article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
structures and relations in algebra and representation theory are sometimes `` explained '' by geometric or topological facts , frequently by facts concerning a category of low - dimensional geometric or topological objects .
as a first illustration , consider the following well - known example of an algebraic relation .
let @xmath0 be a set .
a map @xmath1 is called _ associative _ if for any integer @xmath2 and any given @xmath3-tuple @xmath4 , every way of putting brackets in the expression @xmath5 gives , upon applying @xmath6 repeatedly according to the chosen bracketing , one and the same element of @xmath0 .
this is not the typical the textbook definition of associativity ; the latter is rather based on the a map @xmath1 is associative if and only if the equality @xmath7 holds for all triples @xmath8 .
it should be appreciated that this is really an assertion about binary trees with a finite number of leaves .
indeed , any bracketing of @xmath3 elements corresponds to a binary tree with @xmath3 leaves .
the statement then follows from the observation that any two binary trees can be transformed into each other by applying repeatedly a move that amounts to the associativity relation . in this sense , a combinatorial property of the collection of all binary trees `` explains '' the textbook definition of associativity . to give an example of
a structure `` explained '' or , rather , justified , by facts from geometry and related to quantum field theory , consider the symmetric monoidal category @xmath9 of smooth oriented cobordisms .
the objects of @xmath9 are finite disjoint unions of oriented circles @xmath10 , and its morphisms are diffeomorphism classes of smooth oriented surfaces with boundary . one way to learn more about the category @xmath9 is to _ represent _ it on a symmetric monoidal category @xmath11 , i.e. to study symmetric monoidal functors @xmath12 .
( an important example is @xmath13 , the category of vector spaces over a field @xmath14 . )
such a representation @xmath15 of the category @xmath9 is called a ( 2,1)-dimensional topological field theory . for the rest of this section
we restrict our attention to @xmath14-linear representations , with @xmath14 a field , i.e to the case of @xmath13 .
representations can be constructed using generators and relations for the object to be represented .
the category @xmath9 admits such a description in terms of generators and relations , given by the pair - of - pants decomposition of riemann surfaces with boundary . as a consequence , it suffices to know the vector space @xmath16 associated to the circle together with the linear maps @xmath17 and @xmath18@xmath19 that are obtained from the three - punctured sphere regarded as a cobordism @xmath20 and as @xmath21 , respectively , as well as the maps @xmath22 and @xmath23 obtained from the disk , seen as a cobordism @xmath24 and as @xmath25 , respectively .
it is well known ( see e.g. the textbook @xcite ) that the relations in the pair - of - pants decompositions imply that @xmath26 is a commutative frobenius algebra over @xmath14 . in this sense ,
the structure of @xmath9 together with the idea , inspired from quantum field theory , to study representations of @xmath9 , enforce on us the algebraic notion of a commutative frobenius algebra ( which was , of course , a known structure long before the advent of topological field theory ) .
the result amounts to a classification of topological field theories , which are objects in the functor category @xmath27 : the evaluation on the circle provides an equivalence @xmath28 of categories , where @xmath29 is the groupoid of commutative frobenius algebras over the field @xmath14 .
non - commutative frobenius algebras and their representation categories can be obtained in a similar spirit when one allows for a larger geometric category @xmath30 that has as objects disjoint unions of circles as well as intervals .
one then obtains a so - called open / closed topological field theory . for details
, we refer to @xcite . in the sequel we will present a small , and strongly biased by personal taste , digest of instances in which the interplay of geometric and algebraic structures has lead to a mathematical insight .
all those instances are , in some way , related to the priority programme `` representation theory '' .
next we present a more recent example for how topological field theory can `` explain '' a classical fact from algebra .
the following statements about a finite - dimensional hopf algebra @xmath33 over a field @xmath14 can be found in any textbook on hopf algebras .
the hopf algebra @xmath33 has a one - dimensional subspace @xmath34 of left integrals , with @xmath35 the counit of @xmath33 . by associativity , for @xmath36 and any @xmath37 ,
the element @xmath38 is again a left integral of @xmath33 .
since the space of integrals is one - dimensional , this in turn implies the existence of a linear form @xmath39 satisfying @xmath40 for any @xmath41 .
