquark - quark structure of the nucleon @xcite .
this assumption is based on the iteration of the t hooft interaction in the @xmath3 s - channel .
however , this picture of the quark - quark interaction in baryons is only a small part of a more general one , based on the effective meson - exchange interaction @xcite .
when the t hooft interaction is first iterated in the @xmath3 t - channel it inevitably leads to goldstone boson exchange between constituent quarks , which is drastically different from the initial ( not iterated ) t hooft interaction due to the ( anti)screening effects .
the latter effective meson - exchange interaction does not induce a bound scalar diquark , nor an appreciable diquark - quark clustering in nucleon .
this effective meson exchange interaction is also the most attractive in @xmath4 @xmath3 pairs , but the nature of this attraction is very different from that of the t hooft interaction .
this interaction , however , is not strong enough to bind the scalar diquark . when it is combined with a confining interaction it binds the diquark in the sense that there is no asymptotic state with two free constituent quarks , though the mass of the scalar diquark is a few tens of mev above the two - constituent - quark threshold .
there is no significant diquark clustering in the nucleon either , because the nucleon is intrinsically a three - quark system and the fermionic - nature of the constituent quarks plays an important role .
if the subsystem of quarks 1 and 2 is in the @xmath4 state then due to the antisymmetrization the quark pairs in the subsystems 1 - 3 and 2 - 3 are also partly in the @xmath4 state .
this implies that a strong attraction in @xmath4 quark pair contributes in all quark subsystems simultaneously and makes the nucleon compact , but without appreciable quark - diquark clustering .
this paper consists of two independent , but interrelated parts . in the first one we discuss how the instanton - induced interaction ( or some general nonperturbative gluonic interaction ) leads to the poles when it is iterated in the @xmath3 t - channel .
these pole contributions have an evident meson - exchange interpretation .
the latter meson - exchange interaction is drastically different from the initial ( bare ) t hooft interaction which becomes strongly enhanced in the channel of goldstone boson exchange quantum numbers .
we also discuss the role of instantons in @xmath5 systems .
there is no new wisdom in that the nonperturbative gluonic configurations , e.g. instantons , induce the dynamical breaking of chiral symmetry and explain the low - lying mesons .
we include the latter discussion only with the purpose of showing how the nonperturbative gluonic interaction both explains mesons and at the same time leads to the effective meson exchange picture in the @xmath3 systems . through the latter it also explains the baryon spectra and the nuclear force .
our discussion is rather general , and does not necessarily rely on the instanton - induced interaction picture .
any nonperturbative gluonic interaction , which respects chiral symmetry and induces the rearrangement of the vacuum ( i.e. dynamical breaking of chiral symmetry ) , will automatically explain the @xmath6 mass splitting and will imply a meson - exchange picture in baryons .
the second part of this paper is devoted to a detailed study of diquark clustering in the nucleon , based on the effective meson - exchange interactions in the baryons and the nucleon wave functions obtained from the solution of the semirelativistic three - body schrdinger equation .
we show that there is no appreciable diquark clustering in the nucleon and that the effective meson - exchange interaction , which is adjusted to describe the baryon spectrum @xcite , does not bind the scalar diquark nor the nucleon .
however , when this interaction is combined with the confining interaction , one finds a bound diquark but with a mass above the two - quark threshold and very similar in magnitude to that obtained recently in lattice qcd @xcite .
nevertheless , as soon as the strength of the effective meson - exchange interaction is increased , not by a very big amount , it alone binds a nucleon , even without a confining force .
while the contributions from the confining interaction to the nucleon mass are not small , the nucleon size , calculated with the confining interaction alone and in a full model that includes both confinement and effective meson exchange , is different .
it is substantially smaller in the latter case , showing that there is indeed a soft interval between the scale when confinement becomes active , and the scale where chiral physics starts to work .
however , for excited baryon states , which are much bigger in size , the role of confinement increases .
it has been shown in recent years that a successful explanation of light and strange baryon spectroscopy , especially the correct ordering of the lowest states with positive and negative parity , is achieved if the hyperfine interaction between constituent quarks @xmath7 and @xmath8 has a short - range behaviour which reads schematically @xcite : @xmath9 where @xmath10 is a set of a flavor gell - mann matrices for @xmath11 and @xmath12 .
this interaction is supplied by the short - range parts of goldstone boson exchange ( gbe ) , , @xmath13 and @xmath14 exchanges ; due to the axial anomaly the @xmath15 is not a goldstone boson , but in the large @xmath16 limit it also becomes a goldstone boson , and thus the coupling of @xmath15 to a constituent quark should be essentially different from that of octet mesons . ]
vector - meson - like exchange and/or correlated two - pseudoscalar - meson - like exchange @xcite , etc . it is sometimes stated that the instanton - induced t hooft interaction in @xmath3 pairs could also provide a good baryon spectrum as it contains a flavor- and spin - dependence and , iterated in the @xmath3 s - channel , produces a deeply bound scalar diquark which makes the nucleon lighter than the @xmath17 @xcite .
a similar picture of a deeply bound scalar diquark has been advocated in a generalized nambu and jona - lasinio ( njl ) model @xcite .
then a baryon is constructed as an additive diquark - quark system or by solving `` relativistic diquark - quark faddeev equations '' that take into account the quark exchange between the diquark and quark - spectator @xcite . in this section
we show that such a picture of baryons , based on the iteration of the local 4-fermion interaction in the @xmath3 s - channel is only a small part of a more general picture , based on the meson - exchange interaction .
the reason is that when the t hooft interaction ( or generalized njl one ) is first iterated in the @xmath3 t - channel , it inevitably leads to the effective meson - exchange between constituent quarks , which is drastically different from the initial ( not iterated ) 4-fermion local interaction due to ( anti)screening effects .
the difference is not only in the flavor- and spin - dependence , but sometimes also in the sign of the interaction . to demonstrate this we use a simple @xmath18 t hooft - determinant interaction for two light flavors ( u and d ) , neglecting for the simplicity the tensor coupling term , which is suppressed by the factor @xmath19 @xcite
. we also assume zero masses for the current u and d quarks in this section . for our illustrative purposes
such an approximation is justified .
this hamiltonian reads : @xmath20 . \label{hooft}\ ] ] the _ dimensional _ strength of the interaction @xmath21 as well as the ultraviolet cut - off scale @xmath22 can be related to parameters of the instanton liquid @xcite ( the dimensionless coupling constant is given by @xmath23 ) .
the interaction ( [ hooft ] ) is attractive in the scalar - isoscalar @xmath24 channel ( the first term ) , leading to chiral symmetry breaking , or , which is related , to a massive @xmath25-meson field and the constituent mass @xmath26 of quarks .
this is readily obtained from the schwinger - dyson ( gap ) equation for a quark green function in the hartree - fock approximation .
the interaction in the @xmath27 pseudoscalar - isovector channel is driven by the second term of ( [ hooft ] ) .
it is so strong , that when it is iterated in the @xmath27 s - channel by solving the bethe - salpeter equation , see fig .
1 , it exactly compensates the @xmath28-energy , supplied by the first term in ( [ hooft ] ) , and thus there appear @xmath29 mesons with zero mass as deeply bound relativistic @xmath27 systems - nambu - goldstone bosons .
the nonzero mass of the pseudoscalar mesons is brought about by the nonzero current quark mass as a perturbation , which is well illustrated by the current algebra results ( gell - mann - oakes - renner relations ) .
the first two terms in the hamiltonian ( [ hooft ] ) form in fact the classical njl hamiltonian @xcite and the statement above is a theorem , proved by nambu and jona - lasinio many years ago . this scenario holds if the fixed strength of the interaction @xmath21 exceeds some critical level . in a more sophisticated derivation @xcite the strength of the interaction @xmath21 is not fixed and should be determined after one gets the chirally broken phase .
the hamiltonian ( [ hooft ] ) does not contain any interaction in @xmath27 pairs with vector meson quantum numbers .
so , according to the scenario above , the masses of vector mesons , @xmath1 and @xmath30 , should be approximately @xmath28 , which is well satisfied empirically .
thus , it can not be overemphasized that the @xmath31 mass splitting is brought about not by the perturbative color - magnetic interaction between nonrelativistic constituent quarks system and its nambu - goldstone boson nature ( zero mass in the chiral limit ) can not be obtained from a nonrelativistic reduction of the second term in ( [ hooft ] ) used in the schrdinger equation .
] , but by the detailed balance between the first and second terms in ( [ hooft ] ) , which is determined exclusively by the demand that the gluonic interaction between current quarks must satisfy chiral @xmath32 symmetry ) is color - independent .
one can , of course , rewrite this interaction using the pauli principle in terms of linear combinations of different operators , like @xmath33 , @xmath34 , @xmath35 , @xmath36 , @xmath37 , ... the presence of the @xmath36 structure in this decomposition does not mean that the interaction ( [ hooft ] ) becomes similar in its effect to the color - magnetic component of the one gluon exchange interaction , which is explicitly color - dependent .
it is like an identity @xmath38 does not mean that effect of @xmath39 is contained in @xmath40 . ]
. an important question , which is actively debated nowdays , is which particular _ nonperturbative _ gluonic configurations in qcd , e.g. instantons , or abelian monopoles , or other topological configurations , are intrinsically responsible for the chiral symmetry breaking . among other attractive features of the instanton - induced interaction ( [ hooft ] )
is that it automatically solves the @xmath41 problem , giving a much bigger mass to the pseudoscalar flavor - singlet ( in the present 2-flavor formulation that is isosinglet ) meson @xmath15 @xcite .
this is because of the last term in ( [ hooft ] ) .
only this term contributes in a pseudoscalar flavor - singlet quark - antiquark pair . since this interaction is repulsive , the @xmath15 becomes heavy , contrary to @xmath0 .
we note in passing that the color - magnetic interaction can not explain this big @xmath42 mass splitting .
clearly , the simple hamiltonian ( [ hooft ] ) is only some part of a more complicated physical situation .
for instance , one definitely needs some additionl attractive interaction , e.g. confinement , otherwise the @xmath15 meson or vector mesons will be unbound , while the octet pseudoscalar mesons are probably not affected by the long - range confining interaction .
having mentioned all the positive features of the hamiltonian ( [ hooft ] ) in the quark - antiquark system , we are now going to discuss its implications in quark - quark systems , i.e. in baryons .
what is typically done is a fierz - rearrangement of the hamiltonian ( [ hooft ] ) into diquark @xmath3 channels @xcite ( or , similarly , a fierz - rearrangement of the generalized njl hamiltonian into diquark channels @xcite ) .
then the diquark hamiltonian is iterated in the @xmath3 s - channel , see fig .
the interaction in the scalar @xmath43 diquark turns out to be attractive and it produces a deeply bound scalar diquark . ] .
however , as soon as the hamiltonian ( [ hooft ] ) is iterated first in the @xmath3 t - channel , see fig .
3 , it implies irreducible ( for the @xmath3 s - channel ) pion- and sigma - exchange interactions between quarks .
this statement comes about as a theorem since the iteration of the hamiltonian ( [ hooft ] ) in @xmath3 t - channel is equivalent to its iteration in @xmath27 s - channel .
clearly the set of diagrams in fig .
3 contains all the diagrams of fig . 2 , but in addition it contains many others , and the effect of these additional diagrams is so important that the physics implied by fig . 2 and fig .
3 is drastically different ) is applied in baryons in the framework of the chiral quark - soliton model @xcite . in this case
quark - quark correlations through the self - consistent chiral _
mean field _ and quantization of its rotation take into account some part of the iterations in s - channel of fig . 2 and do not take into account the t - channel ladders of fig .
3 . ] . a simple example of the different physical implications is that fig .
3 suggests a long - range meson - exchange yukawa tail , which is crucial for the interaction of quarks , belonging to different nucleons , while if the picture of fig .
2 were correct the nuclear force would be absent .
another evident difference is that according to the hamiltonian ( [ hooft ] ) and fig .
2 the interaction is absent in flavor - symmetric , @xmath44 , quark pairs ,- resonance and its excitations .
if that were the case , the positive parity state @xmath45 , which belongs to the @xmath46 shell because of its positive parity , would be approximately @xmath47 mev above the negative parity pair @xmath48 .
] while the @xmath3 interaction of fig .
3 does not vanish in this case .
less evident is that even the short - range interaction between quarks is crucially modified in fig . 3 as compared to fig .
we call it the `` ( anti)screening effect '' and illustrate it below . in order to see it one should avoid the fierz - rearrangement of ( [ hooft ] ) into a diquark hamiltonian .
instead , one can use the initial hamiltonian ( [ hooft ] ) , but assume that all initial , intermediate and final state @xmath49 wave functions are explicitly antisymmetric .
consider the first term of ( [ hooft ] ) . in the nambu - goldstone mode of chiral symmetry a fermion field
has a large dynamical ( constituent ) mass @xmath26 . using a @xmath50 expansion ,
one obtains that to leading order ( @xmath51 ) the first term of ( [ hooft ] ) leads to a @xmath52-function type attraction in all quark pairs allowed by pauli principle : @xmath53 the effect of the third term in ( [ hooft ] ) to the same order is @xmath54 the potentials ( [ first ] ) and ( [ third ] ) , combined together , produce @xmath55 note that at this order the second and fourth terms of ( [ hooft ] ) do not contribute . the potential ( [ comb ] ) suggests a strong attraction in the isospin - zero quark pair , and no interaction in @xmath44 quark pairs . assuming relative angular momentum @xmath56 within the @xmath57 quark pair the pauli principle implies that the spins of the quarks should be antiparallel , @xmath58 .
when the theory is sensibly regularized the delta - function attraction is smeared out over the instanton size @xmath59 @xmath60 this substitution arises from a replacement of the static green function of the infinitely heavy particle in ( [ hooft ] ) @xmath61 by the green function of a particle with mass @xmath62 @xmath63 when the strength of the interaction is big enough , the potential ( [ comb])-([delta ] ) , iterated by solving the semirelativistic schrdinger equation ( i.e. when the kinetic energy operator is taken in a relativistic form ) can produce a deeply bound scalar diquark , in agreement with @xcite .
indeed , when one takes the strength @xmath64 gev@xmath65 , @xmath66 , with the instanton size @xmath59 between 0.3 and 0.35 fm and the constituent mass @xmath67 mev @xcite one finds a very deeply bound diqurk . in the illustration above we have used a simplified but transparent nonrelativistic picture that adequately reflects in the present case the essential features of a more rigorous bethe - salpeter approach .
sometimes the potential ( [ comb ] ) is applied to explain the hyperfine splittings in baryons @xcite .
while it can generate the @xmath68 mass splitting , it fails to explain the lowest levels with positive and negative parity because it does not contain the necessary spin - isospin dependence ( [ gbe ] ) .
it will become evident from the discussion below that such an interpretation of the role of instantons in baryons does not survive as soon as the wider class of diagrams in fig .
3 is considered .
what happens when the first term in ( [ hooft ] ) is iterated in the @xmath3 t - channel ?
the corresponding amplitude is @xmath69 where @xmath70 is the loop integral ( bubble ) with the scalar vertex which represents vacuum polarization in the scalar channel .
( [ iter ] ) defines `` running amplitude '' and a negative sign in the denominator implies its antiscreening behaviour .
the expression ( [ iter ] ) is known to have a pole at @xmath71 @xcite , which can be identified with the exchange by scalar meson @xmath25 with the mass @xmath72 in the chiral limit .
the coupling constant of the @xmath25-meson to constituent quark can be obtained as a residue of ( [ iter ] ) at the pole @xmath73 expanding the @xmath74 vertex in @xmath50 , one obtains to leading order @xmath75 the following well - known sigma - exchange potential @xmath76 the equivalence between the t - channel ladder of bubbles in fig .
3 beyond the @xmath25-meson pole in the t - channel and the meson - exchange diagram is achieved only when some form factor @xmath77 is inserted into the meson - quark vertex , i.e. the left hand side of eq .
( [ pole ] ) should be multiplied with @xmath78 .
the form factor is to be normalized @xmath79 . in principle
( [ pole ] ) allows to obtain a functional form for such a form factor .
however , it will be very far from reality because the toy model ( [ hooft ] ) does not contain confinement , which should be important for the interaction between quark and antiquark in the weakly bound system like @xmath25-meson ( note that its mass is just at the `` continuum threshold '' @xmath28 ) .
in this situation the best way is to rely on our general understanding of the low - energy effective theory .
both constituent quarks and chiral meson fields as well as their couplings make a sense only in the nambu - goldstone mode of chiral symmetry .
when momentum transfer at the meson - quark vertex exceeds the chiral symmetry breaking scale @xmath80 ( which within all njl - like models coincides with the regularization scale @xmath22 ) the effective theory should be cut off .
this cut off is accomplished by a form factor in the meson - constituent quark vertex and should be related to the internal structure of both quasiparticles .
but in any case the scale parameter in this form factor should be comparable with @xmath80 . at high
momenta one can use neither constituent quark nor chiral fields and original quark - gluon degrees of freedom should be used instead . approximating this form factor by @xmath81 instead of the potential ( [ sigma ] ) we arrive at @xmath82 note , that any functional form of form factor leads to a similar suppression of the potential at short range but
it does not influence its long - range part which is determined exclusively by the position of the pole .
if one takes a dipole form factor the suppression will be stronger .
what is the fate of the third term in ( [ hooft ] ) , when it is iterated in the @xmath3 t - channel ? in this case one obtains the following amplitude @xmath83 where @xmath84 are isospin indices .
the positive sign in the denominator indicates screening .
for instance , at @xmath71 the strength of the interaction is reduced by the factor 2 versus a bare vertex . still , this suppression of the interaction at low momenta is not realistic , because the toy model ( [ hooft ] ) does not contain confinement and thus there is only a repulsion in the scalar - isovector quark - antiquark system . when confinement is added in the quark - antiquark pairs ,
there appear heavy scalar - isovector mesons and the sign of the amplitude ( [ iter2 ] ) becomes opposite at small momenta !
this low - momentum amplitude corresponds to the exchange by scalar - isovector mesons between quarks .
the corresponding meson - exchange interaction is similar in form to ( [ sigfor ] ) , but with an additional factor @xmath85 .
the expectation value of the operator @xmath86 in the scalar diquark is @xmath87 and thus the interaction ( [ comb ] ) is stronger by the factor 4 through the interaction ( [ third ] ) in the picture of fig .
2 . in contrast , in the picture of fig .
3 the contributions from the scalar and scalar - isovector meson exchanges tend to cancel each other .
thus we see that the initial interaction is screened .
this screening means that the interaction ( [ comb ] ) becomes weaken and that its isospin dependence is modified .
it is trivial to check that the attraction ( [ sigfor ] ) does not lead to a bound scalar diquark with any reasonable coupling constant , sigma - meson mass , cut - off mass @xmath88 and constituent quark mass ( see discussion in the next chapter ) .
the scalar - isovector meson exchange will further reduce this attraction , though the coupling constant of the scalar - isovector mesons to constituent quarks will be essentially smaller . both the scalar - isoscalar exchange and scalar - isovector exchanges between constituent quarks do not contain the flavor - spin dependence ( [ gbe ] ) which is necessary for baryon spectroscopy - exchange is known to be very important for the medium - range attraction in the nn system and it also contributes to binding nucleon .
however it only has a small influence on the splittings via different radial behaviour of baryon wave functions . ] .
now we shall extend our @xmath50 expansion of the hamiltonian ( [ hooft ] ) to the next - to - leading order , taking into account terms @xmath89 .
the first and third terms of ( [ hooft ] ) will give at this order the spin - orbit forces , as well as some small corrections to the interactions ( [ first ] ) and ( [ third ] ) .
the second and fourth terms generate , however , a flavor - spin dependent interaction .
consider the second term in ( [ hooft ] ) .
it gives both the spin - spin and tensor force components .
we ignore below the tensor force as it is irrelevant to our simple discussion and for the @xmath56 @xmath3 pair .
then : @xmath90 again , assuming that the instanton has a finite size @xmath59 , the potential ( [ second ] ) reads : @xmath91 let us now iterate the second term of ( [ hooft ] ) in the @xmath3 t - channel @xmath92 where @xmath93 is a bubble with a pseudoscalar vertex ( vacuum polarization in the pseudoscalar channel ) .
there is a pole at @xmath94 in ( [ iterp ] ) in the chiral limit , which can be identified as a pion - exchange ( beyond the chiral limit it is shifted to a physical pion mass @xmath95 . )
the coupling constant of pion to constituent quark can be obtained as a residue of ( [ iterp ] ) at the pole @xmath96 thus near the pole ( [ iterp ] ) - ( [ polep ] ) represents a pion - exchange potential between quarks , which in the chiral limit , @xmath97 , at the order @xmath98 ( omitting the tensor force component ) is : @xmath99 the difference between ( [ second]-[secmod ] ) and ( [ pion ] ) is obvious : the interaction ( [ pion ] ) is much stronger . and @xmath100 where @xmath101 is a zero order function stemming from confinement . ]
a source of this enhancement is also obvious : near the pole the original bare interaction @xmath102 becomes strongly reinforced .
the interaction is represented by a bare vertex @xmath102 , but at @xmath103 it becomes infinitely enhanced in the channel with goldstone boson exchange quantum numbers . ]
the pion pole is located just near the space - like region and thus strongly influences the quark - quark interaction at not very high momentum transfer .
again , to retain the equivalence between the t - channel ladder of bubbles in fig . 3 and the pion - exchange diagram beyond the pole , one must insert a form factor into the @xmath104 vertex .
the effect of this form factor is to smear out the @xmath52-type interaction in ( [ pion ] ) over the region @xmath105 . if this form factor is choosen in the form ( [ form ] ) , then one obtains @xmath106 the @xmath89 expansion of the fourth term of the hamiltonian ( [ hooft ] ) will give a result similar to second term , without , however , isospin - dependent factor and with the opposite sign .
its iteration in the @xmath3 t - channel will produce screening effects as it is repulsive in the @xmath107 s - channel .
when , however , this term is combined with an additional attractive interaction , e.g. confinement , it will give @xmath15 .
then the iteration in the @xmath3 t - channel will imply @xmath15-exchange between quarks .
the latter interaction is similar to ( [ pion ] ) , except that the factor @xmath108 is not present and that in this case there appears a yukawa part of the potential because the mass of @xmath15 is not zero in the chiral limit .
the discussion above suggests that while for the picture of fig .
2 the most important interaction is ( [ comb ] ) and the interaction ( [ second ] ) is only some very small correction to it , in the case of fig .
3 the most important interaction in baryons becomes ( [ pion ] ) and the one of ( [ sigfor ] ) only plays a modest role for splittings .
this is a consequence of an antiscreening .
the antiscreening implies that if a typical momentum transfer in the meson - quark vertex ( which in qq systems is of the same order as momentum of quarks ) is below the chiral symmetry breaking scale , then the original ( bare ) quark - quark vertex in the pseudoscalar channel is strongly reinforced by the pole which occurs when one iterates it in the t - channel .
these pole contributions represent the goldstone boson exchange interactions between constituent quarks in the nambu - goldstone mode of chiral symmetry . only at a rather high momentum transfer ( i.e. very far from the poles )
there should appear a sensitivity to the original ( bare ) quark - quark vertex . in the latter case
the constituent quarks and chiral fields can not be used as effective degrees of freedom .
so the crucial question is what a typical momentum transfer in the given system is . in the low - lying baryons it is below the chiral symmetry breaking scale thus justifying a use of the effective qq interactions there @xcite .
we hope that the discussion above has been transparent enough to show a dramatic difference between the initial t hooft interaction , taken literally in @xmath3 system , and its implication after iteration in the t - channel , producing meson exchange between constituent quarks .
in fact , what one needs for the chiral symmetry breaking is a scalar interaction between quarks .
any pairwise gluonic interaction between quarks in the local approximation will necessarily contain the first and second terms of ( [ hooft ] ) with fixed relative strength .
this is because of the chiral invariance .
thus all our conclusions are rather general and do not rely necessarily on t hooft interaction .
an important lesson is to see how this nonperturbative gluonic interaction , which induces the dynamical breaking of chiral symmetry , suggests an explanation of both the low - lying mesons and at the same time of baryons and the nuclear force through the effective meson exchange picure in @xmath3 systems . among the various applications of this idea ,
will be to work out how the meson - exchange interaction shifts the transition point from the chiral symmetry broken phase to the color - superconductor phase .
we also mention a recent lattice study @xcite which shows directly that the hyperfine @xmath68 splitting is mostly due to the meson - exchange interaction between quarks . another indirect evidence in favor of the picture in fig .
3 versus that in fig . 2
is that after cooling ( the cooling means that all gluonic configurations , except for instantons , are removed ) the @xmath109 splitting disappears @xcite . while the cooling does not affect the initial t hooft interaction between quarks and thus the whole s - channel ladder of fig .
2 is active , it ruins the t - channel ladder of fig .
the reason is that there are not enough antiquarks in the fock space after cooling as in quenched approximation they are mostly produced by different gluons , including perturbative ones , attached to valence quark lines ( z graphs ) .
we start this section with a short description of the effective meson - exchange interaction model , adjusted to describe baryon spectroscopy within an exact semirelativistic 3-body formulation @xcite .
the hamiltonian of ref . @xcite reads : @xmath110 @xmath111 here the relativistic form of the kinetic - energy operator is used , with @xmath112 the 3-momentum and @xmath113 the masses of the constituent quarks .
the dynamical part consists of the quark - quark interaction @xmath114 the linear pairwise confining potential @xmath115 includes both the color - electric string @xmath116 with the color factor absorbed into the string tension @xmath117 as well as a constant @xmath118 , which is large and negative , and thus effectively includes all possible spin- and flavor - independent attractive interactions between quarks , e.g. @xmath25-exchange ( [ sigfor ] ) , etc .
the flavor- and spin - dependent part of the above hamiltonian is @xmath119[3ex]{}+v_{\eta}(\vec r_{ij } ) \lambda_i^8 \lambda_j^8 + \frac{2}{3}v_{\eta'}(\vec r_{ij})\right ] \vec\sigma_i\cdot\vec\sigma_j,\end{aligned}\ ] ] @xmath120 with @xmath121 ( @xmath122 ) being the individual phenomenological meson masses , and @xmath123 the meson - quark coupling constants .
the constituent mass of the light quarks @xmath124 was fixed in @xcite to a typical value , @xmath125 mev , implied by a simple static quark model formula for the nucleon magnetic moment .
it is astonishing that the same value has been obtained in a lattice measurement @xcite .
all other parameters of the above hamiltonian can be found in ref .
@xcite . in light quark systems , like @xmath126 and @xmath17 , only the @xmath0-like ,
@xmath14-like and @xmath15-like parts of the potential ( [ voct ] ) contribute .
the @xmath0-like exchange interaction is determined by the following matrix elements : @xmath127 while the @xmath14- and @xmath15-like exchanges depend only on the spin @xmath128 of a quark pair . combining all @xmath0 , @xmath14 and @xmath15 interactions
one finds that the potential ( [ voct ] ) is most attractive at short distances in @xmath129 quark pair and essentially less attractive in the @xmath130 diquark system . in other possible color - antitriplet @xmath3 pairs
it is repulsive . applying the hamiltonian ( [ htot ] ) in a color - antitriplet @xmath3 system
, one finds a mass @xmath131 mev for a scalar diquark , @xmath2 , and a mass @xmath132 mev for an axial - vector diquark , @xmath133 . in both cases
the relative orbital angular momentum is @xmath56 , so the total angular momentum coinsides with the spin of two quarks .
these values are very similar to those obtained recently from the lattice `` diquark spectroscopy '' @xcite .
the root - mean - square ( r.m.s . )
radius of the scalar diquark is 0.354 fm and of the axial - vector one - 0.438 fm .
these radii do not include the size of the constituent quark .
it is evident that the confining interaction implies a bound diquark in the sense that there are no asymptotically free constituent quarks .
so it is very instructive to compare the mass of the above diquarks with the unphysical two - constituent - mass threshold , @xmath134 mev .
the scalar diquark mass is a few tens of mev above the threshold , which indicates that the meson - exchange part of the interaction , including @xmath118 , does not bind a diquark without the confining interaction @xmath116 .
this can also be checked explicitly . to this end
we combine the spin- and isospin - dependent interaction ( [ voct ] ) with the @xmath25-exchange potential ( [ sigfor ] ) and drop the confining potential ( [ conf ] ) .
the @xmath135 coupling constant is constrained to be equal to the @xmath104 one , as suggested by chiral symmetry , @xmath136 the sigma mass is taken to be @xmath137 , which is implied by the well known result for all njl - like interactions , @xmath138 . with these constraints
we do not find a bound diquark with any reasonable value for @xmath139 gev . only with @xmath140 gev
does a weakly bound scalar diquark appear .
if one increases the @xmath141 coupling constant by a factor 1.5 , but keeps the @xmath104 coupling constant , then we obtain a bound diquark only at @xmath142 gev .
thus we conclude that the meson - exchange interaction itself does not bind a diquark .
the next question we address in this section is whether the meson - exchange interaction binds nucleon itself , without confinement .
a - priori one can not exclude the possibility that while the diquark is unbound the three quark system will be bound because of genuine 3-body effects ( compare , e.g. , the binding energy of tritium and deuteron ) .
indeed , a full model , including confinement ( [ conf ] ) , produces a nucleon mass which is below the three constituent mass threshold , @xmath143 mev .
hence the positive contribution from the rising potential @xmath116 , 631 mev , is not big compared to the negative contribution from @xmath144 mev in combination with the negative contribution of the spin - dependent part of interaction , -750 mev .
superficially one could thus conclude that the meson - exchange part of the hamiltonian could bind nucleon without any support from confinement .
however , such an interpretation can not be taken for two reasons .
firstly , we do not know which part of the negative constant @xmath118 comes from the @xmath25-exchange , and which - from the genuine color - electric confinement , because the @xmath145-shape of the gauge - invariant 3-body confining interaction can be approximated by a sum of pairwise potentials only when some additional constant contribution is added . secondly , a perfect fit of the baryon spectrum with a quality similar to that in ref .
@xcite can be obtained with a constituent mass smaller than @xmath146 .
so we have performed a direct calculation of the nucleon , replacing the potential ( [ conf ] ) by the @xmath25-exchange potential ( [ sigfor ] ) .
we have found that for @xmath147 mev the nucleon is unbound and becomes bound at higher values of @xmath88 .
if one increases the @xmath141 by a factor 1.5 , then the nucleon becomes bound for @xmath148 mev .
these results indicate that while the nucleon is unbound with the meson - exchange potential parameters fixed of ref .
@xcite , it could be bound as soon as a spin- and isospin - independent @xmath25-like exchange interaction and/or spin- and isospin - dependent interactions are made stronger , not by a big amount .
it is evident that a description of all excited states demands the presence of confinement because all these states are much above the @xmath149 threshold .
we shall use the following set of jacobi coordinates and a coupling scheme with self - evident notation : @xmath150 @xmath151 @xmath152 @xmath153 @xmath154 let @xmath155 be a projector onto a subspace with a given value of spin @xmath156 and isospin @xmath157 of the particles 1 and 2 . the probability density for finding particles 1 and 2 in a spin - isospin state @xmath158 at a relative distance @xmath159
is given by @xmath160 where @xmath161 is an antisymmetric 3-body baryon wave function .
one can similarly define the probability density for finding the particle 3 at a distance @xmath162 from the center of mass of particles 1 and 2 @xmath163 then one can calculate the corresponding moments @xmath164 @xmath165 in table 1 we present the @xmath166 moments for @xmath126 and @xmath17 in two cases : ( i ) full model , ( ii ) no spin - dependent interaction at all ( i.e. only confinement is active ) . comparing the nucleon r.m.s .
radius , @xmath167 = 0.304 fm , with the radius of a scalar diquark , 0.354 fm , we can deduce the role of genuine 3 body effects - they make the nucleon essentially more compact than the diquark .
the empirical mean square charge radius of the proton , @xmath168 @xmath169 , consists of a few contributions : the contribution from the mean square matter radius above , the charge mean square radius of the constituent quark , the meson exchange current contribution @xcite , the proton anomalous magnetic moment contribution , etc .
the rather small value of the matter radius , obtained above , is consistent with large contributions from other sources .
for instance , the charge radius of the constituent quark should be mainly determined by the @xmath1-meson pole in the time - like region ( vector meson dominance ) and thus can be expected to be of the order @xmath170 fm .
the r.m.s .
radius of the @xmath17-resonance , @xmath171 = 0.390 fm , is larger than that of the nucleon .
this result is easy anticipate since the @xmath17-resonance wave function does not contain @xmath172 components , where the potential ( [ voct ] ) is strongly attractive at short range , and thus the size of the @xmath17-resonance is determined mainly by the weak attraction in the @xmath173 quark pairs as well as by the confining interaction .
the bigger size of @xmath17 has a well known experimental consequence : the @xmath174 electromagnetic form factor falls off faster than the nucleon elastic one .
when the meson - exchange interaction is switched off , the nucleon matter radius becomes larger , @xmath167 = 0.442 fm .
this illustrates that there is a soft gap between the scale where chiral physics starts to work and the scale where confinement is important .
the crucial role of three body effects can also be seen from the comparison of the root mean square distance between quarks in the @xmath175 quark pair in the nucleon , 0.354 fm , with the same distance in a free scalar diquark , 0.708 fm .
similarly , the three body effects and the antisymmetrization are responsible for the fact that the root mean square distance in the @xmath175 quark pair in the nucleon , 0.354 fm , is similar to that one in the @xmath176 subsystem in the nucleon , 0.387 fm , while the potential is very different in both cases .
a comparison of the two numbers above gives an idea about how unimportant clustering is in the nucleon .
it can also be seen from fig . 4 and fig . 5 where we show probability density distributions .
with a pure static `` right triangle '' 3q configuration the relation between @xmath177 and @xmath178 would be @xmath179 .
this relation is almost exactly satisfied with the @xmath17 wave function or with the @xmath180 wave function when the meson - exchange interaction is switched off . in the nucleon wave function
there is a deviation from this relation but it is not large . we thus conclude that there is no appreciable clustering in the nucleon . what is the physical reason for the absense of a significant clustering ?
the answer is that there are genuine 3-body effects and the fermi - nature of quarks do not support clustering . indeed ,
if the quarks , say , with numbers 1 and 2 form a pair @xmath175 , the antisymmetry of the wave function suggests that there are at the same time pairs with quantum numbers @xmath181 or @xmath182 ( along with other quantum numbers ) .
thus a strong attraction acts simultaneously in all quark pairs which makes the nucleon compact but not clustered much . only a much stronger and `` sharper '' interaction in the @xmath183 diquark would lead to an appreciable clustering , but at a cost that @xmath68 splitting will become enormous .
here we summarize our main conclusions .
the nonperturbative gluonic interaction between quarks , e.g. instanton - induced one , which is responsible for the dynamical breaking of chiral symmetry in qcd and thus explains the @xmath6 mass splitting , iterated in the @xmath3 t - channel implies a meson - exchange picture between constituent quarks , and through the latter also explains baryons and nuclear force .
this is a simple consequence of crossing symmetry : if one obtains pion as a solution of the bethe - salpeter equation in the quark - antiquark s - channel , then one inevitably obtains a pion - exchange in the quark - quark systems as a result of iterations in the @xmath3 t - channel .
\2 . due to ( anti)screening effects
the implications of this nonperturbative gluonic interaction in @xmath3 systems are drastically different when it is iterated only in the s - channel as compared to a more general case , when it is first iterated in the t - channel , leading to a meson exchange , and only after that iterated in the s - channel .
\3 . the effective meson - exchange interaction in @xmath3 systems
does not bind diquarks without an additional confining force and does not induce any appreciable clustering in the nucleon .
l.ya.g . thanks d. diakonov and d.o .
riska for comments .
he is indebted to kek - tanashi and tokyo institute of technology nuclear theory groups for a warm hospitality .
his work is supported by a foreign visiting guestprofessorship program of the ministry of education , science , sports and culture of japan .
work of k.v .
is supported by the otka grant no .
t17298 ( hungary ) and by the us department of energy , nuclear physics division , under contract no .
w-31 - 109-eng-39 .
r. rapp , t. schfer , e. shuryak , m. velkovsky , phys .
lett . * 81 * , 53 ( 1998 ) .
m. alford , k. rajagopal , f. wilczek , phys .
b422 * , 247 ( 1998 ) .
t. schfer and e. shuryak , rev .
phys . , * 70 * , 323 ( 1998 ) , see also references therein . l. ya .
glozman and d.o .
riska , physics reports * 268 * , 263 ( 1996 ) .
glozman , surveys in high energy physics , * 14 * , 109 ( 1999 ) ; hep - ph/9805345 .
glozman , w. plessas , k. varga , r. f. wagenbrunn , phys . rev . *
d58 * , 094030 ( 1998 ) .
m. he s , f. karsch , e. laermann , i. wetzorke , phys . rev .
* d58 * , 111502 ( 1998 ) .
u. vogl and w. weise , progr .
* 27 * , 195 ( 1991 ) .
r. alkofer and h. reinhard , chiral quark dynamics , chapter 5 , lecture notes in physics , m33 , springer , 1995 .
n. ishii , w. bentz and k. yazaki , nucl .
a587 * , 617 ( 1995 ) .
d. diakonov , in : selected topics in nonperturbative qcd , proc .
enrico fermi school , course cxxx , a. digiacomo and d. diakonov eds . ,
bologna ( 1996 ) ; hep - ph/9602375 , see also references therein . y. nambu and g. jona - lasinio , phys . rev . * 122 * , 345 ( 1961 ) ; ibid . * 124 * 246 t hooft , phys . rev . * d14 * , 3432 ( 1976 ) ; erratum : _ ibid . _ * d18 * , 2199 ( 1978 ) .
d. diakonov , h. forkel , m. lutz , phys . lett . *
b373 * , 147 ( 1996 ) .
d. diakonov , v. petrov , p. pobylitza , nucl . phys . *
b306 * , 809 ( 1988 ) .
glozman , in preparation e.v .
shuryak and j. rosner , phys . lett .
* b218 * , 72 ( 1989 ) . v. h. blask et al , z. phys . * a337 * , 327 ( 1990 ) .
glozman , hep - ph/9908207 k.f .
liu et al , phys . rev .
* d59 * 112001 ( 1999 ) .
chu , j. grandy , s. huang , j. negele , phys .
rev . * d49 * , 6039 ( 1994 ) . | it is shown that the instanton - induced interaction in qq pairs , iterated in t - channel , leads to a meson - exchange interactions between quarks . in this way
one can achieve a simultaneous understanding of low - lying mesons , baryons and the nuclear force .
the discussion is general and does not necessarily rely on the instanton - induced interaction .
any nonperturbative gluonic interaction between quarks , which is a source of the dynamical chiral symmetry breaking and explains the @xmath0 - @xmath1 mass splitting , will imply an effective meson exchange picture in baryons . due to
the ( anti)screening there is a big difference between the initial t hooft interaction and the effective meson - exchange interaction .
it is demonstrated that the effective meson - exchange interaction , adjusted to the baryon spectrum , does not bind the scalar diquark and does not induce any significant quark - diquark clustering in the nucleon because of the nontrivial role played by the pauli principle .
pacs number(s ) : 12.38.lg , 12.39.-x , 14.20.dh |
one of the main lessons from the general theory of relativity is that one can formulate theories that , in their lagrangian description , are diffeomorphism invariant .
this means that one can perform generic diffeomorphism on the spacetime manifold and the theory remains invariant . in most instances diffeo invariance is achieved by formulating the theory as an action principle where the lagrangian density is defined without the use of background structures ; it is only the dynamical fields that appear in the action . in this manner one incorporates the ` stage ' , the gravitational field , as one of the dynamical fields that one can describe .
the fact that one can write a term that captures the dynamics of the gravitational field is interesting by itself .
it is then worth exploring all the freedom available in the definition of an action principle for general relativity .
this is the main task that we shall undertake in this article .
note that we shall restrict ourselves to general relativity and shall not consider generalizations such as scalar - tensor theories nor massive gravity in our analysis .
the first question that we shall address is that of having a well posed variational principle .
this particularly ` tame ' requirement seems , however , to be sometimes overlooked in the literature . it is natural to ask why we need to have a well posed action principle if , at the end of the day , we already ` know ' what the field equations are .
while this is certainly true , one should not forget that the classical theory is only a ( very useful indeed ! ) approximation to a deeper underlying theory that must be quantum in nature .
if , for instance , we think of a quantum theory defined by some path integral , in order for this to be well defined , we need to be able to write a meaningful action for the whole space of histories , and not only for the space of classical solutions .
this observation becomes particularly vexing when the physical situation under consideration involves a spacetime region with boundaries .
one must be particularly careful to extend the formalism in order to incorporate such boundary terms .
another equally important issue in the definition of a physical theory is that of the choice of fundamental variables , specially when gauge symmetries are present .
again , even when the space of solutions might coincide for two formulations , the corresponding actions will in general be different and that will certainly have an effect in the path integral formulation of the quantum theory . in the case of general relativity , the better known formulation is of course in terms of a metric tensor @xmath3 , satisfying second order ( einstein ) equations .
but there are other choices of variables that yield alternative descriptions . here
we shall consider one of those possibilities .
in particular , the choice we shall make is motivated by writing the theory as a _ local _ gauge theory under the lorentz group . it is well known that one can either obtain einstein equations of motion by means of the einstein hilbert action or in terms of the so called palatini action , a first order action in terms of tetrads @xmath4 and a connection @xmath5 valued on a lie algebra of @xmath6 ( see .
e.g. @xcite and @xcite ) and an affine connection @xmath7 @xcite .
the action we are considering here , in the so called vielbein " formalism , was developed in @xcite and in @xcite in the canonical formulation . ] .
also , it is known that we can have a generalization of this action by adding a term , the holst term , that still gives us the same equations of motion and also allows us to express the theory in terms of real su(2 ) connections in its canonical description ( see . e.g. @xcite and @xcite ) .
this action , known as the holst action , is the starting point of loop quantum gravity and some spin foam models . in the same first order scheme one can look for the most general diffeomorphism invariant first order action that classically describes general relativity , which can be written as the palatini action ( including the holst term ) plus topological contributions , namely , the pontryagin , euler and nieh - yan terms .
furthermore , if the spacetime region we are considering possesses boundaries one might have to add extra terms ( apart from the topological terms that can also be seen as boundary terms ) to the action principle .
thus , the most general first order action for gravity has the form , @xmath8 = s_{\rm palatini } + s_{\rm holst term } + s_{\rm pontryagin } + s_{\rm euler } + s_{\rm nieh - yan } + s_{\rm boundary}.\ ] ] it is noteworthy to emphasize that in the standard textbook treatment of hamiltonian systems one usually considers compact spaces without boundary , so there is no need to worry about the boundary terms that come from the integration by parts in the variational principle .
but if one is interested in spacetimes with boundaries we can no longer neglect these boundary terms and it is mandatory to analyze them carefully .
in order to properly study this action in the whole space - time with boundaries , we need the action principle to be well posed , i.e. we want the action to be _ differentiable and finite _ under the appropriate boundary conditions , and under the most general variations compatible with the boundary conditions .
it has been shown that under appropriate boundary conditions , the palatini action plus a boundary term provides a well posed action principle , that is , it is differentiable and finite .
furthermore , in @xcite the analysis for asymptotically flat boundary conditions was extended to include the holst term . here
we will refer to this well posed holst action as the generalized holst action ( gha ) . in order to explore some properties of the theories defined by an action principle
, the covariant hamiltonian formalism seems to be particularly appropriate ( see , e.g. @xcite , @xcite and @xcite ) . in this formalism
, one can introduce the standard hamiltonian structures such as a phase space , symplectic structure , canonical transformations etc , without the need of a @xmath9 decomposition of the theory .
all the relevant objects are _ covariant_. the most attractive feature of this formalism is that one can find all these structures in a unique fashion given the action principle .
furthermore one can , in a ` canonical ' way , find conserved quantities . on the one hand one can derive hamiltonian generators of canonical transformations and ,
on the other hand , noetherian conserved quantities associated to symmetries .
one important and interesting issue is to understand the precise relation between these two sets of quantities .
the purpose of this article is to explore all these issues in a systematic way .
more concretely , we have three main goals .
the first one is to explore the well - posedness of the action principle with generic boundary terms .
for that we shall study two sets of boundary conditions that are physically interesting ; as outer boundary we shall consider configurations that are asymptotically flat , and in an inner boundary , those histories that satisfy isolated horizon boundary conditions .
the second objective is to explore the most basic structures in the covariant phase space formulation .
more precisely , we shall study the existence of the symplectic structure as a finite quantity and its dependence on the various topological and boundary terms . finally , the last goal is to explore the different conserved quantities that can be defined .
concretely , we shall consider hamiltonian conserved charges both at infinity and at the horizon .
finally we shall construct the associated noetherian conserved current and charges .
in both cases we shall study in detail how these quantities depend on the existence of the boundary terms that make the action well defined .
as we shall show , while the hamiltonian charges are insensitive to those quantities , the noether charges _ do _ depend on the form of the boundary terms added .
while some of these results are not new and have appeared somewhere else , we have several new results and clarifications of several issues . equally important
is the fact that in no publication have all the results available been put in a coherent and systematic fashion .
thus , the final goal of this article is to fill this gap and to present the subject in a pedagogical and self - contained manner .
the structure of the paper is as follows : in section [ sec:2 ] we review what it means for an action principle to be well posed , which is when it is finite and differentiable . in section [ sec:3 ]
we use some results discussed in the previous section , to review the covariant hamiltonian formalism taking enough care in the cases when the spacetime has boundaries .
we begin by defining the covariant phase space and its relation with the canonical phase space .
then we introduce the symplectic structure with its ambiguities and its dependence on boundary terms in the action .
finally we define the symplectic current structure , and the hamiltonian and noether charges . in section [ sec:4 ]
we use the covariant hamiltonian formalism to study the action of eq .
( [ 1 ] ) .
we find the generic boundary terms that appear when we vary the different components of the action . in section [ sec:5 ]
we consider particular choices of boundary conditions in the action principle .
in particular we study spacetimes with boundaries : asymptotically flatness at the outer boundaries , and an isolated horizon as an internal one . in section [ sec:6 ]
we study symmetries and their generators for both sets of boundary conditions .
in particular we first compute the hamiltonian conserved charges , and in the second part , the corresponding noetherian quantities are found .
we comment on the difference between them .
we summarize and provide some discussion in the final section [ sec:7 ] .
in this section we review the action principle that plays a fundamental role in the formulation of physical theories . in order to do that we need to be precise about what it means to have a well posed variational principle .
in particular , there are two aspects to it .
the first one is to define the action by itself .
this is done in the first part of this section . in the second part
, we introduce the variational principle that states that physical configurations will be those that make the action stationary .
in particular , we entertain the possibility that the spacetime region under consideration has non - trivial boundaries and that the allowed field configurations are allowed to vary on these boundaries .
these new features require an extension of the standard , textbook , treatment . in particle mechanics
the dynamics is specified by some action , which is a _ function _ of the trajectories of the particle . in turn , the action @xmath10 is the time integral of the lagriangian function @xmath11 that generically depends on the coordinates and velocities of the particles . in field theory
the dynamical variables , the fields , are geometrical objects defined on spacetime ; now the lagrangian has as domain this function space . in both cases ,
this type of objects are known as functionals . in order to properly define the action we will review what is a functional and some of its relevant properties .
a _ functional _ is a map from a _ normed space _ ( a vector space with a non - negative real - valued norm ) into its underlying field , which in physical applications is the field of the real numbers .
this vector space is normally a functional space , which is why sometimes a functional is considered as a function of a function .
a special class of functionals are the definite integrals that define an action by an expression of the form , @xmath12 = \int_{\mathcal{m } } \mathcal{l } ( \phi^{\alpha } , \nabla\phi^\alpha, ...
,\nabla^n\phi^{\alpha})\,\d^4 v,\ ] ] where @xmath13 are fields on spacetime , @xmath14 , @xmath15 is an abstract label for spacetime and internal indices , @xmath16 their first derivatives , and @xmath17 their @xmath18 derivatives , and @xmath19 a volume element on spacetime .
this integral @xmath20 $ ] maps a field history @xmath21 into a real number if the lagrangian density @xmath22 is real - valued .
prior to checking the well posedness of this action , we will review what it means for an action to be finite and differentiable .
we say that an action is _ finite _ iff the integral that defines it is convergent or has a finite value when evaluated in _ histories _ compatible with the boundary conditions . as the minimum action principle states , the classical trajectories followed by the system are those for which the action is a stationary point .
this means that , to first order , the variations of the action vanish .
as is well known , the origin of this emphasis on extremal histories comes from the path integral formalism where one can show that trajectories that extremise the action contribute the most to the path integral .
first , let us consider some definitions : let @xmath23 be a normed space of function .
a functional @xmath24 is called differentiable if we can write the finite change of the action , under the variation @xmath25 , as f [ + ] -f [ ] = f + r , where @xmath26 ( we are assuming here that vectors @xmath27 belong to the space @xmath23 , so it is a linear space ) .
the quantity @xmath28 $ ] depends linearly on @xmath29 , and @xmath30 = { \mathcal o}((\de\phi)^2)$ ] .
the linear part of the increment , @xmath31 , is called the _ variation of the funcional _ @xmath32 ( along @xmath29 ) .
a stationary point @xmath33 of a differentiable funcional @xmath34 $ ] is a function @xmath33 such that @xmath35=0 $ ] for all @xmath29 . as is standard in theoretical physics , we begin with a basic assumption : the dynamics is specified by an action . in most field theories
the action depends only on the fundamental fields and their first derivatives .
interestingly , this is not the case for the einstein hilbert action of general relativity , but it _ is _ true for first order formulations of general relativity , which is the case that we shall analyze in the present work . in general
, we can define an action on a four dimensional manifold @xmath14 , with topology @xmath36 depending on the fields , @xmath37 and their first derivatives , @xmath38 .
thus , we have @xmath39 = \int_{\mathcal{m } } \mathcal{l } \left ( \phi^{\alpha } , \nabla_{\mu } \phi^{\alpha } \right)\ , \d^4 v\ , .\ ] ] its variation @xmath40 is the linear part of _ ^4 v , where @xmath41 .
it follows that @xmath42 = \int_{\mathcal{m } } \left [ \frac{\partial \mathcal{l } } { \partial \phi^{\alpha } } - \nabla_{\mu } \frac{\partial \mathcal{l } } { \partial ( \nabla_{\mu } \phi^{\alpha } ) } \right ] \delta \phi^{\alpha } \,\ , \d^4 v + \int_{\mathcal{m } } \nabla_{\mu } \left ( \frac{\partial \mathcal{l } } { \partial ( \nabla_{\mu } \phi^{\alpha } ) } \ , \delta \phi^{\alpha } \right ) \d^4 v\ , , \ ] ] where we have integrated by parts to obtain the second term .
let us denote the integrand of the first term as : @xmath43 .
note that the second term on the right hand side is a divergence so we can write it as a boundary term using stokes theorem , @xmath44 where we have introduced the quantity @xmath45 that will be relevant in sections to follow .
note that the quantity @xmath46 $ ] can be interpeted as the directional derivative of the funtion(al ) @xmath10 along the vector @xmath29 .
let us introduce the simbol @xmath47 to denote the exterior derivative on the functional space @xmath48 .
then , we can write @xmath49=\ed\ , s(\de\phi)=\de\phi(s)$ ] , where the last equality employs the standard convention of representing the vector field , @xmath29 , acting on the function @xmath10 . as we mentioned before ,
if we want to derive in a consistent way the equations of motion for the system , the action must be differentiable .
in particular , this means that we need the boundary term ( [ standardsymplecticpotential ] ) to be zero . to simply demand that @xmath50 , as is usually done in introductory textbooks ,
becomes too restrictive if we want to allow _ all _ the variations @xmath51 which preserve appropiate boundary conditions and not just variations of compact support .
thus , requiring the action to be stationary with respect to all compatible variations should yield precisely the classical equations of motion , with the respective boundary term vanishing on any allowed variation .
let us now consider the case in which the spacetime region @xmath52 , where the action is defined , has a boundary @xmath53
. this could be the case , for example , is we have @xmath54 and we consider certain asymptotic boundary conditions .
we could then regard ` infinity ' as a boundary of the spacetime region .
another possibility is to consider an internal boundary , as would be the case when there is a black hole present .
now , given an action principle and boundary conditions on the fields , a natural question may arise , on whether the action principle will be well posed .
so far there is no general answer , but there are examples where the introduction of a boundary term is needed to make the action principle well defined , as we shall show in the examples below .
let us then keep the discussion open and consider a generic action principle that we assume to be well defined in a region with boundaries , and with possible contributions to the action by boundary terms .
therefore , the action of such a well posed variational principle will look like , @xmath55 = \int_{\mathcal{m } } \mathcal{l } \left ( \phi^{\alpha } , \nabla_{\mu } \phi^{\alpha } \right)\ , \d^4 v + \int_{\partial \mathcal{m } } \varphi ( \phi^{\alpha } ) \,\ , \d^3 v\ , , \ ] ] where we have considered the possibility that there is contribution to the action coming from the boundary @xmath53 .
thus , the variation of this _ extended _ action becomes , @xmath56 = \int_{\mathcal{m } } e_{\alpha}\ ; \delta \phi^{\alpha } \ ; \d^4 v + \int_{\partial \mathcal{m } } \theta ( \phi^{\alpha } , \delta \phi^{\alpha } ) \ ; \d^3 v + \int_{\partial \mathcal{m } } \delta \varphi ( \phi^{\alpha } ) \ ; \d^3 v\ , .\ ] ] the action principle will be well posed if the first term is finite and @xmath57 is a boundary term that makes the action well defined under appropriate boundary conditions .
that is , when the action is evaluated along _ histories _ that are compatible with the boundary conditions , the numerical value of the integral should be finite , and in the variation ( [ generalwellposedvarprin ] ) , the contribution from the boundary terms must vanish .
now , asking @xmath46=0 $ ] , for arbitrary variations @xmath27 of the fields , implies that the fields must satisfy @xmath58the _ euler - lagrange _ equations of motion .
note that in the `` standard approach '' , i.e. when one simply considers variations , say , of compact support such that @xmath50 , we can always add a term of the form @xmath59 to the lagrangian density , @xmath60 with @xmath61 arbitrary .
the relevant fact here is that this term will _ not _ modify the equations of motion since the variation of the action becomes , @xmath62 thus , by the boundary conditions , @xmath50 , the second term of the right - hand side vanishes , that is , @xmath63 .
therefore , it does not matter which boundary term we add to the action ; it will not modify the equations of motion . on the contrary ,
when one considers variational principles of the form ( [ generalwellposedaction ] ) , consistent with arbitrary ( compatible ) variations in spacetime regions with boundaries , we can not just add arbitrary total divergences / boundary term to the action , but only those that preserve the action principle well - posedness , in the sense mentioned before .
adding to the action any other term that does not satisfy this condition will spoil the differentiability properties of the action and , therefore , one would not obtain the equations of motion in a consistent manner .
this concludes our review of the action principle .
let us now recall how one can get a consistent covariant hamiltonian formulation , once the action principle at hand is well posed .
in this section we give a self - contained review of the _ covariant hamiltonian formalism _
( chf ) taking special care of the cases where boundaries are present .
if the theory under study has a _ well posed initial value formulation _ , then , given the initial data we have a unique solution to the equations of motion . in this way we have an isomorphism @xmath64 between the space of solutions to the equations of motion , @xmath2 , and the space of all valid initial data , the ` canonical phase space ' @xmath65 . in this even dimensional space , particles ,
we need to specify as initial data their initial positions and velocities , so the space of all possible initial data is an even dimensional space .
we can easily extend this to field theory .
] we can construct a nondegenerate , closed @xmath66form @xmath67 , the symplectic form .
together , the phase space and the symplectic form constitute a symplectic manifold @xmath68 .
we can bring the symplectic structure to the space of solutions , via the pullback @xmath69 of @xmath67 and define a corresponding 2-form on @xmath2 .
in this way the space of solutions is equipped with a _
symplectic form , @xmath70 , since the mapping is independent of the reference cauchy surface one is using to define @xmath64 .
together , the space of solutions and its symplectic structure @xmath71 are known as the _ covariant phase space _ ( cps ) . however , most of the field theories of interest present _ gauge _ symmetries . this fact
is reflected on the symplectic form @xmath70 , making it _
degenerate_. when this is the case , @xmath70 is only a pre - symplectic form , to emphasize the degeneracy .
it is only after one gets rid of this degeneracy , by means of an appropriate quotient , that one recovers a physical non - degenerate symplectic structure .
this section has three parts . in the first one ,
we introduce the relevant structure in the definition of the covariant phase space , starting from the action principle .
in particular , we see that boundary terms that appear in the ` variation ' of the action are of particular relevance to the construction of the symplectic structure .
we shall pay special attention to the presence of boundary terms in the original action and how that gets reflected in the hamiltonian formulation . in the second part
, we recall the issue of symmetries of the theory .
that is , when there are certain symmetries of the underlying spacetime , these get reflected in the hamiltonian formalism . of particular relevance
is the construction of the corresponding conserved quantities , that are both conserved and play an important role of being the generators of such symmetries .
in particular we focus our attention on the symmetries generated by certain vector fields , closely related to the issue of diffeomorphism invariance . in the third part we compare and contrast these hamiltonian conserved quantities with the so - called noether symmetries and charges .
we show how they are related and comment on the fact that , contrary to the hamiltonian charges , the corresponding ` noetherian ' quantities _ do _ depend on the existence of boundary terms in the original action .
in this part we shall introduce the relevant objects that define the covariant phase space . before proceeding we shall make some remarks regarding notation .
it has proved to be useful to use differential forms to deal with certain diffeomorphism invariant theories , and we shall do that here . however , when working with differential forms in field theories one has to distinguish between the exterior derivative @xmath47 in the infinite dimensional covariant phase space , and the ` standard ' exterior derivative on the spacetime manifold , denoted by @xmath72 . in this context ,
differential forms in the cps act on vectors tangent to the space of solutions @xmath2 .
we use @xmath73 or @xmath74 to denote tangent vectors , to be consistent with the standard notation used in the literature .
we hope that no confusion should arise by such a choice .
let us now recall some basic constructions on the covariant phase space .
taking as starting point an action principle , let us first consider the action without any additional boundary term as discussed in the previous section , @xmath75 = \int_{\mathcal{m } } \mathbf{l}\ , , \ ] ] we can consider the lagrangian density , @xmath76 , as a @xmath77form and the fields @xmath78 as certain @xmath79forms ( with @xmath80 ) in the @xmath77dimensional spacetime manifold . recall that in the previous section we used @xmath15 as a generic ( abstract ) index that could be space - time or internal .
in the language of forms the spacetime index referring to the nature of the object in space - time will not appear explicitly , so we are left only with internal indices that we shall denote with @xmath81 to distinguish them from spacetime indices @xmath82 .
then , the variation of the action can be written as ( [ generalwellposedvarprin ] ) or , equivalently in terms of forms as , @xmath83 where @xmath84 are the euler - lagrange equations of motion and @xmath85 is an arbitrary vector , that can be thought to point ` in the direction that @xmath78 changes ' .
note that we are using @xmath85 and @xmath73 , to denote _ the same _ object .
for simplicity in the notation , sometimes the @xmath78 part is dropped out .
here we wrote both for clarity . as we mentioned in the previous section
, the second term of the rhs is obtained after integration by parts , and using stokes theorem it can be written as , @xmath86 this term can be seen as a @xmath87form in the covariant phase space , acting on vectors @xmath85 and returning a real number .
also it can be seen as a potential for the symplectic structure , that we already mentioned in the preamble of this section and shall define below .
for such a reason , this term , @xmath88 is known as _
symplectic potential _ , and the integrand , @xmath89 , as the _ symplectic potential current_. note that from eqs .
( [ varactformswithoutboundary ] ) and ( [ symppotentialwhitoutboundary ] ) , in the space of solutions @xmath90 , @xmath91 . as we pointed out in the previous section , this action may not be well defined , and the introduction of a boundary term could be needed . in that case
the action becomes , @xmath92 = \int_{\mathcal{m } } \mathbf{l } + \int_{\mathcal{m } } \d \varphi ( \phi^{a})\ , , \ ] ] and we have a well posed variational principle , @xmath93.\ ] ] when we have added a boundary term , the symplectic potential associated to this well posed action changes as @xmath94 , equivalently we can consider , @xmath95 .\label{sympl_poten_boundary_term}\ ] ] from this equation we can see that besides the boundary term added to the action , to make it well defined , we can always add a term , @xmath96 , to the symplectic potential current that will not change @xmath97 .
thus , the most general symplectic potential can be written as , @xmath98 = : \int_{\partial \mathcal{m } } \tilde{\theta } ( \phi^{a},\delta \phi^{a } ) .\ ] ] now , we take the exterior derivative of the symplectic potential , @xmath99 , acting on tangent vectors @xmath100 and @xmath101 at a point @xmath102 of the phase space , @xmath103})\ , .\ ] ] from this expression we can define a space - time @xmath104form , _ the symplectic current _
@xmath105 , to be the integrand of the rhs of ( [ edtheta ] ) , @xmath106 in particular , when we have added a boundary term to the action , and taking into account the ambiguities , the symplectic current becomes , @xmath107 } ) + \delta_{[1 } \delta_{2 ] } \varphi + \delta_{[1}\d y ( \delta_{2]})\bigr)\ , . \ ] ] now , the term @xmath108 } \varphi $ ] vanishes by antisymmetry , because @xmath100 and @xmath101 commute when acting on functions .
note that the last term of the rhs of ( [ defj ] ) can be written as @xmath109})$ ] due to @xmath72 and @xmath110 commuting . since @xmath72 and @xmath47 act on different spaces , the spacetime and the space of fields , respectively , they are independent . in this way
@xmath105 is determined by the symplectic structure up to , @xmath111 .
this ambiguity will appear explicitly in the examples that we shall consider below .
therefore we conclude that , _
when we add a boundary term to the original action it will not change the symplectic current _ , and this result holds independently of the specific boundary conditions .
recall that in the space of solutions , @xmath112 , therefore from eqs .
( [ edtheta ] ) and ( [ defj ] ) , @xmath113 } ) = \int_{\mathcal{m } } \d j(\delta_{1 } , \delta_{2}).\ ] ] since we are integrating over _ any _ region @xmath14 , we can conclude that @xmath114 is closed , i.e. @xmath115 .
note that @xmath116 } ) + \d \chi ( \delta_{1 } , \delta_{2 } ) ] = 2 \delta_{[1 } \d \theta ( \delta_{2]})$ ] depends only on @xmath45 .
using stokes theorem , and taking into account the orientation of @xmath14 ( see fig .
[ regionm ] ) , we have @xmath117 where @xmath14 is bounded by @xmath118 , @xmath119 and @xmath120 are space - like slices , @xmath121 is an inner boundary and @xmath122 an outer boundary .
.,width=302 ] now consider the following two possible scenario : first , consider the case when there is no internal boundary , only a boundary @xmath122 at infinity . in some examples
the asymptotic conditions ensures that the integral @xmath123 vanishes , in that case from ( [ intjzero ] ) and taking into account the orientations of @xmath124 , @xmath125 which implies that @xmath126 is independent of the cauchy surface .
this allows us to define a _ conserved _ pre - symplectic form over an arbitrary space like surface @xmath127 , @xmath128
let us now consider the case when we have an internal boundary .
now , the integral @xmath129 may no longer vanish under the boundary conditions , as is the case with the isolated horizon boundary conditions ( more about this below ) .
the `` next best thing '' is that this integral is `` controllable '' .
let us be more specific .
if , after imposing boundary conditions , the integral takes the form , @xmath130 we can still define a _ conserved _ pre - symplectic structure . from ( [ intjzero ] ) , and assuming the integral over the outer boundary vanishes , we now have @xmath131 where @xmath132 and @xmath133 are the intersections of space - like surfaces @xmath119 and @xmath120 with the inner boundary @xmath121 , respectively
. therefore we can define the _ conserved _ pre - symplectic structure as , @xmath134 note that by construction , the two form @xmath135 is closed , so it is justified to call it a ( pre-)symplectic structure .
let us now explore how the covariant hamiltonian formulation can deal with the existence of symmetries , and their associated conserved quantities .
before that , let us recall the standard notion of a hamiltonian vector field ( hvf ) in hamiltonian dynamics .
a hamiltonian vector field @xmath136 is defined as a symmetry of the symplectic structure , namely @xmath137 from this condition and the fact that @xmath138 we have , @xmath139 where @xmath140 means the _ contraction _ of the 2-form @xmath70 with the vector field @xmath136 .
note that @xmath141 is a one - form on @xmath2 acting on an arbitrary vector @xmath73 .
we can denote it as @xmath142 . from the previous equation
we can see that @xmath143 is closed , @xmath144 .
it follows from ( [ lieomegazero ] ) and from the poincar lemma that locally ( on the cps ) , there exists a function @xmath145 such that @xmath146 .
we call this function , @xmath145 , the _ hamiltonian _
, that generates the infinitesimal canonical transformation defined by @xmath136 .
furthermore , and by its own definition , @xmath145 is a _ conserved quantity _ along the flow generated by @xmath136 .
note that the directional derivative of the hamiltonian @xmath145 , along an arbitrary vector @xmath73 can be written in several ways , @xmath147 some of which will be used in - distinctively in what follows .
so far this vector field @xmath136 is an arbitrary hamiltonian vector field on @xmath2 . later
on we will relate it to certain space - time symmetries .
for instance , for field theories that possess a symmetry group , such as the poincare group for field theories on minkowski spacetime , there will be corresponding hamiltonian vector fields associated to the generators of the symmetry group .
in this article we are interested in exploring gravity theories that are diffeomorphism invariant .
that is , such that the diffeomorphism group acts as a ( kinematical ) symmetry of the action .
of particular relevance is then to understand the role that these symmetries have in the hamiltonian formulation .
in particular , one expects that diffeomorphisms play the role of _ gauge _ symmetries of the theory
. however , the precise form in which diffeomorphisms can be regarded as gauge or not , depends on the details of the theory , and is dictated by the properties of the corresponding hamiltonian vector fields .
another important issue is to separate those diffeomorphisms that are gauge from those that represent truly physical canonical transformations that change the system .
those true motions could then be associated to symmetries of the theory .
for instance , in the case of asymptotically flat spacetimes , _ some _ diffeomorphism are regarded as gauge , while others represent nontrivial transformations at infinity and can be associated to the generators of the poincar group . in the case when the vector field @xmath136 generates time evolution , one expects @xmath145 to be related to the energy , the adm energy at infinity .
other conserved , hamiltonian charges can thus be found , and correspond to the generators of the asymptotic symmetries of the theory . in what follows
we shall explore the aspects of the theory that allow us to separate the notion of gauge from standard symmetries of the theory .
in the standard treatment of constrained systems , one starts out with the kinematical phase space @xmath148 , and there exists a constrained surface @xmath149 consisting of points that satisfy the constraints present in the theory .
one then notices that the pullback of @xmath70 , the symplectic structure to @xmath149 is degenerate ( for first class constraints ) .
these degenerate directions represent the gauge directions where two points are physically indistinguishable . in the covariant hamiltonian formulation we are considering here , the starting point is the space @xmath2 of solutions to _ all _ the equations of motion , where a ( pre-)symplectic structure is naturally defined , as we saw before .
we call this a pre - symplectic structure since it might be degenerate .
we say that @xmath135 is degenerate if there exist vectors @xmath150 such that @xmath151 for all @xmath152 .
we call @xmath150 a degenerate direction ( or an element of the kernel of @xmath135 ) .
if @xmath135 is degenerate we have a gauge system , with a gauge submanifold generated by the degenerate directions @xmath150 ( it is immediate to see that they satisfy the local integrability conditions to generate a submanifold ) . note that since we are on the space of solutions to the field equations , tangent vectors @xmath152 to @xmath2 must be solutions to the _ linearized _ equations of motion .
since the degenerate directions @xmath153 generate infinitesimal gauge transformations , configurations @xmath154 and @xmath155 on @xmath156 , related by such transformations , are physically indistinguishable .
that is , @xmath157 and , therefore , the quotient @xmath158 constitutes the physical phase space of the system .
it is only in the reduced phase space @xmath159 that one can define a non - degenerate symplectic structure @xmath70 . in the next subsection
we explain how vector fields are the infinitesimal generators of transformations on the space - time in general
. then we will point out when these transformations are diffeomorphisms and moreover , when these are also gauge symmetries of the system .
let us start by recalling the standard notion of a diffeomorphism on the manifold @xmath160 .
later on , we shall see how , for diffeomorphism invariant theories , the induced action on phase space of certain diffeomorphisms becomes gauge transformations .
there is a one - to - one relation between vector fields on a manifold and families of transformations of the manifold onto itself .
let @xmath161 be a one - parameter group of transformations on @xmath14 , the map @xmath162 , defined by @xmath163 , is a differentiable mapping .
if @xmath164 is the infinitesimal generator of @xmath161 and @xmath165 , @xmath166 also belongs to @xmath167 ; then the lie derivative of @xmath168 along @xmath169 , @xmath170 , represents the rate of change of the function @xmath168 under the family of transformations @xmath171 .
that is , the vector field @xmath169 is the generator of infinitesimal diffeomorphisms .
now , given such a vector field , a natural question is whether there exists a vector field @xmath172 on the cps that represents the induced action of the infinitesimal diffeos ?
as one can easily see , the answer is in the affirmative . in order to see that ,
let us go back a bit to section [ sec:2 ] .
the action is defined on the space of histories ( the space of all possible configurations ) and , after taking the variation , the vectors @xmath173 lie on the tangent space to the space of histories .
it is only after we restrict ourselves to the space of solutions @xmath2 , that @xmath174 .
now these @xmath85 represent _ any _ vector on @xmath175 ( tangent space to @xmath2 at the point @xmath78 ) .
as we already mentioned , these @xmath85 can be seen as `` small changes '' in the fields . what happens if we want the infinitesimal change of fields to be generated by a particular group of transformations ( e.g. spatial translations , boosts , rotations , etc ) ?
there is indeed a preferred tangent vector for the kind of theories we are considering .
given @xmath169 , consider _ ^a:=_^a . from the geometric perspective , this is the natural candidate vector field to represent the induced action of infinitesimal diffeomorphisms on @xmath2 .
the first question is whether such objects are indeed tangent vectors to @xmath2 .
it is easy to see that , _ for kinematical diffeomorphism invariant theories , lie derivatives satisfy the linearized equations of motion_. is a symmetry of the background spacetime .
] of course , in the presence of boundaries such vectors must preserve the boundary conditions of the theory in order to be admissible ( more about this below ) . for instance
, in the case of asymptotically flat boundary conditions , the allowed vector fields should preserve the asymptotic conditions .
let us suppose that we have prescribed the phase space and pre - symplectic structure @xmath135 , and a vector field @xmath176 .
the question we would like to pose is : when is such vector a degenerate direction of @xmath135 ?
the equation that such vector @xmath177 must satisfy is then : | ( _ , ) = 0 , .
[ cond - chida ] this equation will , as we shall see in detail below once we consider specific boundary conditions , impose some conditions on the behaviour of @xmath169 on the boundaries .
an important signature of diffeomorphism invariant theories is that eq.([cond - chida ] ) _ only has contributions from the boundaries_. thus , the vanishing of such terms will depend on the behaviour of @xmath169 there . in particular , if @xmath178 on the boundary , the corresponding vector field is guaranteed to be a degenerate direction and therefore to generate _ gauge _ transformations . in some instances , non vanishing vectors at the boundary also satisfy eq .
( [ cond - chida ] ) and therefore define gauge directions .
let us now consider the case when @xmath169 is non vanishing on @xmath179 and eq .
( [ cond - chida ] ) is not zero . in that case
, we should have |(,_)= h _ ( ) = h _ , [ ecuacion - chida ] for some function @xmath180
. this function will be the generator of the symplectic transformation generated by @xmath177 .
in other words , @xmath180 is the _
hamiltonian conserved charge _
associated to the symmetry generated by @xmath169 .
remark : one should make sure that eq .
( [ ecuacion - chida ] ) is indeed well defined , given the degeneracy of @xmath135 . in order to see that , note that one can add to @xmath177 an arbitrary ` gauge vector ' @xmath136 and the result in the same : @xmath181 .
therefore , if such function @xmath180 exists ( and we know that , locally , it does ) , it is insensitive to the existence of the gauge directions so it must be constant along those directions and , therefore , projectable to @xmath159 . thus , one can conclude that even when @xmath180 is defined through a degenerate pre - symplectic structure , it is indeed a physical observable defined on the reduced phase space .
this concludes our review of the covariant phase space methods and the definition of gauge and hamiltonian conserved charges for diffeomorphism invariant theories . in the next part we shall revisit another aspect of symmetries on covariant theories , namely the existence of noether conserved quantities , which are also associated to symmetries of field theories .
let us briefly review some results about the noether current 3-form @xmath182 and its relation to the symplectic current @xmath114 .
for that , we shall rely on @xcite .
we know that to any lagrangian theory invariant under diffeomorphisms we can associate a corresponding noether current 3-form .
consider infinitesimal diffeomorphism generated by a vector field @xmath169 .
these diffeomorphisms induce the infinitesimal change of fields , given by @xmath183 . from ( [ varactformswithoutboundary ] )
it follows that the corresponding change in the lagrangian four - form is given by _ = e_a_^a + ( ^a , _
^a ) .[3.1 ] on the other hand , using cartan s formula , we obtain _ = + ( ) =( ) , [ 3.2 ] since @xmath184 , in a four - dimensional spacetime .
now , we can define a _ noether current 3-form _ as j_n ( _ ) = ( _ ) - , [ noether ] where we are using the simplified notation @xmath185 . from the equations ( [ 3.1 ] ) and ( [ 3.2 ] ) it follows that on the space of solutions , @xmath186 , so at least locally one can define a corresponding noether charge density 2-form @xmath187 relative to @xmath169 as j_n ( _ ) = q _ .[nch ] following @xcite , the integral of @xmath187 over some compact surface @xmath10 is the noether charge of @xmath10 relative to @xmath169 . as we saw in the previous chapter
there are ambiguities in the definition of @xmath45 ( [ generaltheta ] ) , that produce ambiguities in @xmath187 .
as we saw in the section [ sec:3.1 ] , @xmath45 is defined up to an exact form : @xmath188 .
also , the change in lagrangian @xmath189 produces the change @xmath190 .
these transformations affect the symplectic current in the following way j(_1,_2)j(_1,_2)+(_2 y(_1)-_1 y(_2 ) ) .
the contribution of @xmath161 vanishes , as before , and as we have shown in section [ sec:3.1 ] .
the above transformation leaves invariant the symplectic structure .
it is easy to see that the two changes , generated by @xmath191 and @xmath161 contribute to the following change of noether current 3-form [ noether_current_change ] j_n(_)j_n(_)+y(_)+_- , and the corresponding noether charge 2-form changes as [ noether_charge_change ] q_q_+y(_)++z . the last term in the previous expression is due to the ambiguity present in ( [ nch ] ) .
this arbitrariness in @xmath187 was used in @xcite to show that the noether charge form of a general theory of gravity arising from a diffeomorphism invariant lagrangian , in the second order formalism , can be decomposed in a particular way . since @xmath186 it follows , as in ( [ intjzero ] ) , that @xmath192 and we see that if @xmath193 then the previous expression implies the existence of the conserved quantity ( independent on the choice of @xmath127 ) , @xmath194 note that the above results are valid only on shell . in the covariant phase space , and for @xmath169 arbitrary and fixed
, we have @xcite j_n ( _ ) = ( _ ) -=(_)- ( ) . since , @xmath195 and @xmath196 by the definition of the symplectic current @xmath114 ( [ defj ] ) , it follows that the relation between the symplectic current @xmath114 and the noether current 3-form @xmath182 is given by j(,_)=j_n ( _ ) - ( ( ) ) .[nc ] we shall use this relation in the following sections , for the various actions that describe first order general relativity , to clarify the relation between the hamiltonian and noether charges .
we shall see that , in general , a noether charge does not correspond to a hamiltonian charge generating symmetries of the phase space .
as already mentioned in the introduction , we shall consider the most general action for four - dimensional gravity in the first order formalism . the choice of basic variables is the following : a pair of co - tetrads @xmath4 and a lorentz @xmath6 connection @xmath197 on the spacetime @xmath14 , possibly with a boundary . in order for the action to be physically relevant
, it should reproduce the equations of motion for general relativity and be : 1 ) differentiable , 2 ) finite on the configurations with a given asymptotic behaviour and 3 ) invariant under diffeomorphisms and local internal lorentz transformations .
the most general action that gives the desired equations of motion and is compatible with the symmetries of the theory is given by the combination of palatini action , @xmath198 , holst term , @xmath199 , and three topological terms , pontryagin , @xmath200 , euler , @xmath201 , and nieh - yan , @xmath202 , invariants .
as we shall see , the palatini term contains the information of the ordinary einstein - hilbert 2nd order action , so it represents the backbone of the formalism . since we are considering a spacetime region @xmath14 with boundaries , one should pay special attention to boundary conditions . for instance
, it turns out that the palatini action , as well as holst and nieh - yan terms are not differentiable for asymptotically flat spacetimes , and appropriate boundary terms should be provided .
this section has four parts . in those subsections we are going to analyze , one by one , all of the terms of the action .
we shall take the corresponding variation of the terms and identify both their contributions to the equations of motion and to the symplectic current .
since we are not considering yet any particular boundary conditions , the results of this section are generic .
let us start by considering the palatini action with boundary term is given by @xcite , @xmath203 where @xmath204 , @xmath205 , @xmath206 is a curvature two - form of the connection @xmath207 and , as before , @xmath118 .
the boundary term is not manifestly gauge invariant , but , as pointed out in @xcite , it is effectively gauge invariant on the spacelike surfaces @xmath208 and @xmath209 and also in the asymptotic region @xmath122 .
this is due to the fact that the only allowed gauge transformations that preserve the asymptotic conditions are such that the boundary terms remain invariant .
let us consider first the behaviour of this boundary term on @xmath208 ( or @xmath209 ) .
first we ask that the compatibility condition between the co - tetrad and connection should be satisfied on the boundary .
then , we partially fix the gauge on @xmath127 , by fixing the internal time - like tetrad @xmath210 , such that @xmath211 and we restrict field configurations such that @xmath212 is the unit normal to @xmath208 and @xmath209 . under these conditions
it has been shown in @xcite that on @xmath127 , @xmath213 , where @xmath214 is the trace of the extrinsic curvature of @xmath127 .
note that this is the gibbons - hawking surface term that is needed in the einstein - hilbert action , with the constant boundary term equals to zero . on the other hand , at spatial infinity , @xmath122
, we fix the co - tetrads and only permit gauge transformations that reduce to identity at infinity . under these conditions the boundary term is gauge invariant at @xmath208 , @xmath209 and @xmath122 .
we shall show later that it is also invariant under the residual local lorentz transformations at a weakly isolated horizon , when such a boundary exists .
it turns out that at spatial infinity this boundary term does not reduce to gibbons - hawking surface term , the later one is divergent for asymptotically flat spacetimes , as shown in @xcite .
let us mention that there have been other proposals for boundary terms for palatini action , as for example in @xcite and @xcite , that are equivalent to gibbons - hawking action and are obtained without imposing the time gauge condition .
they are manifestly gauge invariant and well defined for finite boundaries , but they are not well defined for asymptotically flat spacetimes . in time gauge they reduce to ( [ palatiniplusboundary ] ) .
the variation of ( [ palatiniplusboundary ] ) is , @xmath215\label{variationpalatini}\ , , \end{aligned}\ ] ] where @xmath216 we shall show later that the contribution of boundary term @xmath217 vanishes at @xmath122 and @xmath218 , so that from ( [ variationpalatini ] ) we obtain the following equations of motion @xmath219 where @xmath220 is a torsion two - form . from ( [ eompalatinide ] )
it follows that @xmath221 , and this is the condition of the compatibility of @xmath222 and @xmath223 , that implies _aij = e^b_[i_a e_bj]+^c_abe_c[ie^b_j ] , where @xmath224 are the christoffel symbols of the metric @xmath225 .
now , the equations ( [ eompalatinife ] ) are equivalent to einstein s equations @xmath226 . from equations
( [ sympl_poten_boundary_term ] ) and ( [ variationpalatini ] ) , the symplectic potential for @xmath227 is given by @xmath228 therefore from ( [ defj ] ) and ( [ symppotentialpalatini ] ) the corresponding symplectic current is , @xmath229 note that the symplectic current is insensitive to the boundary term , as we discussed in sec .
[ sec:3 ] .
as we shall discuss in the following sections , the palatini action , in the asymptotically flat case , is not well defined , but it can be made differentiable and finite after the addition of the corresponding boundary term already discussed @xcite .
furthermore , we shall also show that in the case when the spacetime has as internal boundary an isolated horizon , the contribution at the horizon to the variation of the palatini action , either with a boundary term @xcite or without it @xcite , vanishes . the first additional term to the gravitational action that we shall consider is the so called _ holst _
term @xcite , first introduced with the aim of having a variational principle whose @xmath9 decomposition yielded general relativity in the ashtekar - barbero ( real ) variables @xcite .
it turns out that the holst term , when added to the palatini action , does not change the equations of motion ( although it is not a topological term ) , so that in the hamiltonian formalism its addition corresponds to a canonical transformation .
this transformation leads to the ashtekar - barbero variables that are the basic ingredients in the loop quantum gravity approach . as we shall show in the next chapter ,
the holst term is finite but not differentiable for asymptotically flat spacetimes , so an appropriate boundary term should be added in order to make it well defined .
the result is @xcite , @xmath230 where @xmath231 is the barbero - immirzi parameter .
the variation of the holst term , with its boundary term , is given by @xmath232 and it leads to the following equations of motion in the bulk : @xmath233 and @xmath234 .
the second one is just the bianchi identity , and we see that the holst term does not modify the equations of motion of the palatini action .
the contribution of the boundary term ( that appears in the variation ) should vanish at @xmath122 and @xmath218 , in order to have a well posed variational principle . in the following section we shall see that this is indeed the case .
on the other hand we should also examine the gauge invariance of the boundary term in ( [ holsttermboundary ] ) .
under the same assumptions as in the case of the palatini boundary term we obtain that @xcite _
m_1 ^ij _ij=2_m_1^abce_b^j_ce_aj^3x=_m_1e^ie_i , and this term is not gauge invariant at @xmath208 or @xmath209 . as we shall see in the following section , at the asymptotic region it is gauge invariant , and also at @xmath218 . in the analysis of differentiability of the action and the construction of the symplectic structure and conserved quantities there is no contribution from the spacial surfaces @xmath208 and @xmath209 , and we can argue that the non - invariance of the boundary term in ( [ holsttermboundary ] ) is not important , but it would be desirable to have a boundary term that is compatible with all the symmetries of the theory . as we shall see later , the combination of the holst and neih - yan terms _ is _ differentiable and gauge invariant . it is easy to see that the symplectic potential for @xmath235 is given by @xcite _ ( ) = 12 _
^ij_ij= 1 _ e^ie_i , where in the second line we used the equation of motion @xmath236 . the symplectic current is given by @xmath237 as we have seen in the subsection [ sec:3.1 ] , this term does _ not _ contribute to the symplectic structure , since it is an exterior derivative .
nevertheless , the holst term modifies the noether charge associated to diffeomorphisms , as we shall see in sec .
[ sec:6 ] . in four dimensions
there are three topological invariants constructed from @xmath238 , @xmath239 and @xmath240 , consistent with diffeomorphism and local lorentz invariance .
they are exact forms and do not contribute to the equations of motion , but in order to be well defined they should be finite , and their variation on the boundary of the spacetime region @xmath160 should vanish .
the first two terms , the pontryagin and euler terms are constructed from the curvature @xmath239 and its dual ( in the internal space ) @xmath241 , while the third one , the neih - yan invariant , is related to torsion @xmath242 .
these topological invariants can be thought of as 4-dimensional density lagrangians defined on a manifold @xmath14 , that additionally are exact forms , but they can also be seen as terms living on @xmath243 . in that case
it is obvious that they do not contribute to the equations of motion in the bulk .
but a natural question may arise . if we take the lagrangian density in the bulk and take the variation , what are the corresponding equations of motion in the bulk ?
one can check that , for pontryagin and euler , the resulting equations of motion are trivial in the sense that one only gets the bianchi identities , while for the nieh - yan term they vanish identically .
let us now see how each of this terms contribute to the variation of the action .
the action corresponding to the pontryagin term is given by , [ pontryagin ] s _ = _ f^ij f_ij = 2 _ ( _ ij d^ij + _ ij^ik _ k^j ) . the boundary term is the chern - simons lagrangian density , @xmath244 .
the variation of @xmath245 , calculated from the lhs expression in ( [ pontryagin ] ) , is s_=-2 _
df^ij_ij+ 2_f^ij_ij , [ var_po1 ] so it does not contribute to the equations of motion in the bulk , due to the bianchi identity @xmath246 , and additionally the surface integral in ( [ var_po1 ] ) should vanish for the variational principle to be well defined .
we will show later that this is indeed the case for boundary conditions of interest to us , namely asymptotically flat spacetimes possibly with an isolated horizon .
on the other hand , if we calculate the variation of the pontryagin term directly from the rhs of ( [ pontryagin ] ) , we obtain @xmath247 the two expressions for @xmath248 are , of course , identical since @xmath249 .
the first one ( [ var_po1 ] ) is more convenient for the analysis of the differentiability of the pontryagin term , but the second one ( [ var_po2 ] ) is more suitable for the definition of the symplectic current , which vanishes identically for any boundary contribution to the action , as we have seen in ( [ defj ] ) .
let us now consider the action for the euler term , which is given by , [ euler ] s _ = _ f^ij f_ij= 2 _ ( _ ij d^ij + _ ij^ik _ k^j ) , with variation , calculated from the expression in the bulk , given by s_= -2 _ df^ij_ij+ 2
_ f^ij_ij .[var_eu ] again , the action will only be well defined if the boundary contribution to the variation ( [ var_eu ] ) vanishes . in the following section
we shall see that it indeed vanishes for our boundary conditions .
let us denote by @xmath250 the boundary term on the rhs of ( [ euler ] ) , then we can calculate the variation of @xmath251 from this term directly as s_=2 _ l_cse . finally , as before , the corresponding contribution to the symplectic current vanishes .
the nieh - yan topological invariant is of a different nature from the two previous terms .
it is related to torsion and its contribution to the action is @xcite , [ nieh - yan ] s _ = _ ( d e^i d e_i - ^ij f_ij ) = _
note that the nieh - yan term can be written as @xmath252 where @xmath199 is the holst term ( [ holsttermboundary ] ) without boundary term .
the variation of the term @xmath253 is given by s_=_2de_ie^i - e^ie^j_ij .
contrary to what happens to the euler and pontryagin terms , the nieh - yan term has a different asymptotic behavior . in the next chapter
we will show that the nieh - yan term is finite , but not differentiable , for asymptotically flat spacetimes .
thus , even when it is by itself a boundary term , it has to be supplemented with an appropriate boundary term to make the variational principle well defined .
we shall see that this boundary term coincides precisely with the boundary term in ( [ holsttermboundary ] ) ( up to a multiplicative constant ) , and the resulting well defined neih - yan action is given by @xmath254 it is straightforward to see that the symplectic potential and symplectic current for this action are the same as for ( [ holsttermboundary ] ) ( up to a factor @xmath255 ) .
this relation between @xmath199 and @xmath253 points to another proposal for a boundary term for the holst action , different from that in ( [ holsttermboundary ] ) , that has the advantage of being manifestly gauge invariant .
namely , for asymptotically flat spacetimes , we can see that the surface term in the variation of neih - yan term cancels the surface term in the variation of the holst term , and the action @xmath256 , given by [ hny ] s_= - _ ^ijf_ij- _
de^i e_i=- _ de^ide_i , well defined .
this combination was proposed in @xcite , but since they were interested in finite boundaries the boundary conditions that they considered are different from the ones we use , namely they impose @xmath257 ( @xmath258 is the metric induced on the boundary ) and leave @xmath259 arbitrary , which is not compatible with our condition @xmath236 on the boundary ( and is not compatible with isolated horizons boundary conditions either ) . to end this part , let us also comment that in the presence of fermions one has to generalize the holst action to spacetimes with torsion , that naturaly leads to neih - yan topological term , instead of the holst term , as shown in @xcite .
but , as we saw the neih - yan term is not well defined for our boundary conditions and should be modified as in ( [ well_defined_ny ] ) . so far in this section we have introduced all the ingredients for the `` most general '' first order diffeomorphism invariant action that classically describes general relativity . as we have already mentioned and we shall prove in the following section , both pontryagin and euler terms , @xmath260 and @xmath261 respectively , are well defined in the case of asymptotically flat spacetimes with a weakly isolated horizon .
this means that we can add them to the palatini action with its boundary term , @xmath262 , and the resulting action will be again well defined . as we foresaw in the previous section
, the addition of the nieh - yan term , @xmath263 , could lead to different possibilities for the construction of a well defined action .
therefore the complete action can be written as , @xmath264 = s_{\rm pb } + s_{\rm h } + \alpha_{1 } s_{\rm po } + \alpha_{2}s_{\rm e } + \alpha_{3 } s_{\rm ny } + \alpha_{4}s_{\rm bh}\ , .\ ] ] here @xmath265 are coupling constants .
the coupling constants @xmath266 and @xmath267 , are not fixed by our boundary conditions , while different choices for the holst - nieh - yan sector of the theory , discussed in the previous part , imply particular combinations of @xmath268 and @xmath269 . to see that , consider @xmath270 that represents the boundary term that we need to add to holst term in order to make it well defined . as we have seen in the previous analysis , if @xmath271 then the combination of the holst and nieh - yan terms is well defined and no additional boundary term is needed , so @xmath272 in that case . for every other value of @xmath273
we need to add a boundary term , and in that case @xmath274 .
other than these cases , there is no important relation between the different coupling constants .
this has to be contrasted with other asymptotic conditions studied in the literature ( that we shall , however , not consider here ) .
it turns out that the palatini action with the negative cosmological constant term is not well defined for asymptotically anti - de sitter ( aads ) spacetimes , but it can be made differentiable after the addition of an appropriate boundary term . in @xcite
it is shown that it can be the same boundary term as in the asymptotically flat case , with an appropriately modified coupling constant .
on the other hand , as shown in @xcite and @xcite , one can choose the euler topological term as a boundary term and that choice fixes the value of @xmath275 .
in that case @xmath276 , and the asymptotically flat case can not be obtained in the limit @xmath277 . the differentiability of nieh - yan term has been analyzed in @xcite .
the result is that this term is well defined , for aads space - times , only after the addition of the pontryagin term , with an appropriate coupling constant .
let us also comment that the details of the asymptotic behaviour in @xcite are different than in the other mentioned papers .
we have considered the most general action for general relativity in the first order formalism , including boundaries , in order to have a well defined action principle and covariant hamiltonian formalism .
we have left , until now , the boundary conditions unspecified , other that assuming that there is an outer and a possible inner boundary to the region @xmath278 under consideration . in this section
we shall consider specific boundary conditions that are physically motivated .
for the outer boundary we will specify asymptotically flat boundary conditions that capture the notion of isolated systems . for the inner boundary we will consider isolated horizons boundary conditions . in this way , we allow for the possibility of spacetimes that contain a black hole .
this section has two parts . in the first one , we consider the outer boundary conditions and in the second part , the inner horizon boundary condition . in each case
, we study the finiteness of the action , its variation and its differentiability .
since this manuscript is to be self - contained , we include a detailed discussion of the boundary conditions before analysing the different contributions to the action .
we are interested in spacetimes that at infinity look like a flat spacetime , in other words , whose metric approaches a minkowski metric at infinity ( in some appropriately chosen coordinates ) . here
we will follow the standard definition of asymptotically flat spacetimes in the first order formalism ( see e.g. @xcite , @xcite and for a nice and pedagogical introduction in the metric formulation @xcite and @xcite ) . here
we give a brief introduction into asymptotically flat spacetimes , following closely @xcite . in order to describe the behaviour of the metric at spatial infinity
, we will focus on the region @xmath279 , that is the region outside the light cone of some point @xmath280 .
we define a @xmath77dimensional radial coordinate @xmath281 given by @xmath282 , where @xmath283 are the cartesian coordinates of the minkowski metric @xmath284 on @xmath285 with origin at @xmath280
. we will foliate the asymptotic region by timelike hyperboloids , @xmath286 , given by @xmath287 , that lie in @xmath279 .
spatial infinity @xmath122 corresponds to a limiting hyperboloid when @xmath288 .
the standard angular coordinates on a hyperboloid are denoted by @xmath289 , and the relation between cartesian and hyperbolic coordinates is given by : @xmath290 .
we shall consider functions @xmath168 that _ admit an asymptotic expansion to order @xmath291 _ of the form , @xmath292 where the remainder @xmath293 has the property that @xmath294 a tensor field @xmath295 will be said to admit an asymptotic expansion to order @xmath291 if all its component in the _ cartesian _ chart @xmath283 do so .
its derivatives @xmath296 admit an expansion of order @xmath297 . and @xmath298 respectively.,width=302 ] with these ingredients at hand we can now define an asymptotically flat spacetime in terms of its metric : a smooth spacetime metric @xmath299 on @xmath279 is _ weakly asymptotically flat at spatial infinity _ if there exist a minkowski metric @xmath284 such that outside a spatially compact world tube @xmath300 admits an asymptotic expansion to order 1 and @xmath301 .
in such a space - time the metric in the region @xmath279 takes the form , @xmath302 where @xmath303 , @xmath304 and @xmath305 only depend on the angles @xmath306 and @xmath307 is the metric on the unit time - like hyperboloid in minkowski spacetime : @xmath308 note that also we could have expanded the metric in a chart @xmath309 , associated with a timelike cylinder , or any other chart .
but we chose the chart @xmath310 because it is well adapted to the geometry of the problem and will lead to several simplifications . in the case of a @xmath311decomposition a cylindrical chart could be a better choice .
for this kind of space - times , one can always find another minkowski metric such that its off - diagonal terms @xmath304 vanish in leading order .
in @xcite it is shown with details that the asymptotically flat metric can be written as @xmath312 with @xmath313 .
we also see that @xmath314 .
these two conditions restrict the asymptotic behaviour of the metric , but are necessary in order to reduce the asymptotic symmetries to a poincar group , as demonstrated in @xcite . from the previous discussion and the form of the metric one
can obtain the fall - off conditions for tetrads .
as shown in @xcite in order to have a well defined lorentz angular momentum one needs to admit an expansion of order 2 , therefore we assume that in cartesian coordinates we have the following behaviour @xmath315 where @xmath316 is a fixed co - frame such that @xmath317 is flat and @xmath318
. the sub - leading term @xmath319 can be obtained from ( [ ds - af ] ) and is given by @xcite , @xmath320 where @xmath321 the asymptotic expansion for connection can be obtained from the requirement that the connection be compatible with tetrad on @xmath122 , to appropriate leading order .
this leads to the asymptotic expansion of order 3 for the connection , @xmath322 we require that @xmath242 vanishes , to an appropriate order , more precisely , we ask that the term of order 0 in @xmath242 vanishes @xmath323 and since @xmath324 it follows that @xmath325 .
the term of order 1 should also vanish leading to @xmath326 .
we also ask that the term of order 2 in @xmath242 vanishes , and we obtain @xmath327 and we shall demand compatibility between @xmath0 and @xmath207 only based on this condition . as a result , we obtain @xmath328 } _ { a } ) \\ & = & 2 \rho\ , ( 2 \rho^{[i } \rho_{a } \partial^{j ] } \sigma - \,^{o}e^{[i } _ { a } \partial^{j ] } \sigma - \rho^{-1 } \,^{o}e^{[i } _ { a } \rho^{j ] } \sigma ) \ , .\end{aligned}\ ] ] note that although @xmath281 appears explicitly in the previous expression , it is independent of @xmath281 . therefore , in the asymptotic region we have @xmath329 .
this condition has its repercussions on the behaviour of the holst and neih - yan terms , as we will show in what follows .
now we have all necessary elements in order to prove the finiteness of the palatini action with boundary term , given by ( [ palatiniplusboundary ] ) .
this expression can be re - written as , @xmath330 or in components @xmath331 where @xmath332 is the metric compatible 4-form on @xmath14 .
this volume element is related by @xmath333 , to the levi - civita tensor density of weight + 1 , @xmath334 .
we will prove that taking into account the boundary conditions ( [ affalloff - tetrad ] ) and ( [ affalloff - connection ] ) , the integrand falls off as @xmath335 , while the volume element on any cauchy slice in asymptotic region goes as @xmath336 , so the action is manifestly finite always ( even off - shell ) , if the two cauchy surfaces are asymptotically time - translated with respect to each other . from ( [ metric - hyperboloid ] ) and
( [ ds - af ] ) , we see that the volume element in the asymptotic region takes the form @xmath337 . in order to prove finiteness we will consider the region bounded by cauchy slices @xmath338 instead of @xmath339 , since in the second case for @xmath340 the volumen of the region does not need to converge ( see fig .
[ slices ] ) .
since @xmath341 at the surface with constant @xmath298 we have @xmath342 . substituting this into the metric
we can see that the volume element is @xmath343 .
as @xmath344 , the angle @xmath345 so @xmath346 .
it follows that in the limit @xmath344 the volume of the region @xmath14 behaves as @xmath347 .
now , we need to deduce the asymptotic behavior of @xmath348 . since @xmath349 it follows that @xmath350+o(\rho^{-2})\ , .\ ] ] the partial derivative , with respect to cartesian coordinates , of any function @xmath351 is proportional to @xmath352 , @xmath353 where the explicit expression for @xmath354 can be obtained from the relation between cartesian and hyperbolic coordinates . as a consequence @xmath355 , and since @xmath356 it follows that @xmath357 falls off as @xmath335 , and the palatini action with boundary term is finite .
now let us prove the differentiability of the action ( [ palatiniplusboundary ] ) . as we have commented after ( [ variationpalatini ] ) ,
this action is differentiable if the boundary term that appears in the variation vanishes .
this boundary term is @xmath358 where we decomposed the boundary as @xmath359 , as in fig.[regionm ] . on the cauchy slices , @xmath119 and @xmath120 , we assume @xmath360 so the integrals vanish , and in the following section we will prove that over @xmath121 this integral also vanishes . here
we will focus on the contribution of the asymptotic region @xmath122 . on a time - like hyperboloid @xmath286 , @xmath287 ,
so that its volume element is @xmath361 and the boundary term can be written as , @xmath362 now we can use that , @xmath363 since @xmath281 is constant on a hyperboloid , it follows that @xmath364 is orthogonal to it and @xmath365 .
then , we obtain @xmath366 in this expression the term with a derivative of @xmath303 is proportional to @xmath367 , so that the variation ( [ holst - i ] ) reduces to @xmath368 where @xmath369 is the unit hyperboloid .
so we see that the palatini action with the boundary term is differentiable when we restrict to configurations that satisfy asymptotically flat boundary conditions , such that @xmath370 has the same ( arbitrary ) value for all of them . in that case , the above expression ( [ final_variation ] ) vanishes .
this last condition is not an additional restriction to the permissible configurations , because every one of them ( compatible with our boundary conditions ) corresponds to some fixed value of @xmath371 . here
we want to emphasize the importance of the boundary term added to the action given that , without it , the action fails to be differentiable .
the contribution from the asymptotic region to the variation of the _ palatini action _ is , @xmath372 our boundary conditions imply that @xmath373 , so that the integral behaves as @xmath374 , and in the limit @xmath340 is explicitly divergent .
as we have seen earlier , in the asymptotic region we have @xmath329 . furthermore , as @xmath375
, we have that @xmath376 .
we can see that explicitly by calculating the term of order 3 in this expression @xmath377 the first term in the previous expression vanishes since @xmath378 , due to ( [ compatibility ] ) .
so , we see that the holst term @xmath379 is finite under these asymptotic conditions , since @xmath380 goes as @xmath335 , while the volume element on every cauchy surface goes as @xmath381 .
the variation of the holst term is well defined if the boundary term , obtained as a result of variation , vanishes .
we will analyze the contribution of this term @xmath382 let us examine the term of order 2 of the integrand , it is @xmath383\ , , \ ] ] due to ( [ compatibility ] ) and @xmath384 , and this term does not contribute to ( [ variation_boundary_holst ] ) .
so , the leading term in @xmath385 is of order 3 , and it does not vanish , since it depends also on @xmath386 , which is not fixed by our boundary conditions .
since the volume element on a hyperboloid @xmath286 goes as @xmath387 , it follows that the boundary term ( [ variation_boundary_holst ] ) does not vanish at @xmath122 ( though it is finite ) .
as our analysis shows we should provide a boundary term for the holst term , in order to make it differentiable .
it turns out that this term should be @xcite @xmath388 let us show first that this term is finite and for that we should prove that its term of order 2 vanishes .
this term is @xmath389 due to the same arguments as in ( [ order2 ] ) , and we see that it does not contribute to the boundary term ( [ boundary_term_holst ] ) .
so , the leading term of the integrand is of order 3 , and since the volume element at @xmath286 goes as @xmath387 , it follows that ( [ boundary_term_holst ] ) is finite . the holst term with its boundary term ( [ holsttermboundary ] ) can be written as @xmath390 and also as an integral over @xmath14 @xmath391 as we have seen in ( [ variationholst ] ) , the variation of the holst term with its boundary term is well defined provided that the following boundary contribution @xmath392 vanishes .
we first note that @xmath393 and @xmath394 have the expansion of the same order , the leading term is @xmath395 . using ( [ deltasigma - rho ] ) , the fact that @xmath396 is orthogonal to @xmath122 and @xmath397 , one can see that the leading term in the integrand vanishes in the asymptotic region , so that @xmath398 and the integral over @xmath122 vanishes . in the next section
we will prove that the integral over @xmath121 vanishes , so that holst action with boundary term is well defined .
since we are interested in a generalization of the first order action of general relativity , that includes topological terms , we need to study their asymptotic behaviour .
we begin with pontryagin and euler terms , that turn out to be well defined .
it is straightforward to see that the pontryagin term ( [ pontryagin ] ) is finite for asymptotically flat boundary conditions . since @xmath399 = \int_{\mathcal{m } } f_{ab}^{ij
} \wedge f_{cdij}\epsilon^{abcd}\ , , \ ] ] the finiteness of this expression depends on the asymptotic behavior of @xmath239 .
taking into account ( [ affalloff - connection ] ) , we can see that the leading term of @xmath400 falls off as @xmath401 . since the volume of any cauchy slice is @xmath402 , in the limit when @xmath340 the integral goes to zero . as a result ,
the pontryagin term is finite even off - shell .
the same result holds for the euler term ( [ euler ] ) , since the leading term in the asymptotic form of @xmath241 is of the same order as of @xmath239 .
now we want to prove that both terms are differentiable .
as we have showed in ( [ var_po1 ] ) , the variation of the pontryagin term is , @xmath403 in the following subsection we prove that on @xmath121 the integral vanishes . for @xmath122
, we need to prove that the integral @xmath404 vanishes . taking into account ( [ affalloff - connection ] )
we can see that the leading term of @xmath405 goes as @xmath406 .
therefore the integral falls off as @xmath407 which in the limit @xmath340 goes to zero .
the same holds for the euler term .
the neih - yan topological term is given by @xmath408 and it is finite since the integrand is of order 3 , and the volume element on @xmath286 is @xmath387 , so the contribution at @xmath122 is finite .
the variation of @xmath253 is @xmath409 and we see that the first term vanishes , but the second one is exactly as in ( [ variation_boundary_holst ] ) and we have seen that it does not vanish , so we need to add a boundary term to the nieh - yan action in order to make it differentiable . as we have seen in ( [ niehyan - section4 ] ) the difference between the nieh - yan and holst terms
is given by an expression quadratic in torsion @xmath410 , @xmath411 and we will analyze this term here . as we have seen before , our asymptotic boundary conditions imply that @xmath412 and since the volume element on every cauchy surface goes as @xmath347 , it follows that @xmath410 is finite off - shell .
the variation of @xmath410 is given by @xmath413 and it is easy to see that it is well defined .
namely , the surface term vanishes since we demand @xmath236 on an isolated horizon @xmath121 , while at the spatial infinity the integrand behaves as @xmath414 and the volume element goes as @xmath387 , and in the limit @xmath288 the contribution of this term vanishes . from ( [ relationnyholst ] )
we see that the neih - yan term is not well defined due to the behaviour of the holst term , so one need to provide a corresponding boundary term in order to make it well defined .
the resulting action is given in ( [ well_defined_ny ] ) , or equivalently @xmath415 we shall consider the contribution to the variation of the action at the internal boundary , in this case a weakly isolated horizon .
a weakly isolated horizon is a non - expanding null 3-dimensional hypersurface , with an additional condition that implies that surface gravity is constant on a horizon .
let us specify with some details its definition and basic properties @xcite .
let @xmath121 be a 3-dimensional null surface of @xmath416 , equipped with future directed null normal @xmath417 .
let @xmath418 be the ( degenerate ) induced metric on @xmath218 ( we denote by @xmath419 an equality which holds only on @xmath218 and the arrow under a covariant index denotes the pullback of a corresponding form to @xmath218 ) .
a tensor @xmath420 that satisfies @xmath421 , is called an inverse of @xmath422 .
the expansion of a null normal @xmath417 is defined by @xmath423 , where @xmath424 is a covariant derivative compatible with the metric @xmath3 .
the null hypersurface @xmath218 is called a _ non - expanding horizon _
if it satisfies the following conditions : ( i ) @xmath218 is topologically @xmath425 , ( ii ) @xmath426 for any null normal @xmath417 and ( iii ) all equations of motion hold at @xmath218 and @xmath427 is future directed and causal for any @xmath417 , where @xmath428 is matter stress - energy tensor at @xmath218 .
the second condition implies that the area of the horizon is constant in time , so that the horizon is isolated .
we need one additional condition in order to satisfy the zeroth law of black hole dynamics . in order to introduce it
let us first specify some details of the geometry of the isolated horizon .
it is convenient to use null - tetrads @xmath429 , where a real , future directed null vector field @xmath430 is transverse to @xmath218 and a complex vector field @xmath291 is tangential to @xmath218 , such that @xmath431 , @xmath432 and all the other scalar products vanish .
since @xmath417 is a null normal to @xmath218 it is geodesic and its twist vanishes .
we define surface gravity @xmath433 as the acceleration of @xmath434 l^a_a l^b _ ( l)l^b . we note that @xmath433 is associated to a specific null normal @xmath417 , if we replace @xmath417 by @xmath435 the acceleration changes @xmath436 . the raychaudhuri and einstein s equations together with the condition on the stress - energy tensor imply that every @xmath417 is also shear free and since its expansion and twist vanish there exists a one - form @xmath437 such that @xcite _
l^b _a l^b .[5.1 ] under the rescaling of the null normal @xmath438 , @xmath207 transforms like a connection @xmath439 ( we see that @xmath207 is invariant under constant rescaling ) .
it is also easy to see that the horizon is ` time ' invariant , in the sense that @xmath440 .
furthermore , the area two - form on the cross - sections of @xmath218 , @xmath441 is also preserved in ` time ' , @xmath442 .
since @xmath417 can be rescaled by an arbitrary positive function , in general @xmath433 is not constant on @xmath218 .
if we want to establish the zeroth law of black hole dynamics @xmath443 we need one additional condition , the ` time ' invariance of @xmath207 , @xmath444 now , if we restrict to constant rescaling of @xmath417 , @xmath445 that leaves @xmath207 invariant , then the zeroth law of black hole dynamics follows , for every null normal @xmath417 related to each other by constant rescaling . all null normals related to each other by a constant rescaling form an equivalence class @xmath446 $ ] .
now , we can define a _ weakly isolated horizon _ ( wih ) @xmath447)$ ] as a non - expanding horizon equipped with an equivalence class @xmath446 $ ] , such that @xmath448 , for all @xmath449 $ ] . in order to analyze the contribution to the variation of the action over the internal boundary , which is a wih @xmath218
, we equip @xmath218 with a fixed class of null normals @xmath446 $ ] and fix an internal null tetrads @xmath450 on @xmath218 , such that their derivative with respect to flat derivative operator @xmath451 vanishes.the permissible histories at @xmath218 should satisfy two conditions : ( i ) the vector field @xmath452 should belong to the fixed equivalence class @xmath446 $ ] ( this is a condition on tetrads ) and ( ii ) the tedrads and connection should be such that @xmath447)$ ] constitute a wih . the expression for tetrads on @xmath218 is given by @xcite e_a^i -l^i n_a + ^i m_a + m^i_a , so that ^ij 2l^[in^j]+2i n(m l^[i^j]- l^[im^j ] ) .[hor1 ] the expression for the connection on @xmath218 is given by @xcite _ij -2 l_[in_j ] + 2u l_[i_j ] + 2|u l_[im_j ] + 2v m_[i_j ] , [ hor2 ] where we have introduced two new one - forms , a complex one @xmath453 and purely imaginary one @xmath454 . in @xcite the expression for these one forms is given in terms of newman - penrose ( np ) spin coefficients and null tetrads .
first we have _a = -(+|)n_a + ( |+)_a + ( + |)m_a , where @xmath455 , @xmath456 and @xmath457 are np spin coefficients .
in what follows we do not need their explicit form , we will just write down the expression for @xmath457 since this coefficient will be of special importance , @xmath458 . since @xmath459 it follows that @xmath460 .
also , it can be shown that @xcite g , where @xmath461=$ ] , with @xmath462 , and @xmath463 are the components of the weyl tensor .
now , it is easy to see that the condition @xmath448 leads to @xmath464 . on the other hand we have u_a -|n_a + |m_a + |_a , and v_a -(-|)n_a + ( -|)_a + ( -|)m_a , where @xmath465 , @xmath466 and @xmath467 are additional np spin coefficients .
it has been also shown that v f , where @xmath468 is a function of the riemann curvature and weyl tensor .
then , we can calculate @xmath469 , _
l v = lv+(lv)=2(im ) .[lv ] we shall also need the expression for the pull - back of the curvature two - form on the weakly isolated non - rotating horizon ( the details are given in @xcite ) f^ij -2i m^[i|m^j]+2f^kln_l l^[i ( m^j]m_k + |m^j]|m_k ) , [ curvature_horizon ] where @xmath279 is the scalar curvature of the cross - section of @xmath121 .
now we have all necessary elements in order to calculate the contribution of the variation of the palatini action , the holst term and topological terms at isolated horizon . before that , let us first examine the gauge invariance of the boundary terms of palatini action and holst term , given in ( [ palatiniplusboundary ] ) and ( [ holsttermboundary ] ) , on a weakly isolated horizon @xmath218 . the residual lorentz transformations , compatible with the definition of @xmath218 can be divided in two groups : ones that preserve the direction of a vector @xmath417 and rotate @xmath291 lc l , nn , me^im , [ lorentz1 ] and ones that leave @xmath417 invariant , but change @xmath430 and @xmath291 ll , nn - u m-|u+u|ul , mm-|u l .[lorentz2 ] note that @xmath470 , since @xmath449 $ ] .
we shall also partially fixed @xmath45 by the condition @xmath471 .
we will see below that this condition implies that @xmath472 and then from ( [ lv ] ) it follows that @xmath473 .
there are no restrictions on @xmath474 .
the palatini boundary term on the horizon reduces to _ _
it was shown in @xcite that @xmath207 is invariant under both classes of transformations .
two - form @xmath475 is invariant under ( [ lorentz1 ] ) , and also under ( [ lorentz2 ] ) since this transformation implies @xmath476 , due to @xmath477 .
similarly , the holst boundary term on the horizon is _
_ ij^ij=_v .
it turns out that @xmath454 is invariant under ( [ lorentz2 ] ) , and under ( [ lorentz1 ] ) it transforms as @xmath478 . due to the restriction on @xmath45
, we have that @xmath479 , where @xmath480 is arbitrary , so that @xmath481 . as a consequence @xmath482 is also invariant under gauge transformations on @xmath218 .
in this part we will analyze the variation of the palatini action with boundary term ( [ palatiniplusboundary ] ) , on an isolated horizon @xmath218 s__= _ _ ijkl^ije^ke^l = _ .[var1 ] since @xmath218 is a non - expanding horizon , @xmath483 .
any other permissible configuration of tetrads , @xmath484 , should also satisfy @xmath485 , where @xmath486 $ ] and @xmath487 . for the null normals in the equivalence class @xmath446 $ ] , @xmath488 , and it follows that @xmath489 . in the variational principle all fields are fixed on initial and final cauchy surfaces , @xmath208 and @xmath209 , in particular @xmath490 on two - spheres at the intersection of the initial and final cauchy surface with the wih , @xmath491 ( see fig .
1 ) . furthermore
, @xmath492 does not change along any null normal @xmath417 , so that @xmath493 on the entire horizon ( comprised between the two cauchy surfaces ) and the integral ( [ var1 ] ) vanishes .
we should remark that , in the following parts , we _ will _ use the same argument whenever we have some field configuration whose lie derivative along @xmath417 vanishes on the horizon , to prove that its variation is zero on the horizon .
we note that the variation of the palatini action , without boundary term , at @xmath218 is s _ _ = - _ _
ijkl^ije^ke^l = - _ .[var1b ] in this case , one can argue that the term on the rhs vanishes , because from @xmath448 it follows that @xmath494 ( a similar , but slightly different , argument was used in @xcite ) .
we see that the variational principle for the palatini action is well defined even without boundary terms on the horizon . nevertheless , for the reasons already mentioned in the previous chapter we shall keep the boundary terms in ( [ palatiniplusboundary ] ) on the whole boundary , including the internal one .
let us now consider the contributions coming from the holst term in the presence of an isolated horizon .
the variation of the holst term , with its boundary term , on an isolated horizon is given by s__= _
^ije_ie_j = _ v , [ var2 ] and for the same reasons that we used before , after the equation ( [ var1 ] ) , since @xmath483 , it follows that @xmath493 and the variation ( [ var2 ] ) vanishes . on the other hand , the variation of the holst term , without a boundary term ,
is also well defined on a horizon , since s__= - _
^ije_ie_j = - _ v .[var2b ] now we can not use the same argument as in the case of palatini action since the lie derivative of @xmath454 does not vanish on @xmath218 , as shown in ( [ lv ] ) , @xmath495 . as we commented earlier
, we have a freedom to perform local lorentz transformations in order to make @xmath457 a real function .
namely , the rotation in the @xmath496 plane , given by @xmath497 , where @xmath45 is an arbitrary function , generates the following transformation of the np spin coefficient @xmath457 @xcite : @xmath498 .
so , @xmath457 can be made real after the appropriate rotation that satisfies the condition @xmath471 .
due to this gauge freedom we can always choose a real @xmath457 , and as a result @xmath499 .
when we change the configuration of fields this condition could be violated , but then again one can perform a gauge transformation to obtain @xmath500 .
then , using the same arguments as before we can conclude that @xmath501 , in the variational principle and ( [ var2b ] ) also vanishes . note
that this argument is simpler and departs significantly from that in @xcite .
let us now consider the possible contributions coming from the topological terms .
that is , we shall see whether the above conditions are sufficient to make the variation of the topological terms well defined at @xmath218 .
the variation of the pontryagin term on the horizon is s__= 2_f^ij_ij = 4i _ v , [ var3 ] where we have used the expresions for the curvature at the horizon ( [ curvature_horizon ] ) and for the connection ( [ hor2 ] ) .
the argument just presented in the previous part implies that @xmath502 , so that the variation @xmath503 vanishes at the horizon .
the variation of the euler term on the horizon is s__= 2 _ f^ij_ij= 4 _ , [ var4 ] and it vanishes since @xmath494 .
note that the variational principle for the neih - yan term is also well defined , since it reduces to the variation of the holst term plus an additional term , given in ( [ var_torsion_term ] ) , that vanishes at @xmath218 since part of the boundary conditions defining an isolated horizon require that the equation of motion @xmath504 , should also hold on @xmath218 .
we can then conclude that the inclusion of the topological terms to the action is compatible with a well defined action principle , without the need of adding new boundary terms at the horizon .
so far we have analyzed the action principle for gravity in the first order formulation with two possible boundary conditions .
we have seen that the action principle is well defined for a suitable choice of boundary terms even in the case when topological terms are incorporated . in this section
we shall consider some of the information that comes from the covariant hamiltonian formulation .
in particular , we shall see how one can define conserved quantities . as we have discussed in secs .
[ sec:3.2 ] and [ sec:3.3 ] there are two classes of quantities , namely those that are generators of hamiltonian symmetries and the so called noether charges .
we shall then analyze the relation between hamiltonian and noether charges for the most general first order gravitational action , focusing on the role that the boundary terms play . as one might anticipate , the fact that the boundary terms do not modify the symplectic structure implies that the hamiltonian charges are insensitive to the existence of extra boundary terms .
however , as we shall see in detail , the noetherian quantities _ do _ depend on the boundary terms .
specifically , we are interested in the relation of the noether charge with the energy at the asymptotic region and the energy of the horizon .
let us start with the hamiltonian charge related to diffeomorphisms , in the first order formalism , for asymptotically flat configurations .
the first step is the construction of the pre - symplectic structure @xmath135 for the full theory , given by ( [ complete - action ] ) . as we have seen in ( [ defj ] ) the boundary terms in the action ( topological terms )
do not contribute to the symplectic current @xmath114 , so that the only contributions in our case come from palatini action and the holst term .
for this reason we shall only consider the palatini and holst terms in this part . from the equation ( [ intjzero ] )
one can obtain a conserved pre - symplectic structure , as an integral of @xmath114 over a spatial surface , if the integral of the symplectic current over the asymptotic region vanishes and if the integral over an isolated horizon behaves appropriately . as we saw in ( [ symplectic_current_holst ] ) ,
the symplectic current for the holst term is an exterior derivative so it does not appear in the ( [ intjzero ] ) , nor in the symplectic structure , so the only contribution comes from the palatini action . as shown in @xcite , for asymptotically flat spacetimes , @xmath505 , where @xmath506 is given by ( [ jpalatini ] ) . on the other hand , in the presence of a weakly isolated horizon
it is shown in @xcite that _ j_p(_1,_2)=__1_2 ( ) -_2_1 ( ) , [ jpalhorizon ] where @xmath507 is a potential defined as _
l=_(l ) , = 0 on s_1 , [ psi ] with @xmath508 , 2-sphere at the intersection of a cauchy surface with a horizon . note that the integral ( [ jpalhorizon ] ) reduces to a surface integral over @xmath509 . finally , as explained in the subsection [ sec:3.1 ] ,
the pre - symplectic structure is of the form given in ( [ sympl_struct ] ) , the sum of two integrals , one of them over a cauchy surface and the other one over a 2-sphere at the boundary of an isolated horizon .
taking into account the expression for the symplectic current ( [ jpalatini ] ) , and the result given above in ( [ jpalhorizon ] ) , the pre - symplectic structure for the palatini action , for asymptotically flat spacetimes with weakly isolated horizon , takes the form @xcite @xmath510 we see that the existence of an isolated horizon modifies the symplectic structure of the theory .
let us now construct the conserved charges for this theory .
we shall consider the hamiltonian @xmath180 that is a conserved quantity corresponding to asymptotic symmetries and symmetries on the horizon of a spacetime .
our asymptotic conditions are chosen in such a way that the asymptotic symmetry group be the poincar group .
the corresponding conserved quantities , energy - momentum and relativistic angular momentum , are constructed in @xcite .
the contribution to the energy from a weakly isolated horizon has been analyzed in @xcite , where the first law of mechanics of non - rotating black holes was deduced .
rotating isolated horizons have been the topic of @xcite , where the contribution from the angular momentum of a horizon has been included . in this paper
we restrict our attention to energy and give a review of the principal results presented in @xcite .
let us consider a case when @xmath169 is the infinitesimal generator of asymptotic time translations of the spacetime .
it induces time evolution on the covariant phase space , generated by a vector field @xmath511 . at infinity @xmath169
should approach a time - translation killing vector field of the asymptotically flat spacetime .
on the other hand , if we have a non - rotating horizon @xmath218 , then @xmath169 , at the horizon , should belong to the equivalence class @xmath446 $ ] . in order that @xmath512 represents a phase space symmetry the condition @xmath513 should be satisfied . as we have seen in sec .
[ sec:3.2 ] , @xmath512 is a hamiltonian vector field iff the one - form x_()=| ( , _ ) , is closed , and the hamiltonian @xmath180 is defined as x_()=h _ . in the presence of the isolated horizon , the symplectic structure ( [ pal_ss ] ) has two contributions , one from the cauchy surface @xmath127 and the other one from the two - sphere @xmath514 , at the intersection of @xmath127 with @xmath218 .
this second term does not appear in @xmath515 , since it is equal to -_s _ _
l-_l , [ int ] and we will show that this integral vanishes . when acting on fields @xmath516 , so that @xmath517 . on the other hand , as pointed out in @xcite , @xmath507 is a potential , a given function of the basic variables .
if we define @xmath518 , as in the case of our basic fields , then the boundary condition in ( [ psi ] ) can not be fulfilled .
so we need to define @xmath519 more carefully .
let @xmath520 denote a potential that corresponds to a null vector @xmath521 in the sense that @xmath522 .
since @xmath523 , it follows that @xmath524 .
we define @xmath525 , then _ l(_l)=0 .
we ask @xmath507 to be fixed at @xmath526 , as a result @xmath527 .
then , it is easy to see that the integral ( [ int ] ) vanishes and the only contribution to @xmath515 comes from the integral over the cauchy surface @xmath127 in ( [ pal_ss ] ) .
as shown in @xcite it turns out that the integrand in @xmath515 is a total derivative , so that h_:= | ( , _ ) = -_m(^ij ) _ ij - ( _ ij ) ^ij , [ firstlaw ] where the integration is over the boundaries of the cauchy surface @xmath127 , the two - spheres @xmath528 and @xmath514 .
the integral at infinity vanishes for every permissible variation @xmath529 , if and only if @xmath169 vanishes asymptotically , so that only diffeomorphisms which preserve the boundary conditions and which are identity at infinity are gauge transformations , i.e. they are in the kernel of @xmath530 .
the asymptotic symmetry group is the quotient of the group of space - time diffeomorphisms which preserve the boundary conditions by its subgroup consisting of asymptotically identity diffeomorphisms . in our case
this is the poincar group and its action generates canonical transformations on the covariant phase space whose generating function is @xmath531 .
the situation is similar at the horizon @xmath218 and infinitesimal diffeomorphisms need not be in the kernel of the symplectic structure unless they vanish on @xmath218 and the horizon symmetry group is the quotient of the lie group of all infinitesimal space - time diffeomorphisms which preserve the horizon structure by its subgroup consisting of elements which are identity on the horizon @xcite .
the surface term at infinity in the expression ( [ firstlaw ] ) defines the gravitational energy at the asymptotic region , whose variation is given by @xmath532 since due to the asymptotic behaviour of the tetrad and connection , the first term in the above expression vanishes . as shown in @xcite , after inserting the asymptotic form of the tetrad ( [ affalloff - tetrad ] ) and connection ( [ affalloff - connection ] ) , this integral represents the variation of the adm energy , @xmath533 , associated with the asymptotic time - translation defined by @xmath169 @xmath534 where @xmath535 is the area element of the unit 2-sphere . on
the other hand , the surface term at the horizon in the expression ( [ firstlaw ] ) represents the horizon energy defined by the time translation @xmath169 , whose variation is given by @xmath536 since the second term in the above expression vanishes at the horizon .
the remaining term is of the form @xmath537 where @xmath538 is the area of the horizon .
now we see that the expression ( [ firstlaw ] ) encodes the first law of mechanics for non - rotating black holes , since it follows that [ first law ] h_=e^_-_()a _ .
we see that the necessary condition for the existence of @xmath180 is that surface gravity , @xmath539 , be a function only of a horizon area @xmath538 . in that case
h_=e^_-e^ _ .
in the following section we want to calculate the noether charge that corresponds to time translation for every term of the action ( [ complete - action ] ) .
we have just seen that @xmath540 is an integral over a cauchy surface of the symplectic current @xmath541 . in section [ sec:3.3 ]
we displayed the relation between the symplectic and noether currents , given in ( [ nc ] ) , and using the definition of noether charge @xmath187 ( [ nch ] ) , we obtain the following relation h_= _
m j ( , _ ) = _ mq_- ( ) .[int2 ] there are two contributions to the above expression , one at @xmath528 and the other one at @xmath514 . as before , @xmath542 ,
is the integral at the rhs of ( [ int2 ] ) calculated over @xmath528 , and @xmath543 the same integral calculated over @xmath514 .
note that the necessary and sufficient condition for the existence of @xmath180 is the existence of the form @xmath544 such that _
m()=_mb .[condb ] let us now consider how the different terms appearing in the action contribute to the noether charges . in this section we shall study the noether charges for the most general first order action we have been considering .
it has two parts . in the first one , we focus our attention on the palatini action , while in the second part we look at the holst and topological terms .
let us see what happens in the case of palatini action with boundary term .
we have seen in the section [ sec:4.1 ] that the symplectic potential current in this case is given by _ ( ) = ^ij_ij . in order to calculate the noether current 3-form ( [ noether ] ) , @xmath545
, we need the following two expressions _ ( _ ) = _ ^ij_ij= [ ( ^ij)+^ij]_ij , and l_=- [ ( ^ijf_ij)- ( ^ij_ij ) ] . from these expressions
we obtain the following result for the noether current 3-form j _ ( _ ) = \{(^ij)f_ij + ( _ij)d^ij + } .
we see that on - shell ( @xmath234 and @xmath546 ) , we have @xmath547 , where the corresponding noether charge is given by q _ = ( ^ij)_ij .
we shall show that the contribution of the second term in ( [ int2 ] ) over @xmath528 vanishes .
namely , _ s _ _ ( ) = _ s_(^ij_ij)=0 , since @xmath548 , @xmath549 and the volume element goes as @xmath347 .
it follows that @xmath550 on @xmath528 and the hamiltonian at infinity exists .
the remaining term at infinity in ( [ int2 ] ) is _
s_(^ij)_ij , [ adm ] and since @xmath551 , due to asymptotic behaviour of the fields , the above expression is equal to @xmath552 given in ( [ energy_infinity ] ) , so in this case e^_adm=_s_q _ , up to an additive constant that we choose to be zero .
note that a similar result is obtained in the second order formalism for the einstein hilbert action with the gibbons - hawking term , as shown in @xcite .
on the other hand , at the horizon the situation is different .
in fact , _
s_q _ = _ s_(^ij)_ij=0 , because @xmath553 on the horizon , and due to the expressions for the tetrad ( [ hor1 ] ) and connection ( [ hor2 ] ) at the horizon .
then , @xmath543 is determined by the remaining term _ s _ _ ( ) = _ s _ [ ( ^ij)_ij+ ^ij ( _ij ) ] .
the first term vanishes since on @xmath218 we have @xmath554 , because @xmath555 .
we are then left with the expression given in ( [ energy_horizon ] ) , and the necessary condition for the existence of @xmath556 is that the surface gravity @xmath539 depends only on the area of the horizon @xcite .
it follows also that there exists a form @xmath544 such that ( [ condb ] ) is satisfied .
we see that in this case e_=_s_q _ , e_=_s _ _ ( ) . in globally stationary spacetimes , @xmath557 , so that @xmath558 , and from the first law ( [ first law ] ) it follows @xmath559 .
for palatini action with boundary term this implies that _ s_q_= _ s _ _ ( ) .
this result depends on the particular form of the action , and it is sensitive to the presence of boundary terms .
let us briefly comment the case of palatini action _ without _ boundary term .
we know that this action is not well defined , its symplectic potential @xmath560 diverges , but we can _ formally _ calculate its noether charge and compare it to the previous example . as we showed in the sections [ sec:3.1 ] and [ sec:3.3 ] , the addition of the total derivative to the action changes its noether charge ( [ noether_charge_change ] ) , but leaves the symplectic structure ( [ defj ] ) unaltered . in the previous example the adm energy was determined completely by the integral of the noether charge over the two - sphere at infinity .
now , the situation is different and both terms in ( [ int2 ] ) contribute to @xmath561 .
we first note that _ ( ) = - ^ij_ij , q _ = - ^ij(_ij ) , where @xmath562 and @xmath563 denote the corresponding quantities for palatini action ( without boundary term ) .
it turns out that _ s__()=-_s _
( ^ij_ij ) , since @xmath564 due to our asymptotic conditions . on the other hand ,
_ = -_s_(_ij)^ij = -_s _ ( _ij)^ij , and the combination of the above expressions , as in ( [ int2 ] ) gives the previous expression for @xmath565 ( [ energy_infinity ] ) . at the horizon
both terms contribute , again .
the results are _ s__()=-_()a _ , where we used the fact that @xmath566 and @xmath567 .
we see again , that in order to satisfy the condition ( [ condb ] ) , @xmath539 should be a function of @xmath538 only .
we also obtain [ variation_noether_charge_horizon_palatini ] _ s _ q_=(_()a _ ) , and the combination of the above expressions , as in ( [ int2 ] ) gives the previous result for @xmath568 ( [ energy_horizon ] ) . finally , let us compare these results for the noether charge with the results of @xcite , and to that end we shall recall one of the principal results in @xcite , referring to the variations of a stationary black hole solution , that states that in diffeomorphism invariant theories , in the second order formalism , the noether charge relative to a bifurcate killing horizon @xmath569 is proportional to the entropy of a black hole @xmath10 .
the result is the following @xmath570 where @xmath571 is the killing field that vanishes on @xmath569 and at infinity tends to a stationary killing vector field with unit norm and @xmath572 is the corresponding surface gravity of a stationary black hole . in the proof of this result
it is assumed that @xmath573 .
furthermore , it has been shown that in the case of stationary variations the integral is independent of the choice of horizon cross - section .
our analysis , based on @xcite and @xcite , is different in various aspects : ( 1 ) we consider the first order formalism , ( 2 ) in our case the existence of the internal boundary is consistently treated , as , for example , in the expression for @xmath574 that involves integration over the whole boundary , not only over the asymptotic region , as in @xcite , ( 3 ) our results are valid also for nonstationary configurations and ( 4 ) in our approach the integration is performed over an arbitrary 2-sphere cross section of a weakly isolated horizon , and not restricted to a preferred bifurcation surface .
taking this into account we wanted to see if in our approach the noether charge can be related to the black hole energy .
we already know that in general this is not the case , since nor the holst term nor the topological terms contribute to the energy of black hole , though they modify the noether charge .
we can formally compare the expression ( [ entropy wald ] ) with our result ( [ variation_noether_charge_horizon_palatini ] ) , taking in account all differences between the two approaches .
we see that , if we impose that @xmath575 , then the result in ( [ variation_noether_charge_horizon_palatini ] ) would look like ( [ entropy wald ] ) . but this restriction is not consistent with the result of @xcite , as we saw in the previous subsection , that the surface gravity is a function of the area of the horizon and that this is a necessary condition in order to have a well defined hamiltonian . as we have seen in this subsection in neither of the cases ,
palatini action with or without surface term , the variation of a corresponding noether charge relative to an isolated horizon is proportional to @xmath576 . to end this section ,
let us calculate the noether charges for the holst term and the topological terms .
we will see that in most of these cases the integrals of the corresponding noether charge 2-form over @xmath528 and @xmath514 vanish .
the only exception is the euler term that has a non - trivial contribution to the noether charge .
let us first consider the holst term with its boundary term @xmath235 , given by ( [ holsttermboundary ] ) .
we know that this term does not contribute to the hamiltonian notion of energy , since it does not modify the symplectic structure of the palatini action . as we have seen in section [ sec:4.2 ] , the symplectic potential current of @xmath235 is given by _
( ) = 12^ij_ij=1e^ie_i , where in the second line we used the equation of motion @xmath577 .
the corresponding noether charge 2-form can be obtained by calculating `` on shell '' the expression @xmath578 and the result is given by q _
= 1(^ij)_ij=1(e^i ) e_i .
now we can show that @xmath579 namely @xmath580 = 0\ , , \ ] ] since @xmath169 is constant on @xmath528 and @xmath581 . on the other hand ,
it is also easy to show that @xmath582 since @xmath583 due to the expressions for the tetrad ( [ hor1 ] ) and connection ( [ hor2 ] ) on the horizon . for the pontryagin term ( [ pontryagin ] ) we obtain _ ( ) = 2 f^ij_ij , q _ = 2 ( _ij)f^ij .
we shall show that the integrals of the noether charge 2-form @xmath584 over @xmath528 and @xmath514 vanish .
for the first one we have _ s _ q _
= 2_s _ ( _ij)f^ij=0 .
since @xmath356 , @xmath585 and the volume element goes as @xmath347 . since , from ( [ curvature_horizon ] ) , the pull - back of the connection on @xmath514 is given by f^ij_s_=-2i m^[i|m^j ] , [ pullback_crossection ] we obtain that the integral of the noether 2-form over @xmath514 is [ nch_po_hor ] _ s _ q _
= 4ic _ s _ ( lv ) = 0 , where we have used the form of the connection on the horizon given by ( [ hor2 ] ) .
now , since @xmath586 , and given that we can always choose a real np coefficient @xmath587 ( as we have seen in the sec .
[ sec:5.2 ] ) , the results follows .
similarly , for the euler term we obtain _ ( ) = 2 f^ij_ij , q _
= 2(_ij)f^ij .
then , just as in the case of the pontryagin term it is easy to see that
= 2_s_(_ij)f^ij=0 , due to the asymptotic behaviour of the fields . at the horizon
the situation is different since the dual of the pull - back ( [ pullback_crossection ] ) is given by ^ij_s_=-2 l^[in^j ] , the corresponding noether charge is non vanishing _ s_q _
= 4c _ s _ ( l ) = 16c_(l)[noether_euler_horizon ] since @xmath588 is constant on the horizon and the remaining integral is a topological invariant .
this result is consistent with the expression for the entropy of the euler term in @xcite , obtained in the second order formalism for stationary black holes . though the noether charge of the euler term over a wih is non - vanishing , the corresponding contribution to the energy is nonetheless , zero .
as we have previously seen in section [ sec:6.1 ] , the variation of the energy at the horizon is h^_= _
s_q_- ( ) , with @xmath589 . for the euler term
we obtain _ s _ cl _ ( ) = 4c_s _ l()=16c_(l ) , since @xmath555 and @xmath590 .
we see that this term cancels the variation of ( [ noether_euler_horizon ] ) in the expression for the energy at the horizon .
finally , we have seen in section [ sec:4.3 ] that the variation of the neih - yan term on shell is proportional to the variation of the holst term , so all the results for the noether charge of the holst term apply directly here .
namely , for the neih - yan term , with its boundary term , given in ( [ well_defined_ny ] ) , we obtain that its noether charge 2-form is q_=2q _ , so that its integrals over @xmath528 and @xmath514 vanish as well .
let us end this section with a remark .
as we have shown , the noether charges at infinity of all the topological terms vanish for asymptotically flat boundary conditions , but this is not the case for locally asymptotically anti - de sitter ( aads ) space - times . in @xcite , aads asymptotic conditions are considered and the noether charge at infinity of the palatini action with negative cosmological constant term turns out to be divergent . in that case the euler term is added in order to make the action well defined and finite . with this modification , the non vanishing ( infinite ) noether charge becomes finite for the well defined action .
let us start by summarizing the main results that we have here presented . 1 .
we have analyzed whether the most general first order action for general relativity in four dimensions has a well posed variational principle in spacetimes with boundaries .
we showed that it is necessary to introduce additional boundary terms in order to have a differentiable action , which is finite for the field configurations that satisfy our boundary conditions : asymptotically flat spacetimes with an isolated horizon as an internal boundary .
we discussed the impact of the topological terms and boundary terms added in order to have a well defined variational principle , to the symplectic structure and the conserved hamiltonian and noether charges of the theory .
we showed that neither the topological terms nor the holst term modify the symplectic structure .
thus , the hamiltonian structure of the theory remains unaffected by the introduction of boundary and topological terms .
it is important to note that this result proves incorrect several assertions that have repeatedly appeared in the literature .
we have shown that even when the hamiltonian conserved charges remain insensitive to the addition of boundary and topological terms , the corresponding noetherian charges _ do _ depend on such choices .
this has as a consequence that the identification of noether charges with , say , energy depends on the details of the boundary terms one has added . for instance , if one only had an internal boundary ( and no asymptotic region ) , several possibilities for the action are consistent , and the relation between energy and noether charge depends on such choices .
we have also seen that one particular topological terms , namely the euler term , contributes non - trivially to the noether charge at the horizon .
our analysis was done using the covariant hamiltonian formalism , that has proved to be economical and powerful to unravel the hamiltonian structure of classical gauge field theories
. it should be interesting to perform a similar analysis using a 3 + 1 decomposition of spacetime and taking special care on the effects of boundaries , extending the analysis of @xcite , where the boundaries were ignored .
a natural question is whether the two hamiltonian approaches to treat this system are equivalent and , if not , understand the underlying reasons for that discrepancy .
this work is in progress and will be reported elsewhere .
we would like to thank n. bodendorfer , s. deser , t. jacobson and r. olea for comments .
this work was in part supported by conacyt 0177840 , dgapa - unam in100212 , and nsf phy 1205388 grants , the eberly research funds of penn state and by cic , umsnh .
j. d. romano , `` geometrodynamics versus connection dynamics ( in the context of ( 2 + 1 ) and ( 3 + 1 ) gravity , '' gen . rel .
grav . * 25 * , 759 ( 1993 ) [ gr - qc/9303032 ] .
p. peldan , `` actions for gravity , with generalizations : a review , '' class . quant .
grav . * 11 * , 1087 ( 1994 ) [ gr - qc/9305011 ] .
a. palatini , deduzione invariativa delle equazioni gravitazionali dal principio di hamilton , " rend
. circ . mat .
palermo * 43 * , 203 ( 1919 ) .
r. l. arnowitt , s. deser and c. w. misner , `` the dynamics of general relativity , '' gen .
grav . * 40 * , 1997 ( 2008 ) [ gr - qc/0405109 ] .
r. utiyama , `` invariant theoretical interpretation of interaction , '' phys . rev . * 101 * , 1597 ( 1956 )
. t. w. b. kibble , `` lorentz invariance and the gravitational field , '' j. math .
* 2 * , 212 ( 1961 )
. d. w. sciama , `` the physical structure of general relativity , '' rev
. mod .
phys . * 36 * , 463 ( 1964 ) [ erratum - ibid . *
36 * , 1103 ( 1964 ) ] .
s. deser and c. j. isham , `` canonical vierbein form of general relativity , '' phys . rev .
d * 14 * , 2505 ( 1976 ) .
s. holst , `` barbero s hamiltonian derived from a generalized hilbert - palatini action , '' phys .
d * 53 * , 5966 ( 1996 ) [ gr - qc/9511026 ] .
n. barros e sa , `` hamiltonian analysis of general relativity with the immirzi parameter , '' int .
. phys .
d * 10 * , 261 ( 2001 ) [ gr - qc/0006013 ] .
a. ashtekar , j. engle and d. sloan , `` asymptotics and hamiltonians in a first order formalism '' , class .
grav . * 25 * , 095020 ( 2008 ) .
[ arxiv:0802.2527 [ gr - qc ] ] a. ashtekar , t. pawlowski and c. van den broeck , `` mechanics of higher - dimensional black holes in asymptotically anti - de sitter space - times '' , class .
* 24 * , 625 ( 2007 ) .
[ arxiv : 0611049 [ gr - qc ] ] a. corichi and e. wilson - ewing , `` surface terms , asymptotics and thermodynamics of the holst action , '' class .
* 27 * , 205015 ( 2010 ) [ arxiv:1005.3298 [ gr - qc ] ] .
a. ashtekar , l. bombeli and o. reula , `` the covariant phase space of asymptotically flat gravitational fields '' , _ analysis , geometry and mechanics : 200 years after lagrange _ , eds .
m. francaviglia and d. holm , north - holland , amsterdam ( 1991 ) .
j. lee and r. m. wald , `` local symmetries and constraints , '' j. math . phys .
* 31 * , 725 ( 1990 ) .
a. n. kolmogorov and s. v. fomin , `` elements of the theory of functions and functional analysis '' , volumes 1 and 2 , dover , new york ( 1999 ) .
v. iyer and r. m. wald , `` some properties of noether charge and a proposal for dynamical black hole entropy '' , phys . rev . *
d50 * , 846 ( 1994).[arxiv:9403028 [ gr - qc ] ] ; r. m. wald and a. zoupas , `` a general definition of conserved quantities in general relativity and other theories of gravity , '' phys .
d * 61 * , 084027 ( 2000 ) [ gr - qc/9911095 ] . y. n. obukhov , `` the palatini principle for manifold with boundary '' , class .
quantum grav .
* 4 * , 1085 ( 1987 ) .
j. f. barbero g. , `` real ashtekar variables for lorentzian signature space times , '' phys .
d * 51 * , 5507 ( 1995 ) [ gr - qc/9410014 ] .
h. t. nieh and m. l. yan , `` an identity in riemann cartan geometry '' , j. math .
* 23 * , 373 ( 1982 ) .
d. j. rezende and a. perez , `` the theta parameter in loop quantum gravity : effects on quantum geometry and black hole entropy , '' phys . rev .
d * 78 * , 084025 ( 2008 ) [ arxiv:0711.3107 [ gr - qc ] ] .
a. ashtekar , c. beetle and j. lewandowski , `` mechanics of rotating isolated horizons '' phys .
rev . * d64 * , 044016 ( 2001 ) .
[ arxiv:0103026 [ gr - qc ] ] | we consider first order gravity in four dimensions . in particular , we consider formulations where the fundamental variables are a tetrad @xmath0 and a so(3,1 ) connection @xmath1 .
we study the most general action principle compatible with diffeomorphism invariance .
this implies , in particular , considering besides the standard einstein - hilbert - palatini term , other terms that either do not change the equations of motion , or are topological in nature .
having a well defined action principle also implies adding additional boundary terms , whose detailed form may depend on the particular boundary conditions at hand .
we consider spacetimes that include a boundary at infinity , satisfying asymptotically flat boundary conditions and/or an internal boundary satisfying isolated horizons boundary conditions . for our analysis
we employ the covariant hamiltonian formalism where the phase space @xmath2 is given by solutions to the equation of motion . for each of the possible terms contributing to the action we study the well posedness of the action , its finiteness , the contribution to the symplectic structure , and the hamiltonian and noether charges .
we show that for the chosen boundary conditions , standard boundary terms warrant a well posed theory .
furthermore , we show that the boundary and topological terms do not contribute to the symplectic structure , nor the hamiltonian conserved charges .
the noether conserved charges , on the other hand , do depend on such additional terms .
the aim of the paper is to present a comprehensive and self - contained treatment of the subject , so the style is somewhat pedagogical .
furthermore , we point out and clarify some issues that have not been clearly understood in the literature . |
multigraded hilbert schemes parametrize families of ideals in a polynomial ring that share the same hilbert function with respect to some grading by an abelian group @xcite .
we are interested in the following particular case .
let @xmath3 be a @xmath4-matrix of unknowns .
we fix the polynomial ring @xmath5 $ ] over a field @xmath6 with the @xmath7-grading by column degree , i.e. @xmath8 . in this grading ,
the hilbert function of the polynomial ring @xmath5 $ ] equals @xmath9 we study the multigraded hilbert scheme @xmath10 , which parametrizes @xmath7-homogeneous ideals @xmath11 in @xmath5 $ ] such that @xmath5/i$ ] has the hilbert function @xmath12 the key example is the ideal @xmath13 that is generated by the @xmath0-minors of @xmath14 , and whose quotient is indeed @xmath7-graded with hilbert function ( [ eqn : ourhf ] ) .
the hilbert scheme @xmath10 has the following geometric interpretation .
each @xmath7-homogeneous ideal in @xmath5 $ ] specifies a subscheme of the product of projective spaces @xmath15 .
the subscheme specified by the ideal @xmath13 is the diagonal embedding of @xmath16 in @xmath17 .
our hilbert scheme @xmath10 is a natural parameter space for degenerations of this diagonal in @xmath17 .
the results obtained in this paper are as follows . in section 2
we prove that all ideals @xmath11 in @xmath10 are radical and cohen - macaulay .
this result is derived by identifying a distinguished borel - fixed ideal @xmath18 with these properties .
it confirms a conjecture on multilinear grbner bases made by conca in @xcite . in section 3
we show that @xmath13 and one of its initial monomial ideals are smooth points on @xmath10 .
the irreducible component containing these points is an equivariant compactification of the homogeneous space @xmath19 where @xmath20 , and @xmath21 is the diagonal embedding .
for @xmath22 we recover thaddeus construction in @xcite of the space of complete collineations .
the relationship of our compactification of @xmath19 to those constructed by lafforgue in @xcite will be discussed in remark [ remark : lafforgue ] and example [ example : twothree ] .
section 4 is concerned with the case @xmath23 , and we regard its results to be the main contribution of this paper . we show that @xmath24 is irreducible , but singular for @xmath25 , and we determine its combinatorial structure .
each point in @xmath24 corresponds to a certain tree of projective lines . among these
are precisely @xmath26 monomial ideals , one for each tree on @xmath27 unlabeled vertices and @xmath28 labeled directed edges , and these form a graph . in section 5
we study the case @xmath29 .
these are the smallest parameters for which the multigraded hilbert scheme @xmath10 is reducible .
we show that @xmath30 is the reduced union of seven irreducible components , with the main component of dimension @xmath31 parametrizing degenerations of the diagonal in @xmath32 .
we list all monomial ideals on @xmath30 and their incidence relations .
section 6 outlines a connection to convexity in affine buildings and tropical geometry . extending previous results in @xcite ,
we show how grbner degenerations on @xmath10 can be used to compute special fibers of deligne schemes .
our plan is to derive conca s conjecture 4.2 in @xcite from the following result .
[ thm : radical ] all ideals @xmath11 corresponding to points in @xmath10 are radical ideals .
we may assume that @xmath6 is an infinite field .
let @xmath33 , the group of invertible @xmath34-matrices modulo scaling , let @xmath35 be the borel subgroup of images of upper triangular matrices in @xmath36 , and let @xmath37 be the algebraic torus of images of diagonal matrices in @xmath36 .
then @xmath38 is a maximal torus in @xmath39 , and @xmath40 is a borel subgroup in @xmath39 .
we consider the action of these groups on the hilbert scheme @xmath10 .
the @xmath38-fixed points of @xmath10 are the monomial ideals that have the same @xmath7-graded hilbert function as @xmath13 .
it suffices to assume that @xmath11 is such a monomial ideal because every other ideal @xmath41 can be degenerated to a monomial ideal @xmath42 via grbner bases , and if @xmath43 is radical then so is @xmath44 .
we can further assume that @xmath11 is _ borel - fixed _ , which means that @xmath11 is fixed under the action of @xmath45 .
indeed , if @xmath46 are generic matrices of @xmath36 then we replace the ideal @xmath11 first by its image @xmath47 , and then by the initial monomial ideal @xmath48 .
the ideal @xmath49 is the _ multigraded generic initial ideal_. the same approach as in @xcite shows that @xmath50 is borel - fixed .
moreover , if @xmath50 is radical then so is @xmath11 .
hence , it suffices to show that every borel - fixed ideal @xmath11 in @xmath51 is a radical ideal .
our result will be a direct consequence of the following two claims : claim 1 : _ there is precisely one borel - fixed ideal @xmath18 in @xmath10 .
_ claim 2 : _ the unique borel - fixed ideal @xmath18 is radical .
_ we first describe the ideal @xmath18 and then prove that it has these properties .
let @xmath52 be any vector in the set @xmath53 we write @xmath54 for the ideal generated by all unknowns @xmath55 with @xmath56 and @xmath57 .
this is a borel - fixed prime monomial ideal .
consider the intersection of the prime ideals @xmath54 : @xmath58 the monomial ideal @xmath18 is a radical and borel - fixed .
each of its @xmath59 associated prime ideals @xmath54 has the same codimension @xmath60 .
we now apply conca s results in ( * ? ? ?
* section 5 ) .
he showed that @xmath18 has the same hilbert function as @xmath13 .
therefore , the ideal @xmath18 is the promised borel - fixed ideal in @xmath10 .
more precisely , ( * ? ? ?
* theorem 5.1 ) states that @xmath18 is precisely the generic initial ideal @xmath61 of the ideal of @xmath62-minors .
we claim that @xmath18 is the only borel - fixed monomial ideal in @xmath10 . to show this
, we apply results about the _ multidegree _ in @xcite .
the multidegree of the prime ideal @xmath54 is the monomial @xmath63 . by (
* theorem 8.44 ) , @xmath54 is the only unmixed borel - fixed monomial ideal having multidegree @xmath64 . by ( * ? ? ?
* theorem 8.53 ) , the borel - fixed ideal @xmath65 has the multidegree @xmath66/z;{\bf t}\bigr ) \quad = \quad \sum_{u \in u } { \bf t}^{u } \quad = \quad \sum_{u \in u } t_1^{u_1 } t_2^{u_2 } \cdots t_n^{u_n } .\ ] ] since the multidegree of a homogeneous ideal is determined by its hilbert series ( * ? ? ?
* claim 8.54 ) , we conclude that every ideal @xmath67 has multidegree ( [ eqn : multidegofz ] ) . now
, suppose that @xmath67 is borel - fixed .
since @xmath11 is monomial , each minimal primary component contributes at most one term @xmath68 to the multidegree ( [ eqn : multidegofz ] ) .
thus , by ( * ? ? ? * theorem 8.53 ) , the minimal primes of @xmath11 are precisely the prime ideals @xmath54 where @xmath52 runs over the elements of @xmath69 .
this implies @xmath70 .
however , since @xmath11 and @xmath18 have the same hilbert function in a positive grading , we conclude that @xmath71 , as desired .
our proof of theorem [ thm : radical ] was based on an idea that was suggested to us by michel brion . in @xcite
, brion proves that for any multiplicity - free subvariety of a flag variety , such as the diagonal in a product of projective spaces , there exists a flat degeneration to a reduced union of schubert varieties , which is our @xmath18 .
although theorem [ thm : radical ] only applies to the special case of the diagonal in a product of projective spaces , it establishes reducedness not just for _ some _ degeneration but for _ any _ ideal with the same multigraded hilbert function . our proof combined the nice argument from @xcite with the explicit description of the borel - fixed monomial ideal given by conca in @xcite
we now come to the question asked by conca in ( * ? ? ?
* conjecture 1.1 ) .
given any @xmath34-matrices @xmath46 with entries in @xmath6 , we apply them individually to the @xmath28 columns of the matrix @xmath3 , form the ideal of @xmath62-minors of the resulting @xmath72-matrix , and then take the initial monomial ideal @xmath73 with respect to some term order .
conca conjectures that ( [ eqn : concaideal ] ) is always a squarefree monomial ideal .
he proves this for generic @xmath74 by showing that ( [ eqn : concaideal ] ) equals the borel - fixed ideal @xmath18 constructed above .
theorem [ thm : radical ] implies the same conclusion under the much weaker hypothesis that the @xmath74 are invertible .
[ cor : itssquarefree ] for any invertible @xmath34-matrices @xmath75 and any term order on @xmath5 $ ] , the monomial ideal @xmath76 is squarefree .
applying invertible matrices @xmath74 to @xmath13 corresponds to taking the orbit of @xmath13 under the action of @xmath77 on @xmath10 .
therefore , ( [ eqn : concaideal ] ) is a monomial ideal that lies in @xmath10 . by theorem [ thm : radical ]
, it is radical and hence squarefree .
corollary [ cor : itssquarefree ] implies ( * ? ? ?
* conjecture 4.2 ) .
at present , we do not know how to prove conca s stronger conjecture to the effect that corollary [ cor : itssquarefree ] holds without the hypothesis that the matrices @xmath74 are invertible ( * ? ? ? * conjecture 1.1 ) .
one idea is to extend our study to multigraded hilbert schemes on @xmath5 $ ] whose defining hilbert function is bounded above by ( [ eqn : ourhf ] ) .
[ cor : connected ] the multigraded hilbert scheme @xmath10 is connected .
all ideals in @xmath10 can be connected to their common generic initial ideal @xmath18 by grbner degenerations . in what follows
we take a closer look at the combinatorics of the ideal @xmath18 .
the ideal @xmath18 is generated by all monomials @xmath78 where @xmath79 , @xmath80 , and @xmath81 .
the maximum degree of a minimal generator is @xmath82 .
this is the description of the ideal @xmath18 given by conca @xcite .
all ideals @xmath11 in @xmath10 share the same hilbert series in the ordinary @xmath83-grading , @xmath84/i)_r \cdot z^r \,\,\ , = \,\,\ , \frac{h(z)}{(1-z)^{n+d-1 } } .\ ] ] the _ @xmath85-polynomial _ in the numerator can be seen from the ideal of @xmath0-minors : @xmath86 note that @xmath87 is the common scalar degree of the ideals in @xmath10 .
[ cor : cm ] every ideal @xmath88 in @xmath10 is cohen - macaulay . since @xmath18 is the common generic initial ideal of all ideals
@xmath11 , it suffices to show that the borel - fixed ideal @xmath18 is cohen - macaulay .
let @xmath89 denote the @xmath90-dimensional simplicial complex corresponding to @xmath18 .
the vertices of @xmath89 are the @xmath91 matrix entries @xmath55 , and its facets are the @xmath92 sets @xmath93 which are complementary to the prime ideals @xmath54 .
we order the facets @xmath94 according to the lexicographic order on the vectors @xmath52 .
we claim that this ordering of the facets is a _ shelling _ of @xmath89 .
since the stanley - reisner ring of a shellable simplicial complex is cohen - macaulay ( * ? ?
* theorem iii.2.5 ) , this will imply corollary [ cor : cm ] . to verify the shelling property
we must show that every facet @xmath94 has a unique subset @xmath95 such that the faces of @xmath94 not containing @xmath95 are exactly those appearing as a face of an earlier facet .
if this condition holds then the @xmath85-polynomial can be read off from the shelling as follows : @xmath96 the unique subset of the facet @xmath94 with these desired properties equals @xmath97 indeed , suppose @xmath36 is a face common to @xmath94 and @xmath98 for some @xmath99 .
then @xmath100 for some @xmath101 , so @xmath36 does not contain @xmath102 .
conversely , suppose that @xmath36 is a face of @xmath94 which does not contain @xmath95 , and let @xmath103 be any element of @xmath104 .
since @xmath101 , @xmath105 is a facet of @xmath89 which contains @xmath36 and which comes earlier in our ordering .
the shellability of @xmath89 was mentioned in ( * ? ? ?
* remark 5.12 ) but no details were given there .
it would be interesting to know whether the simplicial complex @xmath106 of every monomial ideal @xmath11 in @xmath10 is shellable .
in this section we relate our multigraded hilbert scheme to classical constructions in algebraic geometry . for @xmath2
we recover the space of complete collineations and its git construction due to thaddeus in @xcite .
brion @xcite extended thaddeus work to the diagonal @xmath107 of any rational projective homogeneous variety @xmath14 . while the present study is restricted to the case @xmath108 , we believe that many of our results will extend to @xmath109 in brion s setting .
[ prop : grothen ] there is a injective morphism from the multigraded hilbert scheme @xmath10 to a connected component of the grothendieck hilbert scheme of subschemes of @xmath17 .
by @xcite , there is a natural morphism from the multigraded hilbert scheme @xmath10 to the grothendieck hilbert scheme .
theorem [ thm : radical ] shows that every point in @xmath10 corresponds to a radical ideal @xmath110 $ ] .
furthermore , the hilbert function tells us that the ideal @xmath111 has ( affine ) dimension @xmath112 , but by corollary [ cor : cm ] , @xmath11 is pure of dimension @xmath113 , so no associated prime of @xmath11 contains @xmath114 .
thus , @xmath11 is uniquely determined by the subscheme it defines in @xmath17 , so the morphism is injective .
we do not know whether the morphism in proposition [ prop : grothen ] is an immersion , nor whether it is always surjective .
recall that the @xmath39-action on @xmath10 transforms ideals as follows : @xmath115 if @xmath116 then the ideal @xmath11 is left invariant , so the stabilizer of any point @xmath67 contains the diagonal subgroup @xmath117 of @xmath39 . moreover , the stabilizer of the determinantal ideal @xmath13 is precisely the diagonal subgroup @xmath36 .
we write @xmath118 for the closure of @xmath119 in the hilbert scheme @xmath10 .
the subscheme @xmath120 is an irreducible component of @xmath10 .
it is a compactification of the homogeneous space @xmath121 , so it has dimension @xmath122 .
this theorem can be deduced from proposition [ prop : grothen ] using standard algebraic geometry arguments concerning the tangent sheaf of grothendieck s hilbert scheme .
what we present below is a more detailed combinatorial proof based on the identification of an explicit smooth point in lemma [ lem : diag ] . clearly , the dimension of the tangent space at @xmath13 is at least @xmath123 , the dimension of @xmath121 . by semi - continuity , it is bounded above by the tangent space dimension of @xmath10 at any initial monomial ideal @xmath124 . in lemma
[ lem : diag ] below , we identify a particular initial ideal for which this dimension equals @xmath123 . from this
we conclude that the tangent space of @xmath10 at @xmath13 has dimension @xmath125 .
this is precisely the dimension of the orbit closure @xmath120 of @xmath13 .
we also conclude that @xmath13 is a smooth point of @xmath10 , and that the unique irreducible component of @xmath10 containing @xmath13 is the compactified space @xmath120 .
let @xmath126 denote the ideal generated by the quadratic monomials @xmath127 for all @xmath128 and @xmath129 .
we call @xmath126 the _ chain ideal _ , because its irreducible components correspond to chains in the grid from @xmath130 to @xmath131 .
it is a point in @xmath132 since @xmath133 in the lexicographic order .
[ lem : diag ] the tangent space of the multigraded hilbert scheme @xmath10 at the chain ideal @xmath126 has dimension @xmath125 .
we claim that the following three classes @xmath134 , @xmath135 and @xmath136 form a basis for the tangent space @xmath137 } ( m , k[x]/m)_0 $ ] : _ class @xmath134 _ :
for each triple of indices @xmath138 with @xmath128 and @xmath139 we define a @xmath5$]-module homomorphism @xmath140/m\,$ ] by setting @xmath141 _ class @xmath135 _ : for each triple of indices @xmath142 with @xmath128 and @xmath143 we define a @xmath5$]-module homomorphism @xmath144/m\,$ ] by setting @xmath145 _ class @xmath136 _ :
for each pair of indices @xmath146 with @xmath147 and @xmath143 we define a @xmath5$]-module homomorphism @xmath148/m\,$ ] by setting @xmath149 the above @xmath5$]-linear maps are @xmath150-homogeneous of degree zero , and they are clearly linearly independent over @xmath6 .
there are @xmath60 maps in the class @xmath136 , and there are @xmath151 each in the classes @xmath134 and @xmath135 .
this adds up to the required total number of @xmath152 .
it remains to be seen that every @xmath150-graded @xmath5$]-module homomorphism from @xmath126 to @xmath5/m$ ] of degree zero is a @xmath6-linear combination of the above .
suppose that @xmath153/m$ ] is a module homomorphism .
then , for @xmath154 and @xmath155 , we can uniquely write @xmath156 as a linear combination of monomials not in @xmath126 . furthermore , by subtracting appropriate multiples of @xmath157 and @xmath158 , we can assume that the monomials in the linear combination do not include @xmath159 or @xmath160 .
suppose that for some @xmath161 , the coefficient of @xmath162 is some non - zero @xmath163 .
either @xmath164 or @xmath165 , and the two cases are symmetric under reversing the order of both the column indices and the row indices , so we assume the former .
for any @xmath166 such that @xmath167 and @xmath168 , the syzygies imply @xmath169 since the first term is non - zero in @xmath5/m$ ] , the monomial must be divisible by @xmath170 .
thus , either @xmath171 , or both @xmath172 and @xmath173 . in the first case ,
taking @xmath174 , and using the assumption that the coefficient of @xmath175 in @xmath176 is zero , we get a contradiction . in the second case ,
if @xmath177 , then we must only have one choice of @xmath166 and this forces @xmath178 .
therefore , @xmath179 is a linear combination of homomorphisms of class @xmath136 , and thus the classes of @xmath134 , @xmath135 , and @xmath136 span the tangent space at @xmath126 .
[ cor : chain - smooth ] the chain ideal @xmath126 is a smooth point on @xmath10 .
the unique irreducible component of @xmath10 containing @xmath126 is the completion @xmath120 .
we now turn to the case @xmath22 which is well - studied in the literature .
the compactification @xmath180 is the classical _ space of complete collineations _ , which was investigated by thaddeus in @xcite .
in fact , we have : the multigraded hilbert scheme @xmath181 is smooth and irreducible .
it coincides with the space of complete collineations : @xmath182 . up to relabeling
, the chain ideal is the only monomial ideal in @xmath181 .
this point is smooth by lemma [ cor : chain - smooth ] , and hence @xmath181 is smooth .
since it is connected by corollary [ cor : connected ] , we conclude that @xmath181 is also irreducible .
the results in @xcite show that the grothendieck hilbert scheme is isomorphic to the space of complete collineations , and in particular smooth and irreducible .
thus , the morphism in proposition [ prop : grothen ] is an isomorphism between @xmath181 and the space of complete collineations .
the representation of @xmath180 as a multigraded hilbert scheme @xmath181 gives rise to nice polynomial equations for the space of complete collineations .
namely , each ideal @xmath11 in @xmath183 is generated by @xmath184 equations of degree @xmath185 .
as there are @xmath186 monomials in @xmath5_{(1,1)}$ ] , this describes an embedding of @xmath181 into the grassmannian @xmath187 .
the subscheme @xmath181 of this grassmannian is cut out by the determinantal equations which are derived by requiring that the ideal @xmath11 has the correct number of first syzygies in degrees @xmath188 and @xmath189 .
[ equations defining @xmath190 we shall realize the @xmath191-dimensional manifold @xmath192 as a closed subscheme of the @xmath193-dimensional grassmannian @xmath194 , by giving explicit equations in the @xmath195 plcker coordinates .
our equations furnish an explicit projective embedding for thaddeus git construction @xcite which is reviewed further below .
fix a @xmath196-matrix of unknowns @xmath197 consider the ideal @xmath11 generated by the three bilinear polynomials in the vector @xmath198 the condition for @xmath11 to be a point in @xmath192 is equivalent to the condition that the rows of the following two @xmath199-matrices are linearly dependent : @xmath200\ ] ] @xmath201\ ] ] these two matrices are obtained by multiplying the generators of @xmath11 with the entries in the two columns of @xmath202 respectively .
this results in nine polynomials of bidegree @xmath189 and nine polynomials of bidegree @xmath188 , each having @xmath193 terms .
these two sets of polynomials must be linearly dependent because each @xmath203 has its first syzygies in these two bidegrees .
the @xmath195 maximal minors of the matrix @xmath204 are the plcker coordinates @xmath205 on the grassmannian @xmath194 , where the indices run from @xmath206 to @xmath207 . using laplace expansion , we write each @xmath208-minor of the two matrices as a cubic polynomial in these plcker coordinates .
the condition that the matrices have rank at most eight translates into a system of homogeneous cubic polynomials in the @xmath195 unknowns @xmath209 , and these cubics define the space of complete collineations , @xmath210 , as a subscheme of @xmath194 .
thaddeus @xcite realizes @xmath192 as the ( chow or git ) quotient of the grassmannian @xmath211 by the one - dimensional subtorus of @xmath212 given by the diagonal matrices with entries @xmath213 .
we can see this in our equations as follows .
let @xmath214 and @xmath215 be @xmath216-matrices of unknowns .
each point in @xmath211 is represented as the row space of the @xmath217-matrix @xmath218 $ ] .
the group @xmath219 acts on @xmath192 by translating the distinguished point @xmath13 to the ideal generated by the three quadrics @xmath220 the entries of the corresponding @xmath196 matrix @xmath204 are @xmath221 writing @xmath222 for the @xmath223-th column of the matrix @xmath69 and @xmath224 for the @xmath225-th column of @xmath226 , this translates into the following parametric representation of @xmath192 : @xmath227 \det[v_{j_2 } , u_{k_1},v_{k_2 } ] - \det[u_{i_1},v_{i_2 } , v_{j_2 } ] \det[u_{j_1 } , u_{k_1},v_{k_2}].\ ] ] these are quadratic polynomials in the plcker coordinates on @xmath211 .
they are invariant under thaddeus torus action and antisymmetric under swapping each of the three index pairs .
note that of the @xmath195 polynomials only @xmath228 are actually plcker binomials .
of the others , @xmath229 are zero ( for example , @xmath230 ) and @xmath231 are plcker monomials ( for example , @xmath232 \det[v_1 , u_3 , v_2]$ ] ) .
this resulting map @xmath233 gives an embedding of thaddeus quotient @xmath234 .
the cubic relations on @xmath194 described above characterize the image of this embedding .
[ remark : lafforgue ] in the introduction of @xcite , lafforgue describes the following compactification of @xmath121 .
we consider the @xmath34-minors of the @xmath235-matrix @xmath236 .
for each minor there is a corresponding vector @xmath237 in the set @xmath238 , namely , @xmath239 is the number of columns of @xmath240 occurring in that minor .
we introduce a new unknown @xmath241 for each @xmath242 , and we multiply each minor by the corresponding unknown @xmath241 .
the scaled minors parametrize a subvariety in an affine space of dimension @xmath243 this affine variety yields a projective variety @xmath244 which compactifies @xmath121 : @xmath245 in light of @xcite , we can identify @xmath244 with the partial multigraded hilbert scheme @xmath246 obtained by restricting @xmath10 to the subset of degrees @xmath247 . hence there is a natural morphism @xmath248 .
this is an isomorphism for @xmath249 and @xmath2 but we do not know whether this is always the case . in general , @xmath244 is singular , and the main result of @xcite is a combinatorial construction that replaces @xmath244 with another less singular
model @xmath250 .
yet , as discussed in the erratum to @xcite , @xmath251 is not smooth for @xmath252 .
this section concerns the case @xmath23 . the hilbert scheme @xmath24 parametrizes degenerations of the projective line in its diagonal embedding @xmath253 .
our goal is to prove the following two theorems about the structure of @xmath24 .
[ thm : irrbutsing ] the multigraded hilbert scheme @xmath24 is irreducible , so it equals the compactification @xmath254 .
however , @xmath24 is singular for @xmath25 .
our second theorem explains why we refer to @xmath24 as a _ space of trees_. the qualifier `` yet another '' has been prepended to emphasize that this is not the _ space of phylogenetic trees_. the latter is familiar to algebraic geometers as a discrete model for @xmath255 ; see ( * ? ? ?
* theorem 1.2 ) for a precise statement .
following @xcite , there is a natural graph structure on any multigraded hilbert scheme , including @xmath24 .
the vertices are the monomial ideals , and for every ideal in @xmath24 with precisely two initial monomial ideals there is an edge between the corresponding vertices . by (
* theorem 11 ) , this is precisely the induced subgraph on @xmath24 of the graph of all monomial ideals .
we note that our graph is not a _ gkm graph _ in the sense of @xcite because the @xmath38-fixed subvarieties corresponding to edges usually have dimension greater than one .
[ thm : yasot ] there are @xmath256 monomial ideals in @xmath24 , one for each tree on @xmath257 unlabeled vertices with @xmath28 labeled directed edges .
two trees are connected by an edge on @xmath24 if they differ by one of the following operations : 1 .
move any subset of the trees attached at a vertex to an adjacent vertex .
swap two edges that meet at a bivalent vertex ( preserving orientation ) . in this section
we use the following notation for our matrix of variables : @xmath258 thus @xmath259 are homogeneous coordinates on the @xmath260-th factor in our ambient space @xmath261 . the common hilbert function ( [ eqn : ourhf ] ) of all ideals @xmath11 in @xmath24 equals @xmath262 the unique borel - fixed ideal in @xmath24 equals @xmath263 .
our first goal is to prove that @xmath24 is irreducible .
this requires a combinatorial description of the subvarieties @xmath264 of @xmath261 corresponding to ideals @xmath265 .
note that each such subvariety is a reduced curve of multidegree @xmath266 in @xmath261 .
[ lem:43 ] the variety @xmath267 defined by any ideal @xmath265 is the reduced union of several copies of the projective line @xmath268 . for each factor of @xmath261
there is exactly one component of @xmath264 which is not constant along this factor , and for this component , the projection induces an isomorphism .
consider the projection from @xmath264 onto the @xmath260-th factor of @xmath261 .
we infer from the hilbert function ( [ eqn : ourhf2 ] ) that this projection is an isomorphism over an open subset of @xmath269 .
hence there exists a rational map from @xmath270 to a unique component @xmath271 of @xmath264 .
the fact that the projection @xmath272 is a regular morphism implies that the curve @xmath271 is smooth .
we conclude that the map @xmath272 is an isomorphism .
each component of @xmath264 can be labeled by the factors onto which it maps isomorphically .
we draw @xmath264 as a set of intersecting lines , labeled with subsets of the factors . by lemma [ lem:43 ] ,
the labels form a partition of @xmath273 . moreover , since @xmath5/i$ ] is cohen - macaulay , @xmath264 is connected , and because only one component is non - constant along any factor , there is no cycle among its components .
hence our picture is an edge - labeled tree .
we have the following converse to this description of the points of @xmath24 .
[ prop : construction - ideals - h2n ] suppose that @xmath274 is a union of projective lines , which is connected and such that each factor of @xmath261 has a unique projective line projecting isomorphically onto it .
then the radical ideal @xmath11 defining @xmath271 is a point in @xmath24 .
we compute the hilbert function and show that it coincides with ( [ eqn : ourhf2 ] ) .
we proceed by induction on the number of components .
if @xmath271 is irreducible then @xmath271 is the translate of the diagonal @xmath275 in @xmath261 with some @xmath276 acting on the @xmath260-th factor , and therefore @xmath277 lies in @xmath24 .
now suppose @xmath271 is reducible , let @xmath278 be one of its components , and @xmath279 the index set of factors of @xmath261 onto which @xmath278 maps isomorphically .
the prime ideal @xmath280 of @xmath278 is generated by linear forms of multidegrees @xmath281 , and by the @xmath62-minors of a @xmath282-matrix @xmath283 which consists of the @xmath284 columns of @xmath14 acted on by some @xmath276 .
the hilbert function of @xmath280 is @xmath285 since @xmath271 is a tree of projective lines , there exists a component @xmath278 which has only one point of intersection with the other components .
let @xmath286 be the union of the other components and @xmath287 the radical ideal defining @xmath288 .
the ideal @xmath280 of @xmath278 contains linear forms of degree @xmath289 for every @xmath290 , while @xmath287 contains linear forms of degree @xmath289 for every @xmath291 .
this implies that @xmath292 is a homogenous prime ideal generated by linear forms .
its variety equals @xmath293 , and hence @xmath294 has constant hilbert function @xmath206 .
we conclude @xmath295 which is the common hilbert function ( [ eqn : ourhf2 ] ) of all ideals in @xmath24 .
our discussion shows that each point in @xmath24 is characterized by the following data .
first , there is a tree of projective lines @xmath296 , labeled by the parts in a partition @xmath297 .
these represent the factors of the ambient space @xmath261 .
the intersection point of two lines determines a marked point on each of the two lines . for each line labeled with more than one factor
, we have a compatible set of isomorphisms between those factors . given these data , we can compute the ideal @xmath280 of a component @xmath278 as follows : 1 .
let @xmath283 be the submatrix of @xmath14 given by the columns indexed by @xmath284 , acted on by the @xmath298-matrices corresponding to the isomorphisms of @xmath269 .
2 . for each @xmath290 locate the intersection point on @xmath299 that is nearest to @xmath278 . let @xmath300 be the linear form defining this intersection point on @xmath299 .
the ideal @xmath280 is generated by these linear forms and the @xmath0 minors of @xmath283 .
the intersection ideal @xmath301 is the desired point in @xmath24 .
the above algorithm implies that there are infinitely many @xmath302-orbits on @xmath24 when @xmath303 .
consider a tree of four lines , @xmath304 , @xmath305 , @xmath306 , and @xmath307 , which meet a fifth line in four distinct points , with coordinates @xmath308 and @xmath309 on @xmath310 .
each of these intersection points is identified with the point @xmath311 on the line @xmath278 .
then we have @xmath312 as @xmath313 varies over the field @xmath6 , the ideals @xmath11 lie in different @xmath314-orbits on @xmath315 because the cross ratio of the four points on @xmath316 is invariant under @xmath317 .
we shall prove that @xmath24 is irreducible .
let @xmath11 be any ideal in @xmath24 .
we use induction on the number of components of @xmath264 to show that @xmath11 is in the closure of the orbit of @xmath13 .
if @xmath264 is irreducible , then @xmath11 is in the orbit of @xmath13 by the above discussion , so we assume that @xmath264 has at least two components .
we shall construct another ideal @xmath318 such that @xmath319 has one fewer component than @xmath264 and such that @xmath44 degenerates to @xmath11 .
consider the tree picture of @xmath264 as described above , and let @xmath304 be a component which has exactly one point of intersection with the other components .
let @xmath305 be one of these components intersecting @xmath304 .
after relabelling the factors and a change of coordinates , we can assume that the isomorphisms of the factors associated to @xmath304 are all the identity map , and the same holds for the isomorphisms of factors of @xmath305 .
furthermore , we can assume that the point @xmath320 is defined in @xmath261 by the ideal @xmath321 .
we now replace @xmath305 with the component @xmath322 labeled by the set @xmath323 ( with the identity isomorphisms ) , and we call its ideal @xmath287 . by proposition [ prop : construction - ideals - h2n ] , @xmath287 is in @xmath24 .
the ideal @xmath324 of @xmath322 is generated by @xmath325 and the @xmath0 minors of the submatrix of @xmath14 indexed by @xmath326 .
for @xmath327 , we consider the ideal formed by replacing @xmath328 by @xmath329 in @xmath287 for @xmath330 and take the flat limit as @xmath313 goes to @xmath331 .
the limit of @xmath324 under this action is @xmath332 , so the limit of @xmath287 is contained in @xmath11 . since @xmath11 and the limit of @xmath287 have the same hilbert function , they must be equal .
this proves the first assertion in theorem [ thm : irrbutsing ] .
the second assertion will follow from corollary [ cor : trivalent - smooth ] below .
we now come to the combinatorial description of monomial ideals @xmath11 on @xmath24 . here
the tree picture can be simplified .
there are precisely @xmath28 components @xmath333 , and the partition is into singletons @xmath334 .
each line @xmath299 has only two points where intersections are possible , namely , @xmath335 and @xmath336 .
we draw @xmath299 as an oriented line segment with the intersection points only at the end points .
the orientation is indicated by an arrow whose tail represents @xmath337 and whose head represents @xmath338 . in this manner , each monomial ideal @xmath11 in @xmath24 is represented uniquely by a tree @xmath37 with @xmath28 directed labeled edges .
the tree @xmath37 has @xmath339 vertices which remain unlabeled .
this establishes the first part of theorem [ thm : yasot ] .
our construction is illustrated for @xmath340 in figure [ fig : claw - chain ] .
see example [ example : twothree ] below for a combinatorial discussion of the two classes of trees shown here . of the @xmath341 monomial ideals on @xmath342 .
the tree on the left corresponds to the ideal @xmath343 and on the right corresponds to the borel - fixed ideal @xmath344 . ]
we next describe a rule for reading off the generators of a monomial ideal @xmath265 from its tree @xmath37 .
for any two distinct indices @xmath260 and @xmath345 in @xmath346 we set @xmath347 if the directed edge @xmath345 is pointing away from the edge @xmath260 in @xmath271 and @xmath348 otherwise .
this means that the ideal @xmath349 of the component @xmath299 is generated by the variables @xmath350 for @xmath351 . by intersecting these ideals for all @xmath260
we obtain the following combinatorial formula for the ideal @xmath11 .
the monomial ideal associated with the tree @xmath37 equals @xmath352 explicitly , the ideal generator corresponding to pair @xmath353 of edges equals @xmath354 note that the borel - fixed ideal @xmath18 corresponds to the star tree with all edges directed outwards .
our next result concerns tangent spaces of @xmath24 .
[ prop : tangentspace ] let @xmath88 be the monomial ideal corresponding to a directed tree @xmath37 as above .
then the dimension of the tangent space of @xmath24 at @xmath11 is @xmath355 where the sum is over all vertices of @xmath37 , and the function @xmath356 is defined by @xmath357 the following corollary to this result completes the proof of theorem [ thm : irrbutsing ] .
[ cor : trivalent - smooth ] the monomial ideal @xmath11 is a smooth point on the hilbert scheme @xmath24 if and only if every vertex in the tree @xmath37 is at most trivalent . the tree @xmath37 has @xmath27 vertices @xmath358 , and the number of edges is @xmath359 since @xmath24 is a compactification of @xmath360 , its dimension equals @xmath361 since @xmath362 , with equality if and only if @xmath363 , the sum in ( [ eqn : fsum ] ) is equal to @xmath364 , if and only if @xmath365 for all vertices @xmath358 . the star tree ideal @xmath366 has tangent space dimension @xmath367 .
in fact , it is the most singular point on @xmath24 , since every other ideal degenerates to @xmath18 . on the other hand ,
the chain ideal @xmath368 is a smooth point on @xmath24 , because the tree for @xmath126 is a chain of @xmath28 directed edges .
this confirms lemma [ lem : diag ] for @xmath23 . for any distinct edges @xmath369 and @xmath370 meeting at a vertex @xmath358 of the tree @xmath37 , we define the following tangent directions for @xmath24 at @xmath11 : @xmath371 here we use the convention that @xmath372 if @xmath348 and @xmath373 if @xmath347 . moreover , if @xmath358 is bivalent , i.e. @xmath369 and @xmath370 are the only edges incident to @xmath358 , define : @xmath374 if the vertex @xmath358 has degree @xmath375 , then we have defined @xmath376 maps . to show that these map are indeed tangent directions
, we exhibit a one - parameter deformation of @xmath11 .
the tangent vector @xmath377 is realized by the curve on @xmath24 gotten by replacing @xmath378 with @xmath379 among the generators of @xmath11 .
the resulting ideal in @xmath24 represents the tree of lines gotten by merging the edges @xmath369 and @xmath370 to a single @xmath269 labeled by @xmath380 .
the tangent vector @xmath381 is realized by replacing @xmath382 with @xmath383 in all prime components of @xmath280 such that @xmath345 connects to @xmath358 via @xmath370 .
the resulting ideal in @xmath24 represents the tree of lines gotten by sliding the subtree at @xmath358 in direction @xmath370 along the edge labeled @xmath369 .
as @xmath358 ranges over all vertices of the tree @xmath358 , and @xmath384 range over all incident edges , we now have a collection of tangent vectors whose cardinality equals ( [ eqn : fsum ] ) .
to see that these vectors are linearly independent , we note that @xmath385 is the only one of these tangent vectors such that the image of @xmath386 has a non - zero coefficient for @xmath387 and that @xmath388 is the only one such that the image of @xmath386 has a non - zero coefficient for @xmath389 .
thus , a non - trivial linear combination of the @xmath385 and @xmath388 ca nt be the zero tangent vector .
it remains to be seen that our tangent vectors span the tangent space .
suppose there exists a tangent vector @xmath179 that is not in the span of the @xmath385 and @xmath388 .
after subtracting suitable multiples of these known tangent vectors , we may assume that for any pair of adjacent edges @xmath260 and @xmath345 there exists a scalar @xmath390 such that @xmath391 , and furthermore @xmath392 if the node @xmath358 shared by @xmath260 and @xmath345 is bivalent .
suppose that @xmath358 has degree at least @xmath207 and let @xmath369 be an edge incident to @xmath358 distinct from @xmath260 and @xmath345 .
then @xmath393 and hence @xmath394 this implies @xmath395 .
we conclude that @xmath396 for any pair of adjacent edges @xmath260 and @xmath345 .
now suppose that @xmath260 and @xmath345 are not adjacent and write @xmath397 let @xmath370 be the edge adjacent to @xmath345 on the path from @xmath260 to @xmath345 .
then @xmath398 and @xmath399 this implies @xmath400 and , by symmetry , @xmath401 .
we conclude that @xmath402 , so our maps @xmath381 and @xmath388 form a basis for the tangent space of @xmath24 at @xmath11 .
we already saw that the monomial ideals on @xmath24 are in bijection with trees @xmath37 with @xmath257 unlabeled vertices and directed edges that are labeled with @xmath346 . to show that there are @xmath403 monomial ideals
, it suffices to show there are @xmath404 edge - labeled trees on @xmath257 vertices . picking an arbitrary node
as the root and shifting the labels to the nodes away from the root gives a rooted , node - labeled tree , of which there are @xmath405 . from the rooted tree
, we can uniquely recover the edge - labeled tree and the choice of the root , so there are @xmath404 edge - labeled trees .
we now need to identify all ideals @xmath11 in @xmath24 that possess precisely two initial monomial ideals .
we already saw two classes of such ideals in the proof of proposition [ prop : tangentspace ] .
first , there was the ideal with generator @xmath379 which realizes the deformation @xmath388 and swap # 2 in the statement of theorem [ thm : yasot ] .
we also exhibited an ideal for the deformation @xmath381 which realizes the move # 1 when the subset of trees is a singleton .
the general case is subsumed by the following argument . in light of theorem
[ thm : radical ] and proposition [ prop : grothen ] , grbner degenerations of ideals @xmath11 in @xmath24 correspond to scheme - theoretic limits of the trees @xmath271 with respect to one - parameter subgroups of the @xmath406-action on @xmath261 .
let @xmath271 be a tree on @xmath24 that has precisely two degenerations to @xmath406-fixed trees .
there are two cases to be considered .
first suppose that some component @xmath299 of @xmath271 is labeled by a subset @xmath407 with @xmath408 .
the component @xmath299 admits @xmath409 distinct degenerations to a @xmath406-fixed tree , and , by proposition [ prop : construction - ideals - h2n ] , each of these lifts to a degeneration of @xmath271 .
this implies that the tree @xmath271 has @xmath410 edges , and the unique non - singleton label @xmath411 has cardinality two .
moreover , each intersection point @xmath412 is a torus - fixed point on both @xmath278 and @xmath413 .
this is precisely the situation in swap # 2 above . in the second case to be considered ,
the tree @xmath271 consists of the @xmath28 lines @xmath414 , each labeled by a singleton .
consider all components @xmath299 with intersection points that are not torus - fixed .
for each such component @xmath299 there exist two torus degenerations of @xmath271 that move the intersection points on @xmath299 to the two torus - fixed points on @xmath299 . under these degenerations ,
the intersection points @xmath412 with @xmath415 remain in their positions on @xmath278 and @xmath413 .
hence , only one component @xmath299 has intersection points that are not torus - fixed .
this is precisely the situation in move # 1 in theorem [ thm : yasot ] .
[ example : twothree ] ( @xmath340 ) the hilbert scheme @xmath342 has @xmath341 monomial ideals , corresponding to the eight orientations on the claw tree and to the eight orientations on each of the three labeled bivalent trees .
representatives for the two classes of trees are shown in figure [ fig : claw - chain ] .
the eight orientations of the claw tree can be arranged into the vertices of a cube .
each edge in the cube is an edge in the graph corresponding to moving two edges at a time between vertices . along each edge , add two vertices corresponding to bivalent trees and four edges from each of these to each adjacent vertex of the cube .
in addition to these operations corresponding to move # 1 , there are are 24 edges corresponding to swap # 2 .
these are arranged into four hexagons .
the six - dimensional manifold @xmath342 coincides with lafforgue s compactification @xmath416 in remark [ remark : lafforgue ] .
here , ( [ lafhook ] ) amounts to an embedding of @xmath417 into @xmath418 .
the equations for this embedding are as follows .
each @xmath419 parametrizes one of the three generators @xmath420 , @xmath421 , of an ideal in @xmath342 .
being a point in @xmath342 means that these ideal generators admit two linearly independent syzygies in degree @xmath422 .
this happens if and only if the following @xmath423-matrix has rank at most four : @xmath424 by saturating its ideal of @xmath425-minors with respect to the irrelevant ideal @xmath426 , we find that the prime ideal of @xmath427 is generated by nine cubics such as @xmath428 .
in this section we study the smallest case where @xmath10 is reducible , namely , @xmath429 . the multigraded hilbert scheme @xmath30 parametrizes degenerations of the projective plane in its diagonal embedding @xmath430 .
we use the following notation for the unknowns @xmath55 in the polynomial ring @xmath5 $ ] .
@xmath431 [ thm : three - planes ] the multigraded hilbert scheme @xmath30 is the reduced union of seven irreducible components , each of which contains a dense @xmath432 orbit : * the @xmath31-dimensional main component @xmath433 is singular . *
three @xmath434-dimensional smooth components are permuted under the @xmath435-action on @xmath436 . at a generic point ,
the subscheme of @xmath436 is the union of the blow - up of @xmath437 at a point , two copies of @xmath437 , and @xmath438 .
an ideal which represents such a point on this component is @xmath439 *
three @xmath440-dimensional smooth components are permuted under the @xmath435-action on @xmath436 .
a generic point looks like the union of three copies of @xmath437 and @xmath437 blown up at three points .
a representative ideal is @xmath441 with some additional notation , we can describe the isomorphism types of the six extra components .
let @xmath442 denote the variety of complete flags in @xmath443 and @xmath444 the tautological bundle of @xmath260-dimensional vector spaces for @xmath445 or @xmath446 .
then the second class of components are isomorphic to the bundle @xmath447 where @xmath448 is a bundle whose fibers are each isomorphic to the hilbert scheme @xmath342 .
a point in this bundle is equivalent to a point @xmath449 in @xmath450 , together with an ideal in the total coordinate ring of @xmath451 with the appropriate hilbert function ( [ eqn : ourhf ] ) . to relate this formulation to the ideal in ( [ eqn : ideal - extra-14 ] ) , we identify linear forms in @xmath5 $ ] with the direct sum of three vector spaces , each of dimension @xmath207 , and @xmath452 as choice of basis for the @xmath260th summand .
the two flag varieties parametrize the duals of the flags @xmath453 for @xmath454 .
the projective space parametrizes the point whose ideal is @xmath455 .
these spaces determine all the linear generators in ( [ eqn : ideal - extra-14 ] ) .
the additional generators of the first component represent a point in @xmath342 , but without a canonical choice of basis .
the third class of components are isomorphic to the projective bundle : @xmath456 where @xmath442 is as before and @xmath35 is the blow - up of @xmath457 along the diagonal .
we think of the blow - up variety @xmath35 as the parameter space of two points in @xmath437 and a line containing them .
the @xmath458-dimensional vector bundle @xmath459 is the sum @xmath460 inside @xmath461 , where @xmath462 and @xmath463 are the pullbacks to @xmath35 of @xmath464 on each of the copies of @xmath437 , which parametrize the two points .
the flag varieties parametrize the duals of @xmath453 for @xmath465 and @xmath35 parametrizes the two points defined by @xmath466 and @xmath467 , with @xmath468 as the line between them . as before ,
these vector spaces determine the linear generators of the components .
the bundle @xmath469 parametrizes the coefficients of the cubic generator of the ideal of the blowup of @xmath437 at two points .
this ideal equals @xmath470 note that the middle two terms are linearly independent even when @xmath471 and @xmath472 coincide . for generic coordinates , after a change of basis , we can take @xmath473 and @xmath474 to be zero , and after rescaling , we take @xmath475 .
thus , the @xmath432 orbit of the ideal ( [ eqn : ideal - extra-13 ] ) is dense in the @xmath440-dimensional component of @xmath30 .
the proof is computational .
it rests on the singular code posted at http://math.berkeley.edu/~dustin/diagonal/h33.sng .
the computation works regardless of what the field characteristic of @xmath6 is .
it suffices to consider an affine neighborhood of the unique borel - fixed ideal @xmath18 in @xmath30 .
we employ the standard method of chosing coordinates by adding trailing terms with indeterminate coefficients to the ten monomial generators of @xmath18 .
the ideal defining @xmath30 is then derived from the syzygies of @xmath18 .
we then derive the prime ideals representing each of the seven components , by translating the geometric descriptions above into local coordinates around @xmath18 .
implicitization using singular yields the seven prime ideals , and we check that their intersection equals the ideal of the hilbert scheme itself .
we now explain how the parametric representations of the seven components are derived .
the main component is , by definition , parametrized by the @xmath432 orbit of the ideal of @xmath476 minors of @xmath14 .
the other components can also be parametrized by @xmath432 orbits of the representative ideals , but it is also possible and computationally more efficient to use parametrizations which do not require localization .
for the @xmath434-dimensional components , we begin by using local coordinates in @xmath342 to define a family of subschemes of @xmath477 over @xmath478 . renaming some of the variables and adding two linear terms , we get the parametrization of the blow - up of @xmath479 and its degenerations in @xmath480 , i.e. a neighborhood of a fiber of ( [ eqn : extra-14-bundle ] ) .
intersecting with the ideals of linear spaces gives the family in @xmath5 $ ] with the appropriate hilbert function .
different choices of flags can be represented by upper triangular changes of coordinates on each of the three columns of @xmath14 .
the parametrization of the @xmath440-dimensional component follows the same pattern . in a neighborhood of @xmath18 , up to change of basis , we can take the two points parametrized by @xmath35 to be @xmath481 and @xmath482 with @xmath468 as the line between them .
in addition to @xmath375 , the other coordinates are the entries of the upper triangular matrix corresponding to the choice of flag and to the coefficients of the cubic generator in ( [ eqn:13-cubic ] ) .
implicitizing these parametrizations reveals the prime ideals for these seven components , and their intersection is found to equal the ideal of the hilbert scheme itself .
.the symmetry classes of monomial ideals in @xmath30 [ cols="<,^,^,^,^,^,^ " , ] in table [ tbl : monomials ] we show that @xmath30 contains @xmath483 monomial ideals .
they come in @xmath31 symmetry classes , and we list them in four groups , corresponding to the dimension ( @xmath228 , @xmath484 , @xmath485 , and @xmath486 ) of the orbit under the action of @xmath432 .
the @xmath31 monomial ideals appear in the same order as their pictorial representation in figure [ fig : monomial - poset ] .
the third column indicates whether or not the picture is planar .
the second column is the tangent space dimension of @xmath30 at that point .
the triple column shows which components the monomial ideals live on .
the rightmost column shows the order of the symmetry group of the ideal .
note that the permutation group acting on @xmath30 has order @xmath487 : it permutes the three factors of @xmath32 as well as the three coordinates @xmath488 of each projective plane .
the total number @xmath483 of monomial ideals on @xmath30 equals @xmath489 times the sum of the reciprocals in the last column of table [ tbl : monomials ] .
every monomial ideal in @xmath30 corresponds to a polyhedral complex in the boundary in the direct product of three triangles , denoted @xmath490 . using the moment map of toric geometry , each such polyhedral complex can be identified with the real positive points of the corresponding subscheme of @xmath491 .
the following conditions characterize those subcomplexes of @xmath30 whose corresponding monomial ideal has the right multigraded hilbert function : * for every vector of non - negative integers @xmath492 summing to @xmath446 , there is exactly one @xmath446-dimensional cell consisting of the product of a @xmath493-dimensional cell of @xmath494 , a @xmath495-dimensional cell , and a @xmath496-dimensional cell . *
the complex contains exactly ten @xmath331-cells , @xmath497 @xmath206-cells , and six @xmath446-cells .
this characterization is sufficient to show that table [ tbl : monomials ] is complete . by the condition on the number of @xmath206-cells ,
each triangle must meet at least two of the squares .
thus , each pair of squares must be adjacent or be connected by a triangle , and so all three squares must meet in a common point . there are four possible configurations of the squares : either @xmath331 , @xmath206 , @xmath446 , or @xmath207 edges common to multiple squares .
the monomial ideals can be enumerated by considering all possible ways to attach the additional triangles to these configurations .
the monomial ideals form a poset based on containment within the closures of their orbits , illustrated in figure [ fig : monomial - poset ] .
each ideal is drawn as a @xmath446-dimensional subcomplex of @xmath498 .
the subcomplex is drawn abstractly , but the embedding amounts to a choice of labellings .
the bold lines indicate that an additional triangle is attached along that edge . by orbits , we mean orbits under the disconnected group which is generated by multiplying the first column by an arbitrary matrix and by the discrete action of permuting the columns .
the number on the lower right is the dimension of the tangent space of @xmath30 at that monomial ideal .
the ranking is by the dimension of ( every component of ) the orbit of the monomial ideal : 9 , 10 , 11 , or 12 . the maximal elements of the poset in figure [ fig : monomial - poset ] correspond to the `` planar '' complexes , i.e. those such that no edge contains more than two 2-cells . by fixing a isomorphisms between the three copies of @xmath494
, we get a projection from @xmath490 onto @xmath499 . in the case of ideals 1 , 2 , 3 , 4 , and 5
the corresponding subcomplexes project to tilings of @xmath499 . in these cases , there is a monomial ideal in the orbit which is in the toric hilbert scheme of @xmath500 , and the corresponding triangulation of @xmath500 is related to the tiling of @xmath499 by the cayley trick .
the tilings in ( * ? ? ?
* figure 5 ) correspond to the monomial ideals 1 , 3 , 4 , 5 , and 2 in this order .
each triangulation is regular , and the ideals are smoothable , even in the toric hilbert scheme . here , the toric hilbert is the subscheme of @xmath30 obtained by fixing the hilbert function of @xmath13 with respect to the finer grading given by both row degrees and column degrees .
for details , references and further information see @xcite and ( * ? ? ?
* theorem 2 ) . ]
the original motivation which started this project was a discussion with annette werner about tropical convexity , and its connection to affine buildings and moduli of hyperplane arrangements as developed by keel and tevelev @xcite .
our aim was to understand the deligne schemes of @xcite and their special fibers in the concrete language of combinatorial commutative algebra .
we found that the multigraded hilbert scheme @xmath10 offers a suitable framework for studying deligne schemes and their arithmetic . in this section
we briefly discuss the set - up and the connection to the combinatorial results in @xcite .
we plan to pursue this further in a joint project with annette werner .
let @xmath6 be an algebraically closed field with a non - trivial non - archimedean absolute value , let @xmath369 be the residue field of @xmath6 , and @xmath501 the valuation ring of @xmath6 . in computational studies ( such as @xcite )
we usually relax the requirement that @xmath6 be algebraically closed , and we work with the grbner - friendly scenario @xmath502 , @xmath503 $ ] and @xmath504 .
let @xmath505 denote the bruhat - tits building associated with the group @xmath506 over @xmath6 as defined in @xcite .
the building @xmath505 is an infinite simplicial complex of dimension @xmath507 whose vertices are the equivalence classes of @xmath501-submodules of @xmath508 having maximal rank @xmath509 .
let @xmath510 be a finite set of vertices of the affine building @xmath505 . following ( * ? ? ?
* definition 1.8 ) , we let @xmath511 denote the corresponding join of projective spaces over @xmath501 , and we write @xmath512 for its special fiber over @xmath369 . in the special case when @xmath271 is a convex subset of @xmath505 , a classical result of mustafin @xcite states that @xmath511 is semi - stable over @xmath501 , which implies that @xmath512 has smooth irreducible components with normal crossings . in this section
we allow @xmath271 to be any finite set of vertices not necessarily convex of the building @xmath505 .
following @xcite , we shall call @xmath511 the _ deligne scheme _ of the subset @xmath513 .
we now describe the deligne scheme @xmath514 and its special fiber @xmath512 in concrete terms .
the configuration @xmath271 is represented by @xmath515 , an @xmath28-tuple of invertible @xmath34-matrices with entries in the field @xmath6 .
this data is the input for the algorithm of @xcite which computes the convex hull of @xmath271 in @xmath516 .
let @xmath13 be the ideal of @xmath0-minors of a @xmath4-matrix of unknowns .
we consider the transformed ideal @xmath517 in @xmath5 $ ] , and we intersect it with @xmath518 $ ] : @xmath519 .\ ] ] we call @xmath520 $ ] the _ deligne ideal _ of the point configuration @xmath521 .
the image of @xmath522 under the specialization @xmath523 is denoted @xmath524 $ ] .
we call @xmath525 the _ special fiber ideal _ of @xmath271 .
this nomenclature is justified as follows .
[ rem : deligne ] the deligne scheme @xmath514 coincides with the subscheme of @xmath526 defined by the deligne ideal @xmath522 , and its special fiber @xmath512 coincides with the subscheme of @xmath527 defined by the special fiber ideal @xmath525 . remark
[ rem : deligne ] implies that the deligne scheme @xmath511 is a point in the multigraded hilbert scheme @xmath528 over the valuation ring @xmath501 , and its special fiber @xmath512 is a point in @xmath529 over its residue field @xmath369 .
thus our study in sections 2 to 5 is relevant for deligne schemes .
in particular , theorem [ thm : radical ] implies : the formula ( [ getspecfib ] ) translates into the following grbner - based algorithm for computing deligne schemes and their special fibers when @xmath531 , k = { \mathbb{q}}$ ] .
the input data is an @xmath28-tuple @xmath532 of invertible @xmath34-matrices whose entries are rational functions in one variable @xmath533 .
the deligne ideal @xmath534 is an ideal in the polynomial ring @xmath535 $ ] .
it is computed as follows .
we replace the @xmath345-th column of the matrix @xmath3 by its left multiplication with the @xmath345-th input matrix @xmath536 .
we then form the @xmath62-minors of the resulting matrix and multiply each generator by a power of @xmath533 to get an ideal @xmath537 in @xmath535 $ ] .
the deligne ideal is then obtained by saturation as follows : @xmath538 the special fiber ideal is obtained by setting @xmath533 to zero in the deligne ideal : @xmath539 this algorithm generalizes the construction of block and yu @xcite which concerns the special case when the configuration @xmath532 lies in one apartment of the bruhat - tits building @xmath505 .
algebraically , this means that the @xmath536 are diagonal matrices , and , geometrically , the apartment of @xmath505 is identified with the standard triangulation of tropical projective space @xmath540 .
if the diagonal entries of the @xmath536 are monomials with sufficiently generic exponents , then ( [ sat1])-([sat2 ] ) simply amounts to a grbner basis computation for the ideal @xmath13 .
block and yu @xcite consider the alexander dual of the initial monomial ideal of @xmath13 , and they showed that minimal free resolution of that alexander dual is cellular .
it represents the convex hull of @xmath532 in @xmath540 , and hence in @xmath516 .
we conjecture that this method can be adapted to compute the convex hull of @xmath532 even if the matrices @xmath536 do not lie in a common apartment of @xmath505 .
we note that the algorithm ( [ sat1])-([sat2 ] ) is not very practical for computing the special fiber ideal @xmath541 .
we found that running the saturation step ( [ sat1 ] ) in naive manner in macaulay 2 is too slow for interesting values of @xmath509 and @xmath28 .
instead , we propose the following approach . if the @xmath532 are `` sufficiently generic , '' then we replace each @xmath536 by a `` nearby '' matrix of the special form @xmath542 here the weights @xmath543 are generic rational numbers that specify a term order @xmath544 on the polynomial ring @xmath545 $ ] , and @xmath546 is a tuple of invertible matrices over @xmath547 .
the precise meaning of the approximation ( [ approx1 ] ) is that @xmath548 when this holds the special fiber @xmath549 of the deligne scheme @xmath550 is given by a squarefree monomial ideal @xmath541 .
the point is that the right hand side of ( [ approx2 ] ) can be computed much faster in practice than evaluating ( [ sat1 ] ) .
it amounts to computing a grbner basis of the transformed ideal @xmath551 in the polynomial ring @xmath545 $ ] . when the @xmath74 are diagonal matrices over @xmath547 this is precisely the algorithm of block and yu @xcite for convex hulls in @xmath540 .
* acknowledgements .
* both authors were supported by the u.s . national science foundation ( dms-0354321 , dms-0456960 and dms-0757236 ) .
we thank michel brion , aldo conca and annette werner for helpful comments .
f. santos : the cayley trick and triangulations of products of simplices , in _
integer points in polyhedra - geometry , number theory , algebra , optimization _ , ( eds .
a. barvinok , m. beck , c. haase , b. reznick , v. welker ) , contemporary mathematics * 374 * , ams , providence , 2005 . | the diagonal in a product of projective spaces is cut out by the ideal of @xmath0-minors of a matrix of unknowns . the multigraded hilbert scheme which classifies its degenerations has a unique borel - fixed ideal .
this hilbert scheme is generally reducible , and its main component is a compactification of @xmath1 . for @xmath2
we recover the manifold of complete collineations . for projective lines
we obtain a novel space of trees that is irreducible but singular .
all ideals in our hilbert scheme are radical .
we also explore connections to affine buildings and deligne schemes . |
nearly 30 years ago , toomre & toomre ( 1972 ) elegantly demonstrated that the tails and bridges emanating from many peculiar galaxies may arise kinematically from dynamically cold disk material torn off of the outer regions of galaxies experiencing strong gravitational interactions .
early spectroscopic studies of gas within the tidal tails of merging galaxies provided observational support for this hypothesis by showing the tails to have the kinematics expected for a gravitational origin ( e.g. stockton 1974a , b ) .
h@xmath0i mapping studies are particularly well suited to such studies , as the tidally ejected disk material is usually rich in neutral hydrogen and can be traced to very large distances from the merging systems ( e.g. van der hulst 1979 ; simkin _ et al . _ 1986
; appleton _ et al . _
1981 , 1987 ; yun _ et al . _ 1994 ) .
once mapped , the tidal kinematics can be used either alone , to disentangle the approximate spin geometry of the encounter ( stockton 1974a , b ; mihos _ et al . _ 1993
; hibbard & van gorkom 1996 , hereafter hvg96 ; mihos & bothun 1998 ) , or in concert with detailed numerical models , to constrain the full encounter geometry ( e.g. combes 1978 ; combes _ et al . _ 1988
; yun 1992 , 1997 ; hibbard & mihos 1995 ; gardiner & noguchi 1996 ) .
however , not all systems can be easily explained by purely gravitational models such as those used by toomre & toomre .
for example , gravitational forces by themselves should not lead to differences between stellar and gaseous tidal components .
numerical models which include hydrodynamical effects do predict a decoupling of the dissipative gaseous and non - dissipative stellar components ( e.g. noguchi 1988 ; barnes & hernquist 1991 , 1996 ; weil & hernquist 1993 ; mihos & hernquist 1996 ; appleton , charmandaris & struck 1996 ; struck 1997 ) , but only in the inner regions or along bridges where gas orbits may physically intersect ( see e.g. fig . 4 of mihos & hernquist 1996 ) .
decoupling of the gaseous and stellar components within the tidal tails is not expected . nonetheless , differences between the optical and gaseous tidal morphologies have been observed .
these differences can be subtle , with the peak optical and h@xmath0i surface brightnesses simply displaced by a few kpc within the tails ( e.g. ngc 4747 , wevers _
et al . _ 1984
; ngc 2782 smith 1994 ; ngc 7714/4 smith _ et al . _ 1997
; arp 295a , ngc 4676b , and ngc 520 southern tail , hibbard 1995 , hvg96 ) , or they can be extreme , with extensive h@xmath0i tidal features apparently decoupled from , or even anti - correlated with , the optical tidal features .
it is this latter category of objects that we wish to address in this paper . in particular
, we address the morphology of the tidal gas and starlight in the merging systems ngc 520 ( arp 157 ) , arp 220 , and arp 299 ( ngc 3690 ) .
the three systems were observed as part of our on - going studies on the tidal morphologies of optically and ir selected mergers ( hibbard 1995 , hvg96 , hibbard & yun 1996 and in prep . ) .
these studies involve moderate resolution ( @xmath3 ) vla h@xmath0ispectral - line mapping observations and deep optical @xmath4 and @xmath5broad - band imaging with large format ccds using the kpno 0.9 m ( ngc 520 ) and the university of hawaii 88 telescopes .
the h@xmath0i and optical observations , reduction , and data products have been presented in hibbard ( 1995 ) and hvg96 for ngc 520 , in hibbard & yun ( 1999 , hereafter hy99 ) for arp 299 , and in yun & hibbard ( 2000 ; see also hibbard & yun 1996 ) for arp 220 .
we refer the reader to these papers for details of the observations and data reduction .
these systems are extremely disturbed , and we can not hope to offer a full description of their peculiarities here . for more information we refer the reader to the above references .
figures [ fig : n520mos][fig : a220mos ] show the optical and atomic gas morphologies of each of the three systems discussed here .
for ngc 520 and arp 220 only the inner regions are shown in order to highlight the differences we wish to address .
panel * ( a ) * presents a greyscale representation of the optical morphology of each system with features of interest labeled .
panel * ( b ) * shows the h@xmath0i distribution .
contours indicate the distribution of h@xmath0imapped at low - resolution ( @xmath6 ) , whereas the greyscales show the h@xmath0i mapped at higher resolution ( @xmath3 ) .
the former is sensitive to diffuse low column density ( @xmath7 ) neutral hydrogen , while the latter delineates the distribution of the higher column density h@xmath0i .
the central region of each h@xmath0i map appears to have a hole ( indicated by the dotted contours ) , which is due to h@xmath0i absorption against the radio continuum associated with the disk - wide starbursts taking place in each galaxy ( see condon _
et al . _
1990 ) . in panel
* ( c ) * , we again present the optical morphology in greyscales , and the higher resolution h@xmath0i distribution as contours .
finally , panel * ( d ) * presents a smoothed , star - subtracted @xmath5-band image contoured upon a greyscale representation of the high - resolution h@xmath0i map . in the final panels of figs .
[ fig : n520mos][fig : a220mos ] dashed lines labeled `` slice '' indicate the locations from which h@xmath0iand optical intensity profiles have been extracted ; these profiles are plotted in figure [ fig : slices ] .
arrows labeled superwind " indicate the position angle ( p.a . ) of h@xmath8 or soft x - ray plumes , believed to arise from a starburst - driven outflow or galactic superwind in each system .
such outflows are common in other ir bright starbursts ( e.g. heckman , armus & miley 1987 , 1990 hereafter ham90 ; armus , heckman , & miley 1990 ; lehnert & heckman 1996 ) , and are thought to arise when the mechanical energy from massive stars and supernovae in the central starburst is sufficient to drive the dense interstellar medium outward along the minor axis ( e.g. chevalier & clegg 1985 ; joseph & wright 1985 ; suchkov _ et al . _ 1994 )
. often , such starbursts are powerful enough to drive a freely expanding wind of hot plasma completely out of the galaxy ( blowout " ; ham90 ) . in the following subsections
we briefly discuss what is known about the dynamical state of each system , and describe the differences between the stellar and gaseous tidal morphologies . throughout this
paper distances and other physical properties are calculated assuming @xmath9 .
ngc 520 ( arp 157 , ugc 966 ) is an intermediate - stage merger , with the two progenitor nuclei separated by 40 ( 5.8 kpc for @xmath10= 30 mpc , 1 = 145 pc ) and embedded within a common luminous envelope ( hvg96 and references therein ; see fig .
[ fig : n520mos]a ) .
there is a bright optical tidal tail stretching 24 kpc to the southeast ( henceforth referred to as the s tail ) which bends sharply eastward and connects onto a broad optical plume .
this plume continues to the north and west for 60 kpc before it appears to connect onto extended light surrounding the dwarf galaxy ugc 957 ( outside the region plotted in fig .
[ fig : n520mos]a ; see stockton & bertola 1980 ) .
the primary nucleus ( the easternmost nucleus in fig .
[ fig : n520mos]a ) possesses a massive ( @xmath11 ) 1 kpc - scale rotating molecular gas disk ( sanders _ et al . _ 1988
; yun & hibbard 1999 ) . an h@xmath0idisk is kinematically centered on this molecular disk , and extends to a radius of @xmath12 20 kpc ( labeled `` inner disk '' in fig .
[ fig : n520mos]b ) . beyond this
there is an intermediate ring of h@xmath0i with a mean radius of @xmath12 30 kpc ( i.e. , the material which contains the feature labeled `` n clump '' in fig .
[ fig : n520mos]b ) , and a nearly complete outer ring of h@xmath0i with a mean radius of 60 kpc , which extends smoothly through the dwarf galaxy ucg 957 ( only partially seen in fig .
[ fig : n520mos]b ) .
there is a kinematic and morphological continuity between the molecular gas disk , the inner h@xmath0i disk , and outer h@xmath0i ring ( yun & hibbard 1999 ) , which suggests that all of this material is associated with the primary nucleus .
the observations suggests that the ngc 520 interaction involved a prograde - retrograde or prograde - polar spin geometry : the linear morphology of the optical tail - to - plume system is typical of features produced by a disk experiencing a prograde encounter ( i.e. , the disk rotates in the same direction as the merging systems orbit each other ) .
the disk - like morphology and rotational kinematics of the large - scale hi and the lack of any aligned linear tidal features , on the other hand , are more typical of polar or retrograde encounter geometries ( i.e. , disk rotation either perpendicular to or opposite the direction of orbital motion ) .
such encounters fail to raise significant tails ( toomre & toomre 1972 ; barnes 1988 ) , and much of the disk material remains close to its original rotational plane .
neither the intermediate nor the outer h@xmath0i ring has an optical counterpart ( @xmath13@xmath14 27 mag arcsec@xmath15 ) . despite smooth rotational kinematics ,
this outer h@xmath0i has a very clumpy and irregular morphology , with notable gaps near the optical minor axis ( labeled `` ne gap '' and `` sw gap '' in fig .
[ fig : n520mos]c ) .
this figure shows that the outer h@xmath0i and optical structures are anti - correlated , with the peak h@xmath0i column densities ( associated with the n clump ) located to one side of the optical plume . in fig .
[ fig : n520mos]d the h@xmath0i clump appears to be bounded on three sides by the optical contours . in fig .
[ fig : slices]a we present an intensity profile at the location indicated by the dotted line in fig .
[ fig : n520mos]d , showing that the gas column density increases precisely where the optical light decreases .
while the h@xmath0i features exhibit a clear rotational kinematic signature , the well - defined edges and nearly - linear structure of the optical plume suggests that its constituent stars are moving predominantly along the plume , rather than in the plane of the sky : any substantial differential rotation would increase the width of the plume and result in a more disk - like morphology .
we therefore conclude that the gas rings and optical plume are both morphologically and kinematically distinct entities .
this suggests that the observed gas / star anti - correlation is either transient ( and fortuitous ) or actively maintained by some process .
a deep h@xmath8 image of ngc 520 shows plumes of ionized gas emerging both north and south along the minor axis and reaching a projected height of 3 kpc from the nucleus ( hvg96 ) .
it has been suggested that this plume represents a starburst - driven outflow of ionized gas ( hvg96 , norman _
et al . _
the position angle of this plume is indicated by an arrow in fig .
[ fig : slices]d ( p.a .
this direction corresponds to the most dramatic h@xmath0i / optical anti - correlations mentioned above , and in the following we suggest that this region of the optical tail actually lies directly in the path of the out - flowing wind .
arp 299 ( ngc 3690/ic 694 , ugc 6471/2 , mrk 171 , vv 118 ) is also an intermediate stage merger , with two disk systems ( ic 694 to the east , ngc 3690 to the west see figure [ fig : a299mos]a ) in close contact but with their respective nuclei separated by 20 ( 4.7 kpc for @xmath10= 48 mpc , 1 = 233 pc ) .
a long , narrow , faint ( @xmath13@xmath16 26mag arcsec@xmath15 ) tidal tail stretches to the north to a radius of @xmath12 125 kpc .
h@xmath0i imaging of this system by hy99 ( see also nordgren et al .
1997 ) shows a rotating gas - rich disk within the inner regions , and a pair of parallel h@xmath0i filaments extending to the north . from the h@xmath0i morphology and kinematics , hy99
deduce that arp 299 is the result of a prograde - retrograde or prograde - polar encounter between two late - type spirals , with the inner h@xmath0i disk associated with the retrograde disk of ic 694 , and the northern optical tail and tidal h@xmath0ifilaments ejected by the prograde disk of ngc 3690 .
the parallel - filament or bifurcated morphology of the tidal h@xmath0i is quite unlike that of the optical tail .
the inner h@xmath0i filament ( so labeled in fig . [ fig : a299mos]b ) is of lower characteristic column density ( @xmath7@xmath17 8) and is associated with the low surface brightness stellar tail ( fig .
[ fig : a299mos]c ) .
the gas in this filament has a more irregular morphology than that in the outer filament ( e.g. , the `` gap '' and `` knot '' in fig . [ fig : a299mos]b ) , and much of this material is detectable only after a substantial smoothing of the data ( fig .
[ fig : a299mos]b ) .
the outer filament is characterized by a higher h@xmath0i column density ( @[email protected] ) but has no optical counterpart ( @xmath13@xmath14 27.5 mag arcsec@xmath15 ) .
this filament is displaced by approximately 20 kpc ( in projection ) to the west of the inner filament for much its length , after which the filaments merge together in a feature labeled the `` n clump '' in fig .
[ fig : a299mos]b .
the parallel filaments have nearly identical kinematics along their entire lengths , and join smoothly at the n clump .
this implies that these features form a single physical structure .
based on preliminary numerical simulations , hy99 suggest that a bifurcated morphology can arise quite naturally during tail formation .
this occurs when the optically faint , gas - rich outer regions of the progenitor disk are projected adjacent to optically brighter regions coming from smaller initial radii ( see also mihos , 2000 ) .
however , this scenario does not explain why the inner filament , presumably drawn from optically bright but still gas - rich material within the optical disk of the progenitor , should lack accompanying h@xmath0i . as in ngc 520
, there is an anti - correlation between the optical and gaseous column densities across the n clump , with the highest gas column densities ( 23 ) located on either side of the optical tail .
this is illustrated in fig .
[ fig : slices]b , where we plot a profile along the position indicated by the dotted line in fig .
[ fig : a299mos]d .
the optical tail emerges above the n clump , and appears to curve exactly around the northern edge of the n clump ( labeled `` hook '' in fig .
[ fig : a299mos]c ) . also labeled in fig .
[ fig : a299mos]c are the three regions with anomalously high h@xmath0ivelocity dispersions noted by hy99 ( @xmath18 1320 km s@xmath19 compared with @xmath20 710 km s@xmath19 for the remainder of the tail ; see fig . 7d of hy99 ) ; we will will refer to these regions in the discussion ( [ sec : rpswind ] ) . within the main body of arp 299
, vigorous star formation is taking place , with an inferred star formation rate ( sfr ) of 50 @xmath21 ( ham90 ) .
recent x - ray observations reported by heckman _
_ ( 1999 ) show evidence for hot gas emerging from the inner regions and reaching 25 kpc to the north , which the authors interpret as evidence for a hot , expanding superwind .
the position angle of this feature ( p.a.=25 ) is indicated by the arrow in fig .
[ fig : a299mos]d , and points towards the inner tidal filament and n clump .
arp 220 ( ugc 9913 , ic 4453/4 ) is the prototypical ultraluminous infrared galaxy with @xmath22 ( soifer et al . 1984 ) .
it is an advanced merger system with two radio and infrared nuclei separated by 0.9 ( 345 pc for @xmath10= 79 mpc ) , and a bright optical plume extending 35 kpc to the nw ( fig .
[ fig : a220mos]a ) .
each of the two nuclei has its own compact molecular disk .
the two nuclear disks are in turn embedded in one larger 1 kpc scale molecular gas disk ( see scoville , yun , & bryant 1997 , downes & solomon 1998 , sakamoto _
et al . _
1999 and references therein ) .
the spin axis of the eastern nucleus is aligned with that of the kpc - scale disk while the western nucleus rotates in the opposite direction .
these observations suggest that arp 220 is the product of a prograde - retrograde merger of two gas rich spiral galaxies ( scoville , yun , & bryant 1997 ) .
an irregular disk - like distribution of neutral hydrogen extends over a 100 kpc diameter region surrounding the optical galaxy ( yun & hibbard 1999a ) .
the overall h@xmath0i kinematics indicates that this material has a component of rotation in the same sense as that for the eastern nucleus and the molecular gas disk , and opposite the rotation of the western nucleus .
this suggests that the h@xmath0i disk and eastern nucleus originated from the retrograde progenitor , while the the western nucleus and nw optical plume ( fig .
[ fig : a220mos]a ) arose from the prograde progenitor .
because of the vigorous star formation occurring within arp 220 ( sfr=340 @xmath21 , ham90 ) , much of the h@xmath0i within the optical body of the system is seen only in absorption against the bright radio continuum emission from the central starburst . beyond this
, the h@xmath0i has high column densities ( @[email protected] ) , but only to the ne and sw .
most notably , there are local h@xmath0i minima to the nw and se ( see gaps in fig . [
fig : a220mos]b ) .
comparison of the h@xmath0i map with the optical image ( fig .