this linear form is a morphism of algebras and thus a group - like element of @xmath42 . by applying the same reasoning to the dual hopf algebra @xmath42
one finds analogously a distinguished group - like element @xmath43 . a classical result by radford is let @xmath33 be a finite - dimensional hopf algebra over a field @xmath14 , and let @xmath43 and @xmath39 be the distinguished group - like elements .
then the fourth power of the antipode @xmath44 of @xmath33 satisfies @xmath45 for all @xmath46 .
here the symbol @xmath47 denotes the left action of @xmath42 on @xmath33 which , using the sweedler notation @xmath48 for the coproduct of @xmath33 , is given by @xmath49 , and analogously @xmath50 is the right @xmath42-action on @xmath33 .
there is a purely algebraic proof for the purely algebraic statement in theorem 2.1 .
the statement admits a categorical reformulation . consider a finite tensor category @xmath51 , i.e. an abelian rigid tensor category over a field @xmath14 in which the monoidal unit is simple and which obeys several finiteness conditions ( finite - dimensional morphism spaces , finite jordan - hlder length for any object , and finitely many isomorphism classes of simple objects each of which has a projective cover ) . as an example , take the rigid monoidal category @xmath52-mod of finite - dimensional modules over a finite - dimensional hopf algebra @xmath33 .
the group - like element @xmath53 is a morphism of algebras and thus defines an invertible object @xmath54 in @xmath52-mod .
such an invertible object can be introduced for any finite tensor category ( * ? ? ?
* lemma2.9 ) . taking into account that the left dual of a module @xmath55-mod is given by @xmath56 and the right dual of @xmath57 by @xmath58 , radford
s theorem amounts to a relation between the double duals @xmath59 and @xmath60 in @xmath52-mod .
indeed , the assertion in theorem 2.1 generalizes @xcite to the following purely categorical statement : let @xmath51 be a finite tensor category . then there is a natural isomorphism @xmath61 of monoidal functors .
a proof of this statement can be given entirely in a category theoretic setting , see @xcite and ( * ? ? ?
* sect.5.2 ) .
we now outline how the theorem can be `` understood '' in the framework of topological field theories .
the definition of a topological field theory as a symmetric monoidal functor can be extended to other cobordism categories , and to higher - categorical versions of cobordism categories . in a rather general framework
one can consider an @xmath62-category @xmath63 of @xmath3-dimensional framed cobordisms , with objects being @xmath3-framed points , 1-morphisms @xmath3-framed 1-manifolds with boundary , 2-morphisms @xmath3-framed 2-manifolds with corners , etc .
for @xmath64 , all @xmath65-morphisms are invertible .
a topological field theory with values in @xmath11 is then a symmetric monoidal functor @xmath66 , with @xmath11 a symmetric monoidal @xmath62-category .
the same reasoning that shows that the vector space carries additional structure can be used to see that @xmath67 , the object assigned to a point , has special properties . the cobordism theorem @xcite states that @xmath67 is a fully dualizable object in @xmath11 and that any such object determines a framed topological field theory .
there is an equivalence @xmath68\ ] ] of @xmath69-groupoids generalizing ( [ frobequiv ] ) , where on the right hand side the symbol @xmath65 indicates that one discards all non - invertible morphisms in the full subcategory of fully dualizable objects of @xmath11 .
we now focus on three - dimensional topological field theories .
an important target @xmath11 is then the symmetric monoidal 3-category @xmath70 which has finite tensor categories as objects , bimodule categories over finite tensor categories as 1-morphisms , and bimodule functors and bimodule natural transformations as 2- and 3-morphisms , respectively .
a fusion category , i.e. a semisimple finite tensor category , is a fully dualizable object in @xmath70 and thus determines a framed three - dimensional topological field theory .
now by change of framing , the orthogonal group @xmath71 acts on @xmath72 and thus on the functor category @xmath73 .
this translates into a homotopy action of @xmath71 on the fully dualizable objects , i.e. on fusion categories : points in @xmath71 give self - equivalences @xmath74 \,{\to}\ , k[\mathcal bimod^{\mathrm{f.d.}}]$ ] , paths in @xmath71 give natural transformations between self - equivalences , etc .
the homotopy groups of the lie group @xmath71 are well - known : the group of connected components is @xmath75 ; the non - trivial component acts on monoidal categories as @xmath76 .
the fundamental group is @xmath77 ; its non - trivial element acts on fusion categories as an autoequivalence , which turns out @xcite to be given by the bidual @xmath78 .
thus from the group - theoretical fact that the non - trivial element of @xmath77 has order 2 one concludes that the quadruple dual @xmath79 is trivial , which in turn can be seen to imply radford s theorem for fusion categories .