[ fig : a220mos]c ) shows that the nw gap occurs exactly at the location of the optical tail .
the relationship between the optical and h@xmath0i surface brightness levels across this feature are illustrated by an intensity profile measured along the dotted line shown in fig .
[ fig : a220mos]d , and plotted in fig .
[ fig : slices]c . as in ngc 520 and arp 299
, the gas column density increases precisely where the optical light from the tail begins to fall off .
there is a similar h@xmath0i gap to the se , but in this case there is no corresponding optical feature associated with it . at even larger radii ,
the h@xmath0i is more diffuse ( @xmath7@xmath123 ) and has no optical counterpart down to @xmath13=27 mag arcsec@xmath15 .
an x - ray image obtained with the rosat hri camera ( heckman _ et al . _
1996 ) reveals an extended central source that is elongated along p.a.=135(indicated by arrows in fig .
[ fig : a220mos ] ) . a deep h@xmath8 + [ n@xmath0ii ]
image of arp 220 reveals ionized gas with a bright linear morphology at this same position angle ( heckman , armus & miley 1987 ) .
the optical emission line kinematics are suggestive of a bipolar outflow ( ham90 ) , and the physical properties of the warm and hot gas strongly support the superwind scenario for this emission ( ham90 , heckman _
et al . _
1996 ) . as in ngc 520 and arp 299 ,
the position angle of the putative expanding superwind is in the same direction as the h@xmath0i minima , i.e. nw and se .
figures [ fig : n520mos][fig : slices ] provide evidence for both small- and large - scale differences in the distributions of the tidal gas and stars in these three systems .
the small - scale differences are of the type illustrated in fig .
[ fig : slices ] , whereby the gas column density falls off just as the optical surface brightness increases at various edges of the tidal features . in ngc 520 and arp 220 ,
the large - scale differences are between the outer h@xmath0i rings and disks ( which have no associated starlight ) and the optical tails and plumes ( which have no associated h@xmath0i ) .
although these features are kinematically decoupled at present ( with the gas rings and disks predominantly in rotation and the optical tails and plumes predominantly in expansion ) , it is possible that they had a common origin and have subsequently decoupled and evolved separately . in arp 299 , on the other hand , the h@xmath0i filaments and optical tail have similar morphologies and continuous kinematics and are therefore part of the same kinematic structure . in this system
we believe the bifurcated tidal morphology results from a progenitor with a warped gaseous disk ( [ sec : a299 ] & hy99 ) , and we seek to understand why the inner filament is gas - poor , given that its progenitor was obviously gas - rich . in this section
we investigate a number of possible explanations for these observations .
in particular , we discuss the possible role played by : differences in the initial radial distribution of the gas and stars ( [ sec : radial ] ) , dust obscuration ( [ sec : dust ] ) , kinematic decoupling of the gas due to collisions within the developing tidal tail ( [ sec : coll ] ) , ram pressure stripping of the gas , either by a halo or by a galactic scale wind ( [ sec : rps ] ) , and photoionization of the gas , either by the starburst or by local sources ( [ sec : ioniz ] ) . in interacting systems
the h@xmath0i is often more widely distributed than the optical light ( see , e.g. the h@xmath0i map of the m81 system by yun _ et al . _ 1994
; see also van der hulst 1979 ; appleton , davies & stephenson 1981 ) .
these gas - rich extensions frequently have no associated starlight down to very faint limits ( e.g. simkin _ et al . _ 1986
; hvg96 ) .
a natural explanation is that such features arise from the h@xmath0i - rich but optically faint outer radii of the progenitor disks .
the relatively short lifetimes of luminous stars and the larger velocity dispersions of less luminous stars , especially with respect to the gas , will further dilute the luminous content of this material , and the h@xmath0i - to - light ratio of the resulting tidal features will increase with time ( hibbard _ et al . _
gaseous tidal extensions with very little detectable starlight would seem to be the natural consequence .
the outer h@xmath0i rings in ngc 520 and arp 220 and the gas - rich outer filament in arp 299 are all likely to have arisen in this manner
. however , gas - rich outer disks can not give rise to gas - poor optical structures , such as the optical plume in ngc 520 , the optical tail in arp 220 , or the inner filament in arp 299 .
since these features presumably arise from optically brighter regions of the progenitor disks ( regions which are characterized by h@xmath0i column densities higher than that of the outer disks ) one would have expected a priori that these features should also be gas - rich .
it is possible that the disks which gave rise to the plumes in ngc 520 and arp 220 were gas - poor at all radii .
however , this would not account for the discontinuities in the outer gaseous features that project near these optical features ( i.e. , ne gap in ngc 520 and nw gap in arp 220 ) .
we therefore seek other explanations for these structures .
the correspondence between rising gas column density and falling optical surface brightness ( fig .
[ fig : slices ] ) suggests that dust associated with the cold gas may attenuate the optical light . to address this possibility ,
we calculate the expected extinction in the @xmath23band for a given column density of h@xmath0i .
we adopt the milky way dust - to - gas ratio determined by bohlin , savage , & drake ( 1978 ; @xmath24= 4.8 mag@xmath19 ) , which is supported by direct imaging of the cold dust in the outer regions of eight disk galaxies ( alton _ et al . _
this is combined with the galactic extinction law of odonnell ( 1994 ; @xmath25 , from table 6 of schlegel _
et al._1998 ) to yield an expected extinction in the @xmath5-band of @xmath26 mag . from fig .
[ fig : slices ] , the peak h@xmath0i column densities on either side of the optical features are @xmath123 .
the predicted extinction is therefore of order 0.2 mag in the @xmath5-band . from fig .
[ fig : slices ] we see that the mean light level drops by about 1.0 mag arcsec@xmath15 for arp 299 ( from 26.5 mag arcsec@xmath15 to below 27.5 mag arcsec@xmath15 ) , about 1.5 mag arcsec@xmath15 for ngc 520 ( from 25 mag arcsec@xmath15 to below 26.5 mag arcsec@xmath15 ) , and by about 2.5 mag arcsec@xmath15 for arp 220 ( from 23.5 mag arcsec@xmath15 to below 26 mag arcsec@xmath15 ) along the extracted slices . to produce this amount of extinction , the tidal gas would have to have a dust - to - gas ratio that is ten times that in the milky way .
the above analysis assumes that the measured neutral gas column density represents the total gas column density .
however , the sharp drop in h@xmath0i column density observed in many tidal features ( hvg96 , hibbard & yun in preparation ) suggests that the tidal gas may be highly ionized by the intergalactic uv field ( see also references in [ sec : ioniz ] ) .
since large dust grains should survive in the presence of this ionizing radiation , the opacity per atom of neutral hydrogen ( @xmath27 ) should increase in regions of increasing ionization fraction .
observations of ngc 5018 ( hilker & kissler - patig 1996 ) , in which blue globular clusters are absent in a region underlying an associated h@xmath0i tidal stream , may support a high @xmath27 ratio for tidal gas .
nevertheless , the lack of obvious reddening of the @xmath28 colors along the slices in arp 299 and ngc 520 ( hibbard 1995 ; hy99 ) argues against a much higher extinction in these regions .
we conclude that extinction might be important for shaping the morphology of the faintest optical features ( e.g. , the `` hook '' and the end of the optical tail of arp 299 , fig . [ fig : a299mos ] , which has @xmath13 near the detection limit of 28 mag arcsec@xmath15 ) , but is insufficient to greatly affect the overall tidal morphology .
however , an anomalously high tidal dust - to - gas ratio remains a possibility .
this question could be resolved by the direct detection of cold dust in tidal tails with sub - millimeter imaging . during the tail formation process , the leading - edge of the tail
is decelerated with respect to the center of mass of the progenitors , while the trailing - edge is accelerated , and the two edges move towards each other ( see toomre & toomre 1972 , fig .
eventually , the two edges appear to cross , forming a caustic ( wallin 1990 ; struck - marcell 1990 ) . in most cases ,
the caustics are simply due to projection effects . only for low - inclination encounters
will these crossings correspond to physical density enhancements , and numerical experiments suggest that in these cases the density will increase by factors of a few ( wallin 1990 ) .
it has been suggested that collisions experienced by the crossing tidal streams in such low - inclination encounters may lead to a separation between the dissipational ( gas ) and non - dissipational ( stellar ) tidal components ( wevers 1984 ; smith _ et al . _
1997 ) .
the present data do not allow us to directly address this question , since the kinematic decoupling presumably took place long ago . however , several arguments lead us to suspect that this collisional process is not important in tidal tails : ( 1 ) large scale decoupling between the stellar and gaseous tidal morphologies is not seen in many systems known to have experienced low - inclination encounters ( e.g. ngc 4038/9 , `` the antennae '' , hibbard _
et al . _ in preparation ; ngc 7252 `` atoms for peace '' hibbard
_ et al . _
1994 ; ngc 4676 `` the mice '' hvg96 ) ; ( 2 ) the broad plume - like morphologies of the optical features in arp 220 and ngc 520 suggest rather inclined encounters ( [ sec : n520 ] , [ sec : a220 ] ) , in which case wide - spread collisions are not expected ; and ( 3 ) the parallel filaments in the arp 299 tail have identical kinematics , whereas one would expect kinematic differences between the stripped and unstripped material .
therefore while gaseous collisions and dissipation might result in differences between gas and stars during tidal development ( particularly along tidal bridges , where the gas streamlines are converging ; e.g. struck 1997 ; ngc 7714/5 smith _
et al . _ 1997
; arp 295 hvg96 ) , we believe that they are not likely to lead to a wide - spread decoupling in the outer regions .
if the tidal features pass through a diffuse warm or hot medium , or if such a medium passes through the tidal features , it is possible that the tidal gas exchanges energy and momentum with this medium due to collisions .
such effects have been proposed to explain the stripping of the cool interstellar medium from spiral galaxies as they move through the hot igm in clusters ( gunn & gott 1972 ) , and is referred to as ram pressure stripping ( rps ) .
tidal features should be relatively easily stripped , as they lack the natural restoring forces present in disk galaxies , except possibly at a small number of self - gravitating regions . in this case , the momentum imparted due to ram pressure is simply added to or subtracted from the momentum of the gaseous tidal features , and a separation of stellar and gaseous components might be expected . in the next two subsections , we investigate two possible sources for ram pressure : an extended halo associated with the progenitors ( [ sec : rpshalo ] ) ; and an expanding starburst driven superwind ( [ sec : rpswind ] ) . our own galaxy is known to have an extended halo of hot gas ( pietz _ et al._1998 ) .
the existence of similar halos around external galaxies has been inferred from observations of absorption line systems around bright galaxies ( e.g. lanzetta _ et al . _
these halos may have sufficient density to strip any low column density gas moving through them .
several investigators have suggested that such stripping is responsible for removing gas from the magellanic clouds as they orbit through the galaxy s halo , producing the purely gaseous magellanic stream ( e.g. meurer , bicknell & gingold 1985 ; sofue 1994 ; moore & davis 1994 ) .
sofue & wakamatsu ( 1993 ) and sofue ( 1994 ) specifically stress that stripping by galaxy halos should also play an important role in the evolution of h@xmath0i tidal tails .
the tidal features in each of our systems have h@xmath0i column densities and velocities similar to those assumed in the numerical models of sofue ( 1994 ) and moore & davis ( 1994 ) , which resulted in rather extreme stripping of the h@xmath0i clouds .
although this may seem to provide an immediate explanation for our observations of gas / star displacements , we point out that these column densities and velocities are typical of all of the tails thus far imaged in h@xmath0i , the great majority of which do not show the extreme displacements we describe here .
there is no reason to believe that the halo properties of ngc 520 , arp 299 and arp 220 are any different from , or that the encounters were any more violent than similar mergers which do not exhibit such dramatic displacements ( e.g. , ngc 4038/9 , ngc 7252 , ngc 4676 hvg96 ; ngc 3628 dahlem
_ et al . _ 1996
; ngc 2623 , ngc 1614 , mrk 273 hibbard & yun in preparation ; ngc 3256 english _ et al . _
in fact , in light of the stripping simulations mentioned above , one wonders why such displacements are not more common . a possible solution to this puzzle
is suggested by the results of numerical simulations of major mergers . in these simulations ,
the material distributed throughout the halos of the progenitors is tidally distended along with the tails , forming a broad sheath around them ( see e.g. the video accompanying barnes 1992 ) .
this sheath has similar kinematics as the colder tail material , resulting in much lower relative velocities than if the tail was moving through a static halo , thereby greatly reducing any relative ram pressure force . in summary ,
while halo stripping might be effective for discrete systems moving through a static halo ( such as the lmc / smc through the halo of the galaxy , or disks through a hot cluster igm ) , the lack of widespread h@xmath0i / optical decoupling in mergers suggests that it is not very effective for removing gas from tidal tails , and it does not appear to be a suitable explanation of the present observations .
the three systems under discussion host massive nuclear starbursts with associated powerful outflows or superwinds " . optical emission lines and/or x - ray emission
reveal that the observed outflows extend for tens of kpc from the nuclear regions .
theoretical calculations suggest that the observed gas plumes represent just the hottest , densest regions of a much more extensive , lower density medium ( wang 1995 ) . in each of these three systems ,
the most extreme gaps in the h@xmath0i distribution appear along the inferred direction of the expanding hot superwind ( figs .
[ fig : n520mos]d , [ fig : a299mos]d , [ fig : a220mos]d ) .
a very similar anti - correlation between an out - flowing wind and tidally disrupted h@xmath0i has been observed in the m82 system ( yun _ et al . _
1993 ; strickland _ et al . _ 1997
; shopbell & bland - hawthorn 1998 ) , the ngc 4631 system ( weliachew _ et al . _ 1978
; donahue _
et al . _ 1995
; wang _ et al . _
1995 ; vogler & pietsch 1996 ) , and possibly the ngc 3073/9 system ( irwin _ et al . _ 1987
; filippenko & sargent 1992 ) .
it has been suggested that this anti - correlation is due to an interaction between the blown - out gas of the superwind and the cold gaseous tidal debris , either as the wind expands outward into the debris , or as the tidal debris passes through the wind ( chevalier & clegg 1985 ; heckman , armus & miley 1987 , and references above ) .
figure [ fig : cartoon ] presents the suggested geometry for the case of arp 299 .
this figure is constructed from our preliminary efforts to model the northern tail of arp 299 using n - body simulations similar to those presented in hibbard & mihos ( 1995 ; i.e. no hydrodynamical effects are included ) .
we found that we could not match the morphology and kinematics of both filaments simultaneously , but could match either one separately .
[ fig : cartoon ] presents the results of combining these two solutions . in this sense
this figure is _ not a self - consistent fit to the data _ , but simply a cartoon which illustrates the proposed relative placement of the tidal tail and the expanding wind . in this figure ,
the wind opening angle is illustrated by the cone , the gas - rich regions of the tails are represented by dark and light grey circles , and the gas - poor regions of the tails are represented by white circles .
the figure illustrates how the restricted opening angle of such a wind ( or of an ionization cone , cf .
[ sec : ionizsb ] ) may intersect only a portion of the ribbon - like tail .
here we estimate whether ram pressure stripping by the nuclear superwind can exert sufficient pressure on gas at the large distances typical of tidal tails .
we use equation ( 5 ) from heckman , lehnert & armus ( 1993 ; see also chevalier & clegg 1985 ) to calculate the expected ram pressure ( @xmath29 ) of the superwind far from the starburst as a function of its bolometric luminosity ( @xmath30 ) : @xmath31 this equation has been shown to fit the pressure profile derived from x - ray and optical emission line data of arp 299 ( heckman _ et al . _ 1999 ) .
it should provide a lower limit to the ram pressure , since it assumes that the wind expands spherically , while observations suggest that the winds are limited in solid angle .
this pressure can be compared with the pressure of the ambient medium in the tidal tail ( @xmath32 ) , given by the energy per unit volume : @xmath33 where @xmath34 is a constant , @xmath35 is the mass density of the gas , and @xmath36 is the velocity dispersion of the gas . for an equation of state of the form @xmath37
, the constant @xmath34 is equal to @xmath38 , and @xmath36 corresponds to gas sound speed , while for a self - gravitating cloud , @xmath39 and @xmath36 is the one - dimensional velocity dispersion of the cloud . in both cases , we assume that the observed line - of - sight velocity dispersion of the h@xmath0i is a suitable measure of @xmath36 .
the mass density of the tidal gas is given by @xmath40 , where the numerical constant accounts for the presence of he , @xmath41 is the mass of a hydrogen atom , and @xmath42 is the number density of atomic hydrogen . if we assume the gas is uniformly distributed with a column density @xmath7 along a length @xmath43 , we have @xmath44 , where the fiducial values are typical of tidal features ( hvg96 , hibbard & yun in preparation ) .
we rewrite eqn .
( [ eq : perfgas ] ) in terms of the observables : @xmath45 the maximum radius out to which we expect material to be stripped ( @xmath46 ) is then given by the requirement that @xmath29(r ) @xmath47(r ) at r@xmath48 .
we replace @xmath30 in eqn .
( [ eq : prps ] ) by the ir luminosity ( @xmath49 ) , under the assumption that the ir luminosity arises from reprocessed uv photons from the starburst ( lonsdale , persson & matthews 1984 , joseph & wright 1985 ) .
the very high ir luminosities of these systems ( @xmath50 ) make it likely that this is indeed the case , and we are probably making an error of @xmath17 10% ( e.g. heckman , lehnert & armus 1993 ) . equating eqns .
( [ eq : prps ] ) and ( [ eq : ptidal ] ) we find : @xmath51 in table [ tab : results ] we provide estimates of @xmath46 for the systems considered here . in calculating @xmath46
we made the very conservative assumption that the values of @xmath7 and @xmath52 for the stripped gas are equal to the maximum values found within the tidal tails ( see table [ tab : results ] ) .
the results of these calculations indicate that , in all cases , @xmath46 is larger than the radii of the observed gaps in the tidal h@xmath0i distributions
. therefore , in principle , the wind should be able to strip the gas from any tidal material in its path .
the above derivation assumes that the tidal gas is at rest with respect to the wind .
it can be easily generalized to the case of a wind impacting an expanding tidal feature by reducing the wind ram pressure by a factor @xmath53 .
for ngc 520 and arp 220 , the tidal gas is primarily in rotation ( _ i.e. _ , moves perpendicular to superwind ) , so we expect the gas to feel the full ram pressure given above . for arp 299 , heckman _
et al . _ ( 1999 ) find @xmath54 = 800 km s@xmath19 , and we estimate a maximum @xmath55 = 240 km s@xmath19 ( hy99 ) .
therefore , the ram pressure could be reduced by 50% , reducing @xmath46 by 70% from that listed in table [ tab : results ] , _ i.e. _ @xmath56 kpc for arp 299 .
this is still large enough to reach to the region of the n clump in fig .
[ fig : a299mos ] .
the regions of high h@xmath0ivelocity dispersion indicated in fig .
[ fig : a299mos]c may be due to the influence of such a wind .
we note that these regions occur on the side of the h@xmath0i features that face the starburst region
. however , no such kinematic signatures are visible in the gas near the wind axis in ngc 520 or arp 220 .
the lack of gaseous / stellar displacements in the tidal tails of many superwind systems might seem to provide a strong argument against the scenario outlined above . however , there are two conditions needed to produce wind - displaced tidal features : the starburst must be of sufficient energy and duration to achieve blowout " , and the tidal h@xmath0i must intersect the path of the expanding wind material . the second condition is not met for the blowout systems ngc 4676 , ngc 3628 , ngc 2623 , ngc 1614 , and ngc 3256 ( references given in
[ sec : rpshalo ] ) . in these systems
the tidal tails appear to lie at large angles with respect to the blowout axis , and their tidal tails should not intersect the wind .
both conditions are met for m82 and ngc 4631 , which both show extreme h@xmath0i / optical displacements .
for these two systems , the high - latitude h@xmath0i appears to be accreted from nearby disturbed companions ( m81 and ngc 4656 , respectively ; yun _ et al._1993 , weliachew _
et al . _ 1978
) , while for the three major mergers under study here the h@xmath0i appears to intercept the path of the wind as a result of a highly inclined encounter geometry .
a high inclination encounter geometry is therefore a prerequisite for such displaced morphologies , and the host of the superwind should be the disk with a retrograde or polar spin geometry .
nevertheless , ram pressure stripping can not provide a complete explanation of the observations .
since the stars are unaffected by the wind , it would be an unusual coincidence for the edges of the optical plumes to correspond with the edge of the cold gas which is presently being ablated .
nor does it seem likely that the wind could be sufficiently collimated to bore " into the northern h@xmath0i clump in arp 299 just where the optical tail appears projected upon it . therefore a second process is still needed to explain the small - scale anti - correlations . in conclusion , the expanding wind _ should _ affect any tidal h@xmath0i in its path ; however this effect alone can not explain all the details of the observations .
disk galaxies are known to exhibit a precipitous drop in neutral hydrogen column density beyond column densities of a few times @xmath57 cm@xmath15 ( corbelli , schneider & salpeter 1989 ; van gorkom 1993 ; hoffman _ et al . _
this drop has been attributed to a rapid change in the ionization fraction of the gas due to influence of the intergalactic uv field ( maloney 1993 , corbelli & salpeter 1993 , dove & shull 1994 ; see also felten & bergeron 1969 , hill 1974 ) , rather than a change in the total column density of h. tidal tails are assembled from the outermost regions of disk galaxies .
since this material is redistributed over a much larger area than it formerly occupied , its surface brightness must decrease accordingly .
therefore , if the progenitors were typical spirals , with h@xmath0i disks extending to column densities of a few times @xmath57 cm@xmath15 , then the resulting tidal tail _ must _ have gas at much lower column densities .
however , tidal tails exhibit a similar edge in their column density distribution _ at a similar column density _ ( hvg96 ; hibbard & yun in preparation ) .
this is one of the most compelling pieces of evidence for an abrupt change in the phase of the gas at low h@xmath0i column density .
the outer tails mapped in h@xmath0i should therefore be the proverbial `` tip of the iceberg '' of a lower column density , mostly ionized medium . with the tidal gas in this very diffuse state , fluctuations in the incident ionizing flux
might be expected to produce accompanying fluctuations in the neutral gas fraction . given these considerations , we examine the possibility that the total hydrogen column density does not change at the regions illustrated in figs .
[ fig : n520mos][fig : slices ] , but that the neutral fraction does , _
i.e. _ that the gas in the regions under study has a higher ionization fraction than adjacent regions . the intergalactic uv field should be isotropic , and would not selectively ionize certain regions of the tails . here
we examine the possibility of two non - isotropic sources of ionizing flux : ( 1 ) leakage of uv flux from the circumnuclear starburst ; ( 2 ) ionization by late b stars and white dwarfs associated with the evolved stellar tidal population .
our procedure is to compare the expected ionizing flux density shortward of 912 to the expected surface recombination rate of the gas .
we assume that in the area of interest the gas is at a temperature of @xmath59 k , for which a case - b recombination coefficient of @xmath60 is appropriate ( spitzer 1956 ) .
we assume that the hydrogen is almost completely ionized , so @xmath61 .
we further assume that the density of ionized gas is the same as the density of the neutral gas in the adjacent regions , @xmath62 , where @xmath42 is calculated as above ( @xmath63 , [ sec : rpswind ] ) .
the detailed ionization state will depend sensitively on the clumpiness of the gas , but a full treatment of this problem is beyond the scope of this paper . here
we wish to investigate if these processes are in principle able to create effects similar to those observed . here
we consider the case that the superwind does not affect the tidal gas by a direct interaction , but influences it by providing a direct path from the tidal regions to the starburst , free of dust and dense gas ( see also fig . [
fig : cartoon ] ) . through these holes ,
ionizing photons from the young hot stars stream out of the nuclear regions and are quickly absorbed by the first neutral atoms they encounter . following felton & bergeron ( 1969 ; see also mahoney 1993 ) , we solve the equation : @xmath64 the right hand side of this equation represents the total ionizing radiation escaping the starburst region along a direction that has been cleared of obscuring material by the superwind .
we express this in terms of the total ionizing flux of a completely unobscured starburst of a given bolometric luminosity @xmath30 by introducing the factor @xmath65 to account for the fact that only a fraction of the photons emitted into a solid angle @xmath66 find their way out of the starburst region , the total fraction of ionized photons emerging from a starburst , and @xmath65 , the fraction emerging along a particular sightline .
@xmath67 is the total angle averaged fraction , i.e. @xmath67 is the integral of @xmath68 over all solid angles , while @xmath65 is the integral over a solid angle cleared by the wind .
most studies in the literature quote values for @xmath67 . ] .
the expected ionizing flux for a starburst of a given bolometric luminosity @xmath30 is calculated from the population synthesis models of bruzual & charlot ( 1993 ; 1995 ) , assuming continuous star formation with a duration longer than 10 myr ( long enough for the burst to achieve blowout ) , a salpeter imf with @xmath69 and @xmath70 , and solar metallicity .
this yields @xmath71 . again making the standard assumption that most of the starburst luminosity is emitted in the far infrared ( _ i.e._@xmath72 , cf .
[ sec : rpswind ] ) , we rearrange eqn .
( [ eq : photosb ] ) to solve for the radius , @xmath73 , out to which the starburst is expected to ionize a given column density of h@xmath0iof thickness @xmath43 : @xmath74 resulting values for @xmath73 are listed in table [ tab : results ] . for this computation
, we have adopted a value of 10% for @xmath65 .
this is equal to the total fraction of ionizing photons , @xmath67 , escaping from a normal disk galaxy as calculated by dove , shull & ferrara ( 1999 ) .
even higher values of @xmath67 are expected in starburst systems ( dove _ et al . _
since we stipulate that a higher fraction of ionizing photons escape along sightlines above the blowout regions than are emitted along other directions , it follows that @xmath75 , and as a result the values of @xmath73 calculated in table [ tab : results ] should be conservative estimates .
table [ tab : results ] shows that under these simplified conditions , @xmath73 is of the order of , or larger than , the tidal radii of interest .
we therefore conclude that the starburst seems quite capable of ionizing tidal h@xmath0i , if indeed there is an unobstructed path from the starburst to the tidal regions .
this might explain the lack of h@xmath0ialong the wind axis in ngc 520 and arp 220 , and the absence of h@xmath0ialong the optical tidal tail in arp 299 .
this process is especially attractive since it can potentially explain the lack of h@xmath0i at the bases of otherwise gas - rich tidal tails in ngc 7252 ( hibbard _ et al . _ 1994 ) , arp 105 ( duc _ et al . _ 1997 ) , and ngc 4039 ( hibbard , van der hulst & barnes in preparation ) .
these systems do _ not _ show evidence for expanding superwinds , which rules out the possibility that rps is playing a role . and each of these systems possesses a level of star formation that , according to eqn .
[ eq : rion ] , is capable of ionizing gas out to the necessary radii .
however , photoionization by the central starburst does not seem capable of explaining all of the observations . as with the wind hypothesis above
, it would be an unusual coincidence for the edges of the optical plumes to correspond with the edge of ionization cone .
therefore a second process is still needed to explain the small - scale anti - correlations .
the fact that the h@xmath0i column density falls off just as the optical surface brightness increases at the edges of various tidal features ( fig .
[ fig : slices ] ) leads us to suspect that there may be local sources of ionization within the stellar features themselves . for ngc 520 and arp 220
, we believe the outer h@xmath0i is in a disk structure which is intersected by the tidal stellar plumes and we wish to investigate whether ionization by evolved sources within the stellar plumes , such as late b stars and white dwarfs , could be responsible for decreasing the neutral fraction of the diffuse outer h@xmath0i . for arp 299
the geometry is more complicated , and we refer the reader to fig . [
fig : cartoon ] . here
we suggest that part of the purely gaseous tidal filament ( the light grey filament in fig . [
fig : cartoon ] ) is ionized by evolved sources near the end of the stellar tail ( the dark grey filament in fig .
[ fig : cartoon ] , especially those regions nearest the gas - rich filament in the right hand panel of this figure ) . as in the previous section ,
we balance the surface recombination rate with the expected ionizing flux density ( eqn .
[ eq : photosb ] ) . in this case , we calculate the ionizing flux density for an evolved population of stars of a given @xmath5-band luminosity density ( @xmath76 , in @xmath77 ) . in order to approximate the stellar populations in the tidal tails , we assume that the tails arise from the outer edges of an sbc progenitor , and that star formation ceased shortly after the tails were launched .
we again use the models of bruzual & charlot ( 1993 , 1995 ) for a salpeter imf over the mass range 0.1125 @xmath78 , and adopt an exponentially decreasing sfr with a time constant of 4 gyr ( typical of an sbc galaxy , bruzual & charlot 1993 ) , which is truncated after 10 gyr and allowed to age another 500 myr .
this simulates the situation in which star formation within the disk is extinguished as the tail forms , and the ejected stellar population passively fades thereafter .
while tidal tails frequently exhibit _ in - situ _ star formation ( e.g. schweizer 1978 ; mirabel , lutz & maza 1991 ) , it is usually not widespread . under these assumptions ,
a population with a projected @xmath5-band surface brightness of 1 @xmath79 pc@xmath15 should produce an ionizing flux of @xmath80
. therefore eqn .
( [ eq : photosb ] ) becomes @xmath81 , which can be rewritten as : @xmath82 noting that 1 @xmath77 corresponds to @xmath13=25.9 mag arcsec@xmath15 , we rewrite this as a condition on the surface brightness of the tidal features : @xmath83 \label{eq : mur}\end{aligned}\ ] ] referring to fig .
[ fig : slices ] , we see that only the northern plume of arp 220 is bright enough to ionize nearby tidal h@xmath0iat the appropriate column densities .
neither the optical plume in the ngc 520 system nor the northern tail in the arp 299 system appears bright enough to ionize the necessary columns of hydrogen unless the tidal features are unreasonably thick ( @xmath1260 kpc ) .
however , since we have no other explanation for the small scale h@xmath0i / optical differences illustrated in fig .
[ fig : slices ] , we are hesitant to abandon this explanation too quickly . a possible solution is to invoke continued star formation even after the tails are ejected . for instance , if we do not truncate the star formation rate after 10 gyr , instead allowing the star formation rate to continue its exponential decline as the tail expands , then the ionizing flux per 1 @xmath79 pc@xmath15 is 70 times higher than the value of @xmath80 used above .
this would lower the fiducial surface brightness in eqn .
( [ eq : mur ] ) from 23.0 mag arcsec@xmath15 to 27.5 mag arcsec@xmath15 , in which case the faint tidal features in ngc 520 ( @xmath84 25 mag arcsec@xmath15 ) and arp 299 ( @xmath84 26.5 mag arcsec@xmath15 ) could indeed ionize the necessary column densities of adjoining h@xmath0i .
the observed broad - band colors of the tidal tails are not of sufficient quality to discriminate between these two star formation histories , since the expected color differences are only of order @xmath28=0.1 mag . however , whether or not the gas is more highly ionized in the regions of interest can be addressed observationally . the expected emission measure ( @xmath85 )
can be parameterized as : @xmath86 since emission measures of order 0.2 @xmath87 pc have been detected with modern ccd detectors ( e.g. donahue _
et al . _ 1995
; hoopes , walterbos & rand 1999 ) , there is some hope of being able to observationally determine if regions of the tidal tails are significantly ionized .
if the gas has a clumpy distribution , then there should be some high density peaks which might be sufficiently bright to yield reliable emission line ratios .
such ratios would allow one to determine the nature of the ionizing source , e.g. photoionization vs. shocks .
therefore , while we can not assert unequivocally that photoionization plays a role in shaping the outer tidal morphologies , it is possible to test this hypothesis with future observations .
the hypothesis that ionizing flux from a stellar tidal feature may ionize gas in a nearby gaseous tidal feature is not necessarily at odds with the observations that many stellar tails are gas - rich .
this is because tails with cospatial gas and stars arise from regions originally located within the stellar disk of the progenitors , while the optical faint gas - rich tidal features likely arise from regions beyond the optical disk ( [ sec : radial ] ) . in normal disk galaxies , the h@xmath0iwithin the optical disk
is dominated by a cooler component with a smaller scale height and velocity dispersion , while the h@xmath0i beyond the optical disk is warmer and more diffuse , with a larger scale height and velocity dispersion ( braun 1995 , 1997 ) . as a result , @xmath43 should be considerably larger for purely gaseous tidal features than for optically bright tidal features .
in this paper we have described differences between gaseous and stellar tidal features .
there are large - scale differences , such as extensive purely gaseous tidal features ( the outer disks in ngc 520 and arp 220 and the outer filament in arp 299 ) and largely gas - poor optical features ( tidal plumes in ngc 520 and arp 220 and the inner filament in arp 299 ) . and there are smaller - scale differences : the anti - correlation between the edges of gaseous and optical features depicted in fig .
[ fig : slices ] .
a similar anti - correlation is observed between h@xmath0i and optical shells in shell galaxies ( schiminovich _ et al . _ 1994a , b , 1999 )
, many of which are believe to be more evolved merger remnants .
we have examined a number of possible explanations for these observations , including dust obscuration , differences in the original distribution of gas and starlight in the progenitor disks , gas cloud collisions within the developing tails , ram pressure stripping due to an extensive hot halo or an expanding superwind , and photoionization by either the central starburst or evolved sources in the tidal tails themselves . however , no one model easily and completely explains the observations , and it is conceivable that all explanations are playing a role at some level . the most likely explanation for the lack of starlight associated with the outer tidal h@xmath0iis that such features arise from the h@xmath0i - rich but optically faint outer radii of the progenitor disks . the relatively short lifetimes of luminous stars and the large velocity dispersions of less luminous stars , especially with respect to the gas , will further dilute the luminous content of this material , and the h@xmath0i - to - light ratio of the resulting tidal features will increase with time ( hibbard _ et al . _
gaseous tidal extensions with very little detectable starlight would seem to be the natural consequence .
the outer h@xmath0i rings in ngc 520 and arp 220 and the gas - rich outer filament in arp 299 are all likely to arise from these gas - rich regions of their progenitor disks . for the gas - poor tidal features we suggest that the starburst has played an important role in shaping the gaseous morphology , either by sweeping the features clear of gas via a high - pressure expanding superwind , or by excavating a clear sightline towards the starburst and allowing ionizing photons to penetrate the tidal regions .
the primary supporting evidence for this conclusion is rather circumstantial : the five galaxies with the most striking h@xmath0i / optical displacements ( the three systems currently under study here , and the h@xmath0i accreting starburst systems m82 and ngc 4631 ) host massive nuclear starbursts with associated powerful outflows or superwinds aligned with the direction of the most extreme h@xmath0i / optical displacements .
ngc 520 , arp 299 , and arp 220 each experienced prograde / polar or prograde / retrograde encounters .
this relative geometry may be a pre - requisite for the morphological differences reported here .
retrograde and polar encounters do not raise extensive tidal tails ( e.g. barnes 1988 ) , leaving large gaseous disks in the inner regions .
these disks should help collimate and `` mass - load '' the superwind ( heckman , lehnert & armus 1993 ; suchkov _ et al . _
1996 ) , which in turn leads to denser and longer - lived winds .
simultaneously , the combination of opposite spin geometries provides the opportunity for the tidal tail from the prograde system to rise above the starburst region in the polar or retrograde system , where it may intersect the escaping superwind or uv radiation .
if this suggestion is correct , only systems hosting a galactic superwind and experiencing a high - inclination encounter geometry should exhibit such extreme differences between their h@xmath0i and optical tidal morphologies . the observations do not allow us to discriminate between either the rps or the photoionization models : simple calculations suggest that either is capable of affecting the diffuse outer gas if the geometry is right
. there might be some evidence for the effects of an impinging wind on the outer material in arp 299 from the increased velocity dispersion at several points ( hy99 ) ; however ngc 520 and arp 220 show no such signatures .
photoionization is an attractive solution , as it offers a means of explaining the lack of tidal h@xmath0i found at the base of otherwise gas - rich tidal tails in mergers which show no evidence of a superwind ( e.g. ngc 7252 , arp 105 , ngc 4039 ; see
[ sec : ionizsb ] ) .
since any ionized hydrogen will emit recombination lines , both explanations can be checked observationally .
the expected emission measure is given by eqn .
( [ eq : em ] ) , which predicts detectable features at the column densities of interest .
the morphology of the ionized gas should reveal the nature of the ionizing source : photoionized gas should be smoothly distributed , while gas excited by rps should be concentrated in dense shocked regions on the edges of the h@xmath0i that are being compressed by the superwind , _
i.e. _ on the edges nearest the wind axis in figs .
[ fig : n520mos]d , [ fig : a299mos]d & [ fig : a220mos]d .
if the gas is clumpy , there may be regions bright enough to allow line ratios to be measured , which should further aid in discriminating between photoionization or shock excitation .
only two scenarios are offered to explain the small - scale anti - correlations : dust obscuration and photoionization due to evolved sources in the optical tails .
dust obscuration likely affects the apparent tidal morphologies at the lowest light levels , but we suspect that the dust content is too low to significantly obscure the brighter tidal features .
however , if the tidal tails are highly ionized , with the neutral gas representing only a small fraction of the total hydrogen column density , it is possible that we are grossly underestimating the expected amount of absorption .
this question can be investigated directly with submm imaging of the cold dust in tidal tails .
the other possibility is that the uv flux from evolved sources in the optical tails is responsible for ionizing nearby diffuse outer h@xmath0i .
a simple calculation suggest that the tidal tail in arp 220 is bright enough to ionize nearby h@xmath0i , but the expected ionization flux from the optical tails in ngc 520 and arp 299 is too low to explain the observed differences , unless significant star formation continued within these features after their tidal ejection .
if this is indeed the case , then the regions where the neutral gas column density drops rapidly ( see fig.[fig : slices ] ) should contain ionized gas which would emit recombination radiation .
the expected levels of emission should be observable with deep imaging techniques ( see above ) .
this situation requires that the gas and stellar features are physically close , and not just close in projection , which can tested with detailed numerical simulations .
we would like to thank lee armus tim heckman for sharing of unpublished results , and rhodri evans , jacqueline van gorkom , dave schiminovich , and josh barnes for useful discussions .
we thank the referee , chris mihos , for a thorough and useful report .
aaronson , m. huchra , j. , mould , j. , schechter , p. l. & tully , r. b. 1982 , apj , 258 , 64 alton , p. b. , trewhella , m. , davies , j. i. , evans , r. , bianchi , s. , gear , w. , thronson , h. , valentijn , e. & witt , a. 1998 , a&a , 335 , 807 appleton , p. n. , charmandaris , v. & struck , c. 1996 , apj , 468 , 532 appleton , p. n. , davies , r. d. & stephenson , r. j. 1981 , mnras , 195 , 327 appleton , p. n. , ghigo , f. d , van gorkom , j. h. , schombert , j. m. , & struck - marcell , c. 1987 , nature , 330 , 140 armus , l. , heckman , t. m , & miley , g. k. 1990 , apj , 364 , 471 barnes , j. e. 1988 , apj , 331 , 699 barnes , j. e. 1992 , apj , 393 , 484 barnes , j. e. & hernquist , l. 1991 , apj , 370 , l65 barnes , j. e. & hernquist , l. 1996 , apj , 471 , 115 bohlin , r. c. , savage , b. d. & drake , j. f. 1978 , apj , 224 , 132 braun , r. 1995 , a&as , 114 , 409 braun , r. 1997 , apj , 484 , 637 bruzual , a. g. & charlot , s. 1993 , apj , 405 , 538 bruzual , a. g. & charlot , s. 1995 , personal communication chevalier , r. a. & clegg a. w. 1985 , nature , 317 , 44 combes , f. 1978 , a&a , 65 , 47 combes , f. , dupraz , c. , casoli , f. , & pagani , l. 1988 , a&a ( lett . ) , 203 , 9 condon , j. j. , helou , g. , sanders , d. b. , & soifer , b. t. 1990 , apjs , 73 , 359 corbelli , e. & salpeter , e. e. 1993 , apj , 419 , 104 corbelli , e. , schneider , s. e. & salpeter , e. e. 1989 , aj , 97 , 390 dahlem , m. , heckman , t. m. , fabbiano , g. , lehnert , m. d. , & gilmore , d. 1996 , apj , 461 , 724 donahue , m. , aldering , g. , & stocke , j. t. 1995 , apj ( lett . ) , 450 , l45 dove , j. b. & shull , j. m. 1994 , apj , 423 , 196 dove , j. b. & shull , j. m. & ferrara 1999 , apj , in press ( astro - ph/9903331 ) downes , d. & solomon , p. m. 1998 , apj , 507 , 615 duc , p. -a . ,
brinks , e. , wink , j. e. , & mirabel , i. f. 1997 , a&a , 326 , 537 english , j _
et al . _ 1999 ,
aj , submitted felton , j. e. & bergeron , j. 1969 , apjl , 4 , 155 filippenko , a. v. & sargent , w. l. w. 1992 , aj , 103 , 28 gardiner , l. t. & noguchi , m. 1996 , mnras , 278 , 191 gunn , j. e. & gott , j. r. 1972 , apj , 176 , 1 heckman , t. m. , armus , l. , miley , g. k. 1987 , aj , 93 , 276 heckman , t. m. , armus , l. , miley , g. k. 1990 , apjs , 74 , 833 * ( ham90 ) * heckman , t. m. , armus , l. , weaver , k. a. & wang , j. 1999 , apj , 517 , 130 heckman , t. m. , dahlem , m. , eales , s. a. , fabbiano , g. , & weaver , k. 1996 , apj , 457 , 616 heckman , t. m. , lehnert , m. & armus , l. 1993 , in `` the evolution of galaxies and their environment '' , edited by h. a. thronson and j. m. shull ( kluwer , dordrecht ) , p. 455
hibbard , j. e. 1995 , ph .
d. thesis , columbia university hibbard , j. e. & van gorkom , j. h. 1996 , aj , 111 , 655 * ( hvg96 ) * hibbard , j. e. , guhathakurta , p. , van gorkom , j. h. , & schweizer , f. 1994 , aj , 107 , 67 hibbard , j. e. & mihos , j. c. 1995 , aj , 110 , 140 hibbard , j. e. & yun , m. s. 1996 , in `` cold gas at high redshift '' , edited by m. bremer , h. rottgering , p. van der werf , and c. l. carilli ( kluwer , dordrecht ) , p. 47
hibbard , j. e. & yun , m. s. 1999 , aj , 118 , 162 * ( hy99 ) * hilker , m. & kissler - patig , m. 1996 , a&a , 314 , 357 hill , j. k. 1974 , a&a , 34 , 431 hoffman , l. g. , lu , n. y. , salpeter , e. e. , farhat , b. , lamphier , c. & roos , t. 1993 , aj , 106 , 39 hoopes , c. , walterbos , r. & rand , r. 1999 , apj , 522 , 669 van der hulst , j. m. 1979 , a&a , 75 , 97 irwin , j. a. , seaquist , e. r. , taylor , a. r. & duric , n. 1987 , apj ( lett . ) , 313 , l91 joseph , r. d. & wright , g. s. 1985 , mnras , 214 , 87 lanzetta , k. m. , bowen , d. v. , tytler , d. & webb , j. k. 1995 , apj , 442 , 538 lehnert , m. d. & heckman , t. m. 1996 , apj , 462 , 651 lonsdale , c. j. , persson , s. e. , & matthews , k. 1984 , apj , 287 , 95 maloney , p. 1993 , apj , 414 , 41 meurer , g. r. , bicknell , g. v. & gingold , r. a. 1985 , pasa , 6 , 195 mihos , j. c. 2000 , in preparation mihos , j. c. & bothun , g. d. 1998 , apj , 500 , 619 mihos , j. c. , bothun , g. d. , richstone , d. o. 1993 , apj , 418 , 82 mihos , j. c. & hernquist , l. 1996 , apj , 464 , 641 mirabel , i. f. , lutz , d. & maza , j. 1991 , a&a , 243 , 367 moore , b. & davis , m. 1994 , mnras , 270 , 209 noguchi , m. 1988 , a&a , 201 , 37 nordgren , t. e. , chengalur , j. n. , salpeter , e. e. , & terzian , y. 1997 , aj , 114 , 77 norman , c. , bowen , d. v. , heckman , t. , blades , c. , & danly , l. 1996 , apj , 472 , 73 odonnell , j. e. 1994 , apj , 422 , 1580 pietz , j. , kerp , j. , kalberla , p.m.w . ,
burton , w.b . ,
hartmann , d. & mebold , u. 1998 , a&a , 332 , 55 sakamoto , k. , scoville , n. z. , yun , m. s. , crosas , m. , genzel , r. , tacconi , l. j. 1999 , apj , 514 , 68 sanders , d. b. , scoville , n. z. , sargent , a. i. & soifer , b. t. 1988 , apj , 324 , l55 sanders , d. b. , scoville , n. z. , & soifer , b. t. 1991 , apj , 370 , 158 schiminovich , d. , van gorkom , j. h. , van der hulst , j. m. , & kasow , s. 1994 , apj , 423 , l101 schiminovich , d. , van gorkom , j. h. , van der hulst , j. m. , & malin , d. f. 1995 , apj , 444 , l77 schiminovich , d. , van gorkom , j. h. , & van der hulst , j. m. 1999 , aj , submitted schlegel , d. j. , finkbeiner , d. p. & davis , m. 1998 , apj , 500 , 525 .
schombert , j. m. , wallin , j. f. & struck - marcell , c. 1990 , aj , 99 , 497 schweizer , f. 1978 , the structure and properties of nearby galaxies " , iau symp .
77 , edited by e. m. berkhuijsen and r. wielebinski ( reidel , dordrecht ) , p. 279 scoville , n. z. , sargent , a. i. , sanders , d. b. , & soifer , b. t. 1991 , apj , 366 , l5 scoville , n. z. , yun , m. s. , & bryant , p. m. 1997 , apj , 484 , 702 shopbell , p. l. , & bland - hawthorn , j. 1998 , apj , 493 , 129 .
simkin s. m. , van gorkom , j. h. , hibbard , j. e. , hong - jun , s. 1987 , science , 235 , 1367 .
smith , b. j. 1994 , aj , 107 , 1695 .
smith , b. j. , struck , c. , & pogge , r. w. 1997 , apj , 483 , 754 .
sofue , y. 1994 , apj , 423 , 207 .
sofue , y. , & wakamatsu , k. 1993 , a&a , 273 , 79 .
soifer , b. t. , helou , g. , lonsdale , c. j. , neugebauer , g. , hacking , p. , houck , j. r. , low , f. j. , rice , w. , & rowan - robinson , m. 1984 , apj , 283 , l1 spitzer , l. 1956 , apj , 124 , 20 .
stanford , s. a. , & balcells , m. 1991 , apj , 370 , 118 stockton , a. 1974a , apj , 187 , 219 stockton , a. 1974b , apj , 190 , l47 stockton , a. & bertola , f. 1980 , apj , 235 , 37 strickland , d. k. , ponman , t. j. , & stevens , i. r. 1997 , a&a , 320 , 378 struck , c. 1997 , apj ( supp . ) , 113 , 269 struck - marcell , c. 1990 , aj , 99 , 71 suchkov , a. a. , berman , v. g. , heckman , t. m. & balsara , d. s. , 1996 , apj , 463 , 528 suchkov , a. a. , balsara , d. s. , heckman , t. m. , leitherer , c. 1994 , apj , 430 , 511 toomre , a. , & toomre , j. 1972 , apj , 178 , 623 van gorkom , j. h. 1993 , in `` the evolution of galaxies and their environment '' , edited by h. a. thronson and j. m. shull ( kluwer , dordrecht ) , p. 345 vogler , a. , & pietsch , w. 1996 , a&a , 311 , 35 wallin , j. f. 1990 , aj , 100 , 1477 wang , b. 1995 , apj , 444 , 590 wang , q. d. , walterbros , r. a. m. , steakley , m. f. , norman , c. a. & braun , r. 1995 , apj , 439 , 176 weil , m. l. , & hernquist , l. 1993 , apj , 405 , 142 weliachew , l. , sancisi , r. , & guelin , m. 1978 , a&a , 65 , 37 wevers , b. h. m. r. , appleton , p. n. , davies , r. d. , & hart , l. 1984 , a&a , 140 , 125 yun , m. s. 1992 , ph.d .
thesis , harvard university yun , m. s. 1997 , in _ `` galaxy interactions at low and high redshfit '' _ , iau symposium no . 186 , eds .
d. sanders & j. barnes , in press yun , m. s. , ho , p. t. p. , & lo , k. y. 1993 , apjl , 411 , l17 yun , m. s. , ho , p. t. p. , & lo , k. y. 1994 , nature , 372 , 530 yun , m. s. & hibbard , j. e. 2000 , in preparation yun , m. s. & hibbard , j. e. 1999 , apj , submitted lrlllc @xmath88 & ( km s@xmath19 ) & 2260 & 3080 & 5400 & + @xmath10 & ( mpc ) & 30 & 48 & 79 &
@xmath89 + @xmath49 & ( l@xmath90 ) & @xmath91 & @xmath92 & @xmath93 & @xmath94 + @xmath95 & ( cm@xmath15 ) & @xmath96 & @xmath97 & @xmath97 & @xmath98 + @xmath99 & ( km s@xmath19 ) & 16 & 20 & 40 & @xmath98 + @xmath100 & ( kpc ) & 25 & 70 & 30 & @xmath101 + + @xmath46 & ( kpc ) & 50 & 140 & 95 & @xmath102 + @xmath103 & ( kpc ) & 25 & 95 & 130 & @xmath104 + [ tab : results ] | as part of several h@xmath0i synthesis mapping studies of merging galaxies , we have mapped the tidal gas in the three disk - disk merger systems arp 157 ( ngc 520 ) , arp 220 , and arp 299 ( ngc 3690 ) .
these systems differ from the majority of the mergers mapped in h@xmath0i , in that their stellar and gaseous tidal features do not coincide .
in particular , they exhibit large stellar tidal features with little if any accompanying neutral gas and large gas - rich tidal features with little if any accompanying starlight
. on smaller scales , there are striking anti - correlations where the gaseous and stellar tidal features appear to cross .
we explore several possible causes for these differences , including dust obscuration , ram pressure stripping , and ionization effects .
no single explanation can account for all of the observed differences .
the fact that each of these systems shows evidence for a starburst driven superwind expanding in the direction of the most striking anti - correlations leads us to suggest that the superwind is primarily responsible for the observed differences , either by sweeping the features clear of gas via ram pressure , or by excavating a clear sightline towards the starburst and allowing uv photons to ionize regions of the tails . if this suggestion is correct , only systems hosting a galactic superwind and experiencing a high - inclination encounter geometry ( such that tidal gas is lifted high above the starburst regions ) should exhibit such extreme differences between their h@xmath0i and optical tidal morphologies .
= cmcsc10 at 12pt = cmcsc10 = cmr12 = cmr10 # 1@xmath1 # 1@xmath2 |
quantum computers are devices that use quantum mechanics for enhanced ways of information processing @xcite .
indeed , it is known that problems such as integer factoring can be performed significantly faster on a quantum computer than on any known classical device @xcite . despite these exciting perspectives , the questions : _ `` what are the essential resources that give quantum computers their computational power '' _ , and _ `` are quantum computers fundamentally more powerful than classical devices ? '' _ remain to date largely unanswered .
the existence of several models for quantum computation , each based on different concepts , indicates that there may not be a straightforward answer to these difficult questions . the new paradigm of _ measurement - based _ , or _
one - way _ quantum computation @xcite , @xcite has lead to novel perspectives in these respects .
the introduction of this model established that certain many - qubit quantum states , such as the _
2d cluster states _
@xcite , exhibit the remarkable property that universal quantum computation can be achieved by simply individually measuring the qubits of the system in a specific order and basis , and by classical processing of the measurement results .
the initial state of the system then serves as the resource for the entire computation which is ( in part ) consumed in the process .
this is in sharp contrast to the quantum circuit model , where computations are realized via unitary evolution . within the measurement - based paradigm for quantum computation ,
fundamental questions regarding the speed - up of quantum with respect to classical computation can be formulated and investigated in an alternative , and in several cases much more concise , way . in particular , the introductory questions of this paper can be restated as _ `` which resource states for measurement - based quantum computation ( mqc ) yield a computational speed - up over classical computers ? '' _
this question will be adressed in the present article .
as entanglement can only decrease in a one - way computation , the enhanced computational power of such a quantum computer ( beyond a classical turing machine ) must originate in the entanglement structure of its resource state .
owing to this insight , a series of papers have recently been devoted to investigating which types of entanglement needs to be present in any resource state which achieves the desired enhanced computational power @xcite @xcite @xcite @xcite @xcite @xcite @xcite @xcite @xcite . in this paper
we establish a new necessary condition for resource states to yield a computational speed - up with respect to classical computers .
the crucial point of this result is that the present criterion is entirely different in nature with respect to previously established requirements in particular , it will _ not _ be stated in terms of entanglement . in the following
we will focus on resource states belonging to the rich class of _ graph states _
, which are generalizations of the 2d cluster states and which play an important role in several applications in quantum information theory ( e.g. , one - way quantum computation , quantum error - correction , multipartite entanglement theory , communication schemes ; see @xcite for a review ) . a graph state on @xmath0 qubits
is defined by means of a mathematical graph on @xmath0 vertices , which completely encodes the correlations in the system .
our main result will be a connection between the possibility of obtaining a computational speed - up w.r.t .
classical computation by performing mqc on graph state resources , and certain properties of the associated graphs in relation with mathematical logic theory more particularly , the _ decidability _ of logic theories .
as is well known since gdel s incompleteness theorems @xcite , every formal system that is sufficiently interesting ( or rich ) contains statements which can neither be proved to be true nor to be false within the axiomatic framework of the system .
gdel s incompleteness theorem not only applies to formal systems relating to natural numbers ( cf .
peano arithmetic ) but also to graphs .
indeed , many interesting graph properties can be expressed within a ( classical ) formal language , denoted by @xmath1 ( the exact definition of this language is stated below ) .
examples of such properties are planarity or 2-colorability of graphs . in this paper
we will show that the computational power of graph states as resources for measurement - based quantum computation is reflected in the expressive power of the formal language @xmath1 defined on the underlying graphs , which encode the set of quantum correlations in the system .
in particular , the following result will be obtained .
_ graph state resources for measurement - based quantum computation can only yield a computational speed - up over classical computers if the formal language @xmath1 defined on the underlying graphs is undecidable .
_ here undecidability is to be interpreted in a sense similar to gdel , meaning that there exist propositions , expressible in the logic @xmath1 , which can not be proven or disproven .
the theorem provides a necessary condition to assess the computational power of graph state resources by considering the underlying graphs , which , by the very definition of graph states , render a classical encoding of the quantum correlations in the states .
the concept of undecidability is to be regarded as a notion of complexity of the graphs and hence of the correlations in the system stated independently of any quantitative ( entanglement ) measure .
this paper aims at connecting two quite different fields of research , namely quantum computation and mathematical logic .
therefore , in the following we will give a brief review of the basic concepts of measurement - based ( one - way ) quantum computation , as well as logic theories on graphs .
this will allow us to reformulate the main theorem in a precise manner .
the proof of the theorem is obtained by combining our previous results regarding entanglement width and mqc on graph states @xcite , and a recent graph theoretic result @xcite .
we will conclude the paper with an interpretation of these results .
we will be concerned with a particular class of multi - party quantum states , called _ graph states _
@xcite , which are generalizations of the 2d cluster states .
a graph state @xmath2 on @xmath3 qubits is the joint fixed point ( i.e. , an eigenvector with eigenvalue 1 ) of @xmath3 commuting correlation operators [ k ] k_a:= _ x^(a)_bn(a ) _ z^(b),where @xmath4 and @xmath5 are pauli matrices , and the upper indices denote on which qubit system these operators act .
moreover , @xmath6 denotes the set of neighbors of qubit @xmath7 in the graph @xmath8 .
thus , a system in a graph state has @xmath9 for every @xmath7 .
for example , a 2d cluster state is obtained if the underlying graph is a @xmath10 square lattice @xmath11 ( thus @xmath12 ) .
the family of 2d cluster states is known to be a _ universal _ resource for measurement based quantum computation , in that any unitary operation can efficiently be simulated by performing measurements on a 2d cluster state of appropriate dimensions @xcite . as graph states
are generally highly entangled , and as they can efficiently be prepared by applying a suitable poly - sized quantum circuit to a product input state , they form natural candidates to serve as resources for mqc .
we envisage a situation where an ( infinitely large ) family of graph states = \{|g_1 , measurements can be performed on arbitrary members of @xmath13 , thus allowing to implement quantum computations of arbitrary length . for example , @xmath13 could be the family of 2d cluster states , @xmath14 . in this
setting we are interested in resources @xmath13 for mqc which yield a computational speed - up over classical devices . given a family of states @xmath13 , we will say that efficient classical simulation of mqc on @xmath13 is possible , if for every state @xmath15 it is possible to simulate every locc protocol ( short for _ local operations and classical communication _ ) on a classical computer with overhead poly@xmath16 , where @xmath17 denotes the number of qubits on which the state @xmath18 is defined @xcite .
evidently , resources which allow a computational speed - up over classical computers do _ not _ allow efficient classical simulation of mqc .
in the following we will focus on the graphs associated to families of graph states , and the formal languages defined on them .
note that by definition ( [ k ] ) the graph @xmath8 is an encoding of the correlations present in the corresponding graph state .
therefore , any property of these graphs reflects a property of the corresponding states , and , more particularly , the correlations in these states .
next we define some basic notions of logic theory which are necessary to state our main results concisely below .
we refer to refs .
@xcite @xcite for an extensive treatment .
we also emphasize that we will favor clarity of the exposition over mathematical rigor . in fundamental aspects of graph theory one
is interested in formal approaches to formulate graph properties such as 2-colorability , connectedness , planarity , etc .
this formalization is obtained by defining a logical calculus in which such graph properties can be expressed .
roughly speaking , a logic on a graph @xmath8 corresponds to a set of rules which determine the basic constituents with which statements regarding @xmath8 can be constructed .
such formalization in terms of logic allows , in principle , artificial devices to mechanically prove or disprove , for a given graph or set of graphs , properties expressible in this logic .
the simplest logic is _ first - order logic _ ,
which is obtained by allowing formulas containing the following elementary components : * quantifications @xmath19 and @xmath20 over vertices @xmath21 of the graph ( i.e. , `` there exists a vertex @xmath21 such that [ ... ] '' , or `` for all vertices @xmath21 it holds that [ ... ] '' ) and connectives @xmath22 , @xmath23 , and @xmath24 , i.e. , the logical `` and '' , `` or '' and `` not '' .
* moreover , it is allowed to express whether two vertices are adjacent in the graph or not .
this is formally achieved by introducing the symbol @xmath25 , which is defined by ( a , b)= \ { + cl & \{a , b } + & , + . for every pair of vertices @xmath26 of the graph .
a simple example of a first - order logic formula on a graph @xmath8 is `` there exist vertices @xmath21 , @xmath27 , and @xmath28 such that @xmath21 is connected to @xmath27 and @xmath27 is connected to @xmath28 '' or , more formally , x y z edge(x , y ) ( y , z ) .
it turns out that first - order logic is often not rich enough to express interesting graph properties .
one therefore extends first - order logic by allowing more elementary symbols .
in particular , one may supplement first - order logic with the following elements : * next to variables @xmath29 denoting vertices of the graph , one also allows _ set variables _
@xmath30 ( indicated by capital letters ) , which denote subsets of vertices . *
furthermore , one adds quantifications @xmath31 , @xmath32 over such sets . * finally , one introduces elementary formulas of the form @xmath33 , which allow one to express that a vertex @xmath21 belongs to a certain subset @xmath34 .
the logical calculus which is thus obtained is called _ monadic second - order logic _ , or _
ms logic _ in short .
ms logic is strictly more expressive than first - order logic , i.e. , there are problems which can be expressed with ms logic which can not be expressed using only first - order logic .
an example of an ms formula on a graph is x y\{z ( zx zy ) + z z ( z , z)(z , zxz , zy ) } .this formula expresses that the graph can properly be colored with 2 colors ( i.e. , the vertices can partitioned in two classes such that no two adjacent vertices are in the same class ) .
many interesting graph properties can be expressed in ms logic , among which there are several np - hard problems ( such as e.g. 3-colorability ) , indicating that ms logic on graphs has a considerable expressive power .
a slight extension of ms logic is obtained by including atomic formulas of the form @xmath35 , indicating that the set @xmath34 has even cardinality . in this way one obtains ms logic with the additional possibility to count modulo two , denoted by _ c@xmath36ms logic _
@xcite , which will be our topic of interest , i.e. , it corresponds to the logic @xmath1 as denoted above .
c@xmath36ms logic is an interesting extension of ms logic which is moreover physically interesting in that it e.g. allows to express whether two graph states are local unitary equivalent @xcite .
next we introduce the fundamental notion of _ decidability _ of logic theories .
we again refer to @xcite @xcite for details .
let @xmath37 be a ( finite or infinite ) family of graphs , and let @xmath1 denote c@xmath36ms logic .
the @xmath1-theory of @xmath38 is defined to be the collection of all formulas @xmath39 , expressed in the logic @xmath1 , which are satisfied ( or `` true '' ) for _ all _ graphs in the family @xmath38 .
the set @xmath38 is said to have a _
_ @xmath1-theory if for every formula @xmath39 expressed in the logic @xmath1 , it is possible to decide ( in finite time ) whether or not @xmath39 belongs to the @xmath1-theory of @xmath38 .
the set @xmath38 is said to have an _
@xmath1-theory if it does not have a decidable @xmath1-theory .
for example , the formula @xmath39 could correspond to graph planarity , 2-colorability , etc . then the question is asked whether all graphs in a given family of graphs @xmath38 are planar , 2-colorable , etc . if every possible such question , that is , every formula expressible in the language @xmath1 , can be answered in finite time , then this family is said to have a decidable @xmath40-theory .
the decidability or undecidability of a logic theory @xmath1 on a set @xmath41 is a reflection of both the expressive power of the logic @xmath1 `` how many properties can be expressed in the logic @xmath40?'' and the complexity ( regarded in a colloquial sense ) of the family @xmath38``how rich is the structure of the graphs in @xmath38 ? '' . before giving examples of ( un)decidable ms theories , is important to make the following two remarks .
first , decidability of a logic theory is not concerned with the _ efficiency _ with which problems can be solved
one only asks whether it is _ in principle _ possible to verify whether a given formula @xmath39 is true , where one does not care about e.g. the computational complexity of a possible verification algorithm .
second , note that _ any _ first - order or ms theory is decidable on _ finite _ families of graphs @xmath42 .
this is simply because , in this finite regime , any formula can be verified by an exhaustive enumeration of cases .
thus , decidability is only relevant when infinite structures are considered .
let us now give some important examples .
first , let @xmath43 be the set of all binary tree graphs , which are regarded as so - called incidence structures .
a milestone result was obtained by rabin , who proved that @xmath44 has a decidable ms theory @xcite .
this result has many important implications in graph theory and computer science .
further , let @xmath45 be the set of all 2d ( @xmath10 ) lattice graphs
. then @xmath45 has an _ undecidable _
c@xmath36ms theory @xcite . as final examples ,
let @xmath46 and @xmath47 be the sets of all triangular lattice graphs and hexagonal lattice graphs , respectively , regarded as adjacency structures
. then also @xmath46 and @xmath48 have _ undecidable _
c@xmath36ms theories @xcite .
keeping in mind that the graph states corresponding to the 2d rectangular , hexagonal and triangular lattices have been shown to be universal resources for mqc @xcite , the above examples already suggest a connection between the computational power of a family of graph states as a resource for mqc , and the c@xmath36ms logic defined on the underlying graphs . this connection will now be fully established , as we are now in a position to precisely state and prove the main result of this paper .
let @xmath49 be an ( infinitely large ) family of graphs and let @xmath50 be the associated family of graph states .
the main theorem can then precisely be formulated as follows .
_ if a family of graphs @xmath38 has a decidable c@xmath51ms logic theory then mqc performed on the graph state resource @xmath50 can classically be simulated efficiently . _
thus , this results states that any family of graphs with a decidable c@xmath51ms logic theory _ can not _ give rise to a graph state resource for mqc which yields a computational speed - up as compared to classical computers .
the proof of the theorem is in fact quickly obtained by invoking previous results of the present authors and a highly nontrivial result from graph theory .
the proof has two main ingredients ( i ) and ( ii ) : * courcelle and oum @xcite proved that every class of graphs @xmath38 which exhibits a divergence with respect to a graph invariant called _ rank - width _ ( we refer to ref .
@xcite for definitions ) must have an undecidable c@xmath51ms theory . * in previous work @xcite , the present and other authors proved that every family of resource states @xmath50 where @xmath38 has a bounded rank - width , allows an efficient simulation of mqc . combining ( i ) and ( ii ) then yields the proof of the theorem .
next we elaborate on the above proof strategy .
let us first introduce the notion `` rank - width '' .
the rank - width rwd(@xmath8 ) @xcite of a graph @xmath8 is a parameter which measures how well a graph can be approximated by means of certain `` tree - like '' structures .
graphs with small rank - width include , e.g. , a one - dimensional chain with open or closed boundary conditions , or a 2-dimensional `` stripe '' , which is a @xmath52 square lattice where @xmath0 may be arbitrarily large , but where @xmath53 is held fixed . graphs of large rank - width include e.g. 2-dimensional @xmath54 lattices ( for growing @xmath0 ) , or lattices of higher dimensions . for completeness , we give here the definition of the rank - width , which is quite technical [ the reader who is not interested in these mathematical details may skip to the next paragraph ] .
let @xmath8 be a graph with vertex set @xmath55 and edge set @xmath56 .
let @xmath57 be the @xmath54 adjacency matrix of @xmath8 , i.e , one has @xmath58 if @xmath59 and @xmath60 otherwise . for every bipartition @xmath61 of the vertex set @xmath62 ,
define @xmath63 to be the @xmath64 submatrix of @xmath57 defined by ( a , b ) : = ( _ ab)_aa , bb.let @xmath65 be a _ subcubic _
tree , which is a tree such that every vertex has exactly 1 or 3 incident edges ( see fig [ subcubic](a ) ) .
the vertices which are incident with exactly one edge are called the _ leaves _ of the tree . for a given fixed graph @xmath8 , we will be interested in the collection of all possible subcubic trees @xmath65 with exactly @xmath0 leaves @xmath66 , which are identified with the @xmath0 vertices of @xmath8 .
letting @xmath67 be an arbitrary edge of such a tree @xmath65 , we denote by @xmath68 the graph obtained by deleting the edge @xmath69 from @xmath65 .
the graph @xmath68 then consists of exactly two connected components , which naturally induce a bipartition @xmath70 of the set @xmath62 ( see fig .
[ subcubic](b ) ) .
the rank - width of the graph @xmath8 is now defined by the following optimization problem : ( g ) = _
t_et _ 2 ( a_t^e , b_t^e ) . here the minimization is taken over all subcubic trees @xmath65 with @xmath0 leaves , which are identified with the @xmath0 vertices in the graph .
moreover , @xmath71 denotes the rank of the matrix @xmath63 when arithmetic is performed modulo 2 .
one notices that the construction involving the subcubic trees is designed to single out a specific class of bipartitions @xmath70 of the vertex set of @xmath8 , over which the min - max optimization problem is performed . for a given tree
@xmath65 one considers the maximum , over all edges @xmath69 in @xmath65 , of the quantity @xmath72 ; then the minimum , over all subcubic trees @xmath65 , of such maxima is computed . as an example
, it can be shown that the rank - width of a 1d chain is equal to 1 ( independent of the length of the chain ) , whereas an @xmath54 square lattice has rank - with of @xmath73 and thus increases with the size of the lattice .
while the rank - width is indeed a rather involved mathematical concept , it turns out to be crucial for the current investigation .
in particular , the following highly nontrivial result by courcelle and oum is of particular interest @xcite : _ every class of graphs @xmath38 with an unbounded rank - width must have an undecidable c@xmath51ms theory . _
this result connects the notion of rank - width of a class of graphs with the decidability of the logic theory of this class .
we will use this result to prove the theorem by juxtaposing it to a result obtained by the present and other authors , which relates the rank - width of a ( family of ) graph(s ) with the computational power of the associated ( family of ) graph state(s ) . for
, it was proved in @xcite that : .
the combination of this result with the one obtained by courcelle and oum , then immediately yields the proof of the theorem .
finally , we conclude this paragraph by remarking that the theorem represents a sufficient condition for a resource @xmath50 to be simulatable , but not a necessary one . for
, there exist graph state resources with an unbounded rank - width and therefore an undecidable c@xmath36ms logic theory for which mqc is nevertheless simulatable ; examples of such resources are given by the so - called `` toric code states '' @xcite , or graph states with logarithmically growing rank - width @xcite .
in the theorem the desired connection between measurement based quantum computation on graph states and mathematical logic theories on the underlying graphs is fully obtained .
it is the authors opinion that the present results should be regarded as conceptual results , aimed at establishing a connection between seemingly remote areas of research , rather than yielding direct practical applications .
we are aware that assessing whether a family of graphs has a decidable c@xmath36ms theory is a formidable task , and that logic theory itself is a dynamic area of research with difficult outstanding problems @xcite
. therefore the present results are not likely to e.g. directly provide new examples of states on which mqc can be simulated efficiently . nevertheless , we believe that our findings present a new , and possibly deep , perspective towards understanding the central issue of what the computational power of quantum computers with respect to classical devices is .
the present connection to logic theory offers an entirely new view on `` how complex '' states need to be in order for them to possibly provide computational speed - ups , next to more standard considerations regarding entanglement .
while the theorem in fact follows from previously obtained results regarding mqc and entanglement width of graph states @xcite @xcite , the resulting logic criteria are entirely different in nature .
finally , one might be inclined to relate ( c@xmath36ms ) logic formulas defined on graphs to the content of the quantum computations ( i.e. , measurement patterns ) implemented on the corresponding graph states .
as far as the authors are aware , there does not seem to be a direct relation between the classical logic defined on graphs and the quantum measurements that might be associated to a quantum logic on the corresponding graph states .
thusfar it seems that the classical logic theories , and the issue of their ( un)decidability , are related to assessing the complexity of the graphs , and hence of the ( correlations in the ) graph states , with no direct correspondence to quantum algorithms .
however , it would be very interesting to investigate this issue in more detail , and we leave this as an open problem .
this is proven by using a result in ref .
@xcite stating that a graph transformation called _ local complementation _ can be expressed in c@xmath36ms logic , and by noting that local complementation of graphs corresponds to local unitary operations on the corresponding graph states @xcite .
as @xmath74 and @xmath75 correspond to universal resources @xcite , these families have unbounded rank width . using @xcite
then again yields the undecidability of the corresponding c@xmath36ms theories . | we establish a connection between measurement - based quantum computation and the field of mathematical logic .
we show that the computational power of an important class of quantum states called _ graph states _ , representing resources for measurement - based quantum computation , is reflected in the expressive power of ( classical ) formal logic languages defined on the underlying mathematical graphs .
in particular , we show that for all graph state resources which can yield a computational speed - up with respect to classical computation , the underlying graphs describing the quantum correlations of the states are associated with _
undecidable _ logic theories . here
undecidability is to be interpreted in a sense similar to gdel s incompleteness results , meaning that there exist propositions , expressible in the above classical formal logic , which can not be proven or disproven . |
let @xmath8 be a simple arrangement formed by @xmath0 hyperplanes in dimension @xmath1 .
we recall that an arrangement is called simple if @xmath9 and any @xmath1 hyperplanes intersect at a distinct point .
the closures of connected components of the complement of the hyperplanes forming @xmath10 are called the cells , or @xmath1-faces , of the arrangement . for @xmath11 ,
the @xmath12-faces of @xmath10 are the @xmath12-faces of its cells .
a facet is a @xmath13-face of @xmath8 , and a facet belonging to exactly one bounded cell is called an external facet .
equivalently , an external facet is a bounded facet which belongs to an unbounded cell . for @xmath14 ,
an external @xmath12-face is a @xmath12-face belonging to an external facet .
let @xmath15 denote the number of external @xmath12-faces of @xmath10 .
the set of all external facets forms the envelope of the arrangement .
it was hypothesized in @xcite that any simple arrangement @xmath8 has at least @xmath2 external facets . in section [ 2d ] , we show that a simple arrangement of @xmath0 lines has at least @xmath4 external facets for @xmath16 , and that this bound is tight . in section [ 3d ] , we show that a simple arrangement of @xmath0 planes has at least @xmath6 external facets for @xmath17 , and exhibit a simple plane arrangement with @xmath7 external facets . for polytopes and arrangements
, we refer to the books of edelsbrunner @xcite , grnbaum @xcite and ziegler @xcite and the references therein .
[ ext2dlb ] for @xmath16 , a simple line arrangement has at least @xmath4 external facets .
the external vertices of a line arrangement can be divided into three types , namely @xmath18 , @xmath19 and @xmath20 , corresponding to external vertices respectively incident to 2 , 3 , and 4 bounded edges .
let us assign to each external vertex @xmath21 a weight of 1 and redistribute it to the 2 lines intersecting at @xmath21 the following way : if @xmath21 is incident to exactly 1 unbounded edge , then give weight 1 to the line containing this edge , and weight 0 to the other line containing @xmath21 ; if @xmath21 is incident to @xmath22 or @xmath23 unbounded edges , then give weight @xmath24 to each of the @xmath22 lines intersecting at @xmath21 .
see figure [ fig_weight_distribution ] for an illustration of the weight distribution .
a total of @xmath25 weights is distributed and we can also count this quantity line - wise .
the end vertices of a line being of type @xmath18 or @xmath19 , we have three types of lines , @xmath26 and @xmath27 , according to the possible types of their end - vertices .
as a line of type @xmath27 contains @xmath22 vertices of type @xmath19 , its weight is at least @xmath22 .
similarly the weight of a line of type @xmath28 weight is at least @xmath29 . remarking that a line of type @xmath30 contains at least one vertex of type @xmath20 yields that the weight of a line of type @xmath30 is at least @xmath31 . for @xmath16
the number of lines of type @xmath28 is at most 2 as otherwise the envelope would be convex which is impossible , see for example @xcite .
therefore , counting the total distributed weight line - wise , we have @xmath32 .
since for a line arrangement the number of external facets @xmath33 is equal to the number of external vertices @xmath34 , we have @xmath35 . for @xmath16 ,
consider the following simple line arrangement : @xmath36 is made of the @xmath22 lines @xmath37 and @xmath38 forming , respectively , the @xmath39 and @xmath40 axis , and @xmath41 lines defined by their intersections with @xmath37 and @xmath38 .
we have @xmath42 and @xmath43 for @xmath44 , and @xmath45 and @xmath46 where @xmath47 is a constant satisfying @xmath48 .
see figure [ a072 ] for an arrangement combinatorially equivalent to @xmath49 .
one can easily check that @xmath49 has @xmath4 external facets and therefore the lower bound given in proposition [ ext2dlb ] is tight . for @xmath16 ,
the minimum possible number of external facets of a simple line arrangement is @xmath4 .
, height=377 ]
let @xmath50 for @xmath51 be the planes forming the arrangement @xmath52 . for @xmath51 , the external vertices of the line arrangement @xmath53 are external vertices of the plane arrangement @xmath52 . for @xmath17 ,
the line arrangement @xmath53 has at least @xmath54 external facets by proposition [ ext2dlb ] , i.e. , at least @xmath54 external vertices .
since an external vertex of @xmath52 belongs to 3 planes , it is counted three times . in other words ,
the number of external vertices of @xmath52 satisfies @xmath55 for @xmath17 . as the union of all of the bounded cells is a piecewise linear ball ,
see @xcite , the euler characteristic of the boundary gives @xmath56 .
since an external vertex belong to at least 3 external edges , we have @xmath57 .
thus , we have @xmath58 . as @xmath55 , it gives @xmath59 for @xmath17 , we consider following simple plane arrangement : @xmath60 is made of the @xmath61 planes @xmath37 , @xmath38 and @xmath62 corresponding , respectively , to @xmath63 , @xmath64 and @xmath65 , and @xmath66 planes defined by their intersections with the @xmath39 , @xmath40 and @xmath67 axis .
we have @xmath68 , @xmath69 and @xmath70 for @xmath71 , and @xmath72 , @xmath73 and @xmath74 where @xmath47 is a constant satisfying @xmath75 .
see figure [ a073 ] for an illustration of an arrangement combinatorially equivalent to @xmath76 where , for clarity , only the bounded cells belonging to the positive orthant are drawn .
we first check by induction that the arrangement @xmath77 formed by the first @xmath0 planes of @xmath78 has @xmath79 external facets .
the arrangement @xmath77 is combinatorially equivalent to the plane cyclic arrangement which is dual to the cyclic polytope , see @xcite for combinatorial properties of the ( projective ) cyclic arrangement in general dimension . see figure [ a63 ] for an illustration of @xmath80 .
let @xmath81 denote the half - space defined by @xmath62 and containing the positive orthant , and @xmath82 the other half - space defined by @xmath62 .
the union of the bounded cells of @xmath83 in @xmath82 is combinatorially equivalent to the bounded cells of @xmath84 and therefore has @xmath85 facets on its boundary by induction hypothesis , including @xmath86 bounded facets contained in @xmath62 .
these @xmath86 bounded facets also belong to a bounded cell of @xmath83 in @xmath82 and therefore are not external facets of @xmath83 .
thus , the number of external facets of @xmath83 belonging to a bounded cell in @xmath82 is @xmath87 .
the union of the bounded cells of @xmath83 in @xmath81 can be viewed as a simplex cut by @xmath88 sliding down planes .
it has @xmath89 facets on its boundary , including the @xmath86 bounded facets contained in @xmath62 belonging to a bounded cell of @xmath83 in @xmath82 .
thus , the number of external facets of @xmath83 belonging to a bounded cell in @xmath81 is @xmath90 .
therefore , @xmath83 has @xmath91 external facets .
we now consider how the addition of @xmath92 to @xmath84 impacts the number of external facets .
this impact is similar in nature to the addition of @xmath92 to the first @xmath93 lines of @xmath36 .
the addition of @xmath92 creates @xmath94 new bounded cells : one above @xmath37 that we call the _ @xmath0-shell _ , and the other ones being below @xmath37 .
the @xmath0-shell turns @xmath88 external facets of @xmath84 above @xmath37 into internal facets of @xmath60 , and adds 3 external facets .
for each external facet of @xmath84 belonging to @xmath37 which is turned into an internal facet of @xmath60 , one external facet of @xmath60 on @xmath92 and not incident to @xmath37 is added .
below @xmath37 , the addition of @xmath92 creates @xmath95 new external facets of @xmath60 with an edge on @xmath37 .
finally , @xmath88 new external facets belonging to @xmath37 and bounded by @xmath92 are created from unbounded facets of @xmath84 .
thus , the total number of external facets of @xmath96 is @xmath97 .
we do not believe that @xmath98 minimizes the number of external facets . among the 43 simple combinatorial types of arrangements formed by 6 planes ,
the minimum number of external facets is @xmath99 while @xmath100 has 23 external facets .
see figure [ a63_29 ] for an illustration of the combinatorial type of one of the two simple arrangements with @xmath101 planes having 22 external facets .
the far away vertex on the right and 3 bounded edges incident to it are cut off ( same for the far away vertex on the left ) so the 10 bounded cells of the arrangement appear not too small .
david bremner + faculty of computer science , + university of new brunswick , new brunswick , canada .
+ _ email _ : [email protected] + + antoine deza , feng xie + department of computing and software , + mcmaster university , hamilton , ontario , canada .
+ _ email _ : deza , [email protected] | a facet of an hyperplane arrangement is called external if it belongs to exactly one bounded cell .
the set of all external facets forms the envelope of the arrangement .
the number of external facets of a simple arrangement defined by @xmath0 hyperplanes in dimension @xmath1 is hypothesized to be at least @xmath2 . in this note
we show that , for simple arrangements of @xmath3 lines or more , the minimum number of external facets is equal to @xmath4 , and for simple arrangements of @xmath5 planes or more , the minimum number of external facets is between @xmath6 and @xmath7 . |
the pioneering experiments of projectile fragmentation at relativistic energies of @xmath0ar and @xmath1ca beams at the lbl bevalac @xcite have demonstrated the potential of this method for the production of exotic nuclei .
based on these ideas , the sis / frs facility @xcite at gsi has used also heavier projectiles like e.g. @xmath2ni , @xmath3kr , @xmath4xe , and @xmath5pb to produce and study exotic nuclei @xcite . for planning such experiments ,
when count - rate predictions are needed , analytical descriptions of fragmentation cross sections are useful .
they are also useful in simulation programs for projectile - fragment separators ( like e.g. intensity @xcite or mocadi @xcite ) .
compared to physical models of high - energy fragmentation reactions , which in general involve time - consuming monte - carlo calculations , the virtue of an analytical formula lies in the short computing time and the possibility to calculate easily sub - microbarn cross sections that are beyond the reach of physical - model calculations . in 1990 , smmerer _ et al . _
@xcite proposed a universal empirical parametrization of fragmentation cross sections ( `` epax '' , ref .
@xcite ) which was based on and similar to previous prescriptions by rudstam @xcite and silberberg _ et al . _
the parametrization was to a large extent based on multi - gev proton - induced spallation cross sections , since only scarce heavy - ion induced experimental data were available at that time .
meanwhile , more precise data from relativistic heavy - ion - induced fragmentation reactions together with recent results from projectile fragmentation of heavy nuclei ( @xmath6au and @xmath5pb ) on h@xmath7 targets @xcite ) allow a more stringent comparison of proton- and heavy - ion induced isotope distributions .
this comparison indicates that for heavy nuclei the two reactions lead to different isotopic distributions , which can not be obtained from each other just by scaling factors .
this can be easily understood since heavy - ion induced reactions are expected to deposit more excitation energy in a nucleus than proton - induced reactions , making the final product distributions after evaporation broader and more neutron - deficient .
nevertheless , the data show that in both cases the isotopic yield distributions can be well described by gaussian - like analytical functions with parameters that vary smoothly as a function of fragment mass @xcite . in the present paper
, we will base the choice of these parameters exclusively on heavy - ion - induced reaction data .
we will first review briefly the basic characteristics of the epax formula and then show which modifications are necessary to improve the accuracy with which the new experimental results can be reproduced
. this will be followed by a brief comparison with similar attempts by other authors .
the basic characteristics of the analytical description of high - energy fragmentation cross sections by the epax formula are the following @xcite : * in the absence of systematic excitation - function measurements of heavy - ion induced fragmentation reactions , the formula is valid only for the so - called `` limiting fragmentation '' regime , i.e. for projectile energies where the fragmentation yields are no longer energy dependent , at least within the accuracy of the formula ( approximately within a factor of 2 ) .
this is certainly true for incident energies considerably above the fermi energy in nuclei ( @xmath8 40 @xmath9 mev ) , in particular for the typical sis energies of 500 to 1000 @xmath9 mev . *
the epax formula is meant to describe the fragmentation of medium- to heavy - mass projectiles ; nucleon - pickup cross sections are not included .
no attempt is made to describe the fragmentation of fissile nuclei .
therefore , the range of validity is limited to projectiles from around argon to below the lead and bismuth isotopes .
predictions for production cross sections of fission products or of fragments below u where fission competition is significant require an elaborate description of the fission process , such as can be found e.g. in a recent publication by benlliure _
_ @xcite . * for fragments sufficiently far away from the projectile ( i.e. for mass losses larger than 15 - 20% of the projectile mass ) , the isotope distributions are largely independent of the original nucleus ; their position , shape , and width depend only on the fragment mass number .
this constitutes what has been termed the `` residue corridor '' and is related to the fact that the isotope distributions are mainly governed by statistical evaporation from highly excited prefragments produced in collisions between relativistic heavy ions . * for fragments that involve only a small mass loss from the projectile , the isotope distributions should be centered close to the projectile and their variance should be small .
therefore , a smooth transition is anticipated between the residue corridor and the projectile .
the parametrization of this smooth transition constitutes the main task in designing the formula . in a first step
, a parameter set has been searched for that describes the fragmentation yields from projectiles located close to the line of @xmath10stability . in a second step , a modification of the yield distributions due to the neutron or proton excess of projectiles located on the neutron- or proton - rich side of the line of @xmath10stability (
the `` memory effect '' ) has been parametrized . as explained in detail in ref .
@xcite , the cross section ( in barn ) of a fragment with mass @xmath9 and charge @xmath11 produced by projectile fragmentation from a projectile @xmath12 impinging on a target @xmath13 is written as @xmath14 the first term , @xmath15 , represents the mass yield , i.e. the sum of the isobaric cross sections with fragment mass @xmath9 .
the second term describes the `` charge dispersion '' , the distribution of elemental cross sections with a given mass around its maximum , @xmath16 .
the shape of the charge dispersion is controlled by the width parameter , @xmath17 , and the exponent , @xmath18(@xmath19 ) , on the neutron-(proton- ) rich side of the residue corridor .
the factor @xmath20 simply serves to normalize the integral of the charge dispersion to unity .
the mass - yield curve is taken to be an exponential as a function of @xmath21 .
the slope of this exponential , @xmath22 , is a function of the projectile mass .
an overall scaling factor , @xmath23 , accounts for the peripheral nature of fragmentation reactions and therefore depends on the circumference of the colliding nuclei : @xmath24\label{eq3 } \\
ln~p & = & p_2 \cdot a_p + p_1 \label{eq4 } \end{aligned}\ ] ] the numerical values of the various constants can be found in table [ tab_const ] .
.constants used in epax version 1 ( ref.@xcite ) and those used in version 2 ( this work ) .
note that in a few cases also the functional form to calculate a parameter has changed ( see text ) .
[ cols="^,^ , < , < " , ] [ tab_const ] the charge dispersion is characterized by the three parameters @xmath17 , @xmath16 , and @xmath25 .
these three parameters are strongly correlated and difficult to obtain uniquely with a least - squares fitting technique .
note that the isobar distributions are not symmetric on the neutron- and proton - rich side , therefore @xmath25 has two different values , @xmath19 and @xmath18 , on the proton- or neutron - rich side of the valley of @xmath26-stability , respectively . in ref .
@xcite , the exponent @xmath25 for the neutron - rich side of the isobar distribution was chosen as @xmath27 , whereas the proton - rich side falls off like a gaussian ( @xmath28 ) .
the maxima of the charge dispersions , @xmath16 , have been parametrized relative to the valley of @xmath26-stability , @xmath29 @xmath30 is approximated by the smooth function @xmath31 @xmath32 is found to be a linear function of the fragment mass , @xmath9 , for heavy fragments ( @xmath33 ) , and is extrapolated quadratically to zero : @xmath34 similar to the parameter @xmath16 just discussed , the width parameter , @xmath17 , is a function of fragment mass only , irrespective of the projectile . in ref .
@xcite it was found that the experimental @xmath17-values can be approximated by an exponential of the form @xmath35 the equations given above are sufficient to describe the `` residue corridor '' , i.e. the yield distributions of projectiles located on the line of @xmath26-stability if the fragment mass is far from the projectile mass .
close to the projectile , the following modifications have to be introduced @xcite : @xmath36 \label{eq9}\\ r & = & r \cdot [ 1+r_1 \cdot ( a / a_p - r_2)^2 ] \label{eq10}\end{aligned}\ ] ] this serves to gradually reduce to zero the offset of @xmath16 from the line of @xmath26-stability and to decrease the width of the charge dispersion when @xmath9 approaches the mass of the projectile , @xmath37 .
a final correction applies if the projectile does not lie on the line of @xmath26-stability . in this case , the @xmath38 ratio of the fragments will to some extent `` remember '' the @xmath39 ratio of the projectile ( `` memory effect '' ) . for an analytical description
, it is simply assumed that the charge dispersions are shifted by an amount @xmath40 which is a certain fraction of the distance of @xmath41 from @xmath42 , the nuclear charge on the line of @xmath26-stability for @xmath37 .
close to the projectile , this fraction is clearly close to unity ( full memory effect ) , whereas it should gradually approach zero with increasing distance of the fragment mass @xmath9 from the projectile ( loss of memory ) .
the shape of the isobar distribution is assumed to be unchanged .
thus @xmath43 where @xmath40 for neutron - rich projectiles is given by @xmath44\cdot ( z_p - z_{\beta p } ) \label{eq12}\ ] ] the corresponding constants for proton - rich projectiles are termed @xmath45 and @xmath46 instead of @xmath47 and @xmath48 .
numerical values for all constants are given in table [ tab_const ] . with this parametrization ,
epax version 1 was rather successful in describing the gross features of isotope distributions of high - energy projectile fragmentation .
this has been visualized e.g. in refs .
@xcite , where comparisons of epax with experimental data over many orders of magnitude in cross section can be found . as a particular example , we plot in fig.[fig_pb ] experimental isotope distributions from 1 @xmath9 gev @xmath5pb fragmentation @xcite in comparison with the epax version 1 parametrization ( dashed curves ) .
this set of data was also chosen to illustrate that the old epax version has problems to reproduce satisfactorily the isotope distributions of very heavy fragments .
as can be seen best for the low-@xmath11 isotope distributions , the dashed lines are centered too much on the neutron - rich side and exhibit too large maxima .
moreover , the cross sections for fragments close to the projectile ( with masses @xmath49 200 ) are predicted much too high .
minor discrepancies were also found for @xmath4xe - fragment yields @xcite where the experimental distributions were found to be wider than the epax predictions .
serious deficiencies were revealed when , compared to the epax prediction , too large cross sections for neutron - deficient isotopes from @xmath2ni fragmentation @xcite and too small cross sections of neutron - deficient sn isotopes from @xmath50sn fragmentation @xcite were measured .
the availability of extensive cross - section measurements for projectile fragmentation of @xmath0ar , @xmath2ni , @xmath3kr , @xmath4xe , and @xmath5pb @xcite down to @xmath51b or nb cross sections allowed in a first step to find a better parametrization of the mass yields @xmath52 . in a second step , we adjusted the parameters describing the residue corridor , i.e. the width parameter , @xmath17 , and the slope constants , @xmath18 and @xmath19 , of the quasi - gaussian charge dispersion together with its centroid , @xmath16 . in a third step ,
the correction factors for isotopic yields close to the projectile were modified .
finally , modifications for projectiles outside the valley of @xmath26-stability were redetermined .
numerical values for the new constants are given in the fourth column of table [ tab_const ] .
according to eqs .
( [ eq2]-[eq4 ] ) , @xmath52 is described by an exponential function depending on the fragment mass , @xmath9 , the projectile mass , @xmath37 , and the target mass , @xmath53 . in epax version 1 , no additional correction close to the projectile has been performed . as shown by the dashed lines in fig .
[ a_yield ] , this yields only a very rough agreement with experimental data .
note that the experimental mass yields had to be complemented by calculated cross sections ( from epax version 2 ) where experimental data points were missing . for the @xmath0ar , @xmath3kr , and @xmath4xe data ,
the additional calculated cross sections contribute only close to the projectile ; in the @xmath5pb case , however , the experimental data are less complete and more significant corrections were necessary . fig .
[ a_yield ] shows that the slope somewhat further away from the projectile is reasonably well reproduced , however , the absolute height as well as the slope close to the projectile mass needed to be modified .
a much better overall normalisation was achieved by reducing the scaling factor @xmath54 by a factor of 0.6 , as demonstrated by the solid lines in fig .
[ a_yield ] . an additional correction introduced in the new epax version consists of an increase in the mass yield , @xmath15 , ( eq . [ eq2 ] ) close to the projectile according to @xmath55 .
\label{eq15}\ ] ] for @xmath56 .
these modifications give a satisfactory agreement with experimental data . in the case of @xmath4xe , the slope
far from the projectile mass deserves further attention .
however , due to the lack of other complete experimental data , we were not able to significantly improve the mass - yield description .
another comparison of integrated cross sections from epax to experimental data can be performed for charge - changing cross sections . for this purpose ,
we use experimental data from binns _ et al . _
[ charge_yield ] shows a comparison of the experimental data to the epax versions 1 and 2 .
as can be seen , nice agreement is achieved with the new epax formula , in particular close to the target . for the @xmath57xe data ,
the excess charge - changing cross sections for lower @xmath11 are most likely due to secondary reactions in the targets which are unsufficiently accounted for . in the upper part of fig .
[ fig_delta_r ] , we show , as a function of the fragment mass , @xmath9 , the new position @xmath58 of the `` residue corridor '' , for @xmath4xe and @xmath5pb , fragmentation . data from @xmath59u fragmentation ( ref .
@xcite ) are also included in the figure , they were not used to fit the parameters @xmath60 and can serve as a check . for clarity , we plot the offset , @xmath32 , from @xmath30 according to eq .
the dashed lines are for epax version 1 ; the full lines represent the new version 2 .
note that for better results close to the projectile , we have added to @xmath16 according to eq .
[ eq5 ] a tiny fragment - mass dependent correction amounting to @xmath61 .
the lines sloping downward above @xmath62 and @xmath63 illustrate the return of @xmath16 to @xmath26-stability close to the @xmath4xe and @xmath5pb projectiles according to eq .
[ eq9 ] . in a similar way , the lower part of fig .
[ fig_delta_r ] visualizes the analytical description of the width parameter , @xmath17 . while the functional form of the correction for @xmath32 , eq .
[ eq9 ] , was kept identical with epax version 1 , the correction for @xmath17 near the projectile is now written as @xmath64 \label{eq13}\ ] ] for @xmath65 , which yields more narrow charge distributions for @xmath9 close to @xmath37 .
only minor changes were introduced for the slope constants , @xmath18 and @xmath19 , of the neutron and proton - rich sides of the charge distributions , respectively .
while @xmath18 was increased from 1.5 to 1.65 ( constant for all @xmath9 ) , @xmath19 is now slightly fragment - mass dependent with @xmath66 which approaches @xmath67 for large @xmath9 , compared to a previous constant value of 2.0 ( gaussian curve ) . a close inspection of the width parameter , @xmath17
, resulting from fits to the measured @xmath3kr isotope distributions @xcite shows that they are slightly but systematically decreased compared to eq .
[ eq8 ] ( see fig . [ fig_delta_r ] ) .
this means that the isobar distributions are slightly wider for a neutron - rich projectile like @xmath3kr than for a @xmath26-stable one .
therefore , we tentatively introduce the following neutron - excess dependence of @xmath17 : @xmath68 where @xmath17 is calculated according to eq .
[ eq8 ] . for neutron - deficient projectiles ,
@xmath17 is not changed . for a typical example , the combined effect of the changes described above is visualized in fig .
[ fig_xe129sn ] , which displays the isotope distributions from 800 @xmath9 mev @xmath4xe fragmentation measured over four orders of magnitude @xcite .
in particular , the data measured near the maximum are now much better reproduced .
the stronger increase in @xmath17 near the projectile mass , @xmath69 , leads to a much faster decrease of the corresponding fragment yields at @xmath70 .
similar observations , though over a smaller vertical range , can be made for the fragment distributions from @xmath5pb fragmentation shown in fig .
[ fig_pb ] .
it is interesting to note that even for @xmath59u fragmentation products in the mass range @xmath71 , produced in violent collisions and correspondingly long chains of statistical evaporation , rather good agreement between experimental data @xcite and epax can be observed . in the previous epax version
@xcite , the parametrization of the `` memory effect '' was fitted , for the case of neutron - rich projectiles , to the results from @xmath1ca fragmentation at 213 @xmath9 mev @xcite ( see eq .
[ eq12 ] ) . for the case of @xmath3kr
, we were able to check the validity of this prescription in a different fragment - mass range @xcite .
we have kept this prescription for the current version 2 , since the agreement with experimental data has not changed significantly ( see fig .
[ fig_kr86_ca48 ] for selected examples ) .
the situation is different , however , for the case of neutron - deficient projectiles .
there , the previous version was unsatifactory from the beginning : firstly , there were only very few radiochemical cross sections available to derive a guess for the memory effect on the neutron - deficient side .
secondly , the resulting formula , @xmath72\cdot ( z_p - z_{\beta p})$ ] , does not yield the limiting value @xmath73 for @xmath74 .
these deficiencies became particularly obvious when schneider _
et al . _
@xcite measured only small cross sections for neutron - removal products from @xmath50sn .
this lead to a different functional form for the memory effect for neutron - deficient projectiles .
the new parametrization reads @xmath75 \cdot ( z_p - z_{\beta p } ) \label{eq17}\ ] ] which is equivalent to a stronger but more rapidly decaying memory effect compared to the polynomial ( eq . [ eq12 ] ) used previously .
the consequences of the new parametrization of the memory effect for neutron - deficient projectiles are visualized in fig .
[ fig_sn ] .
the small measured cross sections for @xmath76sn @xcite are now better reproduced , while the good agreement ( over more than seven orders of magnitude ! ) with data measured in @xmath77xe fragmentation @xcite is maintained .
we note in passing that eq .
[ eq17 ] is very similar to what has been suggested by von egidy and schmidt @xcite to reproduce production cross sections from antiproton annihilation .
the overall agreement with the experimental fragmentation yields studied in our work is worse , however , when we adopt their description of the memory effect , in particular on the neutron - rich side .
moreover , we do not fit any set of parameters separately for a specific projectile as has been done in ref .
@xcite , but rather want to describe all systems with the same parameter set . the yield distributions from the neutron - deficient projectile @xmath2ni , studied in great detail by blank _
et al . _
@xcite , deserve special attention . not only
do they represent a rare example of cross - section measurements down to the sub - nanobarn level , but also a case where severe discrepancies with epax version 1 were observed : the yields of even-@xmath11 isotopes close to the proton drip line where found to be enhanced in experiment by factors of up to 750 @xcite . as a remedy for this deficiency ,
we have chosen to switch over from quasi - gaussian to exponential slopes of the charge distribution above a certain gradient of the cross - section distribution . in order to achieve a steady transition between the two slopes , the transition point and the slope of the exponential have to be adjusted carefully .
for this purpose , we calculate the derivative of the logarithm of the cross section ( eq . [ eq1 ] ) : @xmath78 the transition point to the exponential slope , @xmath79 , can be calculated for the proton - rich side as a function of the fragment mass @xmath9 according to @xmath80 from @xmath79 on , the slope is exponential with the same gradient as eq .
[ eq1 ] at this point .
the gradient for which we switch to the exponential slope has to be parametrized as a function of the fragment mass @xmath9 .
as for the moment only the @xmath2ni data exhibit this exponential trend ( the measured cross sections in the @xmath76sn region reach the same cross section level , but are further away from the proton drip line ) , we tried to adjust the function @xmath81 in such a way to reproduce the @xmath2ni data without deteriorating the good agreement with the measured @xmath77xe data .
the function which fulfills reasonably well these criteria is the following : @xmath82 the result for the @xmath2ni fragment - yield distributions is visualized in fig . [ fig_ni58 ] .
the full line shows the new epax cross sections including the above modification of the slopes , whereas the omission of the exponential slope leads to the dotted cross - section curve . as a consequence of this modification , epax
version 2 predicts a production cross section e.g. for the doubly - magic nucleus @xmath1ni of @xmath83 b , which is within reach of an experiment . as stated earlier , the present adjustment for very proton - rich projectile fragments
is only based on results from @xmath2ni fragmentation @xcite .
therefore , caution is advisable when applying this parametrization especially to very light proton - rich fragmentation products
. it would be interesting to compare the present parametrization to fragmention yields at the drip line from other proton - rich beams , e.g. from @xmath84ar or @xmath85kr . in the @xmath76sn region ,
the present transition to an exponential slope has only a slight influence on the @xmath76sn cross - section prediction from @xmath50sn fragmentation ( @xmath86 b , to be compared to an experimental value of @xmath87 b @xcite )
. it might be interesting to measure fragmentation cross sections for even more proton - rich nuclei in this mass region to compare to the modified epax formula .
besides the comparison between experimental data and epax predictions for individual elemental distributions , we tried also to visualize in a more global manner the overall quality of the new epax parametrization compared to the data and the improvements with respect to epax version 1 . for this purpose ,
we plot the cross section ratios @xmath88 for all projectile - target combinations used in the present paper in fig .
[ quality ] . for most of the projectiles
the logarithm of the cross - section ratio is centered around zero indicating an agreement on average between experiment and prediction .
for the new epax version 2 , the sigma widths of these distributions vary between 0.5 ( in the case of @xmath0ar and @xmath3kr ) and 0.2 ( in the case of @xmath4xe ) .
if we analyse all experimental data together , we obtain a sigma of 0.4 .
this demonstrates that the new epax formula can predict cross - sections in most cases within a factor of two . in almost all cases
we observe a significantly smaller sigma for the new epax parametrization than for the old version .
this is particularly striking in the cases of @xmath2ni , @xmath4xe , and @xmath5pb . only in the case of @xmath1ca
the agreement between experimental data and predictions deteriorates slightly , exhibiting now a shift to smaller predicted cross sections .
it is likely that the memory effect for neutron - rich projectiles is the origin of this discrepancy . as has been mentioned in the introduction and discussed in detail in ref .
@xcite , other empirical parametrizations have been proposed , mainly for proton - induced spallation reactions ( e.g. refs .
contrary to our formula , these parametrizations fit also the energy dependence of the cross sections . the former approach has been extended to describe heavy - ion induced spallation reactions by scaling proton - induced cross sections by an energy - dependent factor @xcite and has achieved good agreement with measured cross sections ( in the 100 mb to 1 mb range ) for fragments from medium - mass nuclei ( i.e. for projectiles up to @xmath89fe , ref .
@xcite ) . for fragmentation cross sections from heavy - mass nuclei , however , the tsao _
_ formula @xcite is less successful .
this can be demonstrated e.g. by comparing in fig .
[ fig_silv_tsao ] their prediction for the pt isotope distribution in the reaction @xmath5pb + @xmath90cu to experimental data @xcite .
as we have mentioned in the introduction , our physical understanding of high - energy heavy - ion reactions suggests that proton - induced reactions produce more neutron - rich isotope distributions of heavy elements than heavy - ion - induced reactions , therefore it is unlikely that even a better fit of the tsao _
_ formula to the data in fig .
[ fig_silv_tsao ] would apply also to the same isotope distribution formed in the p+@xmath5pb reaction .
we believe that separate parameter sets have to be fitted to the respective experimental data . for @xmath81 @xmath9
gev protons impinging on heavy targets , the recent work of refs .
@xcite provides for the first time a comprehensive data set that allows to extend previous work @xcite to heavier fragments .
a complete parametrization of the bombarding - energy dependence , however , has to await more measurements at different energies for the heavy systems .
the parameters of the new epax formula have been adjusted by fitting data from stable projectiles . as explained above , one of the most difficult tasks of the present work was to reasonably parametrize the cross sections for fragments close to the projectile .
if the projectile is very far from the line of @xmath26 stability , this task is even more difficult .
the comparisons between epax predictions and experimental data shown in the present paper indicate that the chosen parametrization is reasonable
. a possible check of its predictive power could be to compare the epax formula to other data from neutron - rich or neutron - deficient projectiles .
however , only very few other high - quality data are available . as one check
, we compare in fig .
[ exotics ] the fragmentation of a stable @xmath91 mg beam @xcite to a radioactive @xmath92 mg beam @xcite and of a stable @xmath0ar beam @xcite to a radioactive @xmath93ar beam @xcite .
the elemental distributions for mg , na , and ne isotopes shown in the left - hand column and those for s , p , and si isotopes shown in the right - hand column nicely show the experimental memory effect and the high quality of the epax predictions . in a similar way we compared the results from fragmentation of @xmath94ru and @xmath94zr @xcite to epax predictions and found reasonable agreement .
however , for projectile beams still further away from stability , the memory effect becomes more and more important .
therefore , we think that some caution should be applied when using epax predictions for projectiles very far from stability like e.g. @xmath57sn . here , no experimental data at all exist to verify the epax predictions .
we have demonstrated that the quality with which epax reproduces measured high - energy fragmentation cross sections could be improved considerably by introducing rather small modifications to the formula .
this makes epax a more reliable tool to predict production rates for secondary - beam experiments with medium- or heavy - mass exotic nuclei . even after introducing these improvements , there are still some discrepancies with measured data which deserve attention in future modifications of the epax formula .
one aspect concerns the odd - even effects which were taken into account by von egidy and schmidt @xcite . at present
, we do not think that such a rather small modification ( of the order of 20 - 30% ) is necessary in view of the overall discrepancies of factors around 2 still observed , but it may become necessary once better precision can be achieved .
another open question is the change in slope observed for very neutron - deficient @xmath2ni fragmentation products . here
it would be desirable to measure systematically the formation cross sections for fragments close to the proton drip line for other medium - mass neutron - deficient projectiles , e.g. @xmath84ar and @xmath85kr . up to now
, we did not use medium - energy data ( from reactions at less then @xmath8100 @xmath9 mev ) which are available from ganil , riken , or msu .
first of all , these data probably do not fulfill the condition of `` limiting fragmentation '' mentioned above .
in fact , it is well known that at these energies nucleon - exchange reactions as well as pick - up reactions become increasingly important and thus alter in particular the cross sections close to the projectile . a striking example for
this is the difference in experimental cross sections for @xmath76sn production from a @xmath50sn beam of roughly a factor of 100 in recent ganil and gsi experiments @xcite .
in addition , these intermediate - energy data are measured with separators having very low transmissions ( of the order of few percent ) which are difficult to measure experimentally .
it would be interesting , however , to compare epax also to cross sections obtained at lower incident energies once high - quality data become available .
a fortran program of the epax formula as well as additional graphs comparing experimental data to epax predictions can be downloaded from the following address : ftp://ftpcenbg.in2p3.fr/pub/nex/epax .
we wish to thank our colleagues jose benlliure , roman gernhuser , manuel de jong , akira ozawa , jrg reinhold , robert schneider , shuying wan , and martin weber for providing their experimental cross sections prior to publication .
one of us ( k.s . ) acknowledges the kind hospitality of the centre detudes nuclaires at bordeaux - gradignan where part of this work was completed . | new experimental data obtained mainly at the gsi / frs facility allow to modify the empirical parametrization of fragmentation cross sections , epax .
it will be shown that minor modifications of the parameters lead to a much better reproduction of measured cross sections .
the most significant changes refer to the description of fragmentation yields close to the projectile and of the memory effect of neutron - deficient projectiles . |
stochastic dynamics is useful to model many biological , chemical , economical and physical systems .
the random driving forces have very different origins , in most cases they are gaussian white or colored noise sources , but often these forces must be considered as non - gaussian ones , for example , in sensory and biological systems @xcite . moreover , in many physical and biological systems a deviation of real statistics of fluctuations from gaussian law , leading to anomalous diffusion , is observed @xcite . a suitable mathematical model to describe the anomalous diffusion is the fractional fokker - planck equation , which can be derived from different theoretical approaches @xcite . in this paper
we show that the increments of generalized wiener process , representing the integral of non - gaussian white noise , have the properties of infinitely divisible random processes .
this enables us to obtain the characteristic functional of non - gaussian white noise .
then , by using functional approach to split the correlation between stochastic functionals , we derive the kolmogorov s equation directly from langevin equation with non - gaussian white noise source . from this general equation
we can obtain the fokker - planck equation for continuous markovian processes , the kolmogorov - feller equation for discontinuous markovian processes and the fractional fokker - planck equation for anomalous diffusion .
it is well known that wiener process @xmath0 has stationary and independent increments , and its time derivative @xmath1 is gaussian white noise .
the correlation function of the increments @xmath2 depends on time difference @xmath3 only , and increments of non - overlapping time intervals are statistically independent .
let us consider now a non - gaussian random process with stationary and independent increments which can be called as _ generalized wiener process_. we can represent the increment @xmath4 as @xmath5 , \label{incr}\ ] ] where @xmath6 . since the random variable @xmath7 can be divided into the sum of arbitrary number of independent and identically distributed random variables , its probability distribution belongs to the class of infinitely divisible distributions .
hence , we can express the logarithm of characteristic function of the random value @xmath7 in the lvy khinchine form @xcite @xmath8 where @xmath9 is the canonical measure which , according to definition , takes positive values on a finite intervals , and the integrals @xmath10 converge for any positive @xmath11 and @xmath3 . introducing the non - negative density @xmath12 of the canonical measure : @xmath13 , defined in generalized sense , we can rewrite eq .
( [ levy - khinchine ] ) as @xmath14 for two consecutive time intervals @xmath3 and @xmath15 , we have @xmath16 and @xmath17 because @xmath18 is an arbitrary parameter , we obtain from eqs .
( [ levy ] ) and ( [ fert ] ) @xmath19 the differentiable solution of eq .
( [ gamel ] ) , regarding parameter @xmath3 , is only linear one @xmath20 so , from eq .
( [ levy ] ) we obtain @xmath21 where the kernel @xmath22 .
the time derivative of generalized wiener process @xmath23 is a stationary random process and has analogy to gaussian white noise .
now we derive the characteristic functional of this non - gaussian delta - correlated noise . by definition , we have @xmath24 & = & \left\langle \exp\left\ { i\int\nolimits_{0 } ^{t}\xi\left ( \tau\right ) u\left ( \tau\right ) d\tau\right\ } \right\rangle = \left\langle \exp\left\ { i \int\nolimits_{0}^{t}u\left ( \tau\right ) d\eta\left ( \tau\right ) \right\ } \right\rangle \nonumber \\ & = & \left\langle \exp\left\ { i\lim\limits_{\delta_{\tau}\rightarrow0 } \sum\limits_{k=1}^{n}u\left ( \vartheta_{k}\right ) \left [ \eta\left ( \tau_{k}\right ) -\eta\left ( \tau_{k-1}\right ) \right ] \right\ } \right\rangle \label{stil } \\ & = & \lim\limits_{\delta_{\tau}\rightarrow0}\left\langle \prod\limits_{k=1 } ^{n}\exp\left\ { iu\left ( \vartheta_{k}\right ) \left [ \eta\left ( \tau _ { k}\right ) -\eta\left ( \tau_{k-1}\right ) \right ] \right \ } \right\rangle = \lim\limits_{\delta_{\tau}\rightarrow0 } \prod\limits_{k=1}^{n}\theta_{\delta\tau_{k}}\left ( u\left ( \vartheta _ { k}\right ) \right ) , \nonumber\end{aligned}\ ] ] where @xmath25 is some internal point of time interval @xmath26 , @xmath27 , @xmath28 ( @xmath29 , @xmath30 ) , and @xmath31 is the characteristic function of increments . to obtain eq .
( [ stil ] ) we used the statistical independence of increments of generalized wiener process @xmath32 . further from eqs .
( [ ch - f ] ) and ( [ stil ] ) we obtain @xmath33 & = & \lim\limits_{\delta_{\tau}\rightarrow0}\prod \limits_{k=1}^{n}\exp\left\ { \delta\tau_{k}\int\nolimits_{-\infty}^{\infty } \frac{e^{iu\left ( \vartheta_{k}\right ) x}-1-iu\left ( \vartheta_{k}\right ) \sin x}{x^{2}}\,\rho\left ( x\right ) dx\right\ } \nonumber \\ & = & \exp\left\ { \lim\limits_{\delta_{\tau}\rightarrow0}\sum\limits_{k=1}% ^{n}\delta\tau_{k}\int\nolimits_{-\infty}^{\infty}\frac{e^{iu\left ( \vartheta_{k}\right ) x}-1-iu\left ( \vartheta_{k}\right ) \sin x}{x^{2}}% \,\rho\left ( x\right ) dx\right\ } \nonumber \\ & = & \exp\left\ { \int\nolimits_{0}^{t}d\tau \int\nolimits_{-\infty}^{\infty}\frac{e^{iu\left ( \tau\right ) x}-1-iu\left ( \tau\right ) \sin x}{x^{2}}\,\rho\left ( x\right ) dx\right\ } .
\label{main}\end{aligned}\ ] ] now we derive a useful correlation formula for stochastic functionals of non - gaussian white noise .
we use the generalization of furutsu - novikov formula @xcite , to split the correlation between a stochastic functional @xmath34 $ ] of non - gaussian random process @xmath35 , defined on the observation interval @xmath36 , and the process @xmath37 itself @xmath38 \right\rangle = \left .
\frac{\dot \phi _ t\left [ u\right ] } { iu\left ( t\right ) } % \right| _ { \,u=\frac{\delta } { i\delta z}}\left\langle r_t\left [ \xi + z\right ] \right\rangle . \label{kly}\ ] ] here @xmath39 is an arbitrary deterministic function , and @xmath40 = \ln \theta _ t\left [ u\right ] $ ] . from eq .
( [ main ] ) we obtain the following expression for variational operator in eq .
( [ kly ] ) @xmath41 } { iu\left ( t\right ) } = \int \nolimits_{-\infty}^{\infty}\frac{e^{iu\left ( t\right ) x}-1-iu\left ( t\right ) \sin x}{iu\left ( t\right ) x^{2}}\rho\left ( x\right ) dx=\int \nolimits_{-\infty}^{\infty } \frac{\rho\left ( x\right ) } { x^{2}}\,dx \int\nolimits_{0}^{x}[e^{iu\left ( t\right ) y}-\cos y]dy.\ ] ] substituting this expression in eq .
( [ kly ] ) we arrive at @xmath42 \right\rangle = \int\nolimits_{-\infty}^{\infty}\frac{\rho\left ( x\right ) } { x^{2}}\,\,dx\int\nolimits_{0}^{x}\left [ \exp\left\ { y\frac{\delta}{\delta z\left ( t\right ) } \right\ } -\cos y\right ] \left\langle r_{t}\left [ \xi+z\right ] \right\rangle dy.\ ] ] by inserting the operator of functional differentiation into the average and by putting @xmath43 , we get finally @xmath44 \right\rangle = \int\nolimits_{-\infty}^{\infty}\frac{\rho\left ( x\right ) } { % x^{2}}\,dx\int\nolimits_{0}^{x}\left [ \left\langle \exp\left\ { y\frac{\delta}{% \delta\xi\left ( t\right ) } \right\ } r_{t}\left [ \xi\right ] \right\rangle -\left\langle r_{t}\left [ \xi\right ] \right\rangle \cos y\right ] dy .
\label{new}\ ] ]
let us consider now the langevin equation with a non - gaussian white noise @xmath45 @xmath46 by differentiating with respect to time the following expression for probability density of the random process @xmath47 @xmath48 and taking into account eq .
( [ lang ] ) , we obtain @xmath49 -\frac{\partial}{\partial x}\,\,g\left ( x , t\right ) \left\langle \xi\left ( t\right ) \delta\left ( x - x\left ( t\right ) \right ) \right\rangle . \label{prel}\ ] ] by using functional differentiation rules and following the same procedure of ref .
@xcite , from eq .
( [ lang ] ) we get @xmath50 thus , the operator @xmath51 of functional differentiation with respect to the function @xmath52 is equivalent to the ordinary differential operator @xmath53 . using eq .
( [ new ] ) in eq .
( [ prel ] ) and taking into account eq .
( [ equi ] ) , we obtain the kolmogorov s equation for nonlinear system ( [ lang ] ) driven by non - gaussian white noise @xmath54 } { \partial x}+\int\nolimits_{-\infty } ^\infty \frac{\rho \left ( z\right ) } { z^2}\left [ \exp \left\ { -z\frac { \partial } { \partial x}g\left ( x , t\right ) \right\ } -1+\sin z\frac { \partial } { \partial x}g\left ( x , t\right ) \right ] dzw , \label{gene}\ ] ] which is the main result of our paper . by series expansion of the exponential operator in eq .
( [ gene ] ) , we have @xmath55 + \sum\limits_{n=1}^{\infty}\frac{\left ( -1\right ) ^{n}a_{n}}{n!}\left [ \frac{\partial}{\partial x}\,\,g\left ( x , t\right ) \right ] ^{n}w , \label{han}\ ] ] where @xmath56 the equation ( [ han ] ) is valid if all integrals in eq .
( [ a - n ] ) converge .
it coincides with the result obtained in ref .
@xcite for poissonian white noise @xmath57 , if we interpret the coefficients @xmath58 as the moments of random jumps of @xmath47 .
moreover , this equation can be recast in the kramers - moyal form @xmath59 , \label{k - m}\ ] ] where @xmath60 , @xmath61 are the kinetic coefficients .
following ref . @xcite we analyze different kernel functions @xmath62 to obtain particular cases of kolmogorov s equation ( [ gene ] ) ,
related to different non - gaussian noise sources .
\(a ) as a first simple case we consider a gaussian white noise @xmath57 .
the corresponding kernel function is @xmath63 .
therefore we obtain the ordinary fokker - planck equation @xmath64 + d\frac \partial { \partial x}\,\,g\left ( x , t\right ) \frac \partial { \partial x}\,\,g\left ( x , t\right ) w. \label{fp}\ ] ] ( b ) for poisson kernel @xmath65 , eq .
( [ gene ] ) becomes @xmath66 w + b\sum\limits_{n=2}^\infty \frac{\left ( -1\right ) ^nq^{n-2}}{n!}\left [ \frac \partial { \partial x}g\left ( x , t\right ) \right ] ^nw .
\label{pois}\ ] ] ( c ) when the increment @xmath67 of generalized wiener process has the gamma distribution with exponent @xmath68 , the kernel is @xmath69 @xmath70 . as a result ,
after integration we find from eq .
( [ gene ] ) @xmath71 w -\mu \ln \left [ 1+\frac 1a\,\frac \partial { \partial x}\,\,g\left ( x , t\right ) \right ] w.\nonumber\ ] ] ( d ) for one - side @xmath72-stable lvy distributions the kernel function is @xmath73 @xmath74 , and we get @xmath55\ ] ] @xmath75 w\left ( x , t\right ) .\ ] ] after rearrangements we obtain @xmath54}{\partial x } + \frac{\rho_{0}}{\alpha}\,\frac{\partial } { \partial x}\,\,g \left ( x , t\right ) \int\nolimits_{0}^{\infty}\frac{dy}{y^{\alpha}}\left [ \cos y-\exp\left\ { -y% \frac{\partial } { \partial x}\,\,g\left ( x , t\right ) \right\ } \right]w . \label{frac}\ ] ] this equation is the fractional fokker - planck equation , and describes anomalous diffusion in the form of asymmetric lvy flights .
\(e ) for additive driving noise @xmath37 , @xmath76 in eq .
( [ lang ] ) , and the exponential operator in eq .
( [ gene ] ) reduces to the space shift operator . as a result
, we find @xmath77 + \int\nolimits_{-\infty } ^\infty \frac{\rho \left ( z\right ) } { z^2}\left [ w\left ( x - z , t\right ) -w\left ( x , t\right ) + \sin z% \frac{\partial w\left ( x , t\right ) } { \partial x}\right ] dz . \label{add}\ ] ] equation ( [ add ] ) is similar to the kolmogorov - feller equation for purely discontinuous markovian processes @xcite @xmath78 where @xmath79 is probability density of jumps step , and @xmath80 is the constant mean rate of jumps . putting @xmath81 and comparing eq .
( [ add ] ) with eq .
( [ kol ] ) we obtain the kernel function for this case @xmath82 for non - gaussian driving force @xmath83 , with symmetric @xmath72-stable lvy distribution , the kernel function is @xmath84 . as a result , eq .
( [ add ] ) takes the following form @xmath85 + d\int\nolimits_{-\infty}^{\infty}\frac{% w\left ( z , t\right ) -w\left ( x , t\right ) } { \left\vert x - z\right\vert ^{\alpha+1}}\,\,dz \label{anom diff}\ ] ] and describes the anomalous diffusion in form of symmetric lvy flights .
\(f ) finally , for multiplicative noise @xmath37 with @xmath86 , the operator in eq .
( [ gene ] ) reduces to a scaling operator , and the generalized kolmogorov s equation becomes @xmath87\,w + \int\nolimits_{-\infty } ^\infty \frac{\rho \left ( z\right ) } { z^2}\left [ w\left ( e^{-z}x , t\right ) -w\left ( x , t\right ) \right ] dz , \label{param}\ ] ] where @xmath88
now we derive the stationary probability distribution @xmath89 for some simple cases of anomalous diffusion in a fixed potential @xmath90 and in overdamped regime ( see also @xcite ) .
the langevin equation for the particle displacement @xmath47 reads @xmath91 where @xmath83 is a symmetric @xmath72-stable lvy process .
the corresponding kolmogorov s equation is eq .
( [ anom diff ] ) , with @xmath92 .
the stationary probability distribution @xmath89 can be found from the following equation @xmath93 + d\int\nolimits_{-\infty } ^\infty \frac{w_\infty \left ( z\right ) -w_\infty \left ( x\right ) } { \left| x - z\right| ^{\alpha + 1}}\,\,dz=0 .
\label{spd}\ ] ] a ) for uniform potential well with two reflecting boundaries at @xmath94 we get , from eq .
( [ spd ] ) , the uniform stationary probability density @xmath95 as for ordinary brownian diffusion .
\b ) for linear system with parabolic potential @xmath96 , by fourier transform of eq . ( [ spd ] )
, we obtain the asymptotic characteristic function of the process @xmath47 as @xcite @xmath97 c ) for symmetric potential @xmath98 , after fourier transform of eq .
( [ spd ] ) we find @xmath99 where @xmath100 functions @xmath101 and @xmath102 are connected by hilbert formula @xmath103 where @xmath104 means the principal value of integral .
after substitution of eq .
( [ hilbert ] ) in eq .
( [ v - shape ] ) we obtain finally the following integral equation for cosine fourier transform of stationary probability density @xmath105 d ) for quartic potential @xmath106 , from eq .
( [ spd ] ) we obtain the following differential equation for stationary characteristic function @xmath107 equations ( [ last ] ) and ( [ quartic ] ) can be solved numerically .
starting from the _ generalized wiener process _ , whose increments have the probability distribution belonging to the class of infinitely divisible distributions , we obtained the new correlation formula for stochastic functionals .
this allows us to derive the kolmogorov s equation ( [ gene ] ) for diffusion , induced by non - gaussian white noise source , directly from langevin equation . the fokker - planck equation for ordinary brownian motion , the kolmogorov - feller equation for purely discontinuous markovian processes and the fractional fokker - planck equation for anomalous diffusion are obtained as particular cases of the kolmogorov s equation ( [ gene ] ) .
this equation can be used for determining the stationary probability distributions , probability distributions of first - passage times and other statistical characteristics of nonlinear dynamical systems .
this work has been supported by intas grant 2001 - 0450 , miur , infm , by russian foundation for basic research ( project 02 - 02 - 17517 ) , by federal program `` scientific schools of russia '' ( project 1729.2003.2 ) , and by scientific program `` universities of russia '' ( project 01.01.020 ) . | we show that the increments of generalized wiener process , useful to describe non - gaussian white noise sources , have the properties of infinitely divisible random processes . using functional approach and the new correlation formula for non - gaussian white noise we derive directly from langevin equation , with such a random source , the kolmogorov s equation for markovian non - gaussian process . from this equation
we obtain the fokker - planck equation for nonlinear system driven by white gaussian noise , the kolmogorov - feller equation for discontinuous markovian processes , and the fractional fokker - planck equation for anomalous diffusion .
the stationary probability distributions for some simple cases of anomalous diffusion are derived . |
it appears as though the universe is expanding at an accelerating rate , as has been deduced from luminosity distance measurements of type ia supernovae , which appear dimmer than one would expect based on general relativity without a cosmological constant @xcite .
this acceleration has also been deduced from measurements of the current matter density @xmath8 , which is too small to close the universe as required by cosmic microwave background radiation ( cmb ) observations with @xmath9 priors from hst @xcite .
explanations for this discrepancy have been put forward , but most employ a modification of general relativity on cosmological scales or the addition of an exotic dark energy " field . there have also been attempts to explain this seemingly anomalous cosmic acceleration as a consequence of subhorizon inhomogeneity , rather than modified gravity or dark energy .
a perturbation is referred to as subhorizon " if its wavelength is small compared to the hubble length : @xmath10 .
it has been suggested that small - scale density perturbations could cause the appearance of accelerated expansion without the need to introduce any form of dark energy , which is an appealing prospect @xcite .
the fact that inhomogeneity can systematically modify our interpretation of cosmological measurements was first realized by ellis , who called it the fitting problem " @xcite .
the basic idea is this : due to the nonlinearity of the einstein equation , the operators for taking spatial averages and for time evolution do not commute .
this means that , although our universe is homogeneous in the mean , it will likely not have the same time evolution as that of the corresponding homogeneous universe .
nevertheless , we routinely fit distance data to frw models , a procedure that introduces errors into the inferred properties of our universe , and these errors will be present even for very large samples of type ia supernovae . our goal in this paper is to calculate the lowest order fitting effect by calculating the cosmological constant density @xmath0 that one would deduce from a perturbed luminosity distance - redshift relation @xmath11 . if we treat cosmological fluctuations perturbatively and as a random process as suggested by the
fair sample hypothesis " @xcite , then this fitting effect should be fundamentally nonlinear in the density contrast @xmath12 , requiring that we work to at least second order in @xmath13 .
this is because the ensemble averages of first order quantities vanish .
we model observations out to some moderate redshift @xmath14 . within the corresponding comoving spherical region
, the hubble flow velocity @xmath15 is bounded above by z_max ~0.1 , allowing us to use post - newtonian expansions .
there are two different velocity scales that occur , the hubble flow velocity @xmath15 and the peculiar velocity @xmath16 .
the corresponding dimensionless small parameters are _ h=~z_max~0.1 and _ p=~ ( ) , where @xmath17 is the wavelength of the dominant perturbation mode . in our computation
, we will treat both of these parameters as being of formally the same order , and we will denote both by @xmath18 " for book keeping purposes . at the end of our computation
we can identify terms that scale as @xmath19 for different values of @xmath20 and @xmath21 . as mentioned above
, we also expand separately in the fractional density perturbation @xmath13 .
we will compute redshifts @xmath22 and luminosity distances @xmath23 as functions of the affine parameter @xmath24 to third order in @xmath18 and to second order in @xmath13 . combining these results to eliminate @xmath24 will yield @xmath25 as a function of @xmath26 . using this expansion method
, we find that the lowest order inhomogeneity - induced correction to the luminosity distance scales as @xmath27 .
we then fit this relation to what one would expect from a homogeneous cosmological model which contains dust with a density @xmath28 and a cosmological constant with a density @xmath0 , d_l(z)= , by maximizing a likelihood function . here
@xmath29 for a flat universe , @xmath30 for an open universe , and @xmath31 for a closed universe .
we find that the result for the cosmological constant density is dependent on the size of the redshift range for which we have supernova data .
these results are summarized in figure [ omega ] .
for data from @xmath3 out to a limiting redshift @xmath4 , we find that the best - fit cosmological constant density is @xmath5 , and @xmath0 tends to get larger as @xmath6 gets smaller .
the best - fit @xmath32 also becomes larger as @xmath33 becomes smaller , since @xmath34 becomes large on small scales .
although this ensemble averaged result is still quite small , we find that the variance can be @xmath35 for a sample of 100 supernovae out to a redshift @xmath36 .
one implication of these results is that precision measurements of the cosmological constant from nearby supernova data require that we measure @xmath11 over a large enough redshift range , with a large enough sample .
one could also try to correct for some of the effects of inhomogeneity , using available information about large scale structure and about our own peculiar velocity @xcite .
the analysis presented here is more realistic than similar analyses within the context of simplified models of structure formation , such as the spherically symmetric lematre - tolman - bondi ( ltb ) models @xcite , swiss cheese models @xcite and their variants @xcite .
this is because we look at the full three dimensional problem , and assume that there are no bulk flows on cosmological length scales .
there have also been analyses of the perturbations to the luminosity distance - redshift relation that go to newtonian order @xcite , that only consider superhorizon perturbation modes @xcite , and that use taylor expansions of the luminosity distance @xcite , which are most appropriate for long - wavelength perturbations .
in contrast , we go to post - newtonian order , we only consider subhorizon modes , and we fit to frw models , so that we may fully address the fitting problem " .
our analysis is also fundamentally different from those in refs .
@xcite , as we choose a different method for obtaining averaged expansion parameters .
these authors average the expansion rate over a constant time slice , whereas we choose to calculate only observable quantities , namely the luminosity distance and the redshift , along the past light cone of the observer .
we then combine these expressions into @xmath37 , average over viewing angles and ensemble average , and then fit the results to what one would expect in a homogeneous model containing dust and a cosmological constant to find the best - fit value for @xmath0 .
this approach better simulates the process of gathering and analyzing supernova data , and it leads to a different result with a stable perturbative expansion . refs .
@xcite base their characterization of the expansion rate of the universe on quantities that are not related to how observers have deduced the existence of dark energy . in these papers
, perturbations are spatially averaged over a constant time slice .
such a spatial average is somewhat arbitrary , as it is dependent on the choice of spatial hypersurface .
this is in contrast to the observable significance of @xmath11 .
@xcite also use the synchronous gauge for their calculations , wherein there are metric perturbations of order @xmath13 . since @xmath38 on small scales , this gauge is particularly ill suited to perturbation theory .
in contrast , in newtonian - type gauges the metric perturbation is of order @xmath39 .
we explore this difference in section vi .
the organization of the paper is as follows : in section ii below , we introduce our coordinate choice , wherein we recast the friedmann - robertson - walker ( frw ) metric as an expansion around flat space , and in section iii we present the fundamental post - newtonian optics equations that we will need for this calculation .
we then explain our method of computation and calculate the necessary unperturbed quantities in section iv . here
we also compute the luminosity distances and redshifts for a perturbed matter dominated universe , finding @xmath26 and @xmath40 to second order in @xmath13 and to third order in @xmath18 , and we find that we may write the lowest order correction to @xmath11 in terms of the peculiar velocity field .
we then fit to a homogeneous model in section v to find the best - fit @xmath0 and its variance .
the detailed redshift and luminosity distance equations are in appendix a , the necessary results of second order perturbation theory are reviewed in appendix b , and the averaging is discussed in appendix c. then , in section vi we discuss the previous results in the synchronous gauge and show that one can choose coordinates and a definition of acceleration " such that it appears as though there could be a larger fitting effect .
we argue that such a result would be unphysical . a detailed discussion of transforming to synchronous coordinates is given in appendix d. finally , in section vii we make our concluding remarks . as usual ,
greek indices will be summed over all four spacetime dimensions while latin indices will be summed only over the three spatial dimensions .
we will also write 3-vectors in boldface and put arrows over 4-vectors .
in general , certain coordinate choices allow us to conveniently recast the local metric as an expansion around flat space , as was first emphasized for the frw metric by peebles @xcite
. we will take advantage of such an expansion so that we may use the standard post - newtonian formalism for this calculation . starting with the usual frw metric with @xmath41 , ds^2=-d^2+a^2()(d^2+^2d^2 )
, we can define the new radial coordinate = a()so that the line element becomes ds^2=-(1-h^2 ^ 2)d^2 - 2hdd+d^2+^2d^2 , where the hubble parameter of a flat and dust - dominated frw universe is @xmath42 ; we will specialize to this case for the remainder of this paper .
now we change coordinates to the standard post - newtonian gauge . in this gauge ,
the metric to first post - newtonian order can be written as ds^2=g_dx^dx^=-(1 + 2 + 2 ^ 2)dt^2 + 2_idx^idt+(1 - 2)_ijdx^idx^j , [ basicmetric ] where @xmath43 is a flat spatial metric , the potential @xmath44 contains both newtonian and post - newtonian pieces , @xmath45 is the usual gravitomagnetic potential , and 3+**=0 is the gauge condition . achieving this form for the metric entails transforming from @xmath46 and @xmath47 to @xmath48 and @xmath49 , defined by = tand = r . then the line element becomes ds^2=-dt^2+drdt+(dr^2+r^2d^2 )
[ metricbackground ] to the necessary order in @xmath49 .
this metric is of the post - newtonian form ( [ basicmetric ] ) if we define _ ( 0)=+ and _
r(0)= . here
( 0 ) " denote unperturbed , background quantities ; we will add cosmological perturbations in subsequent sections .
the unperturbed density in the new coordinates is _ ( 0)= , and the continuity equation tells us that the unperturbed 3-velocity must be of the form @xmath50 , where v_(0)= , and where @xmath51 .
thus , we see that counting orders of @xmath52 is equivalent to counting orders of @xmath53 in these coordinates
. our coordinate choice and expansion method also have the consequence that the analysis of this paper is only valid for small redshifts .
in general in the standard post - newtonian gauge , the connection coefficients are ^t_tt= , [ gamma1 ] ^t_ti=_,i , ^t_ij=-_ij-_(i|j ) , ^i_tt=^ij_,j , ^i_tj=-^i _
j+^ik_[k|j ] , and ^i_jk=^i_jk-_,k^i _ j-_,j^i _
k+_,l^il_jk , [ gammalast ] to the necessary order in @xmath18 , where @xmath54 is the connection associated with the flat spatial metric @xmath43 , which we will choose to be that of standard spherical coordinates @xmath55 , as in ref .
vertical bars represent covariant derivatives with respect to @xmath43 .
we will also need the ricci tensor components r_tt=^2and r_ij=^2_ij .
furthermore , the first post - newtonian hydrodynamic and einstein equations are + * * = 0 , + ( * v*)*v*=-**(+2 ^ 2)--(**)+3 + 4*v*(*v*)-v^2 * * , ^2=4(1 + 2v^2 - 2 ) , and ^2**=16 + * * , in this gauge . the 3-velocity @xmath56 is related to the 4-velocity @xmath57 of the fluid by = ( u^t , u^i)(1,v^i ) , where demanding that @xmath58 yields ^2=1+v^2 - 2 + 2 ^ 2 - 6v^2+v^4 + 2 * * .
[ gammafactor ]
in this section we will review how to compute luminosity distances and redshifts in a general spacetime , as in refs . @xcite .
our analysis is initially similar to that of ref .
@xcite , although they eventually rely on taylor expansions around the observer s location .
such expansions are sensible for long - wavelength perturbations , but not for the short - wavelength perturbations that we consider here .
we focus attention on a particular observer at some event @xmath59 . in our application
to perturbed frw spacetimes , this observer will be at @xmath60 and at @xmath61 for some fixed @xmath62 .
we consider the congruence of geodesics forming this observer s past light cone .
given the connection , we then find ray trajectories @xmath63 by noting that the 4-momentum is @xmath64 , and by using the geodesic equation = k^_k^=-^_k^k^ , [ nullgeo ] where we have defined @xmath65 . here
the affine parameter @xmath24 is chosen such that @xmath66 at the observer and @xmath67 at the source .
we also note that the 4-momentum is null .
the expansion @xmath68 of the congruence of null rays is related to the area @xmath69 of a bundle of rays by = .
we can find @xmath68 by using the raychadhuri equation = -r_k^k^-^2 - 2||^2 , [ raychaud ] where we have defined the shear of the congruence ||^2= , and where we require @xmath70 as @xmath71 , so that the area of the beam goes to zero at @xmath66 .
the shear @xmath72 is given by the differential equation = -+c_k^k^|t^|t^ , [ shear ] where @xmath73 is the weyl tensor , and we have defined a null newman - penrose tetrad composed of the real 4-vectors @xmath74 and @xmath75 , and the complex conjugate 4-vectors @xmath76 and @xmath77 .
these satisfy the orthogonality conditions k^m_=|t^t_=1 and k^k_=m^m_=t^t_=k^t_=m^t_=0 , as in @xcite .
they are chosen at the observer and then extended along each geodesic in the congruence by parallel transport .
we also choose the initial condition @xmath78 at @xmath66 .
once we find @xmath68 , we then find the luminosity distance as a function of the affine parameter at the source , d_l(_s)&=&_0 + & = & -_s(1+z)^2[dlinitial ] where @xmath79 corresponds to the size of the observer s telescope , which we set to zero . the right hand side of eq .
( [ dlinitial ] ) has a well defined , finite , limit as @xmath80 due to the aforementioned initial condition placed on @xmath68 .
note also that the right hand side has an overall minus sign due to our convention that the affine parameter is negative .
the redshift observed at @xmath66 , of the light emitted from the source at @xmath81 , is 1+z(_s)= , [ zinitial ] where u_k^=(g_ttk^t+g_tik^i+g_itv^ik^t+g_ijv^ik^j ) , and where the subscript s " will in general denote quantities evaluated at the source at the emission time and the subscript
o " will denote quantities evaluated at the observer at the observation time . by combining eqs .
( [ dlinitial ] ) and ( [ zinitial ] ) we can , in principle , compute @xmath25 as a function of @xmath26 in a general spacetime .
now we specialize the results of the preceding subsection to a perturbed frw metric in the post - newtonian gauge ( [ basicmetric ] ) .
our goal is to find both @xmath40 and @xmath26 to order @xmath82 . at the observer ,
we have chosen @xmath60 and @xmath61 and we have normalized the 4-momentum such that @xmath83 .
this implies that @xmath84 and @xmath85 to lowest order
. we will thus need to find the right hand side of eq .
( [ dlinitial ] ) to order @xmath86 so that we may find @xmath40 to order @xmath82 .
because of this , we see that we will need the integral in the exponential to order @xmath87 , and therefore we will need to find @xmath88 to order @xmath87 . similarly ,
inspection of eq .
( [ zinitial ] ) tells us to what post - newtonian order we will need to compute the components of @xmath89 .
to lowest order , @xmath90 , @xmath91 , @xmath92 , @xmath93 , and @xmath94 , and therefore we will need @xmath95 to order @xmath82 and we will need the spatial components @xmath96 to order @xmath87
. the post - newtonian pieces of @xmath89 must be as small or smaller than order @xmath87 , as can be seen by noting that @xmath97 in the null geodesic equation ( [ nullgeo ] ) . given this assumption and the normalization of @xmath89 , eq .
( [ nullgeo ] ) reduces to = -+o()=-^_tt+2^_tr-^_rr+o ( ) .
plugging in the connection coefficients from eq .
( [ gamma1])-([gammalast ] ) , we find = 2_,r+_r , r+o ( ) , = o ( ) , ( r k^)=-_,+o ( ) , and ( r k^)=-_,+o ( ) . using the specified initial conditions , the solutions to these equations are k^t=1 - 2-_r-2_0^rdr+o(^4 ) , [ kt ] k^r=-1+o(^3 ) , [ kr ] k^=_0^r_,+o(^3 ) , [ ktheta ] and k^=_0^r_,+o(^3 ) ; [ kphi ] the integrals above are performed along the unperturbed ray , where @xmath98 and @xmath99 .
we can then find the perturbed ray trajectory by integrating eqs .
( [ kt])-([kphi ] ) with respect to @xmath24 .
most notably , eq . ( [ kr ] ) leads to @xmath100 .
this means that we can easily rewrite eq .
( [ dlinitial ] ) in terms of the radial coordinate @xmath49 of the source : d_l = r(1+z)^2+o(r^3 ) .
[ dlagain ] in order to find the expansion @xmath68 , we first need to find the shear , given by eq .
( [ shear ] ) .
the solution to this equation is = _ 0^()^2c_k^k^|t^|t^d ; since @xmath101 , it turns out that the lowest order shear is @xmath102 . inserting @xmath103 into the raychaudhuri equation ( [ raychaud ] ) gives a contribution of order @xmath104 to the expansion @xmath68 .
however , we already know that we only need @xmath68 to order @xmath105 , and so this contribution is negligible for our purposes here . neglecting shear and defining @xmath106 , we rewrite eq .
( [ raychaud ] ) as & = & -r_tt - r_rr-()+o ( ) + & = & -2 ^ 2-()+o ( ) .
the solution to this is = ^r_0(r)^2 ^ 2dr+o ( ) , [ deltatheta ] where we are using @xmath100 . using this result in eq .
( [ dlagain ] ) yields our final result for the post - newtonian luminosity distance d_l = r(1+z)^2+o(r^3 ) .
[ dlend ] we now turn to evaluating the redshift @xmath26 as a function of the affine parameter @xmath24 . equation ( [ zinitial ] ) is the general expression for the redshift , and it depends on @xmath107 at the source and at the observer . to order @xmath82 , using eqs .
( [ basicmetric ] ) , ( [ gammafactor ] ) , and our solutions for @xmath89 , we obtain u_k^&=&g_u^k^ + & = & -1-v^r - v^2++3v^r - v^rv^2 + 2_0^rdr+v_k^+v_k^+o(^4 ) , where @xmath108 and @xmath109 are given by eqs .
( [ ktheta ] ) and ( [ kphi ] ) , respectively .
therefore , the post - newtonian redshift is 1+z&= & + & = & 1+v_s^r - v_o^r+_o-_s+(v_s^2-v_o^2)+(v_o^r)^2-v_o^rv_s^r-2^r_0dr+(v_k^+v_k^)_o-(v_k^+v_k^)_s + & + & _ ov_o^r+_sv_o^r+_ov_s^r-3_sv_s^r - v_o^2(v_s^r - v_o^r)+(v_o^r)^2(v_s^r - v_o^r)+v_s^2(v_s^r - v_o^r)+o(^4 ) .
[ z ] in eqs .
( [ dlend ] ) and ( [ z ] ) , the right hand sides are evaluated at @xmath110 and @xmath111 .
recall that subscripts o " denote quantities evaluated at the observer where @xmath60 and @xmath61 , while subscripts s " denote quantities evaluated at the source @xmath112 .
in this section we apply the formalism of section iii to a spherical region in a perturbed frw spacetime .
we will describe that region using the post - newtonian metric ( [ basicmetric ] ) .
we expand the metric functions @xmath44 and @xmath113 and the fluid 3-velocity @xmath114 in powers of the density contrast @xmath13 as = _ ( 0)+_(1)+_(2)+o(^3 ) , _
i=_i(0)+_i(1)+_i(2)+o(^3 ) , and v^i = v^i_(0)+v^i_(1)+v^i_(2)+o(^3 ) , respectively .
we also expand the null geodesic @xmath115 and 4-momentum @xmath64 as x^=x_(0)^+x_(1)^+x_(2)^+o(^3 ) and k^=k_(0)^+k_(1)^+k_(2)^+o(^3 ) , respectively .
for the remainder of the paper , quantities that are zeroth order in @xmath13 will be denoted by a subscript ( 0 ) " , first order by a subscript ( 1 ) " , and second order by a subscript ( 2 ) " . also henceforth
first order " and second order " will always refer to orders in @xmath13 , not @xmath18 , unless otherwise specified . in the perturbed spacetime
, we will calculate the redshift @xmath26 and luminosity distance @xmath25 as functions of the observation time @xmath62 , of the affine parameter @xmath24 along the past - directed null geodesic , and of the 4-momentum @xmath116 of photons at @xmath60 and @xmath61 .
we parameterize this future - directed null vector @xmath116 in terms of angles @xmath68 and @xmath117 , in such a way that @xmath83 and @xmath116 is in the direction @xmath118 at @xmath60 .
we can thus express @xmath25 and @xmath26 as functions of @xmath24 , @xmath68 , and @xmath117 at fixed @xmath62 , and by eliminating the affine parameter @xmath24 we can compute @xmath37 .
we can then take an average over angles to find @xmath11 , where we must take some care since there are two sets of relevant angles .
there are the angles @xmath119 which parameterize the direction of @xmath116 in the observer s rest frame , and then there are the coordinate angles @xmath118
. we will need to average over @xmath119 .
this means that we will need to know the relationship between the related infinitesimal solid angles @xmath120 and @xmath121 .
we define cartesian coordinates @xmath122 in terms of the polar coordinates @xmath55 in the standard way .
an orthonormal set of basis vectors for the observer s local lorentz frame can be obtained by renormalizing the coordinate basis vectors @xmath123 and @xmath124 and boosting .
the result is _
t=+ [ basist ] and _
[ basisi ] the angles @xmath118 are defined by = k^t - n^i , [ coordk ] with =( , , ) , while the observer s angles @xmath119 are defined by _
t-^i_i , [ orthok ] with =( , , ) . by inserting ( [ basist ] ) and ( [ basisi ] ) into ( [ orthok ] ) and then comparing with ( [ coordk ] )
, we find + _
o**-*v*_o+(*v*_o)*v*_o+o(^3 ) .
this gives d^2=d^2 .
[ jacobian ] after averaging over viewing angles , we find the expected value of @xmath11 by taking an ensemble average , wherein we treat the density perturbation @xmath13 at any fixed time as a homogeneous random process .
once we have the averaged @xmath11 , we can then analyze these data in terms of a homogeneous universe to see if we would find an apparent acceleration .
assuming gaussian uncertainties , we perform a chi - squared fit to a frw model with a matter density @xmath28 and a cosmological constant density @xmath0 . in the unperturbed background , everything is spherically symmetric , and the line element in our coordinates is given by eq .
( [ metricbackground ] ) .
the background four - momentum @xmath125 is purely in the @xmath126 plane , and is given by eqs .
( [ kt ] ) and ( [ kr ] ) to be k_(0)^t(r , t)=1-++o ( ) and k_(0)^r(r , t)=-1-+o ( ) . since @xmath127 and @xmath128 , we can integrate and invert these equations to find the unperturbed ray trajectory ; keeping in mind the conditions that @xmath129 and @xmath61 at the observer , we find t()=t_0+and r()=-in the unperturbed background . using this , we can use the solution ( [ deltatheta ] ) to the raychaudhuri equation to find the background expansion @xmath130 , _
( 0)()=-+o ( )
. then the zeroth - order luminosity distance is given by eq .
( [ dlend ] ) to be d_l(0)&=&(1+z)^2r(1-)+o ( ) + & = & ( 1+z)^2 , [ dlzero ] where we have defined @xmath131 .
the zeroth - order redshift is found from eq .
( [ z ] ) , z_(0)=+++o ( ) , [ zzero ] and we eventually find the expected @xmath132 by inverting eq .
( [ zzero ] ) and plugging the result into eq .
( [ dlzero ] ) : d_l(0)(z)= .
[ dlzero2 ] thus , for the background , the best - fit cosmological constant density is @xmath133 and the deceleration parameter is @xmath134 .
the perturbed post-1-newtonian line element is , from eq .
( [ basicmetric ] ) , ds^2&-&(1 + 2_(0)+2_(0)^2 + 2_(1)+4_(0)_(1)+2_(2)+2_(1)^2 + 4_(0)_(2))dt^2 + & + & 2(_i(0)+_i(1)+_i(2))dx^idt+(1 - 2_(0)-2_(1)-2_(2))_ijdx^idx^j and the perturbed luminosity distance ( [ dlend ] ) is defined to be d_l=(1+z)^2 e_l , where e_l(r,,)&=&e_l(0)(r,,)+e_l(1)(r,,)+e_l(2)(r,,)+o(^3 ) + & = & r+o(^3 ) , and where we have pulled out the factor of @xmath135 for simplicity .
we then find that the order @xmath13 perturbation is e_l(1)=-^r_0^r_0(r)^2 ^ 2_(1 ) dr
, [ dlfirst ] and the order @xmath136 perturbation is e_l(2)=-^r_0^r_0(r)^2 ^ 2_(2 ) dr , [ dlsecond ] where @xmath131 . in general , all of the terms involving potentials and velocities in these equations , and in those that follow , are evaluated along the zeroth - order , unperturbed , geodesic .
we can now calculate the perturbed redshift z(r,,)=z_(0)(r,,)+z_(1)(r,,)+z_(2)(r,,)+o(^3 ) from eq .
( [ z ] ) , using our knowledge of the zeroth - order quantities , to find z_(1)&=&v^r_s(1)-v^r_o(1)+_o(1)-_s(1)+(v^r_s(1)-v^r_o(1))-2^r_0_(1)dr+_o(1 ) + & -&_s(1)-v^r_o(1)+v^r_s(1)+o(^4 ) [ zfirst ] and z_(2)&=&v^r_s(2)-v^r_o(2)+_o(2)-_s(2)+(v^r_s(2)-v^r_o(2))+(v_s(1)^2-v_o(1)^2)+(v^r_o(1))^2 + & -&v^r_o(1)v^r_s(1)-2^r_0_(2)dr+(v_(1)k_(1)^+v_(1)k_(1)^)_o-(v_(1)k_(1)^+v_(1)k_(1)^)_s+_o(2 ) + & -&_s(2)-v^r_o(2)+v^r_s(2)++ + & + & _ o(1)v^r_o(1)+_s(1)v^r_o(1)+_o(1)v^r_s(1)-3_s(1)v^r_s(1)+x_(1)^iv^r_s(1),i+o(^2 ^ 4 ) , [ zsecond ] where the first order perturbation to the null geodesic is x_(1)^i=-_0^rk_(1)^i dr .
all of the quantities above are evaluated along the zeroth - order geodesic , and the integrals are performed along an unperturbed central ray where @xmath99 and @xmath98 .
now we have found the redshift @xmath26 and luminosity distance @xmath40 as functions of affine parameter @xmath24 and initial 4-momentum @xmath137 , to second order in @xmath13 and to third order in @xmath18 . adding the redshift equations ( [ zzero ] ) , ( [ zfirst ] ) , and ( [ zsecond ] ) yields @xmath138 .
similarly , the luminosity distance @xmath139 is found from adding eqs .
( [ dlzero2 ] ) , ( [ dlfirst ] ) , and ( [ dlsecond ] ) , after replacing the factors of @xmath135 . inverting @xmath138 perturbatively , in terms of either @xmath13 or @xmath18 , gives us @xmath24 as a function of @xmath26 .
plugging this into @xmath139 yields an expression for @xmath37 .
we then angle average this and then ensemble average , assuming that density fluctuations at a given cosmic time are a homogeneous random process .
details of this full procedure are given in appendices a , b , and c , and the result is d_l(z)=(1+z - z^2)+d_l(z ) , [ finalanswer ] where @xmath140 depends on the two point correlation function .
we will only need the lowest order piece of this , which is d_l(z)=-f()v^2_o(1 ) + o ( ) , [ deltalow ] using @xmath141 to lowest order .
the function @xmath142 is related to the velocity two point correlation function ( see appendix c ) : f(r)=-1 , [ feqn ] where @xmath143 is a unit vector that defines the viewing direction and @xmath144 is an arbitrary location in space . note that @xmath145 is independent of time , even though @xmath146 does depend on time .
this is because the time dependences of the numerator and denominator cancel .
the perturbation to the luminosity distance is proportional to v_o(1)^2 = ( _ o(1 ) ) ^2 ; this qualitative scaling has been argued for in refs . @xcite and @xcite .
we can fourier transform @xmath147 , in terms of a wavevector @xmath96 ( not to be confused with the previously - defined 4-momentum ) @xcite , _ ( 1 ) = _ * k*e^i*k * [ ft ] so that we may write the average of @xmath148 as a sum over modes : ( _ ( 1))^2 = h_0 ^ 4_0^^2(k ) , [ fourier ] where @xmath149 is the dimensionless power spectrum of matter density fluctuations at the present time , defined by ^2=_-^d(k)(k)^2 .
[ deltaeqn ] we adopt the following power spectrum ^2(k)=c^2()^4t^2 ( ) , [ hz ] where the factor of @xmath150 reflects a harrison - zeldovich flat spectrum , the amplitude @xmath151 is set by observations , and @xmath152 is the transfer function .
the bbks transfer function @xcite is a good fit for @xmath153 in the absence of dark energy , t(y)=^-1/4 , [ transfer ] where y== .
[ yeqn ] here we show the most general form of the transfer function , where @xmath154 ( not to be confused with the expansion @xmath68 ) is the density of relativistic particles divided by the density of photons , @xmath155 is the density of cold dark matter , and @xmath156 .
we choose @xmath157 for our analysis . using this spectrum , ( _ o(1))^2 = ^_0 ydyt^2(y ) , where @xmath158 and the integral is approximately @xmath159 , using the transfer function in eq .
( [ transfer ] ) .
so we finally find ( _ o(1))^2 910 ^ -6h_0 ^ 2 ^ 2 ^ -1 and therefore v^2_o(1 ) 310 ^ -6 ^ 2 ^ -1 .
using the power spectrum ( [ hz ] ) we also find ( * r*_0,t)*n*(*r*_0+r*n*,t ) = _ 0^ydyt^2(y ) , [ nvnv ] where @xmath160 and @xmath161 are spherical bessel functions of the first kind , defined in eqs .
( [ bess0 ] ) and ( [ bess2 ] ) of appendix c. we plot @xmath162 , found by combining eqs .
( [ feqn ] ) and ( [ nvnv ] ) , in figure [ twopoint ] .
note that this becomes negative for @xmath163 .
note also that we have not used any truncation of the power on scales that are nonlinear .
if we instead were to impose a high-@xmath164 cutoff , so as not to include the effects of any modes that have @xmath165 , then this would lead to differences of a factor of about two .
a different approach would be to include the quasi - linear regime , with the power spectrum given from n - body simulations @xcite .
we will specialize to @xmath166 for the rest of this paper , which yields v^2_o(1 ) 8.3410 ^ -6 . in figure [ perts ] ,
we show how the perturbation @xmath140 scales relative to the unperturbed luminosity distance @xmath132 , for the choice @xmath167 .
note that we are plotting the logarithm of the absolute value , as the perturbation changes sign from positive to negative as one looks at larger distances . by inspection , it becomes clear that @xmath140 is not actually a perturbation for very small redshifts , i.e. for where @xmath168 , and thus our computation of @xmath169 is no longer valid in that regime .
indeed , it is well known that the peculiar velocities of objects within the local supercluster are not small when compared to their redshifts . however
, this will not be a problem in practice , as type ia supernovae at such small redshifts are typically not used for cosmological parameter fitting .
we will eventually take this breakdown of perturbation theory into account by imposing a lower cutoff @xmath33 when we fit our data to a theoretical model . by eye
, we see that it should be safe to choose @xmath170 .
we may now find the inferred cosmological constant and deceleration parameter by analyzing eq .
( [ finalanswer ] ) within the context of what one would expect in a homogeneous model .
the lowest order perturbation to the luminosity distance depends on the difference between the peculiar velocities at the source and at the observer , and so the question that we now ask is : how do peculiar velocities and their correlations affect inferences drawn from data about cosmological models ? we can not simply taylor expand eq .
( [ finalanswer ] ) around the observer to find @xmath171 .
this is because @xmath142 varies on short lengthscales of order @xmath172 , so that a taylor series expansion would effectively mean computing @xmath171 from @xmath11 within this unrealistically short lengthscale .
a good alternative then is to fit the perturbed luminosity distance over a finite range of redshifts to what one would expect in a homogeneous model with matter and a cosmological constant .
suppose that the observer can measure redshifts @xmath173 for a set of distant objects arbitrarily well .
from the distance determinations @xmath174 , the observer can compute @xmath175 , and we can therefore take @xmath176 to be the data gathered by the observer .
suppose also that in actuality the universe is spatially flat with hubble parameter @xmath9 and matter only .
let r_i = h_0 ^ -1be the physical value of @xmath177 , where for a flat matter - only cosmology f(z_i)=_0^z_idz(1+z)^3/2=2[feqn ] and @xmath178 ( not to be confused with the matter perturbation power spectrum ) is the non - frw contribution to @xmath177 , from fluctuations via velocity differences . from eq .
( [ deltalow ] ) , we find the ensemble averaged perturbation ( z_i)&&_0^dyt^2(y ) .
[ deltafull ] the observer fits the data to a frw model that is slightly curved and has a small cosmological constant .
the fitted model is then r_i^fit=_0^z_idzh(z)-k6(_0^z_idzh(z))^3 , where @xmath179 and h^2(z)&=&^2 + & = & ^2(1+z)^3 ; here @xmath180 is the fitted hubble parameter , and @xmath181 and @xmath182 are the density parameters for matter and for the cosmological constant , respectively .
let us work to first order in @xmath183 and @xmath182 , a simplification which ought to suffice as long as @xmath184 .
thus , the fitted model is r_i^fit&=&^-1 + & & ^-1 , where @xmath185 is the same as before , and we have defined g(z)=12_0^zdzz(1+z)^5/2 + 16[f(z)]^3 [ geqn ] and i(z)=12_0^zdzz(2+z)(1+z)^9/2 -16[f(z)]^3 .
[ ieqn ] there are three fitting parameters : @xmath180 , @xmath186 and @xmath187 . from the data and our model we can compute a likelihood function .
assuming gaussian uncertainties this will be the exponential of & = & -12_i^2 + & = & -12_i^2 , where @xmath188 is the estimated uncertainty in the value of @xmath177 inferred from observations and @xmath189 for @xmath190 .
the next step is to maximize @xmath191 with respect to the parameters of the fit , which will lead to a set of coupled nonlinear equations .
to simplify , let us linearize in the small parameters @xmath192 , @xmath193 , @xmath194 and @xmath195 .
the resulting equations are = + -h , = + -h , and = + -h , where we have defined the average @xmath196 . solving for the parameters of the fit , we get & = & d^-1 , [ epsleqn ] & = & d^-1 , [ epsmeqn ] and h&=&-d^-1 , [ heqn ] where = ^2 - 2-+^2 + ^2 .
[ deqn ] these are fairly general for small @xmath197 , and show that there may be contributions to @xmath193 , @xmath192 , and @xmath198 from velocity fluctuations .
next , we need to compute the averages . to do this
, we recall that @xmath199 corresponds to comoving radial coordinate , modulo a factor of @xmath200 . to the order of approximation underlying our calculations
, we can take the comoving source density to be uniform .
moreover , we do not need to worry about malmquist bias , at least for type ia supernovae , which are very bright .
let us also assume that all of the @xmath201 are the same , to keep the problem as simple as possible .
then @xmath202 drops out of our expressions for @xmath192 , @xmath193 , and @xmath198 , although it remains in their uncertainties .
we suppose that our source catalog extends to some maximum value @xmath203 , with a corresponding maximum redshift @xmath6 .
it is worth remembering that @xmath204 is an absolute upper bound , and that for @xmath205 , @xmath206 , so we will be dealing with relatively small values of @xmath199 typically .
moreover , as we have already noted in figure [ perts ] , our small @xmath207 assumption breaks down below a minimum redshift @xmath208 , but this is not a problem as no supernovae below this redshift have ever been used for cosmological model fitting @xcite .
so we will assume a lower cutoff for all of our sums of @xmath209 .
then , for example , = _ f_min^f_max dff^2f^2 , and eqs .
( [ feqn ] ) , ( [ geqn ] ) , ( [ ieqn ] ) , and ( [ deqn ] ) give the lowest order result , assuming that @xmath210 , -f_max^12 . keeping only lowest order terms in @xmath203 in the numerators of eqs .
( [ epsleqn ] ) , ( [ epsmeqn ] ) , and ( [ heqn ] ) as well , we get - , [ epsleqn2 ] -2 , [ correspondence ] and h- .
we see that if @xmath211 , then @xmath193 is zero , because the three terms in eq .
( [ epsleqn2 ] ) cancel .
this means that if @xmath197 arises from velocity correlations , it is only the correlation function of velocities at two separated points that matters , not the rms velocity at a point .
also note that , for this fitting procedure , the deceleration parameter is still @xmath134 , since q_0&=&q_0 - 12= -12-(a a / h^2)_0 + & = & 12(-1 - 2)= 12(--2)=12(2 - 2)=0 from eq .
( [ correspondence ] ) , in agreement with refs .
@xcite , @xcite , and @xcite .
the perturbation @xmath197 , given in eq .
( [ deltafull ] ) , depends on the correlation function @xmath145 , and so it does contribute to @xmath193 . for @xmath3 and @xmath4 , we numerically integrate to find that the best - fit cosmological constant density is @xmath5 .
table [ lambda ] gives a few more results for the best - fit values for @xmath193 , @xmath192 , and @xmath198 as a function of the two limiting redshifts @xmath33 and @xmath6 in the continuum limit , where we have made the assumption that the number of sources @xmath212 is very large : @xmath213 . in this limit , @xmath214 and = _
f_min^f_maxdf w(f)(f ) , where we have the weighting function w(f)-(-+ ) .
we also plot these results in figure [ omega ] , in the introduction .
note that @xmath0 may be positive or negative , depending on the redshift range , since @xmath169 changes sign in the region of interest .
.best - fit parameters in the continuum limit for a few values of the source catalog limiting redshifts @xmath33 and @xmath6 , also for the choice that the dominant perturbation wavelength is @xmath215 times smaller than the hubble scale : @xmath167 . [ cols="^,^,^,^,^",options="header " , ] [ lambda2 ] although the best - fit values for @xmath0 of the previous subsection are very small , we must keep in mind that they are derived from the ensemble averaged perturbation to the luminosity distance . for a given source
, this ensemble averaged perturbation will be far smaller than the leading order perturbation , which depends linearly on the peculiar velocity .
this linear perturbation will be the main source of the variance in the best - fit parameters , and this variance should overwhelm the systematic error for typical supernova sample sizes .
this complication was pointed out by ref . @xcite and it was shown to cause errors of @xmath216 for a sample of actual nearby supernovae in ref .
@xcite .
consider our expression for the best - fit @xmath0 , in terms of n discrete sources , rewritten as a weighted sum , _ = _
i w(f_i)_i .
what we have computed is the ensemble average of this , _
i w(f_i)_i .
the variance is then _
^2&=&(_-_)^2 = _
^2+o(^3 ) + & = & _ i , jw(f_i)w(f_j)_i_j , which has two types of terms contributing : those with @xmath217 and those with @xmath218 .
separating these , we have @xmath219 , where _ 1 ^ 2_iw^2(f_i)_i^2 and _ 2 ^ 2_ijw(f_i)w(f_j)_i_j .
[ var2 ] in the continuum limit @xmath213 , the first piece of the variance becomes _
1 ^ 2_0^f_maxf^2dfw^2(f)^2(f ) [ contvar ] where , from eq .
( [ deltafull ] ) , ^2(f ) = ^2(h_0 r ) = ~v_o^2 .
the integrand in eq .
( [ contvar ] ) is integrable as @xmath220 , and so the quantity @xmath221 is to a good approximation independent of @xmath33 for small @xmath33 .
thus we can for simplicity take @xmath222 . after integrating ,
we find _ 1 ^ 2~()()^-6 .
[ var1 ] for a source catalog of 100 sources out to a limiting redshift @xmath223 , we find that this variance is significant : @xmath224
. the second piece ( [ var2 ] ) of the variance does not depend on the sample size , although it does depend on @xmath203 . in the continuum limit , _ 2 ^ 2_0^f_maxf^2dfw(f)_0^f_max(f)^2dfw(f)(f)(f ) [ sigma2 ] where ( f)(f ) = v_o(1)^2 .
[ deltaffp ] plugging eq .
( [ deltaffp ] ) into eq .
( [ sigma2 ] ) , then using eqs .
( [ feqn ] ) and ( [ nvnv ] ) , and then finally doing some rearranging , we find _ 2 ^ 2()^2_0^yt^2(y)^2 where i(q)_0 ^ 1dx(x - x^2+x^3)(qx - qxqx ) .
this result for @xmath225 does not depend on the sample size , as it only depends on the size of the redshift range @xmath203 , making it a measure of cosmic variance . by integrating numerically , we find that it scales roughly as @xmath226 and _ 2 ^ 2~0.03()^-8 .
[ var2 ] for comparison , ref .
@xcite uses a sample of 115 supernovae up to a redshift @xmath227 , and they find an error from the data of @xmath228 . for this same scenario , we estimate @xmath229 , from the sum of eqs .
( [ var1 ] ) and ( [ var2 ] ) .
the method of analysis that we have presented in the previous sections differs from that of refs .
this is because of ( i ) a difference in gauge choice and ( ii ) a fundamental difference in the definition of what constitutes acceleration " .
we have chosen to use the standard post - newtonian gauge , and to define acceleration as being based on fitting the luminosity distance - redshift relation to that of a homogeneous model containing dust and a cosmological constant . as this definition of acceleration
is based only on observable quantities , performing our calculation in other gauges gives us the same results .
in contrast , refs .
@xcite calculate the cosmological expansion rate , averaged over a constant time slice .
the motivation for doing this comes from the spatially - averaged friedmann equations , also called the buchert equations @xcite .
in particular , ref .
@xcite defines the effective coarse - grained scale factor @xmath230 in terms of the average matter density : @xmath231 , where the angle brackets @xmath232 , with subscript @xmath233 , denote an average over a spatial hypersurface @xmath233 at a given time . then ref .
@xcite defines the coarse - grained hubble rate h_d==_d and the effective deceleration parameter q=--1 .
[ qkolb ] these measures of acceleration
are somewhat arbitrary since the deceleration parameter ( [ qkolb ] ) depends on the spatial hypersurface over which one averages . refs .
@xcite use constant time slices in the comoving synchronous gauge . in this gauge ,
the perturbation to the the expansion @xmath68 is related quite simply to the perturbations to the trace of the connection ; from ref .
@xcite , _ ( 1 ) _
d = ^i_ti(1 ) _ d , [ kolbtrace ] and similarly for @xmath234 . ref .
@xcite claims that spatially averaged perturbations could become quite large , which implies that our perception of the expansion rate of the universe is significantly affected by inhomogeneity .
the culprit is the appearance of terms in @xmath235 with large numbers of spatial gradients , which naturally appear in the synchronous gauge
. these higher derivative terms , which do not appear in our method above , lead to a perturbative instability , wherein terms higher order in perturbation theory do not get smaller as expected .
although the results of the previous sections appear to differ from the claims of refs .
@xcite , in fact the large fitting effect claimed in those papers arises at a higher post - newtonian order than we have computed . in this section
we show that our results are consistent with theirs to the order we have computed .
our method of computation could be extended to higher post - newtonian order , which would allow for a detailed confrontation with their claims . however ,
we believe that our result of a small fitting effect is robust , in the sense that it will not be altered by the inclusion of effects that are higher order in @xmath236 and/or @xmath13 .
this belief is based on the structure of the post - newtonian expansion of einstein s equations , and on the fact that we are computing a gauge - invariant observable . if this is true , then our conclusion is in disagreement with refs .
@xcite .
we believe the most likely reason for the disagreement is that we compute a gauge - invariant observable that is directly and uniquely related to supernova observations , whereas the quantities computed in refs . @xcite
have some arbitrariness and are not directly related to observations .
the proposal of refs .
@xcite that there might be a large backreaction effect in terms of @xmath237 does not necessarily imply that observers will measure large deviations from frw dynamics .
as mentioned above , spatially averaged perturbations are dependent on one s coordinate choice , in the sense that a constant time hypersurface in one coordinate system is most likely not going to be a constant time hypersurface in a different coordinate system .
these averages are unlikely to be directly observable , and are not uniquely related to the cosmic acceleration inferred from cosmological observations . as hirata and seljak @xcite remarked , we can not cover the entire universe with astronomers so as to measure spatially averaged quantities " such as @xmath238 .
it is possible that the measure of acceleration ( [ qkolb ] ) could be large while the observed acceleration is small .
we now turn to showing consistency of our results with those of refs .
@xcite to the order we have computed .
we take our metric ( [ basicmetric ] ) and transform it from the post - newtonian gauge to the synchronous gauge .
we then compute from the transformed metric the perturbation to the hubble rate .
the relative size of the difference between @xmath238 and the expected frw value @xmath239 determines whether or not there will be a large fitting effect . as an example
, we will now compute the ratio = [ kolb ] where the spatial average involves integrating with respect to the perturbed volume element @xmath240 , where @xmath241 is the determinant of the spatial part of the metric .
note that the quantity that we define as @xmath242 differs from what is computed in refs .
@xcite , although we do find the same qualitative result at the end of the day .
below we show that this quantity is small to newtonian order , in correspondence with what was found in @xcite , even though it involves a sum of terms that can be large individually .
the reason these terms are large is that in synchronous coordinates metric perturbations can be of order @xmath13 , which may be of considerable size even though there are no large gravitational potentials anywhere in the universe .
by contrast , in our calculation based on standard post - newtonian coordinates , metric perturbations are at most of order @xmath243 , which is always small . in this sense , perturbation expansions are much better behaved in the standard post - newtonian coordinates than in synchronous coordinates .
we start by reviewing the transformation from standard post - newtonian coordinates ( [ basicmetric ] ) to synchronous coordinates ; a detailed discussion is presented in appendix d. begin with the second order perturbed frw metric in the gauge ds^2=a^2 ( ) , where we are now using conformal and cartesian coordinates for simplicity , and we will only need to work to newtonian order . we can then define the new coordinates @xmath46 and @xmath244 by = + o(_0 ^ 4)+o(_0 ^ 3 ) and x^i=^i-_(1),i-_(2),i+_(1),ij_(1),j+o(^i^2)+o(^i^3 ) , where these potentials are fixed physical quantities , evaluated at @xmath245 , and these spatial derivatives are in terms of the new coordinates .
we are also assuming that we have the growing mode only , for which we have the power law scalings @xmath246 and @xmath247
. then the line element becomes , to lowest order in @xmath18 , ds^2&=&a^2 ( ) + & = & a^2()\{-d^2+d^id^j } , which is now in a synchronous gauge .
note that the metric now has perturbations of order @xmath248 .
these order @xmath13 perturbations will lead to the appearance of large terms in @xmath242 , which will cancel when averaged .
then we find = a^3 ( ) .
the spatial trace of the connection is ^i_i=^ij(a^2 _
ij ) _ , which receives the first and second order perturbations ^i_i(1)=a()_(1)=^ij_ij(1),=-^2_(1)+o(^2 ) and ^i_i(2)=a()_(2)&=&^ij(1)_ij(1),+^ij_ij(2 ) , + & = & -_(1),ij_(1),ij-^2_(2)+(^2_(1))_,k_(1),k+o(^2 ^ 2 ) . using the fourier transformation ( [ ft ] ) , taking an ensemble average , and using the result that @xmath249 ( see appendix c ) , we find from eqs .
( [ kolb ] ) and ( [ kolbtrace ] ) & & ( ^2_(1))^2-_(1),ij_(1),ij+(^2_(1))_,k_(1),k + & = & ( ^2_(1))^2-_(1),ij_(1),ij + & = & _ , i , [ kolb2 ] which is consistent with the lowest order result of ref . @xcite .
this spatial average is a boundary term , whose ensemble average vanishes .
although ( [ kolb2 ] ) vanishes , it contains terms with two more powers of @xmath250 than what one would find in the post - newtonian gauge .
it is these terms that refs .
@xcite argue will lead to a large effect at higher order in perturbation theory .
in other words , using the synchronous gauge and defining acceleration in terms of spatially averaged expansion parameters can lead to a conceivably large correction .
this is in contrast to our earlier method , wherein we calculate the observable effect , which is very small .
note that our expansion ( [ dllong ] ) for @xmath11 contains no four - derivative terms like those in ( [ kolb2 ] ) .
we have computed the inhomogeneity - induced perturbations to the redshifts and luminosity distances that a comoving observer would measure to first post - newtonian order , i.e. we have computed @xmath26 and @xmath40 to order @xmath251 , and to second order in the density perturbation @xmath12 . assuming a flat and matter - dominated background cosmology , the perturbed luminosity distance - redshift relation is given by eq .
( [ finalanswer ] ) . the perturbations to @xmath11 depend on the correlation between the peculiar velocities at the observer and at the source .
roughly speaking , these perturbations are of order @xmath252 when @xmath2 .
the luminosity distance - redshift relation was averaged over viewing angles and over an ensemble of realizations of the density perturbation .
the result is gauge invariant , as it corresponds to a measurable quantity .
we then fit this function to what one would expect in a homogeneous frw cosmology , containing dust and a cosmological constant , to deduce the corresponding perturbation to the inferred cosmological constant density .
the inferred @xmath0 depends on the limiting redshifts @xmath33 and @xmath6 of the sample , and we summarize the best - fit values of @xmath0 for different values of these limiting redshifts in figure [ omega ] and table [ lambda ] .
these ensemble averaged results indicate that we are justified in fitting low-@xmath26 supernova data to homogeneous models , as long as we use supernova data that spans a large enough redshift range .
for instance , assuming that we have luminosities and redshifts from @xmath3 out to @xmath253 , the errors induced by the fitting problem " are small : @xmath254 .
such errors are not large enough to explain the measured value @xmath7 .
this is what we would expect , since we have other evidence to suggest that our universe contains dark energy from large scale structure surveys , from the cmb power spectrum , and from weak lensing .
in contrast to the small value of the best - fit @xmath0 for the ensemble averaged luminosity distance - redshift relation , we find that relatively large errors are possible due to fluctuations in @xmath11 , specifically from terms that are linear in peculiar velocities .
this effect was noted in ref .
@xcite and then calculated in ref .
@xcite for an actual nearby supernova data set .
we find that the associated variance in @xmath0 has two components , one that depends on the number of sources @xmath212 , @xmath255 , and one that does not , @xmath256 .
it should be stressed that our goal in this paper was only to find a rough estimate of the fitting effect .
one potential weakness of our analysis is that we have assumed that @xmath257 , and thus we do not address the effects of highly nonlinear structures .
such nonlinear modes could be included by using the full nonlinear power spectrum from n - body simulations @xcite , and we estimate that this would change the result by approximately a factor of two .
furthermore , we have assumed that the observer is in a random location in the universe , and has no knowledge of his / her own peculiar velocity .
one can redo the calculation for an observer who knows and corrects for this velocity .
it has been claimed that there exists a perturbative instability , where successive orders in an expansion in powers of @xmath13 do not get smaller @xcite .
we do not see any indications of such an instability with our method . when one defines
acceleration " in terms of only directly observable quantities , as we did in sections ii through v , the fitting effect one obtains is small .
is supported by an american dissertation fellowship from the aauw educational foundation .
this research was supported in part by nsf grants phy-0457200 and phy-0555216 .
we also thank syksy rsnen for pointing out an error in an earlier version of this manuscript .
adding the redshift equations ( [ zzero ] ) , ( [ zfirst ] ) , and ( [ zsecond ] ) yields z(,,)&= & + & & + + & & + \{v^r_s(2)-v^r_o(2)+_o(2)-_s(2)+(v^r_s(2)-v^r_o(2))+(v_s(1)^2-v_o(1)^2)+(v^r_o(1))^2 + & & -v^r_o(1)v^r_s(1)-2^r_0_(2)dr+(v_(1)k_(1)^+v_(1)k_(1)^)_o-(v_(1)k_(1)^+v_(1)k_(1)^)_s+_o(2 ) + & & -_s(2)-v^r_o(2)+v^r_s(2)++ + & & + _
o(1)v^r_o(1)+_s(1)v^r_o(1)+_o(1)v^r_s(1)-3_s(1)v^r_s(1)+x_(1)^iv^r_s(1),i+o(^4 ^ 2)}+o(^3 ) , [ zadded ] where the right hand side is evaluated at @xmath258 and @xmath259 . to point out a few of the above effects , the terms linear in velocity and linear in @xmath44 correspond to the doppler effect and the gravitational redshift , respectively .
we also see the second order doppler shift with the @xmath260 terms , and the integrated sachs - wolfe effect with the integrated terms . the perturbed luminosity distance is found from eqs .
( [ dlzero2 ] ) , ( [ dlfirst ] ) , and ( [ dlsecond ] ) to be d_l(,,)&=&\ { + & & - + & & -+o(^3 ) } .
[ dladded ] here we can see the effects of weak gravitational lensing . note that as the cosmological portion of the redshift goes to zero , and hence @xmath261 , the luminosity distance also goes to zero , as expected . by combining eqs .
( [ zadded ] ) and ( [ dladded ] ) , we can eliminate @xmath24 and compute @xmath25 as a function of @xmath26 , @xmath68 , and @xmath117 .
this computation can be carried out explicitly by using the fact that the expressions are power series in @xmath18 and @xmath13 .
this procedure gives : d_l(z,,)&&\{z - z^2+z^3+(-1+z - z^2)(v^r_s(1)+v^r_s(2 ) ) + & & + ( 1-z+z^2)(v^r_o(1)+v^r_o(2))+(1-z)(_s(1)+_s(2))+(-1+z)(_o(1)+_o(2 ) ) + & & + ( -z)(v_o(1)^2-v_s(1)^2)+(-+z)(v^r_o(1))^2+(-+z)(v^r_s(1))^2+(-z)v^r_o(1)v^r_s(1 ) + & & + v^r_o(1)(_o(1)-_s(1))+v^r_s(1)(_s(1)-_o(1))-x_(1)^iv^r_s(1),i+^r_0(_(1)+_(2))dr + & & + ( v_(1)k_(1)^+v_(1)k_(1)^)_s-(v_(1)k_(1)^+v_(1)k_(1)^)_o-(z+v^r_o(1)-v^r_s(1))^r_0^r_0(r)^2 ^ 2_(1 ) dr + & & -z^r_0^r_0(r)^2 ^ 2_(2 ) dr+(v^r_s(1)-v^r_o(1 ) ) + & & + _ s(1 ) + & & - + & & v^r_s(1 ) } , [ dllong ] where , to leading order , @xmath262 .
the functions of @xmath49 and @xmath48 that appear on the right hand side of eq .
( [ dllong ] ) are evaluated at @xmath263 and @xmath264 .
note that the redshift @xmath26 here is the full redshift as measured by the observer .
next we need to average @xmath37 over viewing angles in the observer s rest frame , and also take an ensemble average . in doing so , the averages of first order quantities will vanish .
we also will find that we will only need the second order velocities and potentials to newtonian order , so that we may compute the lowest - order effect .
in terms of comoving coordinates @xmath265 @xcite , the equations of newtonian hydrodynamics are + = 0 , [ euler1 ] + _
p+(*v*_p)*v*_p=- , [ euler2 ] and ^2_p=4_0 a^2 , [ euler3 ] where @xmath266 is the peculiar velocity , @xmath267 is the perturbation to the newtonian gravitational potential , the density contrast is @xmath268/\rho_0(t)$ ] , and the zeroth order quantities are given in section ii .
the newtonian first order results are very well known ; for a detailed review , see peebles @xcite . for a newtonian analysis to second order in @xmath13 , see ref .
@xcite .
the first order result is that the density contrast consists of mode that grows with time , and one that decays with time : _ ( 1)(*r*,t)=f(*r*)t^2/3+g(*r*)t^-1 , [ modes ] where @xmath142 and @xmath269 are functions of the spatial coordinates .
we will only consider the growing mode .
it is useful to rewrite the hydrodynamic equations in terms of their fourier modes .
writing = _ * k*e^i*k * and _ p = _ * k*e^i*k * , [ ft ] eq . ( [ euler3 ] ) becomes k^2_*k*=4_0 a^2_*k * .
[ fpoisson ] the second order density contrast is _
( 2)=(_(1),ij_(1),j+^2_(1)_(1),i)_,i ; [ sod ] this result came from perturbing eqs .
( [ euler1])-([euler3 ] ) to second order and then solving these by using the first order solutions , eqs .
( [ modes ] ) and ( [ ft ] ) .
it can be seen that the expected value of @xmath270 vanishes by substituting the mode expansion of @xmath147 into eq .
( [ sod ] ) : @xmath271 .
we also see from eq .
( [ fpoisson ] ) that @xmath272 depends only on boundary conditions ; we can choose to add overall constants to @xmath44 at each order in @xmath13 , and it is natural to choose these constants to satisfy @xmath273 .
assuming that we only have the growing mode solution of eq .
( [ modes ] ) , we find that the first order peculiar velocity is related to the newtonian potential , _ ( 1)(*r*,t)=-_(1)=-t^1/3t_0 ^ 2/3**_(1 ) .
this averages to zero but its square does not .
the second order velocity perturbation is v^i_(2)=-_(1),ij_(1),j which also averages to zero : @xmath274 .
note that these averages are ensemble averages , not spatial averages .
now we can scrutinize the terms of eq .
( [ dllong ] ) , so that we may find their angular and ensemble averages .
note that the angular averages will be performed with respect to the observer s angles @xmath119 , and so we will need to use the jacobian given in eq .
( [ jacobian ] ) .
the first three terms of eq .
( [ dllong ] ) only depend on the background cosmology , and are unchanged after averaging , and all terms that are to first order in @xmath13 will have a vanishing ensemble average .
as shown in appendix b , terms that depend on @xmath275 and @xmath276 also average to zero .
in addition , there are many terms that have vanishing ensemble averages because they contain an odd number of spatial derivatives of the potential , such as v^r_o(1)_o(1 ) = v^r_s(1)_s(1 ) = 0 , v^r_s(1)v^r_s(1 ) = 0 , x^i_(1)v^r_s(1),i = 0 , v^r_s(1)^r_0_(1)dr = 0 , et cetera .
we also find that ( v_(1)k_(1)^+v_(1)k_(1)^)_s-(v_(1)k_(1)^+v_(1)k_(1)^)_o ~o(^4 ) , since @xmath277 , and taking the difference of the averages at the source and at the observer introduces another factor of @xmath278 .
we can further rewrite the average @xmath279 by exploiting the power law scaling @xmath280 , to find v_s(1)^2 & = & ( t_0 ^ 2/3t^1/3_(1))^2 ( t_0 ^ 2/3_(1))^2(t_0-r)^2/3 + & = & v_o(1)^2 .
we also use @xmath281 , and introduce the two point correlation function @xmath145 , v^r_s(1)v^r_o(1 ) = v_o(1)^2 ( 1-z ) , where @xmath145 is defined by ( * r*_0,t)*n*(*r*_0+r*n*,t ) = v_o(1)^2 , and @xmath143 is a unit vector that defines the viewing direction . we can write this correlation function in terms of a more general correlation function @xmath282 , using the fourier transform of eq .
( [ ft ] ) and eqs .
( [ deltaeqn])-([yeqn ] ) : v_o(1)^2c_ij(r)v_i(*r*_0,t_0)v_j(*r*_0+*r*,t_0 ) = _ 0^ . this function can be rewritten as v_o(1)^2c_ij(r)h_0 ^ 2 , where a(r)=_0^e^-i*k*=_0^and b(r)=_0^e^-i*k*=_0^ , and where we are using spherical bessel functions of the first kind : j_0(x)= [ bess0 ] and j_2(x)=(-)x - x .
[ bess2 ] it follows that v_o(1)^2&= & v_o(1)^2n^in^jc_ij(r)=h_0 ^ 2 + & = & _ 0^ydyt^2(y ) , where @xmath283 .
we plot @xmath162 in figure [ twopoint ] ; we see that it falls to approximately zero for @xmath284 , and thus we do not expect it to be important when measuring the distances to supernovae at redshifts @xmath2 .
note also that @xmath142 becomes negative for large enough @xmath49 .
using these simplifications , we finally get d_l(z)=(1+z - z^2)+d^rms_l(z)+d^corr_l(z ) , where @xmath285 is the perturbation that depends on rms quantities at a given point , which vanishes : d^rms_l(z)=0 , and @xmath286 is the perturbation that depends on @xmath142 . to subleading order ,
this is d^corr_l(z ) + o ( ) , where the subleading terms are suppressed by a factor of @xmath287 or @xmath26 .
we will only use the lowest order piece , d_l(z)&=&d_l^corr(z)-f ( ) + & = & _ 0^dyt^2(y ) .
in the standard post - newtonian gauge discussed in section ii , we can rewrite the metric in terms of conformal coordinates , ds^2=a^2 ( ) , where we will only need this to newtonian order , and now the scale factor is @xmath288 .
we will define @xmath289 to be the conformal time today .
this new time coordinate is related to that of sections ii - v by = 3()^-2/3t = t+o(t^2 ) , and the radial coordinates are related by r=()^-2/3r=+o(r^2 ) , where @xmath290 .
thus , we see that the potentials are the same as before , to newtonian order , except that they now are in terms of comoving distance @xmath291 and conformal time @xmath292 .
we also now use cartesian coordinates for simplicity .
our goal is to transform to the synchronous gauge , with new coordinates @xmath293 , where the line element has the form ds^2=a^2()_d^^=a^2 ( ) . in this gauge , @xmath294 and @xmath295 .
we make the following ansatz for the new coordinates : = + f_(1)(,)+f_(2)(,)+o(_0 ^ 4 ) [ tau ] and x^i=^i+h_(1)^i(,)+h_(2)^i(,)+o(^i^2 ) , [ tildex ] where @xmath296 , @xmath297 , @xmath298 , @xmath299 , and @xmath300 is the time today .
we are also assuming that we have the growing mode only , for which we have the power law scalings @xmath246 and @xmath247 . in order to find the new metric
, we will need the relations a^2()=a^2()and _
( 1)(,x ) + _ ( 2)(,x ) = _
( 1)(,x ) + _ ( 2)(,x ) + _ ( 1),i h^i_(1 ) + o(^4 ) + o(^3 ) .
using these and the coordinate transformations ( [ tau ] ) and ( [ tildex ] ) , we find _
= -(1+f_(1)+f_(2)+2_(1)+2_(1),ih_(1)^i+2_(2)+2_(1)+2_(2))+h_(1)^i h_(1)^i=-1 , implying f_(1)+_(1)+_(1)=0 [ diff1 ] and f_(2)+2_(1),ih_(1)^i+2_(2)+2_(2)-h_(1)^i h_(1)^i=0 .
[ diff2 ] similarly , the time - space component of the new metric is _
i =- f_(1),i - f_(2),i+_(1)^i+_(2)^i+h^j_(1),i_(1)^j+o(^3)=0 and this implies -f_(1),i+_(1)^i=0 [ diff3 ] and -f_(2),i+_(2)^i+h^j_(1),i_(1)^j=0 .
[ diff4 ] equations ( [ diff1 ] ) , ( [ diff2 ] ) , ( [ diff3 ] ) and ( [ diff4 ] ) are solved by f_(1)=-_(1)+ , f_(2)=-_(2)+(_(1))^2 + - h_0^i _
( 1),i , h_(1)^i=-_(1),i+h_0^i ( ) , and h_(2)^i=-_(2),i+_(1),ij_(1),j- _
( 1),ji h_0^j + _ ( 1),j h_0,i^j + h_0^i(x ) , where the arbitrary constants @xmath301 and @xmath302 and functions @xmath303 and @xmath304 represent residual gauge freedoms associated with synchronous coordinates . setting @xmath301 and @xmath302 to zero will give us comoving coordinates .
we can imagine comoving coordinates to be fixed on some spacelike hypersurface from which the worldlines of freely falling particles emanate .
if we set all of the clocks carried by these particles to the same time on this spacelike hypersurface , then @xmath305 . the residual functions @xmath306 and @xmath307 correspond to simply changing the coordinates on the spacelike hypersurface from which worldlines emanate , and we will set @xmath308 . using this solution for the appropriate coordinate transformation ,
we find the spatial part of the new metric to be _ ij&=&_ij - f_(1),i f_(1),j+ h_(1)i , j + h_(1)j , i + & & + h_(2)i , j + h_(2)j , i + h_(1)k , i h_(1)k , j+ + o(^4 ) + o(^3 ) + & = & _ ij-_(1),ij-_(2),ij+_(1),ijk_(1),k+_(1),ik_(1),jk+o(^2)+o(^3 ) . | we explicitly calculate the lowest order systematic inhomogeneity - induced corrections to the cosmological constant that one would infer from an analysis of the luminosities and redshifts of type ia supernovae , assuming a homogeneous universe .
the calculation entails a post - newtonian expansion within the framework of second order perturbation theory , wherein we consider the effects of subhorizon density perturbations in a flat , dust dominated universe . within this formalism ,
we calculate luminosity distances and redshifts along the past light cone of an observer .
the luminosity distance - redshift relation is then averaged over viewing angles and ensemble averaged , assuming that density fluctuations at a given cosmic time are a homogeneous random process .
the resulting relation is fit to that of a homogeneous model containing dust and a cosmological constant , in order to deduce the best - fit cosmological constant density @xmath0 .
we find that the luminosity distance - redshift relation is indeed modified , even for large sample sizes , but only by a very small fraction , of order @xmath1 for @xmath2 .
this lowest order deviation depends on the peculiar velocities of the source and the observer . however , when fitting this perturbed relation to that of a homogeneous universe , via maximizing a likelihood function , we find that the inferred cosmological constant can be surprisingly large , depending on the range of redshifts sampled . for a sample of supernovae extending from @xmath3 out to a limiting redshift @xmath4 , we find that @xmath5 .
the value of @xmath0 has a large variance , and its magnitude tends to get progressively larger as the limiting redshift @xmath6 gets smaller , implying that precision measurements of @xmath0 from nearby supernova data will require taking this effect into account .
this effect has been referred to in the past as the fitting problem " , and more recently as subhorizon backreaction " .
we find that it is likely too small to explain the observed value @xmath7 .
there have been previous claims of much larger backreaction effects .
by contrast to those calculations , our work is directly related to how observers deduce cosmological parameters from astronomical data . |
in a series of papers , de boer et al . @xcite have put forward the idea that the well - known egret excess of diffuse galactic gamma rays @xcite could be well explained by dark matter annihilations .
the idea that excess from the region near the galactic center may be due to dark matter annihilations has a long history ( at least @xcite ) but de boer et al .
have extended this idea to claim that all the diffuse galactic gamma rays detected above 1 gev by the egret satellite , irrespective of the direction , has a sizeable dark matter contribution ( for a similar , but less explicit , proposal , see @xcite ) .
de boer et al .
propose specific supersymmetric models as examples of viable models with correct relic density , and the gamma - ray fluxes are made to fit the observations .
the price they have to pay , however , is a rather peculiar dark matter halo of the milky way , containing massive , disc concentrated rings of dark matter besides the customary smooth halo .
in addition , they have to crank up the predicted gamma - ray flux by considerable boost factors " .
we will here examine these hypotheses a bit closer .
in particular , we will discuss the astrophysical properties and possible inconsistencies of the model .
we point out that , besides the strange features of the halo model ( that we judge difficult to achieve with non - dissipative dark matter ) , supersymmetric models with large gamma - ray flux are essentially always accompanied by a large antiproton flux ( see , e.g. , @xcite ) .
we investigate what the antiproton fluxes would be in the same halo model , using model - by - model the same boost factors as needed to fit the gamma - ray spectrum .
we find that low - mass models ( masses less than 100 gev ) that have low boost factors tend to overproduce antiprotons by a factor of around ten .
higher - mass models ( above a few hundred gev ) have a lower antiproton rate , so the overproduction is slightly less .
however , they give hardly any improvements to the fits to the gamma - ray spectrum .
we will perform this analysis in a general minimal supersymmetric standard model ( mssm ) , but as the correlation between gamma rays and antiprotons is a general feature , our results will be more general .
our conclusion is that the proposal of de boer et al .
@xcite to explain the gamma excess in all sky directions is , at present , not viable . although of course
we can not exclude a small contribution to the diffuse gamma - ray flux from dark matter annihilations .
gamma rays have the advantage of pointing back directly to their sources in the galaxy and not to suffer from energy losses .
this , together with known gamma - ray spectral shape from dark matter annihilation ( distinct from the conventional background ) , permit to extract the sky - projected dark matter distribution from the egret observations . taking this one step
further de boer et al .
propose a specific dark matter profile , with 18 free parameters , and do a best fit to the egret data ( for details , see @xcite ) . the density profile de boer et al .
obtain consists of a dark matter halo with the following ingredients : * a triaxial smooth halo , * an inner ring at about 4.15 kpc with a density falling off as @xmath0 , and * an outer ring at about 12.9 kpc with a density falling off as @xmath1 . where @xmath2 is the height above the galactic plane .
the triaxial halo is a modified isothermal sphere , but flattened in the direction of the earth and in the @xmath2-direction .
the inner ring is rather broad , but very closely located to the baryonic disc , with an exponential fall - off above the galactic plane .
the outer ring is more massive and slightly narrower and also exhibits an exponential fall - off above the galactic plane .
the outer ring is also taken to have a sharper fall - off on the inside than the outside .
both rings are elliptic .
the details of the parameterization of the dark matter model can be found in @xcite . in fig .
[ fig : deboer - density ] , we show the strong concentration of dark matter to the disc ( upper panel ) as well as the ring structure of the model ( lower panel ) . the steep exponential density fall - off away from the disc , originating from the two rings , can most clearly be seen in the upper - right plot . since conventional models of the diffuse gamma - ray emission employ scattering of cosmic rays on gas and dust in the galactic disc
, we may already note that this model will produce a gamma - ray flux that has an angular distribution very similar to the generally accepted cosmic ray contribution @xcite . in fact , besides the need for measurements with a better energy resolution such as will be given by glast , the desired spectral shape can also be obtained by adjusting the balance between the contributions from proton and electron cosmic rays ( whose intensity is very poorly constrained away from the solar neighborhood ) @xcite .
in @xcite it was shown that one can get a good agreement with egret data by adjusting the electron and proton injection spectra ( without any need for a dark matter contribution ) .
even though the dark matter halo profile by de boer et al
. explains the egret data very well , we will here go through some of the astrophysical objections to this model .
first , one may notice that the model of the dark matter really is very close to the `` standard '' model for the baryons of the milky way , containing a thin and a thick disc and a central bulge ( see , e.g. , @xcite ) .
since the dark halo is much more massive than the baryonic one , one of the first things one should investigate is whether there is room to place as much unseen matter in the vicinity of the disc as de boer et al .
do . by observations of the dynamics and density fall - off of stars in the disc , one can get a measure of the gravitational pull perpendicular to the galactic plane .
this in turn can be converted to an allowed disc surface mass density , a method pioneered by bahcall @xcite .
recent analyses @xcite of the disc surface mass density at the solar system location have converged to a model with little room for a concentration of dark matter in the disc .
observations are well described by a smooth dark matter halo and a disc of identified matter ( mainly containing stars , white and brown dwarfs and interstellar matter in form of cold and hot gases ) . .measured local surface densities @xmath3 , within heights @xmath4 ,
compared to the amount of dark matter in the model of de boer et al .
the amount of dark matter exceeds the allowed span for unidentified gravitational matter in the inner part of the galactic disc ( i.e. around @xmath5 ) . @xcite
[ tab : density ] [ cols="^,>,<,^,>,<,^,^ " , ] let us now focus on the de boer et al . profile .
the most important parts for the antiproton signal are the rings that decay exponentially with the height above the galactic plane .
hence , this source distribution is much more concentrated to the galactic disc than a smooth halo profile is .
we therefore expect a more modest change of the fluxes with the extreme values of the diffusion parameters than found in @xcite . as a complement to our calculations with darksusy
, we have derived the antiproton fluxes in the de boer model for the extreme diffusion parameters given in @xcite and which we recall in table [ tab : propagation ] .
this calculation has been done with the propagation code in @xcite . for illustration
, we show in fig . [
fig : pbar_prim_vs_sec ] the full yield of primary antiprotons for a selected supersymmetric configuration for which the agreement with the egret data is very good @xmath6 of order 3 ( neutralino mass of 50.1 gev and derived boost factor of 69 ) . for the sake of completeness
, we have taken into account tertiary antiprotons as in @xcite and computed the effect of diffusive reacceleration with the help of the same crank - nicholson scheme as in @xcite .
solar modulation has been modeled with the simple force
field approximation where a fisk parameter of @xmath7 mv has been assumed .
the red solid curve in fig . [ fig : pbar_prim_vs_sec ] corresponds to the median cosmic ray configuration of table [ tab : propagation ] .
the yellow band is delimited by the extreme configurations and gives an indication on how well the flux of neutralino induced antiprotons can be derived in the case of the de boer et al . dark matter distribution . for the maximal cosmic ray model
, we observe an increase in the antiproton flux by a factor of 2.5 whereas for the minimal configuration , we find a decrease of a factor of 2.6 both at energies lying between 0.4 and 0.6 gev .
the total width corresponds therefore to an overall factor of @xmath8 to be compared to a factor of @xmath9 in the case of an nfw dark matter halo .
as expected , the uncertainties are much smaller in the de boer et al model compared to more conventional halo profile .
again , the main reason for this is that the dark matter in the de boer model is located much closer to the galactic plane and its distribution is reminiscent of the gaseous disc of the milky way that is responsible for the conventional secondary background .
the latter is featured in fig .
[ fig : pbar_prim_vs_sec ] as the narrow green band that has been derived in @xcite from the observed b / c ratio . from this calculation
we also get an estimate of the possible uncertainties arising from different propagation codes . in fig .
[ fig : pbar_prim_vs_sec ] , the long dashed black curve has been computed with the darksusy package using a typical cosmic ray propagation ( as described in @xcite ) . because diffusive reacceleration has not been implemented in that case , the flux falls more steeply close the neutralino mass .
notice however that it is fairly similar to the red solid line , calculate with the propagation code in @xcite .
we can therefore conclude that diffusive reacceleration does not substantially modify the shape of the antiproton spectrum especially in the energy range of interest .
we also note that the two different propagation codes , darksusy and the code of @xcite , gives very good agreement and should not constitute any additional theoretical error . .
predictions spread over the yellow band as the cosmic ray propagation parameters are varied from the minimal to maximal configurations .
the long - dashed black curve is calculated with darksusy for a standard set of propagation parameters ( see @xcite for details ) .
the narrow green band stands for the secondary component . as is evident from this figure , the antiproton fluxes for this example model clearly overshoots the data.,scaledwidth=80.0% ] -0.3 cm notice finally that the primary yellow uncertainty band is at least an order of magnitude above the secondary green component .
the latter alone can already account for the antiproton observations as is clear in fig .
[ fig : pbar_prim_vs_sec ] .
we conclude that the supersymmetric model that has been selected here because it provides a good fit to the egret excess should be excluded as it by far overproduces antiprotons . in fig . [
fig : pbar_prim_vs_sec ] we also compare with the bess data at other energies than our previously selected energy bin .
it should be evident from the figure that the model overproduces antiprotons at essentially every measured energy .
we have here compared with one example model , but this argument can be made more general . in fig .
[ fig : deboer - pbar]b we show with a solid line how far down we could shift the models down by going to the extreme minimal model .
as can be seen , the antiprotons are still overproduced by a factor of 210 for the models with good fits to egret data .
it is therefore difficult to see how the dark matter interpretation of the egret data could be compatible with the antiproton measurements . above we mentioned the antiproton uncertainties coming from our lack of knowledge of the propagation model .
we also have uncertainties arising from the assumptions in the approach of de boer et al .
e.g. , the boost factor is assumed to be independent of position in the galaxy .
this is probably not a very well - justified assumption as we would expect the boost factor to depend on the formation history of that particular region in the galaxy .
if the boost factors do depend on the position in the galaxy , then the boost factors for the flux at _ reception _ at earth can be different for gamma rays and antiprotons since the signal in general come from different parts of the galaxy ( the gamma signal can e.g. be dominated by annihilation at the galactic center , whereas the antiprotons could come from a more nearby structure ) .
we will here investigate what the uncertainties on the boost of the antiproton flux _ as seen at the earth _ could be if we relax the assumption of space - independent boost factors .
in particular , galactic tidal interactions should destroy clumps@xcite in the inner ( older ) parts of the milky way and especially in the inner ring , which is the most prominent feature in the de boer model when it comes to the fits to the gamma - ray fluxes . on the other hand ,
the antiprotons that are detected at the earth originate from a broad region of the milky way halo .
a significant portion is produced in the inner ring , but a substantial part comes also from the outskirts where clumps have survived . the correct _ effective _ boost factor which should have been implemented in the antiproton calculations should therefore probably be larger than the value derived from assuming a spatially constant boost factor .
we can quantify these uncertainties though .
if we take away the inner ring , the antiproton fluxes goes down a factor of 2.0 .
if we take away the outer ring , the antiproton fluxes go down by a factor of 1.4 . and ,
finally , if we would take away the smooth triaxial halo , the antiproton fluxes would go down by a factor of 2.2 .
hence , we can conclude that the antiproton fluxes at the earth are dominated by annihilation in the inner ring and the smooth triaxial halo .
since the antiprotons coming from the inner ring originate in the same place as the gamma rays that dominate the flux in region a , the boost factor derived from the gamma flux should be very close to the _ effective _ boost factor for these antiprotons .
as noted above , this makes up half of the antiproton flux at earth . for the other half , coming mostly from the triaxial smooth halo , our use of the same _ effective _ boost factor as derived from gamma rays from region a is most likely not correct .
however , as noted above , we would expect the boost factor to be larger in these less dense regions , so our ( and de boer s ) assumption of space independent boost factors would lead to an underestimate of the antiproton flux . another argument
slightly more technical points in the same direction .
the _ local _ boost factor @xmath10 of a clump of mass @xmath11 is defined @xcite with respect to some value of reference @xmath12 through the relation @xmath13 where @xmath14 is the dark matter density inside the substructure .
if large boost factors are needed to explain the egret data , this means that the dark matter is tightly packed even inside the inner ring .
that region is already fairly dense with a density of reference @xmath15 gev @xmath16 . on the contrary ,
antiprotons originate from a larger domain where the neutralino density is certainly smaller on average than inside the inner ring . if the clumps have basically the same mass and density profile all over the milky way halo which is probably correct in the case where very small substructures dominate , as they should have survived the galactic tides the product @xmath17 remains constant . we are forced to the conclusion that the actual _ effective _ boost value which should be used for antiprotons is once again larger than in the case of the egret data . therefore assuming identical _ effective _ boost factors ( as seen at earth ) for the photon and antiproton flux is very conservative .
the lower antiproton limit featured in fig . [ fig : deboer - pbar]b should actually be shifted upwards , even further above the bess 98 stripes .
with the enhanced dark matter densities in the de boer et al .
model one may wonder if these models would nt have been seen already in direct detection experiments . to make a simple estimate of this , we plot the scattering cross section on protons versus mass in fig . [ fig : deboer - dir ] .
we compare with the latest cdms results @xcite , but since those results apply for a local dark matter density of @xmath18 gev/@xmath19 , we have rescaled the limits with the local density in the de boer et al . model . to be more specific ,
we have lowered the limits by @xmath20 , which is the ratio of the local dark matter density in the de boer et al .
model and a ` standard ' halo model . as can be seen
, most models would seem to be excluded by the latest cdms results @xcite .
however , in deriving these limits , a standard gaussian velocity distribution has been assumed .
this is most probably a reasonable approximation for the triaxial halo in the de boer et al .
model , but most of the local density comes from the inner ring , where not much is known about the velocity . if the inner ring would rotate along with the galaxy , we expect lower rates than indicated in the figure since the relative velocity between the dark matter and us would be lower . however , if it counter - rotates , we would expect even higher rates .
if we would be conservative and assume that the inner ring would produce no significant rates in direct detection experiments , the limits from cdms would be reduced by a factor of @xmath21 ( as the local density from the smooth halo is 0.5 gev/@xmath19 in the de boer et al .
model ) . in fig .
[ fig : deboer - dir ] we indicate this exclusion limit with a dashed line .
as can be seen , some of the models of them would then fall below the exclusion limit , but some would still be above . hence , one can conclude that the de boer et al .
models produce significant rates in direct detection experiments , and some of the good models are already excluded even if uncertainties of the velocity distributions are taken into account .
however , we can not exclude all of the good models , due to these uncertainties , but many of the models giving the best fits are well within reach with future improved direct detection experiments .
we have here investigated the idea put forward by de boer et al.that the observed egret excess of diffuse gamma rays could be due to dark matter annihilation in the milky way .
this is an interesting idea , but as far as we can see the model does not seem very plausible when other constraints are taken into account .
the strongest constraint is probably the antiproton flux , which would be overproduced far above the flux measured by bess . even if the uncertainties of the antiproton flux are included , the antiproton flux is still more than a factor of 210 above the bess measurements . from the analysis in this paper one also find ( see , fig .
[ fig : deboer - pbar]b ) that models that would be compatible with the antiproton data would always have very bad fits to the egret data ( reduced @xmath22 @xmath23 60/6 ) .
actually , the optimized background model in @xcite would produce a much better fit ( reduced @xmath22 @xmath24 22/6 ; which can also be read of at the high - mass end in fig .
[ fig : deboer - boost]b , where the signal does not contribute much to the fits )
. hence , the models with acceptable antiproton fluxes are not very interesting as they give worse fits to the gamma excess than the more conventional models .
other objections to the model come from astrophysics .
e.g. , the density in the disc in the de boer model is a little bit too high to be compatible with stellar motions in the solar neighborhood .
this could of course be circumvented by adjusting the densities in the de boer et al .
model to have a dip at our location in the solar system .
such fine tunings are not very appealing though , especially as the model already as it stands has a local minimum at the solar neighborhood .
one has to be careful interpreting these results on the disc surface density though as there are large uncertainties and model dependencies . to be conservative
, we interpret these too high disc surface mass densities as a hint of problems with the model .
the actual density profiles of the rings ( exponential fall - off away from the disc ) is also not what you expect from dark matter .
in fact , we believe that the de boer et al . model is most likely a fit of the _ baryonic _ matter distribution of the galaxy , and not the dark matter density .
one could also wonder if the background models could be improved further and provide a even better fit to the egret data .
in fact , the optimized background model in @xcite already gives a good fit to the egret data especially when conventional / larger uncertainties of for egret s observed gamma fluxes are adopted ( reduced @xmath22 @xmath24 5/6 ) .
this optimized model was accomplished by tuning the electron and proton injection spectrum without violating any of the cosmic ray constraints , such as b / c .
hence , there is no real need for dark matter to explain the egret data .
new experiments are on their way though with glast flying in 2007 and with new and better data , we will hopefully get a better understanding of the gamma - ray sky and if there are any ( clear ) signals of dark matter out there .
m.g . would like to thank c. flynn and j. holmberg for useful discussions regarding the local galactic surface density . l.b . and
j.e want to thank i. moskalenko and w. de boer for useful discussions and are grateful to the swedish science research council ( vr ) for support .
also thanks a. strong and o. reimer for discussions .
p.s . acknowledges a support from the french programme national de cosmologie pnc
. w. de boer , c. sander , v. zhukov , a. v. gladyshev and d. i. kazakov , _ the supersymmetric interpretation of the egret excess of diffuse galactic gamma rays _ , phys .
b * 636 * , 13 ( 2006 ) [ arxiv : hep - ph/0511154 ] .
a. w. strong , i. v. moskalenko and o. reimer , _ diffuse galactic continuum gamma rays . a model compatible with egret data and cosmic - ray measurements _ , astrophys .
j. * 613 * , 962 ( 2004 ) [ arxiv : astro - ph/0406254 ] .
v. i. korchagin , t. m. girard , t. v. borkova , d. i. dinescu and w. f. van altena , _ local surface density of the galactic disk from a 3-d stellar velocity sample _ , astronomical journal , * 126 * , 2896 ( 2003 ) [ arxiv : astro - ph/0308276 ] .
o. bienayme , c. soubiran , t. v. mishenina , v. v. kovtyukh and a. siebert , _ vertical distribution of galactic disc stars : iii .
the galactic disc surface mass density from red clump giants _ , arxiv : astro - ph/0510431 . | we investigate the internal consistency of the halo dark matter model which has been proposed by de boer et al .
to explain the excess of diffuse galactic gamma rays observed by the egret experiment .
any model based on dark matter annihilation into quark jets , such as the supersymmetric model proposed by de boer et al .
, inevitably also predicts a primary flux of antiprotons from the same jets . since propagation of the antiprotons in the unconventional , disk - dominated type of halo model used by de boer et al .
is strongly constrained by the measured ratio of boron to carbon nuclei in cosmic rays , we investigate the viability of the model using the darksusy package to compute the gamma - ray and antiproton fluxes .
we are able to show that their model is excluded by a wide margin from the measured flux of antiprotons .
we therefore find that a model of the type suggested by moskalenko et al . , where the intensities of protons and electrons in the cosmic rays vary with galactic position , is far more plausible to explain the gamma excess .
lapth1141/06 0.2 cm * is the dark matter interpretation of the egret gamma excess * * compatible with antiproton measurements ? *
lars bergstrm , joakim edsj , michael gustafsson _ department of physics , stockholm university , albanova university center , _ _ se-106 91 stockholm , sweden _ pierre salati _ laboratoire dannecy - le - vieux de physique thorique lapth , cnrs and universit de savoie , _
_ 9 , chemin de bellevue , b.p.110 74941 annecy - le - vieux , france _ |
in multi - user wideband wireless systems , two fundamental obstacles limit the gains in spectral efficiency : * inter - symbol interference ( isi ) : in a wideband communication system that uses a band - limited signal , isi is fundamentally unavoidable when the transmission bandwidth @xmath1 exceeds the coherence bandwidth @xmath2 of the channel .
isi distorts the signal between subsequent transmitted symbols ; this effect limits the spectral efficiency of a wideband communication system . * inter - user interference ( iui ) : in a multi - cell communication system that uses universal frequency reuse per cell , iui ( or inter - cell interference ) is also inevitable .
iui limits the spectral efficiency of the multi - cell system , because it reduces signal - to - interference plus noise ratio ( sinr ) of the users .
the problem of mitigating both isi and iui simultaneously is challenging in general .
the common approach has been to deal with the problem of mitigating them separately .
for example , orthogonal frequency division multiplexing ( ofdm)@xcite is a well - known method that successfully removes isi in single - user wideband communication systems .
the key principle of the ofdm is to change the linear convolution relation between input signals and the impulse response of the isi channel to be the circular convolution between them by adding an appropriate cyclic prefix .
then , using the inverse discrete fourier transform ( idft ) as transmit eigen - beamforming , and the discrete fourier transform ( dft ) as receive eigen - beamforming , the isi channel is decomposed into a set of orthogonal subchannels ( subcarriers ) .
essentially , this channel decomposition creates multiple parallel gaussian channels , each experiencing isi - free narrowband flat fading .
it has shown that this strategy achieves the capacity of the isi channel with a simple water - filling power allocation @xcite in an asymptotic sense , as the number of subcarriers tends to infinity .
the concept of this channel decomposition creating parallel subchannels has been extensively extended to multi - user scenarios @xcite by incorporating iui .
for instance , by allowing two transmitters to share an ofdm symbol to communicate with a receiver , the capacity region for such multiple access channel ( mac ) with isi was characterized in @xcite .
similarly , the capacity for the two - user interference channel with isi is also established , especially when iui is strong @xcite .
the common achievable scheme revealing the capacity results is the use of separate coding over each subchannel with multiuser water - filling algorithms based on game - theoretic optimization . for the general case of the @xmath0-user interference channel with isi , which can be transformed into the @xmath0-user parallel gaussian interference channel by ofdm @xcite ,
the capacity remains open .
it has shown in @xcite that the optimal sum degrees of freedom ( sum - dof ) of the parallel @xmath0-user interference channel is @xmath3 , assuming the channel coefficients are independent across subchannels .
the key to this sum - dof characterization is a novel iui management method , referred to as _ interference alignment _ ( ia ) .
in contrast to the gaussian point - to - point , multiple access , and broadcast channels with isi in which separate coding is optimal @xcite , joint encoding across subchannels has shown to provide a significant spectral efficiency gain over separate coding when signal - to - noise ratio ( snr ) is high for the parallel gaussian interference channel @xcite . despite the significant gains , ia requires global and instantaneous channel state information at transmitter ( csit ) , so
ia is too difficult to be used as a practical interference management solution in many multi - user wideband wireless systems . in the absence of knowledge about csit
, the sum - dof of numerous wireless networks is the same as what is achievable by time - division multiple access ( tdma ) among users @xcite .
recent research on ia has made progress towards using limited csit to realize the gains of ia in practical wireless systems @xcite .
blind interference alignment @xcite is a representative technique that uses the knowledge of channel coherence patterns at the transmitters instead of instantaneous channel coefficients themselves .
however , diversity of coherence intervals among different receivers can naturally occur in wireless channels ( or can be induced artificially using reconfigurable antennas @xcite ) ; by exploiting this diversity , one can achieve the same dof in an siso interference channel as that attained with perfect knowledge of csit @xcite .
the impediment to implementing these techniques is that to construct a particular super - symbol structure , the cooperative transmitters must know the channel coherence pattern information ( second - order channel statistics ) from all receivers .
due to insertion loss in rf circuits , the use of switching - reconfigurable antennas @xcite is also not preferable to current wireless systems . in this paper , we consider a single - input - single - output ( siso ) @xmath0-user interference channel with isi as illustrated in fig .
[ fig:1 ] .
it is important to notice that this channel model is not necessarily equivalent to the parallel ( or multi - carrier ) gaussian interference channel .
one can transform the interference channel with isi to the corresponding multi - carrier interference channel by using the decomposition approach by ofdm @xcite , but not necessarily vice versa . throughout the paper , we focus on assumptions that the channel coefficients are linear time - invariant ( lti ) with finite - length impulse response .
most importantly , we assume that transmitters have completely lack csit ; i.e. , that they do not even know channel coherence patterns that are essential for implementing the conventional blind interference alignment @xcite . in this
setting , _ the fundamental question we address in this paper is whether the sum - spectral efficiency can be made to scale linearly with @xmath0 , even with complete lack of csit . _
our major contribution of this paper is to show that the answer to this question is affirmative in some isi conditions .
more precisely , we demonstrate that the sum - dof of the @xmath0-user interference channel with isi is @xmath4 where @xmath5 and @xmath6 are the maximum impulse response lengths of the desired and interfering links , respectively .
for example , when @xmath7 and @xmath8 , @xmath3 sum - dof can be achieved in a channel that has isi .
this is a remarkable result because it has shown that the optimal sum - dof of the multi - carrier @xmath0-user siso interference channel , @xmath3 , is achievable only when global and perfect csit is available @xcite or the specialized channel realizations occur @xcite .
our result , however , shows that , with complete lack of csit , the sum - dof increases linearly with @xmath0 no matter how the channel coefficients realized , provided that @xmath9 . to show the achievability of this result , we introduce a novel communication method , called _ interference - free ofdm ( if - ofdm)_. the principal idea of if - ofdm is to convert the effective channel matrices of all interfering wireless links to circulant channel matrices , while keeping the non - circulant structure for the channel matrix of the desired link by judiciously choosing the cyclic prefix size .
using the fact that column vectors of an idft matrix are the common right eigenvectors of circulant matrices , each transmitter sends its information along a set of linearly independent column vectors of the idft matrix .
this strategy essentially allows aligning all the interference within the same subspace at each receiver , while the desired signals remain out of the interference subspace . by projecting the received signals onto the subspace that is orthogonal to the interference subspace ,
all iui are removed perfectly . once iui is eliminated ,
each receiver decodes the desired information symbols by removing the remaining inter - subcarrier interference . by using this method
, we also characterize an achievable sum - spectral efficiency of the @xmath0-user interference channel with isi under two csit assumptions on the desired link . for the case of completely absent csit
, we obtain a closed - form expression of the sum - spectral efficiency that achieved by the concatenation of two transmission methods : 1 ) if - ofdm for mitigating iui and 2 ) zero - forcing based successive interference cancellation ( zf - sic ) to remove inter - subcarrier interference .
furthermore , when a transmitter can learn csit from its paired receiver only , we also yield a characterization of an achievable sum - spectral efficiency of the @xmath0-user interference channel with isi . to derive this result ,
a two - stage precoding method is presented , which essentially concatenates if - ofdm to mitigate iui and singular value decomposition ( svd ) based vector coding with water - filling power allocation to reject inter - carrier interference .
one remarkable feature of the proposed method is the use of separate encoding across subchannels ; this implies that a data symbol of each transmitter is sent through a subcarrier .
this fact is both practically relevant and theoretically interesting , because it reveals that separate encoding achieves a significant spectral efficiency gain in a multi - carrier interference channel .
the rest of the paper is organized as follows .
we describe a system model of the siso @xmath0-user interference channel with isi in section ii .
we illustrate the key idea of the proposed if - ofdm through simple examples in section iii .
we present the main theorem giving the sum - dof of the siso @xmath0-user interference channel with isi in section iv .
then , in section v , we provide the characterization of achievable sum - spectral efficiency under two assumptions on the desired link s csit .
we also provide some discussion to apply applications of the proposed method under practical isi conditions in section vi .
the paper concludes with future directions in section vii . -user interference channel with isi when @xmath10 , @xmath11 , and @xmath12.,width=336 ]
we consider a @xmath0-user siso interference channel with isi . as illustrated in fig . [ fig:1 ]
, transmitter @xmath13 uses the shared spectrum to communicate with its associated receiver @xmath14 .
we assume that all transmitters and receivers are equipped with a single antenna .
let @xmath15 $ ] be the transmitting signal of transmitter @xmath14 at time slot @xmath16 with an average power constraint @xmath17|^2]=p$ ] .
we also denote the channel impulse response from transmitter @xmath18 to receiver @xmath19 by @xmath20\}_{\ell=1}^{l_{k , i}}$ ] , in which @xmath21 is the effective number of multi - paths between transmitter @xmath18 and receiver @xmath19 .
then , the received signal of receiver @xmath22 at time slot @xmath23 is @xmath24=\sum_{i=1}^k\sum_{\ell=1}^{l_{k , i}}h_{k , i}[\ell]x_{i}[n-\ell+1]+ z_k[n ] , \label{eq : received_system_model}\end{aligned}\ ] ] where @xmath25 $ ] is the noise at receiver @xmath18 during time slot @xmath23 . here , the number of channel - taps of the wireless channel from transmitter @xmath26 to receiver @xmath18 is typically defined as @xmath27 where @xmath28 is the delay spread of the channel from transmitter @xmath26 to receiver @xmath18
, @xmath1 is the transmission bandwidth of a signal , and @xmath29 is the smallest integer greater than or equal to @xmath30 . for example , if symbol duration @xmath31 is larger than delay spread @xmath28 , i.e. , @xmath32 for all @xmath33 , this model is equivalent to the @xmath0-user siso frequency - flat ( narrowband ) interference channel . in our model , we assume that fading coefficients of @xmath20\}$ ] are time - invariant for a communication duration , i.e. , a block - fading assumption . in each block , all the channel coefficients @xmath20\}$ ] are drawn from a continuous distribution . in a rich - scattering propagation environment , for instance , each channel coefficient can be modeled as an independent complex gaussian random variable with zero - mean and variance that decays exponentially as the number of taps increases , i.e. , @xmath34\simeq \mathcal{cn}(0,e^{-\beta(\ell-1)})$ ] , where @xmath35 is the constant that determines the rate of channel power reduction .
furthermore , @xmath25 $ ] are assumed to be independent and identically - distributed ( iid ) complex gaussian random variables with zero mean and variance @xmath36 , i.e. , @xmath37 . throughout the paper ,
we assume that all transmitters have no csit , and that each receiver has only knowledge of the channel impulse response between itself and the paired transmitter ; i.e. , no knowledge of channel state information at receiver ( csir ) associated with other receivers .
this is the minimum channel knowledge to perform coherent detection at receivers .
transmitter @xmath18 sends an independent message @xmath38 to intended receiver @xmath18 at rate @xmath39 for @xmath40 .
then , it is said that rate @xmath41 is achievable if receiver @xmath18 can decode the desired message with an error probability that is arbitrarily small for sufficient channel uses @xmath42 .
the sum - dof that characterizes an approximate sum - spectral efficiency at high is defined as @xmath43 we aim at using the sum - dof metric in this paper to provide a clear understanding of our proposed transmission method in section iii and iv . in section
v , we will characterize the expressions of the sum - spectral efficiency when the proposed method is applied .
in this section , we present the key idea behind if - ofdm starting with some example cases . gaining insights from this section , we provide our main result for general scenarios in section iv .