( further , there are weakenings on the side of topological field theories which allow one to make statements about the quadruple dual of finite tensor categories , see @xcite . ) in this way , topological field theory provides a highly surprising connection between homotopy groups of lie groups and radford s @xmath31-theorem for finite - dimensional hopf algebras .
in the previous section we used topological field theories in three dimensions that are fully extended , i.e. down to the point . in the sequel we study a different variant of three - dimensional topological field theories , which is only extended down to one - dimensional manifolds .
it is of independent interest for representation theory .
consider the symmetric monoidal bicategory @xmath80 whose objects are closed oriented smooth one - dimensional manifolds , 1-morphisms are oriented surfaces with boundary and 2-morphisms are oriented three - manifolds with corners . to represent this bicategory ,
consider a symmetric monoidal 2-functor @xmath81@xmath82 with values in a symmetric monoidal bicategory @xmath11 : this is a called a 3 - 2 - 1-extended topological field theory .
one possible , and particularly important , choice for @xmath11 is the symmetric monoidal bicategory @xmath83-@xmath84 of so - called 2-vector spaces , which has finite semisimple @xmath14-linear abelian categories as objects , @xmath14-linear functors as 1-morphisms and natural transformations as 2-morphisms . in analogy to the situation with @xmath9 , the category @xmath85
is endowed with additional algebraic structure , and that structure should determine the three - dimensional topological field theory . analyzing this structure on @xmath85 with the help of a presentation of @xmath80 in terms of generators and relations @xcite , the category @xmath85 turns out to be a modular tensor category , i.e. a braided monoidal semisimple category , together with dualities and a ribbon twist , with non - degenerate braiding .
the non - degeneracy condition on the braiding can be formulated in several equivalent ways . a converse statement , that a modular tensor category determines a 3 - 2 - 1-extended topological field theory ,
has already been established long ago @xcite .
topological field theory thus suggests to regard modular tensor categories as a categorified notion of ( semisimple ) commutative frobenius algebras .
this is indeed a fruitful point of view .
there are several representation - theoretic sources of modular tensor categories : + * left modules over a connected factorizable ribbon weak hopf algebra with haar integral over an algebraically closed field @xcite . * local sectors of a net of von neumann algebras on @xmath86 that has finite @xmath87-index and is strongly additive and split @xcite .
* representations of a selfdual @xmath88-cofinite vertex algebra with an additional finiteness condition on the homogeneous components and with semisimple representation category @xcite . + ( the last two items constitute different mathematical formalizations of chiral conformal field theories . ) despite these many sources , modular tensor categories are `` rare '' objects . as a potential further source of modular tensor categories , crossed modules of finite groups have been proposed @xcite some time ago .
a crossed module consists of two finite groups @xmath89 and @xmath90 , an action of @xmath90 on @xmath89 by group automorphisms , and a group homomorphism @xmath91 such that @xmath92 for all @xmath93 and @xmath94 .
these data determine a braided monoidal category @xmath95 whose objects are @xmath89-graded finite - dimensional vector spaces carrying an action of @xmath90 such that @xmath96 , with the braiding given by @xmath97 for @xmath98 and @xmath99 homogeneous elements of degree @xmath100 , respectively .
the category @xmath95 is a premodular tensor category , i.e. has all the properties of a modular tensor category except that the braiding may be degenerate .
it is modular if and only if @xmath101 is an isomorphism , in which case one obtains the representation category of the double of a finite group , i.e. an already known type of modular tensor category . in the general case
the category can be modularized according to a standard procedure @xcite .
somewhat disappointingly , also this modularization turns out to be braided equivalent to the representation category of the drinfeld double of a finite group .
the modularization leads to a category with a certain weak group action , suggesting that it is actually the neutral part of an equivariant modular category @xcite . to find equivariant extensions of a given modular tensor category @xmath51
is an interesting algebraic problem with deep links to the brauer - picard group @xcite of @xmath51 .
again topological field theory , in this case a twisted version of dijkgraaf - witten theories , provides one solution to this algebraic problem . as a simplifying assumption ,
let us suppose that the group homomorphism @xmath91 appearing in the crossed module of finite groups is an injective group homomorphism , so that the group @xmath89 is a normal subgroup of @xmath90 .
the corresponding sequence @xmath102 of groups is not necessarily split .
but still the group @xmath103 acts weakly on @xmath89 , which can be seen as follows .
choose a set - theoretic section @xmath104 of @xmath105 .
then for any @xmath106 the automorphism @xmath107 of @xmath89 obeys @xmath108 with @xmath109 .