* example 1 : * suppose @xmath44 and @xmath45 for @xmath33 .
in this example case , our goal is to show that each transmitter is able to send one data symbol to the associated receiver using four time slots , i.e. , @xmath46 .
the key to demonstrating this result is the proposed transmission method , referred as to _ if - ofdm_. the principal idea of if - ofdm is to create a circulant channel structure for all interfering wireless links , while keeping the non - circulant structure for the desired wireless link . because of this relativity of the channel structures between the interfering and the desired links
, very simple precoding can be used to align all interference signals in a subspace while preserving the desired signals in a different subspace .
let @xmath47^{\top}$ ] be the beamforming vector carrying information symbol @xmath48 for @xmath14 .
this beamforming vector will be used for all transmitters .
the transmitted signals of transmitter @xmath18 during two time slots can be interpreted as a simple repetition transmission method , namely , @xmath49 \\ { x}_{k}[2 ] \\
\end{array}% \right ] = { \bf f}_1s_k,\end{aligned}\ ] ] where @xmath14 .
then , from the received signals of receiver @xmath18 during four time slots are given by @xmath50 \\ { y}_{k}[2 ] \\ { y}_{k}[3 ] \\ { y}_{k}[4 ] \\
\end{array}% \right]}_{{\bf y}_k } \!&\!= \underbrace{\left[% \begin{array}{cc } { h}_{k , k}[1 ] \!&\ ! 0 \\ { h}_{k , k}[2 ] \!&\ ! { h}_{k , k}[1 ] \\ { h}_{k , k}[3 ] \!&\ ! { h}_{k , k}[2 ] \\ 0 \!&\ ! { h}_{k , k}[3 ] \\
\end{array}% \right]}_{{\bf h}_{k , k } } \left[% \begin{array}{c } { x}_{k}[1 ] \\ { x}_{k}[2 ] \\
\end{array}% \right ] \nonumber \\\!&\!+ \sum_{i\neq k}^k\underbrace{\left[\!\!% \begin{array}{cc } { h}_{k , i}[1 ] \!&\ ! 0 \\ { h}_{k , i}[2 ] \!&\ ! { h}_{k , i}[1 ] \\ 0 \!&\ ! { h}_{k , i}[2 ] \\ 0 \!&\ ! 0 \\ \end{array}% \!\!\right ] } _ { { \bf h}_{k , i}}\!\!\ ! \left[\!\!% \begin{array}{c } { x}_{i}[1 ] \\ { x}_{i}[2 ] \\ \end{array}% \!\!\right ] \!\!+\!\ ! \left[\!\!% \begin{array}{c } { z}_{k}[1 ] \\ { z}_{k}[2 ] \\ { z}_{k}[3 ] \\ { z}_{k}[4 ] \\ \end{array}% \!\!\right ]
. \label{eq : example_rx2}\end{aligned}\ ] ] as mentioned previously , the key idea of if - ofdm is to 1 ) make all the channel matrices of interfering links @xmath51 be circulant matrices , 2 ) while ensuring that desired channel matrix @xmath52 does not hold the circulant matrix structure . here , we present an idea of how to accomplish these above two conditions .
let @xmath53 be a combiner matrix that is applied for receiver @xmath18 .
in this example , we particularly construct @xmath54 for @xmath14 as follows : @xmath55.\end{aligned}\ ] ] applying this linear combiner to @xmath56 in , we obtain output vector @xmath57 as @xmath58 \!+\ ! { y}_{k}[3 ] \\ { y}_{k}[2 ] \\
\end{array}% \!\!\!\right ] } _ { { \bf \bar y}_{k } } = \!\!\underbrace{\left[\!\!\!% \begin{array}{cc } { h}_{k , k}[1]\!+\!{h}_{k , k}[3 ] \!\!&\!\ ! { h}_{k , k}[2 ] \\ { h}_{k , k}[2 ] \!\!&\!\ ! { h}_{k , k}[1 ] \\
\end{array}% \!\!\!\right]}_{{\bf \bar h}_{k , k } } \!\underbrace { \left[\!\!\!% \begin{array}{c } { x}_{k}[1 ] \\ { x}_{k}[2 ] \\
\end{array}% \!\!\!\right]}_{{\bf x}_k } \nonumber \\\!&\!\!\!\!\!\!\!\!\!\!+ \sum_{\ell\neq k}^k\underbrace{\left[\!\!% \begin{array}{cc } { h}_{k , i}[1 ] \!\!&\!\ ! { h}_{k , i}[2 ] \\ { h}_{k , i}[2 ] \!\!&\!\ ! { h}_{k , i}[1 ] \\
\end{array}% \!\!\right ] } _ { { \bf \bar h}_{k , i } } \!\underbrace{\left[\!\!\!% \begin{array}{c } { x}_{i}[1 ] \\ { x}_{i}[2 ] \\ \end{array}% \!\!\!\right]}_{{\bf x}_{i } } \!+\ !
\underbrace{\left[\!\!\!\!% \begin{array}{c } { z}_{k}[1 ] \!+\!{z}_{k}[3 ] \\ { z}_{k}[2 ] \\ \end{array}% \!\!\!\!\right]}_{{\bf \bar z}_k}\!\!. \label{eq : output_combiner}\end{aligned}\ ] ] in , all the effective interference channel matrices @xmath59 are circulant , whereas the effective desired matrix @xmath60 is not a circulant matrix .
it is well - known that a circulant matrix @xmath61 can be diagonalized by a discrete - time fourier transform ( dft ) matrix that obeys @xmath62 \left[% \begin{array}{cc } \lambda^{1}_{k , i } \!&\ ! 0 \\ 0 \!&\ !
\lambda^{2}_{k , i } \end{array}% \right]\left[% \begin{array}{cc } { \bf f}_1 \!&\ ! { \bf f}_2 \end{array}% \right]^h,\end{aligned}\ ] ] where @xmath63^{\top}$ ] and @xmath64^{\top}$ ] are the eigenvectors , and @xmath65 denotes the eigenvalue associated with eigenvector @xmath66 for @xmath67 .
recall that all transmitters have applied the same linear beamforming vector for sending each data symbol , i.e. , @xmath68 for @xmath14 . as a result
, all interference signals are aligned in the signal direction @xmath69 , i.e. , @xmath70 for all @xmath71 and @xmath72 .
this fact allows us to rewrite @xmath73 in as follows : @xmath74 because @xmath69 and @xmath75 are mutually orthogonal to each other , we multiply receive beamforming vector @xmath76 to @xmath73 ; by this process , it possible to eliminate all the aligned interference at receiver @xmath18 can be eliminated . accordingly , the interference - free received signal output is given by @xmath77 therefore , the spectral efficiency of the @xmath18th communication link is @xmath78|^2}{6\sigma^2}\right),\end{aligned}\ ] ] where @xmath79=\frac{3}{2}\sigma^2 $ ] , @xmath80=p$ ] , and @xmath81}{2}$ ] . because we have used four time slots to deliver @xmath0 independent data symbols , the sum - dof of @xmath82 is achievable , i.e. , @xmath46 .
* example 2 : * suppose @xmath45 and @xmath83 .
we assume that @xmath84 \neq 0 $ ] for @xmath85 while , @xmath86=h_{k , i}[2]=0 $ ] and @xmath87 \neq 0 $ ] for @xmath88 .
therefore , @xmath89 , h_{k , k}[2 ] , h_{k , k}[3]\right ) |=3 $ ] and @xmath90 , h_{k , i}[2],\ldots , h_{k , i}[4]\right ) |=2 $ ] , where @xmath91 denotes a support set of @xmath92 .
this example corresponds to the scenario in which the arrival - times of the desired signal and interference signals differ due to the propagation delay between the desired and the interfering links . in this example
, we will show that each transmitter sends one information symbol by spanning five time slots .
let @xmath93^{\top}$ ] be the beamforming vector carrying information symbol @xmath48 for @xmath14 .
then , from , the received signals of receiver @xmath18 during five time slots are given by @xmath50 \\ { y}_{k}[2 ] \\ { y}_{k}[3 ] \\ { y}_{k}[4 ] \\ { y}_{k}[5 ] \\
\end{array}% \right]}_{{\bf y}_k } \!&\!= \underbrace{\left[% \begin{array}{cc } { h}_{k , k}[1 ] \!&\ ! 0 \\ { h}_{k , k}[2 ] \!&\ ! { h}_{k , k}[1 ] \\ { h}_{k , k}[3 ] \!&\ ! { h}_{k , k}[2 ] \\ 0 \!&\ ! { h}_{k , k}[3 ] \\ 0 \!&\ ! 0 \\ \end{array}% \right]}_{{\bf h}_{k , k } } \left[% \begin{array}{c } { x}_{k}[1 ] \\ { x}_{k}[2 ] \\
\end{array}% \right ] \nonumber \\\!&\!+ \sum_{i\neq k}^k\underbrace{\left[\!\!% \begin{array}{cc } 0 \!&\ ! 0 \\ 0 \!&\ ! 0 \\ { h}_{k , i}[3 ] \!&\ ! 0 \\ { h}_{k , i}[4 ] \!&\ ! { h}_{k , i}[3 ] \\ 0 \!&\ ! { h}_{k , i}[4 ]
\\ \end{array}% \!\!\right ] } _ { { \bf h}_{k , i}}\!\!\ ! \left[\!\!% \begin{array}{c } { x}_{i}[1 ] \\ { x}_{i}[2 ] \\
\end{array}% \!\!\right ] \!\!+\!\ ! \left[\!\!% \begin{array}{c } { z}_{k}[1 ] \\ { z}_{k}[2 ] \\ { z}_{k}[3 ] \\ { z}_{k}[4 ] \\ { z}_{k}[5 ] \\
\end{array}% \!\!\right ] .
\label{eq : example_rx}\end{aligned}\ ] ] by taking two output signals @xmath94+y_k[5]$ ] and @xmath95 $ ] , we obtain @xmath96 \!+\
! { y}_{k}[5 ] \\ { y}_{k}[4 ] \\
\end{array}% \!\!\!\right ] } _ { { \bf \bar y}_{k } } \!&\!= \!\!\underbrace{\left[\!\!\!% \begin{array}{cc } { h}_{k , k}[3 ] \!\!&\!\ !
{ h}_{k , k}[2 ] \\ 0 \!\!&\!\ ! { h}_{k , k}[3 ] \\
\end{array}% \!\!\!\right]}_{{\bf \bar h}_{k , k } } \!\underbrace { \left[\!\!\!% \begin{array}{c } { x}_{k}[1 ] \\ { x}_{k}[2 ] \\
\end{array}% \!\!\!\right]}_{{\bf x}_k } \nonumber \\\!&\!\!\!\!\!\!\!\!\!\!+ \sum_{\ell\neq k}^k\underbrace{\left[\!\!% \begin{array}{cc } { h}_{k , i}[1 ] \!\!&\!\ ! { h}_{k , i}[2 ] \\ { h}_{k , i}[2 ] \!\!&\!\ ! { h}_{k , i}[1 ] \\
\end{array}% \!\!\right ] } _ { { \bf \bar h}_{k , i } } \!\underbrace{\left[\!\!\!% \begin{array}{c } { x}_{i}[1 ] \\ { x}_{i}[2 ] \\
\end{array}% \!\!\!\right]}_{{\bf x}_{i } } \!+\ ! \underbrace{\left[\!\!\!\!% \begin{array}{c } { z}_{k}[3 ] \!+\!{z}_{k}[5 ] \\ { z}_{k}[4 ] \\
\end{array}% \!\!\!\!\right]}_{{\bf \bar z}_k } , \nonumber\end{aligned}\ ] ] which reveals that the channel matrices for the interfering links can be made circulant , while the matrix for the desired link is kept non - circulant . by applying the same decoding strategy
as explained in example 1 , the sum - dof of @xmath97 is achievable .
as shown in example 2 , one can still use the proposed idea in the case of @xmath98 , provided that the support sets @xmath99 , h_{k , k}[2],\ldots , h_{k , k}[l_{\rm d}]\}$ ] and @xmath100 , h_{k , i}[2],\ldots , h_{k , i}[l_{{\rm i } } ] \right\}$ ] differs , i.e. , @xmath90 , h_{k , i}[2],\ldots , h_{k , i}[l_{{\rm i}}]\right ) | < |{\rm supp } \left ( { h}_{k , k}[1 ] , h_{k , k}[2],\ldots , h_{k , k}[l_{{\rm d}}]\right ) | $ ] . in the rest of this paper , however , we will focus on the case of @xmath101 by assuming no relative propagation delay occurs between the desired and the interfering links .
this assumption implies that all channel coefficients have non - zero entries , i.e. , @xmath84\neq 0 $ ] for @xmath102 and @xmath34\neq 0 $ ] for @xmath103 .
this assumption made for mathematical convenience to explain the proposed idea clearly , but does not preclude application of the proposed idea in the case of @xmath104 with relative propagation delay .
* remark 1 ( no csit and no csir for interfering links ) : * the most prominent feature of if - ofdm is that all iui signals are canceled with the common and the predefined transmit beamforming vector @xmath69 and receive beamforming vector @xmath75 .
this feature implies that all transmitters and receivers do not require to have any knowledge in both either csit and csir to the designing of @xmath69 and @xmath75 .
this fact makes the proposed idea extremely practical .
* remark 2 ( interpretation of the proposed method with the lens of multi - carrier systems ) : * as illustrated in example 1 and 2 , all transmitters send a data symbol along with the first column vector of the 2-point idft matrix @xmath105^{\top}$ ] .
this beamforming method can be interpreted as the spreading operation of a data symbol ( in the first subcarrier ) to a time - domain signal vector with length two ; this is similar to a conventional multicarrier transmission method . because @xmath69 is the first eigenvector of @xmath106 , all the interference signals at receiver @xmath14 are aligned to its direction
; this alignment implies that they remain in the first subcarrier .
meanwhile , the non - circulant channel matrix @xmath107 spreads the desired signal ( sent through the first subcarrier ) to the both subcarriers .
receiver @xmath18 discards the aligned interference signal in the first - subcarrier by multiplying @xmath76 , which is the second eigenvector of @xmath106 . accordingly , the desired signal contained in the second subcarrier can be obtained .
in this section , we generalize the idea of if - ofdm introduced in the previous section .
the following theorem is the main result of this paper .
[ theorem1 ] consider a @xmath0-user wideband siso interference channel with @xmath21 .
let @xmath108 and @xmath109 for @xmath110 , respectively .
then , the achievable sum - dof with completely no csit is @xmath111 where @xmath112 and @xmath113 .
we start by providing a lemma that is essential for demonstrating our achievability result .
the following lemma yields the matrix factorization of a circulant matrix .
[ lemma1 ] let @xmath114 be a circulant matrix comprised of elements @xmath115 .
then , @xmath116 is decomposed as @xmath117 where @xmath118\in\mathbb{c}^{n\times n}$ ] is the @xmath23-point idft matrix whose @xmath18th column vector is defined as @xmath119^{\top},\end{aligned}\ ] ] @xmath120 and @xmath121 are the @xmath23-th roots of unity .
the eigenvalue associated with @xmath122 is then given by @xmath123 see @xcite .
this lemma is used in the proof of theorem [ theorem1 ] .
in this proof , we focus on the case of @xmath124 for some @xmath14 because it is trivial to achieving the sum - dof@xmath125 by tdma - ofdm when @xmath126 .
the proposed achievable method uses a block transmission technique .
let @xmath42 be the total number of time slots per transmission block .
the transmission block with @xmath42 time slots consists of @xmath127 sub - blocks ( or @xmath127 ofdm symbols ) each with @xmath128 time slots , where @xmath129 . at the end of the transmission block , we span @xmath130 additional time slots to prevent from inter - block interference
therefore , the total number of time slots needed for a block transmission is @xmath131 in this proof , we first demonstrate how each communication link achieves dof of @xmath132 when the number of subblocks @xmath127 is fixed .
we present a transmission method that uses an ofdm technique based on cyclic prefixes , because it is a more systematic transmission method than the one explained in section iii .
let @xmath133 be the transmitted signal vector of the @xmath134th sub - block with length @xmath135 , which is defined as @xmath136 , \!\!&\!\!\cdots , \!\!&\!\ !
x_k[(b-1){\bar n}\!+\!n ] \\ \end{array}% \!\!\right]^{\!\!\top},\end{aligned}\ ] ] where @xmath137 .
we also use @xmath138 to denote a set of subcarrier indices for delivering data symbols of transmitter @xmath18 .
we create @xmath133 by the superposition of @xmath139 information symbols along with precoding vectors @xmath140 , which is @xmath141 where @xmath142^{\top}$ ] is the @xmath23th eigenvector of a circulant matrix with size of an @xmath143 by @xmath143 as shown in lemma [ lemma1 ] .
one possible interpretation of this precoding method is that it spreads @xmath144 information symbols in the frequency domain using a set of column vectors of the @xmath143-point idft matrix to multiplex the signals in the time domain .
we generate the input data vector of the @xmath134th sub - block with length @xmath145 by adding a cyclic prefix with length @xmath146 to @xmath147 , i.e. , @xmath148^{\!\top},\end{aligned}\ ] ] where @xmath149,\ldots , x_k[(b\!-\!1){\bar n}\!+\!n]\right]^{\!\!\top}\!\!$ ] . to prevent from inter - block - interference of the desired link ,
we provide a guard interval with length @xmath130 zeroes .
finally , the transmitted signal vector from transmitter k during a block transmission is finally , the transmitted signal vector from transmitter @xmath18 during a block transmission is @xmath150^{\!\top}\!\!\!. \end{aligned}\ ] ] then , the received signal of receiver @xmath18 in time slot @xmath23 is expressed by the superposition of @xmath0 linear convolution operations between the transmitted signals and the channel impulse responses . without loss of generality , we focus on the output signals of the @xmath134th sub - block at receiver @xmath18 , which are @xmath151 & \!= \sum_{\ell=1}^{l_{k , k}}h_{k , k}[\ell]x_{k}[(b\!-\!1){\bar n } + n-\ell+1 ] \nonumber \\
\!&\!+\!\!\sum_{i \in\mathcal{k}/\{k\ } } \sum_{\ell=1}^{l_{k , i}}h_{k , i}[\ell]x_{i}[(b\!-\!1){\bar n}\!+\!n\!-\!\ell+1 ] \nonumber \\
n } + t ] , \nonumber \end{aligned}\ ] ] where @xmath152 and @xmath153 .
let @xmath154,\cdots , y_k[(b\!-\!1)b\!+\!{\bar n } ] \right]^{\!\top}\in\mathbb{c}^n$ ] and @xmath155 , \cdots , z_k[(b\!-\!1)b\!+\!{\bar n } \right]^{\!\!\top } \in\mathbb{c}^n$ ] be the effective received signal and noise vectors of the @xmath134th sub - block after discarding the cyclic prefix , respectively .
then , the effective input - output relationship of the @xmath134th sub - block in a matrix form is given by @xmath156 where @xmath157 \!\!&\!\ ! \cdots \!\!&\!\ ! { h}_{k , i}[l_{k , i } ] \!\!&\!\ ! \cdots \!\!&\!\ ! { h}_{k , i}[3 ] \!\!&\!\
{ h}_{k , i}[2 ] \\ { h}_{k , i}[2 ] \!\!&\!\ ! { h}_{k , i}[1 ] \!\!&\!\
! \cdots \!\!&\!\ !
{ h}_{k , i}[l_{k , i } ] \!\!&\!\ ! \cdots \!\!&\!\ ! { h}_{k , i}[3]\\ \vdots \!\!&\!\ ! \vdots \!\!&\!\ ! \ddots\!\!&\!\
! \ddots \!\!&\!\ ! \ddots \!\!&\!\
! \vdots \\ { h}_{k , i}[l_{k , i } ] \!\!&\!\ ! 0 \!\!&\!\ ! \ddots \!\!&\!\ ! \cdots \!\!&\!\ ! 0 \!\!&\!\ ! { h}_{k , i}[l_{k , i}]\\ 0 \!\!&\!\ ! { h}_{k , i}[l_{k , i } ] \!\!&\!\ ! 0 \!\!&\!\
! \cdots \!\!&\!\ ! { h}_{k , i}[1 ] \!\!&\!\ !
0\\ 0 \!\!&\!\ ! 0 \!\!&\!\ ! { h}_{k , i}[l_{k , i}\ ! ] \!\!&\!\
! \cdots \!\!&\!\ ! { h}_{k , i}[2 ] \!\!&\!\ ! { h}_{k , i}[1]\\ \end{array}% \!\!\!\!\right ] \nonumber\end{aligned}\ ] ]
is the effective channel matrix from transmitter @xmath26 to receiver @xmath18 with size of @xmath158 for all @xmath33 and @xmath159 .
it is important note that @xmath160 is a circulant matrix , because the length of @xmath161 was selected with length @xmath162 .
however , the effective channel matrix that carries the desired information symbols , @xmath163 does not hold the circulant matrix structure , because the length of cyclic prefix vector @xmath161 was chosen to be shorter than the length of channel - taps of direct links , i.e. , @xmath164 .
this creates _ the relativity of alignment between direct channels and interference channels by a matrix structure . _
* inter - user interference cancellation : * we have constructed the input data vector of the @xmath134th sub - block as the superposition of @xmath165 data streams by using linear beamforming , i.e. , @xmath166 to perform the if - ofdm transmission . utilizing this ,
the received signal expression in is rewritten as @xmath167 where the second equality follows from the fact that @xmath168 for all @xmath169 , @xmath159 , and @xmath170 .
notice that the interference signals sent from transmitter @xmath26 are confined at receiver @xmath14 in an @xmath171-dimensional subspace defined as @xmath172\right).\end{aligned}\ ] ] because the union of @xmath173 belongs to the total interference subspace , namely @xmath174\right),\end{aligned}\ ] ] we eliminate all iui signals sent from @xmath175 transmitters at receiver @xmath18 by multiplying @xmath176^h\in\mathbb{c}^{(l_{k , k}-l_{\rm i } ) \times n } $ ] , which spans the null space of @xmath177 , to @xmath73 ; thereby , the output signal vector that only contains the desired data symbols @xmath178^{\top}$ ] can be obtained as @xmath179 where @xmath180\in\mathbb{c}^{n\times ( l_{k , k}-l_{\rm i})}$ ] and @xmath181 is the effective noise vector at receiver @xmath18 whose distribution is invariant with @xmath182 because @xmath183 is a unitary transformation matrix .
* decodability of subblock data : * now , the remaining key step to end this proof is to show that @xmath184 almost surely .
we prove this by considering two different cases : 1 ) @xmath185 and 2 ) @xmath186 .
we first consider the case of @xmath185 , i.e. , @xmath187 . in this case , by the definition , @xmath188 can be written as in .
@xmath189 \!\!&\!\ ! 0 \!\!&\!\ ! \cdots\!\!&\!\ !
\cdots & \!\ ! \cdots \!\!&\!\ ! { h}_{k , k}[l_{\rm i } ] \!\!&\!\
! \cdots \!\!&\!\ ! { h}_{k , k}[2 ] \\
{ h}_{k , k}[2 ] \!\!&\!\ ! { h}_{k , k}[1 ] \!\!&\!\ ! 0 \!\!&\!\ !
\cdots & \!\ ! \cdots \!\!&\!\ ! { h}_{k , k}[l_{\rm i}-1 ] \!\!&\!\
\cdots \!\!&\!\ ! { h}_{k , k}[3]\\ { h}_{k , k}[3 ] \!\!&\!\ ! { h}_{k , k}[2 ] \!\!&\!\ ! { h}_{k , k}[1 ] \!\!&\!\
! \ddots \!\!&\!\ ! \vdots \!\!&\!\ ! \cdots&\!\
! \cdots \!\!&\!\ !
\vdots \!\!&\!\ ! \vdots \!\!&\!\ ! \ddots \!\!&\!\ !
\ddots & \!\ ! \cdots \!\!&\!\ !
h_{k , k}[l_{k , k } ] \!\!&\!\
! \ddots \!\!&\!\ ! \vdots\\
\vdots \!\!&\!\ ! \vdots \!\!&\!\ !
\ddots & \!\ ! \cdots \!\!&\!\
! \ddots \!\!&\!\ ! 0 \!\!&\!\ ! \ddots \!\!&\!\ ! { h}_{k , k}[l_{k , k}-1]\\ \vdots \!\!&\!\ ! \vdots \!\!&\!\ ! \ddots \!\!&\!\
! \ddots \!\!&\!\ ! \ddots \!\!&\!\ !
\vdots & \!\ ! \cdots \!\!&\!\ ! { h}_{k , k}[l_{k , k}]\\ { h}_{k , k}[l_{k , k } ] \!\!&\!\ ! { h}_{k , k}[l_{k , k}\!-\!1 ] \!\!&\!\ ! \ddots \!\!&\!\ ! \ddots \!\!&\!\ !
\cdots & \!\ ! 0 \!\!&\!\
! \cdots \!\!&\!\ !
0\\ 0 \!\!&\!\ ! { h}_{k , k}[l_{k , k } ] \!\!&\!\
! \ddots \!\!&\!\
! \ddots \!\!&\!\ ! \vdots&\!\ !
\vdots \!\!&\!\ ! \cdots \!\!&\!\ ! 0\\
\vdots \!\!&\!\ !
\vdots \!\!&\!\ ! \ddots\!\!&\!\ !
\ddots \!\!&\!\ !
\ddots & \!\ ! 0 \!\!&\!\ !
\vdots \!\!&\!\ ! \vdots \\ 0 \!\!&\!\ ! 0 \!\!&\!\ ! { h}_{k , k}[l_{k , k } ] \!\!&\!\
h_{k , k}[l_{k , k}\!-\!1 ] & \!\ ! \cdots \!\!&\!\
! \cdots \!\!&\!\ ! \cdots \!\!&\!\
! { h}_{k , k}[1]\\\hline \end{array}% \!\!\!\right]\in \mathbb{c}^{n\times n}. \vspace{-0.3cm}\label{eq : direct_channel_matrix1}\ ] ] by the linearity of a linear - time - invariant system , @xmath188 in can be decomposed into the difference of the circulant and the non - circulant matrices as @xmath190 where @xmath191 + { \bf p}^1h_{k , k}[2 ] + \cdots + { \bf p}^{l_{\rm d}}h_{k , k}[l_{k , k } ] \nonumber\end{aligned}\ ] ] and @xmath192 is the @xmath26th power of the @xmath193 cyclic permutation matrix @xmath194 .
\nonumber\end{aligned}\ ] ] furthermore , the non - circulant matrix @xmath195 is defined as @xmath196 , \nonumber\end{aligned}\ ] ] where the submatrix @xmath197 is an upper triangular matrix , namely , @xmath198 \!\!&\!\ ! h_{k , k}[l_{k , k}\!-\!1 ] \!\!&\!\ !
\cdots \!\!&\!\ ! h_{k , k } [ l_{\rm i}+1 ] \\ 0 \!\!&\!\ !
h_{k , k}[l_{k , k } ] \!\!&\!\ !
\cdots \!\!&\!\ !
h_{k , k}[l_{\rm i}+2 ] \\ 0 \!\!&\!\ ! 0 \!\!&\!\ !
\cdots \!\!&\!\ !
h_{k , k}[l_{\rm i}+3 ] \\
\vdots \!\!&\!\ !
\vdots \!\!&\!\ ! \ddots\!\!&\!\ !
\vdots \\ 0 \!\!&\!\!0 & \cdots \!\!&\!\ ! h_{k , k}[l_{k , k } ] \\
\end{array}% \!\!\right ] .
\label{eq : non_circulant_1 } \end{aligned}\ ] ] from this decomposition , it implies that @xmath199 notice that @xmath200,h_{k , k}[l_{\rm i}+2],\ldots , h_{k , k}[l_{k , k}]$ ] in were selected from iid complex random variables , i.e. , @xmath201 .
furthermore , because @xmath202 and @xmath183 are the sub - matrices of the @xmath143-point idft and dft matrix , then with high probability , the rank of the resultant channel matrix is @xmath203 in this case , after discarding the cyclic prefix with size @xmath204 , the effective channel matrix between transmitter @xmath18 and receiver @xmath18 can be represented as the sum of two matrices , namely , @xmath205 where @xmath206 \!\!&\!\ ! 0 \!\!&\!\ ! \cdots \!\!&\!\ ! 0 \!\!&\!\ ! 0 \!\!&\!\ ! 0 \\ { h}_{k , k}[2 ] \!\!&\!\ ! { h}_{k , k}[1 ] \!\!&\!\ ! 0 \!\!&\!\ ! 0 \!\!&\!\ ! \cdots \!\!&\!\ ! 0 \\ \vdots \!\!&\!\ ! \ddots \!\!&\!\ ! \ddots \!\!&\!\ ! \vdots \!\!&\!\ !
\vdots \!\!&\!\ ! \vdots
\\ \vdots \!\!&\!\ !
\ddots \!\!&\!\ ! \ddots\!\!&\!\ !
\ddots \!\!&\!\ ! 0 \!\!&\!\ ! \vdots \\ { h}_{k , k}[n\!\!-\!\!1 ] \!\!&\!\ ! \ddots\!\!&\!\
! \ddots \!\!&\!\ ! { h}_{k , k}[2 ] \!\!&\!\ ! { h}_{k , k}[1 ] \!\!&\!\ ! 0 \\ { h}_{k , k}[n ] \!\!&\!\ ! { h}_{k , k}[n\!\!-\!\!1 ] \!\!&\!\ ! \cdots \!\!&\!\ ! \cdots \!\!&\!\ !
h_{k , k}[2 ] \!\!&\!\ ! { h}_{k , k}[1 ] \\
\end{array}% \!\!\!\!\right ] , \nonumber \end{aligned}\ ] ] denotes the data matrix , and @xmath207 \!\!\!&\!\!\ !
{ h}_{k , k}[\!l_{\rm i}\!\!-\!\!1\ ! ] \!\!&\!\ !
\cdots \!\!\!&\!\!\ !
h_{k , k}[3 ] \!\!\!&\!\!\!h_{k , k}[2 ] \\ 0 \!\!&\!\ !
\cdots \!\!&\!\ ! 0 \!\!&\!\
! \vdots \!\!&\!\ ! \ddots \!\!\!&\!\!\ ! \ddots \!\!\!&\!\!\!h_{k , k}[4 ] \!\!\!&\!\!\ ! h_{k , k}[3 ]
\\ 0 \!\!&\!\ ! \cdots \!\!&\!\ ! 0 \!\!&\!\ ! { h}_{k , k}[\!l_{k , k}\!\!-\!\!1\ ! ] \!\!\!&\!\!\ !
\ddots \!\!\!&\!\!\ ! \vdots \!\!\!&\!\!\ !
\ddots \!\!\!&\!\!\ !
h_{k , k}[4 ] \\ 0 \!\!&\!\ !
\cdots \!\!&\!\ ! 0 \!\!&\!\ ! { h}_{k , k}[l_{k , k } ] \!\!\!&\!\!\ !
\vdots \!\!\!&\!\!\!\ddots \!\!\!&\!\!\ ! \vdots \!\!\!&\!\!\ !
\vdots \\ 0 \!\!&\!\
! \cdots \!\!&\!\ ! 0 \!\!&\!\
! \vdots \!\!\!&\!\!\ !
\ddots \!\!\!&\!\!\ !
\vdots \!\!\!&\!\!\ ! { h}_{k , k}[\!n\!\!+\!\!1\ ! ] \!\!\!&\!\!\ ! { h}_{k , k}[\!n\ ! ] \\ 0 \!\!&\!\ ! \cdots \!\!&\!\ ! 0 \!\!&\!\ !
0\!\!\!&\!\!\ ! \cdots \!\!\!&\!\!\!{h}_{k , k}[\!l_{k , k}\ !
] \!\!\!&\!\!\ !
\cdots \!\!\!&\!\!\ ! { h}_{k , k}[\!n\!\!+\!\!1\ ! ] \\ \end{array}% \!\!\!\!\right]\!\
! \nonumber \end{aligned}\ ] ] denotes the cyclic prefix addition matrix .
similar to the previous case , we decompose this effective channel matrix into the sum of a circulant and a non - circulant matrix as @xmath208 where @xmath209 + { \bf p}^1h_{k , k}[2 ] + \cdots + { \bf p}^{n-1}h_{k , k}[n ] .
\nonumber\end{aligned}\ ] ] the non - circulant matrix is given as in .
@xmath210 \!\!&\!\ !
\cdots \!\!&\!\ ! \cdots \!\!&\!\ ! -{h}_{k , k}[\!l_{\rm i}\!+\!1\ ! ] \!\!&\!\ ! 0 \!\!\!&\!\!\ ! 0 \!\!&\!\ !
\cdots \!\!\!&\!\!\ !
\cdots \!\!\!&\!\!\!0 \\ 0 \!\!&\!\
! 0 \!\!&\!\ ! -{h}_{k , k}[n ] \!\!&\!\
! \cdots \!\!&\!\ !
-{h}_{k , k}[\!l_{\rm i } ] \!\!&\!\ ! 0 \!\!\!&\!\!\ ! 0 \!\!&\!\ ! \cdots \!\!\!&\!\!\ !
\cdots \!\!\!&\!\!\!0 \\ 0 \!\!&\!\ ! 0 \!\!&\!\ ! \cdots \!\!&\!\ ! \ddots \!\!&\!\ ! \vdots \!\!&\!\ ! \vdots \!\!&\!\ !
\ddots \!\!\!&\!\!\ ! \ddots \!\!\!&\!\!\!0\!\!\!&\!\!\ ! 0 \\ 0 \!\!&\!\ ! 0 \!\!&\!\ !
\cdots \!\!&\!\ ! 0 \!\!&\!\ !
-h_{k , k}[n ] \!\!&\!\ ! \vdots \!\!&\!\ !
\ddots \!\!\!&\!\!\ ! \ddots \!\!\!&\!\!\!0\!\!\!&\!\!\ ! 0 \\ 0 \!\!&\!\ ! 0 \!\!&\!\ !
\cdots \!\!&\!\ ! 0 \!\!&\!\ ! 0 \!\!&\!\ ! 0\!\!&\!\ !
\cdots \!\!\!&\!\!\ ! \ddots \!\!\!&\!\!\!0\!\!\!&\!\!\ ! 0 \\ 0 \!\!&\!\!0 \!\!&\!\ ! 0 \!\!&\!\ ! \cdots \!\!&\!\ ! 0 \!\!&\!\ ! { h}_{k , k}[\!n\!\!+\!\!1\ ! ] \!\!\!&\!\!\
\ddots \!\!\!&\!\!\ ! \vdots \!\!\!&\!\!\ !
\ddots \!\!\!&\!\!\ ! 0 \\
0 \!\!&\!\!0 \!\!&\!\ ! 0 \!\!&\!\
! \cdots \!\!&\!\ ! 0 \!\!&\!\ ! \vdots \!\!\!&\!\!\ !
\ddots \!\!\!&\!\!\ ! \vdots \!\!\!&\!\!\ !
\ddots \!\!\!&\!\!\ ! 0 \\ 0 \!\!&\!\!0 \!\!&\!\ ! 0 \!\!&\!\ !
\cdots \!\!&\!\ ! 0 \!\!&\!\ ! { h}_{k , k}[l_{k , k } ] \!\!\!&\!\!\ !
\vdots \!\!\!&\!\!\!\ddots \!\!\!&\!\!\ ! \vdots \!\!\!&\!\!\ !
\vdots \\ 0 \!\!&\!\!0 \!\!&\!\ ! 0 \!\!&\!\
! \cdots \!\!&\!\ ! 0 \!\!&\!\
! \vdots \!\!\!&\!\!\ !
\ddots \!\!\!&\!\!\ !
\vdots \!\!\!&\!\!\ ! { h}_{k , k}[\!n\!\!+\!\!1\ ! ] \!\!\!&\!\!\ ! 0 \\ 0 \!\!&\!\!0 \!\!&\!\ ! 0 \!\!&\!\ !
\cdots \!\!&\!\ ! 0 \!\!&\!\ ! 0\!\!\!&\!\!\ !
\cdots \!\!\!&\!\!\!{h}_{k , k}[\!l_{k , k}\ ! ] \!\!\!&\!\!\ !
\cdots \!\!\!&\!\!\ ! { h}_{k , k}[\!n\!\!+\!\!1\ ! ] \\
\end{array}% \!\!\right]\!\!. \label{eq : nc_2 } \end{aligned}\ ] ] from this decomposition , the resultant channel matrix of the @xmath18th link is @xmath211 since @xmath200,h_{k , k}[l_{\rm i}+2],\ldots , h_{k , k}[l_{k , k}]$ ] were chosen to be iid complex random variables , @xmath212 .
accordingly , we conclude that @xmath213 * inter - subblock interference cancellation : * we have shown that @xmath144 independent data symbols in the @xmath134th sub - block are decodable after canceling the inter - user interference without csit , assuming that inter - subblock interference does not occur .
unfortunately , the inter - subblock interference is unavoidable between the @xmath134th subblock and its previous @xmath214th subblock for @xmath215 because the cyclic prefix size is shorter than the number of channel - taps of the direct link , i.e. , @xmath216 . by concatenating @xmath217 for @xmath153 after discarding @xmath130 zeros at the end of the transmission block , the total input - output relationship during a transmission block when ignoring noise is @xmath218\!\!&=\!\ ! \left[\!\!\!% \begin{array}{cccc}\small { \bf f}_{\mathcal{s}_k^c}^h{\bf \bar h}^1_{k , k}{\bf f}_{\mathcal{s}_k}\!&\ ! { \bf 0 } \!&\!\cdots \!&\ ! { \bf 0 } \\ { \bf f}_{\mathcal{s}_k^c}^h{\bf \bar h}^{1,2}_{k , k}{\bf f}_{\mathcal{s}_k}\!&\!{\bf f}_{\mathcal{s}_k^c}^h{\bf \bar h}^2_{k , k}{\bf f}_{\mathcal{s}_k}\!&\ !
\cdots \!&\
! \!{\bf 0 } \\ { \bf 0}\!&\!{\bf f}_{\mathcal{s}_k^c}^h{\bf \bar h}^{3,2}_{k , k}{\bf f}_{\mathcal{s}_k}\!&\ ! \cdots \!&\ !
{ \bf 0 } \\ \vdots \!&\ ! \ddots \!&\ ! \ddots \!&\ !
\vdots \\ { \bf 0}\!&\ ! { \bf 0}\!&\
! \cdots \!&\ ! { \bf f}_{\mathcal{s}_k^c}^h{\bf \bar h}^{b}_{k , k}{\bf f}_{\mathcal{s}_k}\\ \end{array}% \!\!\!\!\!\right]\!\!\!\!\left[\!\!\!\!% \begin{array}{c } { \bf s}_k^1 \\ { \bf s}_k^2 \\ \vdots \\ { \bf s}_k^{b } \\ \end{array}% \!\!\!\!\right]\!\ ! , \label{eq : concentrating_rx_signals}%\ !
\nonumber\\ & % + \!\left[\!\!% % \begin{array}{c } % { \bf \tilde z}^1_k \\ % { \bf \tilde z}^2_k \\ % \vdots \\ % { \bf \tilde z}^{b}_k \\ % \end{array}% % \!\right]\!\!,\end{aligned}\ ] ] where @xmath219 is the inter - subblock interference channel matrix of the @xmath18th link . by the definition
, one can easily verify that it is the same matrix as in , @xmath220.\end{aligned}\ ] ] because no inter - subblock interference occurs in the the first subblock output vector , i.e. , @xmath221 and @xmath222 , @xmath223 can be reliably decoded when ignoring noise @xmath224 .
once @xmath223 is decoded , @xmath225 can be decoded using @xmath226 by successive interference cancellation . by applying this strategy over @xmath127 subblocks recursively ,
receiver @xmath18 is capable of decoding @xmath227 independent data symbols , @xmath228 , using @xmath229 channel uses . by symmetry ,
all the other receivers decode @xmath227 independent data symbols spanning @xmath42 channel uses . as a result , as @xmath127 tends to infinity , we conclude that the achievable dof of the @xmath18th communication link is @xmath230 since @xmath231 , we reach the expression in theorem 1 ; this completes the proof .
we provide some remarks about our result .
* remark 3 ( precoding matrix design and subcarrier selection ) : * the proposed transmission method suggests that all transmitters use the same precoding vectors when carrying their own data symbols . in our proof , for ease of exposition , we construct the precoding matrix by taking the first @xmath144 column vectors among the first @xmath232 column vectors of a @xmath158 idft matrix , i.e. , @xmath233 .
one may , however , construct a set of the transmit precoding vectors by selecting arbitrary @xmath234 column vectors from the idft matrix .
once the set of the precoding vectors is determined , all receivers exploit the decoding matrix that lies in the null space of the subspace spanned by the precoding vectors . as a result , with the predetermined decoding dft matrix , all receivers can eliminate the aligned interference signals .
interesting future work would be to find the optimal @xmath234 column vectors ( subcarriers ) among @xmath143 column vectors in @xmath235 that maximize the achievable sum - spectral efficiency by exploiting csit feedback .
* remark 4 ( symmetric case ) : * suppose the symmetric scenario in which @xmath236 for @xmath14 and @xmath237 for @xmath110 and @xmath238 . in this case , the achievable sum - dof simplifies to @xmath239 shows that in the extreme case of @xmath240 and @xmath241 , @xmath242 can be achieved . in another extreme case with @xmath243 , one can achieve @xmath244 .
these results imply that the sum - dof converges to @xmath46 as @xmath245 tends to infinity .
thus , we can expect the linear gain with respect to @xmath0 in the both cases .
* remark 5 ( difference with blind interference alignment @xcite ) : * the proposed method is similar to blind interference alignment because the both align the interference signals without using csit .
nevertheless , the proposed method differs from blind interference alignment in that it does not need a special channel coherence pattern to accomplish interference alignment .
the only requirement for if - ordm is that the number of channel taps ( i.e. , the delay spread ) be larger for the desired links than for the interfering links .
* remark 6 ( comparison with tdma - ofdm ) : * the proposed communication strategy resembles with the ofdm transmission method , which is widely used in the contemporary wireless systems ( e.g. lte and wifi ) .
for example , if each transmitter - and - receiver pair sends @xmath42 data symbols by using @xmath42-point ifft operation with the addition of cyclic prefix length @xmath246 in a round - robin fashion , the sum - dof of the @xmath0-user interference with isi is @xmath247 as @xmath42 goes infinity , it converges to unity .
this trend implies that the ofdm method does not scale with the number of @xmath0 .
the proposed transmission method can be interpreted as a variant of the conventional ofdm transmission , because if - ofdm also uses idft to transform a set of data symbols .
nevertheless , the two methods differ .
the key differences are : 1 ) the selection of cyclic prefix size and 2 ) the use of the subcarriers to remove iui .
if - ofdm selects the cyclic prefix size to create the relativity of the channel structure , whereas ofdm chooses the size to create parallel subchannels of the desired link by removing ici .
furthermore , the proposed method uses a set of column vectors of the idft matrix during transmit mode , and the receiver applies the non - overlapping column vectors of the dft matrix to reject iui . the conventional ofdm transmission , however , uses the same set of column vectors of idft and dft matrices in transmitting and receiving mode .
it is instructive to consider an example to provide a better understanding our result .
* example 3 : * suppose a symmetric case in which with @xmath248 , @xmath45 , and @xmath249 ; these parameters correspond to the scenario where two dominant multi - paths exist with the line - of - sight path in the desired channels , but no effective multi - paths are resolved in the interference channels . because @xmath241 , no cyclic prefix is needed in this example .
furthermore , we set the size of each subblock ( ofdm symbol ) to @xmath250 . because @xmath251 , transmitter @xmath18 uses a 4-point idtf matrix for precoding and sends two information symbols in during the @xmath134th subblock where @xmath252 . in this example , suppose the active subcarrier set that carries two data symbols in each subblock is @xmath253 .
applying the first and third column vectors of the 4-point idtf matrix as precoding vectors @xmath254^{\top}$ ] and @xmath255^{\top}$ ] , we create the channel input vector of each subblock as follows : @xmath256 \\ { x}_{k}[2 ] \\ { x}_{k}[3 ] \\ { x}_{k}[4 ] \\
\end{array}% \right ] = \frac{1}{\sqrt{4 } } \left[% \begin{array}{c } 1 \\ 1 \\ 1 \\ 1
\\ \end{array}% \right ] s_{k,1}^1 + \frac{1}{\sqrt{4 } } \left[% \begin{array}{c } 1 \\
-1 \\ 1 \\
\end{array}% \right ] s_{k,3}^1\nonumber \\ { \bf x}_k^{2}&= \left[% \begin{array}{c } { x}_{k}[5 ] \\ { x}_{k}[6 ] \\ { x}_{k}[7 ] \\ { x}_{k}[8 ] \\
\end{array}% \right ] = \frac{1}{\sqrt{4 } } \left[% \begin{array}{c } 1 \\ 1 \\ 1 \\ 1
\\ \end{array}% \right ] s_{k,1}^2 + \frac{1}{\sqrt{4 } } \left[% \begin{array}{c } 1 \\ -1 \\ 1 \\
-1 \\ \end{array}% \right ] s_{k,3}^2 .
\end{aligned}\ ] ] then , the receive signal at receiver @xmath14 until time slot 8 is given as in ( [ eq : example2 ] ) .
@xmath257 \\ { y}_{k}[2 ] \\ { y}_{k}[3 ] \\ { y}_{k}[4 ] \\ { y}_{k}[5 ] \\ { y}_{k}[6 ] \\ { y}_{k}[7 ] \\ { y}_{k}[8 ] \\
\end{array}% \!\!\!\right ] \!\!&\!=\ !
\!\small \left[\!\!\!% \begin{array}{cccccccc } { h}_{k , k}[1 ] \!\!&\!\!0 \!\!&\!\!0 \!\!&\!\!0 \!\!&\!\!0\!\!&\!\ ! 0\!\!&\!\!0\!\!&\!\ ! 0 \\ { h}_{k , k}[2 ] \!\!&\!\ ! { h}_{k , k}[1 ] \!\!&\!\
! 0 \!\!&\!\!0 \!\!&\!\!0 \!\!&\!\ ! 0 \!\!&\!\!0 \!\!&\!\ ! 0 \\ { h}_{k , k}[3 ] \!\!&\!\!{h}_{k , k}[2]\!\!&\!\ ! { h}_{k , k}[1 ] \!&\ ! 0 \!&\ ! 0 \!\!&\!\ ! 0 \!&\ ! 0 \!\!&\!\ !
0\\ 0 \!\!&\!\ ! { h}_{k , k}[3]\!\!&\!\ ! { h}_{k , k}[2 ] \!\!&\!\ ! { h}_{k , k}[1 ] \!\!&\!\ ! 0 \!\!&\!\ ! 0 \!\!&\!\ ! 0 \!\!&\!\ ! 0 \\ 0 \!\!&\!\ ! 0\!\!&\!\ !
{ h}_{k , k}[3 ] \!\!&\!\!{h}_{k , k}[2 ] \!\!&\!\ ! { h}_{k , k}[1 ] \!&\ ! 0 \!\!&\!\ ! 0 \!&\ ! 0 \\ 0 \!&\ ! 0 \!\!&\!\!0 \!\!&\!\!{h}_{k , k}[3 ] \!\!&\!\ !
{ h}_{k , k}[2 ] \!\!&\!\ ! { h}_{k , k}[1 ] \!\!&\!\ ! 0 \!\!&\!\ ! 0 \\ 0 \!\!&\!\!0\!\!&\!\ ! 0\!\!&\!\ ! 0\!\!&\!\ ! { h}_{k , k}[3 ] \!\!&\!\ ! { h}_{k , k}[2 ] & \!\ ! { h}_{k , k}[1 ] \!\!&\!\
! 0 \\ 0 \!\!&\!\!0\!\!&\!\ ! 0\!\!&\!\ ! 0\!\!&\!\ ! 0\!\!&\!\ ! { h}_{k , k}[3 ] \!\!&\!\ ! { h}_{k , k}[2 ] & \!\ ! { h}_{k , k}[1 ] \\
\end{array}% \!\!\!\!\right ] \!\!\ ! \left[\!\!\!% \begin{array}{c } { x}_{k}[1 ] \\ { x}_{k}[2 ] \\ { x}_{k}[3 ] \\ { x}_{k}[4 ] \\ { x}_{k}[5 ] \\ { x}_{k}[6 ] \\ { x}_{k}[7 ] \\ { x}_{k}[8 ] \\
\end{array}% \!\!\!\right ] \!+\ ! \sum_{\ell\neq k}^k h_{k , i}[1]{\bf i}_{8\times 8}\!\ ! \left[% \begin{array}{c } { x}_{i}[1 ] \\ { x}_{i}[2 ] \\ { x}_{i}[3 ] \\ { x}_{i}[4 ] \\ { x}_{i}[5 ] \\ { x}_{i}[6 ] \\ { x}_{i}[7 ] \\ { x}_{i}[8 ] \\
\end{array}% \right ] \!+\ !
\small\left[% \begin{array}{c } { z}_{k}[1 ] \\ { z}_{k}[2 ] \\ { z}_{k}[3 ] \\ { z}_{k}[4 ] \\ { z}_{k}[5 ] \\ { z}_{k}[6 ] \\ { z}_{k}[7 ] \\ { z}_{k}[8 ] \\
\end{array}% \right ] .
\label{eq : example2}\end{aligned}\ ] ] in the first subblock , the effective input - output relationship is given by @xmath258 \\ { y}_{k}[2 ] \\ { y}_{k}[3 ] \\ { y}_{k}[4 ] \\
\end{array}%dd \!\!\!\right ] \!\!= \left[\!\!% \begin{array}{cccc } { h}_{k , k}[1 ] \!&\ ! 0 & 0 \!&\ ! 0 \\ { h}_{k , k}[2 ] \!&\ ! { h}_{k , k}[1 ] & 0 \!&\ ! 0 \\ { h}_{k , k}[3 ] \!&\ ! { h}_{k , k}[2 ] & { h}_{k , k}[1 ] \!&\ ! 0 \\ 0 \!&\ ! { h}_{k , k}[3 ] & { h}_{k , k}[2 ] \!&\ ! { h}_{k , k}[1 ] \\
\end{array}% \!\!\right ] \left[\!\!% \begin{array}{c } { x}_{k}[1 ] \\ { x}_{k}[2 ] \\ { x}_{k}[3 ] \\ { x}_{k}[4 ] \\
\end{array}% \!\!\right ] \nonumber \\\!&\!\!+\ ! \sum_{i\neq k}^k
\left[\!\!% \begin{array}{cccc } { h}_{k , i}[1 ] \!&\ ! 0 \!&\ ! 0 \!&\ ! 0 \\ 0\!&\ ! { h}_{k , i}[1 ] \!&\ ! 0 \!&\ !
0\\ 0 \!&\ ! 0 \!&\ ! { h}_{k , i}[1 ] \!&\ ! 0 \\ 0\!&\ ! 0 \!&\ ! 0 \!&\ ! { h}_{k , i}[1]\\ \end{array}% \!\!\right ] \!\!\ !
\left[\!\!% \begin{array}{c } { x}_{i}[1 ] \\ { x}_{i}[2 ] \\ { x}_{i}[3 ] \\ { x}_{i}[4 ] \\
\end{array}% \!\!\right ] \!\!+\!\ !
\left[\!\!% \begin{array}{c } { z}_{k}[1 ] \\ { z}_{k}[2 ] \\ { z}_{k}[3 ] \\ { z}_{k}[4 ] \\
\end{array}% \!\!\!\right ] \label{eq : simple1}\!\!.\end{aligned}\ ] ] since @xmath259 for @xmath14 , is can be written as @xmath260 \\ { y}_{k}[2 ] \\ { y}_{k}[3 ] \\ { y}_{k}[4 ] \\
\end{array}%dd \right ] \!&\!= { \bf h}_{k , k } ( { \bf f}_1 s_{k,1}^1 + { \bf f}_3 s_{k,3}^1 ) \nonumber \\\!&\!+
\sum_{i \neq k}^k { h}_{k , i}[1 ] ( { \bf f}_1s_{i,1}^1+{\bf f}_3s_{i,3}^1)+ \left[% \begin{array}{c } { z}_{k}[1 ] \\ { z}_{k}[2 ] \\ { z}_{k}[3 ] \\ { z}_{k}[4 ] \\ \end{array}% \right]\label{eq : simple2}.\end{aligned}\ ] ] by this procedure , all the interference signals are aligned in the direction of @xmath69 and @xmath261 , whereas the desired signal is not aligned to the interference subspace . by multiplying @xmath262^{\top}$ ] and @xmath263^{\top}$ ] to @xmath264,~ y_k[2],~ y_k[3],~ y_k[4]]^{\top}$ ] , the aligned iui signals are canceled , and we finally obtain the output : @xmath265 } _ { { \bf \tilde h}_k } \left[\!\!% \begin{array}{c } s^1_{k,1 } \\ s^1_{k,3 } \\
\end{array}%dd \!\!\!\right ] + \left[\!\!% \begin{array}{c } { \bf f}_2^{\top}{\bf z}_k\\ { \bf f}_4^{\top}{\bf z}_k\\ \end{array}%dd \!\!\!\right ] .
\end{aligned}\ ] ] since @xmath266 with high probability , @xmath267 and @xmath268 can be reliably decoded by using the zf decoder to eliminating ici .
we now consider the received signals in the second subblock , which are written as the superposition of the desired signal , the inter - subblock interference , and the iui .
the effective input - output relationship of the @xmath134th block is given by @xmath269 \\ { y}_{k}[6 ] \\ { y}_{k}[7 ] \\ { y}_{k}[8 ] \\
\end{array}%dd \!\!\!\right ] = \underbrace{{\bf h}_{k , k } ( { \bf f}_1s_{k,1}^2 + { \bf f}_3s_{k,3}^2)}_{{\rm desired~ signal}}+ \underbrace{\sum_{i \neq k}^k { h}_{k , i}[1 ] ( { \bf f}_1s_{i,1}^2+{\bf f}_3s_{i,3}^2)}_{{\rm inter - cell~interference } } \nonumber \nonumber \\ & + \underbrace{\left[\!\!% \begin{array}{cccc } 0 \!&\ ! 0 & { h}_{k , k}[3 ] \!&\ !
{ h}_{k , k}[2 ] \\ 0 \!&\ ! 0 & 0 \!&\ ! { h}_{k , k}[3 ] \\ 0 \!&\ ! 0 & 0 \!&\ ! 0 \\ 0 \!&\!0 & 0 \!&\!0 \\ \end{array}% \!\!\right ] \left[\!\!% \begin{array}{c } { x}_{k}[1 ] \\ { x}_{k}[2 ] \\ { x}_{k}[3 ] \\ { x}_{k}[4 ] \\ \end{array}% \!\!\right]}_{{\rm inter - subblock ~interference } } + \left[\!\!% \begin{array}{c } { z}_{k}[1 ] \\ { z}_{k}[2 ] \\ { z}_{k}[3 ] \\ { z}_{k}[4 ] \\ \end{array}% \!\!\!\right],\end{aligned}\ ] ] this equation shows that inter - subblock interference terms @xmath270{x}_{k}[3]+{h}_{k , k}[2]{x}_{k}[4]$ ] and @xmath270{x}_{k}[4]$ ] can be cancelled , as receiver @xmath14 has decoded these information symbols in the previous subblock . accordingly , the same input - output relationship after eliminating the inter - subblock interference is given by @xmath271 \\ { y}_{k}[6 ] \\ { y}_{k}[7 ] \\ { y}_{k}[8 ] \\
\end{array}%dd \right ] -\left[\!\!% \begin{array}{cccc } 0 \!&\ ! 0 & { h}_{k , k}[3 ] \!&\ ! { h}_{k , k}[2 ] \\ 0 \!&\ ! 0 & 0 \!&\ ! { h}_{k , k}[3 ] \\ 0 \!&\ ! 0 & 0 \!&\ ! 0 \\ 0 \!&\!0 & 0 \!&\!0 \\ \end{array}% \!\!\right ] \left[\!\!% \begin{array}{c } { x}_{k}[1 ] \\ { x}_{k}[2 ] \\ { x}_{k}[3 ] \\ { x}_{k}[4 ] \\ \end{array}% \!\!\right ] \nonumber \\
\!&\!= { \bf h}_{k , k } ( { \bf f}_1 s_{k,1}^2 + { \bf f}_3 s_{k,3}^2 ) \nonumber \\\!&\!+ \sum_{i \neq k}^k { h}_{k , i}[1 ] ( { \bf f}_1s_{i,2}^2+{\bf f}_3s_{i,2}^2)+ \left[% \begin{array}{c } { z}_{k}[1 ] \\ { z}_{k}[2 ] \\ { z}_{k}[3 ] \\ { z}_{k}[4 ] \\ \end{array}% \right].\end{aligned}\ ] ] applying the same decoding strategy as in the first subblock , @xmath272 and @xmath273 can also be decoded . as a result ,
receiver @xmath18 can decode the four independent data symbols @xmath274 with the ten channel uses .
when @xmath127 subblocks are used , the achievable rate of the @xmath18th link is @xmath275 consequently , as @xmath127 goes to infinity , the achievable sum - spectral efficiency of this channel is asymptotically given by @xmath276 this example demonstrates that the proposed communication strategy achieves the sum - dof of the channel without csit , @xmath277 .
[ fig:2 ] illustrates one possible transmitter and receiver architecture to implement the proposed method explained in this example .
* remark 7 ( inter - subblock interference cancellation ) : * one may implement the proposed method without using inter - subblock interference cancellation operation by including @xmath234 guard - time ( zero ) slots with the cyclic prefix between consecutive subblocks
. this can reduce the computational complexity of interference cancellation algorithms at the cost of the increased underutilization of the resources .
in this section , we characterize the achievable sum - spectral efficiency of the siso frequency - selective interference channel with 1 ) complete lack of csit and 2 ) the desired link s csit .
we first provide a characterization of the sum - spectral efficiency by using the proposed method explained in the previous section , which requires no csit . from , the output signal vector of the @xmath134th subblock at receiver @xmath18 after canceling the inter - subblock interference is given by @xmath278 applying the qr decomposition @xcite , the effective channel matrix @xmath279 can be represented by a product of a unitary matrix @xmath280 and an upper - triangular matrix @xmath281 , namely , @xmath282 under the premise that each transmitter knows the effective snr and uses adaptive modulation / coding to select the right rate , the achievable rate of the @xmath18th message @xmath283 sent over @xmath284 time slots by zf - sic is computed as @xmath285 where @xmath286 is the @xmath23th diagonal element of @xmath287 and @xmath288 because a fraction @xmath289 of power is only used for the data transmission with @xmath290=\frac{np}{l_{k , k}-l_{\rm i}}$ ] to satisfy @xmath291=np$ ] .
notice that @xmath292 for all @xmath293 because the channel coefficients in @xmath294 are invariant over different subblocks . by symmetry , the achievable sum - spectral efficiency is @xmath295 as @xmath127 tends to infinity .
[ fig:3 ] shows that the ergodic sum - spectral efficiency increases linearly with @xmath0 when @xmath7 and @xmath8 .
it is also notable that if - ofdm outperforms tdma - ofdm in all snr regimes . when @xmath7 and @xmath8 , in which @xmath34 $ ] is generated from @xmath296 for all @xmath297 .
, width=336 ] we now consider the case when csit of the desired link is available , i.e. , @xmath298 is known to transmitter @xmath18 . under this premise
, we provide a characterization of the achievable sum - spectral efficiency by using the proposed two - stage precoding and decoding method with the optimal power allocation via water - filling .
it is well - known that the effective channel matrix @xmath298 can be diagonalized by two unitary matrices @xmath299 and @xmath300 by using svd @xcite , namely , @xmath301 where @xmath302\right)$ ] is a diagonal matrix that consists of the singular values of effective channel matrix @xmath303 . from this decomposition ,
we use the proposed two - stage precoding method to construct transmit signal vector @xmath304 as @xmath305 where @xmath306 is the inner precoding matrix that creates multiple parallel subchannels by eliminating iui ; @xmath202 is the outer precoding matrix that aligns inter - user - interference .
@xmath307\right)$ ] is the diagonal matrix in which the @xmath23th diagonal element @xmath308 is power allocated to data symbol @xmath309 .
since this two - stage precoding method does not change the subspace occupied by the iui signals , the iui signals are removed by multiplying @xmath183 to the received vector @xmath73 after discarding the cyclic prefix . multiplying @xmath310 to @xmath311 , we have @xmath144 parallel sub - channel output vector : @xmath312 then , assuming that a separate capacity - achieving awgn code is used to communicate over each of parallel subchannels , the maximum sum - spectral efficiency achievable using this scheme for large enough @xmath127 is @xmath313 where @xmath314 is the optimal power allocated to the @xmath23th subchannel based on _ water - filling strategy _ in @xcite , namely , @xmath315 where @xmath316 is the parameter associated with a lagrange multiplier that is selected to satisfy the power constraint .
this two - stage transmission method provides a better achievable sum - spectral efficiency than the case with completely no csit , whereas the same sum - dof achieves with it .
[ fig:4 ] illustrates the ergodic sum - spectral efficiency when @xmath317 , @xmath318 , and @xmath319 .
as can be seen , csit feedback for the direct link provides the gain in terms of the ergodic sum - spectral efficiency in the low snr regime when @xmath320 . for @xmath321 ,
the proposed two - stage transmission can substantially improve the sum - spectral efficiency , even in the high snr regime .
* remark 8 ( separability ) : * our two - stage precoding communication strategy is practically relevant because it provides a significant spectral efficiency gain in a multi - carrier interference channel with separate encoding over each carrier and a power allocation across carriers .
$ ] is generated from @xmath296 for all @xmath297 .
, width=316 ]
the fundamental principle of if - ofdm is to exploit the circulant matrix property so that iui is aligned in the subspace spanned by a certain set of column vectors of a fourier matrix .
this principle can be applicable in numerous communication scenarios . in this section ,
we first explain some scenarios in which the proposed method shows a better sum - dof gain in interference networks when csit is not available . in the sequel , we discuss some practical ways how to apply the proposed method when @xmath98 . in this subsection , we consider two channel models : 1 ) multiple - input - single - output ( miso ) broadcast channel with isi and 2 ) multiple - input - multiple - output ( mimo ) frequency - flat @xmath0-user interference channel with a circulant structure .
one interesting observation is that the proposed method is directly applicable to a miso broadcast channel with isi when csit is not available .
suppose that @xmath0 transmitters cooperate to form a virtual transmitter with @xmath0 distributed antennas such as cloud radio access networks ( c - ran ) @xcite . in this case , the proposed method enables simultaneous support for the @xmath0 users , even with completely no csit , provided that the channel condition of the scheduled user set satisfies the condition of @xmath322 .
in what follows , we present this claim precisely .
* corollary 1 : * suppose a miso broadcast channel with isi in which a transmitter equipped with @xmath0 distributed antennas communicates with @xmath0 users , each with a single antenna .
let @xmath21 be the number of channel - taps from the @xmath26th transmit antenna to the @xmath18th user . with completely no csit ,
the sum - dof of this channel is lower bounded by @xmath323 since the cooperation of @xmath0 antennas at the transmitter does not degrade the sum - dof , corollary 1 holds from theorem [ theorem1 ] .
consider a frequency - flat ( narrowband ) mimo @xmath0-user interference channel in which all transmitters and receivers have @xmath143 antennas .
let @xmath324 be the channel matrix from transmitter @xmath26 to receiver @xmath18 .
the following corollary shows the sum - dof of the mimo interference channel with a special channel structure . *
corollary 2:*suppose @xmath325 be a circulant matrix for @xmath159 and @xmath326 be a non - circulant matrix composed of elements drawn from a continuous distribution .
then , when csit is not available , the proposed method achieves the sum - dof of this channel as @xmath327 the proof is direct from theorem [ theorem1 ] .
when @xmath104 , one can also apply the proposed idea by treating a fraction of isi as additional noise .
this approach is of interest to the case when a distance - dependent large - scale model is incorporated in multi - cell wireless systems .
* example 4 : * let @xmath328 be the distance between transmitter @xmath26 to receiver @xmath18 in the network .
then , the received signal of receiver @xmath18 can be rewritten by incorporating the distance - dependent large - scale model as @xmath24&=d_{k , k}^{-\frac{\alpha}{2 } } \left(\sum_{\ell=1}^{l_{d } } h_{k , k}[\ell]x_{k}[n-\ell+1]\right ) \nonumber \\ & + \sum_{i\neq k}^k d_{k , i}^{-\frac{\alpha}{2 } } \left(\sum_{\ell=1}^{l_{\rm i}}h_{k , i}[\ell]x_{i}[n-\ell+1]\right ) \!+\ ! { z}_k[n ] , \end{aligned}\ ] ] where @xmath329 is the path - loss exponent .
when each receiver is in the cell center area , i.e. , @xmath330 , receiver @xmath18 can ignore the last isi terms in @xmath34x_i[n-\ell]$ ] for @xmath331 by treating them as the additional noise , i.e. , @xmath332&= \left(\sum_{\ell=1}^{l_{d } } h_{k , k}[\ell]x_{k}[n-\ell+1]\right ) \nonumber \\ & + \sum_{i\neq k}^k\ !
\left(\frac{d_{k , k}}{d_{k , i}}\right)^{\!\!\!-\frac{\alpha}{2}}\!\ ! \left(\sum_{\ell=1}^{l_{d}-1}h_{k , i}[\ell]x_{i}[n-\ell+1]\right ) \!+\ ! { z}^{{\rm eff}}_k[n ] , \nonumber\end{aligned}\ ] ]
where @xmath333= \sum_{i\neq k}^k \left(\frac{d_{k , k}}{d_{k , i}}\right)^{\!\!\!-\frac{\alpha}{2}}\sum_{\ell = l_d}^{l_{i}}h_{k , i}[\ell]x_{i}[n-\ell+1]+ d_{k , k}^{\frac{\alpha}{2}}z_k[n ] .
\nonumber\end{aligned}\ ] ] assuming that @xmath34 $ ] is statistically independent for all different @xmath297 , and its power decays exponentially as the number of channel - taps increases , i.e. , @xmath334|^2]=e^{-\beta ( \ell-1)}$ ] , the variance of the effective noise is @xmath335|^2\right ] = \sum_{i\neq k}^k\left(\frac{d_{k , k}}{d_{k , i}}\right)^{\!\!\!-\alpha } \sum_{\ell = l_d}^{l_{i } } e^{-\beta ( \ell-1 ) } p + d_{k , k}^{\alpha}\sigma^2,\end{aligned}\ ] ] this effective noise is negligible when @xmath336 , so the use of the proposed idea in multi - cell systems can still guarantee a reasonable performance .
this example has had the purpose of showing the feasibility of the proposed method when @xmath98 by incorporating a large - scale fading effect .
more rigorous system - level performance evaluation should be conducted to gauge the practical gain ; we leave this for future work .
we have demonstrated that the sum - dof of the siso @xmath0-user interference channel with isi can increase linearly with @xmath0 , even when csit is totally absent .
this result is surprising because , in the siso interference channel even without isi , this linear sum - dof gain with respect to @xmath0 is achievable when global and perfect csit are available across all transmitters or special channel patterns occur ; this requirement has been a major obstacle to implementing interference alignment in practice .
our main result is proven by introducing a novel communication strategy , if - ofdm .
the main idea of the scheme is to create relativity of alignment based on different matrix structures , i.e. , circulant and non - circulant structures , which are not subject to the realization of the channel coefficients . a promising direction for future work is to consider a multi - antenna setting , ( e.g. , a multiple - input - multiple - output ( mimo ) @xmath0-user interference channel with isi ) , to determine whether the sum - dof can increase with the both the numbers of antennas and of users when csit is unavailable .
other directions for future work include a study of system - level performance for our proposed scheme in cooperative cellular networks .
in particular , by using a stochastic geometry framework @xcite , one can obtain a closed - form expression of the sinr distribution and gauge the gains in terms of spatially - averaged ergodic spectral efficiency to revisit the fundamental limit of cooperation @xcite .
furthermore , one could seek ways to exploit the relativity of alignment based on circulant matrix property to design a new communication strategy in numerous multi - user communication scenarios when csit is not available .
lastly , a derivation of the sum - dof outer bounds for the @xmath0-user interference channel with isi would be useful from an information - theoretic perspective . s. b. weinstein and p. m. ebert , `` data transmission by frequency division multiplexing using the discrete fourier transform , '' _ ieee transactions on communication technology _ , vol .
628 - 634 , oct .
1971 . c. y. wong , r. s. cheng , k. b. letaief , and r. d. murch , `` multicarrier ofdm with adaptive subcarrier ,
bit , and power allocation , '' _ ieee journal on selected areas in communications , _ vol .
1747 - 1758 , oct .
w. rhee and j. m. cioffi , `` increasing in capacity of multiuser ofdm system using dynamic subchannel allocation , '' _ in proc .
ieee international vehicular technology conference _ , vol .
2 , pp . 1085 - 1089 , may 2000 .
s. t. chung and j. cioffi , `` the capacity region of frequency - selective gaussian interference channels under strong interference , '' _ ieee transactions on information theory _ , vol .
1812 - 1821 , sept . 2007 .
v. r. cadambe and s. a. jafar , `` interference alignment and the degrees of freedom of the @xmath0 user interference channel , '' _ ieee transactions on information theory _ , vol .
54 , no . 8 , pp . 3425 - 3441 , aug .
2008 .
n. lee and r. w. heath jr .
, `` space - time interference alignment and degrees of freedom regions for the miso broadcast channel with periodic csi feedback , '' vol .
515 - 528 , jan . 2014 .
r. tandon , s. a. jafar , s. shamai , and h. v. poor , `` on the synergistic benefits of alternating csit for the miso - bc , '' , vol .
59 , no . 7 , pp . 4106 - 4128 , jul . 2013 .
j. chen and p. elia , `` toward the performance versus feedback tradeoff for the two - user miso broadcast channel , '' _ ieee transactions on information theory _
8336 - 8356 , dec . 2013 | this paper considers a @xmath0-user single - input - single - output interference channel with inter - symbol interference ( isi ) , in which the channel coefficients are assumed to be linear time - invariant with finite - length impulse response .
the primary finding of this paper is that , with no channel state information at a transmitter ( csit ) , the sum - spectral efficiency can be made to scale linearly with @xmath0 , provided that the desired links have longer impulse response than do the interfering links .
this linear gain is achieved by a novel multi - carrier communication scheme which we call _ interference - free orthogonal frequency division multiplexing ( if - ofdm)_. furthermore , when a transmitter is able to learn csit from its paired receiver only , a higher sum - spectral efficiency can be achieved by a two - stage transmission method that concatenates if - ofdm and vector coding based on singular value decomposition with water - filling power allocation .
a major implication of the derived results is that separate encoding across subcarriers per link is sufficient to linearly increase the sum - spectral efficiency with @xmath0 in the interference channel with isi .
simulation results support this claim . |
observations of spatial patterns at various length scales frequently are the only point where the physical world meets theoretical models . in many cases
these patterns consist of a number of comparable objects distributed in space such as pores in a sandstone , or craters on the surface of a planet .
another example is given in figure [ fig : kerscher_galaxies - circles ] , where we display the galaxy distribution as traced by a recent galaxy catalogue .
the galaxies are represented as circles centered at their positions , whereas the size of the circles mirrors the luminosity of a galaxy . in order to test to which extent theoretical predictions fit the empirically found structures of that type
, one has to rely on quantitative measures describing the physical information .
since theoretical models mostly do not try to explain the structures individually , but rather predict some of their generic properties , one has to adopt a _
statistical point of view _ and to interpret the data as a realization of a random process . in a first step one often confines oneself to the spatial distribution of the objects constituting the patterns and investigates their clustering thereby thinking of it as a realization of a _
point process_. assuming that perspective , however , one neglects a possible linkage between the spatial clustering and the intrinsic properties of the objects .
for instance , there are strong indications that the clustering of galaxies depends on their luminosity as well as on their morphological type . considering figure [ fig : kerscher_galaxies - circles ]
, one might infer that luminous galaxies are more strongly correlated than dim ones .
effects like that are referred to as _ mark segregation _ and provide insight into the generation and interactions of , e.g. , galaxies or other objects under consideration . the appropriate statistical framework to describe the relation between the spatial distribution of physical objects and their inner properties are _ marked point processes _ , where discrete , scalar- , or vector - valued marks are attached to the random points .
+ in this contribution we outline how to describe marked point processes ; along that line we discuss two notions of independence ( section [ sec : kerscher_basic ] ) and define corresponding statistics that allow us to quantify possible dependencies . after having shown that some empirical data sets show significant signals of mark segregation ( section[sec : kerscher_data ] ) , we turn to analytical models , both motivated by mathematical and physical considerations ( section [ sec : kerscher_models ] ) .
+ contact distribution functions as presented in the contribution by d. hug et al . in this volume
are an alternative technique to measure and statistically quantify distances which finally can be used to relate physical properties to spatial structures .
mark correlation functions are useful to quantify molecular orientations in liquid crystals ( see the contribution by f. schmid and n. h. phuong in this volume ) or in self - assembling amphiphilic systems ( see the contribution by u. s. schwarz and g. gompper in this volume ) . but also to study anisotropies in composite or porous materials , which are essential for elastic and transport properties ( see the contributions by d. jeulin , c. arns et al . and
vogel in this volume ) , mark correlations may be relevant .
the empirical data the positions @xmath1 of some objects together with their intrinsic properties @xmath2 are interpreted as a realization of a marked point process @xmath3 ( stoyan , kendall and mecke , 1995 ) . for simplicity
we restrict ourselves to homogeneous and isotropic processes .
+ the hierarchy of joint probability densities provides a suitable tool to describe the stochastic properties of a marked point process .
thus , let @xmath4 denote the probability density of finding a point at @xmath5 with a mark @xmath6
. for a homogeneous process this splits into @xmath7 where @xmath8 denotes the mean number density of points in space and @xmath9 is the probability density of finding the mark @xmath6 on an arbitrary point . later on we need moments of this mark distribution ; for real - valued marks the @xmath10th - moment of the mark - distribution is defined as @xmath11 the mark variance is @xmath12 .
+ accordingly , @xmath13 quantifies the probability density to find two points at @xmath14 and @xmath15 with marks @xmath16 and @xmath17 , respectively ( for second - order theory of marked point processes see @xcite ) .
it effectively depends only on @xmath16 , @xmath17 , and the pair separation @xmath18 for a homogeneous and isotropic process .
two - point properties certainly are the simplest non - trivial quantities for homogeneous random processes , but it may be necessary to move on to higher correlations in order to discriminate between certain models . in the following we will discuss two notions of independence , which may arise for marked point patterns . for this , consider two renaissance families , call them the sforza and the gonzaga .
they used to build castles spread out more or less homogeneously over italy . in order to describe this example in terms of a marked point process
, we consider the locations of the castles as points on a map of italy , and treat a castle s owner as a discrete mark , @xmath19 and @xmath20 , respectively .
there are many ways how the castles can be built and related to each other .
[ [ independent - sub - point - processes ] ] independent sub - point processes : + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + for example , the sforza may build their castles regardless of the gonzaga castles . in that case the probability of finding a sforza castle at @xmath14 and a gonzaga castle at @xmath15 factorizes into two one - point probabilities and we can think of the sforza and the gonzaga castles as uncorrelated sub - point processes . in the language of marked point processes this means , e.g. , that @xmath21 for any @xmath22 .
if all the joint @xmath23-point densities factorize into a product of @xmath24-point densities of one type each , then we speak of _ independent sub - point processes_. dependent sub - point processes indicate _ interactions _ between points of different marks ; for instance , the gonzaga may build their castles close to the sforza ones in order to avoid that a region becomes dominated by the other family s castles .
[ [ mark - independent - clustering ] ] mark - independent clustering : + + + + + + + + + + + + + + + + + + + + + + + + + + + + a second type of independence refers to the question whether the different families have different styles to plan their castles .
for instance , the gonzaga may distribute their castles in a grid - like manner over italy , whereas the sforza may incline to build a second castle close to each castle they own .
rather than asking whether two sub - point processes ( namely the gonzaga and the sforza castles , respectively ) are independent ( `` independent sub - point processes '' ) , we are now discussing whether they are _ different _ as regards their statistical clustering properties .
any such difference means that the clustering _ depends _ on the intrinsic mark of a point .
+ whenever the two - point probability density of finding two objects at @xmath14 and @xmath15 depends on the objects intrinsic properties we speak of _ mark - dependent clustering_. it is useful to rephrase this statement by using bayes theorem and the conditional mark probability density @xmath25 in case the spatial product density @xmath26 does not vanish .
@xmath27 is the probability density of finding the marks @xmath16 and @xmath17 on objects located at @xmath14 and @xmath15 , given that there are objects at these points .
clearly , @xmath27 depends only on the pair separation @xmath28 for homogeneous and isotropic point processes .
we speak of _ mark - independent _ clustering , if @xmath29 factorizes @xmath30 and thus does not depend on the pair separation .
that means that regarding their marks , pairs with a separation @xmath31 are not different from any other pairs . on the contrary , mark - dependent clustering or _ mark segregation _
implies that the marks on certain pairs show deviations from the global mark distribution .
+ in order to distinguish between both sorts of independencies , let us consider the case where we are given a map of italy only showing the gonzaga castles . if the distribution of castles in italy can be understood as consisting of independent sub - point processes , we can not infer anything about the sforza castles from the gonzaga ones .
however , if @xmath32 , sforza castles are likely to be found close to gonzaga ones .
here , @xmath33 and @xmath34 are the probabilities that a castle belongs to the sforza or gonzaga family .
if , on the other hand , mark - independent clustering applies , typical clustering properties such as the spatial clustering strength are equal for both castle distributions , and the gonzaga castles are in the statistical sense already representative of the whole castle distribution in italy .
that means in particular that , if the gonzaga castles are clustered , so are the sforza ones .
before we turn to applications , we have to develop practical test quantities in order to test for segregation effects in real data and to describe them in more detail . to investigate correlations between sub - point processes ,
suitably extended nearest neighbor distribution functions or @xmath35-functions have been employed @xcite .
also the ( conditional ) cross - correlation functions can be used ( see eq .
[ eq : kerscher_cond - crosscorr ] ) , for a further test see @xcite , p. 302 . here
we consider a multivariate extension of the @xmath36-function @xcite , as suggested by @xcite .
+ for this , consider the nearest neighbor s distance distribution from an object with mark @xmath2 to other objects with mark @xmath37 , @xmath38 ( `` @xmath39 to @xmath40 '' , for details see @xcite ) .
let @xmath41 denote the distribution of the nearest neighbor s distance from an object of type @xmath39 to any other object ( denoted by @xmath42 ) . finally , @xmath43 is the nearest neighbor distribution of all points .
similar extensions of the empty space function are possible , too .
let @xmath44 denote the distribution of the nearest @xmath39-object s distance from an arbitrary position , whereas @xmath45 is the nearest object s distance distribution from a random point in space to any object in the sample .
we consider the following quantities : @xmath46 they are defined whenever @xmath47 .
if two sub - point processes , defined by marks @xmath48 , are independent then one gets @xcite @xmath49 note , that the @xmath50 depend on higher - order correlations functions , similar to the @xmath36-function @xcite .
suitable estimators for these @xmath36-functions are derived from estimators of the @xmath51 and @xmath20-functions @xcite . in order to quantify the mark - dependent clustering or to look for the mark segregation
, it proves useful to integrate the conditional probability density @xmath29 over the marks weighting with a test function @xmath52 @xcite .
this procedure reduces the number of variables and leaves us with the weighted pair average : @xmath53 the choice of an appropriate weight - function depends on whether the marks are non - quantitative labels or continuous physical quantities .
1 . for labels only combinations of indicator - functions are possible , the integral degenerates into a sum over the labels . supposed
the marks of our objects belong to classes labelled with @xmath54 , the conditional cross - correlation functions are given by @xmath55 with the kronecker @xmath56 for @xmath57 and zero otherwise .
mark segregation is indicated by @xmath58 for @xmath59 and @xmath60 , where @xmath61 denotes the number density of points with label @xmath39 .
the @xmath62 are cross - correlation functions under the _ condition _ that two points are separated by a distance of @xmath31 ( compare @xcite , p. 264
, for applications see the martian crater distribution studied in sect . [ sec : kerscher_martian ] and figure [ fig : kerscher_mars - markcorr ] in particular ) .
2 . for positive real - valued marks @xmath6 ,
the following pair averages prove to be powerful and distinctive @xcite : + 1 .
one of the most simplest weights to be used is the mean mark : @xmath63 quantifies the deviation of the mean mark on pairs with separation @xmath31 from the overall mean mark @xmath64 .
a @xmath65 indicates mark segregation for point pairs with a separation @xmath31 , specifically their mean mark is then larger than the overall mark average .
+ closely related is stoyan s @xmath66 function using the squared geometric mean of the marks as a weight @xcite @xmath67 2 .
accordingly , higher moments of the marks may be used to quantify mark segregation , like the mark fluctuations @xmath68 or the mark - variogram @xcite : @xmath69 3 .
the mark covariance @xcite is @xmath70 mark segregation can be detected by looking whether @xmath71 differs from zero .
a @xmath71 larger than zero , e.g. , indicates that points with separation @xmath31 tend to have similar marks .
sometimes the mark covariance is normalized by the fluctuations @xcite : @xmath72 .
+ these conditional mark correlation functions can be calculated from only three independent pair averages @xcite : @xmath73 , @xmath74 , and @xmath75 .
thus the above mentioned characteristics are not independent , e.g.@xmath76 .
+ we apply these mark correlation functions to the galaxy distribution in section [ sec : kerscher_gal ] ( figure [ fig : kerscher_ssrs - lum ] ) , to martian craters in section [ sec : kerscher_martian ] ( figure [ fig : kerscher_mars - markcorr ] ) and to pores in sandstones considered in section [ sec : kerscher_sandstone ] .
3 . also vector - valued information @xmath77 , describing , e.g. , the orientation of an anisotropic object at position @xmath1 may be available .
it is therefore interesting to consider vector marks such as done by @xcite who use a mark correlation function to quantify the alignment of vector marks .
here we suggest three mark correlation functions quantifying geometrically different possibilities of an alignment . in order to ensure coordinate - independence of our descriptors , we focus on scalar combinations of the vector marks in using the scalar product @xmath78 and the cross product @xmath79 .
different from the case of scalar marks , it is a non - trivial task to find a set of vector - mark correlation functions which contain all possible information ( at least up to a fixed order in mark space ) . we provide a systematic account of how to construct suitable vector - mark correlation functions in a complete and unique way for general dimensions in the appendix .
+ here we only cite the most important results . for
that we need the distance vector between two points , @xmath80 , the normalized distance vector , @xmath81 , and the normalized vector mark : @xmath82 with @xmath83 .
the following conditional mark correlation functions will be used to quantify alignment effects : 1 .
@xmath84 quantifies the @xmath85lignment of the two vector marks @xmath86 and @xmath87 : @xmath88 it is proportional to the cosine of the angle between @xmath86 and @xmath87 .
we normalize with the mean @xmath89 . for purely independent vector marks @xmath84
is zero , whereas @xmath90 means that the marks of pairs separated by @xmath31 tend to align parallel to each other .
in some applications , e.g. for the orientations of ellipsoidal objects , the vector mark is only defined up to a sign , i.e. @xmath91 and @xmath92 mean actually the same . in this case
the absolute value of the scalar product is useful : @xmath93 for uncorrelated random vectors we get @xmath94 .
@xmath85 and @xmath95 can readily be generalized to any dimension @xmath96 , where we expect @xmath97 @xmath98 for uncorrelated random orientations . in two dimensions
@xmath95 is proportional to @xmath99 as defined by @xcite .
@xmath100 quantifies the @xmath101ilamentary alignment of the vectors @xmath86 and @xmath87 with respect to the line connecting both halo positions : @xmath102 @xmath100 is proportional to the cosine of the angle between @xmath86 and the distance vector @xmath103 connecting the points . for uncorrelated random vector marks ,
we expect again @xmath104 ; @xmath100 becomes larger than that , whenever the vector marks of the objects tend to point to objects separated by @xmath31
an example is provided by rod - like metallic grains in an electric field : they concentrate along the field lines and orient themselves parallel to the field lines . 3 . @xmath105 quantifies the @xmath106lanar alignment of the vectors and the distance vector .
@xmath105 is proportional to the volume of the rhomb defined by @xmath86 , @xmath87 and @xmath103 : @xmath107 quite obviously , this quantity can not be generalized to arbitrary dimensions ; the deeper reason for that will become clear in the appendix . we get @xmath108 for randomly oriented vectors , whereas it is becoming larger for the case that @xmath87 is perpendicular to @xmath86 as well as to @xmath103 . + applications of vector marks
can be found in section [ sec : kerscher_halos ] ( figure [ fig : kerscher_halos - orientation ] ) where we consider the orientation of dark matter halos in cosmological simulations .
but one can think of other applications : mark correlation functions may serve as orientational order parameters in liquid crystals in order to discriminate between nemetic and smectic phases ( see the contribution by f. schmid and n. h. phuong in this volume ) .
they can also quantify the local orientation and order in liquids such as the recently measured five - fold local symmetry found in liquid lead @xcite . as a further application one could try to measure
the signature of hexatic phases in two - dimensional colloidal dispersions and in 2d melting scenarios occurring in experiments and simulations of hard - disk systems ( for a review on hard sphere models see @xcite . finally , the orientations of anisotropic channels in sandstone ( see the contribution by c. arns et al . in this volume ) are relevant for macroscopic transport properties , therefore their quantitative characterization in terms of mark correlation functions might be interesting .
before we move on to applications a few general remarks are in order : first , the definition of these mark characteristics based on the conditional density @xmath109 leads to ambiguities at @xmath31 equal zero as discussed by @xcite , but there is no problem for @xmath110 .
furthermore , suitable estimators for our test quantities are based on estimators for the usual two - point correlation function @xcite .
+ mark - dependent clustering can also be defined at any @xmath23-point level .
mark - independent clustering at every order is called the random labelling property @xcite .
mark correlation functions based on the @xmath23-point densities may be used .
for discrete marks the multivariate @xmath36-functions ( see eq . ) are an interesting alternative , sensitive to higher - order correlations .
the random labelling property then leads to the relation @xmath111 which may be used as a test @xcite .
in many cases already the question whether one or the other type of dependence as outlined above applies to certain data sets is a controversial issue . in the following we will apply our test quantities to a couple of data sets in order to probe whether there is an interplay between some objects marks and their positions in space .
other applications to biological , ecological , mineralogical , geological data can be found in @xcite .
the distribution of galaxies in space shows a couple of interesting features and challenges theoretical models trying to understand cosmological structure formation ( see e.g.@xcite ) . there has been a long debate , whether and how strongly the clustering of galaxies depends on their luminosity and their morphological type ( see , e.g.@xcite ) .
the methods which have been used so far to establish such claims were based on the spatial two - point correlation function ; it was estimated from different subsamples that were drawn from a catalogue and defined by morphology or luminosity .
however , some authors claimed that the signal of luminosity segregation observed by others was a spurious effect , caused by inhomogeneities in the sample and an inadequate choice of the statistics @xcite .
@xcite could show that methods based on the mark - correlation functions , as discussed in sect .
[ sec : kerscher_mark - segregation ] , are not impaired by inhomogeneities , and found a clear signal of luminosity and morphology segregation . + in order to quantify segregation effects in the galaxy distribution we consider the southern sky redshift survey 2 ( ssrs 2 , @xcite ) , which maps a significant fraction of the sky and provides us with the angular sky positions , the distances ( determined via the redshifts ) , and some intrinsic properties of the galaxies such as their flux and their morphological type . as marks
we consider either a galaxy s luminosity estimated from its distance and flux , or its morphological type . in the latter case
we effectively divide our sample into early - type galaxies ( mainly elliptical galaxies ) and late - type galaxies ( mainly spirals ) . in order to analyze homogeneous samples ,
we focus on a volume - limited sample of @xmath112 depth equals roughly @xmath113 million light years .
the number @xmath114 accounts for the uncertainty in the measured hubble constant and is about @xmath115 .
volume - limited samples are defined by a limiting depth and a limiting luminosity .
one considers only those galaxies which could have been observed if they were located at the limiting depth of the sample . ]
@xcite .
+ in a first step we ask whether the early- and the late - type galaxies form independent sub - processes . in figure
[ fig : kerscher_morph_ind ] we show @xmath116 as function of the distance @xmath31 being far away from the value of one .
recalling eq .
, we conclude that the morphological types of galaxies are not distributed independently on the sky .
not surprisingly , the inequality @xmath117 indicates positive interactions between the galaxies of both morphological types ; indeed galaxies attract each other through gravity irrespective of their morphological types .
+ after having confirmed the presence of interactions between the different types of galaxies , we tackle the issue whether the clustering of galaxies is different for different galaxies .
we consider the luminosities as marks ( see fig .
[ fig : kerscher_galaxies - circles ] ) . in figure
[ fig : kerscher_ssrs - lum ] we show some of the mark - weighted conditional correlation functions . already at first glance , they show evidence for luminosity segregation , relevant on scales up to @xmath118 . to strengthen our claims
, we redistribute the luminosities of the galaxies within our sample randomly , holding the galaxy positions fixed . in that way we mimic a marked point process with the same spatial clustering and the same one - point distribution of the luminosities , but without luminosity segregation . comparing with the fluctuations around this null hypothesis
, we see that the signal within the ssrs 2 is significant .
+ the details of the mark correlation functions provide some further insight into the segregation effects .
the mean mark @xmath119 indicates that the luminous galaxies are more strongly clustered than the dim ones .
our signal is scale - dependent and decreasing for higher pair separations .
the stronger clustering of luminous galaxies is in agreement with earlier claims comparing the correlation amplitude of several volume - limited samples @xcite .
+ the @xmath120 being larger than the mark variance of the whole sample , @xmath121 , shows that on galaxy pairs with separations smaller than @xmath118 the luminosity fluctuations are enhanced . the fact that the mark segregation effect extends to scales of up to @xmath118 is interesting on its own . in particular
, it indicates that galaxy clusters are not the only source of luminosity segregation , since typically galaxy clusters are of the size of @xmath122 .
+ the signal for the covariance @xmath71 , however , could be due to galaxy pairs inside clusters . it is relevant mainly on scales up to @xmath123 indicating that the luminosities on galaxy pairs with small separations tend to assume similar values .
our results in part confirm claims by @xcite , who compared the correlation functions @xmath124 for different volume - limited subsamples and different luminosity classes of the ssrs 2 catalog ( see also @xcite ) .
many structures found in the universe such as galaxies and galaxy clusters show anisotropic features . therefore one can assign orientations to them and ask whether these orientations are correlated and form coherent patterns . here
we discuss a similar question on the base of numerical simulations of large scale structure ( e.g. , @xcite ) .
+ in such simulations the trajectories of massive particles are numerically integrated .
these particles represent the dominant mass component in the universe , the dark matter . through gravitational instability high density peaks ( `` halos '' ) form in the distribution of the particles ; these halos are likely to be the places where galaxies originate . in the following
we will report on alignment correlations between such halos @xcite , for a further application of mark correlation functions in this field see @xcite .
+ the halos used by @xcite stem from a @xmath0-body simulation in a periodic box with a side length of 500 .
the initial and boundary conditions were fixed according to a @xmath125cdm cosmology ( for a discussion of cosmological models see @xcite ) .
halos were identified using a friend - of - friends algorithm in the dark matter distribution .
not all of the halos found were taken into account ; rather the mass range and the spatial number density of the selected halos were chosen to resemble the properties of observed galaxy clusters in the reflex catalogue @xcite .
typically our halos show a prolate distribution of their dark matter particles .
+ for each halo the direction of the elongation is determined from the major axis of the mass - ellipsoid .
this leads to a marked point set where the orientation @xmath77 is attached to each halo position @xmath1 as a vector mark with @xmath126 .
details can be founds in @xcite .
+ in fig .
[ fig : kerscher_halos - orientation ] the vector - mark correlation functions as defined in eqs . , , and are shown . since only the orientation of the mass ellipsoids can be determined , we use @xmath127 ( eq .
[ eq : kerscher_def - vector - corr - absolute ] ) instead of @xmath84 .
the signal in @xmath127 indicates that pairs of halos with a distance smaller than 30 show a tendency of parallel alignment of their orientations @xmath128 .
the deviation from a pure random alignment is in the percent range but clearly outside the random fluctuations .
the alignment of the halos orientations @xmath128 with the connecting vector @xmath103 quantified by @xmath100 is significantly stronger ; it is particularly interesting that this alignment effect extends to scales of about 100 .
+ in a qualitative picture this may be explained by halos aligned along the filaments of the large scale structure .
indeed such filaments are prominent features found in the galaxy distribution @xcite and in @xmath0-body simulations @xcite , often with a length of up to 100 .
the lowered @xmath105 indicates that the volume of the rhomboid given by @xmath128 and @xmath103 is reduced for halo pairs with a separation below 80 .
already a preferred alignment of @xmath128 along @xmath103 leads to such a reduction , similar to a plane - like arrangement of @xmath129 . for the halo distribution the signal in @xmath105 seems to be dominated by the filamentary alignment .
+ the question whether there are non - trivial orientation patterns for galaxies or galaxy clusters has been discussed for a long time .
@xcite reported a significant alignment of the observed galaxy clusters out to 100 .
@xcite , however claimed that this effect is small and likely to be caused by systematics ; @xcite find no indication for alignment effects at all .
subsequently several authors purported to have found signs of alignments in the galaxy and galaxy cluster distribution ( see e.g.@xcite ) .
our fig .
[ fig : kerscher_halos - orientation ] shows that from simulations significant large - scale correlations are to be expected in the orientations of galaxy clusters , in agreement with the results by @xcite .
these results are also supported by a simulation study carried out by @xcite .
let us now turn to another , still astrophysical , but significantly closer object : the mars ( see figure [ fig : kerscher_mars_eg ] ) .
many planets surfaces display impact craters with diameters up to @xmath130 km and a broad range of inner morphologies .
these craters are surrounded by ejecta forming different types of patterns .
the craters and their ejecta are likely to be caused by asteroids and periodic comets crossing the planets orbits , falling down onto the planet s surface , and spreading some of the underlying surfaces material around the original impact crater . a variety of different crater morphologies and a wide range of ejecta patterns can be found . in principle ,
either the different impact objects ( especially their energies ) or the various surface types of the planet may explain the repertory of patterns observed . whereas the energy variations of impact objects do not cause any peculiarities in the spatial distribution of the craters ( apart from a possible latitude dependence ) , geographic inhomogeneities
are expected to originate inhomogeneities in the craters morphological properties .
+ we try to answer the question for the ejecta patterns origin using data collected by @xcite who already found correlations between crater characteristics and the local surface type employing geologic maps of the mars .
complementary to their approach , we investigate two - point properties without any reference to geologic mars maps .
we restrict ourselves only to craters which have a diameter larger than @xmath131 km and whose ejecta pattern could be classified , ending up with @xmath132 craters spread out all over the martian surface .
we use spherical distances for our analysis of pairs .
+ in a first step we divide the ejecta patterns into two broad classes consisting of either the simple patterns ( single and double lobe morphology , i.e. sl and dl in terms of the classification by @xcite ; we speak of `` simple craters '' ) or the remaining , more complex configurations ( `` complex craters '' ) .
using our conditional cross correlation functions @xmath62 as defined in equation , we see a highly significant signal for mark correlations ( figure [ fig : kerscher_mars - crosscorr ] ) . at small separations ,
crater pairs are disproportionally built up of simple craters at the expense of cross correlations .
this can be explained assuming that crater formation depends on the local surface type : if the simple craters are more frequent in certain geological environments than in others , then there are also more pairs of them to be found as far as one focuses on distances smaller than the typical scale of one geological surface type .
cross pairs are suppressed , since typical pairs with small separations belong to one geological setting where the simple craters either dominate or do not . only a small , positive segregation signal occurs for the complex craters .
hence our analysis indicates that the broad class of complex craters is distributed quite homogeneously over all of the geologies . on top of this
there are probably simple craters , their frequency significantly depending on the surface type .
+ if the ejecta patterns were independent of the surface , no mark segregation could be observed ( other sources of mark segregation are unlikely , since the martian craters are a result of a long bombardment history diluting any eventual peculiar crater correlations ) . in this sense , the signal observed indicates a surface - dependence of crater formation .
this result is remarkable , given that we did not use any geological information on the mars at all .
the picture emerging could be described using the random field model , where a field ( here the surface type ) determines the mark of the points ( see below ) .
+ in a second step , we analyze the interplay between the craters diameters and their spatial clustering .
now the diameter serves as a continuous mark .
the results in figure [ fig : kerscher_mars - markcorr ] show a clear signal for mark segregation in @xmath133 and @xmath134 at small scales .
the latter signals that pairs with separations in a broad range up to @xmath135 km tend to have similar diameters ; this is in agreement with the earlier picture : as @xcite showed , the simple craters are mostly small - sized .
pairs with relatively small separations thus often stem from the same geological setting and therefore have similar diameters and similar morphological type .
+ also the signal of @xmath133 seems to support this picture : since the simple craters are more strongly clustered than the other ones and since they have smaller diameters , one could expect @xmath136 . as we shall see in sect .
[ sec : kerscher_models ] , however , a @xmath137 contradicts the random field model ; therefore , the mark - dependence on the underlying surface type ( thought of as a random field ) can not account for the signal observed .
thus , we have to look for an alternative explanation : it seems reasonable , that , whenever a crater is found somewhere , no other crater can be observed close nearby ( because an impact close to an existing crater will either destroy the old one or cover it with ejecta such that it is not likely to be observed as a crater ) .
this results in a sort of effective hard - core repulsion .
this repulsion should be larger for larger craters .
thus , pairs with very small separations can only be formed by small craters , therefore @xmath136 for tiny @xmath31 .
the scale beyond which @xmath138 should somehow be hidden within the crater diameter distribution . indeed , at about @xmath139 km the segregation vanishes , which is about twice the largest diameter in our sample .
taking into account that the ejecta patterns extend beyond the crater , this seems to be a reasonable agreement . as shown in sect .
[ sec : kerscher_boolean - depletion ] a model based on these consideration is able to produce such a depletion in the @xmath140 .
this effect could also in turn explain part of the cross correlations observed earlier in figure [ fig : kerscher_mars - crosscorr ] .
a similar effect is to be expected for the mark variance .
close pairs are only accessible to craters with a smaller range of diameters ; therefore , their variance is diminished in comparison to the whole sample . however , an effect like this is barely visible in the data .
+ altogether , the crater distribution is dominated by two effects : the type of the ejecta pattern and the crater diameter depend on the surface , in addition , there is a sort of repulsion effect on small scales .
now we turn to systems on smaller scales .
sandstone is an example of a porous medium and has extensively been investigated , mainly because oil was found in the pore network of similar stones . in order to extract the oil from the stone one can try to wash it out using a second liquid , e.g. water .
therefore , one tries to understand from a theoretical point of view , how the microscopic geometry of the pore network determines the macroscopic properties of such a multi - phase flow .
especially the topology and connectivity of the microcaves and tunnels prove to be crucial for the flow properties at macroscopic scales .
details are given , for instance , in the contributions by c. arns et al .
, h .- j .
vogel et al . and j. ohser in this volume .
a sensible physical model , therefore , in the first place has to rely on a thorough description of the pore pattern .
+ one way to understand the pore network is to think of it as a union of simple geometrical bodies .
following @xcite , one can identify distinct pores together with their position and their pore radius or extension .
this allows us to understand the pore structure in terms of a marked point process , where the marks are the pore radii .
+ in the following , we consider three - dimensional data taken from one of the fontainbleau sandstone samples through synchrotron x - ray tomography .
these data trace a @xmath141 mm diameter cylindrical core extracted from a block with bulk porosity @xmath142 , where the bulk porosity is the volume fraction occupied by the pores . a piece with @xmath143 mm length ( resulting in a @xmath144 mm@xmath145 volume ) of the core was imaged and tomographically reconstructed @xcite .
further details of this sample are presented in the contribution by c. arns et al . in this volume .
based on the reconstructed images the positions of pores and their radii were identified as described in @xcite .
+ in our results for the mark correlation functions a strong depletion of @xmath140 and @xmath146 is visible for @xmath147 m in fig .
[ fig : kerscher_sand ] .
this small - scale effect may be explained similarly to the martian craters : large pores are never found close to each others , since they have to be separated by at least the sum of their radii .
the histogram of the pore radii in fig .
[ fig : kerscher_sand - hist ] shows that most of the pores have radii smaller than 100@xmath148 m , and consequently this effect is confined to @xmath147 m . in sect .
[ sec : kerscher_boolean - depletion ] we discuss the boolean depletion model which is based on this geometric constraints and is able to produce such a reduction in the @xmath140 .
this purely geometric constraint also explains the reduced @xmath146 and increased covariance @xmath149 . for separations larger than @xmath150
m there is no signal from the covariance , but both @xmath140 and @xmath146 show a small increase out to @xmath151 m .
this indicate that pairs of pores out to these separations tend to be larger in size and show slightly increased fluctuations .
however , this effect is small ( of the order of @xmath152 ) and may be explained by the definition of the holes , which may lead to `` artificial small pores '' as `` bridges '' between larger ones .
this hypothesis has to be tested using different hole definitions . in any case
the main conclusion seems to be that apart from the depletion effect at small scales there are no other mark correlations .
given the significant mark correlations found in various applications , one may ask how these signals can be understood in terms of stochastic models . a thorough understanding of course requires a physical modeling of the individual situation .
there are , however , some generic models , which we will focus on in the following : in sect . [ sec : kerscher_boolean - depletion ] we introduce the boolean depletion model , which is able to explain some of the features observed in the distribution of craters and pores in sandstone .
another generic model is the _ random field model _ where the marks of the points stem from an independent random field ( section [ sec : kerscher_rfm ] ) . in sect .
[ sec : kerscher_cox - rf - model ] we generalize the idea behind the random field model further in order to get the _ cox random field model _ , which allows for correlations between the point set and the random field . other model classes and their applications
are discussed by e.g.@xcite . in our analysis of the martian craters and the holes in sandstone , we found that for small separations only small craters , or small holes in the sandstone , could be found .
we interpreted this as a pure geometric selection effect .
the boolean depletion model is able to quantify this effect , but also shows further interesting features .
+ the starting point is the boolean model of overlapping spheres @xmath153 ( see also the contributions by c. arns et al . and d. hug in this volume as well as @xcite ) .
for that , the spheres centers @xmath1 are generated randomly and independently , i.e. according to a poisson process of number density @xmath154 .
the radii @xmath155 of the spheres are then chosen independently according to a distribution function @xmath156 , i.e. with probability density @xmath157 .
the main idea behind the depletion is to delete spheres which are covered by other spheres . to make this procedure unique we remove only those spheres which are completely covered by a ( notably larger ) sphere .
the positions and radii of the remaining spheres define a marked point process .
note , that this depletion mechanism is minimal in the sense that a lot of overlapping spheres may remain .
this boolean depletion model may be considered as the low - density limit of the well - known widom - rowlinson model , or ( more generally ) of non - additive hard sphere mixtures ( see @xcite ) .
+ the probability that a sphere of radius @xmath155 is not removed is then given by @xmath158 with the step function @xmath159 for @xmath160 and @xmath161 otherwise , and the volume of the @xmath96-dimensional unit ball @xmath162 ( @xmath163 , @xmath164 , @xmath165 ) .
the limit in eq .
is performed by keeping @xmath166 constant , with @xmath0 the initial number of spheres and @xmath167 the volume of the domain .
+ the number density of the remaining spheres reads @xmath168 where the one - point probability density @xmath169 that a sphere has radius @xmath155 is given by @xmath170 the probability that one or both of the spheres @xmath171 and @xmath172 are not removed is given by @xmath173 with @xmath174 and @xmath175 at this point we have to consider the set union of two spheres with radii @xmath176 and @xmath177 , respectively ; the volume of this geometrical configuration can be calculated ; in three dimensions , e.g. , we have : @xmath178 for @xmath179
. otherwise this volume reduces either to the volume of the larger sphere ( @xmath180 ) or to the sum of both spherical volumes ( @xmath181 ) .
+ similarly as in eq . the spatial two - point density turns out to be @xmath182 such that the conditional two - point mark density simply reads @xmath183 from this we can derive all of the mark correlation functions from sect . [
sec : kerscher_mark - segregation ] . in order to get an analytically tractable model we adopt a bimodal radius distribution in the original boolean model and start therefore with @xmath184 where we assume that @xmath185 . due to the depletion the number density @xmath8 of the spheres as well as the probability @xmath186
to find the smaller radius @xmath187 at a given point are then lowered ; we get @xmath188 with @xmath189 .
altogether , the bimodal model can be parameterized in terms of the radii @xmath190 @xmath187 , the ratio @xmath191 $ ] and the density @xmath192 .
the latter two quantities , however are not observable from the final point process , therefore we convert them into the parameters @xmath193 $ ] and @xmath194 , so that all other quantities can be expressed in terms of these , for instance , @xmath195 , and @xmath196 , + from eq .
we determine the mean mark , i.e. the mean radius of the spheres @xmath197 and from eq . the spatial product density @xmath198 1+\alpha^2 \left[\exp\left(ni(x)\right)-1\right ] & 1 \leq x < 2 , \\[1ex ] 1 & 2 \leq x , \end{cases}\ ] ] with the normalized inter - sectional volume @xmath199 of two spheres and @xmath200 .
finally , using eq .
one can calculate the mark correlation functions , e.g. @xmath201 1-\alpha^2(1-\alpha ) \frac{r_2-r_1}{\overline{r } } \frac{\exp\left(n i(x ) \right)-1 } { 1+\alpha^2\left[\exp\left(n i(x ) \right)-1\right ] } & 1 \leq x < 2 , \\[1ex ] 1 & 2\leq x .
\end{cases}\label{eq : kerscher_km_depl}\ ] ] in fig .
[ fig : kerscher_depletion ] the @xmath140 function from the boolean depletion model is shown .
the model with the solid line illustrates that a reduced @xmath140 for small radii can be obtained by simply removing smaller spheres .
at least qualitatively this model is able to explain the depletion effects we have seen both in the distribution of martian craters ( fig .
[ fig : kerscher_mars - markcorr ] ) and in the distribution of pores in sandstone ( fig .