( we refrain from exhibiting the coherence relations satisfied by these group elements @xmath110 . )
this weak action induces a ( weak ) @xmath103-action on the modularization of the category @xmath95 which in the case at hand is the category @xmath111-mod of modules over the double of @xmath89 . for any finite group @xmath112
, the dijkgraaf - witten theory provides a 3 - 2 - 1-extended topological field theory .
the construction is based on the following facts ( for a recent presentation , see @xcite and references therein ) . to any manifold @xmath113 there is associated the groupoid @xmath114 of @xmath112-principal bundles .
this groupoid is essentially finite if @xmath113 is compact .
( extended ) cobordisms are spans of manifolds , possibly with corners ; pull back of bundles yields spans of groupoids .
taking functor categories into the symmetric monoidal category
@xmath84 amounts to a linearization and results in a 3 - 2 - 1-extended topological field theory .
in particular one obtains @xmath115 the main idea of the construction of a @xmath103-equivariant modular category whose neutral component is the modularization of @xmath95 is geometric @xcite : replace the categories @xmath114 by categories of twisted bundles .
[ def : twisted ] let @xmath103 act weakly on @xmath89 , such that there is an exact sequence @xmath102 of finite groups , and let @xmath116 be a @xmath103-cover over a smooth manifold @xmath57 . + ( i ) a _ @xmath117-twisted @xmath89-bundle over @xmath57 _ is a pair @xmath118 consisting of a @xmath90-bundle @xmath119 over @xmath57 and a smooth map @xmath120 over @xmath57 that obeys @xmath121 for all @xmath122 and all @xmath123 . + ( ii ) a _ morphism @xmath124 of @xmath117-twisted bundles _ is a morphism @xmath125 of @xmath90-bundles such that @xmath126 .
+ ( iii ) the category of @xmath117-twisted @xmath89-bundles is denoted by @xmath127 .
a @xmath103-equivariant topological field theory can now be obtained explicitly by a direct generalization of the construction of ordinary dijkgraaf - witten theories . in particular , for any group element @xmath106 there is a @xmath103-cover @xmath128 of @xmath129 with monodromy @xmath130 and thus a category @xmath131 $ ] . then the category @xmath132 can be shown @xcite to be a @xmath103-equivariant modular category .
its orbifold category is the category of modules over the drinfeld double of the group @xmath90 .
the idea to extend the dijkgraaf - witten construction to categories of generalizations of bundles turns out to be quite fruitful . as an illustration consider topological field theories with defects .
such theories are defined on cobordism categories consisting of manifolds with singularities , see e.g. ( * ? ? ?
* sect.4.3 ) .
they turn out to be most intimately linked to the theory of module categories over fusion categories @xcite .
dijkgraaf - witten theories with defects can be realized using a generalization of relative bundles @xcite . in this way
a combination of field - theoretic and geometric arguments allows one to recover pure algebraic results @xcite on module categories over the fusion category of finite - dimensional vector spaces that are graded by a finite group .
the program , initiated by andruskiewitsch and schneider , of classifying finite - dimensional pointed hopf algebras over a field has grown into one of the more remarkable recent developments in algebra ; see @xcite for a review .
one of the surprises that were encountered in the realization of this program is a close connection between pointed hopf algebras and lie theory .
the crucial ingredient of this connection is the theory of weyl groupoids ( see @xcite , as well as @xcite ) . in this groupoid
a construction is central that , while being purely algebraic , turns out to have important ramifications in topological field theory . in the form
introduced originally @xcite , this construction makes extensive use of smash products ; a convenient formulation in the framework of braided monoidal categories proceeds as follows .
we first summarize a few facts that are needed in the construction @xcite : * the notions of a hopf algebra @xmath0 , of a hopf pairing @xmath133 , and of the category @xmath134 of yetter - drinfeld modules over @xmath0 make sense in any braided category @xmath51 .