[ fig : kerscher_sand ] ) .
the jump at @xmath202 is a relict of the strictly bimodal distribution with only two radii .
[ fig : kerscher_depletion ] also shows that the boolean depletion model is quite flexible , allowing for a @xmath203 , but also @xmath119 is possible . + without ignoring the considerable difference of this boolean depletion model to the pore size distribution in real sandstones ( see figures [ fig : kerscher_sandstone_eg]-[fig : kerscher_sand ] ) one may still recognize some interesting similarities : this simple model explains naturally a decrease of @xmath140 if the distribution of the radii is symmetric ( @xmath204 ) . as visible in figure
[ fig : kerscher_sand - hist ] this is approximately the case for the pore radii .
moreover , note that even quantitative features are captured correctly indicating that the decrease of @xmath140 visible in figure [ fig : kerscher_sand ] is indeed due to a depletion effect .
for instance , the decrease starts at @xmath205 where @xmath206 is the largest occurring radius ( see the histogram in figure [ fig : kerscher_sand - hist ] ) and the value of @xmath207 at @xmath208 is in accordance with equation assuming that @xmath209 and the normalized density of pores @xmath210 necessary for a connected network .
of course a more detailed analysis is necessary based on eqs . and and the histogram shown in figure [ fig : kerscher_sand - hist ] .
the `` random - field model '' covers a class of models motivated from fields such as geology ( see , e.g. , @xcite ) .
the level of the ground water , for instance , is thought of as a realization of a random field which may be directly sampled at points ( hopefully ) independent from the value of the field or which may influence the size of a tree in a forest .
+ in general , a realization of the random field model is constructed from a realization of a point process and a realization of a random field @xmath211 .
the mark of each object located at @xmath1 traces the accompanying random field via @xmath212 .
the crucial assumption is that the point process is stochastically independent from the random field .
+ we denote the mean value of the homogeneous random field by @xmath213=\overline{u^1}$ ] and the moments by @xmath214 , with the one - point probability density @xmath215 of the random field and @xmath216 the expectation over realizations of the random field .
the product density of the random field is @xmath217 $ ] with @xmath28 . for a general discussion of random field models , see @xcite .
+ in this model the one - point density of the marks is @xmath218 , and @xmath219 etc .
the conditional mark density is given by @xmath220,\ ] ] where @xmath221 is the dirac delta distribution .
clearly , this expression is only well - defined under a suitable integral over the marks . with eq .
one obtains @xmath222 and the mark - correlation functions defined in sect .
[ sec : kerscher_mark - segregation ] read @xmath223 therefore , there are some explicit predictions for the random field model : an empirically determined @xmath133 significantly differing from one not only indicates mark segregation , but also that the data is incompatible with the random field model . looking at figure [ fig : kerscher_ssrs - lum ] we see immediately that the galaxy data are not consistent with the random field model .
similar tests based on the relation between @xmath66 and the mark - variogram @xmath224 were investigated by @xcite and @xcite .
the failure of the random field model to describe the luminosity segregation in the galaxy distribution allows the following plausible physical interpretation : the galaxies do not merely trace an independent luminosity field ; rather the luminosities of galaxies depend on the clustering of the galaxies . we shall try to account for this with a better model in the following section . in the random field model , the field
was only used to generate the points marks . in the cox random field model , on the contrary
, the random field determines the spatial distribution of the points as well . as before , consider a homogeneous and isotropic random field @xmath225 .
the point process is constructed as a cox - process ( see e.g. @xcite ) .
the mean number of points in a set @xmath226 is given by the intensity measure @xmath227 where @xmath228 is a proportionality factor fixing the mean number density @xmath229 .
the ( spatial ) product density of the point distribution is @xmath230 where again @xmath231 denotes the product density of the random field .
@xmath232 is the normalized two - point cumulant of the random field ( see below ) .
we will also need the @xmath23-point densities of the random field : @xmath233.\ ] ] like in the random field model , the marks trace the field , but this time rather in a probabilistic way than in a deterministic one : the mark @xmath2 on a galaxy located at @xmath1 is a random variable with the probability density @xmath234 depending on the value of the field @xmath235 at @xmath1 .
this can be used as a stochastic model for the genesis of galaxies depending on the local matter density .
+ in order to calculate the conditional mark correlation functions we define the conditional moments of the mark distribution given the value @xmath236 of the random field : @xmath237 the spatial mark product - density is @xmath238.\ ] ] and with eq .
@xmath239 , \ ] ] for @xmath240 and zero otherwise .
the mark correlation functions can therefore be expressed in terms of weighted correlations of the random field : @xmath241,\nonumber\\ { \left\langle m^2 \right\rangle_{\rm p}}(r ) = \frac{1}{\rho_2^u(r ) } { \mathbb{e}}\big [ \overline{m^2}(u({\mathbf{x}}_1))\ u({\mathbf{x}}_1 ) u({\mathbf{x}}_2)\big],\\ { \left\langle m_1 m_2 \right\rangle_{\rm p}}(r ) = \frac{1}{\rho_2^u(r ) } { \mathbb{e}}\big [ \overline{m}(u({\mathbf{x}}_1))\overline{m}(u({\mathbf{x}}_2))\ u({\mathbf{x}}_1 ) u({\mathbf{x}}_2)\big].\nonumber\end{gathered}\ ] ] to proceed further , we have to specify @xmath242 . as a simple example we choose @xmath2 equal to the value of the field @xmath235 at the point @xmath1 , such as in the random field model .
thinking of the random field as a mass density field and the mark of a galaxy luminosity , that means that the galaxies trace the density field and that their luminosities are directly proportional to the value of the field . with @xmath243
the conditional mark moments become @xmath244 .
the moments of the unconstrained mark distribution read @xmath245 , and the three basic pair averages are @xmath246 hence , the mark correlation functions defined in sect . [ sec : kerscher_mark - segregation ] are determined by the higher - order correlations of the random field . with the cox random field model we go beyond the random field model , e.g. @xmath247 is not equal to one any more .
at this point , we have to specify the correlations of the random field @xmath211 .
the simplest choice , a gaussian random field , is not feasible here , since a number density ( cp .
[ eq : kerscher_cox - definition ] ) has to be strictly positive , whereas the gaussian model allows for negative values . instead
, we will use the hierarchical ansatz : we first express the two- and three - point correlations in terms of normalized cumulants @xmath124 and @xmath248 ( see , e.g.,@xcite ) , @xmath249 in order to eliminate @xmath250 we use the hierarchical ansatz ( see e.g. @xcite ) : @xmath251 this ansatz is in reasonable agreement with data from the galaxy distribution , provided @xmath252 is of the order of unity ( @xcite ) . several choices for @xmath253 and @xmath252 lead to well - defined cox point process models based on the random field @xmath211 @xcite .
now we can express @xmath140 from eq . entirely in terms of the two - point correlation function @xmath254 of the random field : @xmath255 where we made use of the fact that @xmath256 .
inserting typical parameters found from the spatial clustering of the galaxy distribution we see from fig .
[ fig : kerscher_km - coxrf ] that the cox random field model allows us to qualitatively describe the observed luminosity segregation in fig.[fig : kerscher_ssrs - lum ] .
but the amplitude of @xmath133 predicted by this model is too high .
the cox random field model , however , is quite flexible in allowing for different choices for @xmath242 ; also different models for the higher - order correlations of the random field may be used , e.g. a log - normal random field @xcite .
clearly more work is needed to turn this into viable model for the galaxy distribution .
whenever objects are sampled together with their spatial positions and some of their intrinsic properties , marked point processes are the stochastic models for those data sets .
combining the spatial information and the objects inner properties one can constrain their generation mechanism and their interactions .
+ developing the framework of marked point processes further and outlining some of their general notions is thus of interest for physical applications .
let us therefore look at mark correlations again from both a statistical and a physical perspective .
we focused on two kinds of dependencies .
+ on the one hand , one can always ask , whether objects of different types `` know '' from each other . from a statistical point of view , this is the question whether the marked point process consists of two completely independent sub - point processes . physically , this concerns the question whether the objects have been generated together and whether they interact with each other .
+ on the other hand , it is often interesting to know whether the spatial distribution of the objects changes with their inner properties . for the statistician
, this translates into the question whether mark segregation or mark - independent clustering is present . for the physicist such
a dependency is interesting since one can learn from them whether and how the interactions distinguish between different object classes or whether the formation of the objects mark depends on the environment .
+ we discussed statistics capable of probing to which extent mark correlations are present in a given data set , and showed how to assess the statistical significance . applying our statistics to real data ,
we could demonstrate , that the clustering of galaxies depends on their luminosities .
large scale correlations of the orientations of dark matter halos were found . using the mars data we could validate a picture of crater generation on the martian surface :
mainly , the local geological setting determines the crater type .
we also could show that the sizes of pores in sandstone are correlated .
+ in order to understand empirical data sets in detail , we need models to compare to . as generic models the boolean depletion model , the random field model and its extension , the cox random field models are of interest . + further application of the mark correlations properties may inspire the development of further models .
it seems therefore that marked point processes could spark interesting interactions between physicists and mathematicians . certainly , the distributions of physicists and mathematicians in coffee breaks at the wuppertal conference were clustered , each . but could one observe positive cross - correlations ? using mark correlations we argue , that , even more , there is lots of space for positive interactions. .
we would like to thank andreas faltenbacher , stefan gottlber and volker mller for allowing us to present some results from the orientation analysis of the dark matter halos ( sect .
[ sec : kerscher_halos ] ) . for providing the sandstone data ( sect .
[ sec : kerscher_sandstone ] ) and discussion we thank mark knackstedt .
herbert wagner provided constant support and encouragement , especially we would like to thank him for introducing us to the concepts of geometric algebra as used in the appendix .
in order to form versatile test functions for describing mark segregation effects , we integrated the conditional mark probability density @xmath29 twice in mark space thereby weighting with a function of the marks @xmath52 ( see eq .
[ eq : kerscher_def - paverage ] ) .
such a pair - averaging reduces the full information present in @xmath29 .
so one may ask , whether or in which sense the mark correlation functions give a complete picture of the present two - point mark correlations .
+ for scalar marks @xmath2 this task is trivial . with a polynomial weighting function @xmath257 ( @xmath258 ) we consider moments of @xmath29 , hence , we can be complete only up to a given polynomial order in the marks @xmath16 and @xmath17 . at first order
there is only the mean @xmath73 . at second order
we have @xmath75 and @xmath259 .
all the mark correlation functions discussed in sect . [ sec : kerscher_mark - segregation ] can be constructed from these three pair averages and @xmath259 at the two - point level , however , does not imply that one should not consider linear combinations of them . for instance , it may well be the case , that only certain linear combinations yield significant results . ] .
higher - order moments of the marks involve more and more cross - terms .
+ for vector - valued marks , however , it is not obvious that the test quantities proposed in sect . [ sec : kerscher_mark - segregation ] trace all possible correlations between the vectors up to third order . to settle this case we have to consider the framework of geometric algebra , also called clifford algebra .
a detailed introduction to geometric algebra is given in @xcite , shorter introductions are @xcite . in geometric algebra
one assigns a unique meaning to the geometric product ( or clifford product ) of quantities like vectors , directed areas , directed volumes , etc .
the geometric product @xmath260 of two vectors @xmath261 and @xmath262 splits into its symmetric and antisymmetric part @xmath263 here @xmath264 denotes the usual scalar product ; in three dimensions , the wedge product @xmath265 is closely related to the cross product between these two vectors .
however , @xmath265 is not a vector like @xmath266 , but a bivector a directed area .
higher products of vectors can be simplified according to the rules of geometric algebra ( for details see @xcite ) .
+ let us consider the situation where objects situated at @xmath14 and @xmath15 bear vector marks @xmath86 and @xmath87 , respectively , and let the normalized distance vector be @xmath267 .
note , that @xmath103 is not a mark at all , rather it can be thought of as another vector which may be useful for constructing mark correlation functions . + for many applications it is reasonable to assume isotropy in mark space , i.e. all of the mark correlation functions are invariant under common rotations of the marks . for galaxies , e.g. , there does not seem to be an a priori preferred direction for their orientation .
in more detail we have then @xmath268 and so on , where @xmath155 is an arbitrary rotation in mark space .
this means that the mark correlation functions depend only on rotationally invariant combinations of the vector marks .
therefore , only rotationally invariant combinations of vectors are sensible building blocks for weighting functions .
we thus can restrict ourselves to scalar weighting functions , which result in coordinate - independent vector - mark correlation functions .
+ again we proceed by considering mixed moments as basic combinations . we restrict ourselves to scalar quantities being polynomial in the vector components .
one may also discuss moments in a broader sense allowing for vector moduli . in this wider sense , for example , @xmath269 or @xmath270 would be allowed .
we do not consider such quantities here , because they are not polynomial in the vector components .
their squares anyway appear at higher orders .
furthermore , it turns out that the characterization we will provide depends on the embedding dimension .
the first- and second - order moments are identical in two and three dimensions , but at the third order they start to differ . 1 . in the strict sense of scalar quantities being linear in the vector components
there are no first - order moments for vectors .
at second order we encounter the following products : @xmath271 , @xmath272 , @xmath273 , @xmath274 .
note , that , e.g. , @xmath275 and @xmath276 do not make any difference as regards the mark correlation functions , since the pair averages implicitly render the indices symmetric ; moreover , although the geometrical product is non - commutative , @xmath277 and @xmath278 do not lead to different mark correlation functions .
furthermore , @xmath279 .
@xmath280 provides us with higher moments of the modulus of the vectors . to investigate these kinds of correlations
already scalar marks would be sufficient .
new information is encoded in the other products .
+ consider @xmath281 .
the symmetric part @xmath282 is clearly a scalar and defines the alignment @xmath283 ( eq . [ eq : kerscher_align ] ) .
the antisymmetric part @xmath284 is a bivector .
its unique
modulus ( see again @xcite ) , @xmath285 , may be useful , but is no longer a polynomial in the vector components .
@xmath286 appears at the fourth order . in a completely analogous way we can treat @xmath287 .
the symmetric part @xmath288 defines @xmath100 .
hence at second order , the only possible vector - mark correlation functions are @xmath84 and @xmath100 .
3 . at third order
we have to consider products of three vectors . in general the product of three vectors
@xmath289 splits into @xmath290 i.e. , a vector ( consisting of the three first terms ) , and a pseudo - scalar , a directed volume . in two dimensions the pseudo - scalar @xmath291 vanishes .
+ now we have to form all possible products of the three vectors @xmath129 and to derive scalars . in three dimensions
the only new combination is the pseudo - scalar @xmath292 giving the oriented volume @xmath293 .
unfortunately , this oriented volume averages out to zero .
thus , in a strict sense , there are no interesting third - order quantities .
closely related , however , is the modulus of the pseudoscalar @xmath294 proportional to our @xmath105 .
this expression is invariant under permutations of the vectors .
4 . at third order and in two dimensions
all of the relevant combinations are products of first- and second - order combinations ; no specifically new combination appears .
this is different from the case of three dimensions , where at third order an entirely new geometric object , the pseudo - scalar @xmath292 can be constructed .
there is a general scheme behind this argument : since in @xmath96 dimensions any geometrical product of more than @xmath96 vectors vanishes , all relevant combinations of vectors at orders higher than @xmath96 are essentially products of combinations of lower - order factors .
baddeley , a. j. ( 1999 ) : `` sampling and censoring '' . in : _ stochastic geometry , likelihood and computation _ , ed . by o.
barndorff - nielsen , w. kendall , m. van lieshout , volume 80 of _ monographs on statistics and applied probability _
, chapter 2 ( chapman and hall , london ) bhringer , h. , p. schuecker , l. guzzo , c. collins , w. voges , s. schindler , d. neumann , g. chincharini , r. cruddace , a. edge , h. macgillivray , p. shaver ( 2001 ) : `` the rosta - eso flux limited x - ray ( reflex ) galaxy cluster survey i : the construction of the cluster sample '' , _ _ , p. 826 , l. n. , c. n. a. willmer , p. pellegrini , o. l. chaves , c. rite , m. a. g. maia , m. j. geller , d. w. latham , m. j. kurtz , j. p. huchra , m. ramella , a. p. fairall , c. smith , s. lipari ( 1998 ) : `` the southern sky redshift survey '' , _ _ * 116 * , pp .
17 djorgovski , s. ( 1987 ) : `` coherent orientation effects of galaxies and clusters '' . in : _ nearly normal galaxies . from the planck time to the present _ , ed . by s.
m. faber ( springer , new york ) , pp .
227233 guzzo , l. , j. bartlett , a. cappi , s. maurogordato , e. zucca , g. zamorani , c. balkowski , a. blanchard , v. cayatte , g. chincarini , c. collins , d. maccagni , h. macgillivray , r. merighi , m. mignoli , d. proust , m. ramella , r. scaramella , g. stirpe , g. vettolani ( 2000 ) : `` the eso slice project ( esp ) galaxy redshift survey .
the redshift and real - space correlation functions '' , _ _ * 355 * , pp .
116 hermit , s. , b. x. santiago , o. lahav , m. a. strauss , m. davis , a. dressler , j. p. huchra ( 1996 ) : `` the two point correlation function and the morphological segregation in the optical redshift survey '' , _ _ * 283 * , p. 709
isham , v. ( 1985 ) : `` marked point processes and their correlations '' . in : _ spatial processes and spatial time series analysis _
, ed . by f.
droesbeke ( publications des facults universitaires sain - louis , bruxelles ) kerscher , m. ( 2000 ) : `` statistical analysis of large scale structure in the universe '' . in : _ statistical physics and spatial statistics : the art of analyzing and modeling spatial structures and pattern formation _ , ed . by k.
r. mecke , d. stoyan , number 554 in lecture notes in physics ( springer , berlin ) , astro - ph/9912329 lwen , h. ( 2000 ) : `` fun with hard spheres '' . in : _ statistical physics and spatial statistics : the art of analyzing and modeling spatial structures and pattern formation _ ,
ed . by k.
r. mecke , d. stoyan , number 554 in lecture notes in physics ( springer , berlin ) mecke , k. ( 2000 ) : `` additivity , convexity , and beyond : application of minkowski functionals in statistical physics '' . in : _ statistical physics and spatial statistics : the art of analyzing and modeling spatial structures and pattern formation _
, ed . by k.
r. mecke , d. stoyan , number 554 in lecture notes in physics ( springer , berlin ) sok , r. m. , m. a. knackstedt , a. p. sheppard , w. v. pinczewski , w. b. lindquist , a. v. a , , l. paterson ( 2000 ) : `` direct and stochastic generation of network models from tomographic images ; effect of topology on two - phase flow properties '' .
in : _ proc . upscaling downunder _ , ed . by l.
paterson , ( kluwer academic , dordrecht ) stoyan , d. ( 2000 ) : `` basic ideas of spatial statistics '' . in : _ statistical physics and spatial statistics : the art of analyzing and modeling spatial structures and pattern formation _
, ed . by k.
r. mecke , d. stoyan , number 554 in lecture notes in physics ( springer , berlin ) | mark correlations provide a systematic approach to look at objects both distributed in space and bearing intrinsic information , for instance on physical properties .
the interplay of the objects properties ( marks ) with the spatial clustering is of vivid interest for many applications ; are , e.g. , galaxies with high luminosities more strongly clustered than dim ones ?
do neighbored pores in a sandstone have similar sizes ?
how does the shape of impact craters on a planet depend on the geological surface properties ? in this article
, we give an introduction into the appropriate mathematical framework to deal with such questions , i.e. the theory of marked point processes . after having clarified the notion of segregation effects
, we define universal test quantities applicable to realizations of a marked point processes .
we show their power using concrete data sets in analyzing the luminosity - dependence of the galaxy clustering , the alignment of dark matter halos in gravitational @xmath0-body simulations , the morphology- and diameter - dependence of the martian crater distribution and the size correlations of pores in sandstone . in order to understand our data in more detail
, we discuss the boolean depletion model , the random field model and the cox random field model .
the first model describes depletion effects in the distribution of martian craters and pores in sandstone , whereas the last one accounts at least qualitatively for the observed luminosity - dependence of the galaxy clustering . draft |
vector mesons ( @xmath3 , @xmath4 , etc . ) play a significant role in hadronic physics .
their interactions , though not constrained by low - energy theorems , apparently follow the broad pattern of vector meson dominance ( vmd ) @xcite .
there have been numerous efforts to incorporate vector mesons into field - theoretical frameworks .
historically , the yang - mills theory was discovered in an early attempt to treat the @xmath2 meson @xcite .
more recently , interesting schemes based on `` hidden local symmetries '' ( hls ) were developed by bando _
et al . _
@xcite . in the original model @xcite ,
the @xmath2 meson is the boson of a spontaneously broken gauge group .
the model has been extended to two hidden gauge groups @xcite ; then it also incorporates the lowest axial vector meson @xmath5 . with suitable parameters
, these models can be quite successful phenomenologically , although they can not be systematically derived from qcd ( except in the limit of very light @xmath2 , if such a limit could be reached @xcite ) . in this paper
we explore theories with very large , and even infinite number @xmath0 of hidden local symmetries .
our motivation is twofold .
first and most straightforwardly , there are excited states in the vector and axial vector channels ( @xmath6 , @xmath7 , @xmath8 , etc .
@xcite ) , which must become narrow resonances in the limit of large number of colors @xmath9 .
it is tempting to treat them as gauge bosons of additional broken gauge groups . ,
@xmath10 , etc . as a `` chain structure ''
was made in ref .
@xcite . ]
the second motivation comes from recent theoretical developments .
many strongly coupled gauge theories are found to have a dual description in terms of theories with gravity in higher dimensions @xcite .
it was suggested that the string theory dual to large-@xmath9 qcd must have strings propagating in five dimensions , in which the fifth dimension has the physical meaning of the energy scale @xcite . in the framework of field theory ,
the fifth dimension can be `` deconstructed '' in models with a large number of gauge fields @xcite .
we discovered that the continuum limit @xmath1 can lead to results that qualitatively , and in many cases even quantitatively , agree with phenomenology .
most remarkably , the vector meson dominance , which in the hls theories required a tuning of parameters , becomes a natural consequence of the @xmath1 limit .
another advantage of the limit @xmath1 is the possibility of matching to the asymptotic behavior of the current - current correlator known from perturbative qcd .
as anticipated , a natural interpretation of this limit is a discretization , or deconstruction , of a 5-dimensional gauge theory .
further , to our amusement , in the calculation of current - current correlators we found a relation very similar to the one employed in the ads / cft correspondence : the current - current correlator in 4d theory is expressed in terms of the variations of the classical 5d action with respect to the boundary values of the bulk gauge fields on the 4d boundaries .
we limit our discussion to the isospin-1 sector of qcd .
it is straightforward to extend the discussion to the isospin-0 sector ( @xmath11 , @xmath12 , and @xmath13 mesons ) .
the detailed treatment of the @xmath14 problem , chiral anomaly , wess - zumino - witten term , and baryons is deferred to future work .
the paper is organized as follows . in section
[ sec : model ] we describe the open moose model . in section [ sec : observables ] we compute different physical observables : the vector meson mass spectrum , the decay constants of the pion and the vector mesons , the coupling between the vector mesons and the pions , and the pion electromagnetic form factor .
we also check the validity of weinberg s spectral sum rules , and discover that the limit @xmath1 automatically leads to exact vmd for the pion formfactor . in section [ sec :
continuum ] we take the limit of infinite number of the hidden groups @xmath1 .
we show that the theory can be understood as a 5d yang - mills theory in an external metric and dilaton background .
we establish an ads / cft - type prescription for calculating the current - current correlators .
we consider two concrete realizations of the open moose in section [ sec : examples ] .
we find that a `` cosh '' background metric interpolating between two ads boundaries leads to correct asymptotic behavior of the current - current correlator .
this allows us to establish a relationship between hadron parameters such as @xmath15 , @xmath16 , and the qcd parameter @xmath9 . in section [ sec : baryon ] we show that the instanton , which is a quasiparticle in @xmath17 dimensions , becomes a skyrmion upon reduction to 4d , and thus describes the baryon .
section [ sec : concl ] contains concluding remarks .
the model under consideration is described by the following lagrangian , but write all indices as lower indices for simplicity , unless it could lead to a confusion . ]
@xmath18 the covariant derivatives are defined as @xmath19 a shorthand notation is used for the product of the gauge field @xmath20 and its coupling constant : @xmath21 . if we assume @xmath22 , then eqs .
( [ dsigma1 ] ) and ( [ dsigma3 ] ) become special cases of eq .
( [ dsigma ] ) for @xmath23 and @xmath24 .
the model contains @xmath25 nonlinear sigma model fields @xmath26 ( or , in general , @xmath27 ) , interacting via @xmath0 `` hidden '' gauge bosons @xmath28 .
the model has a chiral @xmath29 symmetry and an @xmath30 local symmetry : @xmath31 in particular , the product @xmath32 is the pion field , which can be seen from its transformation properties , @xmath33 the parameters entering ( [ l ] ) are @xmath25 decay constants @xmath34 and @xmath0 gauge couplings @xmath35 .
we shall assume they are invariant under a reflection with respect to the middle of the chain , @xmath36 which ensures parity is a symmetry in the theory ( [ l ] ) .
in the case @xmath37 the model reduces to the chiral lagrangian . for @xmath38
it is the version of the hidden local symmetry realized in the limit of very light @xmath2
s @xcite .
the model with @xmath39 and a particular choice of parameters @xmath40 , @xmath41 has been considered in ref .
graphically , the model can be represented by a `` theory - space '' diagram shown in fig .
[ fig : k ] .
since this diagram is the usual `` moose diagram '' cut open , we shall call the model ( [ l ] ) the `` open moose '' theory .
fig : k note that ( [ l ] ) is not the most general lagrangian satisfying all the symmetries and limited to lowest derivatives .
in fact , terms of the following type are not forbidden : @xmath42 as well as analogous expressions containing products of more than two consecutive @xmath43 s . in order to restrict the lagrangian to the form ( [ l ] ) an additional condition of nearest neighbor locality in the @xmath44 space should be imposed .
it is this condition that enables us later to interpret this theory as a dimensionally deconstructed 5d gauge theory in the limit @xmath1 .
in this section we derive expressions for physical observables , such as pion and vector meson decay constants , mass spectrum , and pion - vector couplings in terms of the parameters of the moose @xmath34 and @xmath35 . for computations ,
the following gauge is the most convenient , @xmath45 where @xmath46 is a function of all @xmath34 , which we shall specify in a moment ( in eq .
( [ f ] ) ) .
the advantage of this gauge is that the pion field @xmath47 does not mix with other fields : @xmath48 = 0.\ ] ] the value of @xmath46 is fixed by requiring that the kinetic term for @xmath49 is canonically normalized : @xmath50 therefore @xmath51 to determine @xmath15 we use noether s theorem to construct the axial current @xmath52 .
let us consider an infinitesimal axial su(2 ) transformation .
it acts only on @xmath53 and @xmath54 at the ends of the moose : @xmath55 if the parameter @xmath56 depends on coordinates , the lagrangian changes by @xmath57 . on the other hand from ( [ l ] ) one finds @xmath58 which means @xmath59 , i.e. , @xmath60 .
equation ( [ f ] ) becomes @xmath61 it is a simple exercise to verify that for @xmath62 this general formula is in agreement with corresponding results in these theories .
it is also perhaps useful to observe that sending @xmath63 on one of the links effectively sets gauge fields on the ends of this link equal to each other ( @xmath64 ) , effectively eliminating this link and reducing @xmath0 by one .
the formula ( [ fpi ] ) obviously reflects this reduction the corresponding term @xmath65 drops out . in our gauge ,
the vacuum is @xmath67 for all @xmath44 . expanding to second order in @xmath28 , we find the terms that determine the masses of the vector mesons : @xmath68 ^ 2 \equiv \sum_{\substack{k=1\\k'=1}}^{k } ( m^2)_{kk ' } { { \rm tr } } a^k_\mu a^{k'}_\mu\,.\ ] ] the mass matrix can be diagonalized by using an orthogonal matrix @xmath69 satisfying @xmath70 in terms of new vector fields @xmath71 defined as @xmath72 the mass term ( [ lmass ] ) is diagonal . using eq .
( [ lmass ] ) and the orthogonality of @xmath73 , we can write the equation determining @xmath69 and @xmath66 as @xmath74 - f_k^2[(gb_n)^k-(gb_n)^{k-1 } ] = - \frac{m_n^2 b_n^k}{g_k}\,.\ ] ] this is essentially a discretized version of a sturm - liouville problem
. we shall write the corresponding differential equation in section [ sec : continuum ] when we consider the continuum limit @xmath1 .
we shall also use the discrete equation ( [ evdisc ] ) in section [ sec : vv ] . without solving eq .
( [ evdisc ] ) , we can conclude right away that there is a tower of eigenvalues @xmath66 , @xmath75 , corresponding to the masses of vector and axial vector mesons .
the lowest @xmath76 and @xmath77 states correspond to the @xmath2 and @xmath78 mesons .
moreover , states with opposite parity alternate in the spectrum : @xmath79 odd @xmath80 states correspond to vector mesons and even @xmath80 to axial vector mesons . in the real world , the trend of alternating parity can be seen in the hadronic spectrum for the few first vector and axial vector states .
let us compute the coupling of @xmath80th vector meson to a pion pair , @xmath81 .
expanding the lagrangian in @xmath82 and @xmath83 , isolating @xmath84 terms , and using eq .
( [ abalpha ] ) , we find @xmath85 [ ( ga_\mu)^{k-1 } + ( ga_\mu)^k ] \\ & = i \sum_{n=1}^k \sum_{k=1}^{k+1 } \frac{f_\pi^2 } { 4f_k^2 } [ ( gb_n)^{k-1 } + ( gb_n)^k ] { { \rm tr } } [ \partial_\mu \pi,\pi ] \alpha^n_\mu \ , . \end{split}\ ] ] recall that we use the matrix notation where @xmath86 and @xmath87 . if we normalize @xmath81 so that the relevant coupling is @xmath88 then @xmath89.\ ] ] note that this @xmath90 coupling vanishes for axial mesons @xmath91 , as required by parity , because their `` wave functions '' @xmath92 are odd under @xmath93 ( eq .
( [ parity ] ) ) .
sec : vv we define the decay constants for the vector and axial vector mesons via the matrix elements of the vector and axial vector currents between the vacuum and the one - meson states , @xmath96 here @xmath97 is a single - particle state of the @xmath80-th vector boson ( @xmath98 , @xmath99 , etc . ) with isospin @xmath100 and polarization @xmath101 .
both @xmath94 and @xmath95 have the dimension of [ mass@xmath102 .
@xmath103 for axial vector mesons ( @xmath80 even ) and @xmath104 for vector mesons ( @xmath80 odd ) .
it is convenient to compute @xmath94 by looking at the vector current - current correlator @xmath105 .
the residues at the poles are easily related to @xmath94 .
the correlator can be obtained by gauging the corresponding su(2)@xmath106 transformation and differentiating the action with respect to the gauge field @xmath107 : @xmath108}{\delta b^a_\mu(x)\delta b^b_\nu(y ) } \,,\ ] ] were @xmath109 $ ] is the vacuum energy functional in the presence of the external field @xmath110 .
the su(2)@xmath106 transformation only affects the two @xmath43 links at the ends of the chain according to eq .
( [ usigmau ] ) : @xmath111 and therefore the terms containing the gauge field @xmath107 are only from @xmath23 and @xmath24 : @xmath112 keeping only terms bilinear in the fields we find ( the term @xmath113 is absent due to parity ) : @xmath114 + { \cal l}_{a^2 } + { \cal l}_{f^2}\,.\ ] ] extremizing the action with respect to @xmath82 at fixed @xmath110 and then taking the second derivative with respect to @xmath110 we find ( this is also equivalent to taking the gaussian integral over @xmath82 and then differentiating the logarithm of this integral ) @xmath115(x ) [ ( ga^b_\nu)^1 + ( ga^b_\nu)^k ] ( y ) \rangle,\ ] ] where the @xmath116 is the propagator of @xmath82 , i.e. , the inverse of the quadratic form found in @xmath117 . diagonalizing this expression using eq .
( [ abalpha ] ) and performing a fourier transformation with respect to the four - dimensional coordinate @xmath118 , we find @xmath119 where the decay constants @xmath94 are determined to be @xmath120 . \ ] ] note that @xmath103 for axial mesons for which @xmath121 .
one can also use eq .
( [ evdisc ] ) together with eq .
( ( [ gnv ] ) ) to write a different representation for @xmath94 : @xmath122 = m_n^2\sum_{k=1}^k \frac{b_n^k}{g_k } \,.\ ] ] it is perhaps easier to look at this equation as a discretized version of integration by parts as we shall do in section [ sec : continuum ] . ) or ( [ gnvk ] ) can be also understood in the following way ( the reader might recognize the discussion given in refs .
one should realize that because of the mixing of @xmath110 with @xmath82 the actual photon is not the field @xmath110 , but a linear combination of @xmath110 and all @xmath82 that leaves vacuum @xmath67 invariant .
this is similar to the mixing of the standard model hypercharge boson and weak isospin vector boson to produce the photon .
the corresponding linear combination of the @xmath82 fields has a `` wave function '' proportional to @xmath123 ( each @xmath124 enters with weight @xmath123 ) .
the mixing of the actual photon is now entirely through derivative terms in the expansion of @xmath125 .
the corresponding coefficients are given by the overlap of the photon `` wave function ''
@xmath126 and the @xmath80th vector meson `` wave function '' @xmath127 .
the factor @xmath128 is from the derivatives evaluated using the equation of motion for the @xmath80th meson . ]
the calculation of the decay constants of the axial vector mesons @xmath95 is completely analogous .
we introduce an auxiliary gauge field @xmath129 coupled to the axial current @xmath130 : @xmath131 in addition to the @xmath132 mixing and @xmath133 contact terms there is now also the mixing with pion field @xmath134 . differentiating the logarithm of the partition function twice , we obtain @xmath135 where @xmath136.\ ] ] note that , as expected , @xmath104 for vector mesons for which @xmath137 .
weinberg @xcite has derived two sum rules for the weighted integrals of the difference of spectral functions of the vector and axial vector current correlators , @xmath138 and @xmath139 .
we shall now verify that both sum rules hold by using transversality of the current - current correlators . from eq .
( [ gnvk ] ) it is easy to show that the @xmath140 correlator is transverse .
transversality requires that the contact term in eq .
( [ vv ] ) is related to the pole terms through the following sum rule : @xmath141 this can be checked by using eq .
( [ gnvk ] ) and orthogonality of @xmath69 .
we indeed find @xmath142\frac{b_n^k}{g_k}= 2 f_1 ^ 2.\ ] ] by using eq .
( [ gnvsum ] ) one can rewrite the correlator in a manifestly transverse form , @xmath143 the transversality of the @xmath144 correlator ( [ aa ] ) amounts to the following sum rule : @xmath145 by comparing eqs .
( [ gnvsum ] ) and ( [ gnasum ] ) we conclude that @xmath146 which is one of weinberg s sum rules . ] where @xmath147 is defined via @xmath138 in eq .
( [ vvpi ] ) .
a similar equation defines @xmath148 .
the sum rules state that @xmath149\mu^{-2}d\mu^2=f_\pi^2 ; \qquad ( ii ) \int [ \rho_v(\mu^2 ) - \rho_a(\mu^2)]d\mu^2 = 0.\ ] ] in our theory , according to eq .
( [ vvpi ] ) , @xmath150 .
] this sum rule holds for any @xmath0 . note that , for @xmath1 both sums ( [ gnvsum ] ) and ( [ gnasum ] ) must diverge , since @xmath34 must become infinite at the ends of the moose ( to ensure convergence in eq .
( [ fpi ] ) ) . however , their difference is finite . the second weinberg sum rule @xmath151 also holds .
it is easy to prove by using the definitions ( [ gnv ] ) and ( [ gna ] ) and the orthogonality of @xmath69 : @xmath152 which vanishes for all @xmath153 . in the case
@xmath38 , there is only one meson
@xmath2 , and no axial mesons at all . the pion form factor ( defined to be the isovector part of the electromagnetic form factor )
, @xmath154 can be found isolating terms linear in @xmath110 in the lagrangian ( [ lb ] ) .
there are two contributions to the form factor the direct interaction , given by the term @xmath155 in the lagrangian and the interaction mediated by vector mesons given by the @xmath156 mixing terms and the couplings @xmath157 .
one finds @xmath158 using the expressions ( [ gnvk ] ) and ( [ gnpipik ] ) for @xmath94 and @xmath81 and the orthogonality of the matrix @xmath69 , the sum rule related to the total charge of pion can be verified , @xmath159 if we understand vmd as the statement that @xmath160 is saturated by a sum over resonances ( i.e. , dominance by the whole tower of mesons ) , then in our model vmd is valid when the contribution of the direct interaction is negligible , @xmath161 .
thus vmd is a natural consequence of the @xmath1 limit ( due to eq .
( [ fpi ] ) ) . a stronger statement that @xmath160 is saturated by a single @xmath2 pole
is not , in general , valid ( see , however , section [ sec : cosh ] ) .
sec : continuum in the preceding section we derived formulas that are valid for an arbitrary @xmath0 .
now we wish to consider the limit @xmath1 . in this limit
the expressions that we found can be simplified , provided that @xmath34 and @xmath35 are sufficiently smooth functions of @xmath44 . in this case
we can consider replacing the discrete variable @xmath44 by a continuum variable that we shall call @xmath162 : @xmath163 here @xmath164 plays the role of the `` lattice spacing . ''
if the limit @xmath1 is performed in the following way , @xmath165 then @xmath162 becomes a continuum replacement for @xmath44 .
if @xmath34 and @xmath35 are smooth functions of @xmath44 , we can also replace them by functions of @xmath162 , @xmath166 for the resonance `` wave functions '' @xmath69 that vary smoothly we can write @xmath167 so that orthogonality of the matrix @xmath69 translates into orthonormality of the functions @xmath168 , @xmath169 the wave functions of sufficiently high resonances with @xmath170 can not be expected to be smooth , so they must be treated discretely . we shall always be interested in a finite number of lowest resonances , while @xmath1 .
sec : contform let us now write continuum limits for the main formulas we have derived in the preceding section . from eq .
( [ fpi ] ) , @xmath171 from eq .
( [ evdisc ] ) , @xmath172 with dirichlet boundary conditions @xmath173 ( since we set @xmath22 ) . from eq .
( [ gnpipik ] ) , @xmath174 from eq .
( [ gnv ] ) , using the fact that @xmath175 , @xmath176_{-u_0}^{+u_0}.\ ] ] by using eq .
( [ eigen ] ) , we find the continuum limit of eq .
( [ gnvk ] ) : @xmath177_{-u_0}^{+u_0}= m_n^2\int_{-u_0}^{+u_0 } \!du\ , \frac{b_n(u)}{g(u)}\,.\ ] ] analogously , eq . ( [ gna ] ) becomes @xmath178|_{+u_0}+[f^2(u ) ( g(u)b(u))']|_{-u_0}\,.\ ] ] it is very interesting that the physical observables we calculated are all well behaved in the continuum limit @xmath1 , @xmath179 ( provided the corresponding integrals over @xmath162 converge ) . for reference ,
the equations for other vertex couplings are presented in appendix [ app : mnp ] .
our long - moose theory with @xmath181 can be also considered as a discretized ( or deconstructed ) five - dimensional continuum gauge theory in curved spacetime .
the variable @xmath162 plays the role of the fifth , deconstructed , dimension . the smoothly varying fields @xmath43 s can be interpreted as the link variables along the fifth dimension @xmath182 , @xmath183 for this equation and for remainder of section [ sec : continuum ] we shall make a temporary switch of notations , absorbing the gauge coupling constants @xmath184 into the fields @xmath82 : @xmath185 .
then the action ( [ l ] ) can be written in the 5d notations as @xmath186 we now compare this action to the action of a gauge field in a background of curved spacetime and a dilaton field . in the following , @xmath187 denotes the determinant of the metric tensor .
the action is taken in the form : @xmath188 where @xmath189 are 5d lorentz indices .
the coupling to the dilaton field is written so that the effective gauge coupling is @xmath190 . in our simple model
we consider the metric and the dilaton as classical background fields with no dynamics of their own . taking the dilaton field to be dependent only on the fifth coordinate @xmath162 , @xmath191 , and the metric to be of the warped form , @xmath192 the action ( [ 5ds ] )
can be expanded as @xmath193 equation ( [ 5dsexp ] ) coincides with eq .
( [ sfgf ] ) if one makes the following identification : fgwarp @xmath194 notice that the warp factor @xmath195 is equal to @xmath196 , i.e. , @xmath197 it is also easy to see that the wave equation for the spin-1 mesons ( [ eigen ] ) is one of the yang - mills / maxwell equations for the 5d ( massive ) gauge field in the curved background .
notice also that the pion field ( [ sigmaproduct ] ) is now the wilson line stretching between the two boundaries , @xmath198 we now show that correlators of conserved currents in our theory can be computed by using a prescription essentially identical to the ads / cft one .
namely , the generating functional for the correlation functions of the currents is equal to the action of a solution to the classical field equations , with the sources serving as the boundary values for the classical fields .
recall our calculation of the current - current correlators in section [ sec : vv ] . instead of the vector field @xmath107 and
@xmath199 let us introduce two separate fields @xmath200 and @xmath201 , corresponding to gauging the su(2)@xmath202 and su(2)@xmath203 global symmetries of the theory .
the appearance of these fields modify the first and last terms in the moose , @xmath204 in this section we also absorb the coupling @xmath35 into the field @xmath82 .
remember that there are no dynamical fields associated with the ends of the moose @xmath205 and @xmath25 .
we can treat the ends of the moose more equally with the other points by thinking that the values of the field @xmath206 at the ends of the moose , at @xmath205 and @xmath25 , are fixed at given values : @xmath207 if the field @xmath206 is smooth , we can translate this into the continuum limit by setting boundary conditions on the continuous 5d field @xmath208 : @xmath209 at tree level , the generating functional is thus equal to @xmath210 = e^{is_{\textrm{cl}}[a_\mu^{\textrm{cl}}]}\ ] ] where @xmath211 is the solution to the classical field equation that satisfies the boundary conditions ( [ bc ] ) .
this formula is of the same form as the formula for ads / cft correspondence : the sources for the boundary theory ( in our case @xmath212 ) serve as the boundary values for the bulk field . in particular , in order to compute the correlation functions for the conserved currents @xmath213 or @xmath214 one just needs to differentiate the classical action with respect to the corresponding boundary values , e.g. , @xmath215 } { \delta a^l_\mu(x ) \delta a^l_\nu(y)}\ ] ] the fact that we have arrived at an ads / cft - like formula ( [ adscft ] ) makes one wonder if the hidden local symmetry models for the @xmath2 and @xmath78 vector mesons @xcite are ( very coarsely ) discretized versions of a 5d theory dual to qcd .
this could explain why these models enjoy certain phenomenological success , and why in the @xmath39 model @xcite one is driven to choose the parameters so that it becomes a moose theory ( i.e. , nearest neighbor local ) .
a difference from the usual ads / cft correspondence is that there are _ two _ boundaries in the open moose theory .
however , if so desired , one can reformulate the 5d theory ( [ 5ds ] ) and ( [ warped ] ) in the spatial region @xmath216 ( which is one half of the original @xmath217 ) at the price of having two gauge fields obeying a matching condition at @xmath218 .
then the spacetime will have only one boundary at @xmath219 .
so far , our discussion has been general and valid for any choice of @xmath34 and @xmath35 . in this section
we shall consider two concrete realizations of the open - moose theory .
our goal is to illustrate the general formulas , and to compare the results with the phenomenology of vector mesons .
the two examples are chosen because they are exactly solvable : the spectrum of the vector mesons and the coupling constants can be found in the closed form .
the first example is also the simplest possible model , but it has a significant physical drawback that we point out at the end .
we think nevertheless that it is a useful reference point for comparison and for understanding the robustness / sensitivity of the results towards the change of the background parameters @xmath220 and @xmath221 .
consider a moose with parameters @xmath34 and @xmath35 independent of @xmath44 . , but we shall only consider @xmath181 . ] in the continuum limit @xmath181 the corresponding functions are therefore constant , does not affect the results ; it is equivalent to rescaling @xmath46 and @xmath184 . ]
@xmath222 let us now apply general formulas from section [ sec : contform ] to determine the properties of this theory in terms of the parameters @xmath46 and @xmath184 . from ( [ fpiu ] )
@xmath223 the spectrum and wave functions of the spin-1 mesons are given by eq .
( [ eigen ] ) , which becomes @xmath224 this means @xmath225 the few first wave functions are plotted in fig .
[ fig : sin ] .
fig : sin from eq .
( [ gnpipiu ] ) , for @xmath226 , @xmath227 consider the @xmath2 meson , @xmath76 .
the ratio of @xmath228 to @xmath229 is dimensionless and is equal to @xmath230 the coupling @xmath231 can be found from the width of the @xmath2 , which decays predominantly to two pions : @xmath232 , where @xmath233 is the velocity of the final - state pions . using @xmath234 mev
, we find the ratio ( [ ksrf ] ) to be around 1.9 in nature .
for comparison , the ksrf relation @xcite corresponds to this ratio being equal to 2 , and the value in georgi s vector limit ( i.e. , @xmath38 moose theory ) is 4 @xcite .
therefore , our model would underpredict @xmath235 from experimental @xmath16 and @xmath15 . the decay constants @xmath94 and @xmath95 are given by eqs .
( [ gnvbg ] ) and ( [ gnabg ] ) and are equal to @xmath236 in eq .
( [ gnv1 ] ) @xmath237 refers to @xmath94 for odd @xmath80 s and @xmath95 for even @xmath80 s . for @xmath76 , we find @xmath238 .
we can now predict the rate of the electromagnetic decay @xmath239 , using @xmath240 and the experimental values for @xmath15 and @xmath16 .
we find @xmath241 kev , which is somewhat smaller than the measured value @xmath242 kev @xcite .
it is also interesting to consider the contribution of the @xmath2 meson to pion form factor at @xmath243 ( eq . ( [ gvpipi0 ] ) ) , @xmath244 thus the single @xmath2-meson dominance holds to within 20% .
the vmd is , however , exact if all vector mesons are included in the limit @xmath1 . indeed
the direct pion - photon interaction in eq .
( [ gvpipi ] ) vanishes : @xmath245 . ) .
each contribution is proportional to @xmath246 and @xmath247 . ]
the drawback of this model is that it fails to satisfy the asymptotic condition on @xmath248 that follows from qcd : @xmath249 when @xmath250 . instead , @xmath248 in this model vanishes as @xmath251 when @xmath250 .
indeed , according to eq .
( [ vvpi ] ) , with values of @xmath94 and @xmath66 found in eqs .
( [ gnv1 ] ) and ( [ mn1 ] ) , @xmath252 we shall now consider an exactly solvable model that will satisfy the condition @xmath253 at large @xmath254 . this model is given by [ coshbg ] @xmath255 according to eqs .
( [ fgwarp ] ) , this corresponds to a constant dilaton background and the following background metric , @xmath256 the two boundaries are located at @xmath257 .
near the boundaries the metric becomes asymptotically ads@xmath258 .
according to the ads / cft philosophy , @xmath162 has the physical meaning of the energy scale ; large @xmath162 s correspond to short distances .
therefore one can expect that the current correlators has the conformal form at short distance , i.e. , as @xmath250 , @xmath259 the main reason for choosing the background ( [ coshbg ] ) is that @xmath260 is the simplest function interpolating between @xmath261 and @xmath262 , and that the mass spectrum can be found exactly ( see below ) .
otherwise , we have no reason to prefer this background over any other that has two ads@xmath258 boundaries . )
coincides with the 5d part of the induced metric on a probe d7 brane in @xmath263 @xcite . ]
applying eq .
( [ fpiu ] ) , one finds @xmath264 the wave equation for the vector mesons is @xmath265 which implies the following spectrum : @xmath266 in particular , @xmath267 and @xmath268 .
taking @xmath16 as an input , this predicts @xmath269 mev , which is not far from the observed @xmath270 mev .
however , the masses of higher excitations grow faster with @xmath80 than in the real world
. the 5d eigenfunctions of the vector mesons are @xmath271 where @xmath272 are the associated legendre functions .
the first few wave functions are ( see fig .
[ fig : cosh ] ) : @xmath273 fig : cosh in order to establish eq .
( [ piexp ] ) , we compute the decay constants of vector mesons from eqs .
( [ gnvbg ] ) and ( [ gnabg ] ) , @xmath274 the correlation function for the vector current is found from eq .
( [ vvpi ] ) , @xmath275 at large @xmath276 , the sum can be replaced by an integral , which is logarithmically divergent . ) , subtracting a constant equal to @xmath277 for @xmath181 .
of course , the equation is only valid for @xmath278 .
] one thus finds for large @xmath254
@xmath279 the asymptotic behavior of @xmath280 is the same .
thus the current correlators have the correct asymptotics at large @xmath254 .
moreover , they obey weinberg s sum rules , as proven in section [ sec : weinberg ] .
the constraints imposed by the @xmath250 behavior and weinberg s sum rules on the masses @xmath66 and decay constants @xmath237 are quite nontrivial @xcite .
it is remarkable that the open - moose construction generates examples that automatically satisfy these constraints .
one can match the asymtotics ( [ pivlog ] ) with the result found from qcd , @xmath281 where @xmath9 is the number of colors , to obtain @xmath282 by using this relationship between @xmath283 and @xmath9 together with @xmath284 , we can now express all quantities in the model via a single mass @xmath16 and the number of colors @xmath9 .
a short summary is given in appendix [ app : cosh ] .
for example , for @xmath15 we find from eq .
( [ fpilambdag ] ) @xmath285 for @xmath286 eq .
( [ mrhofpinc ] ) predicts @xmath287 87 mev , rather close to the experimental value of 93 mev .
interestingly , eq
. ( [ mrhofpinc ] ) coincides with the one obtained from qcd sum rules @xcite .
the large @xmath9 scaling in ( [ mrhofpinc ] ) also matches : @xmath288 , @xmath289 .
another distinct feature of the model is that the pion form factor is dominated by a single @xmath2 pole .
indeed , the coupling @xmath90 is found by substituting eq .
( [ bncosh ] ) into eq .
( [ gnpipiu ] ) , @xmath290 which vanishes for all @xmath291 .
this is due to the orthogonality of legendre functions and the fact that @xmath292 therefore @xmath2 meson dominance is exact for the pion form factor : @xmath293 . for @xmath76
one finds @xmath294 the ksrf ratio in this model is equal to 3 .
this means that the @xmath2 width is underpredicted by a factor of about 2/3 .
however , it is still interesting to compute @xmath235 for arbitrary @xmath9 , in the chiral limit , @xmath295 the rate of the electromagnetic decay @xmath239 in this model , @xmath296 is equal to 6.5 kev at @xmath286 , which is rather close to the observed @xmath242 kev . interestingly ,
the prediction from qcd sum rules @xcite is close but different in this case : @xmath297 .
the phenomenology of the @xmath78 meson in this model is discussed in appendix [ app : a1 ] .
the excitations with @xmath298 have an unrealistic mass spectrum in our model , so we shall not discuss their phenomenology .
the baryon appears in the framework of chiral lagrangians as a solitonic object : a skyrmion @xcite .
one wonders : what is the corresponding object in 5d that can describe the baryon .
an obvious candidate is the instanton , which can be `` lifted '' to become a quasiparticle in 5d .
here we show that the instanton appears from the point of view of 4d as a skyrmion .
we are interested only in topological aspects , and defer the question of stability of such a solution to future work . ,
@xmath2 , and @xmath78 mesons has been studied @xcite .
it was determined that vector mesons not only stabilize the skyrmion , but also noticeably improve agreement with phenomenology . ] on an intuitive level , to see the relation between the instanton and the skyrmion , one can consider as an example the well - known instanton solution in the singular gauge ( in the flat background metric ) : @xmath300 in this solution we shall think of @xmath301 as a four - vector with coordinates @xmath302 running through 1 , 2 , 3 , and 5 ( i.e. , @xmath118 , @xmath303 , @xmath304 , and @xmath162 ) and @xmath305 . note that the metric signature for these 4 coordinates is euclidean @xmath306 ( see ( [ metric ] ) ) .
the solution we wish to use to describe a baryon is static , i.e. , @xmath307 , and there is no dependence on @xmath308 . to see the behavior of the pion field we need to look at @xmath309 ( see ( [ sigmacont ] ) ) : @xmath310 we see that at every fixed @xmath162 the solution is a hedgehog , thus having the same topology as the skyrmion made of the pion field
. we shall now show that for an arbitrary background metric the topological charge of the instanton is equal to the baryon charge of the pion - field skyrmion .
our discussion is very similar to that of refs .
the 5d yang - mills theory possesses a conserved topological current , @xmath311 here @xmath312 is defined so that its elements are @xmath313 . for simplicity
we assume that @xmath283 is a constant and absorb it into the gauge field , and drop the hat in the 5d lorentz indices in subsequent formulas .
the boundaries are assumed to be @xmath257 .
that this current is conserved , @xmath314 , can be shown by using the bianchi identity @xmath315}=0 $ ] .
the topological charge of a static solution is @xmath316 the numerical coefficient in eq .
( [ 5dtopcurr ] ) was chosen so that the static instanton has unit total charge .
now consider a field configuration where the pion field , given by the wilson line along the @xmath162 coordinate ( [ sigmacont ] ) has a nontrivial winding , and @xmath317 goes to 0 at the boundaries . to compute the topological charge of this configuration , it is convenient to perform a gauge transformation to set @xmath318 .
explicitly , @xmath319 according to ( [ sigmacont ] ) , @xmath320 .
thus while @xmath321 remains 0 at the left boundary @xmath322 , it becomes nonzero at the right boundary : @xmath323 by using the identity @xmath324 one can rewrite the topological charge ( [ qff ] ) as @xmath325 since at @xmath322 @xmath326 and @xmath327 . by using eq .
( [ aibound ] ) one transforms this expression to @xmath328.\ ] ] it is now obvious that the topological charge becomes the winding number of the pion field .
therefore , the instanton becomes a skyrmion , and corresponds to the physical baryon .
we considered a theory of an `` open moose '' given by lagrangian ( [ l ] ) illustrated in fig .
[ fig : k ] .
this model describes a multiplet of massless goldstone bosons and a tower of vector and axial vector mesons . we developed a formalism for calculating the mass spectrum and the coupling constants in this theory for arbitrary parameters of the moose , @xmath34 and @xmath35 , and determine their values in the continuum limit , when the number of hidden symmetry groups @xmath0 tends to infinity .
we applied this formalism to two exactly solvable realizations of the model and found that the physics of the lowest modes match quite well with the phenomenology of the @xmath83 , @xmath2 and @xmath78 mesons .
we also find that the open - moose theory naturally incorporates the phenomenon of vector meson dominance .
for example , the pion form factor is saturated by poles from a tower of vector mesons . moreover , since couplings between mesons are given by overlap integrals , the couplings of highly excited @xmath2 s to the pion are suppressed by the oscillations of their wave functions in the fifth dimension .
this means that the pion form factor should be well approximated by the sum of contributions from a few lowest @xmath2 s . in the second example we considered ( the `` cosh '' background ) the situation
is brought to an extreme : the pion form factor is saturated by a single pole @xmath2-meson dominance .
we verified that both weinberg s spectral sum rules are automatically obeyed , in a nontrivial way , in any open - moose theory .
one of our original motivations was to include the excited vector mesons beyond the lowest @xmath78 .
with respect to that goal , we achieved only limited success , at least within the two exactly solvable models we considered . on the one hand , we do find that vector and axial vector mesons alternate in the spectrum , as it seems to be the case in qcd , at least for a few excited states . on the other hand , in both our simple models ,
the mass of an @xmath80-th state @xmath66 is @xmath329 at large @xmath80 , which seems to be in contradiction with the real world , and with the theoretical prejudice that @xmath330 .
further study of different backgrounds might provide a model that reproduces desired features of excited mesons and help understand constraints that phenomenology and qcd theorems impose on functions @xmath220 and @xmath221 .
alternatively , it is also possible that the excited vector mesons have `` stringy '' nature and can not , in principle , be incorporated into our field - theoretical scheme . by a suitable choice of background , even an exactly solvable one .
but we did not find such models viable in other respects . ]
the success that the model enjoys in describing the lowest states can be attributed to an apparent property of low - energy qcd : at intermediate distances correlation functions are reasonably well saturated by a single pole . in the `` cosh '' model the excited mesons ensure the correct behavior of the ( averaged ) spectral densities , thus playing the role of the continuum .
this explains why some results of qcd sum rules are well reproduced .
we hope that the study of the open moose theories will deepen our understanding of qcd at the fundamental level .
one intriguing fact discovered in these theories is the similarity to the ads / cft correspondence .
the procedure of calculating current - current correlators is essentially equivalent to the well - known ads / cft prescription : the correlators are given by the variational derivatives of the classical 5d action of the dual theory with respect to the sources living on the 4d boundary .
there is overwhelming evidence that the @xmath331 supersymmetric yang - mills theory is described by a string theory .
perhaps , an open moose theory is a low - energy limit of the string theory dual to qcd . in this regard ,
the result we found in the `` cosh '' model , @xmath332 is reassuring in view of a general expectation that such a dual theory should have a coupling proportional to @xmath333 in the t hooft limit . among the questions left for further study
is the detailed phenomenology of isoscalar mesons ( @xmath11 , @xmath12 , @xmath13 , etc . ) .
these mesons are described by an additional 5d abelian gauge field , which should be introduced into the action ( [ 5ds ] ) .
most of our results should generalize straighforwardly to this case .
however , there is an important new issue that the isoscalar sector brings into the theory .
the global u(1)@xmath334 symmetry must be explicitly broken , e.g. , @xmath11 should not be massless .
it is very encouraging that the 5d formulation of the theory provides a very natural mechanism for this .
it is the topological 5d chern - simons term of the form @xmath335 where @xmath336 is the 5d vector field describing isoscalars .
this term breaks the u(1)@xmath334 symmetry in the desired way .
in particular , it is not invariant under u(1)@xmath334 transformations on the 4d boundary ( although it is invariant under local transformations in the bulk of 5d ) .
it is easy to see that it also provides @xmath337 and other anomalous processes in qcd .
the coefficient of the term ( [ aff ] ) can be fixed by matching to qcd chiral anomaly , and is therefore proportional to @xmath9 .
the term ( [ aff ] ) also couples the @xmath12 meson field to the baryon current , providing a hard - core repulsion between baryons , and preventing the baryon / instanton from shrinking to zero size ( this effect is the origin of the stabilization of the skyrmion observed in ref .
. it would be also interesting to see how the open - moose theory realizes di vecchia - veneziano - witten lagrangian @xcite and the corresponding phenomenology .
other avenues for future study are the incorporation of finite quark masses , extension to three flavors and realization of the wess - zumino - witten topological term ( which does require a 5th dimension @xcite ) .
* acknowledgments * the authors thank s. r. beane , g. gabadadze , w - y .
keung , and especially t. imbo and m. strassler for discussions .
the authors also wish to acknowledge the review talk by r. l. jaffe at the _ qcd and string theory _ workshop at the institute for nuclear theory in seattle , which provided much of the motivation for this work .
m.a.s . acknowledges the hospitality of the institute for nuclear theory , university of washington , where part of this work has been done , and thanks riken bnl center and u.s .
department of energy [ de - ac02 - 98ch10886 ] for providing facilities essential for the completion of this work .
d.t.s . is supported , in part , by doe grant no.doe-er-41132 and the alfred p. sloan foundation .
m.a.s . is supported , in part , by a doe oji grant and by the alfred p. sloan foundation .
app : mnp for reference , we provide here some additional formulas for interaction vertices in the continuum limit of an arbitrary open - moose model .
let us define @xmath338 , @xmath339 , @xmath340 , and @xmath341 so that the lagrangian contains @xmath342 then @xmath343,{\label}{pimn}\\ g_{\pi\pi mn } & = & -\frac{f_\pi^2}8\int\!\frac{du}{f^2(u)}\ , b_m b_n,{\label}{pipimn}\\ g_{mnp } & = & \int\!du\ , gb_m b_n b_p,{\label}{mnp}\\ g_{mnpq } & = & \int\!du\ , g^2 b_m b_n b_p b_q,{\label}{mnpq}\end{aligned}\ ] ] direct couplings to external currents are suppressed in the continuum limit ( this includes , in particular , vector - meson dominance by a whole tower of mesons ) .
a simple qualitative interpretation of these couplings exists in terms of the overlaps of the wave functions in the @xmath162 space , which reflects the property of locality of the theory ( [ l ] ) or ( [ 5ds ] ) .
it is straightforward for the last two , resonance - resonance , couplings ( [ mnp ] ) and ( [ mnpq ] ) .
these terms come from the second , @xmath344 term in ( [ l ] ) . for the pion - resonance couplings ( [ pimn ] ) , ( [ pipimn ] ) and ( [ gnpipiu ] )
, one should bear in mind that the strength of the coupling is proportional to @xmath345 , and think of the pion wave function as being proportional to @xmath346 ( looking at ( [ sigmapi ] ) ) .
the @xmath162 derivatives in ( [ pimn ] ) are necessary to account for the fact that , although the pion wave function is even in @xmath347 , the pion itself is a pseudoscalar .
instead of expressing the results in terms of the parameters of the model @xmath348 and @xmath283 , we will use @xmath16 and @xmath9 .
the relations are @xmath349 @xmath350 @xmath351 @xmath352 rho - meson dominance of the pion form factor : @xmath353 @xmath354 there is a `` @xmath355 rule '' for pion emission : @xmath356 @xmath357 there is also a `` triangle rule '' for triple resonance vertex : @xmath358 i.e. , the amplitude vanishes if a triangle ( even a degenerate one ) with sides @xmath359 , @xmath360 and @xmath361 does not exist .
app : a1 let us discuss the phenomenology of the lowest axial vector meson ( the @xmath77 excitation in the open moose ) . from eq .
( [ mncosh ] ) , the mass of the @xmath78 meson in the `` cosh '' model is @xmath362 .
the @xmath78 decays into @xmath363 with the coupling ( [ gnn+1 ] ) @xmath364 by using the formula @xcite @xmath365 we find @xmath366 experimentally , the total width of @xmath78 is 250 to 600 mev , of which about 60% comes from @xmath367 @xcite .
the @xmath78 decay constant in our model is @xmath368 a lattice measurement of this constant yields ( in our normalization ) @xmath369
gev@xmath370 @xcite , while an analysis of hadronic @xmath371 decays gives @xmath372 gev@xmath370 @xcite .
the agreement is fair , but not exceptionally good .
it can be checked that the two terms cancel each other exactly , so the amplitude vanishes . on the other hand ,
the partial width of this decay is quoted to be @xmath376 kev @xcite .
the simplest @xmath39 hidden local symmetry model also suffers from the same problem ; in ref .
@xcite this was cured by adding higher - derivative terms to the action .
it would be interesting to see if this rate can be made nonzero by adding more terms to the action ( [ 5ds ] ) . c. n. yang and r. l. mills , phys .
* 96 * , 191 ( 1954 ) .
m. bando , t. kugo , s. uehara , k. yamawaki and t. yanagida , phys .
lett . * 54 * , 1215 ( 1985 ) .
m. bando , t. fujiwara and k. yamawaki , prog .
* 79 * , 1140 ( 1988 ) . for a review see m. bando , t. kugo and k. yamawaki , phys .
rept . * 164 * , 217 ( 1988 ) .
h. georgi , nucl .
b * 331 * , 311 ( 1990 ) .
k. hagiwara _ et al .
_ [ particle data group collaboration ] , phys . rev .
d * 66 * , 010001 ( 2002 ) .
m. b. halpern and w. siegel , phys .
d * 11 * , 2967 ( 1975 ) .
j. m. maldacena , adv .
theor . math .
* 2 * , 231 ( 1998 ) [ int .
j. theor .
phys . * 38 * , 1113 ( 1999 ) ] [ arxiv : hep - th/9711200 ] .
s. s. gubser , i. r. klebanov and a. m. polyakov , phys .
b * 428 * , 105 ( 1998 ) [ arxiv : hep - th/9802109 ] .
e. witten , adv .
* 2 * , 253 ( 1998 ) [ arxiv : hep - th/9802150 ] . for a review see o. aharony , s. s. gubser , j. m. maldacena , h. ooguri and y. oz , phys . rept . *
323 * , 183 ( 2000 ) [ arxiv : hep - th/9905111 ] .
a. m. polyakov , int .
j. mod .
phys . a * 14 * , 645 ( 1999 ) [ arxiv : hep - th/9809057 ] .
n. arkani - hamed , a. g. cohen and h. georgi , phys .
* 86 * , 4757 ( 2001 ) [ arxiv : hep - th/0104005 ] .
c. t. hill , s. pokorski and j. wang , phys .
d * 64 * , 105005 ( 2001 ) [ arxiv : hep - th/0104035 ] .
n. m. kroll , t. d. lee and b. zumino , phys .
rev . * 157 * , 1376 ( 1967 ) .
s. weinberg , phys .
lett . * 18 * , 507 ( 1967 ) .
k. sfetsos , nucl .
b * 612 * , 191 ( 2001 ) [ arxiv : hep - th/0106126 ] .
a. falkowski and h. d. kim , jhep * 0208 * , 052 ( 2002 ) [ arxiv : hep - ph/0208058 ] .
l. randall , y. shadmi and n. weiner , jhep * 0301 * , 055 ( 2003 ) [ arxiv : hep - th/0208120 ] . k. kawarabayashi and m. suzuki , phys .
* 16 * , 255 ( 1966 ) .
riazuddin and fayyazuddin , phys .
rev . * 147 * , 1071 ( 1966 ) .
m. kruczenski , d. mateos , r. c. myers and d. j. winters , jhep * 0307 * , 049 ( 2003 ) [ arxiv : hep - th/0304032 ] .
s. r. beane , phys . rev .
d * 64 * , 116010 ( 2001 ) [ arxiv : hep - ph/0106022 ] . m. a. shifman , a. i. vainshtein and v. i. zakharov , nucl .
b * 147 * , 448 ( 1979 ) .
t. h. skyrme , nucl .
phys . * 31 * , 556 ( 1962 ) .
g. s. adkins and c. r. nappi , phys .
b * 137 * , 251 ( 1984 ) .
y. igarashi , m. johmura , a. kobayashi , h. otsu , t. sato and s. sawada , nucl .
b * 259 * , 721 ( 1985 ) .
meissner and i. zahed , phys .
lett . * 56 * , 1035 ( 1986 ) .
for a review see also i. zahed and g. e. brown , phys .
* 142 * , 1 ( 1986 ) . c. t. hill and p. ramond , nucl .
b * 596 * , 243 ( 2001 ) [ arxiv : hep - th/0007221 ] .
c. t. hill , phys .
lett . * 88 * , 041601 ( 2002 ) [ arxiv : hep - th/0109068 ] .
a. karch and e. katz , jhep * 0206 * , 043 ( 2002 ) [ arxiv : hep - th/0205236 ] .
p. di vecchia and g. veneziano , nucl .
b * 171 * , 253 ( 1980 ) .
e. witten , nucl .
b * 223 * , 422 ( 1983 ) .
m. wingate , t. degrand , s. collins and u. m. heller , phys .
lett . *
74 * , 4596 ( 1995 ) [ arxiv : hep - ph/9502274 ] .
n. isgur , c. morningstar and c. reader , phys .
d * 39 * , 1357 ( 1989 ) . | motivated by phenomenological models of hidden local symmetries and the ideas of dimensional deconstruction and gauge / gravity duality , we consider the model of an `` open moose . ''
such a model has a large number @xmath0 of hidden gauge groups as well as a global chiral symmetry . in the continuum limit
@xmath1 the model becomes a 4 + 1 dimensional theory of a gauge field propagating in a dilaton background and an external space - time metric with two boundaries .
we show that the model reproduces several well known phenomenological and theoretical aspects of low - energy hadron dynamics , such as vector meson dominance .
we derive the general formulas for the mass spectrum , the decay constants of the pion and vector mesons , and the couplings between mesons .
we then consider two simple realizations , one with a flat metric and another with a `` cosh '' metric interpolating between two ads boundaries .
for the pion form factor , the single pole @xmath2-meson dominance is exact in the latter case and approximate in the former case .
we discover that an ads / cft - like prescription emerges in the computation of current - current correlators .
we speculate on the role of the model in the theory dual to qcd . |
there has been much recent interest in the study of spatiotemporal dynamical processes on nonuniform or complex networks . in this context ,
coupled map lattices @xcite have provided fruitful and computationally efficient models for the investigation of a variety of dynamical processes in spatially distributed systems . in particular , the discrete - space character of coupled map systems makes them specially appropriate for the investigation of spatiotemporal dynamics on nonuniform networks that can represent models of heterogeneous media .
the nonuniformity may be due to the intrinsic heterogeneous nature of the substratum , typical of pattern formation in biological contexts , or it may arise from random imperfections or fluctuations in the medium at some length scales .
such heterogeneities can have significant effects on the formation of spatial patterns , for example , they can induce reberverators in excitable media and defects can serve as nucleation sites for domain growth processes .
recently , the study of the phase - ordering properties of systems of coupled chaotic maps and their relationship with ising models in statistical physics has been a focus of attention @xcite .
these works have mainly assumed the phase competition dynamics taking place on an ordered spatial support .
this article investigates the phenomenon of phase growth in coupled chaotic maps on lattices with randomly distributed impurities as a model for studying this process on nonuniform media .
in particular , this model of coupled maps on nonuniform networks yields a scenario to explore the role that the local configurations of the underlying lattice play on the statistical properties of phase ordering processes on spatiotemporal systems .
in addition , this model can be seen as a simpler but computationally more efficient alternative to the study of phase separation phenomena on media with impurities than conventional computational fluid dynamics techniques . in sec .
ii , we present a procedure for the construction of a lattice with impurities and define the coupled map model on this network .
the phase growth dynamics in the presence of impurities is studied in sec .
the phase diagram of the system on the space of its parameters is obtained in this section .
section iv contains the conclusions of this work .
to generate a lattice with impurities , we start from a two - dimensional array of cells of size @xmath0 with periodic boundary conditions and remove a given fraction @xmath1 of cells at random .
the removed sites can be considered as impurities , defects , or random imperfections on the spatial support .
we define an impurity as a non - active site , i.e. , a site that possesses no dynamics .
the fraction of impurities in the lattice can be characterized in terms of the minimum euclidean distance @xmath2 between impurities , as shown in figure 1 . in this way
, the density @xmath1 scales with the minimum distance between impurities as @xmath3 .
we shall use values @xmath4 in order to avoid contiguous impurities on the lattice .
( left ) spatial support showing active sites ( @xmath5 ) and impurities ( @xmath6 ) .
impurities are randomly placed keeping a minimum distance @xmath2 between them .
( right ) the density of impurities scales as @xmath3 . ] the @xmath7 active sites can be enumerated by an index @xmath8 .
the equations describing the dynamics of the diffusively coupled map system defined on such nonuniform lattice are @xmath9 where @xmath10 gives the state of an active site @xmath11 at time @xmath12 ; @xmath13 is the set of the nearest active neighbors ( von neumann neighborhood ) of site @xmath11 , and @xmath14 is the cardinality of @xmath13 ; the parameter @xmath15 $ ] measures the coupling strength , and @xmath16 is a map that expresses the local dynamics .
impurities are not coupled to any other cell in the lattice .
the above coupled map equations on a lattice with impurities can be generalized to include other coupling schemes and neighborhoods , higher dimensions or continuous - time local dynamics .
different spatiotemporal phenomena can be also be studied on such nonuniform structures by providing appropriate local dynamics and couplings . in order to describe a bistable dynamics
, we assume a piecewise local map @xcite @xmath17 \\ \mu x , & \mbox{if } & x \in [ -1/3,1/3 ] \\ 2\mu/3 - \mu x , & \mbox{if } & x \in [ 1/3,1 ] \;.
\end{array}\right.\ ] ] when the parameter @xmath18 the map has two symmetric chaotic band attractors , one with values @xmath19 and the other with @xmath20 , separated by a finite gap about the origen .
then the local states have two well defined symmetric phases that can be characterized by spin variables defined as the sign of the state at time @xmath12 , @xmath21 .
we fix the local parameter at @xmath22 and set the initial conditions as follows : one half of the active sites are randomly chosen and assigned random values uniformly distributed on the positive attractor while the other half are similarly assigned values on the negative attractor . if the number of active sites is odd , then the state of the remaining site is assigned at random on either attractor . in regular lattices ( @xmath23 )
phase growth occurs for values @xmath24 , where @xmath25 @xcite .
in contrast , in the medium with impurities there exist a minimum value of @xmath1 for which the domains formed by the two phases reach a frozen configuration for all values of the coupling @xmath26 .
figure 2 shows stationary patterns emerging in the system for different values of parameters .
the top panels show that the average size of domains decreases when the density of impurities in the system is increased .
the bottom panels reveal the presence of complex domains where both phases coexists in a chessboard ( also called antiferromagnetic ) configuration for large values of the coupling strength . with impurities , for different values of parameters .
top left : @xmath27 , @xmath28 .
top right : @xmath29 , @xmath28 .
bottom left : @xmath27 , @xmath30 .
bottom right : @xmath27 , @xmath31.,scaledwidth=40.0% ] to characterize the phase ordering properties of the system eq .
( [ cml ] ) we use the normalized size of one phase domain as a function of time as an order parameter , defined as @xmath32 where @xmath33 is the euclidean distance between sites @xmath11 and @xmath34 .
figure 3 shows the average asymptotic value @xmath35 as a function of @xmath26 in the absence of impurities .
the quantity @xmath35 exhibits a continuous phase transition from a heterogeneous state , characterized by a small value of @xmath36 , to a homogeneous state , for which @xmath37 , at the critical value @xmath38 of the coupling parameter reported in @xcite . as a function of @xmath26 for @xmath39 ,
averaged over @xmath40 realizations of initial conditions .
error bars correspond to the standard deviations.,scaledwidth=40.0% ] before reaching its stationary value , the average domain size increases in time as @xmath41 , where the exponent @xmath42 characterizes the rate of phase growth .
figure 4 shows the evolution of @xmath43 with fixed value of @xmath26 for different values of the distance @xmath2 between impurities . vs. @xmath12 for @xmath44 ( empty circles ) , and @xmath45 ( solid circles ) , with fixed @xmath28 , averaged over @xmath40 realizations .
the values of the scaling exponents @xmath46 and @xmath47 , as well as a typical error bar are indicated on each curve.,scaledwidth=40.0% ] we observe that the phase growth process follows two well differentiated regimes during its time evolution . at early times
@xmath48 , we find @xmath49 , with an exponent @xmath50 whose value is not appreciably affected by the presence of impurities . for later times , the scaling behavior changes to @xmath51 , with @xmath52 indicating that domains grow faster in this regime , and where @xmath47 depends on the density of impurities .
impurities inhibit the growth of the phase domains and prevent the system from reaching a homogeneous state for values of @xmath53 .
figure 4 shows that the average size of the resulting domains in the inhomogeneous state are smaller in the presence of impurities , in agreement with the behavior observed in other bistable systems on media with impurities @xcite .
theoretical models and experiments with binary fluids have also shown that phase growth is restricted when impurities are present @xcite the domain growth regimes characterized by the exponents @xmath46 and @xmath47 in fig .
4 can be understood in terms of the stability of local configurations of the two phases . with this aim
, we define the fraction of sites in a given phase that have @xmath54 neighbors in that same phase at time @xmath12 , given by @xmath55 with @xmath56 ; where @xmath57 note that @xmath58 .
figure 5 shows the local spatial configurations corresponding to @xmath56 .
the patterns associated to @xmath59 represent configurations where the two phases are in contact , while the homogeneous state corresponds to @xmath60 .
, for @xmath59 , and @xmath61.,scaledwidth=40.0% ] figure 6 shows the time evolution of the fractions @xmath62 for two different values of @xmath2 .
we observe that in both cases vs. @xmath12 for different values of @xmath2 , with fixed @xmath28 .
@xmath63 ( continuous line ) ; @xmath64 ( long - dashed line ) ; @xmath65 ( short - dashed line ) ; and @xmath66 ( dotted line ) . ( a ) @xmath67 .
( b ) @xmath68.,scaledwidth=40.0% ] vs. @xmath12 for different values of @xmath2 , with fixed @xmath28 .
@xmath63 ( continuous line ) ; @xmath64 ( long - dashed line ) ; @xmath65 ( short - dashed line ) ; and @xmath66 ( dotted line ) . ( a ) @xmath67 .
( b ) @xmath68.,scaledwidth=40.0% ] @xmath63 and @xmath69 vanish at @xmath70 .
thus , the decay of in the number of local configurations associated to @xmath71 and @xmath72 determines the regime of slow domain growth characterized by the exponent @xmath46 in fig .
phase growth in this earlier regime occurs by the addition of single cells to domains whose size is of the order of one cell .
the spatial scale involved in this process is smaller than the minimum distance between impurities @xmath2 and , therefore , the presence of impurities does not affect this mechanism of phase growth . as a consequence ,
the exponent @xmath46 does not depend on @xmath2 . for @xmath73 , the fractions @xmath65 and @xmath66
keep decaying until they reach their asymptotic values .
phase growth in this second regime occurs when domains in the same phase and whose average sizes are greater than one cell enter in contact .
this process is illustrated in fig .
7 . at the local level , when two domains of equal phase get separated by a one - cell thick layer of cells in the opposite phase , there are cells on this layer that posses a local configuration described by @xmath74 .
the coupled map dynamics , eqs .
( [ cml ] ) , acting on a cell in such local configuration will produce a change of its phase , yielding first another configuration of type @xmath74 , then transforming into a configuration of type @xmath75 , and finally into the configuration associated to @xmath60 . as a result ,
the domains join together forming a domain whose average size is much larger than the sizes of the initial domains . through this growth mechanism , the average size of domains in the system increases faster than by the process the successive additions of single cells that takes place at earlier times .
the difference in the growth velocity of phase domains in these two regimes is manifested by the fact that @xmath76 . when impurities are present , the average size of domains is comparable to the minimum distance @xmath2 .
thus , there is a high probability that some impurities lie on the interface .
these impurities constrain the transformations of the local configurations @xmath74 into @xmath75 , and into @xmath60 described above .
consequently , the phase growth process becomes slower , and the fractions @xmath65 and @xmath66 reach greater asymptotic values in the presence of impurities , as manifested in fig .
therefore , the asymptotic average size of domains for @xmath77 is smaller than that for @xmath23 , as it was shown in fig .
figure 8 shows the growth exponents @xmath46 and @xmath47 as functions of the coupling parameter @xmath26 , for two different values of the density of impurities @xmath1 .
we observe that @xmath46 increases slowly with @xmath26 and , as it was already manifested in fig . 4 ,
its behavior is unaffected by the presence of impurities .
in contrast , the exponent @xmath47 appears above some threshold value @xmath78 and its behavior depends on both @xmath26 and @xmath1 .
( dashed line ) and @xmath47 ( continuous lines ) vs. @xmath26 for different values of @xmath2 .
the values shown are averages obtained over @xmath40 realizations of initial conditions foe each value of @xmath26 .
empty circles : @xmath46 for @xmath44 ; empty squares : @xmath46 for @xmath45 ; solid squares : @xmath47 for @xmath67 ; solid circles : @xmath47 for @xmath68 .
typical error bars are shown on each curve.,scaledwidth=40.0% ] figure 9 shows the growth exponent @xmath47 calculated on the space of parameters @xmath79 .
there is a critical boundary @xmath80 that separates the regimes where fast growth occurs , characterized by @xmath81 , from the regime where only slow growth takes place , corresponding to @xmath82 , on the plane @xmath79 .
near the critical boundary @xmath80 , the exponent @xmath47 can be described by the scaling relation @xmath83 , where @xmath84 also depends on @xmath2 or , equivalently , on the density of impurities @xmath1 . as a function of @xmath26 and @xmath2 .
the critical boundary @xmath80 that separates the fast growth regime from slow growth regime is indicated by a thick continuous line on the plane @xmath85 .
the values are obtained as in fig .
8.,scaledwidth=40.0% ] figure 10 shows the critical exponent @xmath84 versus @xmath2 .
the exponent @xmath84 can be well fitted by the scaling relation @xmath86 , where @xmath87 and @xmath88 .
below @xmath89 there is no regime of fast phase growth . as a function of the distance between impurities @xmath2 , obtained as averages over @xmath40 realizations of initial conditions for each parameter value .
error bars are shown.,scaledwidth=40.0% ] our results show that it is possible to control both the growth velocity and the size of phase domains in bistable media by the inclusion of impurities in the spatial support of the system . when the density of impurities increases above some critical value for a given strength of the coupling , both the velocity at which domains grow and their stationary sizes reach smaller values than those corresponding to the absence of impurities . to characterize the emergence of chessboard domains for large values of the coupling @xmath26 , as those shown in the bottom panels of fig . 2
, we calculate the asymptotic fractions @xmath90 as functions of @xmath26 in fig . 11 . versus @xmath26 for @xmath23 .
@xmath91 ( continuous line ) ; @xmath92 ( long - dashed line ) ; @xmath93 ( short - dashed line ) ; and @xmath94 ( dotted line ) . values shown correspond to averages over @xmath40 realizations of initial conditions.,scaledwidth=40.0% ] note that @xmath91 , which vanished rapidly for small values of @xmath26 , becomes different from zero and increases for @xmath95 .
the local configuration corresponding to @xmath71 is effectively associated to a chessboard pattern .
thus the appearance of a finite value of the fraction @xmath91 at this large value of @xmath26 signals the onset of chessboard domains in the system .
furthermore , we have found that the presence of impurities has little effect on the behavior of @xmath91 , suggesting that the existence of chessboard domains is mainly associated to large values of the coupling parameter and to the properties of the local dynamics .
figure 12 summarizes the collective behavior of the coupled map lattice with bistable dynamics in the presence of impurities , eq .
( [ cml ] ) .
this figure shows the phase diagram of this system on the space of parameters @xmath96 .
the regions where the homogeneous ( one phase ) and heterogeneous ( two coexisting phases ) , as well as the region where chessboard ( antiferromagnetic ) states occur are indicated .
the critical boundary @xmath97 for the onset of the regime for fast growth of domains @xmath98 on this plane is marked by a dashed line . ) on the space of parameters @xmath96 .
the regions where homogeneous , heterogeneous , and chessboard phases exist are indicated by labels .
the dashed line indicates the critical boundary for the onset of the regime for fast growth of domains characterized by @xmath99 .
the critical value @xmath38 for the occurrence of domain growth in the absence of impurities is marked on the vertical axis.,scaledwidth=40.0% ]
we have considered a system of chaotic maps coupled on a lattice with randomly distributed empty sites as a model to study phase ordering processes on media with impurities .
the density of impurities @xmath1 is related to the minimum distance between impurities @xmath2 .
we have shown that the rate at which phase domains grow becomes smaller when impurities are present in the system . on the other hand ,
the average size of the resulting domains in the inhomogeneous state of the system decreases when the density of impurities is increased .
we have calculated the critical boundary that separates the regime of slow growth of domains from the regime of fast growth on the space of parameters of the system @xmath85 . along this critical boundary ,
the exponent for fast growth @xmath47 exhibits scaling properties along both parameter axes , manifested by the existence of the critical exponents @xmath84 and @xmath100 in figs . 9 and 10 .
we have explained the transition between these two growth regimes in terms of the stability properties of the local configuration measured by the fractions @xmath62 . for large values of the coupling strength @xmath26 , the system displays a complex heterogeneous state consisting of domains of the two phases coexisting with spatial domains having a chessboard ( antiferromagnetic ) configuration .
we have found that the appearance of chessboard patterns is not appreciably affected by the presence of impurities .
the phase diagram showing the regions where the different behaviors of the system occur on the space of parameters @xmath96 was obtained .
our results indicate that the inclusion of impurities can be used as a control mechanism for the size and growth velocity of domains forming in bistable media .
recent studies of chaotic maps on complex networks @xcite indicate that topology may play a decisive role in determining emerging collective behaviors .
the present results suggest that spatial inhomogeneities may also be employed as a selection mechanism for patterns arising in general spatiotemporal systems .
this work was supported in part by grant i-886 - 05 - 02-a from cdcht , universidad de los andes , mrida , venezuela .
c.e . acknowledges support from decanato de investigacin , universidad nacional experimental del tchira , san cristbal , venezuela
. 99 chaos 2 ( 1992 ) 279 , focus issue on coupled map lattices ; edited by k. kaneko .
a. lemaitre and h. chat , phys .
lett . * 82 * , 1140 ( 1999 ) .
j. kockelkoren , a. lemaitre , and h. chat , physica a * 288 * , 326 ( 2000 ) .
w. wang , z. liu , and b. hu , phys .
lett . * 84 * , 2610 ( 2000 ) .
l. angelini , m. pellicoro , and s. stramaglia , phys .
a * 285 * , 293 ( 2001 ) .
f. schmser , w. just , and h. kantz , phys .
e * 61 * , 3675 ( 2000 ) .
l. angelini , phys .
a * 307 * , 41 ( 2003 ) .
k. tucci , m. g. cosenza and o. alvarez - llamoza , phys .
e * 68 * , 027202 ( 2003 ) .
j. miller and d. huse , phys .
e * 48 * , 2528 ( 1993 ) .
r. paul , s. puri and h. rieger , phys .
e * 71 * , 61109 ( 2005 ) .
k. yurekli , a. karim , e. j. amis y r. krishnamoorti , macromolecules * 36 * , 7256 ( 2003 ) .
f. qiu , g. peng , v. v. ginzburg , a. c. balazs , h. y. chen , d. jasnow , j. chem .
phys . * 115 * , 3779 ( 2001 ) .
m. g. cosenza and k. tucci , phys .
e * 65 * , 036223 ( 2002 ) . | a system of coupled chaotic bistable maps on a lattice with randomly distributed impurities is investigated as a model for studying the phenomenon of phase growth in nonuniform media .
the statistical properties of the system are characterized by means of the average size of spatial domains of equivalent spin variables that define the phases .
it is found that the rate at which phase domains grow becomes smaller when impurities are present and that the average size of the resulting domains in the inhomogeneous state of the system decreases when the density of impurities is increased .
the phase diagram showing regions where homogeneous , heterogeneous , and chessboard patterns occur on the space of parameters of the system is obtained . a critical boundary that separates the regime of slow growth of domains from the regime of fast growth in the heterogeneous region of the phase diagram is calculated .
the transition between these two growth regimes is explained in terms of the stability properties of the local phase configurations .
our results show that the inclusion of spatial inhomogeneities can be used as a control mechanism for the size and growth velocity of phase domains forming in spatiotemporal systems . |
_ aperiodic tiles _ are the set of prototiles which tile the space with their isomorphic images by euclidean motions ( composition of translations , rotations and reflections ) but only in non - periodic way .
there have been many examples of aperiodic tiles and study on them @xcite .
two of well - known examples of aperiodic tiles with simple prototiles up to euclidean motions are penrose tiles and ammann tiles which are uncovered in the mid 70 s .
these sets consist of two prototiles and it has been the smallest number of prototiles which form aperiodic tiles until recently . since then , people have been interested in finding a single prototile for an aperiodic tile .
this problem is coined as ` mono - tile ' problem or ` einstein ' problem ( one stone in german ) .
it had taken quite some time before taylor and socolar announced in 2010 the existence of an aperiodic mono - tile .
their tile is a hexagonal tile with colored decorations and matching rules which can be embedded onto a single tile using shape only .
penrose had found earlier a mono - tile using matching rules which is the reformulation of @xmath2 aperiodic tiles given in @xcite , but in this case it can not be replaced by a single tile using only the shape .
one needs two other tiles to replace the matching rules in the penrose mono - tile .
both of taylor - socolar mono - tile and penrose functional mono - tile are based on hexagonal shape .
but they make different tilings .
it would be interesting to understand the tiling space of penrose mono - tile tilings and compare with the tiling space of taylor - socolar tilings . around the mid 80 s , shechtman @xcite discovered a quasicrystal with forbidden rotational symmetry of crystal diffraction pattern . after the confirmation of the existence of quasicrystals ,
crystal is redefined as a material whose diffraction patterns consist essentially of bright peaks ( c.f .
since tilings made by aperiodic tiles , which we define as aperiodic tilings , are not periodic , they have served as good models for the structures of quasicrystals when they show the diffraction patterns consisting of pure point diffraction spectrum , i.e. , bragg peaks only without diffuse background .
many examples of known aperiodic tiling like penrose tiling and ammann tiling show the pure point diffraction spectrum .
the objective of this article is investigating whether the aperiodic taylor - socolar tiling , which is a fixed point of a substitution , has pure point diffraction spectrum . mathematically the pure point diffraction spectrum is quite often studied through the spectrum of the dynamics of the _ dynamical hull _
, that is , a compact space generated by the closure of translation orbits of the tiling .
the two notions of pure pointedness in diffraction and dynamical spectra are equivalent in quite a general setting @xcite . in general tilings ,
almost periodicity of tilings is an equivalent criterion for the pure point spectrum .
when it is restricted on substitution tilings , the almost periodicity can be easily checked by overlap coincidence .
briefly it means the following : when two tiles in a tiling intersect in the interior after shifting one tile by a translation of two other same type tiles in the tiling , one can observe a pair of same type subtiles in the same position in the common interior ( see subsec.[overlap ] ) @xcite .
an aperiodic taylor - socolar tiling itself does not follow tile - substitution rule strictly .
but as it is mentioned in @xcite , half - hexagonal tiles satisfy tile - substitution .
we should note here that the taylor - socolar half - hexagonal tiling is mutually locally derivable from the taylor - socolar tiling .
we consider the taylor - socolar half - hexagonal substitution tiling whose identical image in taylor - socolar tilings belongs to the dynamical hull generated by a repetitive taylor - socolar tiling .
we apply the substitution data on the half - hexagonal tiles of taylor - socolar tiles to the algorithm for checking the overlap coincidence .
the algorithm can be found in @xcite . as the result
, we were able to check that the half - hexagonal taylor - socolar substitution tiling has overlap coincidence .
so we can conclude that the aperiodic taylor - socolar tiling has pure point spectrum .
one can also note that a dynamical hull of taylor - socolar tilings is invariant under the action of rotations of @xmath3 . in the diffraction pattern of a taylor - socolar tiling , we observed six - fold rotational symmetry .
the tiling space of taylor - socolar tilings with the matching rules is slightly bigger than the tiling space of a taylor - socolar substitution tiling .
but it is shown in @xcite that the tilings in the difference have pure point spectrum , computing the total index of cosets .
thus any taylor - socolar tiling under the same matching rules have the pure point spectrum . in the case of self - similar tilings , the discrete part of the diffraction pattern , which is called bragg spectrum , can be characterized in terms of the fourier modules .
there are three types of bragg spectra- limit - periodic , quasiperiodic , limit - quasiperiodic .
the taylor - socolar tilings belong to limit - periodic structure , since the expansion factor is rational .
the fourier module of this structure is an aperiodic structure which is the limit of a sequence of periodic structures ( see @xcite ) .
a substitution tiling with half - hexagonal shapes was known much earlier ( see ( * ? ? ?
* ex.10.1 ) ) .
it is known from @xcite that the substitution point set representing the half - hexagonal substitution tiling is a cut - and - project set and so it has pure point spectrum .
however taylor - socolar tiling differs from @xcite in the sense that we consider a substitution tiling reflecting the aperiodicity of taylor - socolar mono - tile and have to distinguish protiles by their colors .
various other ways to observe the pure point spectrum are pointed out in @xcite .
one can observe that there is an one - to - one almost everywhere map from a dynamical hull of taylor - socolar tilings to a dynamical hull of half - hexagonal substitution tilings .
then this induces the pure point spectrum of taylor - socolar tilings using the result of @xcite .
the other observation would be through checking the modular coincidence which has been introduced in @xcite .
one can see in the figures of @xcite and ( * ? ? ?
15 ) that @xmath4 or @xmath5 type tiles form a sublattice structure of a whole hexagonal lattice with the expansion factor of @xmath6 . it would be sufficient to check if the modular coincidence occurs with these @xmath4 and @xmath5 type tiles .
furthermore @xcite provides a geometrical way to observe the limit - periodic structure in taylor - socolar tilings which shows that tiling can be decomposed into a superposition of periodic structure ( see fig.19 and thm.7.1 in @xcite )
. therefore the pure point spectrum may not be so surprising in the case of taylor - socolar tilings .
however comparing with the above methods , the biggest advantage of our method is that it is almost _
automatic _ and it can be applied to many variations of taylor - socolar substitutions based on hexagonal lattices , such as penrose mono - tile tiling , with minor changes of the substitution data and investigate the diffraction spectrum of a tiling generated by it .
furthermore it can be applied to substitution tilings whose underlying structures are not even on lattices .
we briefly mention the notions of tilings and tile - substitution in @xmath7 that we use in this paper . for more about tilings and tile - substitutions , see @xcite .
we begin with a set of types ( or colors ) @xmath8 .
a _ tile _ in @xmath7 is defined as a pair @xmath9 where @xmath10 ( the support of @xmath11 ) is a compact set in @xmath7 , which is the closure of its interior , and @xmath12 is the type of @xmath11 .
a _ tiling _ of @xmath7 is a set @xmath13 of tiles such that @xmath14 and distinct tiles have disjoint interiors .
we always assume that any two @xmath13-tiles with the same color are translationally equivalent .
let @xmath15 .
we say that a set @xmath16 of tiles is a _ patch _ if the number of tiles in @xmath16 is finite and the tiles of @xmath16 have mutually disjoint interiors .
we define @xmath17 for a bounded set @xmath18 .
we say that @xmath13 is _ repetitive _ if for every compact set @xmath19 , @xmath20 is relatively dense .
we say that a tiling @xmath13 has _ finite local complexity ( flc ) _
if for each radius @xmath21 there are only finitely many translational classes of patches whose support lies in some ball of radius @xmath22 . a _ multi - color set _ or _ @xmath23-multi - color set _ in @xmath24 is a subset @xmath25 ( @xmath23 copies ) where @xmath26 .
we also write @xmath27 . recall that a delone set is a relatively dense and uniformly discrete subset of @xmath24 .
we say that @xmath28 is a _
delone multi - color set _ in @xmath24 if each @xmath29 is delone and @xmath30 is delone .
we say that @xmath31 is a _
meyer set _ if it is a delone set and @xmath32 is uniformly discrete ( @xcite ) .
the types ( or colors ) of points on delone multi - color sets have the same concept as the colors of tiles on tilings .
we say that a linear map @xmath33 is _ expansive _ if all the eigenvalues of @xmath34 lie outside the closed unit disk in @xmath35 .
[ def - subst ] _ let @xmath36 be a finite set of tiles in @xmath7 such that @xmath37 ; we will call them _ prototiles_. denote by @xmath38 the set of non empty patches .
we say that @xmath39 is a _ tile - substitution _ ( or simply _ substitution _ ) with an expansive map @xmath34 if there exist finite sets @xmath40 for @xmath41 such that @xmath42 with @xmath43 here all sets in the right - hand side must have disjoint interiors ; it is possible for some of the @xmath44 to be empty . _
the substitution ( [ subdiv ] ) is extended to all translates of prototiles by @xmath45 and to patches and tilings by @xmath46 .
the substitution @xmath47 can be iterated , producing larger and larger patches @xmath48 .
we say that @xmath13 is a _ substitution tiling _ if @xmath13 is a tiling and @xmath49 with some substitution @xmath47 . in this case
, we also say that @xmath13 is a _
fixed point _ of @xmath47 .
we say that a substitution tiling is _ primitive _ if the corresponding substitution matrix @xmath50 , with @xmath51 , is primitive . a repetitive fixed point of a primitive tile - substitution with flc
is called a _ self - affine tiling_. if @xmath52 , we say that @xmath16 is a _
generating patch_. we say that @xmath28 is a _
delone multi - color set _ in @xmath7 if each @xmath29 is delone and @xmath53 is delone .
any tiling @xmath13 can be converted into a delone multi - color set by simply choosing a point @xmath54 for each tile @xmath55 so that the chosen points for tiles of the same type are in the same relative position in the tile : @xmath56 .
we define @xmath57 and @xmath58 .
clearly @xmath13 can be reconstructed from @xmath59 given the information about how the points lie in their respective tiles .
this bijection establishes a topological conjugacy of @xmath60 and @xmath61 .
concepts and theorems can clearly be interpreted in either language ( flc , repetitivity , pure point dynamical spectrum , etc . ) . if a self - affine tiling @xmath62 is given , we get an associated substitution delone multi - color set @xmath63 of @xmath13 ( see ( * ? ? ?
* lemma5.4 ) ) .
there are two notions of pure pointness in the study of tilings - pure point dynamical spectrum and pure point diffraction spectrum .
we briefly give the definitions of them .
let @xmath13 be a tiling in @xmath7 .
we define the space of tilings as the orbit closure of @xmath13 under the translation action : @xmath64 , in the well - known `` local topology '' : for a small @xmath65 two point sets @xmath66 are @xmath67-close if @xmath68 and @xmath69 agree on the ball of radius @xmath70 around the origin , after a translation of size less than @xmath67 .
the group @xmath7 acts on @xmath71 by translations which are obviously homeomorphisms , and we get a topological dynamical system @xmath61 .
let @xmath72 be an ergodic invariant borel probability measure for the dynamical system @xmath61 .
we consider the associated group of unitary operators @xmath73 on @xmath74 for which @xmath75 .
the dynamical system @xmath76 is said to have _
pure point(or pure discrete ) spectrum _ if the linear span of the eigenfunctions is dense in @xmath77 . on the other hand ,
there is a notion of pure point diffraction spectrum which characterizes quasicrystals .
let @xmath78 be a multi - color point set in @xmath7 .
we consider a measure of the form @xmath79 , where @xmath80 and @xmath81 .
the autocorrelation of @xmath82 is @xmath83 where @xmath84 is a measure of @xmath82 restricted on the ball @xmath85 of radius @xmath86 and @xmath87 is the measure , defined by @xmath88 , where @xmath89 is a continuous function with compact support and @xmath90 .
the diffraction measure of @xmath82 is the fourier transform @xmath91 of the autocorrelation ( see @xcite ) .
when the diffraction measure @xmath91 is a pure point measure , we say that @xmath59 has pure point diffraction spectrum and so @xmath59 has the structures of quasicrystals .
it turns out that the two notions of pure pointedness are same , i.e , the pure point dynamical spectrum of @xmath92 is equivalent to the pure point diffractivity of @xmath93 @xcite . a triple @xmath94 , with @xmath95 and @xmath96 ,
is called an _ overlap _ if @xmath97 where @xmath98 and @xmath99 .
we define @xmath100 the _ support of an overlap _ @xmath101 and denote it by @xmath102 .
an overlap @xmath94 is a _ coincidence _ if @xmath103 let @xmath104 be an overlap in @xmath13 , we define _ @xmath105-th
inflated overlap _
@xmath106 [ def - overlapcoincidence ] _ we say that a self - affine tiling @xmath13 admits an _ overlap coincidence _ if there exists @xmath107 such that for each overlap @xmath108 in @xmath13 , @xmath109 contains a coincidence . _
when @xmath13 is a self - affine tiling in @xmath7 such that @xmath110 is a meyer set , @xmath92 has a pure point dynamical spectrum if and only if @xmath13 admits an overlap coincidence@xcite .
so we will check the pure point spectrum of taylor - socolar tilings through the computation of overlap coincidence in the next subsection . in the half - hexagonal substitution tiling
, there are 14 half - hexagonal prototiles which come from dividing 7 hexagonal prototiles @xmath111 into the left and the right ( see ( * ? ? ?
* fig . 15 ) ) . since the substitution tiling we defined in ( [ def - subst ] )
requires finite prototiles up to only translations , we need to treat the rotated types and reflected types of the 14 half - hexagonal prototiles as different prototiles .
so we consider a substitution tiling with 168 prototiles .
using the algorithm in @xcite which is originated from @xcite and made for the computation of the taylor - socolar tilings , we check if the half - hexagonal substitution taylor - socolar tiling has pure point spectrum .
now a question would be `` when the dynamics of the half - hexagonal substitution tiling has pure point spectrum , can we infer that the dynamics of the original hexagonal tiling also has pure point spectrum ? '' .
let @xmath13 be a fixed point of a primitive substitution and @xmath112 .
it is shown in ( * ? ? ?
* lemma a.6 . ) that @xmath13 has uniform cluster frequencies(ucf ) , i.e. for any @xmath13-patch @xmath16 , there exists @xmath113 uniformly in @xmath114 . from (
* theorem 3.2 ) , the measure @xmath115 has pure point diffraction spectrum , for any choice of complex numbers @xmath116 if and only if each measure @xmath117 has pure point diffraction spectrum . by the construction of the half - hexagonal taylor - socolar tiling , we can take a substitution point set @xmath59 representing the half - hexagonal taylor - socolar tiling to include a substitution point set @xmath118 representing the original hexagonal taylor - socolar tiling .
so from ( * ? ? ?
* theorem 3.2 ) , we can conclude that taylor - socolar tiling has pure point diffraction spectrum . as a generating patch
, one can start with the patch as shown in figure [ ts - generating - patch ] .
since this patch is contained in the next inflated patch after the substitution , it gives a fixed tiling under the substitution .
for the computational reason , we give a tile - substitution whose expansion involves rotation and reflection .
it is possible to use the second iteration of half - hexagons without rotation and reflection parts on expansions .
but in this case , the substitution gets bigger . the tile - substitution for @xmath119
is shown in figure [ ts - substitution ] .
for other half - hexagonal tiles , the figures of tile - substitution is similar .
we give the precise tile - substitution below .
one can check the computational algorithm in @xcite .
, @xmath120 , @xmath121 , and @xmath122 .
the black dot indicates the origin.,scaledwidth=90.0% ] let @xmath123 where @xmath124 is a rotation of @xmath125 counter clockwise through the origin and @xmath126 is a reflection through @xmath127-axis .
the tile substitution is given as follows ; let @xmath128 and @xmath129 .
we denote @xmath130 , where @xmath131 , @xmath132 , and @xmath133 .
the tile - substitution rule for @xmath134 and @xmath121 is the following .
for @xmath135 , @xmath136 the tile substitution rule for other types of tiles are similar .
we give the rules for the convenience to the readers .
@xmath137 @xmath138 @xmath139 @xmath140 @xmath141 @xmath142 let us describe how the algorithm works in general term . when tile - substitution date @xmath47 is given , we consider a substitution delone multi - color set @xmath59 which is fixed under the substitution . to build the set @xmath59 , we need to find a point @xmath143 which is fixed under the substitution .
applying the substitutions to @xmath144 infinitly many times , we can easily obtain a point set @xmath59 which is fixed under the substitution .
it is sufficient to check the overlap coincidence for all the overlaps which occur by finite translation vectors of same type tiles in the tiling . from the meyer property ,
the number of overlaps are finite . after collecting all the overlaps
, we can check overlap coincidence for each overlap applying the substitution many times . here
the number of times of applying the substitution can be limited by the number of overlaps .
so the algorithm will be terminated .
the detail is given in @xcite .
the computation of overlap coincidence of taylor - socolar tiling takes rather long time comparing to other examples in @xcite .
we guess that it is due to the number of prototiles ( 168 ) which is much more than other examples .
it is an interesting question whether the complexity of the substitution rule is related with the computation time of the algorithm when the number of prototiles is same .
we are grateful to michael baake and franz ghler for their interest and helpful suggestions . | recently taylor and socolar introduced an aperiodic mono - tile .
the associated tiling can be viewed as a substitution tiling .
we use the substitution rule for this tiling and apply the algorithm of @xcite to check overlap coincidence .
it turns out that the tiling has overlap coincidence .
so the tiling dynamics has pure point spectrum and we can conclude that this tiling has a quasicrystalline structure .
* keywords * : mono - tile tiling , quasicrystals , pure point diffraction / dynamical spectrum , self - affine tilings , overlap coincidence .
shigeki akiyama @xmath0 and jeong - yup lee @xmath1 |
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 ( antonucci 1993 ) .
optical narrow emission lines present in both type 1 and type 2 seyfert s spectra are produced in the narrow line region ( nlr ) at @xmath8 pc scale from the nucleus .
optical broad emission lines originate in the broad line region ( blr ) at sub - pc scale .
the latter are observed only in type 1 seyfert s spectra since , in type 2 seyferts , they are obscured by a molecular torus .
much evidence has been found in favor of this picture , such as the larger amount of absorbing material measured from x - ray observations in seyfert 2s with respect to seyfert 1s ( risaliti et al .
1999 , awaki et al .
1991 ) .
however , in the last few years the number of cases in which observations do not match with unified models is increasing both in the local and in the distant universe .
type 1 active galactic nuclei ( agn ) with significant absorption have been found ( cappi et al .
2006 , mateos et al .
2005 , fiore et al .
2001 ) as well as type 2 agn without x - ray absorption ( brightman & nandra 2008 , bianchi et al .
2008 , wolter at al .
2005 , corral et al .
2005 , caccianiga et al .
2004 , barcons , carrera & ceballos 2003 , panessa & bassani 2002 , pappa et al .
2001 ) .
[ sam ] [ cols="<,^,^,^,^,^,^,^,^,^,^,^,^,^ " , ] notes : col ( 1 ) : galaxy name ; col ( 2)-(3 ) optical position in epoch j2000 ; col .
( 4 ) redshift ; col . ( 5 ) [ oiii]/h@xmath3 ratio ; col .
( 6 ) : amplitude of the @xmath9 band measured over the period from 1974 to 2002 ; col .
( 7 ) : observation date ; col .
( 8) : mos1/mos2/pn observation exposures ; col .
( 9 ) : mos1/mos2/pn filters , m = medium . columns ( 1 ) , ( 4)-(6 ) are from hawkins ( 2004 ) .
several explanations have been proposed to reconcile the unified model paradigm with these pieces of evidence .
for instance , the broad emission lines could fade away in response to a decrease of the continuum emission ( guainazzi et al .
2005 , matt et al .
alternatively , the blr is covered by clumpy and variable obscuring material , as in ngc4388 where a variation of a factor of 100 in column density has been observed from x - ray observations ( elvis et al .
2004 ) and in ngc1365 where a spectral change from compton - thin to compton - thick and back to compton - thin has happened in four days ( risaliti et al . 2007 ) . in the above mentioned cases ,
the misleading x - ray and optical behavior is basically due to non - simultaneity of the observations . finally , the blr could be weak or absent and its formation linked to the accretion physics ( wang & zhang 2007 , elitzur & shlosman 2006 , nicastro et al .
2003 , nicastro 2000 , williams et al .
simultaneous optical and x - rays observations have confirmed that ngc3147 is blr - free and without x - ray absorption ( bianchi et al .
the strategy of observing the source simultaneously in the two bands has been proved to be fundamental in order to avoid mismatches due to spectral or flux variability of the source at x - rays and optical wavebands .
hawkins ( 2004 ) presented the results from a long term monitoring of about 800 quasars . among them a new class of agn is reported , i.e. the naked " agn , where the absence of broad emission lines is accompanied by strong optical variability , suggesting that the nucleus is seen directly .
subsequently _ chandra _ snapshot observations of three naked " agn ( q2130 - 431 , q2131 - 427 and q2122 - 444 ) by gliozzi et al .
( 2007 ) confirmed this hypothesis given the absence of significant absorption in the x - ray spectra , though of low statistics . in this work we present quasi - simultaneous x - ray and optical spectroscopic observations with xmm-_newton _ and emmi / ntt ( _ la silla observatory _ , lso , eso ) of two sources ( q2130 - 431 and q2131 - 427 ) out of the six naked " agn sample defined in hawkins ( 2004 ) and
discuss their nature with respect to unified models and recent developments .
we also report on the serendipitous detection of the cluster in appendix a. in the following , we adopt @xmath10 km s@xmath11 mpc@xmath11 , @xmath12 and @xmath13 ( spergel et al . 2003 ) .
hawkins ( 2004 ) carried out a yearly photometrical large - scale monitoring programme for agn over the last 25 years .
the survey was based on a long series of photographic plates from the uk 1.2 m schmidt telescope .
candidate agn were selected from a catalogue of 200,000 objects to perform follow - up spectroscopic study .
optical spectra ( obtained in july 2002 ) pinpoint a class of objects ( 6 in that sample ) which show narrow and weak h@xmath3 emission line , large [ oiii]@xmath145007/h@xmath3 ratios typical of type 2 seyferts ( i.e. , [ oiii]@xmath145007/h@xmath3 @xmath15 3 , shuder & osterbrock 1981 ) , and no sign of broad emission lines , could not be determined since this line was redshifted out of the optical passband . ] . for this class of objects ,
the difference between the maximum and minimum light in magnitudes in the b@xmath16 passband over a period of 25 years , @xmath17b , reveals large amplitude variations normally only found in type 1 objects .
their brightness varied at least by a factor of 3 on a timescale of 5 - 10 years and also on shorter time scales ( see light curves in hawkins 2004 ) .
two among the brightest sources of the six naked " agn sample have been observed in this work ( q2130 - 431 and q2131 - 427 ) .
in particular , hawkins ( 2004 ) reported also on a previous optical observation of q2131 - 427 , in 1991 with efosc on the 3.6 m at eso , when the source was 0.7 magnitudes brighter than in 2002 .
its optical spectrum showed no sign of broad emission lines and weak h@xmath3 , basically consistent with the spectrum observed afterward in 2002 , apart from some additional flux at the blue end of the continuum . in table [ sam ]
we report the two observed naked " agn , where redshift , [ oiii]@xmath145007/h@xmath3 and @xmath17b are taken from hawkins ( 2004 ) .
[ table = opt_spec ] q2130 - 431 and q2131 - 427 were observed at the ntt telescope in la silla ( chile ) in director discretionary time ( i d : 278.b-5021 ) to obtain a quasi - simultaneous observation with the xmm-_newton _ ones .
observations for both sources have been performed on 2006 , december 20th with the eso multi - mode instrument ( emmi ) with grating # 4 and a 1-arcsec slit width .
the condition of the night was good ( i.e. , clear to photometric ) .
the exposure for each spectrum was of 1200s ( air mass 1.70 for q2130 - 431 and 1.92 for q2131 - 427 ) .
the wavelength resolution measured from unblended arc lines was 3 , in the range 6000 - 9000 .
a standard reduction process was applied using midas and iraf tasks .
the raw data were bias subtracted and corrected for pixel - to - pixel variations ( flat - field ) .
object spectra were extracted and sky subtracted .
wavelength calibrations were carried out by comparison with exposures of he and ar lamps , with an accuracy @xmath4 1 .
relative flux calibration was carried out by observations of the spectrophotometric standard star sa95_42 ( oke 1990 ) .
we estimate an error on the flux calibration @xmath4 10% from the standard adjustment during the calibration procedure .
statistical errors were propagated throughout the data reduction process .
the spectral fitting was carried out using the qdp fitting routines via @xmath18 minimisation .
the spectrum has been fitted with a linear component in order to reproduce the continuum which is always almost flat .
the line intensities and widths were measured by fitting gaussian profiles and the full width half maximum ( fwhm ) calculated taking into account the spectrograph resolution .
the nlr is stratified as it is the blr , but emission lines coming from the same region ( like the doublets and the balmer lines ) should display similar widths . for this reason , velocity widths for narrow lines were tied together to the same value for each doublet as well as for the narrow components ( and broad , if any ) of the balmer lines .
the results from the spectral fitting of the two objects are shown in table [ table = opt_spec ] and the optical fitted spectra are shown in figures [ opt1 ] and [ opt2 ] . in the case of q2130 - 431 a broad component with a fwhm
@xmath19 9700 km s@xmath11 is clearly detected in the h@xmath2 and h@xmath3 emission lines . according to the whittle ( 1992 ) classification criteria for intermediate type seyfert galaxies ( based on the f@xmath1}$]/f@xmath20 ratio ) ,
q2130 - 431 should be classified as seyfert 1.5 .
however , f@xmath1}$]/f@xmath20 = 1.0 is close to the boundary between the seyfert 1.5 and seyfert 1.8 classification ; since the seyfert 1.5 classification is mainly used for sources with highly variable spectra , we adopt a seyfert 1.8 classification for q2130 - 431 .
conversely , no broad components ( h@xmath2 nor h@xmath3 ) are significantly detected in the spectra of q2131 - 427 .
we compute an upper limit on the h@xmath2 and h@xmath3 broad emission line components by fixing the line widths to 4000 km s@xmath11 , a conservative value for blr widths ( ho et al .
we are confident that in this source the lack of detection of broad line components is not due to the host galaxy contamination .
indeed , our spectra are quite flat and do not show any significant contribution from the host galaxy .
since the data are taken in long - slit mode , we are able to extract only the most central part of the source , thus avoiding as much as possible the host galaxy contribution .
we calculate the fractional contribution of the h@xmath2 broad component to the total ( broad@xmath21narrow ) h@xmath2 emission by using the estimated upper limit .
this corresponds to at most 30 per cent , a low value if compared to the average value in the ho et al .
( 1997 ) seyfert sample which is around 65 per cent .
note that the latter is a conservative estimate since the average fwhm for broad line components in ho et al .
( 1997 ) is around 2700 km s@xmath11 while our upper limit has been calculated for a fwhm of 4000 km s@xmath11 . in the classical [ oiii]/h@xmath3 vs. [ nii]/h@xmath2 diagnostic diagram plot used to classify narrow emission line galaxies ( kewley et al .
2006 ) , both sources are placed in the seyfert region , well above the liner - starburst limits . in conclusion ,
the optical classification for q2130 - 431 and q2131 - 427 is seyfert 1.8 and seyfert 2 , respectively .
it is interesting to compare our spectral results with those obtained from hawkins ( 2004 ) and here reported within parenthesis in table [ table = opt_spec ] .
the [ oiii ] and h@xmath3 fhwm measured for q2130 - 431 are consistent with the ones reported in hawkins ( 2004 ) , after spectral resolution correction , while our measured [ oiii]/h@xmath3 ratio of @xmath19 11 is smaller compared to @xmath19 16 ( see table [ sam ] ) . in the optical spectrum of q2131 - 427 the major difference results in the h@xmath3 fwhm significantly smaller than the one measured by hawkins ( 2004 ) ( @xmath19 830 km s@xmath11 ) . on the other hand ,
the [ oiii]/h@xmath3 ratio @xmath19 9 and the [ oiii ] width are consistent within errors .
we note that , this comparison exercise should be taken with caution , since the quality of the spectroscopy observations in hawkins ( 2004 ) , performed with 2df multi - fibre spectrograph , can not provide accurate error estimates for the fwhm and could also be problematic when calculating line ratios , given the unreliability of sky subtraction and flux calibration .
[ table = obs_spec ] q2130 - 431 and q2131 - 427 were observed by xmm-_newton _ on november 2006 ( obsid : 0402460201 ) .
data have been processed starting from the observation data fits files with sas 7.0.0 .
x - ray events corresponding to patterns 0 - 12 and 0 - 4 were selected from the epic - mos and epic - pn cameras , respectively .
we used the most updated calibration files available at the time of the reduction for each source data .
source light curves and spectra were extracted from circular regions of 25@xmath22 centered on the source , while background products were obtained from off - set regions close to the source .
exposure times have been filtered out for periods of high background and the good time interval " exposures are reported in table [ sam ] as well as the observation date and the instruments filters .
spectra were binned according to the counts collected in each source .
the ancillary and detector response matrices were generated using the xmm-_newton _ sas arfgen and rmfgen tasks .
the x - ray light - curves have been examined and no significant variability has been found .
the pn , mos1 and mos2 spectra were analyzed using xspec v.12.4.0 . galactic absorption
is implicitly included in all spectral models .
abundances are those of anders & grevesse ( 1989 ) . the errors , lower and upper limits quoted correspond to 90% confidence range for one interesting parameter ( i.e. , @xmath23 ; avni 1976 ) , unless otherwise stated .
the x - ray spectrum of q2130 - 431 is well represented by a single power - law with photon spectral index @xmath24 ( @xmath25 = 150/144 ) .
no extra absorption apart from the galactic one is needed ( @xmath26 @xmath4 2 @xmath5 @xmath6 @xmath7 ) .
q2131 - 427 shows a very weak spectrum , modeled by a single power - law ( @xmath27 @xmath19 2 ) and an upper limit ( at 90% confidence level ) on the intrinsic absorption of @xmath26 @xmath4 9 @xmath5 @xmath6 @xmath7 .
we fixed a narrow gaussian emission line at the rest energy of the fek@xmath2 line , 6.4 kev , and we measured an upper limit on the equivalent width of the line of @xmath19 2 kev .
however , this large value is probably due to the significant residuals visible above 5 kev in the pn spectrum ; actually the very low photon statistics around 6 - 7 kev and the background dominance in this range prevent us from drawing any strong conclusions based on this limit . in figures
[ q2130 ] and [ q2131 ] best fit spectra are shown on the left panels , while column density versus photon index contour plots are on the right panels . in table
[ table = obs_spec ] we show the best fit parameters together with the model fluxes and luminosities in the 0.5 - 2 kev and 2 - 10 kev bands .
we also report the black hole mass derived from the @xmath28 @xmath29 @xmath30 relation ( tremaine et al .
2002 ) , where @xmath30 is the stellar velocity dispersion obtained from the fwhm([oiii ] ) as shown in greene & ho ( 2005 ) .
the eddington ratio has been calculated assuming that @xmath31 @xmath32 20 ( elvis et al .
1994 , vasudevan & fabian 2009 ) .
the naked " agn as discovered by hawkins ( 2004 ) are among the best candidates to look for true " unabsorbed seyfert 2 galaxies ( panessa & bassani 2002 ) .
indeed , the monitored optical variability over 25 years points to a nucleus seen directly where , given the absence of broad optical emission lines in the optical spectra , the blr may lack .
as successfully proven for ngc3147 ( bianchi et al .
2008 ) , the key strategy to spot this class of agn is through simultaneous optical and x - ray observations in order to rule out variability as the cause of the optical and x - ray mismatch .
we have therefore observed two naked " agn candidates , q2130 - 431 and q2131 - 427 , simultaneously in the x - ray and optical band . for q2131 - 427 ,
the photometric measurements obtained over @xmath19 25 years , as reported in hawkins ( 2004 ) , show a slow decrease of the emission , from b @xmath19 20 mag in 1974 to b @xmath19 21 mag in 2002 .
the optical monitor on board xmm-_newton _ ( mason et al .
2001 ) measured b = 20.83 @xmath33 0.10 , only slightly above the last measurement in the b@xmath34 light curve ( in 2001 ) b@xmath34 = 21.2 @xmath33 0.3 , confirming that the source is still in a low flux state .
interestingly , q2131 - 427 shows no signs of broad emission line components in its optical spectrum and the faint x - ray spectrum of this source is unabsorbed , with an upper limit on the intrinsic column density of @xmath26 @xmath4 9 @xmath5 @xmath6 @xmath7 , consistent with that from a host galaxy .
this picture is hard to reconcile within the unified model predictions , where the blr is obscured by a toroidal absorbing medium of dust and gas , the former responsible of the optical extinction of the broad emission lines and the latter measurable in x - rays .
one possibility is that the blr is hidden behind a compton thick medium whose column density can not be measured in x - ray spectra below 10 kev ( matt et al .
the upper limit on the ew of the fek@xmath2 line at 6.4 kev does not rule out the compton thick hypothesis , however this measurement is not reliable , given the low statistics spectrum above 5 kev . instead , the f@xmath35/f@xmath1}$ ] ratio}$ ] has been reddening corrected ( bassani et al .
1999 ) . ] of @xmath19 6 is typical of compton thin sources .
tran ( 2001 ) studied the properties of type 2 seyfert galaxies with polarized ( hidden ) broad - line regions ( hblr ) seen in reflected light compared to the non - hblr seyfert 2 galaxies , introducing the use of a diagnostic diagram involving [ oiii]@xmath145007/h@xmath3 .
the median values of [ oiii]@xmath145007/h@xmath3 found are 6.8 @xmath33 1.5 and 9.9 @xmath33 1.3 for the non - hblr and hblr seyfert 2 galaxies respectively .
the [ oiii]@xmath145007/h@xmath3 ratio of q2131 - 427 is @xmath19 9 , suggesting that this source might harbor a hidden blr . in this hypothesis
, some sort of dusty medium should obscure the optical broad emission lines and be gas - free .
as already invoked for the unabsorbed seyfert 2 in panessa & bassani ( 2002 ) , a dust - to - gas ratio 10 - 50 times higher than galactic would be enough to obscure the blr .
however , this is an unlikely scenario as a dust obscuration is often observed to be lower than the associated gas column density ( see maiolino et al .
moreover , the balmer decrement of the nlr , i.e. ( h@xmath2/h@xmath3)@xmath36 @xmath19 4.2 , suggests that the environment is not particularly dusty ( ward et al . 1987 ) and the optical extinction a@xmath37 @xmath19 1.3 mag derived from the balmer decrement is still too small to be able to completely obscure a standard blr .
the hypothesis of a pc and/or sub - pc scale clumpy obscuring medium which results in a variable column density in x - ray spectra ( risaliti et al .
2002 , elvis et al .
2004 , risaliti et al .
2005 , bianchi et al .
2009 ) can not help to explain the complete absence of broad components in the optical spectrum . under the assumption that the x - ray absorption originates from the blr clouds ( see risaliti et al .
2009 ) , we should expect variable column density in the x - ray spectra but still broad emission lines in the optical spectrum , indeed at these spatial scales the gas should be dust - free being within the sublimation radius .
the most likely explanation for the observed optical / x - ray discrepancy is to assume that in this source the blr does not exist or its emission is gradually declining .
indeed , the well documented optical flux decrease may favor the hypothesis that the blr fade away in response to a decrease of the continuum emission .
recent models propose a dependence of the blr properties with luminosity and/or accretion rate ( wang & zhang 2007 , elitzur & shlosman 2006 , nicastro et al .
2003 , nicastro 2000 , williams et al .
according to elitzur & shlosman ( 2006 ) , the classical molecular torus is just the outer part of a clumpy wind that comes out off the accretion disk . at
bolometric luminosities lower than @xmath19 @xmath38 ergs sec@xmath11 , the mass accretion can no longer sustain the cloud outflow rate causing the disappearance of the torus and , at lower luminosities , of the broad line clouds too . according to the above scheme , the absence of a blr in q2131 - 427 is not justified given the high bolometric luminosity ( l@xmath39 @xmath19 8 @xmath5 10@xmath40 ergs sec@xmath11 ) ; however the low eddington ratio measured ( l@xmath41/l@xmath42 = 0.0024 ) may favor the scenario in which the blr is formed in a vertical disk wind originating at a critical distance in the accretion disk , at accretion rates higher than a minimum value ( @xmath43 , nicastro 2000 ) . according to this model , below the critical accretion rate the blr can not be formed . in the case of q2131 - 427 , this is one of the best hypothesis to interpret the observational mismatch , taking into account that the eddington ratio is around this critical value and it is probably affected by uncertainties related to the determination of the black hole mass and of the bolometric correction .
if the blr properties actually depend on the luminosity and/or accretion physics , this must have strong implications for agn unification models and also for population studies of obscured and unobscured agn samples ( e.g. fraction of type 2 objects versus luminosity ) as discussed in elitzur & shlosman ( 2006 ) ( see also wang & zhang ( 2007 ) for an evolutionary sequence of seyfert galaxies ) . the q2130 - 431 optical spectrum shown here
reveals that this source is a type 1.8 seyfert galaxy , i.e. we detect h@xmath2 and h@xmath3 broad line components ( see figure [ opt1 ] ) .
the h@xmath2 broad component was not detected by hawkins ( 2004 ) since , at the redshift of the source , the line is shifted out of their optical spectra . also the h@xmath3 broad component was not detected , maybe because it is difficult to identify given its width , being hidden under a galaxy continuum ( see figure 6 in hawkins 2004 ) .
alternatively , it may be variable .
note that the fwhm of the h@xmath3 narrow line component has not varied significantly between 2002 ( @xmath19 426 km s@xmath11 ) and 2006 ( 320@xmath44 km s@xmath11 ) observations .
the simultaneous ultraviolet imaging with the optical monitor , allows us to detect the source in the uvw1 filter ( 245 - 320 nm ) , with a galactic reddening corrected magnitude of @xmath45 , slightly brighter when compared to the last measurement of @xmath46 mag , as shown in the 25 years light curve ( see fig . 8 , hawkins 2004 ) . in x - rays ,
the 2 - 10 kev luminosity of @xmath19 4 @xmath5 10@xmath47 ergs s@xmath11 suggests that this is a typical type 1 seyfert galaxy with a standard photon index ( @xmath27 = 1.8 - 2 ; e.g. , piconcelli et al . 2005 ) and no intrinsic absorption ( @xmath4 2 @xmath5 @xmath6 @xmath7 , at 90% confidence level ) .
the f@xmath35/f@xmath1}$ ] ratio is @xmath19 77 , well within the regime of compton thin objects ( cappi et al .
2006 , bassani et al .
our x - ray flux is in agreement with the _ chandra _ one found by gliozzi et al .
( 2007 ) suggesting no particular variability in x - rays .
according to the above observational evidence , q2130 - 431 should not be considered as a pure naked " agn , in the sense that it displays optical broad emission line components and , therefore , it has a blr .
however , intermediate type seyfert galaxies are characterized by large balmer decrements consistent with reddening of the blr and determining the 1.8/1.9 spectroscopic type ( osterbrock 1981 ) .
the weakness of the broad balmer lines is often ascribed to dust absorption ( maiolino & rieke 1995 ) and , indeed , moderate absorption is observed in x - rays ( risaliti et al .
1999 ) . in q2130 - 431
, the optical blr balmer decrement of 3.4 ( almost the same value of the nlr balmer decrement ) suggests that the blr does not suffer from heavy reddening .
in addition , in the x - ray spectrum , no absorption has been measured , strengthening the idea that the blr is not obscured and that the weakness of the broad optical lines can be intrinsic , e.g. due to a small amount of gas in the clouds .
a weak blr can be due to a luminosity / accretion rate related disappearance , as suggested for q2131 - 427 .
interestingly the existence of this kind of objects should be hypothesized within an evolutionary sequence , as the one drawn in elitzur & shlosman ( 2006 ) , from type 1 objects to true " type 2s .
however , the high x - ray luminosity and the eddington ratio ( l@xmath39/l@xmath48 @xmath19 0.37 ) disfavor the latter hypothesis .
alternatively , the lack of a prominent blr in q2130 - 431 may simply be a peculiar intrinsic property of this object . in both scenarios ,
q2130 - 431 can represent a case of true " intermediate seyfert galaxy , in which the blr is intrinsically weak instead of being obscured by dust or gas .
we have obtained quasi - simultaneous x - ray ( xmm - newton ) and optical ( ntt - emmi ) spectra of two seyfert galaxies , q2130 - 431 and q2131 - 427 , belonging to the class of naked " agn characterized by hawkins ( 2004 ) and , currently , composed of six members . the strong optical brightness variability and the lack of broad emission line components in their spectra
have suggested that in these agn the nucleus is seen directly and it is taken off its blr .
our quasi - simultaneous observations have proven fundamental to confirm the naked " hypothesis in the case of q2131 - 427 , for which we have ruled out the compton thick nature based on the low f@xmath35/f@xmath1}$ ] ratio . moreover
the balmer decrement in the nlr excludes a high dust - to - gas ratio as a possible explanation of the broad emission lines absence . instead
, the present observations have confirmed that no broad components are found associated to the optical spectrum emission lines and simultaneously no absorption is measured in the x - ray spectrum .
either a dusty medium is absorbing the blr being adequately gas free to let the x - ray continuum escape , however this is an unlikely scenario , or the source is accreting below a critical accretion rate such that the blr can not be formed ( nicastro 2000 ) .
the naked " hypothesis is instead discarded for q2130 - 431 , which indeed displays h@xmath2 and h@xmath3 broad line components in its optical spectrum . however , no sign of mild absorption in its x - ray continuum is found , contrary to what expected for intermediate seyfert galaxies ( in order to justify the partial obscuration of the blr ) , neither the blr itself seems to suffer from heavy reddening .
the presence of a small amount of gas to photoionize can be the reason for the intermediate optical spectrum , making it a case of a potential true " intermediate seyfert galaxy . at last , in the appendix
, we present the results on the cluster abell 3783 serendipitously detected in the q2131 - 427 epic fov providing a description of its fundamental parameters .
we thank the anonymous referee for the useful comments . f.p .
acknowledges support by asi - inaf i/08/07/0 grants .
fjc and xb acknowledge support by the spanish ministry of science and innovation , under grants esp2006 - 13608-c02 - 01 . this work benefited from an italian - spanish integrated action ( hi 2006 - 0079 ) .
thanks m.t .
ceballos for the technical support .
humphrey is thanked for the use of his surface brightness fitting code .
this research has made use of data obtained from the high energy astrophysics science archive research center ( heasarc ) , provided by nasa s goddard space flight center .
this research has also made use of the nasa / ipac extragalactic database ( _ ned _ ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with nasa .
99 s. w. , fabian a. c. , 1998 , , 297 , l57 r. , 1993 , , 31 , 473 k. a. , 1996 , in asp conf .
ser . 101 : astronomical data analysis software and systems v , vol . 5 , 17 h. , koyama k. , inoue h. , halpern j. p. , 1991
, , 43 , 195 bassani l. , dadina m. , maiolino r. , salvati m. , risaliti g. , della ceca r. , matt g. , zamorani g. , 1999 , apjs , 121 , 473 x. , carrera f. j. , ceballos m. t. , 2003 , , 339 , 757 m. , steinmetz m. , 1996 , , 283 , 431 bianchi , s. , piconcelli , e. , chiaberge , m. , bailn , e. j. , matt , g. , & fiore , f. 2009 , , 695 , 781 bianchi s. , corral a. , panessa f. , barcons x. , matt g. , bassani l. , carrera f. j. , jimnez - bailn e. , 2008 , mnras , 385 , 195 s. , guainazzi m. , matt g. , fonseca bonilla n. , 2007 , , 467 , l19 s. , matt g. , 2002 , , 387 , 76 s. , matt g. , nicastro f. , porquet d. , dubau j. , 2005 , , 357 , 599 brightman m. , nandra k. , 2008 , mnras , 390 , 1241 g. l. , norman m. l. , 1998 , , 495 , 80 caccianiga a. , et al . , 2004 , a&a , 416 , 901 cappi m. , et al . , 2006 , a&a , 446 , 459 a. , fusco - femiano r. , 1976 , , 49 , 137 a. , barcons x. , carrera f. j. , ceballos m. t. , mateos s. , 2005 , , 431 , 97 s. , molendi s. , 2001 , , 551 , 153 j. m. , lockman f. j. , 1990 , , 28 , 215 x. y. , de robertis m. m. , 2006 , , 131 , 1236 m. , shlosman i. , 2006 , , 648 , l101 m. , risaliti g. , nicastro f. , miller j. m. , fiore f. , puccetti s. , 2004 , , 615 , l25
m. et al . , 1994 , , 95 , 1 s. , 2000 , , 311 , 313 s. , balestra i. , 2009 , , 496 , 343 fiore f. , et al . , 2001 ,
mnras , 327 , 771 f. et al . , 2007 , , 669 , 158 f. et al . , 2008 , , 673 , 176 m. , sambruna r. m. , foschini l. , 2007 , , 662 , 878 greene j. e. , ho l. c. , 2005 , apj , 627 , 721 n. , sauval a. j. , 1998 , space science reviews , 85 , 161 q. , huang j. , 2002 , , 579 , 205 m. , fabian a. c. , iwasawa k. , matt g. , fiore f. , 2005 , , 356 , 295 m. , 2008 , xmm - soc - cal - tn-0018 issue 2.7 m. r. s. , 2004 , , 424 , 519 l. c. , filippenko a. v. , sargent w. l. , 1995 , , 98 , 477 l. c. , filippenko a. v. , sargent w. l. w. , peng c. y. , 1997 , , 112 , 391 k. d. , snowden s. l. , 2008 , , 478 , 575 kewley , l. j. , groves , b. , kauffmann , g. , & heckman , t. 2006 , , 372 , 961 lin , y .- t . , & mohr , j. j , 2004 , , 617 , 879 r. , marconi a. , salvati m. , risaliti g. , severgnini p. , oliva e. , la franca f. , vanzi l. , 2001 , , 365 , 28 maiolino , r. , & rieke , g. h. 1995 , , 454 , 95 m. , 1998 , , 504 , 27 k. o. et al . , 2001 , , 365 , l36 mateos s. , barcons x. , carrera f. j. , ceballos m. t. , hasinger g. , lehmann i. , fabian a. c. , streblyanska a. , 2005 , a&a , 444 , 79 s. et al . 2009 , , 496 , 789 g. , guainazzi m. , maiolino r. , 2003 , , 342 , 422 matt g. , pompilio f. , la franca f. , 1999 , newa , 4 , 191 f. , 2000 , , 530 , l65 f. , martocchia a. , matt g. , 2003 , , 589 , l13 oke , j. b. 1990 , , 99 , 1621 osterbrock d. e. , 1981 , apj , 249 ,
462 f. , bassani l. , 2002 , , 394 , 435 a. , georgantopoulos i. , stewart g. c. , zezas a. l. , 2001 , , 326 , 995 peterson b. m. , pogge r. w. , wanders i. , smith s. m. , romanishin w. , 1995 , pasp , 107 , 579 e. , jimenez - bailn e. , guainazzi m. , schartel n. , rodrguez - pascual p. m. , santos - lle m. , 2005 , , 432 , 15 risaliti g. , et al . , 2009 , mnras , 393 , l1 risaliti g. , elvis m. , fabbiano g. , baldi a. , zezas a. , salvati m. , 2007 , apj , 659 , l111
g. , elvis m. , fabbiano g. , baldi a. , zezas a. , 2005 , , 623 , l93 risaliti g. , elvis m. , 2004 , assl , 308 , 187 risaliti g. , elvis m. , nicastro f. , 2002 , apj , 571 , 234 risaliti , g. , maiolino , r. , & salvati , m. 1999 , , 522 , 157 shuder j. m. , osterbrock d. e. , 1981 , apj , 250 , 55 spergel , d. n. , et al .
2003 , , 148 , 175 tran h. d. , 2001 , apj , 554 , l19 tremaine s. , et al . , 2002 ,
apj , 574 , 740 vasudevan , r. v. , & fabian , a. c. 2009 , , 392 , 1124 a. et al . , 1999 , , 525 , 47 a. et al . , 2006 , , 640 , 691 w. et al . , 1999 , , 349 , 389 wang , j .-
, & zhang , e .- p . 2007 , , 660 , 1072 ward m. , elvis m. , fabbiano g. , carleton n. p. , willner s. p. , lawrence a. , 1987 , apj , 315 , 74 west , r. m. , & fradsen , s. 1981 , , 44 , 329 whittle , m. 1992 , , 79 , 49 a. , gioia i. m. , henry j. p. , mullis c. r. , 2005 , , 444 , 165
an extended source , of a @xmath49 arc - minute scale , in the field of q2131 - 427 is clearly detected at 10.2 off - axis ( see fig.[fig.1 ] ) .
the x - ray peak of of the source in equatorial coordinates is and it is associated with the cluster of galaxies ( see the overlay of the x - ray contours on the dss image in fig.[fig.1 ] ) at the redshift of @xmath50 ( west & fradsen 1981 ) . with the caveat of the psf at this off - axis angle (
@xmath51 is the 50% encircled energy fraction radius at 1.5 kev ) the peak is consistent with being associated with the brightest galaxy ( bcg ) of the cluster , 2masx j21335944 - 4238448 , rather than the likely interloper 2masx j21340071 - 4238528 with position at a redshift @xmath52 .
these are the only galaxies with redshifts available listed in
_ ned _ within 3 of the source .
considerations of the redshift derived from the fe k@xmath53 complex in the x - ray spectrum , the luminosity and the temperature of the cluster give further strength to the association . in the following
we will therefore assume @xmath50 as the redshift of the source for which 1 corresponds to 194 kpc .
all the errors quoted in this appendix are at the 68% confidence limit .
the data reduction and preparation follow the procedure of gastaldello et al .
( 2008 ) where more details can be found ; here we provide just a brief description .
the data were re - processed with sas v8.0
. for each detector we created images in the 0.5 - 2 kev band with point sources masked using circular regions of 25 radius centered at the source position .
the images have been exposure corrected and a radial surface brightness profile was extracted from a circular region of 7 of radius centered on the cluster centroid .
we account for the x - ray background in the surface brightness analysis by including a constant - background component .
the data were grouped to have at least 20 counts per bin in order to apply the @xmath18 statistic .
the fitted model is convolved with the psf .
the joint best - fit @xmath54-model ( cavaliere & fusco femiano 1976 ) has a core radius of @xmath55 kpc ( @xmath56@xmath57 ) and @xmath58 for a @xmath18/d.o.f .
= 210/141 ( see fig.[fig.2 ] )
. fits to the profiles of the individual detectors give consistent results within 1@xmath59 of the combined - fit result and in the case of the mos2 detector is formally acceptable ( 49/47 ) .
the main contribution to the @xmath18 comes mainly from the pn ( @xmath18/d.o.f .
= 102/61 for the fit to the individual profile ) and its origin is instrumental ( ccd gaps and pixels under - sampling the psf ) , whereas a large area of the radial profile at @xmath60 is lost for mos1 ( @xmath18/d.o.f .
= 55/29 ) due to the missing ccd-6 , hence there is no need for more complicated models . for spectral fitting , we extracted spectra for each detector from a 2.5region centered on the peak of the emission , to maximize the s / n over the background . redistribution matrix files ( rmfs ) and ancillary response files ( arfs ) were generated using the sas tasks _ rmfgen _ and _ arfgen _ in extended source mode with appropriate flux weighting .
the background was estimated locally using spectra extracted from a source free region of the same extent at the same off - axis angle ( avoiding ccd-4 for mos1 and ccd-5 for mos 2 because in this observation they display anomalous background for energies below 1 kev , see kuntz & snowden 2008 ) .
the spectra from the three detectors were re - binned to ensure a signal - to - noise ratio of at least 3 and a minimum 20 counts per bin and they were jointly fitted with an apec thermal plasma modified by galactic absorption ( dickey & lockman 1990 ) .
the spectral fitting was performed in the 0.5 - 7 kev band .
the spectra are shown in fig.[fig.3 ] : the best fit parameters are @xmath61 kev and @xmath62 for a @xmath18/d.o.f .
= 418/396 .
the source photons correspond to about 78% of the total events ( @xmath63 counts in each mos and 5700 in the pn ) .
if we leave the redshift parameter free we obtain @xmath64 , in good agreement with the redshift determined optically : the quality of the data is such to tightly constrain the redshift of the cluster ; in fig.[fig.4 ] we show the temperature - redshift and redshift - metal abundance @xmath18 confidence contours .
a likely previous x - ray identification is the source 1rxs j213401 - 423833 in the all - sky bright source catalogue ( voges et al .
1999 ) and associated with .
the count - rate of @xmath65 ct / s which corresponds to an unabsorbed flux in the 0.1 - 2.4 kev band of @xmath66 ( for a 5 extraction radius ) is in good agreement with our determination of @xmath67 in the 0.1 - 2.4 kev band ( according to the derived surface brightness model the 2.5 - 5 region contributes only @xmath68% of the flux within a 5 region ) . to investigate possible spatial variation in the spectral parameters of the cluster
, we extracted two annular regions of radii 0 - 1 and 1 - 3 .
the derived spectral parameters are : @xmath69 kev and @xmath70 with @xmath18/d.o.f .
= 226/177 for the inner annulus ; @xmath71 kev and @xmath72 with @xmath18/d.o.f .
= 313/314 for the outer annulus .
the width of the bins have been chosen in order to avoid bias in the temperature measurement caused by scattered flux by the psf .
it is difficult , given the off - axis position of the cluster , to asses the existence of gradients in the temperature : given the quality of the data the cluster is consistent with being isothermal over the explored radial range .
the abundance gradient , signature of relaxed cool - core cluster ( e.g. , de grandi & molendi 2001 ) , is on the contrary significant at the 2.6@xmath59 level . under the assumption of isothermality and assuming that the cluster follows the best fit model to the surface brightness profile derived above we can calculate the total mass profile using the best - fit @xmath54-model for which the gas density and total mass profiles can be expressed by simple analytical formula ( e.g. , ettori 2000 ) .
we evaluated @xmath73 as the radius at which the density is 500 times the the critical density and the virial radius as the radius at which the density corresponds to @xmath74 , as obtained by bryan & norman ( 1998 ) where @xmath75 , @xmath76 and @xmath77^{1/2}$ ] ] for the concordance cosmological model used in this paper . to evaluate the errors on the estimated quantities we used the same procedure as above repeating the measurements for 10000 random selections drawn from gaussian distributions for the temperature and parameters of the surface brightness profile .
for @xmath78 we obtained , @xmath79 within @xmath80 kpc ; the virial mass is , @xmath81 , within the virial radius @xmath82 kpc .
if the identification of 2masx j21335944 - 4238448 with the bcg is correct , the optical luminosity calculated in the @xmath83 following lin & mohr 2004 is @xmath84 corresponding to @xmath85 , consistent with the central galaxy luminosity - host halo mass relation ( and its scatter ) derived in the literature ( e.g. , lin & mohr 2004 ) .
we calculated the gas mass to be @xmath86 .
for , the aperture of 485 kpc used for spectroscopy encloses 74% of the flux within @xmath73 .
the derived bolometric luminosity within @xmath73 is @xmath87 : lies very close to the - relation derived from the cluster sample of objects with @xmath88 kev of markevitch ( 1998 ) , though given the angular resolution of the data it has not been possible to correct for the effect of an eventual presence of a cool core ( allen & fabian 1998 ; markevitch 1998 ) . finally , as done routinely in our previous analysis
( e.g. , gastaldello et al .
2007 , 2008 ) we have studied the sensitivity of our results to various systematic uncertainties and data analysis choices that may impact our results , like varying or using a different energy band ( 1 - 7 kev ) or standard blank background fields for background subtraction , finding always agreement within the errors with the results reported above . below we summarize two possible sources of errors not previously explored in our work .
_ cross - calibration : _ improvements have been made in the status of the cross - calibration between the epic cameras in the recent sas releases , but some calibration differences are still present which can affect the spectral parameters of an extended source positioned off - axis like in the present observation . in the 2.5 aperture
the fit to the individual mos1 data returns @xmath89 , @xmath90 and @xmath91 ( @xmath18/d.o.f .
= 111/105 ) ; for mos 2 @xmath92 , @xmath93 and @xmath94 ( @xmath18/d.o.f .
= 96/114 ) ; for pn @xmath95 , @xmath96 and @xmath97(@xmath18/d.o.f .
= 176/186 ) .
the above results correspond to an unabsorbed flux in the 0.5 - 2 kev band of @xmath98 and @xmath99 for mos1 , @xmath100 and @xmath101 for mos2 , @xmath102 and @xmath103 for pn .
the results are within 1@xmath59 of the best joint fit and the overall trend of higher fluxes of the mos compared to the pn is consistent with the current calibration uncertainties ( guainazzi 2008 , mateos et al .
2009 ) .
_ radial range used in the surface brightness fitting : _ simulations and simple analytic models pointed out how the @xmath54 model overestimates gas mass because it returns a biased low @xmath54 due to the restricted range of radii where the fit is performed ( e.g. bartelmann & steinmetz 1996 ; roncarelli et al .
indeed recent analyses investigating surface brightness profiles of clusters with and find evidence for a steepening of the gas density slope with radius for clusters ( vikhlinin et al .
1999 , 2006 ; ettori & balestra 2009 ) .
we investigated fitting the surface brightness profile in a narrower range ( 5 ) and wider range ( 9 ) compared to the choice made in the analysis .
using the narrower range we obtain a core radius of @xmath104 kpc ( @xmath105@xmath106 ) and @xmath107 for a @xmath18/d.o.f .
= 149/129 and using the wider range we obtain @xmath108 kpc ( @xmath109@xmath106 ) and @xmath110 for a @xmath18/d.o.f .
= 338/156 .
the use of a radial range as large as possible is clearly preferred in the determination of the @xmath54 model parameters , but we did not adopt the fit obtained in the 9 aperture because of the large residuals due to a not optimal fitting of the background level .
this is due to the fact that , because of the off - axis position of the source , the increased radial range is obtained with only a partial azimuthal coverage due to the increasing out of field of view area ( only @xmath111% of the 6 - 9annulus is covered by data in the mos1 and pn detectors ) . had we used the parameters obtained in the narrower range we would have derived @xmath112 within @xmath113 kpc and @xmath114 , within the virial radius @xmath115 kpc whereas using the wider range we would have derived @xmath116 within @xmath117 kpc and @xmath118 , within the virial radius @xmath119 kpc . | hawkins ( 2004 ) reported on a class of naked " agn characterized by strong amplitude optical brightness variability and the complete absence of broad emission lines in the optical spectrum .
the variability suggests that the nucleus is seen directly , however the absence of broad lines contradicts the simple formulation of unified models for agn .
we present the results of quasi - simultaneous spectroscopic observations with xmm-_newton _ and ntt ( _ la silla _ ) of two naked " agn .
we confirm the naked " nature of q2131 - 427 for which no broad emission line components have been detected in the optical spectrum and its x - ray spectrum shows no signs of intrinsic absorption .
the optical and x - ray mismatch in this source can not be ascribed to a high nuclear dust - to - gas ratio and a compton thick nature is ruled out on the basis of the high f@xmath0/f@xmath1}$ ] ratio .
the broad line region ( blr ) may be completely absent in this source , possibly as a consequence of its low eddington ratio . on the other hand ,
the optical spectrum of q2130 - 431 shows h@xmath2 and h@xmath3 broad emission line components , revealing the presence of a blr .
a mild x - ray absorption is expected in intermediate type 1.8 seyfert galaxies like q2130 - 431 , however we put a very low upper limit on the column density ( @xmath4 2 @xmath5 @xmath6 @xmath7 ) , also the low balmer decrement suggests that the blr itself does not suffer from reddening .
we propose that in this object the blr is intrinsically weak , making a case of true " intermediate seyfert galaxy .
we also report on the x - ray detection of the abell 3783 galaxy cluster in the xmm-_newton _ field - of - view of the q2131 - 427 observation .
[ firstpage ] galaxies : active - galaxies : seyfert - x - rays : galaxies - galaxies : clusters : general |
stone - wales defect @xcite @xmath1(55 - 77 ) is the simplest example of topological disorder in graphene and other @xmath2-hybridized carbon systems .
it can be formed by rotating a c - c bond by @xmath3 with regard to the midpoint of the bond referred to as the sw transformation so that four hexagons are turned into two pentagons and two heptagons .
this defect has received considerable amount of attention @xcite , because it has the lowest formation energy among all intrinsic defects in graphenic systems , and because it presumably plays an important role in plastic deformation of carbon nanotubes ( cnt ) under tension @xcite by relieving strain @xcite .
it can also act as a source for dislocation dipoles @xcite . regardless of being the lowest energy defect in graphene @xcite and other @xmath2-hybridized carbon nanostructures ,
the @xmath1(55 - 77 ) needs about 5 ev to appear in graphene @xcite , and 35 ev in cnts with a diameter above 1 nm @xcite , which should lead to a negligible equilibrium concentration of such defects at room temperature .
however , recent advances in hrtem have allowed the identification of such defects in graphene @xcite and cnts @xcite . moreover
, sw transformations play an important role in the response of graphene to electron irradiation @xcite , leading to changes in the morphology of vacancy - type defects @xcite and to their migration .
such changes are equally surprising , because the barrier for bond rotation is about 5 ev @xcite , which should exclude thermal activation as a cause for sw transformation at room temperature during experimentally relevant time scales .
regarding irradiation effects , previous simulations @xcite showed that an energy of @xmath4 ev must be transferred to a c atom in graphene in the in - plane direction for a bond rotation to occur .
also this can not explain the frequently observed sw transformations under the usual tem imaging conditions , since with typical acceleration voltages ( @xmath5 kv ) the transferred kinetic energy in the direction almost perpendicular to the electron beam will remain significantly below 10 ev . here , by combining aberration - corrected ( ac- ) hrtem with atomistic computer simulations ,
we show that topological defects associated with the sw transformation can be formed in @xmath2-hybridized carbon nanostructures by impacts of individual electrons at energies even _ below _ the threshold for a carbon atom displacement .
we further study in detail the mechanisms of irradiation - driven bond rotations for pre - existing vacancy - type defect structures and how they transform and migrate due to electron impacts . at the same time
we explain why electron irradiation at moderate energies ( @xmath6 kev ) tends to rather amorphize @xcite than perforate graphene .
we also show via simulations that the @xmath1(55 - 77 ) can appear in curved graphitic structures due to `` incomplete '' recombination of irradiation - induced frenkel defects , reminiscent of the formation of wigner - type defects in silicon @xcite .
graphene membranes used in our experiments were prepared by mechanical exfoliation of graphite on si / sio@xmath7 substrates and transfer of the resulting graphene sheets onto tem grids as described previously @xcite . for tem experiments we used an fei titan
@xmath8 equipped with an image - side aberration corrector , operated at 80 kv .
the spherical aberration was set to 15 @xmath9 m and images were recorded at scherzer defocus .
the extraction voltage of the source was reduced to 2 kv and the condensor lens c2 was switched off in order to minimize the energy spread . under these conditions ,
dark contrast in the images can be directly interpreted in terms of the atomic structure .
image sequences were recorded on a ccd camera with exposure times of 1 s and intervals of approximately 2 s. we carried out atomistic computer simulations based on the spin - polarized density functional theory ( dft ) implemented in the plane wave basis set vasp code @xcite . the projector augmented wave potentials @xcite were used to describe the core electrons , and the generalized gradient approximation of perdew , burke and ernzernhof @xcite for exchange and correlation .
we included plane waves up to a kinetic energy of 300 ev .
the @xmath10-point sampling of the brillouin zone was performed using the scheme of monkhorst - pack @xcite for the periodic dimensions .
structure relaxation calculations were combined with molecular dynamics ( dft - md ) simulations with a lower kinetic energy threshold and fewer @xmath10-points . due to the high computational cost of the dft - md method , only a few simulations were carried out at this level .
whenever statistics needed to be gathered , we calculated the forces using the non - orthogonal dft - based tight binding ( dftb ) method @xcite .
the main results were checked against dft - md . in total
, we carried out @xmath11 dynamical dftb - md simulations .
the simulated structures consisted of 120200 atoms and were fully optimized . for the displacement threshold simulations ,
one of the atoms was assigned a kinetic energy @xmath12 with the initial velocity vector pointing to a preselected direction .
the initial temperature of the system was set to 5 k , although we observed no differences when carrying out the simulations for initially stationary atoms .
displacement threshold @xmath13 ( minimum kinetic energy required to eject the atom ) was found to be 22.50 ev , in a good agreement with earlier dftb results @xcite .
it is also close to the dft value ( 22.03 ev ) @xcite . for the annihilation simulations ,
various system temperatures were studied ( 5001200 k ) both to fasten the migration of the adatoms and to understand the effect of an elevated temperature ( as will be mentioned below ) .
we begin the presentation of our results with the description of experimental observations of @xmath1(55 - 77 ) in hrtem images .
several long image sequences , typically containing hundreds of images from clean and initially defect - free graphene membranes , were recorded at 80 kv .
occasionally , @xmath1(55 - 77 ) defects appear in individual exposures , as in the example shown in figure [ px : swexp]b ( figure [ px : swexp]c with structure overlay ) .
remarkably , in most of the observed cases , isolated @xmath1(55 - 77 ) appeared in pristine graphene for one 1 s exposure , only to disappear in the following frame .
hence , the lifetime of this defect under the 80 kv electron beam in terms of irradiation dose is of the order of @xmath14 , the dose used for a single exposure .
( 55 - 77 ) ( b ) .
frame ( c ) shows the same image with a structure overlay .
@xmath1(55 - 77 ) disappears in the following frame ( d ) .
scale bar is 1 nm .
( see also video s1 in ref .
@xcite . ) ] to understand the appearance and disappearance of @xmath1(55 - 77 ) , we carried out atomistic simulations for individual displacement events under the electron beam .
after calculating @xmath13 , we extended this calculation to all in - plane ( @xmath15 ) and out - of - plane angles in the range @xmath16 $ ] , figure [ px::sect ] .
displacements with @xmath17 would result in transferred kinetic energies of more than 2 ev below @xmath13 for an electron beam even at 120 kev .
since the displacement threshold increases for increasing @xmath18 , it is unlikely that this restriction would lead us to miss any significant electron beam - induced structural changes , especially for electron energies similar to those used in this study ( 80 kev ) .
the calculated displacement thresholds are shown in figure [ px::sect ] in a relative scale along with the space angles for which we observed the formation of @xmath1(55 - 77 ) . as a function of the displacement space angle ( @xmath18 , @xmath15 ) .
crosses mark the angles for which we observed impact - induced sw transformations . ]
it is evident from figure [ px::sect ] that the sw transformation is a very likely event at displacement angles slightly away from the graphene plane normal ( @xmath19 ) .
the transferred kinetic energies ( @xmath12 ) required for this process are below the displacement threshold for the corresponding @xmath15 and @xmath18 ( @xmath20 ) since no actual removal of the recoil atom is required for the bond rotation to occur .
typically , @xmath21 ev resulted in the @xmath1(55 - 77 ) formation , although for some space angles even @xmath22 ev was enough .
the probability for @xmath1(55 - 77 ) formation is particularly high for certain space angles , which is related to different mechanisms of sw transformation , as described below .
the above - presented result is in clear contrast with the earlier simulation results for graphite @xcite where no @xmath1(55 - 77 ) formation was observed for low @xmath18 .
this discrepancy is caused by the neighboring graphene planes in the case of graphite : the displaced atom gets attached to the adjacent layer and does not therefore initiate a bond rotation . in figure [ px::swprocs ]
we show the two processes which account for the majority of the sw transformations observed during our simulations . in the `` circle '' process ( figure [ px::swprocs]a ) , the displaced atom circles around its neighbor , whereas in the `` nudge '' process ( figure [ px::swprocs]b ) it nudges the neighbor to cause the bond rotation .
note that the example cases are for the same @xmath15 and almost same @xmath12 , but for different @xmath18 .
the resulting process for each displacement is an interplay of all three variables ( @xmath12 , @xmath18 , @xmath15 ) .
similar mechanisms also exist for cnts .
however , as two new parameters ( tube diameter and chirality ) should be introduced for quantitative analysis of sw transformation , we did not study this process in nanotubes at length due to unreasonably high computational cost . ):
( a ) the `` circle '' process , and ( b ) the `` nudge '' process .
the black spheres correspond to the recoil ( displaced ) atoms , and the mono - vacancy structure is highlighted with numbered carbon rings in the second panels .
the last panel shows schematically the route of the displaced atom .
( see also videos s8 and s9 in ref .
@xcite . ) ] since we frequently observed formation of vacancy adatom pairs ( adatoms play the role of interstitials in graphene and cnts ) in our simulations of electron impacts onto graphene and earlier in cnts @xcite , we also explored another possible mechanism of @xmath1(55 - 77 ) formation , which is based on `` incomplete '' annihilation of a frenkel defect .
this study was motivated by the peculiarities of the recombination of such a defect in bulk silicon . in that covalently - bonded material
the recombination can give rise to either annihilation of the defect and restoration of the perfect crystal lattice , or to a wigner - type defect @xcite .
such topological defects are imperfections in the crystal lattice with the locally `` correct '' number of atoms ( as opposed to vacancies and interstitials ) , with the atomic configuration separated from the perfect structure by a finite potential barrier .
such defects are deemed to also exist in graphite @xcite , and @xmath1(55 - 77 ) can clearly be classified into this group .
while carbon adatoms on graphene @xcite and cnts @xcite ( especially those inside nanotubes ) are mobile at room temperature , they can easily find vacancies in the system and annihilate .
indeed , we occasionally observed disappearance of vacancies in hrtem image sequences . figure [ px::expannih ] shows an example of a mono - vacancy that disappears during observation .
this proves that mobile carbon atoms are present under our experimental conditions , and may recombine with vacancy - type defects .
however , we never noticed the creation of a @xmath1(55 - 77 ) after an observed mono - vacancy .
( 5 - 9 ) mono - vacancy , with an overlay in the inset , and ( b ) shows the same region in a later exposure , with no defect visible .
scale bar is 1 nm .
( see also video s7 in ref .
@xcite . ) ] to simulate the annihilation process , we created nearby frenkel defects ( separated by a few ngstrms ) in a graphene layer and small ( 6,6 ) zigzag and ( 10,0 ) armchair cnts ( radii @xmath23 and @xmath24 , respectively ) .
we then heated the structures and collected statistics on the evolution of each system by running dynamical atomistic simulations at various temperatures ( 5001200 k ) .
three possible outcomes emerged from the simulations : ( 1 ) perfect annihilation to the pristine structure ( similar to the experimental images in figure [ px::expannih ] ) , ( 2 ) formation of a @xmath1(55 - 77 ) ( figure [ px::swannih]a ) and , surprisingly , ( 3 ) sputtering of a c@xmath7 dimer with a remaining reconstructed di - vacancy @xmath25(5 - 8 - 5 ) ( figure [ px::swannih]b ) .
( 55 - 77 ) creation process as an outcome of a vacancy adatom annihilation in a ( 6,6 ) armchair nanotube .
( b ) carbon dimer sputtering process in a ( 10,0 ) zigzag nanotube .
the black spheres stand for adatoms and the gray ones denote the dimer to be sputtered [ marked in the first panel of ( b ) ] . the nanotube structures are shown from inside the tube .
the tube axis is in the horizontal direction .
( see also videos s10 and s11 in ref .
@xcite . ) ] for graphene , we always observed perfect annihilation in accordance with the experiments .
however , if the adatom in graphene was placed on top of one of those two bonds in the nine - membered carbon ring which are right next to the pentagon , a @xmath1(55 - 77 ) was spontaneously formed without an energy barrier .
thus @xmath1(55 - 77 ) may also form in graphene due to recombination of frenkel defects , but the probability for this process must be much lower than in cnts . for ( 10,0 ) cnt , we obtained perfect annihilation in approximately 54% of cases , @xmath1(55 - 77 ) was formed in approximately 34% of cases and dimer sputtering occurred in approximately 12% of cases . for ( 6,6 ) cnt , the values were 53% , 42% and 4% , respectively .
with increasing temperature the probability to sputter a dimer showed a slight tendency to increase .
we also ran the calculations for a ( 8,8 ) cnt ( @xmath26 ) at 800 k in order to estimate the curvature dependency of the results .
the values did not significantly differ from those for the ( 6,6 ) cnt , except for a somewhat increased tendency to perfect annihilation and decreased sputtering ( with probabilities of 56% and 2% , respectively ) .
the formation energy of a spatially separated frenkel defect in graphene within the dftb model is approximately 11.1 ev .
transformation to form sw defect from this initial setup leads to an energy gain of 5.4 ev . by sputtering a dimer
, graphene would instead gain 0.4 ev .
hence , all three observed outcomes are energetically reasonable also for graphene . for nanotubes ,
the formation energies of both @xmath1(55 - 77 ) and a di - vacancy are lowered due to the curvature and stronger c - c bonds at pentagons @xcite .
the corresponding energy gains are also higher , which can explain why the probability for @xmath1(55 - 77 ) defect formation and dimer evolution is higher in curved carbon nanostructures .
the actual energies depend on the local curvature .
it is also plausible that the dimer evaporation process plays a role in shrinking fullerenes under electron irradiation @xcite .
@xmath1(55 - 77 ) defect represents the elementary case of a topological change in the graphene structure , _
i.e. _ , a single bond rotation in the otherwise perfect structure . more abundant , however , are changes in the atomic configuration through bond rotations in the reconstructed vacancy defects , as recent experiments indicate @xcite . in presence of a ( multi-)vacancy , the atomic configuration of a defect can be transformed between different metastable structures via bond rotations .
it was presumed @xcite that such sw transformations of vacancy - type defects would be stimulated by electron impacts , but the actual atomistic mechanism has not been hitherto unraveled . in order to get microscopic insight into irradiation - stimulated bond rotations near vacancy - type defects in graphene
, we carried out a set of experiments and dedicated simulations aimed at assessing the probability of sw transformations in the defect structures . in our experiments , we initially generated `` defective '' graphene by brief 150 kv electron irradiation , and then recorded image sequences of di - vacancies using 80 kv ac - hrtem .
as observed previously @xcite , the vacancies can transform between different configurations under the influence of the 80 kv electron beam .
moreover , the di - vacancies migrate and transition between the different reconstructed configurations via sw transformations , typically until they cluster into larger defects ( see videos s5 and s6 in ref .
@xcite ) . in figure [ px::mvacexp ]
, we show an example of a di - vacancy defect that transforms between different reconstructed shapes [ @xmath25(5 - 8 - 5 ) , @xmath25(555 - 777 ) , @xmath25(5555 - 6 - 7777 ) ] under the electron beam .
the changes in the atomic structure of these di - vacancy configurations can be described by sw transformations at the defect . moreover ,
multiple transformations allow migration of the di - vacancy .
similar to @xmath1(55 - 77 ) defect formation discussed above , the activation energy for these transitions is far too high to allow a thermally activated process with an observable rate at room temperature .
hence , on the basis of the observations , the activation energy for the transition must be provided by the electron beam .
( 5555 - 6 - 7777 ) transforming into @xmath25(555 - 777 ) ( b ) , and @xmath25(5 - 8 - 5 ) ( c ) .
each of these transitions can be explained by a single bond rotation . in a later frame ( d ) , the defect is again a @xmath25(5555 - 6 - 7777 ) , but shifted by one lattice parameter .
scale bar is 1 nm .
( see also videos s24 in ref .
@xcite . ) ] in order to confirm that the sw transformations at these defects are caused by single electron impacts , we carried out atomistic simulations of such impacts onto atoms near di - vacancies .
due to the computational cost related to many non - equivalent atoms present in the system and a large number of possible atomic configurations , we could not repeat the detailed analysis of the role of initial space angle of the displacement similar to pristine graphene , and therefore limited our simulations to the @xmath27 case for all non - equivalent atoms at reconstructed di - vacancy structures .
however , because the defects break the symmetry of the lattice , a directional preference for the displacements arises ( atom displaced in the perpendicular direction will change its direction due to local strain ) .
this effect is strong enough to facilitate bond rotations in reconstructed di - vacancy structures . in figure [ px::mvacsim ]
we present examples of such processes for both @xmath25(5 - 8 - 5)@xmath28(555 - 777 ) and @xmath25(555 - 777)@xmath28(5555 - 6 - 7777 ) transformations . within these simulations
, we never observed @xmath25(555 - 777)@xmath29(5 - 8 - 5 ) transformations .
because the symmetry of the @xmath25(555 - 777 ) defect around the middle atom is the same as that of pristine graphene ( the middle atom is represented as a black sphere in the last panel of figure [ px::mvacsim]a ) , one would also expect that a @xmath30 displacement is required for this transformation , similar to the @xmath1(55 - 77 ) case , although in principle the surrounding atoms could also cause this transformation .
we noticed that the most likely di - vacancy transformation , at least for @xmath31 , is the @xmath25(5555 - 6 - 7777)@xmath29(555 - 777 ) .
the @xmath25(5 - 8 - 5)@xmath28(555 - 777 ) was the least likely one of those observed .
we never observed a @xmath25(5 - 8 - 5)@xmath32(5 - 7 ) transformation @xcite during our simulations , which we also attribute to the limited simulated conditions ( @xmath31 ) .
curiously , however , we did observe one transformation in which a @xmath25(5 - 8 - 5 ) di - vacancy directly migrated one step in the zigzag lattice direction . ) ; ( a ) @xmath25(5 - 8 - 5)@xmath28(555 - 777 ) and ( b ) @xmath25(555 - 777)@xmath28(5555 - 6 - 7777 ) .
the black spheres indicate the recoil ( displaced ) atoms .
structurally equivalent atoms to the displaced ones are marked with circles in first panels . in the last panels , the circled atoms are those which could cause a backward transformation upon displacement in addition to the recoil atom .
( see also videos s1214 in ref .
@xcite . ) ] another interesting observation originating from these simulations is the fact that the displacement threshold for atoms in the central part of the reconstructed defects [ @xmath25(555 - 777 ) and @xmath25(5555 - 6 - 7777 ) ] are higher than that for pristine graphene ( by as much as 5% ) .
this may explain why defect structures tend to grow into larger and larger amorphous patches instead of collapsing into holes under continuous electron irradiation at low voltages ( @xmath33 kev ) @xcite : even when atoms are removed from the defected area , the displacements occur at the edges of the existing defects rather than at the central part where the local atomic density is already lower .
clearly , since the core structure of these defects consists of carbon hexagons , there must exist a limiting size above which the displacement threshold becomes similar to that of ideal graphene .
to conclude , by combining ac - hrtem experiments and atomistic simulations , we have shown that the bond rotations which lead to creation of topological defects in carbon nanostructures are caused by single electron impacts or incomplete annihilation of frenkel defects .
this explains the discrepancy between experimental observations of stone - wales defects and their relatively high formation energy and even higher energy barrier for bond rotation .
the sw transformation in graphene can be initiated at least in two different ways upon electron impact ( involving a `` circling '' or `` nudging '' motion ) , and for almost any space angle , provided that enough energy is transferred from the electron to the target atom .
our simulations indicate that @xmath1(55 - 77 ) can appear as a result of `` incomplete '' recombination of a frenkel defect reminiscent of the formation of wigner - type defects in silicon @xcite .
however , this is much more likely in the case of local curvature , as in nanotubes .
more surprisingly , we also observed sputtering of c@xmath7 dimer as a result of annihilation of a frenkel defect in carbon structures with high curvature .
moreover , we showed that the displacement threshold of atoms in the central area of reconstructed defects is higher than that of pristine graphene , by as much as 5% , which explains why defected graphene under low - energy electron irradiation ( @xmath34 kev ) tends to turn graphene into a two - dimensional amorphous structure @xcite instead of a perforated membrane . for
different di - vacancy structures , even displacements in direction perpendicular to the graphene layer can initiate sw transformation and thus local structural changes and defect migration .
our results provide microscopic insight into the irradiation - induced changes in the atomic structure of carbon nanosystems under electron irradiation , and taking into account the interesting electronic properties of defects associated with sw transformations @xcite , may open new avenues for irradiation - mediated engineering @xcite of carbon nanostructures with next - generation electron microscopes .
we acknowledge financial support by the german research foundation ( dfg ) , the german ministry of science , research and arts ( mwk ) of the state baden - wuerttemberg within the salve ( sub angstrom low voltage electron microscopy ) project and academy of finland through several projects .
we also thank csc , espoo , finland , for generous grants on computer time . | observations of topological defects associated with stone - wales type transformations ( _ i.e. _ , bond rotations ) in high resolution transmission electron microscopy ( hrtem ) images of carbon nanostructures are at odds with the equilibrium thermodynamics of these systems . here , by combining aberration - corrected hrtem experiments and atomistic simulations , we show that such defects can be formed by single electron impacts , and remarkably , at electron energies below the threshold for atomic displacements .
we further study the mechanisms of irradiation - driven bond rotations , and explain why electron irradiation at moderate electron energies ( @xmath0100 kev ) tends to amorphize rather than perforate graphene .
we also show via simulations that stone - wales defects can appear in curved graphitic structures due to incomplete recombination of irradiation - induced frenkel defects , similar to formation of wigner - type defects in silicon . |
the past century witnessed the quantum leap of the semiconductor industries which gave birth to the computer science and information technology .
we are now in an era in which information keeps being generated at a skyrocketing pace such that the net volume of information produced per day might be comparable to that accumulated after years one century ago . as an important participant ,
magnetic data recording now has assumed the major task of information documentation , through video tapes , hard disks , etc . in order to cope with the exponentially growing information volume ,
the need to develop data storage devices with higher capacity and faster read / write operation speed is demanded .
this intrigue the development of spintronics - the pursuit to employ , in addition to the charge of electrons , their spin properties into applications .
as one major branch , magnetic domain wall ( dw ) propagation along nanowires has attracted considerable attention in recent years due to its potential in achieving , for instance , high - intensity information storage , nonvolatile random access memory and dw logic circuit @xcite . it has already been known for almost 40 years that the 1 + 1 ( spatial and time ) dimensional landau - lifshitz - gilbert ( llg ) equation @xcite , which universally governs magnetization dynamics , admits a well - known exact walker propagating dw solution for a biaxial nanowire .
it predicts that , in the presence of an external magnetic field , the dw subjects to a rigid - body translational motion which is valid when the magnetic field is in a proper regime . despite its attractive simplicity and elegance , and
the fact that this walker solution has played a pivotal role in our current understanding of both current - driven and field - driven dw propagation in magnetic nanowires @xcite , whether or not this solution is genuine , i.e. , describing a realistic physical system , is still an open question .
as one necessary touchstone , genuine solution of a physical system must be stable against small perturbations . by now there is no proof of the stability of the walker solution and the validity of it for a 1d wire is always taken as self - evident .
any deviation in experiments or numerical simulations are assumed to be attributed to the quasi-1d nature or other effects @xcite .
however , there are signs @xcite that this solution may be unstable . for instance , in reference @xcite , it is shown that under a huge hard - axis anisotropy , a dw motion damped by spin - wave emission occurs after the field exceeds a critical value .
in addition , only stern waves were observed therein . in reference
@xcite , a propagating dw dressed with spin - waves was also captured both in the absence and presence of the gilbert damping , and unlike @xcite , the spin - waves observed emit both stern and bow waves .
moreover , apparent deviation of dw velocity and deformation of dw profile from walker predicted values were also observed . unlike the microscopic dw profile which is sensitive to any errors incurred in simulations , the speed of dw manifests collective behavior of spins composing the dw ; thus it is capable of reflecting the macroscopic physics that are invulnerable to the self - averaging microscopic perturbations when a large number of spins are involved .
therefore , this velocity deviation , as a more conspicuous fingerprint of the dws destabilization , shall also be addressed in regards to its origin in order for a deep understanding of dw propagation in nanowires . on the other hand ,
applications of spintronics devices require accurate description of dw motion @xcite .
thus , the stability of the walker propagating dw solution becomes vital in our understanding of dw propagation along a magnetic wire . in a nanowire , with easy axis along @xmath0 and hard axis along @xmath1 . in the absence of external magnetic field ( upper ) ,
a static dw exists between two domains with @xmath2 . under a field parallel to the easy axis , the walker propagating dw moves towards the energy minimum state ( @xmath3 ) at a speed @xmath4 while the dw profile is preserved .
( b ) illustration of coordinate system .
( c ) solid line : dw profile @xmath5 vs. @xmath6 that varies noticeably around the dw center . dashed line : local perturbation to @xmath5.,title="fig:",width=3 ] + however , unlike stability analysis of solutions of linear and nonlinear ordinary differential equations which can be easily done by using the linearization techniques and lyapunov - exponent concept @xcite , it is hard in general for nonlinear partial differential equations like the llg equation . although numerical approaches can provide clues and hints , an analytical approach is lack for walker solution until recently the progress of traveling wave analysis suggests a feasible way . in this paper , we shall present in detail the method and results of our stability analysis of the walker exact propagating dw solution of a 1 + 1 dimensional llg equation @xcite .
it is shown that a propagating dw is always dressed with spin waves so that the walker solution is not stable against spin - wave emission . in the low field region ,
only stern spin waves are emitted while both stern and bow waves emerge under high field .
when the transverse magnetic anisotropy is larger than a critical value and the external field is sufficiently high , the solution is convective or absolute unstable , corresponding to severe distortion of the propagating dw profile .
this shall lead to noticeable deviation of dw speed from the walker formula besides that the dw is dressed with spin waves .
the paper is organized as follows .
the model and theoretical formulation are explained in the next section .
section iii is the results and discussions , and the conclusion is in section iv .
to study the stability of walker exact propagating dw solution under an external field , we consider the dimensionless llg equation @xcite , @xmath7 this llg equation describes the dynamics of the magnetization @xmath8 of a magnetic nanowire schematically shown in fig .
[ fig1 ] . with the easy axis along the wire ( @xmath0 direction ) and the width and thickness being smaller than the exchange interaction length
, exchange interaction dominates the stray field energy caused by magnetic charges on the edges ; the dw structure tends to be homogeneous in the transverse direction @xcite , i.e. , behaves effectively 1d .
we are interested in the behavior of a head - to - head dw under an external field shown in fig .
[ fig1 ] . in eq .
, @xmath9 is the unit direction of the local magnetization @xmath10 with saturation magnetization @xmath11 and @xmath12 is the phenomenological gilbert damping constant .
the effective field ( in units of @xmath11 ) is @xmath13 where @xmath14 , @xmath15 , and @xmath16 are respectively the easy axis anisotropy coefficient , the hard axis anisotropy coefficient , and the exchange coefficient .
@xmath17 is the external magnetic field parallel to @xmath18 .
the time unit is @xmath19 , where @xmath20 is the gyromagnetic ratio . using polar angle @xmath21 and azimuthal angle @xmath22 for @xmath9 as shown in fig .
[ fig1 ] , this llg equation has a well known walker propagating dw solution @xcite , @xmath23 here @xmath24 is the walker breakdown field and @xmath25 is the dw width which will be used as the length unit ( @xmath26 ) in the analysis below .
@xmath27 is the walker rigid - body dw speed that is linear in the external field and the dw width , and inversely proportional to the gilbert damping constant .
solution is exact for @xmath28 . in the following analysis , the meaning of stability / instability of the dw
is confined to lyapunov definition , i.e. , the dw is stable if any other solution of eq .
starting close enough to the walker solution will remain close to it forever ; otherwise it is unstable .
we will prove the instability of solution against spin - wave emission by performing a spectrum analysis according to a recent developed theory for a general travelling front , such as a propagating head - to - head dw shown in fig .
[ fig1 ] . to prove the instability of solution
, we follow a recently developed theory ( sandstede and scheel @xcite and fiedler and scheel @xcite ) for stability of a general traveling front , that is , a solution connecting two homogeneous states , such as a propagating head - to - head dw shown in fig .
modus operandi _ is to perturb eq . by a small deviation @xmath29 from the solution via which an evolution equation governing this deviation can be derived .
note that if we directly perturb eq .
by @xmath30 , @xmath31 and @xmath32 , with @xmath33 ( @xmath34 , @xmath35 and @xmath36 are components of eq . in the cartesian
coordinates ) , the three components of @xmath29 were not independent due to the preservation of @xmath37 .
a convenient way to circumvent this problem is , instead of analyzing in the cartesian space , to work with the polar - coordinate form of eq . in which the two variables @xmath38 and @xmath39 satisfy @xcite : @xmath40 where single and double prime denote the first and the second derivatives with respect to @xmath41 . by assuming @xmath42 and @xmath43 the solution of eq . with @xmath44 , @xmath45 , and by keeping only the terms of the first order in @xmath38 and @xmath46 , the linearized equations of @xmath38 and @xmath46 in the moving dw frame of velocity @xmath4 ( with the coordinate transformation @xmath47 and @xmath48 , where @xmath49 ) are , in a two - component form of @xmath50 ( superscript @xmath51 means transpose ) , @xmath52 where @xmath53 , @xmath54 , and @xmath55 are @xmath56 matrices that depend on @xmath57 through @xmath58 .
@xmath53 has the following matrix elements : @xmath59+k_\bot(\sqrt{1 -{\rho^2}}-1 ) \cos [ 2g(\xi)]/2 + ( h\alpha -k_\bot\rho /2)\tanh\xi
\}/(1 + \alpha^2)$ ] , @xmath60+k_\bot(\sqrt{1-\rho^2}-1 ) \cosh\xi\cos [ 2g(\xi)]/2+(h + k_\bot\alpha\rho /2)\sinh\xi \ } /(1+\alpha^2)$ ] , @xmath61sesh@xmath62 , @xmath63 . here
@xmath64 is the gudermannian function and @xmath65 .
@xmath54 , @xmath55 can be expressed explicitly in terms of @xmath57 as : @xmath66 eq .
is a linearized equation , and its general solutions are linear combinations of basic solutions of the form , @xmath67 where @xmath68 is a proper complex number that supports nontrivial solutions ( not constant zero ) for equation @xmath69 where @xmath70 .
then all such @xmath71 define the spectrum of @xmath72 .
it is straightforward to verify that , due to translational invariance of solution , @xmath73 always belongs to the spectrum , with the corresponding eigenfunction @xmath74 if none of @xmath71 in the spectrum has positive real part , the spectrum is said to be stable ; otherwise it is unstable . for a stable spectrum
, any moderate deviations from the walker solution must either decay exponentially with time [ @xmath75 or undergo periodic motion by retaining its amplitude [ @xmath76@xmath77 .
when the spectrum encroaches the left half plane , exponentially growing modes ( @xmath76@xmath78 ) exist .
we shall use so - called essential and absolute spectra of @xmath79 to decide the stabilities / instabilities of domains and dw profile .
in order to compute the spectrum of @xmath79 , it is convenient to rewrite eq . in the first order differential form by using @xmath80 , @xmath81 where @xmath82 @xmath83 is the @xmath56 identity matrix .
all @xmath71 that supports nontrivial solutions to eq .
form its spectrum .
eq . have the same spectrum because they are equivalent .
we shall focus hereafter on the spectrum of eq . .
to do so , we need to obtain the conditions under which eq .
has nontrivial solutions .
let us first divide @xmath6 axis into four regions : @xmath84 , @xmath85 , @xmath86 and @xmath87 with @xmath88 .
notice that @xmath89 depends on @xmath6 only through @xmath5 that varies with @xmath6 only within the dw , eq .
is essentially , @xmath90 in region @xmath91 and @xmath92 in region @xmath93 .
the two asymptotic matrices @xmath94 are , @xmath95 @xmath94 can be directly obtained from eq . by replacing @xmath96 with @xmath97 for @xmath98 and with @xmath99 for @xmath100 . in region @xmath91 ( @xmath93 ) and for each given @xmath71 , @xmath101 has 4 eigenvalue and eigenvector pairs , @xmath102 @xmath103 , and eq .
( ) has solution of form @xmath104 ( @xmath105 ) .
@xmath71 can then be denoted by @xmath106 for @xmath107 ( @xmath108 ) being the number of @xmath109 with positive ( negative ) real parts . obviously , we have @xmath110 except on the so - called fredholm borders explained below in detail .
@xmath111 can then be ordered descending by their real parts as @xmath112 .
each solution @xmath113 ( @xmath114 ) in region @xmath91
( @xmath93 ) can be continued into region @xmath115 ( @xmath86 ) as @xmath116 ( @xmath117 ) .
suppose we are interested in a nontrivial bounded solution @xmath118 of eq .
, i.e. , @xmath119 , then @xmath118 must be the linear superposition of those @xmath120 [ @xmath117 ] in @xmath115 ( @xmath86 ) whose corresponding eigenvalues @xmath122
( @xmath123 ) have positive ( negative ) real parts .
note that the number of @xmath122 ( @xmath123 ) with re@xmath124 ( re@xmath125 ) is @xmath126 ( @xmath127 ) , whether or not such @xmath118 exists is equivalent to whether or not we can find nontrivial solution @xmath128 satisfying @xmath129 this is the condition of the continuation of @xmath118 at @xmath130 .
the spectrum of eq .
is the set of all @xmath68 such that eq .
has at least one nonzero solution of @xmath128 for @xmath131 and @xmath132 . obviously , there are @xmath133 variables and 4 equations .
the existence of such a solution is then @xmath134 .
the explict solutions of @xmath128 require the knowledge of @xmath135 that is normally not known analytically because of the complicate @xmath6-dependence of @xmath136 .
numerical method such as the shooting algorithm used in the schrodinger equation may be use here by numerically integrating eq . starting from @xmath137 ( where all linear independent solutions of eqs . and are known ) and ending at @xmath130 .
correct set of @xmath128 shall make the shooting of @xmath118 from @xmath137 end with the same value at @xmath130 .
the shooting algorithm , proved to be efficient for the schrodinger equation whose spectrum is on the real line , may become excessively arduous for the llg eq . where the spectrum extends to the whole complex plane . as we shall see , this formidable task can be partly dodged as far as only the essential instability is considered which is pertinent to spin wave emissions .
similar to the energy spectrum of a quantum system , the spectrum @xmath71 of eq . can be discrete and continuum .
the continuum @xmath71 is also called the essential spectrum . the essential spectrum is not sensitive to the so called relatively compact perturbations to eq . .
here a relatively compact perturbation can be understood , in some senses , as a local perturbation to a schrodinger equation @xmath138\psi = e\psi .\ ] ] this continuous spectrum will not be changed by a @xmath139 of finite potential range , such as a potential well or barrier , although wave functions are altered and point spectrum may be introduced . according to @xcite , a similar local perturbation to eq .
( or in general to any linearized equation of a system around a front solution ) preserve the essential spectrum so that we can replace @xmath5 ( fig .
[ fig1 ] ( c ) ) by @xmath140 , where @xmath141 is the heaviside step function . the new equation with the same essential spectrum as that of eq .
is @xmath142 where @xmath143 @xmath144 have already been defined in eq . .
since @xmath145 is constant in each region of @xmath146 and @xmath147 , the corresponding @xmath148 in eq .
are just eigenvectors @xmath149 of @xmath94 .
. becomes @xmath150 eq .
have nonzero solution if the number of variables @xmath151 and @xmath152 , @xmath153 , is greater than 4 .
thus , eq . has nontrivial solution bounded at @xmath154 for all @xmath71 whose @xmath155 .
if one allow other types of solutions at @xmath154 , then the general condition is @xmath156 . indeed according to the theory of refs .
@xcite , the essential spectrum of @xmath72 ( also of eq . and/or eq . )
is the union of all closed sets of @xmath71 ( boundaries included ) whose indices @xmath157 and @xmath158 satisfy @xmath159 .
the boundaries of each region , known as the fredholm borders , must be those lines crossing which either @xmath127 or @xmath126 changes its value by 1 . then along each fredholm border
, either @xmath160 or @xmath161 must possess pure imaginary eigenvalue ( not a hyperbolic matrix ) ; thus these lines can be determined by @xmath162=0 $ ] with @xmath163 @xcite .
each of the two equations has two branches of allowed @xmath71 denoted as @xmath164 .
note that eqs . and admit pure plane wave solution @xmath165 when @xmath71 is on @xmath166 ( @xmath167 ) .
therefore an encroachment of these borders to the right half plane implies spin wave emission .
we refer to the type of instability characterized by the presence of essential spectrum on the right half plane as the essential instability .
( a ) and @xmath168 ( b ) .
the fredholm borders are @xmath164 .
solid border lines correspond to spin waves with negative group velocities while the dashed border lines are for the spin waves with positive group velocities .
propagating dw wall emit stern waves in low fields [ right of ( a ) ] , and stern and bow waves in higher field ( @xmath169 ) [ right of ( b ) ] .
the green dots are zero group velocity modes .
@xmath170 is used .
( c ) illustration of transient instability ( i ) in which the unstable mode travels to only one direction ; convective instability ( ii ) in which the unstable mode travels to both directions ; and the absolute instability ( iii ) in which the unstable mode does not travel ( stationary in the moving frame ) . , title="fig:",width=3 ] + in order to understand numerical results in ref .
@xcite , parameters of yttrium iron garnet ( yig ) @xcite are assumed in our analysis with @xmath171 , @xmath172 , @xmath173 , and @xmath174 .
@xmath175 is used and @xmath176 is a varying parameter .
[ fig2 ] plots the essential spectrum for @xmath170 ( in units of @xmath177 that is about 10 times larger than @xmath178 ) .
the qualitative results are very similar to the early results @xcite : in the absence of an external field , the two branches of the spectrum of @xmath94 are the same , @xmath179 , shown in fig . [ fig2](a ) .
since the spectrum encroaches the right half plane , unstable plane waves shall exist and spin wave emission are expected .
similar conclusion was also obtained in early study @xcite , but for @xmath180 .
solid lines are for negative group velocity [ determined by im@xmath181 , thus these are stern modes .
the dashed lines indicate positive group velocity , corresponding to bow modes .
the green dots are zero group velocity points . according to fig .
[ fig2](a ) , all unstable modes have negative group velocities so that dw can only emit stern waves in the low fields . as the external field increases , @xmath182 and @xmath183 will separate , and the area of essential spectrum in @xmath184-plane becomes bigger and bigger ( shadowed regimes in fig .
[ fig2](b ) .
the green dots also moves toward im@xmath185-axis and cross it at @xmath186 [ fig .
[ fig2](b ) ] . upon further increase of @xmath17 ,
the unstable modes have both positive and negative group velocities although the most of them have the negative ones .
one shall have propagating dw to emit both stern and bow waves .
the stern waves should be stronger than the bow waves as schematically shown in the right figure of fig .
[ fig2](b ) .
this is exactly what were observed in numerical simulations for stern wave emission in low field @xcite and stern - and - bow wave emission in high field @xcite . in a realistic wire with damping ,
emitted spin waves will be dissipated after a short distance , and are hard to be observed in experiments .
the essential spectrum decides the instability of domains .
dw propagation will generate spin waves in domains when the essential spectrum encroaches the right half of the @xmath71 plane .
the fact that the essential spectrum is not affected by the variation of dw profile means that the essential spectrum can not determine the instability of dw profile that is important for many quantities such as the dw velocity .
interestingly , dw instability is determined by the so - called absolute spectrum explained below .
it can be classified into three categories .
absolute instability [ ai , fig .
[ fig2](c)iii ] occurs when at each fixed point on the @xmath6-axis , the disturbance grows exponentially with time .
it is associated with the emergence of nontraveling unstable modes in the absolute frame ( the moving frame that we adopted ) ; thus coins its name .
this point - wise growth feature of ai is in sharp contrast with the other two types of instability , which albeit grows in the total norm , decays locally at each fixed point on the @xmath6-axis .
they happens when all unstable modes are transported to infinities at fast enough velocities .
it is called a transient instability ( ti , fig . [ fig2](c)i ) if the disturbance generated locally in the dw region transports to infinity in one direction ( either towards @xmath187 or @xmath188 ) , while it is called convective ( ci , fig . [ fig2](c)ii ) if it can transport in both directions .
intuitively , transient instability shall have the least influence on the dw property since once generated , it will leave the dw region quickly and will not interact with the dw hereafter .
convective instability is stronger than the transient one since although transported outside the dw region , it could influence the dw through second order effect in which new bidirectional unstable modes excited by the convecting wave packets can collide and interact with the dw again .
absolute instability is the most severe one in the sense that once a nontraveling disturbance is generated , it can stay within and keep interacting with the dw , leading to dramatic modification on the dw profile .
for this reason , physical quantities depending on dw profile , such as the dw velocity , are expected to be strongly affected .
the three types of transportation behavior , either unidirectional , bidirectional or non - travelling , are determined by the so - called absolute spectrum and the branching points @xcite . to introduce the absolute spectrum and the branching points
, we recall that , for each @xmath71 in the complex plane , there are four @xmath189 ( @xmath190 ) for @xmath191 , ordered by their real parts as @xmath192 .
then @xmath71 is said to belong to the absolute spectrum ( @xmath193 ) if and only if @xmath194=\hbox{re}[\kappa_3^+(\lambda)]$ ] or @xmath195=\hbox{re}[\kappa_3 ^ -(\lambda)]$ ] .
the branching points are special points in the absolute spectrum , denoted as @xmath196 , satisfying @xmath197 .
to have a better feeling about the absolute spectrum and differences in unidirectional / bidirectional transportation and nontraveling modes of a wave @xmath198 , we introduce the concept of pointwise decay and growth .
a wave @xmath198 is said to be pointwise decay iff @xmath199 for any fixed @xmath200 .
the opposite ( @xmath187 instead of 0 ) is said to be pointwise growth .
let us first consider a wavelet that may exemplify a transient disturbance transporting to the right along the z axis : @xmath201 this is an unstable mode if @xmath202 .
at each fixed point @xmath200 , @xmath199 ( @xmath187 ) if @xmath203 ( @xmath204 ) . in another word ,
an unstable disturbance moving fast enough can lead to pointwise decay ( vanish in a long time at each fixed point ) although its norm @xmath205 increases exponentially with time .
interestedly , can be brought to be stable when @xmath206 if an exponential weight @xmath207 is used @xmath208 where @xmath209 and @xmath210 note that the integral @xmath211 is finite whenever @xmath212 .
therefore any @xmath213 $ ] makes @xmath214 such that the new norm decay exponentially with time .
however , if @xmath204 , either diverges with time or @xmath211 is infinity for any @xmath215 . in another word ,
the mode becomes stable under a proper exponential weight @xmath216 for @xmath203 , and a transient disturbance traveling towards @xmath188 fast enough can be stabilized by a positive @xmath215 since then the multiplier @xmath207 balance the growing modes at @xmath188 . in regions divided by @xmath217 for @xmath170 , @xmath218 ( a ) and @xmath219 ( b ) .
no absolute spectrum presents before @xmath220 tangents at the real axis while in ( b ) , unstable absolute spectrum presents in the region enclosed by the solid circle .
( c ) enlarged description of the region enclosed by the solid circle in ( b ) .
the absolute spectrum is between the two branching points @xmath221 and @xmath222 ( green dots ) .
( d ) plot of @xmath223 and @xmath224 vs. @xmath71 between @xmath221 and @xmath222 . at @xmath225 , @xmath226 .
( e ) phase diagram of transient ( ti ) and absolute / convective ( ai / ci ) instabilities .
the boundary is the bifurcation line between ti / ci - and - ci instabilities in @xmath15 and @xmath227 plane .
the bifurcation line is only plotted for @xmath228 here @xmath229 at which @xmath230 ( @xmath231 ) .
noted that our analysis is valid for fields below the walker breakdown value.,title="fig:",width=3 ] + in general , the exponentially weighted norm denoted by @xmath232 for a real number @xmath215 is defined as @xmath233 the transient and convective instabilities behave very differently under the norm . for a given @xmath71 ,
its eigenmode is transient unstable if it has an exponentially growing factor that travels towards @xmath188 ( or @xmath187 ) . under an exponentially weighted norm with a proper choice of @xmath234 ( @xmath235 ) for mode traveling to @xmath188 ( @xmath187 ) , the growth at @xmath188 ( @xmath187 )
can be absorbed by the multiply @xmath236 .
therefore the essential spectrum calculated under the exponential norm with @xmath234 ( @xmath235 ) can be transferred to the left half of the @xmath71 plane for the unidirectional modes traveling towards @xmath188 ( @xmath187 ) .
mathematically , this corresponds to a proper choice of the origin of the @xmath71 plane in some sense .
thus , with the proper definition of the norm by choosing a large enough @xmath237 , all unstable unidirectional eigenmodes of eigenvalues @xmath71 ( essential spectrum in the right half of the @xmath71 plane ) are removable because all such @xmath71 can be transferred to the left half of the @xmath71 plane .
this treatment fail to the modes traveling to both directions of @xmath238 ( bidirectional eigenmodes ) .
they are not removable since an exponential weight can only suppress the growth in one direction and blow up in the other direction .
the ability / inability of using an exponential weight to stabilize / destabilize transient / convective modes leads to the following properties : ti occurs if all unstable essential spectrum can be move to the left half @xmath71 plane under a proper exponentially weighted norm while it is ci or ai if part of the unstable essential spectrum can not be stabilized by the norm .
a naturally raised question then is which part of the essential spectrum can not be removed by this weight .
the answer is quite simple : the absolute spectrum can not be moved around in the @xmath71 plane by introducing an exponential weight .
it must locate to the left of the rightmost fredholm border . if it encroaches the right half of the @xmath71 plane , then the essential spectrum can not be stabilize no matter how one chose the exponential weight @xmath215 . to see why this is
so , it is noticed that we need to introduce a weight @xmath239 [ @xmath240 in order to move the fredholm border determined by eq .
thus , by using the exponential weight of @xmath241 , it is equivalent to shift the eigenvalues of @xmath94 @xcite by @xmath242 and accordingly the indices of the @xmath71 are transformed as @xmath243 now suppose @xmath71 with @xmath202 belongs to the essential spectrum but not on the fredholm border , which means @xmath156 . for the llg equation and independent of the norm we use , @xmath71 in the right hand side of the rightmost fredholm border
has the indices of @xmath244 , then all possible combinations of @xmath245 in the regions right after passing through the rightmost fredholm border can only be one of the four cases : @xmath246 , @xmath247 , @xmath248 , @xmath249 .
consider for instance @xmath250 , then obviously @xmath251 and @xmath252 .
if we also have @xmath253 , i.e. @xmath254 , we could always find the aforementioned proper weight as , for instance : @xmath255 which means that , for essential spectrum calculated under this norm , @xmath71 is well to the right of the essential spectrum .
we can thus remove all unstable @xmath71 ( i.e. , @xmath202 ) in this way if there is no @xmath193 on the right half of the @xmath71 plane .
however , if @xmath71 belongs to @xmath193 such that @xmath256 , it is easy to verify that no such pair of @xmath241 exist .
this absolute spectrum is exactly the set of @xmath71 which could not be stabilized by the aforementioned proper weights @xmath241 .
therefore we conclude that the absence of @xmath193 in the right half of the @xmath71 plane indicate transient instability in which all eigenmodes are unidirectional while the presence of the absolute spectrum means emergence of bidirectional eigenmodes .
finally , the presence of unstable non - traveling modes is associated with the branching points s presence on the right half plane .
it is straightforward to verify that the eigenmodes associated with these branching points have zero group velocity , as follows .
denote the secular polynomial of @xmath101 as : @xmath257.$ ] then for @xmath196 satisfying @xmath258 , it must hold that : @xmath259 where @xmath260 .
then the group velocity @xmath4 of modes associated with @xmath196 is @xmath261 thus , branching points @xmath196 are non - travelling eigenmode @xcite . for @xmath170 ,
the absolute spectrum in the right half of the @xmath71 plane is generated by @xmath161 .
[ fig3](a ) shows two branches @xmath262 .
they are well separated by the real axis for @xmath218 and no absolute spectrum could be found in the right half plane .
as the field increases , the two branches get closer with each other and at an onset field @xmath263 , depending on @xmath15 , two branches tangent at the real axis and then separate again in horizontal direction as shown in fig .
[ fig3](b ) for @xmath264 .
at this moment , unstable absolute spectrum begins to emerge on the real axis ( enclosed by the dashed circle ) .
[ fig3](c ) is the enlarged vision showing the absolute spectrum ( the segment between two branching points @xmath225 ( green solid dots ) ) .
the dependence of @xmath265 or @xmath266 on @xmath267 between these two points is shown in fig . [ fig3](d ) . according to refs .
@xcite , wavepackets would be emitted if the essential spectrum encroaches the right half @xmath71-plane .
there are three types of instability @xcite .
the instability is called transient ( ti ) if the essential spectrum encroaches the right half plane and absolute spectrum are either in the left half plane or does not exist .
the propagating dw emits stern waves shown in fig .
. the instability is called convective if both essential and absolute spectrum encroaches the right half @xmath71-plane . in this case
, the emitted waves can propagate in both direction as shown by fig .
[ fig2](c)ii . for an convective instability ,
if any branching point is also in the right half @xmath71-plane , the instability is called absolute .
an absolute instability can then emit non - traveling ( zero group velocity ) waves as illustrated in fig .
[ fig2](c)iii . for llg equation , since the absolute spectrum is the segment connecting two branching points @xmath268 and @xmath269 [ fig .
[ fig3 ] ( c ) , ( d ) ] , the absolute instability ( ai ) and convective instability ( ci ) co - exist .
it is known that transient instability is very weak that can be removed under proper mathematical treatment @xcite .
thus , we should not expect to have great physical consequences . on the other hand ,
the absolute instability move with the dw , and cause the change of dw profile @xcite .
it is known @xcite that field - induced dw propagating speed is proportional to the energy damping rate that is sensitive to dw profile . therefore absolute instability , which deform propagating dw profile , shall substantially alter dw speed .
this may explain why the field - induced dw speed start to deviate from the walker result only when the field is large enough to emit both stern and bow waves in simulations @xcite .
[ fig3](e ) is the calculated phase diagram in @xmath15 and @xmath227 plane .
a transition from transient instability ( denoted as ti in the figure ) to absolute / convective instability ( ai / ci ) occur at a critical field @xmath263 as lng as @xmath270 at which @xmath230 .
it means no absolute / convective instability exist for @xmath271 , and one shall not see noticeable change in famous walker propagation speed mentioned early .
this may explain why many previous numerical simulations on permalloy , which have small transverse magnetic anisotropy , are consistent with walker formula .
a snapshot of the convecting wavepackets could be identified in fig . 2 in reference @xcite where wavepackets can be seen in the vicinity of the traveling dw and travel to both directions .
it should be noticed that the effects of point spectrum have not been analyzed . in principle
, it can also affect the stability of the walker solution , and should be a very interesting subject too .
unfortunately , there are not many theorems on the point spectrum yet .
thus , one can only rely on a numerical method to find a point spectrum of operator @xmath72 and to find out whether it can also induce any instability on a propagating dw .
in conclusion , we present a powerful recipe for analyzing the stability of a front of partial differential equation . for the walker propagating dw solution of the llf equation in 1 + 1 dimension , it is found that dw will always emit stern waves in a low field , and both stern and bow waves in a higher field .
thus the exact walker solution of llg equation is not stable .
the true propagating dw is always dressed with spin waves . in a real experiment .
the emitted spin waves shall be damped away during their propagation , and make them hard to be detected in realistic wires . for a realistic wire with its transverse magnetic anisotropy larger than a critical value and when the applied external field is larger than certain value , a propagating dw may undergo simultaneous convective and absolute instabilities . as a consequence ,
the propagating dw will not only emit both spin waves and spin wavepackets , but also change significantly its profile .
thus , the corresponding walker dw propagating speed will deviate from its predicted value , agreeing very well with recent simulations .
this work is supported by hong kong rgc grants ( 604109 and 605413 ) .
schryer and l.r .
walker , j. appl . phys . * 45 * , 5406 ( 1974 ) .
parkin , m. hayashi , and l. thomas , science * 320 * , 190 ( 2008 ) .
d. a. allwood et al . , science * 309 * , 1688 ( 2005 ) .
v. cros et al .
, wo patent 2,006,064,022 , ( 2006 ) .
allwood , g. xiong , c.c .
faulkner , d. atkinson , d. petit , and r.p .
cowburn , science * 309 * , 1688 ( 2005 ) .
beach , c. knutson , c. nistor , m. tsoi , and j.l .
erskine , phys .
lett . * 97 * , 057203 ( 2006 ) .
wang , p. yan , j. lu and c. he , ann . phys .
* 324 * , 1815 ( 2009 ) ; x.r .
wang , p. yan , and j. lu , epl * 86 * , 67001 ( 2009 ) .
l. landau , and e. lifshitz , phys .
z. sowjetunion * 8 * , 153 ( 1953 ) ; t.l .
gilbert , phys . rev . *
100 * , 1243 ( 1955 ) . z. li and s. zhang , phys .
lett . * 92 * , 207203 ( 2004 ) .
a. thiaville1 , s. rohart , ju.v .
cros and a. fert , epl * 100 * , 57002 ( 2012 ) .
j. linder , phys .
b * 87 * , 054434 ( 2013 ) .
r. wieser , e.y .
vedmedenko , and r. wiesendanger , phys .
b * 81 * , 024405 ( 2010 ) .
wang , p. yan , y.h .
shen , g.e.w .
bauer , and x.r .
wang , phys .
* 109 * , 167209 ( 2012 ) . j. sthr and h.c .
siegmann , _ magnetism : from fundamentals to nanoscale dynamics _ , ( springer - verlag , berlin , 2006 ) .
m. hatami , g.e.w .
bauer , q. zhang , and p.j .
kelly , phys .
lett . * 99 * , 066603 ( 2007 ) .
han , s.k .
kim , j.y .
lee , s.j .
hermsdoerfer , h. schultheiss , b. leven , and b. hillebrands , appl .
. leet . * 94 * , 112502 ( 2009 ) .
p. yan , x.s .
wang , and x.r .
wang , phys .
lett . * 107 * , 177207 ( 2011 ) .
b. hu and x.r .
wang , phys .
b * 87 * , 035311 ( 2013 ) .
wang , j.n .
wang , b.q .
sun , and d.s .
jiang , phys .
b * 61 * , 7261 ( 2000 ) ; z.z .
sun , h.t .
wang , s.d .
wang , and x.r .
wang , phys .
b * 69 * , 045315 ( 2004 ) .
b. hu and x.r .
wang , phys .
lett . * 111 * , 027205 ( 2013 ) .
d.g . porter and m.j .
donahue , j. appl .
95 * , 6729 ( 2004 ) .
palmer , proc .
soc . * 104 * , 149 ( 1988 ) .
b. sandstede and a. scheel , dynam . syst . * 16 * , 1 ( 2001 ) ; b. sandstede , and a. scheel , math . nachr . * 232 * , 39 ( 2001 ) .
b. fiedler and a. scheel , in_trends in nonlinear analysis _ , edited by m. kirkilionis , s. kromker , r. rannacher , and f. toni ( springer , 2003 ) .
d. henry , _ geometric theory of semilinear parabolic equations _ , ( springer , 1981 ) .
l. perko , _ differential equations and dynamical systems _ , ( springer , 2001 ) ; t. kato , _ perturbation theory for linear operators _ , ( springer , 1995 ) .
d. bouzidi and h. suhl , phys .
lett . * 65 * , 2587 ( 1990 ) .
chomaz , phys .
lett . * 69 * , 1931 ( 1992 ) ; a. couairon , and j.m .
chomaz , physica d * 108 * , 236 ( 1997 ) .
l. brevdo and t. j. bridges , philos .
london , ser .
a * 354 * , 1027 ( 1996 ) .
b. sandstede , _ handbook of dynamical systems ii _ , ( north - holland , 2002 ) ; b. sandstede and a. scheel , phys .
e. * 62 * , 7708 ( 2000 ) .
pego and m.i .
weinstein , commun .
phys . * 164 * , 305 ( 1994 ) .
j. humpherys , b. sandstede , and k. zumbrun , numer . math . * 103 * , 631 ( 2006 ) ; j. humpherys and k. zumbrun , physica d * 220 * , 116 ( 2006 ) .
rademacher , j. appl .
* 5 * , 634 ( 2006 ) | a powerful mathematical method for front instability analysis that was recently developed in the field of nonlinear dynamics is applied to the 1 + 1 ( spatial and time ) dimensional landau - lifshitz - gilbert ( llg ) equation . from the essential spectrum of the llg equation
, it is shown that the famous walker rigid body propagating domain wall ( dw ) is not stable against the spin wave emission . in the low field region
only stern spin waves are emitted while both stern and bow waves are generated under high fields . by using the properties of the absolute spectrum of the llg equation
, it is concluded that in a high enough field , but below the walker breakdown field , the walker solution could be convective / absolute unstable if the transverse magnetic anisotropy is larger than a critical value , corresponding to a significant modification of the dw profile and dw propagating speed . since the walker solution of 1 + 1 dimensional llg equation can be realized in experiments , our results could be also used to test the mathematical method in a controlled manner . |
the free energy ( or potential of mean force ) is the thermodynamic force driving structural processes such as conformational changes of macromolecules in aqueous solution , ligand binding at the active site of an enzyme , protein - protein association , etc .
the free energy gives information about both the rate at which these processes occur and the mechanism by which they occur .
this makes free energy calculations a central issue in biophysics .
molecular dynamics ( md ) simulations provide a tool for performing such calculations on a computer in a way which is potentially both precise and inexpensive ( e.g. @xcite ) .
since a free energy is in essence the logarithm of a probability density function ( see ( [ eq : free ] ) below for a precise definition ) it can in principle be calculated by histogram methods based on the binning of an md trajectory .
this direct approach , however , turns out to be unpractical in general because the time scale required for the trajectory to explore all the relevant regions of configuration space is prohibitively long .
probably the best known and most widely used technique to get around this difficulty is the weighted histogram analysis method ( wham ) @xcite . following @xcite , wham adds artificial biasing potentials to maintain the md system in certain umbrella sampling windows .
wham then recombines in an optimal way the histograms from all the biased simulations to compute the free energy .
wham is much more efficient than the direct sampling approach , and generalizations such as @xcite alleviate somewhat the problem of where to put the umbrella windows ( usually , this requires some _ a priori _ knowledge of the free energy landscape ) . in practice
, however , wham remains computationally demanding and it only works to compute the free energy in 2 or 3 variables .
an interesting alternative to wham is metadynamics @xcite . in essence
metadynamics is a way to use an md trajectory to place inverted umbrella sampling windows on - the - fly and use these windows both to bias the md simulation and as histogram bins to sample the free energy directly ( thereby bypassing the need of further histogram analysis in each window ) .
both wham and metadynamics compute the free energy directly by histogram methods , but an alternative approach is possible . unlike the free energy which is a global quantity , its negative gradient ( known as the mean force ) can be expressed in terms of a local expectation and thereby computed at a given point in the free energy landscape .
this is the essence of the blue moon sampling strategy @xcite and it offers the possibility to calculate first the mean force at a given set of locations , then use this information to reconstruct the free energy globally . in one dimension , this approach is known as thermodynamic integration and it goes back to kirkwood @xcite . in higher dimensions , however , this way to compute free energies has been impeded by two issues .
the first is where to place the points at which to compute the mean force , and the second is how to reconstruct the free energy from these data in this paper , we propose a method , termed single - sweep method , which addresses both of these issues in two complementary but independent steps . in a first step ,
we use the temperature - accelerated molecular dynamics ( tamd ) proposed in @xcite ( see also @xcite ) to quickly sweep through the important regions of the free energy landscape and identify points in these regions where to compute the mean force . in the second step
we then reconstruct the free energy globally from the mean force by representing the free energy using radial - basis functions , and adjusting the parameters in this representation via minimization of an objective function .
the single - sweep method is easy to use and implement , does not require _ a priori _ knowledge of the free energy landscape , and can be applied to map free energies in several variables ( up to four , as demonstrated here , and probably more ) .
the single - sweep method is also very efficient , especially since the mean force calculations can be performed using independent calculations on distributed processors ( i.e. using grid computing facilities @xcite ) .
the reminder of this paper is organized as follows . in sec .
[ sec : theory ] , we describe the two steps of the single - sweep method in detail , starting with the second one for convenience . in sec .
[ sec : app1 ] we illustrate the method on a simple two - dimensional example .
this example is then used for comparison with metadynamics in sec .
[ sec : metadynamics ] . in sec .
[ sec : ad ] we use the single - sweep method to compute the free energy of alanine dipeptide ( ad ) in solution in two and in four of its dihedral angles . finally , concluding remarks are made in sec .
[ sec : conclu ] and the details of the md calculation on ad are given in appendix [ sec : md ]
we shall consider a molecular system with @xmath0 degrees of freedom whose position in configuration space @xmath1 will be denoted by @xmath2 .
we also introduce a set of @xmath3 collective variables @xmath4 which are functions of @xmath2 such as torsion angles , interatomic distances , etc .
if @xmath5 denotes the potential energy of the system and @xmath6 its temperature , the free energy @xmath7 in the variables @xmath8 is defined as @xmath9 so that @xmath10 is , up to a proportionality constant , the probability density function ( pdf ) of the variables @xmath8 .
as mentioned in the introduction , the negative gradient of the free energy , @xmath11 , is known as the mean force , and it can be computed locally at point @xmath12 via calculation of an expectation ( see ( [ eq : meanforceapprox ] ) below ) . in this section , we shall suppose that we have obtained an estimate of @xmath13 at points @xmath14 , and we focus on the reconstruction of the free energy @xmath7 from these data . a specific way to pick these points and compute @xmath15 will be given in sec .
[ sec : tamd ] , but it is worth pointing out that the reconstruction method proposed here works with data set collected in any other ways . our reconstruction method uses a radial - basis function representation for the free energy @xmath7 with centers at @xmath16 @xcite : @xmath17 here @xmath18 is a constant used to adjust the overall height of @xmath19 but is otherwise irrelevant , @xmath20 denotes the euclidean norm in @xmath21 , and @xmath22 where @xmath23 is a radial - basis function ; a convenient choice is to use the gaussian packet @xmath24 though other radial - basis functions ( multiquadric , sobolev splines , wendland , etc .
@xcite ) can be used as well , see sec .
[ sec : ad ] . in ( [ eq : radialbasis ] ) the heights @xmath25 and the radial - basis function width @xmath26 are adjustable parameters which we determine by minimizing over @xmath25 and @xmath27 the following objective function , which measures the discrepancy between the negative gradient of the function @xmath28 in ( [ eq : radialbasis ] ) at the centers @xmath29 , @xmath30 , and the mean force @xmath31 estimated at these centers : @xmath32 before explaining how we perform this minimization , let us give several reasons why the radial - basis representation ( [ eq : radialbasis ] ) for @xmath7 is natural and convenient . first , the centers @xmath29 in ( [ eq : radialbasis ] ) do not have to lie on a regular grid , which permits to use mean force data collected anywhere .
second , the representation ( [ eq : radialbasis ] ) can be used in any dimension .
third , this representation has very good convergence properties , i.e. a small number of centers gives an accurate representation of @xmath7 .
in fact , unlike standard representations based e.g. on linear interpolation on a regular grid , the rate of convergence in @xmath33 of the representation in ( [ eq : radialbasis ] ) can be made independent of @xmath3 ( a feature which the radial - basis representations share with sparse grids @xcite ) .
going back to the minimization of @xmath34 , it can be performed as follows . for fixed @xmath27 , the @xmath35 minimizing ( [ eq : objective ] ) solve the following linear algebraic system @xmath36 where @xmath37 and @xmath38 are given by @xmath39 given the centers @xmath29 and the estimates @xmath40 of the mean force at these centers , the coefficients @xmath41 and @xmath42 can be easily computed , and the linear system ( [ eq : linalg ] ) can be solved by any standard technique , e.g. gaussian elimination .
once the solution @xmath43 of ( [ eq : linalg ] ) is determined , to find the optimal @xmath44 satisfying @xmath45 we compute the residual @xmath46 for increasing values of @xmath27 starting from the distance between the centers .
more sophisticated procedures could be used to minimize @xmath46 over @xmath27 , but the brute force method that we used proved to be efficient enough because computing successive solutions of ( [ eq : linalg ] ) for various @xmath27 is very fast . to measure the error in the approximation
, we used the residual per center defined as @xmath47 which reaches its minimum value at the same @xmath44 as @xmath46 .
overall , the procedure is simple and inexpensive since the determination of @xmath35 at fixed @xmath27 is computationally straightforward and cheap , and can be easily repeated to perform the one - dimensional minimization over @xmath27 .
one caveat that we should mention , however , is that the condition number of the matrix @xmath37 increases rapidly when the number of centers and/or @xmath27 increase .
this is a known problem of radial - basis functions @xcite . to avoid any problems , we capped the admissible condition number at @xmath48 and , in situations where this threshold value was reached while @xmath49 was still decreasing , picked for @xmath44 the corresponding value of @xmath27 .
these situations only occurred in the two - dimensional example ( see sec .
[ sec : app1 ] ) when a lot of centers were used ( 500 or more , i.e. much more than what will be used in the ad example ) , and even in these cases , such a large condition number did not lead to any noticeable loss of accuracy in the results ( even though the coefficients @xmath35 were then very large ) .
we also observed that , given a number of centers @xmath29 and a value of @xmath27 , the condition number is typically lower when the dimension of @xmath12 is larger .
finally , we observed that the condition number was much lower with the wendland radial - basis function ( see ( [ eq : wend ] ) in sec .
[ sec : ad ] ) than with the gaussian radial - basis function ( [ eq : kernel ] ) .
it remains to explain how to identify the centers @xmath50 and estimate the mean force at these points .
following @xcite , we will do so using the extended system @xmath51 & \quad + \text{thermostat terms at $ \beta^{-1}$ } \end{aligned } \\[20pt ] \displaystyle\gamma \dot \zz = \kappa ( \ttheta(\xx ) - \zz ) + \sqrt{2 \gamma \bar \beta^{-1}}\ , \eeta(t ) \end{cases}\ ] ] where @xmath52 is the mass matrix , @xmath53 is a white - noise , i.e. a gaussian process with mean 0 and covariance @xmath54 , and @xmath55 , the friction coefficient @xmath56 and the artificial inverse temperature @xmath57 are parameters whose role we explain now .
the system in ( [ eq : tamd ] ) describes the motion of @xmath2 and @xmath12 over the extended potential @xmath58 as shown in @xcite , by adjusting the parameter @xmath59 so that @xmath60 and the friction coefficient @xmath61 so that @xmath12 moves slower than @xmath2 , one can generate a trajectory @xmath62 in @xmath63-space which effectively moves at the artificial temperature @xmath64 on the free energy computed at the physical temperature @xmath6 . by taking @xmath65 , the @xmath62 trajectory visits rapidly the regions where the free energy is relatively low ( i.e. within a range of a few @xmath64 ) even if these regions are separated by barriers which the system would take a long time to cross at the physical temperature @xmath6 .
this gives us a way to determine automatically where are the relevant regions in free energy space .
in @xcite , the extended system in ( [ eq : tamd ] ) was proposed to sample the free energy landscape directly . here
, we make a different use of ( [ eq : tamd ] ) : we utilize the trajectory @xmath62 to rapidly sweep through @xmath12-space and generate the centers @xmath16 used in the radial - basis representation ( [ eq : radialbasis ] ) .
specifically , we start from @xmath66 , then deposit a new center @xmath29 along @xmath62 each time @xmath62 reaches a point which is more than a prescribed distance @xmath67 away from all the previous centers , where @xmath67 is a parameter controlling the density of the covering by the centers ( the smaller @xmath67 , the higher the number of centers deposited ) . at the same time , at each of these centers @xmath29 , we launch a simulation of ( [ eq : tamd ] ) with @xmath68 fixed , i.e we use @xmath69 & \quad + \text{thermostat terms at $ \beta^{-1}$ } \end{aligned}\ ] ] and compute : @xmath70 the calculations of these time averages are independent of each other , and hence they can be distributed , using ( ideally ) at least one processor per center @xmath29 , an approach that optimally fits with the purposes of grid computing @xcite .
the estimator in ( [ eq : meanforceapprox ] ) has the advantage of being simple , but it introduces an error due to the finiteness of @xmath71 .
this error can be decreased by using @xmath71 in ( [ eq : tamd2 ] ) and ( [ eq : meanforceapprox ] ) larger than @xmath59 in ( [ eq : tamd ] ) , or even eliminated by using constrained instead of restrained simulations and using the blue - moon estimator for the mean force @xcite .
once the centers @xmath72 have been deposited and the estimates @xmath73 of the mean force at these centers have been obtained , we use the reconstruction procedure explained in sec.[sec : reconstruct ] and compute the optimal set of coefficients @xmath35 and the optimal @xmath44 to use in the representation ( [ eq : radialbasis ] ) for @xmath7 .
we conclude this section by stressing that using the extended dynamics ( [ eq : tamd ] ) to sweep through @xmath12-space and deposit the centers @xmath29 is very different than using it to sample @xmath7 directly , which makes our approach very different from wham or metadynamics .
unlike with sampling , revisiting twice a region in @xmath12-space is unnecessary and even undesirable since no new center will be deposited .
the accuracy of the reconstruction depends on the number of centers and the accuracy at which the mean force is computed in ( [ eq : meanforceapprox ] ) much more than the precise locations where the centers are deposited .
an important practical consequence is that it is rather straightforward to pick the parameters @xmath59 and @xmath61 in ( [ eq : tamd ] ) since the final result is robust against variations in these parameters . ) by forward euler with a time - step @xmath74 for @xmath75 steps ( white curve ) shown above the contour plot of the mueller potential ( with @xmath76 level sets evenly distributed between @xmath77 and @xmath78 in a scale where the minimum of the potential is @xmath77 ) .
the red circles are the locations of the centers deposited along the trajectory using @xmath79 . in this run , @xmath80 centers were deposited.,width=312 ]
since , given the location of the centers @xmath29 , the mean force estimation at these points is quite standard , as a first illustration we use a two - dimensional example for which @xmath81 and @xmath82 where @xmath5 is the mueller potential @xcite . in this case , there is no need to extend the system as in ( [ eq : tamd ] ) , and the temperature accelerated dynamics simply reduces to ( setting @xmath83 by appropriate rescaling of time ) @xmath84 fig .
[ fig:1 ] shows a tamd trajectory generated by solving ( [ eq : tamdex ] ) by forward euler with the initial condition @xmath85 and a time - step of @xmath86 for @xmath75 time - steps at @xmath87 ( for comparison the energy barrier between the two main minima of the mueller potential is about @xmath88 ) .
also shown are the centers @xmath89 obtained by depositing a new center along the trajectory each time the trajectory reaches a point which is @xmath79 away for all the previous centers . in this run ,
@xmath80 centers were deposited . at the centers
, we used @xmath90 as estimate of the `` mean force '' ( i.e. there is no sampling error in the present example ) .
we then used these data to reconstruct the free energy as explained in sec .
[ sec : reconstruct ] .
[ fig : muelsweepres ] shows the residual per center @xmath49 defined in ( [ eq : resid ] ) .
the optimal @xmath27 for this run was @xmath91 and the condition number at this @xmath44 was @xmath92 .
the level sets of the reconstructed potential are shown in fig .
[ fig:2 ] and compared to those of the original mueller potential , while fig .
[ fig : muelsweeptrj1 ] compares the values of the original and reconstructed mueller potential along the tamd trajectory shown in fig .
[ fig:1 ] . as a simple estimate of the error
, we used : @xmath93 where @xmath94 denotes the reconstructed potential and @xmath95 is the domain in which the original potential remains less than @xmath96 above its minimum value .
the error defined in ( [ eq : error1 ] ) for this calculation was @xmath97 .
these results , which are already very good , can be improved by diminishing @xmath67 and thereby increasing the number of centers without having to increase the length of the tamd trajectory .
for example , by taking @xmath98 , we obtained @xmath99 centers in a trajectory still @xmath100 steps long .
using these centers to reconstruct the mueller potential , we obtained @xmath101 .
the level sets of the reconstructed and original potential defined in fig .
[ fig:2 ] now overimposed so perfectly that they could not be distinguished on the scale of fig .
[ fig:2 ] ( data not shown ) .
the optimal @xmath27 for this calculation was @xmath102 , at which value the residual was @xmath103 and the condition number was @xmath104 .
finally , we note that the result can also by improved by keeping the same distance @xmath98 between centers but increasing the length of the tamd trajectory .
for instance , increasing the number of steps to @xmath105 produced a reconstruction with an error @xmath106 ( data not shown ) .
defined in ( [ eq : resid ] ) for the reconstruction of the mueller potential with the @xmath107 steps single - sweep trajectory shown in fig .
[ fig:1 ] .
the optimal @xmath27 for this run was @xmath91.,width=312 ] ) ( black curve ) .
here we use the @xmath80 centers shown in fig .
[ fig:1 ] .
the optimal @xmath27 is @xmath91 , see fig .
[ fig : muelsweepres ] .
we show 29 level sets evenly distributed between @xmath77 and @xmath78 .
the level sets of the reconstructed potential and the original one are in so close agreement that they can only be distinguished in some localized regions ( e.g. near the saddle point between the two minima in the lower right corner ) . , width=312 ]
because metadynamics @xcite also uses an extended dynamical system for @xmath2 and @xmath12 and the gaussian packet ( [ eq : kernel ] ) to represent @xmath7 , the single - sweep method bears similarities with it . yet
, there is an essential difference between the two methods . unlike the single - sweep method , metadynamics
does not use the mean force , and estimates @xmath7 by direct sampling , which turns out to be a less efficient way to proceed .
let us elaborate on this claim .
recall that metadynamics uses an extended system like ( [ eq : tamd ] ) but where the equation for @xmath12 is replaced by @xcite @xmath108 here @xmath6 is now the physical temperature of the system , and @xmath59 and @xmath61 are parameters playing the same role as in ( [ eq : tamd ] ) .
the integral term in ( [ eq : zzmeta ] ) is a flooding term ( with @xmath109 controlling the flooding rate ) which deposits gaussian packets @xmath110 on the energy landscape wherever @xmath62 goes , thereby progressively leveling the effective free energy landscape felt by @xmath62 .
the negative of the integral of the gaussian packets deposited then gives an approximation of the free energy . for a trajectory with @xmath111 time - steps of size @xmath112 ,
the time - discretized approximation of this integral reads ( compare ( [ eq : radialbasis ] ) ) @xmath113 where @xmath18 is a constant used to adjust the height of @xmath19 .
steps single - sweep trajectory shown in fig .
[ fig:1 ] .
the green line shows the absolute value of the difference between the two.,width=312 ] despite the fact that ( [ eq : freemeta ] ) uses gaussian packets which are radial - basis functions , the representation ( [ eq : freemeta ] ) is very different from the standard radial - basis representation ( [ eq : radialbasis ] ) used in the single - sweep method . in particular , there are no coefficients @xmath25 to adjust in ( [ eq : freemeta ] )
this has the important consequence that , instead of requiring a single sweep across @xmath12-space to get an accurate estimate of the free energy , metadynamics requires that the trajectory revisits many times the same locations in @xmath12-space to deposit centers ( i.e. @xmath111 in ( [ eq : freemeta ] ) must be much larger than @xmath33 in ( [ eq : radialbasis ] ) to achieve the same accuracy ) .
this is because the leveling out achieved by the integral term in ( [ eq : zzmeta ] ) and , hence , the convergence of the representation ( [ eq : freemeta ] ) , only occur statistically @xcite ( in contrast , the mean force data used at each center in the single sweep method contains already all the statistical information needed at that center ) .
this is consistent with metadynamics being in essence an histogram method , albeit one where the histogram windows are adjusted on - the - fly .
steps overimposed on the original mueller potential.,width=312 ] what this entails in terms of efficiency can be illustrated on the two - dimensional mueller example considered before . in this example , to generate the metadynamics trajectory we used ( [ eq : tamdex ] ) with flooding terms added as in ( [ eq : zzmeta ] ) , consistent with what was done in ref .
@xcite to test the efficiency of metadynamics in a similar set - up .
[ fig : muelmetaxy ] shows the metadynamics trajectory obtained by integrating ( [ eq : zzmeta ] ) for @xmath75 timesteps with a time - step of @xmath114 ( same as in the single - sweep method for the results shown in figs .
[ fig:1 ] to [ fig : muelsweepres ] ) . as can be seen in fig .
[ fig : muelmetaxy ] , this number of timesteps was enough for the trajectory to visit the important regions of the potential . the reconstructed free energy from this calculation is compared in fig .
[ fig:6 ] to the original one .
the metadynamics result is also compared in fig . [ fig:6b ] to the one obtained by the single - sweep method with @xmath80 centers .
the error ( [ eq : error1 ] ) for this metadynamics calculation was @xmath115 , i.e. almost two orders of magnitude higher than with the single - sweep method .
[ fig : muelmetatrj1 ] shows the original and reconstructed mueller potential along the metadynamics trajectory ( blue and red lines , respectively ) , together with the absolute value of their difference ( green line ) . by comparing figs .
[ fig : muelsweeptrj1 ] and [ fig : muelmetatrj1 ] , it can be seen that the discrepancy between the original and reconstructed potential is much larger with metadynamics than with the single - sweep method on trajectories of the same length .
note that these results clearly indicates that it is not sufficient that the metadynamics trajectory visits once a region on phase space to get an accurate representation of the free energy in this region .
this was already noted in refs .
@xcite .
steps with a time - step of @xmath74 ( same as in figs.[fig:1 ] and [ fig:2 ] ) .
we only use 10 level sets evenly distributed between @xmath77 and @xmath78 because the differences between the maps are much bigger than with the single - sweep method and drawing more level sets makes the figure difficult to read .
( there are only seven black level sets in the energy reconstructed by metadynamics because it levels off around @xmath116.),width=312 ] steps and a time - step of @xmath74 .
other representation of these contourplots were already shown in figs .
[ fig:2 ] and [ fig:6 ] respectively .
the map reconstructed by the single - sweep method is very close to the map of the original mueller potential .
the colormaps used in both panels are the same and the reconstructed potentials are shifted so that their minimum is @xmath77 ; the white region in the left panel is where the energy is above @xmath96 and is not shown ( the result of metadynamics shown in the right panel levels off around @xmath116 which is why there is no white).,width=312 ] steps and a time - step of @xmath74 .
other representation of these contourplots were already shown in figs .
[ fig:2 ] and [ fig:6 ] respectively .
the map reconstructed by the single - sweep method is very close to the map of the original mueller potential .
the colormaps used in both panels are the same and the reconstructed potentials are shifted so that their minimum is @xmath77 ; the white region in the left panel is where the energy is above @xmath96 and is not shown ( the result of metadynamics shown in the right panel levels off around @xmath116 which is why there is no white).,width=312 ] we were able to improve the metadynamics result by extending the simulation to @xmath117 steps .
the covering of the important regions in the potential was now extensive ( data not shown ) , and the reconstructed potential ( data not shown ) looked visually better than the one obtained with the shorter trajectory . yet
the error ( [ eq : error1 ] ) was @xmath118 , i.e. still two orders of magnitude larger than the highest error we obtained with the single - sweep method using a 10 times shorter trajectory .
we did not attempt to go to longer runs with metadynamics because the memory term in ( [ eq : zzmeta ] ) makes such simulations increasingly expensive ( their cost scales as the square of the number of timesteps ) .
it should also be stressed that in all these calculations , we optimized the parameters @xmath119 , @xmath27 and @xmath6 the best we could .
this optimization , however , turns out to be complicated since there is no systematic way to perform it because , unlike with the single - sweep method , there is no objective function to minimize in metadynamics .
the results shown in figs . [ fig : muelmetaxy][fig : muelmetatrj1 ] were obtained with @xmath120 , @xmath121 and @xmath122 . to be fair , we should conclude this comparison by mentioning that the simplicity of the mueller potential example tends to exaggerate the gain that the single sweep method provides over metadynamics .
indeed , in realistic situations , the single - sweep method also requires to compute the mean force via ( [ eq : meanforceapprox ] ) , an operation which was unnecessary in the mueller example since the force was readily available . computing the mean force adds an extra cost to the method .
it is worth stressing again , however , that the computation of the time averages in ( [ eq : meanforceapprox ] ) can be distributed over several processors .
this means that in the ideal situation where the user has at least one processor per center , the effective time to compute all of the mean forces is the same as the one for computing a single one of these forces , i.e. we have perfect scalability .
metadynamics can be parallelized per replica as well , as was proposed in refs .
@xcite , but not as straightforwardly and not with perfect scalability .
indeed , in all of these versions of metadynamics , the simulated replica are never completely independent from each other .
steps metadynamics shown in fig .
[ fig : muelmetaxy ] .
the green line shows the absolute value of the difference between the two .
the green line here should be compared with the one in fig .
[ fig : muelsweeptrj1 ] for the single - sweep method : the discrepancy between the original and reconstructed potential is always larger with metadynamics than with the single - sweep method.,width=312 ]
in this section , we use the single - sweep method to reconstruct the free energy of the solvated alanine dipeptide ( ad ) molecule in two and four torsion angles at @xmath123 k. while ad is not an example of biochemical interest _ per se _ , we study it because it has been extensively used as a benchmark example for free energy calculations in the literature @xcite . on top of this
the system is simple enough that we can use it to systematically investigate how the accuracy of the reconstruction method depends on the number of centers and how robust the method is with respect to statistical errors in the input data for the mean forces .
another question we investigate in this section is the robustness of the method against the choice of radial - basis functions . specifically , we compare results obtained using the gaussian packet ( [ eq : kernel ] ) and the wendland function @xmath124 where @xmath125 if @xmath126 , and @xmath127 otherwise .
( [ eq : wend ] ) is another well - known example of radial - basis function which has the pleasant property that it is compactly supported .
this property is appealing in the calculations since it limits the range over which centers interact in ( [ eq : objective ] ) .
all md simulations reported below were performed with a version of the moil code @xcite suitably modified by us , and the amber / opls @xcite force field ( for details of the md set - up see appendix [ sec : md ] ) .
we use the standard dihedral angles @xmath128 and @xmath129 . at @xmath123 k ,
the system is confined in a region of the @xmath130 space with @xmath131 by energy barriers higher than @xmath6 . in order to overcome these barriers and sweep through the whole @xmath132
^ 2 $ ] space , we generated a trajectory by using ( [ eq : tamd ] ) with @xmath133 , @xmath134 kcal / mol / rad@xmath135 , a friction coefficient @xmath136 kcal@xmath137ps / mol / rad@xmath135 and an artificial temperature @xmath138 kcal / mol . with this choice of the parameters , the important regions of the @xmath132 ^ 2 $ ] space
were visited in @xmath139 steps ( @xmath140 ps in the time units of the md variables ) .
the time series of @xmath128 and @xmath129 along this trajectory are shown in fig .
[ fig : adtrj ] . variations in @xmath59 , @xmath61 and @xmath141 led to qualitatively similar tamd trajectories , indicating that the the method is robust with regard to the choice of these parameters .
sets with a different number @xmath33 of centers were deposited along the tamd trajectory afterwards by processing this trajectory using various distances @xmath67 between the centers . specifically , we generated sets of @xmath142 , @xmath143 , @xmath144 , @xmath145 , @xmath146 and @xmath147 centers using , respectively , @xmath148 , @xmath149 , @xmath150 , @xmath151 , @xmath152 , @xmath153 . and @xmath129 along the @xmath140 ps long tamd trajectory for the solvated alanine dipeptide ( ad ) . , width=312 ]
given a set of @xmath33 centers @xmath154 , we computed the mean forces via @xmath33 independent md simulations with restraints at @xmath155 , i.e. by simulating ( [ eq : tamd2 ] ) in the isokinetic ensemble at @xmath123 k , and estimated the mean force via ( [ eq : meanforceapprox ] ) with @xmath156 kcal / mol / rad@xmath135 .
this value of @xmath157 was high enough since we checked that the reconstructed free energy remained invariant with higher values of @xmath157 ( we did so up to @xmath158 kcal / mol / rad@xmath135 ) .
we then used this data in the reconstruction procedure explained in sec .
[ sec : reconstruct ] .
note that since the free energy in the @xmath130 angle is periodic , we have to periodically extend the centers for the representation .
this amounts to changing the representation in ( [ eq : radialbasis ] ) into @xmath159 where @xmath160 is the unit vector in @xmath21 . in practice ,
only few periodic replica of the centers are needed ( i.e. @xmath161 for @xmath162 ) because the radial - basis functions centered at the centers further away from the cell under consideration make negligible contributions to the result in this cell .
we first detail our result with this choice of parameters to pick an example which led to a good balance between accuracy and efficiency .
other choices of parameters are discussed below .
thus , fig .
[ fig:7 ] shows the reconstructed free energy map obtained with @xmath163 ( @xmath145 centers ) and by computing the mean forces from ( [ eq : meanforceapprox ] ) with @xmath164 ps .
the optimal @xmath27 in this calculation was @xmath165 . in the figure
, the minimum of the free energy is set at @xmath166 kcal / mol , and contour levels are plotted at @xmath167 kcal / mol , @xmath168 kcal / mol and then every @xmath168 kcal / mol .
the centers are represented as white circles , and the mean forces at the centers as arrows . and @xmath129 dihedral angles at @xmath123 k calculated with the single - sweep method by using @xmath145 centers deposited at a distance of @xmath163 from each other .
units for the free energy are kcal / mol , and contour levels are plotted at @xmath167 kcal / mol , @xmath168 kcal / mol , and then every @xmath168 kcal / mol .
the optimal @xmath27 in this reconstruction was @xmath169 .
the centers are represented as white circles .
at every center , the corresponding mean force vector is also shown .
mean forces were calculated by using ( [ eq : meanforceapprox ] ) with @xmath156 kcal / mol / rad@xmath135 and @xmath164 ps.,width=312 ] since the free energy map depends on the force field used , comparison with results in the literature is difficult . to assess the accuracy of our result self - consistently , we compared it with the free energy calculated by computing the pdf of @xmath128 and @xmath129 from a direct md simulation ( dmds ) of about @xmath170 ns .
while this trajectory does not cover all the @xmath132 ^ 2 $ ] space , it covers the important regions and allows for an unbiased estimation of the free energy in these regions which can be used as benchmark .
the left panel in fig .
[ fig:8 ] shows the contour levels of the free energy from the single - sweep ( black lines ) and that from dmds ( red lines ) .
the contour levels are plotted at @xmath171 kcal / mol ( dotted lines ) , every @xmath167 kcal / mol from @xmath167 to @xmath172 kcal / mol , and then every @xmath173 kcal / mol . as can be seen , single - sweep results agree remarkably well with those of the dmds . in terms of cost , to generate the result shown in fig .
[ fig:7 ] , we had to make one simulation run of @xmath140 ps to generate the tamd trajectory , plus @xmath145 independent runs of @xmath174 ps distributed on different nodes ( the additional cost of estimating the parameters @xmath35 and @xmath44 to use in ( [ eq : radialbasis ] ) is insignificant ) .
this makes for a total of @xmath175 ns of absolute simulation time .
however , after distribution , the effective simulation time needed is only @xmath142 ps . on top of this , we show below that a good estimate of the free energy can be obtained with as low as 90 centers ( i.e. with an absolute simulation time of @xmath176 ns and the same effective simulation time , @xmath142 ps ) . for comparison , in ref .
@xcite ensing _ et al .
_ report a @xmath172 ns calculation performed with metadynamics to estimate the free energy of ad in @xmath128 and @xmath129 .
it is not clear how well this metadynamics calculation can be parallelized to reduce its effective cost ( it was not parallelized in ref .
in addition , the result of this metadynamics calculation is unlikely to be as accurate as the one in fig .
[ fig:7 ] ( in ref . @xcite
no comparison like the one shown in fig .
[ fig:8 ] is provided ) centers , the right panel the one with @xmath147 centers .
the contour levels of the free energy are plotted at @xmath171 kcal / mol ( dotted lines ) , from @xmath167 to @xmath177 kcal / mol separated by @xmath167 kcal / mol , and then separated by @xmath173 kcal / mol.,width=325 ] next we analyze how robust are the results with respect to the statistical error in the mean force data and the choice of radial - basis function .
we also analyze convergence in function of the number of centers . as reference value , we take the free energy reconstructed with @xmath178 ( @xmath147 centers ) and @xmath179 ps of time averaging in ( [ eq : meanforceapprox ] ) .
the map of the free energy calculated with these parameters ( data not shown ) is visually very similar to the one shown in fig .
[ fig:7 ] , but it is more accurate . the residual error can be estimated from the right panel of fig .
[ fig:8 ] which shows the contour levels of the free energy from the single - sweep ( black lines ) and that from dmds ( red lines ) : these level sets coincide up to statistical errors in the dmds , indicating that the free energy provided by the single sweep method with @xmath147 centers and @xmath179 ps can indeed be taken as an `` exact '' benchmark .
and @xmath129 reconstructed with gaussian functions using @xmath147 centers and @xmath179 ps . the figures in the different panels correspond to various number of centers and length of time averaging for the mean forces , as indicated .
units are kcal / mol .
note that the scale of the colormap is different from the one in fig .
[ fig:7 ] . in particular , the differences are mostly below @xmath167 kcal / mol with @xmath144 centers and @xmath174 ps simulations already.,width=312 ] fig .
[ fig : diffsad ] shows the differences between the map of the reference free energy reconstructed with @xmath178 ( @xmath147 centers ) and @xmath179 ps and those reconstructed with less centers and shorter restrained simulations .
the largest errors are in the regions corresponding to the highest peaks of the free energy ( these are also the regions were the least centers were deposited ) .
the differences never exceed @xmath180 kcal / mol with @xmath142 centers and @xmath164 ps and they fall mostly below @xmath167 kcal / mol with @xmath144 centers and @xmath164 ps already .
we also compared the quality of the reconstruction of the free energy when using gaussian ( [ eq : kernel ] ) and wendland ( [ eq : wend ] ) basis functions .
[ fig : resad ] shows the residual per center versus the number of centers for ad , by using gaussian ( filled symbols ) and wendland ( empty symbols ) basis functions , using @xmath164 ps ( diamonds ) and @xmath179 ps ( circles ) long restrained simulations to estimate the mean forces . at equal values of @xmath33 and @xmath181 , the reconstruction is slightly more accurate with gaussian than with wendland functions , though these differences turn out to be quite small in terms of the free energy maps themseves ( data not shown ) . and @xmath129 angles .
data are from calculations with different time averaging length , using gaussian ( g ) and wendland ( w ) basis functions.,width=312 ] using longer simulations for the mean force ( which means a smaller random error on these forces ) also improves the results . with @xmath147 centers and @xmath179 ps long simulations for the mean forces , the maps reconstructed with gaussian and wendland basis functions were almost identical ( data not shown ) , and they were not significantly different from the map shown in fig .
[ fig:7 ] .
these results , however , were obtained at very different values of optimal @xmath27 : @xmath182 with gaussians and @xmath183 with wendland functions .
the condition numbers at the optimal @xmath27 were @xmath184 and @xmath185 , respectively , which is low .
note that the same trends here described were observed in a periodic test case for which the potential was known exactly ( data not shown ) . as a second more challenging test , we computed the free energy of ad in the four torsion angles , @xmath128 , @xmath129 , @xmath186 , and @xmath187 .
a tamd trajectory of @xmath188 ps was generated by using ( [ eq : tamd ] ) with @xmath189 , @xmath134 kcal / mol / rad@xmath135 , an artificial temperature @xmath190 kcal / mol , friction coefficients @xmath136 kcal@xmath137ps / mol / rad@xmath135 for @xmath128 and @xmath129 and @xmath83 kcal@xmath137ps / mol / rad@xmath135 for @xmath186 and @xmath187 .
the md potential keeps the amide planes in _
trans _ configuration , and so @xmath186 and @xmath187 were varying in the range @xmath191 $ ] . in @xmath188
ps , the tamd trajectory covered well the accessible state space for @xmath128 , @xmath129 , @xmath186 and @xmath187 , in the sense that the time series for these angles were similar in their respective state space to those shown in fig .
[ fig : adtrj ] ( notice however that extensive coverage of the four - dimensional space is unlikely in so short a run ) . along the tamd trajectory ,
@xmath192 centers at a distance of @xmath193 were deposited . at these centers , the mean forces @xmath40 were computed by using ( [ eq : meanforceapprox ] ) with @xmath156 kcal / mol / rad@xmath135 and @xmath164 ps ( i.e. the absolute time of simulation was about @xmath194 ns , but the effective time after distribution was @xmath195 ps only ) .
we used these @xmath40 in the objective function ( [ eq : objective ] ) to finally get the representation ( [ eq : radialbasis ] ) of the four - dimensional free energy @xmath196 .
the optimal @xmath27 in this representation was @xmath197 and the condition number at this value of @xmath27 was @xmath198 . , @xmath129 , @xmath186 angles obtained from the marginal in these angles of the pdf associated with the free energy in four angles @xmath196 .
data are represented for @xmath199 $ ] .
note that the scale of the colormap is different from the one in fig .
[ fig:7],width=312 ] since a full graphical representation of @xmath200 is not possible , we did several tests to validate our result .
[ fig:9 ] shows the three dimensional free energy @xmath201 obtained from the marginal in these angles of the pdf associated with @xmath200 .
this marginal was calculated _ a posteriori _ by numerical integration over @xmath187 of @xmath202 with the full @xmath200 reconstructed by the single - sweep method .
the map is reasonable , and shows nontrivial features in all three directions .
[ fig:10 ] shows the two dimensional free energy @xmath203 obtained from the marginal in these angles of the pdf associated with @xmath196 .
the map is in remarkably good agreement with the one in fig .
[ fig:7 ] . as a further test of accuracy ,
we re - calculated the mean force using ( [ eq : tamd2 ] ) and ( [ eq : meanforceapprox ] ) using a different set of centers than those used in ( [ eq : radialbasis ] ) .
then we estimated the relative error between these mean forces and the ones obtained by taking the negative gradient of the reconstructed @xmath196 using the original set of centers and mean forces : @xmath204 where @xmath205 are the new centers and @xmath206 are the mean forces at these centers .
the new centers were 20 points chosen at random in the domains @xmath207 $ ] , @xmath208 $ ] . , @xmath129 angles obtained from the marginal in these angles of the probability density associated with the free energy in four angles @xmath196 .
contour levels are as in fig .
[ fig:7 ] .
note the remarkable agreement between this map and the one shown in fig .
[ fig:7].,width=312 ] fig .
[ fig : randtest ] , top panel , shows , for each of these centers , the distance from the closest of the @xmath192 centers ( black line ) .
data are compared to the minimal distance between the @xmath192 centers ( red dashed line ) .
this result shows that , with @xmath193 , these centers fill properly the four dimensional domain in the sense that every new center is always a distance about @xmath67 to one of the original centers .
[ fig : randtest ] , middle panel , shows the relative error @xmath209 for @xmath210 ( black solid line ) , when @xmath164 ps long restrained simulations are used to compute the mean forces .
the mean value of @xmath209 ( black dashed line ) is @xmath211 , with standard deviation @xmath212 and maximum value @xmath213 . for comparison , the mean value of the relative residual per center ( red dashed line ) is @xmath212 , with standard deviation @xmath214 and maximum value @xmath215 ( red dashed - dotted line ) .
[ fig : randtest ] , bottom panel , shows @xmath209 when @xmath179 ps long restrained simulations are used to compute the mean forces . in this case , the mean value of @xmath209 ( black dashed line ) is @xmath216 , with standard deviation @xmath217 and its maximum value is @xmath218 . for comparison ,
the mean value of the relative residual per center ( red dashed line ) is @xmath212 , with standard deviation @xmath214 and maximum value @xmath219 ( red dashed - dotted line ) .
these results show that , in points away from the original centers , the reconstructed free energy is as accurate as it is at the centers , which is clearly the best we can hope for . centers ( black line ) , compared with the minimal distance between the @xmath192 centers ( red line ) .
middle and lower panel , relative error @xmath209 defined in ( [ eq : relerror ] ) for mean forces computed respectively from @xmath174 and @xmath220 ps restrained simulations : the error @xmath209 ( black solid line ) and its mean value ( black dashed line ) , compared with the mean value of the relative residual per center for the @xmath192 centers set ( red dashed line ) and its maximum value ( red dashed - dotted line).,width=312 ] centers ( black line ) , compared with the minimal distance between the @xmath192 centers ( red line ) .
middle and lower panel , relative error @xmath209 defined in ( [ eq : relerror ] ) for mean forces computed respectively from @xmath174 and @xmath220 ps restrained simulations : the error @xmath209 ( black solid line ) and its mean value ( black dashed line ) , compared with the mean value of the relative residual per center for the @xmath192 centers set ( red dashed line ) and its maximum value ( red dashed - dotted line).,width=312 ]
in summary , we have proposed a method for the calculation of free energies which is simple , accurate , and efficient .
unlike standard histogram methods such as wham and metadynamics , the single - sweep method uses the mean force computed at a set of centers to reconstruct the free energy .
this set of centers is determined using tamd to rapidly sweep through the important regions of the free energy , and the mean forces at these centers are estimated in a standard way via the computation of a conditional expectation using time - averaging along restrained or constrained simulations . from these data , the free energy @xmath7
is then reconstructed globally by minimization of an objective function to determine the coefficients in a radial - basis function representation of @xmath7 .
if convenient , this reconstruction step can use data for the centers and the mean forces obtained by other means than tamd .
compared with histogram methods and metadynamics , the single - sweep technique combines several advantages : * it does not require _ a priori _ knowledge of the free energy since it uses tamd to find the important regions in the landscape automatically .
* the most costly step of the calculation , namely the computation of the mean forces at the centers , can be straightforwardly distributed on different , independent , processors . *
the reconstruction step is variational , i.e. the optimal coefficients in the free energy representation are determined automatically , which limits the number of parameters to adjust beforehand . *
the results can be easily monitored for convergence , and systematically improved if desired .
in particular , new centers can be added on top of previous ones along the same tamd trajectory to increase the accuracy without having to repeat the previous calculation . *
the method can be used in more than 2 dimensions and its computational complexity is the same regardless of the dimension .
we believe that these features make the single - sweep method appealing to calculate the free energy of systems more complicated but also more interesting than the ones studied in this paper .
we thank giovanni ciccotti and david chandler for carefully reading the manuscript ; weinan e for pointing out sparse grid methods which prompted us to test our method in four dimensions ; ron elber and anthony west for their help with the moil code ; sara bonella , simone meloni , michele monteferrante and maddalena venturoli for useful discussions ; and finally , eric darve for suggesting the test using ( [ eq : relerror ] ) .
this work was partially supported by nsf grants dms02 - 09959 and dms02 - 39625 , and by onr grant n00014 - 04 - 1 - 0565 .
all md simulations were performed with the moil code @xcite , and the amber / opls @xcite force field as implemented in the code .
a starting structure for the ad molecule ( ch@xmath221-co - nh - c@xmath222hch@xmath221-co - nh - ch@xmath221 ) was solvated in a box of @xmath223 water molecules of volume @xmath224 .
periodic boundary conditions were used .
van der waals interactions were truncated at @xmath225 .
electrostatic interactions were treated with the particle mesh ewald method @xcite with real space cutoff @xmath225 , a grid of @xmath226 points , and @xmath172-th order _
b_-splines for the interpolation of the structure factor ( in order to be in the high accuracy range @xcite ) .
the tip3 model @xcite was used for the water molecules .
non - bonded interaction lists were updated every @xmath194 steps .
all chemical bonds in the system were kept fixed with the shake algorithm @xcite .
amide planes were restrained to be always in _
trans _ configuration .
the velocity verlet algorithm was used for the dynamics of the cartesian variables with time - step @xmath168 fs , and all velocities were scaled at every step to keep the temperature at @xmath123 k. in order to obtain the initial configuration for the temperature accelerated md simulation ( tamd ) , the system was first equilibrated for @xmath88 ps by keeping the solute molecule fixed ( i.e. by zeroing forces and velocities of its atoms ) and by assigning to all water atoms at every step velocities sampled from a maxwell distribution at @xmath123 k. then , the whole system was simulated for @xmath227 ps for equilibration .
the torsion angles used in the simulations are defined by the quadruplets of atoms ( c , n , c@xmath222,c ) and ( n , c@xmath222,c , n ) for @xmath128 and @xmath129 , and ( o , c , n , c@xmath222 ) and ( c@xmath222,c , n , h ) for @xmath186 and @xmath187 . in the tamd simulation ,
the equations of motion of the collective variables were integrated with the forward euler scheme with time - step @xmath168 fs , @xmath136 kcal@xmath137ps / mol / rad@xmath135 for @xmath128 and @xmath129 and @xmath83 kcal@xmath137ps / mol / rad@xmath135 for @xmath186 and @xmath187 .
the force constant for the restraint potential was @xmath134 kcal / mol / rad@xmath135 , and the effective temperature such that @xmath228 kcal / mol .
the cartesian coordinates of water and ad atoms were saved during the tamd simulation . in this way , for every center @xmath29 deposited along the trajectory in collective variables space
, there is a corresponding configuration @xmath229 of the system in cartesian space such that @xmath230 .
we used these configurations as initial conditions for the restrained simulations at the centers .
data for the mean force calculations were accumulated after further relaxation of the system for @xmath231 ps . | a simple , efficient , and accurate method is proposed to map multi - dimensional free energy landscapes .
the method combines the temperature - accelerated molecular dynamics ( tamd ) proposed in [ maragliano & vanden - eijnden , chem .
. lett . * 426 * , 168 ( 2006 ) ] with a variational reconstruction method using radial - basis functions for the representation of the free energy .
tamd is used to rapidly sweep through the important regions of the free energy landscape and compute the gradient of the free energy locally at points in these regions .
the variational method is then used to reconstruct the free energy globally from the mean force at these points .
the algorithmic aspects of the single - sweep method are explained in detail , and the method is tested on simple examples , compared to metadynamics , and finally used to compute the free energy of the solvated alanine dipeptide in two and four dihedral angles . |
figure [ fig1 ] illustrates the general features of an epidemic outbreak on a population structured in different communities . starting from an index case a disease spreads widely inside a community thanks to the frequent intra - community interactions .
in addition the disease is transmitted to other communities via individuals belonging to different communities .
while the inter - community interactions may be rare they are determinant to understand the overall outbreak progression . based on this picture
i divide the population in two types or classes .
the _ locals _ belonging to a single community and the social _ bridges _ belonging to different communities . in a first approximation
i assume that ( i ) all communities are statistically equivalent , ( ii ) the mixing between the local and bridges is homogeneous , and ( iii ) social bridges belong to two populations . while these assumptions are off course approximations they allow us to gain insight into the problem . they could be relaxed in future works to include other factors such as degree correlations among interacting individuals @xcite and more realistic mixing patterns @xcite .
an epidemic outbreak taking place inside a community is then modeled by a a multi - type branching process @xcite starting from an index case ( see fig . [ fig1 ] ) .
the key intra - community magnitudes are the reproductive number and the generation times @xcite .
the reproductive number is the average number of secondary cases generated by a primary case .
the disease transmission introduces some biases towards individuals that interact more often .
therefore , i make an explicit distinction between the index case and other primary cases and denote their expected reproductive numbers by @xmath0 and @xmath1 , respectively .
the generation time @xmath2 is the time elapse from the infection of a primary case and the infection of a secondary case .
it is a random variable characterized by the generation time distribution function @xmath3 .
these magnitudes can be calculated for different models such as the susceptible infected recovered ( sir ) model and they can be estimated from empirical data as well .
finally , a community outbreak is represented by a causal true rooted at the index case @xcite . in this tree
the generation of an infected case is given by the distance to the index case .
furthermore , the tree can have at most @xmath4 generations , where @xmath4 is the average distance between individuals inside a community . #
1#2#3#4#5 @font ( 4350,4045)(0,-10 ) ( 750,2058)(4050,2058 ) ( 3690.000,1968.000)(4050.000,2058.000)(3690.000,2148.000 ) ( 750,2058)(4050,258 ) ( 3690.861,351.376)(4050.000,258.000)(3777.054,509.397 ) ( 750,2058)(4050,3858 ) ( 3777.054,3606.603)(4050.000,3858.000)(3690.861,3764.624 ) ( 4050,258 ) ( 4050,258 ) ( 0,2358)(0,0)[lb]@xmath5 ( 2550,3558)(0,0)[lb]@xmath6 ( 2550,1158)(0,0)[lb]@xmath7 ( 4350,3708)(0,0)[lb]@xmath8 ( 4350,1908)(0,0)[lb]@xmath8 ( 2550,2208)(0,0)[lb]@xmath6 ( 750,2058 ) ( 750,2058 ) ( 4350,108)(0,0)[lb]@xmath9 ( 4050,2058 ) ( 4050,2058 ) ( 4050,3858 ) ( 4050,3858 )
let us focus on a primary case at generation @xmath10 and its secondary cases at the following generation ( see fig .
[ fig2 ] ) .
let @xmath5 denote the expected number of descendants of the primary case at generation @xmath10 .
in particular @xmath11 gives the expected number of descendants from the index case , i.e. the expected outbreak size . in turn , @xmath12 is the expected number of descendants generated by a local secondary case at generation @xmath13 .
otherwise , if the secondary case is a bridge , it starts a new outbreak in a different community with expected outbreak size @xmath11 .
putting together the contribution of locals and bridges we obtain the recursive equation @xmath14\ , & d=0\\ \\
( 1-\beta)\left [ 1 + \tilde{r } \int_0^tdg(\tau ) n_{d+1}(t-\tau)\right ] + \beta n_0(t)\ , & 0<d < d\\ \\ 1-\beta + \beta n_0(t)\ , & d = d\ .
\end{array } \right . \label{nn}\ ] ] iterating this equation from @xmath15 to @xmath16 we obtain @xmath17 where @xmath18^{d-1 } g^{\star d}(t ) \label{ft}\ ] ] and @xmath19 denotes the @xmath10-order convolution of @xmath20 , i.e. @xmath21 and @xmath22 .
@xmath23 represents the expected outbreak size inside a community at time @xmath24 and @xmath25 is the final expected outbreak size inside a community . when @xmath26 it coincides with the expected outbreak size inside a community @xcite .
when @xmath27 ( [ n0n0 ] ) provides a self - consistent equation to determine the overall expected outbreak size after taking into account the inter - community transmissions . to calculate @xmath11 i use the laplace transform method
consider the incidence @xmath28 and its laplace transform @xmath29 substituting the recursive equation ( [ n0n0 ] ) in ( [ nomega ] ) i obtain @xmath30 where @xmath31 the validity of ( [ nomega ] ) is restricted to @xmath32 values satisfying @xmath33 , resulting in different scenarios depending on the value of the parameter @xmath34 _ local outbreaks : _ when @xmath35 then @xmath36 is defined for all @xmath37 and @xmath38 is obtained inverting the laplace transform in ( [ nomega ] ) .
furthermore , since @xmath39 is defined from ( [ fomega1 ] ) it follows that @xmath38 decreases to zero when @xmath40 , i.e. the epidemic outbreak dies out . _
global outbreaks : _ when @xmath41 the incidence grows exponentially @xmath42 , where @xmath43 is the positive root of the equation @xmath44 these two scenarios are equivalent to those obtained for a single community @xcite .
@xmath45 represents the effective community s reproductive number and the threshold condition @xmath46 delimits the local and global scenarios . to go beyond the final outbreak
i analyze the progression of the inter - communities outbreak .
i assume that the disease is transmitted at a constant rate @xmath47 from a primary case to a secondary case independently of their type . in this case
the intra - community incidence is given by @xcite @xmath48 for @xmath49 , where @xmath50 calculating the inverse laplace transform of ( [ nomega ] ) i finally obtain @xmath51}\ , \label{ntgamma}\ ] ] as a function of time in units of the local disease transmission rate @xmath47 , for @xmath52 ( dash - dotted ) , 0.5 ( dotted ) , 0.9 ( dashed ) , 1.0 ( solid ) and 1.1 ( dash - dash - dotted ) .
the panels from top to bottom corresponds to different average distances @xmath4 between individuals inside a community.,width=336 ] where @xmath53 is the gamma function .
figure [ fig4 ] shows the progression of the incidence as obtained from ( [ ntgamma ] ) .
as predicted above , the outbreak dies out when @xmath35 while when @xmath41 it grows exponentially .
more important , the incidence exhibits oscillations at the early stages , their number increasing with increasing @xmath4 .
for example , we distinguish about two oscillations for @xmath54 while for @xmath55 several oscillations are observed .
these oscillations represent resurgent epidemics , which are often observed in real outbreaks @xcite and simulations @xcite . for the different usa states .
@xmath7 was computed as the ratio between the number of state out - immigrants and the total state population according to the 1995 - 2000 usa census ( http://www.census.gov ) .
@xmath56 was computed as the number of habitants living with aids according to the 2005 statistics published by the us department of health ( http://www.hhs.gov ) .
b ) and c ) aids incidence in the usa b ) and uk c ) by year , as reported by the us department of health and the uk health protection agency ( http://www.hpa.org.uk ) , respectively.,width=576 ]
to understand the relevance of these results in a real world scenario i analyze data reported for the aids epidemics .
first , i estimate the parameter @xmath45 determining the spreading regime , local or global .
figure [ fig3]a shows the value of @xmath45 across the usa by state . for most states
@xmath41 , reaching significantly large values for several states .
for example , @xmath45 exceeds 1,000 for california and new york .
these numbers indicate that the usa aids epidemics is in the global spread scenario ( @xmath41 ) , in agreement with the general believe .
second , i analyze the temporal evolution of the aids incidence .
[ fig3]b and c show the aids incidence in usa and uk by year , indicating a similar temporal pattern .
the epidemics started with an increasing tendency of the incidence which , after reaching a maximum , switched to a decreasing trend .
after some years , however , the epidemics resurges with a new incidence increase .
this picture coincides with the model predictions in fig .
therefore , a possible explanation of the observed multiple peaks is the existence of a community structure , which can be attributed to geographical location and other factors .
@xmath45 in ( [ betac ] ) represents the expected number of infected individuals leaving their community .
the numerical simulations reported in @xcite indicated the existence of a transition at @xmath57 , from local outbreaks when @xmath35 to global epidemics when @xmath58 .
i have demonstrated that there is indeed a phase transition at @xmath57 .
furthermore , the analytical solution provides an expression of @xmath45 as a function of the bridge s fraction and the intra - community expected outbreak size ( [ p0 ] ) .
@xmath45 represents a measure of the reproductive number at the inter - community level .
its value can be estimated from the expected outbreak size inside a community and the bridge s fraction . based on the resulting estimate
we can determine if an epidemics is in the local or global epidemics scenario and react accordingly .
the inter - community disease transmission is characterized by oscillations at the early stages which represents resurgent epidemics , the number of these resurgencies being determined by the characteristic distance between individuals within a community .
in essence , when @xmath4 is small the time scale characterizing the outbreak progression within a community is very small @xcite .
therefore , the time it takes to observe the infection of a social bridge is very small as well , resulting in the mixing between the intra- and inter - community transmissions .
in contrast , when @xmath4 is large it takes a longer time to observe the infection of a social bridge and by that time the intra - community outbreak has significantly developed . therefore , in this last case the outbreak within communities is partially segregated in time . when multi - agent models are not available these results allow us to evaluate the potential progression of an epidemic outbreak and consequently determine the magnitude of our response to halt it .
they are also valuable when a detailed metapolpulation model is available , funneling the search for key quantities among the several model parameters .
more important , this work open avenues for future analytical works that side by side with multi - agent models will increase our chances to control global epidemics .
vazquez , a. ( 2006_a _ ) .
causal tree of disease transmission and the spreading of infectious diseases . in _ discrete methods in epidemiology
70 , of _ dimacs series in discrete mathematics and theoretical computer science _
. 163179 . ams providence | our chances to halt epidemic outbreaks rely on how accurately we represent the population structure underlying the disease spread .
when analyzing global epidemics this force us to consider metapopulation models taking into account intra- and inter - community interactions .
recently watts _
et al _ introduced a metapopulation model which accounts for several features observed in real outbreaks [ watts _ et al _ , pnas 102 , 11157 ( 2005 ) ] . in this work
i provide an analytical solution to this model , enhancing our understanding of the model and the epidemic outbreaks it represents .
first , i demonstrate that depending on the intra - community expected outbreak size and the fraction of social bridges the epidemic outbreaks die out or there is a finite probability to observe a global epidemics .
second , i show that the global scenario is characterized by resurgent epidemics , their number increasing with increasing the intra - community average distance between individuals .
finally , i present empirical data for the aids epidemics supporting the model predictions .
human populations are structured in communities representing geographical locations and other factors leading to partial segregation .
this population structure has a strong impact on the spreading patterns of infectious diseases among humans , forcing us to consider metapopulation models making an explicit distinction between the intra- inter - community interactions @xcite .
the increase in model realism is paid , however , by an increase in model complexity .
detailed metapopulation models are difficult to build and as a consequence they are available for a few locations in the world @xcite or they cover a single route of global transmission @xcite .
recently watts _
et al _
@xcite introduced a simple metapopulation model making an explicit distinction between the intra- and inter - community interactions . in spite of the model simplicity it accounts for several features observed in real epidemic outbreaks .
in particular , the numerical results indicate the existence of a transition from local to global epidemics when the expected number of infected individuals changing community reaches one @xcite .
i go a step forward and provide an analytical solution to the watts _ et al _ metapopulation model .
i demonstrate that there is indeed a phase transition when the expected number of infected individuals changing community reaches one .
this analytical solution allow us to obtain a much deeper insight into the main features of global epidemic outbreaks . #
1#2#3#4#5 @font ( 10226,10535)(0,-10 ) ( 8713,2708 ) ( 835,9012)(1435,8712 ) ( 1193.505,8765.666)(1435.000,8712.000)(1247.170,8872.997 ) ( 1435,8712)(2035,8412 ) ( 1793.505,8465.666)(2035.000,8412.000)(1847.170,8572.997 ) ( 1435,8712)(2035,9012 ) ( 1847.170,8851.003)(2035.000,9012.000)(1793.505,8958.334 ) ( 835,9012)(1435,9312 ) ( 1247.170,9151.003)(1435.000,9312.000)(1193.505,9258.334 ) ( 2035,9012)(2635,9012 ) ( 2395.000,8952.000)(2635.000,9012.000)(2395.000,9072.000 ) ( 2035,9012)(2635,8712 ) ( 2393.505,8765.666)(2635.000,8712.000)(2447.170,8872.997 ) ( 2635,9012)(3235,9012 ) ( 2995.000,8952.000)(3235.000,9012.000)(2995.000,9072.000 ) ( 1435,9312)(2035,9312 ) ( 1795.000,9252.000)(2035.000,9312.000)(1795.000,9372.000 ) ( 2035,9312)(2635,9312 ) ( 2395.000,9252.000)(2635.000,9312.000)(2395.000,9372.000 ) ( 2035,9312)(2635,9612 ) ( 2447.170,9451.003)(2635.000,9612.000)(2393.505,9558.334 ) ( 2635,9312)(3235,9312 ) ( 2995.000,9252.000)(3235.000,9312.000)(2995.000,9372.000 ) ( 1435,9312)(2035,9612 ) ( 1847.170,9451.003)(2035.000,9612.000)(1793.505,9558.334 ) ( 2635,8712)(3235,8712 ) ( 2995.000,8652.000)(3235.000,8712.000)(2995.000,8772.000 ) ( 2635,8712)(3235,8412 ) ( 2993.505,8465.666)(3235.000,8412.000)(3047.170,8572.997 ) ( 3235,9012)(4435,9012 ) ( 4075.000,8922.000)(4435.000,9012.000)(4075.000,9102.000)(4183.000,9012.000)(4075.000,8922.000 ) ( 2035,8412)(235,5712 ) ( 359.808,6061.461)(235.000,5712.000)(509.577,5961.615)(374.784,5921.677)(359.808,6061.461 ) ( 2635,5712)(3835,5412 ) ( 3463.920,5412.000)(3835.000,5412.000)(3507.577,5586.626)(3590.524,5473.119)(3463.920,5412.000 ) ( 2035,4812)(535,2262 ) ( 639.953,2617.928)(535.000,2262.000)(795.101,2526.665)(662.769,2479.207)(639.953,2617.928 ) ( 5635,4812)(7435,2712 ) ( 7132.382,2926.761)(7435.000,2712.000)(7269.048,3043.904)(7271.001,2903.333)(7132.382,2926.761 ) ( 235,5712)(835,5412 ) ( 593.505,5465.666)(835.000,5412.000)(647.170,5572.997 ) ( 835,5412)(1435,5112 ) ( 1193.505,5165.666)(1435.000,5112.000)(1247.170,5272.997 ) ( 835,5412)(1435,5712 ) ( 1247.170,5551.003)(1435.000,5712.000)(1193.505,5658.334 ) ( 235,5712)(835,6012 ) ( 647.170,5851.003)(835.000,6012.000)(593.505,5958.334 ) ( 1435,5712)(2035,5712 ) ( 1795.000,5652.000)(2035.000,5712.000)(1795.000,5772.000 ) ( 1435,5712)(2035,5412 ) ( 1793.505,5465.666)(2035.000,5412.000)(1847.170,5572.997 ) ( 2035,5712)(2635,5712 ) ( 2395.000,5652.000)(2635.000,5712.000)(2395.000,5772.000 ) ( 835,6012)(1435,6012 ) ( 1195.000,5952.000)(1435.000,6012.000)(1195.000,6072.000 ) ( 1435,6012)(2035,6012 ) ( 1795.000,5952.000)(2035.000,6012.000)(1795.000,6072.000 ) ( 1435,6012)(2035,6312 ) ( 1847.170,6151.003)(2035.000,6312.000)(1793.505,6258.334 ) ( 2035,6012)(2635,6012 ) ( 2395.000,5952.000)(2635.000,6012.000)(2395.000,6072.000 ) ( 835,6012)(1435,6312 ) ( 1247.170,6151.003)(1435.000,6312.000)(1193.505,6258.334 ) ( 2035,5412)(2635,5412 ) ( 2395.000,5352.000)(2635.000,5412.000)(2395.000,5472.000 ) ( 2035,5412)(2635,5112 ) ( 2393.505,5165.666)(2635.000,5112.000)(2447.170,5272.997 ) ( 1435,5112)(2035,5112 ) ( 1795.000,5052.000)(2035.000,5112.000)(1795.000,5172.000 ) ( 1435,5112)(2035,4812 ) ( 1793.505,4865.666)(2035.000,4812.000)(1847.170,4972.997 ) ( 535,2262)(1135,1962 ) ( 893.505,2015.666)(1135.000,1962.000)(947.170,2122.997 ) ( 1135,1962)(1735,1662 ) ( 1493.505,1715.666)(1735.000,1662.000)(1547.170,1822.997 ) ( 1135,1962)(1735,2262 ) ( 1547.170,2101.003)(1735.000,2262.000)(1493.505,2208.334 ) ( 535,2262)(1135,2562 ) ( 947.170,2401.003)(1135.000,2562.000)(893.505,2508.334 ) ( 1735,2262)(2335,2262 ) ( 2095.000,2202.000)(2335.000,2262.000)(2095.000,2322.000 ) ( 1135,2562)(1735,2562 ) ( 1495.000,2502.000)(1735.000,2562.000)(1495.000,2622.000 ) ( 1735,2562)(2335,2562 ) ( 2095.000,2502.000)(2335.000,2562.000)(2095.000,2622.000 ) ( 1735,2562)(2335,2862 ) ( 2147.170,2701.003)(2335.000,2862.000)(2093.505,2808.334 ) ( 2335,2562)(2935,2562 ) ( 2695.000,2502.000)(2935.000,2562.000)(2695.000,2622.000 ) ( 1135,2562)(1735,2862 ) ( 1547.170,2701.003)(1735.000,2862.000)(1493.505,2808.334 ) ( 1735,1662)(2335,1662 ) ( 2095.000,1602.000)(2335.000,1662.000)(2095.000,1722.000 ) ( 1735,1662)(2335,1362 ) ( 2093.505,1415.666)(2335.000,1362.000)(2147.170,1522.997 ) ( 3835,5412)(4435,5112 ) ( 4193.505,5165.666)(4435.000,5112.000)(4247.170,5272.997 ) ( 4435,5112)(5035,4812 ) ( 4793.505,4865.666)(5035.000,4812.000)(4847.170,4972.997 ) ( 3835,5412)(4435,5712 ) ( 4247.170,5551.003)(4435.000,5712.000)(4193.505,5658.334 ) ( 5635,5412)(6235,5412 ) ( 5995.000,5352.000)(6235.000,5412.000)(5995.000,5472.000 ) ( 4435,5712)(5035,5712 ) ( 4795.000,5652.000)(5035.000,5712.000)(4795.000,5772.000 ) ( 5035,5712)(5635,5712 ) ( 5395.000,5652.000)(5635.000,5712.000)(5395.000,5772.000 ) ( 5035,5712)(5635,6012 ) ( 5447.170,5851.003)(5635.000,6012.000)(5393.505,5958.334 ) ( 5635,5712)(6235,5712 ) ( 5995.000,5652.000)(6235.000,5712.000)(5995.000,5772.000 ) ( 4435,5712)(5035,6012 ) ( 4847.170,5851.003)(5035.000,6012.000)(4793.505,5958.334 ) ( 5635,5112)(6235,5112 ) ( 5995.000,5052.000)(6235.000,5112.000)(5995.000,5172.000 ) ( 5635,5112)(6235,4812 ) ( 5993.505,4865.666)(6235.000,4812.000)(6047.170,4972.997 ) ( 5035,4812)(5635,4812 ) ( 5395.000,4752.000)(5635.000,4812.000)(5395.000,4872.000 ) ( 5035,4812)(5635,4512 ) ( 5393.505,4565.666)(5635.000,4512.000)(5447.170,4672.997 ) ( 4435,9012)(5035,8712 ) ( 4793.505,8765.666)(5035.000,8712.000)(4847.170,8872.997 ) ( 5035,8712)(5635,8412 ) ( 5393.505,8465.666)(5635.000,8412.000)(5447.170,8572.997 ) ( 5035,8712)(5635,9012 ) ( 5447.170,8851.003)(5635.000,9012.000)(5393.505,8958.334 ) ( 4435,9012)(5035,9312 ) ( 4847.170,9151.003)(5035.000,9312.000)(4793.505,9258.334 ) ( 5635,9012)(6235,9012 ) ( 5995.000,8952.000)(6235.000,9012.000)(5995.000,9072.000 ) ( 6235,9012)(6835,9012 ) ( 6595.000,8952.000)(6835.000,9012.000)(6595.000,9072.000 ) ( 5035,9312)(5635,9312 ) ( 5395.000,9252.000)(5635.000,9312.000)(5395.000,9372.000 ) ( 5635,9312)(6235,9312 ) ( 5995.000,9252.000)(6235.000,9312.000)(5995.000,9372.000 ) ( 5635,9312)(6235,9612 ) ( 6047.170,9451.003)(6235.000,9612.000)(5993.505,9558.334 ) ( 6235,9312)(6835,9312 ) ( 6595.000,9252.000)(6835.000,9312.000)(6595.000,9372.000 ) ( 5035,9312)(5635,9612 ) ( 5447.170,9451.003)(5635.000,9612.000)(5393.505,9558.334 ) ( 5635,8412)(6235,8412 ) ( 5995.000,8352.000)(6235.000,8412.000)(5995.000,8472.000 ) ( 5635,8412)(6235,8112 ) ( 5993.505,8165.666)(6235.000,8112.000)(6047.170,8272.997 ) ( 7435,2712)(8035,2412 ) ( 7793.505,2465.666)(8035.000,2412.000)(7847.170,2572.997 ) ( 8035,2412)(8635,2112 ) ( 8393.505,2165.666)(8635.000,2112.000)(8447.170,2272.997 ) ( 7435,2712)(8035,3012 ) ( 7847.170,2851.003)(8035.000,3012.000)(7793.505,2958.334 ) ( 8035,3012)(8635,3012 ) ( 8395.000,2952.000)(8635.000,3012.000)(8395.000,3072.000 ) ( 8635,3012)(9235,3012 ) ( 8995.000,2952.000)(9235.000,3012.000)(8995.000,3072.000 ) ( 8635,3012)(9235,3312 ) ( 9047.170,3151.003)(9235.000,3312.000)(8993.505,3258.334 ) ( 9235,3012)(9835,3012 ) ( 9595.000,2952.000)(9835.000,3012.000)(9595.000,3072.000 ) ( 8035,3012)(8635,3312 ) ( 8447.170,3151.003)(8635.000,3312.000)(8393.505,3258.334 ) ( 8635,2112)(9235,2112 ) ( 8995.000,2052.000)(9235.000,2112.000)(8995.000,2172.000 ) ( 8635,2112)(9235,1812 ) ( 8993.505,1865.666)(9235.000,1812.000)(9047.170,1972.997 ) ( 5035,4812)(5635,5112 ) ( 5447.170,4951.003)(5635.000,5112.000)(5393.505,5058.334 ) ( 5035,4812)(5635,5412 ) ( 5507.721,5199.868)(5635.000,5412.000)(5422.868,5284.721 ) ( 5713,9008 ) ( 2113,9008 ) ( 8413,6608 ) ( 1513,5708 ) ( 5335,1512 ) ( 1813,2258 ) ( 5113,5408 ) ( 9235,3312 ) ( 9235,3312 ) ( 2035,8412 ) ( 2035,8412 ) ( 835,9012 ) ( 835,9012 ) ( 1435,9312 ) ( 1435,9312 ) ( 2635,9012 ) ( 2635,9012 ) ( 3235,9012 ) ( 3235,9012 ) ( 2635,9312 ) ( 2635,9312 ) ( 3235,9312 ) ( 3235,9312 ) ( 2635,8712 ) ( 2635,8712 ) ( 3235,8712 ) ( 3235,8712 ) ( 3235,8412 ) ( 3235,8412 ) ( 2635,9612 ) ( 2635,9612 ) ( 2035,9612 ) ( 2035,9612 ) ( 2035,9312 ) ( 2035,9312 ) ( 2035,9012 ) ( 2035,9012 ) ( 835,5412 ) ( 835,5412 ) ( 1435,5112 ) ( 1435,5112 ) ( 235,5712 ) ( 235,5712 ) ( 835,6012 ) ( 835,6012 ) ( 2035,5712 ) ( 2035,5712 ) ( 2635,5712 ) ( 2635,5712 ) ( 2035,6012 ) ( 2035,6012 ) ( 2635,6012 ) ( 2635,6012 ) ( 2035,5412 ) ( 2035,5412 ) ( 2635,5412 ) ( 2635,5412 ) ( 2635,5112 ) ( 2635,5112 ) ( 2035,5112 ) ( 2035,5112 ) ( 2035,4812 ) ( 2035,4812 ) ( 2035,6312 ) ( 2035,6312 ) ( 1435,6312 ) ( 1435,6312 ) ( 1435,6012 ) ( 1435,6012 ) ( 1435,5712 ) ( 1435,5712 ) ( 1135,1962 ) ( 1135,1962 ) ( 1735,1662 ) ( 1735,1662 ) ( 535,2262 ) ( 535,2262 ) ( 1135,2562 ) ( 1135,2562 ) ( 2335,2262 ) ( 2335,2262 ) ( 2335,2562 ) ( 2335,2562 ) ( 2935,2562 ) ( 2935,2562 ) ( 2335,1662 ) ( 2335,1662 ) ( 2335,1362 ) ( 2335,1362 ) ( 2335,2862 ) ( 2335,2862 ) ( 1735,2862 ) ( 1735,2862 ) ( 1735,2562 ) ( 1735,2562 ) ( 1735,2262 ) ( 1735,2262 ) ( 4435,5112 ) ( 4435,5112 ) ( 5035,4812 ) ( 5035,4812 ) ( 3835,5412 ) ( 3835,5412 ) ( 4435,5712 ) ( 4435,5712 ) ( 5635,5412 ) ( 5635,5412 ) ( 6235,5412 ) ( 6235,5412 ) ( 5635,5712 ) ( 5635,5712 ) ( 6235,5712 ) ( 6235,5712 ) ( 5635,5112 ) ( 5635,5112 ) ( 6235,5112 ) ( 6235,5112 ) ( 6235,4812 ) ( 6235,4812 ) ( 5635,4812 ) ( 5635,4812 ) ( 5635,4512 ) ( 5635,4512 ) ( 5635,6012 ) ( 5635,6012 ) ( 5035,6012 ) ( 5035,6012 ) ( 5035,5712 ) ( 5035,5712 ) ( 5035,8712 ) ( 5035,8712 ) ( 5635,8412 ) ( 5635,8412 ) ( 4435,9012 ) ( 4435,9012 ) ( 5035,9312 ) ( 5035,9312 ) ( 6235,9012 ) ( 6235,9012 ) ( 6835,9012 ) ( 6835,9012 ) ( 6235,9312 ) ( 6235,9312 ) ( 6835,9312 ) ( 6835,9312 ) ( 6235,8412 ) ( 6235,8412 ) ( 6235,8112 ) ( 6235,8112 ) ( 6235,9612 ) ( 6235,9612 ) ( 5635,9612 ) ( 5635,9612 ) ( 5635,9312 ) ( 5635,9312 ) ( 5635,9012 ) ( 5635,9012 ) ( 8035,2412 ) ( 8035,2412 ) ( 8635,2112 ) ( 8635,2112 ) ( 7435,2712 ) ( 7435,2712 ) ( 8035,3012 ) ( 8035,3012 ) ( 9235,3012 ) ( 9235,3012 ) ( 9835,3012 ) ( 9835,3012 ) ( 9235,2112 ) ( 9235,2112 ) ( 9235,1812 ) ( 9235,1812 ) ( 1435,8712 ) ( 1435,8712 ) ( 8635,3312 ) ( 8635,3312 ) ( 8635,3012 ) ( 8635,3012 ) |
fast radio bursts ( frbs ) are intense radio flashes occurring at high galactic latitudes with anomalously high dispersion measure ( dm ) @xcite .
duo to the lack of distance information , their physical origin is unknown .
some people suggested that the high dm is dominated by the ionized intergalactic medium , which implies that frbs may occur at cosmological distances .
recently , keane et al .
( 2016 ) claimed to discover the first frb host galaxy , which is an elliptical galaxy at @xmath0 @xcite .
however , this conclusion was questioned by some subsequent papers @xcite .
more recently , using fast - dump interferometry with the karl g. jansky very large array ( vla ) , the host galaxy of repeating frb 121102 was discovered @xcite .
optical imaging and spectroscopy identify frb 121102 a redshift of @xmath1 @xcite .
the cosmological origin of frb 121102 is confirmed .
therefore frbs are promising cosmological probes .
however , the physical origin of frbs is mysterious until now .
many theoretical models for frbs are proposed , including collapses of supra - massive neutron star into black hole @xcite , magnetar pulse - wind interactions @xcite , charged black hole binary mergers @xcite , giant pulse emissions from pulsars @xcite , giant flares from magnetars @xcite , unipolar inductor model @xcite , and double neutron stars mergers @xcite .
the frb 121102 is repeating , which disfavors models involving cataclysmic events @xcite .
additional six bursts @xcite and nine bursts @xcite for frb 121102 are detected .
so there may be two populations of frbs @xcite .
dai et al .
( 2016 ) proposed that the repeating bursts are produced by lots of asteroids encountering with highly magnetized pulsar @xcite .
a neutron star - white dwarf binary model also has been proposed for the repeating frb 121102 @xcite .
until now , twenty six bursts of frb 121102 have been observed .
however , the nine bursts discovered by vla are not observed by arecibo observatory . in this paper
, we investigate the frequency distributions of peak flux , fluence , duration and waiting time for frb 121102 .
we also test the proposed models for frbs using the derived distributions .
this paper is organized as follows .
the frequency distributions are shown in section 2 . in section 3 ,
we test theoretical models using the statistical results .
finally , the conclusion and discussions are given in section 4 .
for frb 121102 , we use the parameters of eleven bursts from @xcite and six bursts from @xcite , which are listed in table 1 . because the nine bursts observed by vla in the 2.5 - 3.5 ghz @xcite , and these bursts are not detected by arecibo , only the upper limit is given .
these nine bursts are not considered in our analysis .
the eleven bursts in @xcite are discovered by william e. gordon telescope at the arecibo observatory and the 7-beam arecibo l - band feed array ( alfa ) .
the alfa is a seven - beam receiver operating at 1.4 ghz with 0.3 ghz bandwidth @xcite .
the antenna gains for these beams are different , i.e. , 10.4 k jy@xmath2 for the central beam at low zenith angles and 8.2 k jy@xmath2 for the other six beams @xcite . because the bursts could be detected by different beams , the observed flux or fluence must be corrected .
only the last six bursts are pointing to the central beam @xcite , so the fluxes and fluences of other five bursts are normalized to the central beam by multiplying a factor of @xmath3 .
the additional six bursts are observed by green bank telescope and the single - pixel l - wide receiver at arecibo observatory @xcite .
therefore , the fluxes of these bursts are intrinsic . for each bursts ,
column 2 gives the peak time of each burst listed in column 1 .
the peak flux is presented in column 3 in unit of jy .
column 4 gives the fluence @xmath4 of each burst in unit of jy ms .
the observed duration time of burst is given in column 5 .
the waiting time is given in column 6 .
the waiting time @xmath5 is defined as the difference of occurring times for two adjacent bursts , and can be calculated from the time difference of column 2 . only the continues observation is considered . when calculating the waiting time , the peak flux limit 0.02 jy is considered . because the detection threshold of alfa is about 0.02 jy @xcite .
the definition of waiting time is widely used in solar physics and astrophysics .
the number of bursts @xmath6 with fluence between @xmath4 and @xmath7 can be expressed by @xmath8 where @xmath9 is the power - law index .
the number of bursts for frb 121102 is small . rather than examining the differential distribution directly
, it is preferable to plot the cumulative distribution , which can avoid binning of the data . because the width of binning can affect the fitting result . integrating equation ( [ fluencedis ] )
, we obtain the cumulative distribution of fluence @xmath10 for the peak flux @xmath11 , the differential frequency distribution is @xmath12 so the number of frbs with peak flux larger than @xmath11 is @xmath13 we apply the markov chain monte carlo ( mcmc ) method to derive the best - fitting parameters . in astrophysical observations ,
count statistics is often limited .
the bursts of frb 121102 is 17 .
such low count number does not fulfill the condition required for the gaussian approximation , a well approximation is the poisson distribution .
consider the number of observed events @xmath14 following poisson distribution , the likelihood function for mcmc method can be expressed as @xmath15 where @xmath16 is the parameter in the model to be constrained by the observed data , @xmath17 is the @xmath18th observed data , and @xmath19 is the theoretical number predicted by model . for the cumulative distribution , it has @xmath20 .
therefore , the likelihood can be re - expressed as @xmath21 , where @xmath22 is the total number of observed events .
we use a python package pymc @xcite to apply the mcmc method to optimize the parameters of theoretical distributions . in the fitting
, we consider the priors of all the parameters @xmath16 as uniform distributions in a relatively large range , because the priors are not important when sampling enough samples with mcmc method .
we must note that the events in each bin of the differential distribution are independent , but the number of events @xmath23 in the cumulative distribution are statistically dependent .
fortunately , we use a logarithmic binning , the fluctuations of events for cumulative distribution in each bin , may follow approximately the same random statistics @xmath24 in each bin as for the differential distribution .
so the likelihood function of equation ( 5 ) may be a well approximation .
this problem has been extensively discussed in @xcite .
figure 1 shows the cumulative distributions of fluence ( left panel ) and peak flux ( right panel ) for seventeen bursts of frb 121102 , respectively .
the power - law index for fluence is @xmath25 with @xmath26 confidence level .
the value of @xmath9 is from 1.5 to 2.2 @xcite . while , for peak flux , the power - law index is @xmath27 with @xmath26 confidence level .
the differential distribution of duration time @xmath28 can be expressed as @xmath29 so the cumulative distribution of duration time @xmath28 is @xmath30 where @xmath31 is the maximal duration time .
the markov chain monte carlo ( mcmc ) method is also used to derive the best - fitting parameters simultaneously .
left panel of figure 2 presents the cumulative distribution of duration for frb 121102 . from this panel , a maximal duration time is obviously shown .
the best - fitting power - law index and maximal duration time are @xmath32 , and @xmath33 with @xmath26 confidence level , respectively .
it should be noted that the observed duration time will be broadened when radio waves propagate through a plasma .
the scatter - broadening time of a pulsed signal depends on the dm and the observing frequency , and an empirical function is given @xcite .
however , there is no clear evidence for scatter broadening of frb 121102 @xcite .
if the burst rate is constant , the waiting - time distribution is the poisson interval distribution @xcite @xmath34 where @xmath5 is the interval between events , and @xmath35 is the burst rate .
if the burst rate is time varying , the waiting time distribution can be treated as a combination of piecewise constant poisson processes .
generally , for most forms of @xmath36 , the waiting time distribution can be shown as power - law form @xcite @xmath37 in order to avoid binning of the data , the cumulative waiting time distribution is given by @xmath38 for frb 121102 , the observation is not continues @xcite .
the detailed observations by different telescopes are shown in figure 1 of @xcite .
therefore , in order to obtain reliable waiting times , we select the waiting times during periods of continuous observation .
we use the waiting times presented in table 1 of @xcite .
there are ten waiting times from tens to hundreds of seconds .
right panel of figure 2 shows the cumulative waiting time distribution of frb 121102 .
the best - fitting power - law index and maximal waiting time are @xmath39 , and @xmath40 s with @xmath26 confidence level .
the power - law distributions indicates the stochastic engine for frb 121102 .
there are many models proposed to explain the properties of frbs . in this section
, we will test theoretical models predictions with statistical results .
dai et al .
( 2016 ) proposed that the repeating bursts can be produced from lots of asteroids encountering with highly magnetized pulsar @xcite . in order to explain observation , the diameters of asteroids are small , i.e. , @xmath41 km @xcite . from their equation
( 2 ) , the differential frequency distribution of diameter @xmath42 of asteroids is predicted to @xmath43 , with duration time @xmath28 .
so if the index for differential frequency distribution of duration is @xmath44 , the value is @xmath45 . from the observation of sloan digital sky survey ,
the a broken power law was found with @xmath46 for large asteroids ( 5 - 50 km ) and @xmath47 for smaller asteroids ( 0.5 - 5 km ) @xcite .
the differential size distribution of small asteroids is @xmath48 @xcite .
these value are well consistent with the model prediction .
cordes and wasserman ( 2016 ) suggested that frbs originate from crab - like giant pulses of extragalactic neutron stars @xcite .
the index of peak flux cumulative distribution @xmath49 is from @xmath50 to @xmath51 .
the low limit is a little larger than the best - fitting value .
lyutikov et al .
( 2016 ) argued that frbs , including repeating and non - repeating frbs , are from giant pulses of young rapidly rotating pulsars @xcite . in their model ,
the intrinsic luminosity of an frb is proportional to the spin - down power of neutron star .
so the predicted distribution of frb flux is @xmath52 @xcite . from our statistical study of 17 bursts of frb 121102
, the cumulative distribution @xmath53 is found , which is different from their model prediction .
so the repeating frb 121102 may disfavor the rotationally powered model .
the distance measurement of the repeating frb 121102 also ruled out rotationally - powered radio emission @xcite .
katz ( 2016 ) proposed that frbs are generated by magnetic energy released in magnetar magnetospheres @xcite .
metzger et al . ( 2017 ) also argued that the repeating frb 121102 is powered by millisecond magnetar , through its rotational or magnetic energy @xcite . from observational constraint ,
the magnetic energy is favored @xcite . generally , the soft gamma repeater ( sgr ) outbursts result from the dissipation of magnetostatic energy in the magnetosphere of magnetars .
so we compare the statistical properties of frb 121102 and sgr 1806 - 20 .
figure 3 shows the differential distributions of duration ( left panel ) , and waiting time ( right panel ) for sgr 1806 - 20 .
we use the waiting time data from @xcite , and duration time data from @xcite .
the best - fitting indices are @xmath54 and @xmath55 for duration time and waiting time , respectively .
the distribution of sgr 1806 - 20 burst energies follows a power - law @xmath56 with @xmath57 @xcite .
these indices are well consistent with those of frb 121102 .
this may indicate that repeating frbs may be related to extremely magnetized neutron stars . besides these statistic distributions , there are some phenomenological similarity between frbs and sgrs .
first , they are both repeating .
at least , frb 121102 show repeating bursts @xcite .
second , the duty factor @xmath58 with flux @xmath59 is similar , @xmath60 for sgrs and @xmath61 for frbs @xcite .
this value denotes the the fraction of the time in which a source emits at close to its peak flux .
in this paper , we study the statistical properties of repeating frb 121102 , including peak flux , fluence , duration time and waiting time .
the cumulative distributions of peak flux , fluence and duration show power - law forms .
the waiting time distribution also shows power - law distribution , and is consistent with a non - stationary poisson process .
power - law size distributions have been discovered in many astrophysical phenomena , which may indicate a stochastic central engine .
we also compare the statistical results with theoretical models predictions .
the duration distribution from theoretical model relating asteroids encountering with highly magnetized pulsar is consistent with observations .
similar distributions between frb 121102 and sgr 1806 - 20 , such as fluence , duration and waiting time , also support the models proposed by @xcite , in which the magnetic energy releases in magnetar magnetospheres . so more observation is needed to distinguish these two models .
power - law distributions of events have been discovered in a large number of astrophysical phenomena in many wavelengths ( for a recent review , see * ? ? ?
the power - law frequency distributions , including peak flux , fluence , duration and waiting time , are predicted by self - organized criticality ( soc ) systems @xcite .
these distributions also satisfy the criteria that define a soc system @xcite , which occurs in many natural systems that exhibit nonlinear energy dissipation @xcite .
therefore , frbs may also be avalanche events . in future
, some facilities , such as chinese five - hundred - meter aperture spherical radio telescope ( fast ) @xcite , canadian hydrogen intensity mapping experiment ( chime ) @xcite , the square kilometer array , or other upcoming wide - field telescopes , will collect a large number of frbs .
the statistics of frbs will give constraints on the nature of central engine .
we thank the anonymous referee for useful comments and suggestions .
we also thank z. g. dai for helpful discussion .
this work is supported by the national basic research program of china ( 973 program , grant no .
2014cb845800 ) and the national natural science foundation of china ( grants 11422325 and 11373022 ) , the excellent youth foundation of jiangsu province ( bk20140016 ) .
lorimer , d. r. , bailes , m. , mclaughlin , m. a. , narkevic , d. j. , & crawford , f. 2007 , sci , 318 , 777 thornton , d. , stappers , b. , bailes , m. et al .
2013 , sci , 341 , 53 masui , k. , lin , h. , sievers , j. et al .
2015 , natur , 528 , 523 katz , j. i. 2016 , mod .
lett . a , 31 , 1630013 ravi , v. et al . , 2016 , science , 354 , 1249 keane , e. f. , johnston , s. , bhandari , s. et al . , 2016 , natur , 530 , 453 williams , p. k. g. & berger , e. , 2016 , apjl , 822 , l22 vedantham , h. k. , et al . , 2016 ,
apjl , 824 , l19 chatterjee , s. et al . , 2017 , nature , 541 , 58 tendulkar , s. p. et al .
, 2017 , apj , 834 , l7 falcke , h. , & rezzolla , l. 2014 , a&a , 562 , a137 zhang , b. 2014 , apjl , 780 , l21 ravi , v. , & lasky , p. d. 2014 , mnras , 441 , 2433 lyubarsky , y. 2014 , mnras , 442 , l9 zhang , b. 2016 , apjl , 822 , l14 cordes , j. m. & wasserman , i. , 2016 , mnras , 457 , 232 popov , s. b. & postnov , k. a. 2013 , arxiv : 1307.4924 .
kulkarni , s. r. , et al .
, 2014 , apj , 797 , 70 katz , j. i. 2014 , phys . rev .
d , 89 , 103009 pen , u. l. , & connor , l. 2015 , apj , 807 , 179 kulkarni , s. r. , ofek , e. o. & neill , j. d. 2016 , arxiv : 1511.09137 wang , j. s. , et al .
, 2016 , apjl , 822 , l7 totani , t. 2013 , pasj , 65 , l12 spitler , l. g. , et al .
2016 , natur , 790 , 101 scholz , p. , et al . ,
2016 , apj , 833 , 177 champion , d. j. , et al . 2016 , mnras , 460 , l30 li , l. b. et al . , 2017 , raa , 17 , 6 dai , z. g. , et al .
, 2016 , apj , 829 , 27 gu , w. m. , et al .
, 2016 , apjl , 823 , l28 cordes , j. m. et al . , 2006 ,
apj , 637 , 446 https://github.com/pymc-devs/pymc aschwanden , m. j. , 2015 , apj , 814 , 19 lu , w. b. & kumar , p. 2016 , mnras , 461 , l122 bhat n. d. r. , et al .
, 2004 , apj , 605 , 759 wheatland , m. s. , sturrock , p. a. & mctiernan , j. m. , 1998 , apj , 509 , 448 aschwanden , m. j. & mctiernan , j. m. , 2010 , apj , 717 , 683 ivezic , z. , et al . , 2001 , aj , 122 , 2749 yoshida , f. & nakamura t. 2007 , planet . space science , 55 , 113 lyutikov , m. , burzawa , l. & popov , s. b. , 2016 , mnras , 462 , 941 lyutikov , m. , 2017 , arxiv : 1701.02003 katz , j. i. 2016 , apj , 826 , 226 metzger , b. d. , berger , e. , margalit , b. 2017 , arxiv : 1701.02370 gogus , e. , woods , p. m. , kouveliotou , c. , et al .
2000 , apj , 532 , l121 gogus , e. , kouveliotou , c. , woods , p. m. , et al .
2001 , apj , 558 , 228 cheng , b. , et al .
1996 , nature , 382 , 518 prieskorn , z. & kaaret , p. , 2012 ,
apj , 755 , 1 law , c. j. , et al . , 2015 , apj , 807 , 16 aschwanden , m. j. , 2011 , self - organized criticality in astrophysics : the statistics of nonlinear processes in the universe , springer - verlag : berlin bak , p. , tang , c. , & wiesenfeld , k. 1987 , phys .
59 , 381 katz , j. i. 1986 , j. geophys . res .
91 , 10412 wang , f. y. & dai , z. g. , 2013 , nature physics , 9 , 465 wang , f. y. , dai , z. g. , yi , s. x. & xi , s. q. , 2015 , apjs , 216 , 8 nan , r. d. , li , d. , jin , c. , et al .
, 2011 , ijmpd , 20 , 989 kaspi , v. m. , 2016 , science , 354 , 1230 lllllllll no . & peak time & peak flux & fluence & width & waiting time + & mjd & jy & jy ms & ms & s + 1 & 56233.282837008 & 0.05 & 0.13 & [email protected] & + 2 & 57159.737600835 & 0.038 & 0.13 & [email protected] & + 3 & 57159.744223619 & 0.038 & 0.13 & [email protected] & 572.2 + 4 & 57175.693143232 & 0.05 & 0.25 & [email protected] & + 5 & 57175.699727826 & 0.025 & 0.11 & [email protected] & 568.9 + 6 & 57175.742576706 & 0.02 & 0.06 & [email protected] & + 7 & 57175.742839344 & 0.02 & 0.06 & [email protected] & 22.7 + 8 & 57175.743510388 & 0.14 & 0.9 & [email protected] & 58.0 + 9 & 57175.745665832 & 0.05 & 0.3 & [email protected] & 186.2 + 10 & 57175.747624851 & 0.05 & 0.2 & [email protected] & 169.3 + 11 & 57175.748287265 & 0.31 & 1.0 & [email protected] & 57.2 + 12 & 57339.356046005 & 0.04 & 0.2 & [email protected] & + 13 & 57345.447691250 & 0.06 & 0.4 & [email protected] & + 14 & 57345.452487925 & 0.04 & 0.2 & [email protected] & 414.4 + 15 & 57345.457595303 & 0.02 & 0.08 & [email protected] & 441.3 + 16 & 57345.462413106 & 0.09 & 0.6 & [email protected] & 416.3 + 17 & 57364.204632665 & 0.03 & 0.09 & [email protected] & + | fast radio bursts ( frbs ) are millisecond - duration radio signals occurring at cosmological distances . however the physical model of frbs is mystery , many models have been proposed . here
we study the frequency distributions of peak flux , fluence , duration and waiting time for the repeating frb 121102 .
the cumulative distributions of peak flux , fluence and duration show power - law forms .
the waiting time distribution also shows power - law distribution , and is consistent with a non - stationary poisson process .
these distributions are similar as those of soft gamma repeaters ( sgrs ) .
we also use the statistical results to test the proposed models for frbs .
these distributions are consistent with the predictions from avalanche models of slowly driven nonlinear dissipative systems . |
the world of 2 qubits is the simplest setting where the notions of entanglement @xcite , bell inequalities @xcite and their witnesses @xcite , first appear
. it would be nice if , like the bloch sphere for one qubit @xcite , they could also be visualized geometrically .
however , the world of two qubits is represented by @xmath0 hermitian matrices and being 16 dimensional it is not readily visualized .
it can , however , be visualized by introducing an appropriate equivalence relation .
this idea has been used in @xcite to describe the separable and entangled states . here
we show that the entanglement witnesses and the chsh bell inequalities can be incorporated in this descriptions as well .
the geometric description is faithful to the duality between separable states and witnesses as we shall explain .
this allows for elementary and elegant proofs of non - elementary results . any @xmath0 hermitian operator @xmath1 can be represented by a @xmath0 real matrix @xmath2 using the pauli matrices as the basis : @xmath3 greek indices run on @xmath4 , roman indices on @xmath5 .
@xmath6 is the identity and @xmath7 are the pauli matrices .
summation over a pair of repeated indices is always implied , and indices are raised and lowered using the minkowski metric tensor @xmath8 . to reduce the number of components from 16 to 3 one relies on notions of equivalence . in particular , forgetting about the overall normalization of operators reduces the dimension by 1 .
an effective notion of equivalence comes from allowing alice and bob to operate on their respective qubits @xmath9 we shall focus on the case @xmath10 where the operation is invertible but not trace preserving .
the physical interpretation of this is that states which are accessible by local , reversible filtering are identified .
it is known as slocc @xcite and is briefly reviewed in section [ slocc ] . since @xmath11 the slocc equivalence reduces the dimension by 12 . as a consequence
, the slocc equivalence classes of unnormalized 2 qubits states can be visualized in 3 dimensions .
as we shall see , the slocc equivalence classes of entanglement witnesses are represented by the cube , the states by the tetrahedron and the separable states by the octahedron of fig .
[ states3d ] .
the octahedron and tetrahedron have been identified as the slocc representation in @xcite . adding the cube as a representation of the slocc equivalence classes of entanglement witnesses ,
shows that the natural duality relation between witnesses and separable states is preserved in the visualization of the slocc equivalence classes : the cube is the dual of the octahedron in the usual sense of duality of convex sets @xcite .
in particular , the number of faces in one is the number of vertices in the other .
the tetrahedron is , of course , its own dual . since the work of the horodeckis , @xcite , fig .
[ states3d ] has been widely used in quantum information theory for the special cases of states with maximally mixed subsystems @xcite .
this is a 9 dimensional family of states with @xmath12 in eq .
( [ observable ] ) .
since this family has lower dimension , it can be visualized in 3 dimensions using a more restrictive notion of equivalence than slocc : alice and bob are allowed to perform only unitary operations on their respective qubits with @xmath13 in eq .
( [ eq : equivalence - states ] ) .
this in arguably the most fundamental notion of equivalence in quantum information theory and is known as locc @xcite .
it is trace preserving , which expresses the fact that , unlike slocc , it is not lossy , ( no state is ever discard ) . since @xmath14 the locc equivalence classes of this 9 dimensional family of states can be represented in 3 dimensions @xcite .
it is remarkable that both the visualization and the interpretation of fig .
[ states3d ] remains the same when one goes from the 9 dimensional family to the 16 dimensional family of general 2 qubits states .
all that changes is the notion of equivalence .
[ states3d ] turns out to play a significant role also in the theory of quantum communication .
namely , it characterizes the stochastic properties of certain _ single qubit _ quantum channels as shown in @xcite .
this rather different interpretation of the figure follows from a deep relation , known as the choi - jamiolkwosky isomorphism @xcite , between linear operators acting on the hilbert space of alice and bob , and linear maps on single qubit states . using this , one finds , @xcite that ( for unital and trace preserving channels ) , the octahedron represents channels that destroy entanglement , the tetrahedron represents the completely positive maps and the cube the positive maps .
+ in section [ cto ] we shall review the slocc interpretation of fig .
[ states3d ] from a perspective that focuses on the duality relations between the sets in the figure .
the main new results concern the visualization of entanglement witnesses , duality and of the chsh bell inequalities in sections [ sec : vis - bell ] .
the local mapping of a two qubit state @xmath15 given by eq .
( [ eq : equivalence - states ] ) preserves positivity and takes a product state @xmath16 to a product state .
it therefore maps any separable state a convex combination of product states to a separable state .
this makes the equivalence @xmath17 , a useful notion in studying the entanglement of two qubits @xcite . since the operation does not preserve the normalization of the state it is convenient to consider states up to normalization .
the operations performed by alice and bob can be interpreted as _ probabilistically reversible filtering _ associated with the povm e_1^(m)= , e_2^(m)=1-e_1^(m ) ( @xmath18 is not a local operator .
local povm would require four @xmath19 s . )
the probability of successfully filtering the state @xmath20 is strictly positive and is given by @xmath21 .
when @xmath22 is unitary the filtering succeeds with probability one and no state is lost .
the filtering is probabilistically reversible since the original state @xmath15 can be recovered , with non - zero probability , from @xmath23 using the filter @xmath24 .
one can broaden the notion of equivalence under slocc from states to observable and in particular , to witnesses @xmath1 .
we take the action on witnesses to be contragradient to that of states : @xmath25 ( in case @xmath22 is unitary states and observables transform the same way ) . the motivation for
this choice is to have @xmath26 transforms by a similarity transformation so its trace and therefore the associated expectation and probability , is left invariant .
we shall say that @xmath27 is a _ potential entanglement witness _ if @xmath28 for all separable states @xmath29 .
since the set of separable states @xmath30 is a convex cone in the space of @xmath0 matrices , the set of potential entanglement witnesses @xmath31 is the dual convex cone of @xmath30 .
the set of states , @xmath32 , is a convex cone as well , and the three convex cones are evidently nested @xmath33 the cones @xmath32 , @xmath30 and @xmath31 all lie in 16 dimensions , which is not very useful for visualization .
slocc takes a potential entanglement witness to a potential entanglement witness .
this follows from @xmath34 which shows that @xmath35 is an entanglement witness if @xmath27 is .
slocc allows one to reduce the study of states , separable states , and ( potential ) entanglement witnesses to the study of the corresponding equivalence classes .
as will be explained in section [ cto ] , the equivalence classes of the three cones are described by the three polyhedra shown in fig .
[ states3d ] .
similar ideas can be used to visualize bell inequalities as we now proceed to show .
every bell inequality has a corresponding witness @xcite .
let @xmath36 be a witness for a specific bell inequality .
a state @xmath37 satisfies this bell inequality if @xmath38 since the separable states satisfy all types of bell inequalities @xcite it is clear that @xmath36 belongs to the family of potential entanglement witnesses .
the set @xmath39 of states satisfying ( [ bell ] ) for a specific type of bell inequalities ( e.g. the chsh family ) forms a convex cone .
however in general it is larger then the cone @xmath30 of separable states .
the chsh bell inequalities describe a situation where alice may choose to measure her qubits in one of two directions , @xmath40 and bob may similarly choose one of the directions , @xmath41 .
it is represented by the witness @xcite @xmath42 where @xmath43 is the chsh operator @xcite : @xmath44 @xmath45 .
the chsh bell inequality then takes the form of eq .
( [ bell ] ) .
the family of all chsh witnesses is an a - priory 8 dimensional family associated with the 4 directions alice and bob choose .
( in fact , an implicit degeneracy in eq .
( [ bchsh ] ) makes it is only 7 dimensional . )
locc takes a chsh witness corresponding to the directions @xmath46 to a witness associated with rotated directions while keeping @xmath47 and @xmath48 fixed .
this reduces the dimension by 6 and allows us to visualize the equivalence classes of chsh witnesses . as we explain in section [ sec : vis - bell ] , they turn out to be the three circles shown in fig .
[ bell_picture ] .
local operations , eq . ( [ eq : equivalence - states ] ) , are guaranteed to take separable states to separable states .
this reflects the fact that no entangled state can ever be ( locally ) filtered from a separable state .
this is not the case for states that satisfy bell inequalities .
in fact , @xcite gave examples of states satisfying all the chsh inequalities whose filtration violate the inequality .
this is consistent because the positivity of @xmath49 _ does not _ imply positivity of @xmath50 .
slocc does not act nicely on chsh .
this can also be seen from the fact that the family of chsh witnesses _ is not _
mapped on itself by slocc .
it is therefore _ not possible _ to represent the states that satisfy chsh inequalities in terms of their slocc equivalence classes .
it seems clear however that if a state @xmath23 , filtered from @xmath15 ( using only local operations ) , breaks a certain bell inequality then the properties of @xmath15 itself are inconsistent with ( more general ) local hidden variables .
indeed if @xmath15 would have been describable in terms of hidden variables that would have implied that the results of any experiment done on it including one involving local filtration ( which is also a type of measurement ) should be explainable in terms of these hidden variables consisting of filtration plus subsequent spin measurement can be described using hidden variables only if the choice of alice whether to measure @xmath51 or @xmath52 can influence the result of the filtration which was completed prior to it .
this breaks natural causality assumptions .
the fact that @xmath15 does not itself break bell s inequality may then be traced to the fact that the standard derivation of bell s inequalities does not involve these causality assumptions . ] .
this motivates the introduction of a notion of states that satisfy bell inequality in a slocc sense by requiring not only the state @xmath37 satisfy the bell inequality @xmath36 , but also that all states that can be _ probabilistically filtered _ from @xmath37 do .
this is clearly a slocc invariant notion .
mathematically , it is expressed by the requirement @xmath53 we shall denote this set @xmath54 .
it is , of course , a smaller set , @xmath55 , but it is not empty .
this set is slocc invariant and so visualized in three dimensions .
the corresponding 3-dimensional set is the intersection of @xmath56 cylinders shown in fig .
[ bell3d ] , as we shall explain in section [ sec : vis - bell ] .
for any fixed witness @xmath36 the equivalence class @xmath54 is a convex cone since _
m tr ( w_b ( _ 1+_2)^m ) _ m tr ( w_b _ 1^m)+ _ m tr ( w_b _ 2^m ) since the intersection of convex cones is a convex cone , it follows that the states that satisfy _ a family _ of bell inequalities in the slocc sense also form a convex cone .
in particular , this is so for the chsh family .
the intersection of the cone with the hyperplane @xmath57 is then evidently a convex set .
similarly , the dual ( convex ) cone to @xmath54 is slocc invariant by eq .
( [ covariance ] ) , i.e. if @xmath36 is a witness for a given slocc family , so is @xmath58 .
these notions of witnesses and states conform to the notion of slocc .
they can therefore be represented in terms of their equivalence classes and can be visualized in three dimensions .
to describe the slocc equivalence classes of qubits it is convenient to use their lorentz description .
any single qubit observable @xmath59 can be written as @xmath60 the observable @xmath59 is then represented by the real 4-vector @xmath61 .
@xmath59 is positive , and so is a state , if its trace and determinant are positive . since @xmath62 and @xmath63 ,
states are described by 4-vectors @xmath61 that lie in the forward light - cone .
consider @xmath64 since @xmath65 , it follows that the action of @xmath66 on the observable @xmath59 can be implemented by an ( orthochronous ) lorentz transformation of the four vector @xmath61 .
namely , @xcite , @xmath67 similarly , any observable @xmath1 in the world of 2 qubits can be represented by by a @xmath0 real matrix @xmath2 as in eq .
( [ observable ] ) .
this representation allows a simple geometric characterization of potential entanglement witnesses in terms of matrices @xmath2 that map the forward light - cone into itself .
this follows from the fact that @xmath1 is an entanglement witness iff for any product state , represented by time - like vectors @xmath68 : @xmath69 this characterization of potential witnesses will play a role in what follows . to describe the slocc equivalence classes we shall consider invariants under the action ( [ eq : equivalence - states ] ) .
the pair @xmath10 associated with @xmath70 gives rise to a pair of lorentz transformations @xmath71 and @xmath72 such that @xcite : @xmath73 since @xmath74 , @xmath75 is an invariant .
a more interesting and powerful invariant is constructed as follows : the minkowski adjoint of @xmath2 is defined as : @xmath76 where @xmath77 is the minkowski metric tensor .
@xmath78 transforms contragradiently to @xmath2 under @xmath22 .
this follows easily from the defining relations of the lorentz transformation @xmath79 : @xmath80 it follows that @xmath81 undergoes a similarity transformation under the action of @xmath22 , so its spectrum is a slocc invariant . for a general observable ,
@xmath82 is not guaranteed to be a hermitian matrix , and its spectrum therefore may not be real . however
, if @xmath1 is a potential entanglement witness , a simplification occurs .
in particular , the eigenvalues of @xmath82 are guaranteed to be positive @xcite .
this can be seen from the following argument : suppose @xmath1 is a _ strict _ witness of entanglement , so that eq .
( [ forward - lc ] ) holds with strict inequality .
@xmath81 then maps the forward light - cone into its interior , since for any causal vector @xmath15 0<()_()^=_(^ ) ^ this implies , by a fixed point argument , that the largest eigenvalue of @xmath81 is positive and has a time - like eigenvector .
the lorentz orthogonal subspace to this eigenvector is a space - like invariant subspace .
restricted to this subspace , the minkowsky adjoint coincides with the ordinary adjoint .
this makes the remaining eigenvalues positive as well by @xmath83 above then it might happen that the largest eigenvalue has an eigenvector which is light - like .
this case is much more complicated , but for our aims here can be handled by a limiting argument . ] .
up to sign the _ lorentz singular values _
@xcite are defined as the roots of the eigenvalues of @xmath84 .
we denote them by @xmath85 .
they are the lorentz analog of the singular values of a matrix are not the covariant components of a lorenzian 4-vector . ] .
as the above argument shows the largest singular value which will be denoted @xmath86 corresponds generically to a timelike eigenvector ( and in degenerate cases to a null one ) while @xmath87 generically correspond to a spacelike eigenvector ( or possibly a null one in degenerate cases ) . defining @xmath85 as the square roots of the ( necessarily positive ) eigenvalues of @xmath84 still leaves a sign ambiguity . a unique determination @xmath85
is achieved by letting @xmath88 take the sign of @xmath75 and choosing all others non - negative .
furthermore , one orders them according to @xmath89 unless @xmath86 happens to vanish the slocc equivalence class may be characterized , up to scaling , by the three vector [ coord ] = 1 _ 0(_1 , _ 2 , _ 3 ) the slocc equivalence classes of potential entanglement witnesses would then be represented by the pyramid [ eq : pyramid ] \{(x , y , z ) | xy z0 } remarks : * if @xmath1 is a strict potential witness ( implying @xmath90 ) then it turns out @xcite that in analogy to the usual singular value decomposition one can find a pair of lorentz transformations that bring @xmath2 to its canonical form @xmath91 this in turn imply that the singular values @xmath85 completely determine @xmath1 s slocc equivalence class .
* however if @xmath92 ( corresponding to a non - strict potential witness ) then there is no a - priori guarantee that such a pair of lorentz transformations exist .
witnesses @xmath1 associated with this boundary case split into nonequivalent classes : those having the canonical form ( [ canonical - form ] ) and others having more complicated canonical forms @xcite .
thus in the boundary case @xmath85 do not completely determine the slocc equivalence class .
note that the condition @xmath93 is enough to guarantee that @xmath94 takes the forward light - cone to itself and hence to guarantee that a potential entanglement witness @xmath95 having @xmath85 as singular values exists .
operators @xmath95 which differ by permutation of @xmath87 or by flipping a sign of a pair of @xmath96 s are slocc equivalent .
there are 24 such operations corresponding to the tetrahedral group .
strictly , therefore , the slocc equivalence classes of @xmath1 are represented by the pyramid of ( [ ww ] ) .
however for the purpose of drawing pictures it is more aesthetic to symmetrize and give up ( [ ww ] ) . now
each ( generic ) equivalence class is represented by 24 points in the @xmath96 s space([coord ] ) . in particular , the potential entanglement witnesses are then represented by the unit cube .
the following fact @xcite allows one to translate the duality relation between potential witnesses and separable state from 16 dimensions to 3 : [ oded_lemma ] let @xmath1 and @xmath97 be two potential entanglement witnesses . then : @xmath98 where @xmath85 and @xmath99 are the lorentz singular values of @xmath1 and @xmath97 respectively , ordered according to eq .
( [ ww ] ) . from a geometric point of view
it may be more aesthetic to use an equivalent formulation of the theorem which allows using any of the 24 possible representatives @xmath85 ( not necessarily satisfying ( [ ww ] ) ) .
this is easily achieved by replacing the r.h.s .
( [ duality ] ) by @xmath100 where the minimum is taken over the 24 possible representatives @xmath101 . in particular , given a potential witness @xmath1 and a separable state @xmath29
, we have for any representatives as in eq .
( [ coord ] ) , @xmath102 and @xmath103 , the inequality [ eq : cube - octa - duality ] 0 4(1 + ) fig .
[ fff ] demonstrates this inequality for a particular choice of @xmath1 .
since the positivity of the right hand side is the standard duality relation between convex sets in 3 dimensions @xcite we see that the theorem translates the duality , eq .
( [ ent - witness ] ) , between the 16 dimensional cones , to the duality between convex sets in 3 dimensions @xcite . .
the green triangle lies on the plane @xmath104 .
one may think of points near the corner of the tetrahedron that lie beyond the green triangle as representing the states that are incriminated as entangled by @xmath105 .
the chosen witness is optimal in the sense that no other entanglement witness detects a larger set of entangled states.,title="fig:",width=302 ] + [ fff ] letting the tetrahedral group act on the pyramid of eq .
( [ eq : pyramid ] ) gives the unit cube .
since the cube is the dual ( also known as polar @xcite ) of the octahedron one learns from eq .
( [ eq : cube - octa - duality ] ) that the slocc equivalence classes of the separable states are represented by the octahedron ( up to the tetrahedral symmetries ) .
the 16 dimensional set of states is self - dual . by eq .
( [ eq : cube - octa - duality ] ) the corresponding slocc equivalence classes must be represented by a self - dual convex set in three dimensions , which turns out to be the tetrahedron . to see this note that an operator @xmath15 in the canonical form , eq .
( [ canonical - form ] ) , is a sum of mutually commuting operators with one relation , ( @xmath106 ) .
it follows that its eigenvalues are \{_0+_1 _ 1 + _ 2 _ 2 - _ 1_2 _
3}where @xmath107 . requiring @xmath15 to be positive restricts to the intersection of four half - spaces which evidently yield the tetrahedron .
this completes the derivation of fig .
[ states3d ] .
one nice consequence of the geometric construction is a `` proof by inspection '' that for 2 qubits the peres separability test is iff @xcite .
it is easy to see @xcite that if @xmath15 is a separable state , then its partial transpose is positive .
the converse is not true in general , but is true for 2 qubits .
however the proof @xcite rests on non - elementary facts from operator algebras .
the proof by inspection goes as follows @xcite : denote by @xmath108 the partial transposition of @xmath15 . since @xmath109 is antisymmetric , while the remaining @xmath110 are symmetric , one has ( ^p)_=\ { rl -_2 , & + _ , & + . on the slocc equivalence classes of states @xmath111
, the partial transposition then acts as a reflection in the 2 axis : replacing @xmath112 with @xmath113 , since the octahedron of separable states is the intersection of the tetrahedron with its reflection through the @xmath114 plane the result follows .
the chsh witnesses and inequalities were described in section [ sec : chsh ] . for the sake of simplicity in notation we shall now stick with the plus sign in the witness of eq .
( [ eq : chsh - witness ] ) . ] .
+ if a state violates a chsh inequality then it is necessarily entangled , but the opposite claim is false ; there are entangled states that do not violate any chsh inequality .
our aim is to visualize these .
the chsh witnesses have the property that @xmath115 .
this family is invariant under the action of locc .
the associated equivalence classes then live in three dimensions and are in 1 - 1 correspondence with the 3 singular values of the @xmath116 matrix @xmath117 [ 3by3mtarix ] ( _ b)_ij= ( _ b)_ijthe singular values are the roots of the three eigenvalues of @xmath118 . to find the explicit dependence of the singular values on the locc invariants @xmath119 and @xmath120the angles between the two directions alice and bob choose we introduce the pair of @xmath121 matrices : a=(a , a ) b=(b , b ) one checks that 2_b= a(b+b)^t+a(b - b)^t= a ( rr 1 & 1 + 1 & -1 + ) b^tsince @xmath122 is manifestly a @xmath116 matrix with rank @xmath123 , one of its eigenvalues is zero .
its remaining nonzero eigenvalues equal to those of the @xmath124 matrix [ two - by - two ] ( b^t b ) ( rr 1 & 1 + 1 & -1 + ) ( a^t a ) ( rr 1 & 1 + 1 & -1 + ) evidently a^t a= ( cc 1 & + & 1 + ) , b^t b= ( cc 1 & + & 1 + ) the matrix in eq .
( [ two - by - two ] ) now takes the form [ trace_one ] ( cc ^2 ( ) & ^2 ( ) + ^2 ( ) & ^2 ( ) + ) it has a unit trace , so the singular values of @xmath125 lie on the unit circle @xmath126 . solving for the eigenvalues we find : @xmath127 as @xmath128 and @xmath129 vary from @xmath130 to @xmath131 this gives one eighth of the unit circle where @xmath132 .
if we adjoint to it the twenty four representatives of the same equivalence class , we get the three mutually intersecting unit circles , shown in fig . [ bell_picture ] , representing the locc equivalence classes of chsh witnesses .
the dual set to the three unit circles ( more precisely to their convex hull ) then represents the locc equivalence classes of the states that satisfy all the chsh inequalitiesa - priori , only states with completely mixed subsystems are accommodated in a 3-d locc diagram .
however , it is easy to see that @xmath133 and @xmath134 do not affect @xmath135 . ] . to describe this geometrically ,
note that the dual set of the unit circle in the @xmath136 plane , for example , is the cylinder along the @xmath137 axis , with a unit radius .
the locc equivalence classes of states that satisfy all the chsh inequalities , is the intersection of three cylinders along the @xmath138 , @xmath139 and @xmath137 axes , with a unit radius .
this set ( see figure [ bell3d ] ) is bigger than the set of separable states represented by the octahedron .
+ in @xcite r. horodecki , p. horodecki and m. horodecki solved the problem of finding the optimal chsh witness for a given ( normalized ) state @xmath15 .
they show that tr(w_b)1- and the inequality is saturated for an appropriate choice of angles @xmath140 . here
@xmath141 are the two largest singular values of the @xmath116 matrix @xmath142 constructed from the spatial components of the matrix elements @xmath143 as in eq .
( [ 3by3mtarix ] ) .
in particular , the state @xmath15 violates a chsh inequality if and only if @xmath144 .
this result can be derived , essentially by inspection , from the geometric description of the previous section .
recall , ( see footnote [ f4 ] ) , that the state @xmath15 may be assumed , without loss , to be one where the subsystems are completely mixed . the locc equivalence class of @xmath15 is represented by the three singular values of @xmath142 , which we denote by @xmath145 . a bell witness ,
is represented by the three singular values of @xmath125 , which we denote by @xmath146 .
the vector @xmath146 takes values on the three circles in the figure . for a normalized state @xmath15
tr(w_b)= 1+_b it is clear that the optimal choice of a witness ( a minimizer ) is to choose the witness @xmath117 so that the vector @xmath146 is as anti - parallel to @xmath145 as possible .
( recall that @xmath146 is constrained to lie in one of the principle planes . )
the minimizer is then smallest entry among \ { 1-|x| , 1-|y|,1-|z| } this reproduces the result of the horodeckies .
[ bell3d ] also admits the following slocc interpretation : the states represented by points lying in the intersection of the three cylinders have the property that they , and all that can be filtered from them , satisfy all the chsh inequalities .
this is an immediate consequence of theorem [ oded_lemma ] which guarantees that @xmath147 attains its minimum value when @xmath15 takes its canonical form .
the slocc equivalence classes of states that lie outside the intersection of the cylinders have the property that they can always be filtered to yield states that violate some chsh inequality .
the chsh inequality constrains alice and bob to two dichotomic tests each .
a general theory of bell inequalities allows alice and bob @xmath148 and @xmath149 tests , having @xmath150 and @xmath151 possible results .
( a geometric framework for deriving such generalized bell inequalities is described in @xcite . )
let @xmath152 denote a corresponding bell inequality .
the chsh inequality is then @xmath153 .
von neumann tests on qubits restricts the outcomes of each test to two , @xmath154 , however , the number of tests , @xmath155 can be arbitrary .
one naively expects that by increasing the number of tests one might be able to incriminate some of the entangled states that pass the chsh test . indeed , d. collins and n. gisin @xcite present an example of a state @xmath15 that violates an @xmath156 inequality but does not violate any chsh inequality . here
we shall present numerical evidence which shows that when bell inequalities are interpreted in the slocc sense then @xmath156 ( with three dichotomic tests ) , is strictly weaker than @xmath153 .
it follows that the state found by collins and gisin can be filtered to a state that violates chsh .
the @xmath156 inequality derived by d. collins and n. gisin @xcite takes the form @xmath157 where @xmath158 is the probability that when alice chooses the @xmath159th measurement and bob chooses the @xmath160th measurement , they both get the outcome @xmath130 . allowing alice and
bob three experiments to choose from gives them more freedom .
in particular , they are always free to disconnect the third experiment .
this means that the chsh , or @xmath153 , must be a special case of @xmath156 .
indeed , the chsh @xmath153 inequality , @xmath161 is obtained from @xmath156 by alice disconnecting her third experiment , @xmath162 , and bob disconnecting his first experiment , @xmath163 . renaming @xmath164 and @xmath165 as @xmath166 and @xmath164
respectively gives chsh .
dichotomic , von neumann , tests of alice and bob are described by projection operators , namely setting in the above p(a_ib_j)22 where @xmath167 and @xmath168 are interpreted as directions in a measurement of the spin projection .
note that the case of a disconnected experiment _ is not _ of this form : it is not a dichotomic von neumann measurement .
the witness corresponding to @xmath169 with three dichotomic von neumann measurements for both alice and bob is then @xmath170 @xmath171 , implies that the state @xmath15 violates a bell inequality .
@xmath172 represents an ( a - priori ) 12 dimensional family of witnesses .
it has six locc invariant parameters , the angles @xmath173 and @xmath174 , @xmath175 .
we can visualize this family in 3 dimensions by representing @xmath176 by its canonical form under @xmath177 .
let us introduce the following direction matricesa= ( cccc 1 & 0 & 0 & 0 + 0 & a_1x & a_2_x & a_3_x + 0 & a_1y & a_2_y & a_3_y + 0 & a_1z & a_2_z & a_3_z + ) b= ( cccc 1 & 0 & 0 & 0 + 0 & b_1_x & b_2_x & b_3_x + 0 & b_1_y & b_2_y & b_3_y + 0 & b_1_z & b_2_z & b_3_z + ) and w_0= ( ccrr 4 & -1 & -1 & 0 + 1 & -1 & -1 & -1 + 1 & -1 & -1 & 1 + 0 & -1 & 1 & 0 + ) so @xmath178 . the lorentz singular values are the roots of the eigenvalues of @xmath179 .
one finds a^t a= ( cccc 1 & 0 & 0 & 0 + 0 & -1 & -(_12 ) & -(_13 ) + 0 & -(_12 ) & -1 & -(_23 ) + 0 & -(_13 ) & -(_23 ) & -1 + ) b^t b= ( cccc 1 & 0 & 0 & 0 + 0 & -1 & -(_12 ) & -(_13 ) + 0 & -(_12 ) & -1 & -(_23 ) + 0 & -(_13 ) & -(_23 ) & -1 + ) the lorentz singular values were calculated numerically .
the intersection of the resulting set with the @xmath180 plane , shown in fig .
[ gisin_z ] is clearly contained in the chsh unit circle .
similarly all points outside this plane were found to lie inside the convex hull of the three chsh circles , implying they represent weaker witnesses .
this means that under slocc the chsh inequality is stronger than @xmath181 ( when alice and bob are constrained to measure three spin directions each ) .
thus , if the two parties are allowed to filter then by letting them choose from @xmath56 possible experiments , we gain less information than by restricting them to choose from @xmath123 possible experiments .
witnesses in the @xmath180 plane : the circle represent chsh witnesses while all the dots inside represent @xmath181 witnesses in this plane .
since the dots lie inside the circle they represent weaker witnesses.,title="fig:",width=377 ] +
many of the important concepts in quantum information , such as entangled and separable states , entanglement witnesses , the chsh bell inequalities etc . can be visualized in three dimensions by introducing an appropriate equivalence relation .
the visualization allows us to give a `` proof by inspection '' of the non - elementary fact @xcite that the peres separability test for 2 qubits is iff .
it also allows us to `` solve by inspection '' the problem of optimizing the chsh bell inequality , which was solved by analytical methods in @xcite .
we have introduced the notion of states that satisfy bell inequalities in the slocc sense .
we gave numerical evidence which shows that allowing alice and bob an additional _ dichotomic von neumann _ test does not enable them to shrink the set shown in fig .
[ bell3d ] , obtained by filtering and chsh .
it is an interesting open question whether four or more dichotomic tests , or more general povm tests , can further shrink the set shown in fig .
[ bell3d ] .
* acknowledgment : * this work is partially supported by the isf .
we thank michael burman for his help with drawing figure [ bell3d ] , netanel lindner for discussions and mary beth ruskai for a helpful correspondence and several useful suggestions . | the notions of entanglement witnesses , separable and entangled states for two qubits system can be visualized in three dimensions using the slocc equivalence classes .
this visualization preserves the duality relations between the various sets and allows us to give `` proof by inspection '' of a non - elementary result of the horodeckies that for two qubits , peres separability test is iff .
we then show that the chsh bell inequalities can be visualized as circles and cylinders in the same diagram .
this allows us to give a geometric proof of yet another result of the horodeckies , which optimizes the violation of the chsh bell inequality .
finally , we give numerical evidence that , remarkably , allowing alice and bob to use three rather than two measurements each , does not help them to distinguish any new entangled slocc equivalence class beyond the chsh class . |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.