( recall that a yetter - drinfeld module has the structure of both a module and of a comodule , subject to a certain consistency condition .
yetter - drinfeld modules form a braided monoidal category . )
* suppose that we are given two hopf algebras @xmath0 and @xmath135 in @xmath51 together with a non - degenerate hopf pairing @xmath136 .
using @xmath137 and its inverse , we can turn a left @xmath0-action into a left @xmath135-coaction and a left @xmath0-coaction into a left @xmath135-action .
hereby we obtain an isomorphism @xmath138 of braided categories .
the input of the construction performed in @xcite is a _ partial dualization datum _
@xmath139 for a hopf algebra @xmath33 in a braided category @xmath51 ; this consists of a projection @xmath140 to a hopf subalgebra , and a hopf algebra @xmath135 together with a non - degenerate hopf pairing @xmath141 .
a partial dualization datum @xmath139 gives rise to a new hopf algebra @xmath142 in @xmath51 as follows .
the radford projection theorem , applied to the projection @xmath140 , allows us to write the hopf algebra @xmath33 as a radford biproduct @xmath143 , with @xmath144 the hopf algebra in the braided category @xmath134 given by the coinvariants of @xmath33 with respect to the projection @xmath105 .
the image @xmath145 of the hopf algebra @xmath144 in the braided category @xmath134 under the braided monoidal equivalence @xmath146 is a hopf algebra in the braided category @xmath147 .
3 . the partially dualized hopf algebra @xmath148 is defined as the radford biproduct @xmath149 of @xmath150 .
this is a hopf algebra in the braided category @xmath151 . in short ,
we simultaneously dualize a hopf subalgebra @xmath0 of @xmath33 and transport the coinvariants @xmath152 covariantly to @xmath153 .
note that this combines a contra- and a covariant operation ; thus partial dualization is not functorial in the hopf algebra @xmath33 . from the partial dualization datum @xmath139 for @xmath33 one
can obtain a partial dualization datum @xmath154 of the hopf algebra @xmath148 .
there is then a canonical isomorphism @xmath155 of hopf algebras in @xmath151 , showing that partial dualization is essentially involutive .
the partially dualized hopf algebra @xmath148 is rather different from @xmath33 , in general . as an illustration
consider the taft algebra @xmath156 for some primitive @xmath157th root of unity @xmath158 .
@xmath156 is a hopf algebra over the complex numbers that is generated by a group - like element @xmath159 of order @xmath160 and a skew - primitive element @xmath161 having coproduct @xmath162 .
there is a projection @xmath105 from the taft algebra @xmath156 to the hopf subalgebra @xmath163 $ ] generated by @xmath159 .
when taking this projection to obtain a partial dualization datum , the resulting partial dualization @xmath142 is in fact isomorphic to the taft algebra itself , with the isomorphism depending on a choice of a hopf pairing @xmath164 , and thereby on a choice of a primitive @xmath157th root of unity . by considering central extensions @xmath165 of @xmath156 by group - like elements one can get hopf algebras with non - trivial partial dualization :
the partial dualization @xmath166 no longer contains any central group - like elements .
rather , the coproduct of the skew - primitive element of @xmath167 gets modified , in such a way its partial dualization has additional central characters . a priori
the categories of left modules in @xmath51 over the hopf algebra @xmath33 and over its partial dualization @xmath142 are thus rather different .
however , there does exist a representation - theoretic relation between them @xcite : the categories of yetter - drinfeld modules over a hopf algebra @xmath33 in @xmath151 and of those over @xmath142 are braided equivalent : we have @xmath168 here the first and third isomorphisms come from the isomorphism @xmath169 of braided categories for a biproduct .
the equivalence of categories of yetter - drinfeld modules implies a relation between the hopf algebra @xmath33 and and its partial dualization @xmath142 .
let us discuss this for the case of a finite - dimensional hopf algebra over a field .
the category of yetter - drinfeld modules over @xmath33 is then the drinfeld center of the category of @xmath33-modules .
now semisimple algebras having isomorphic centers are morita equivalent .
this statement allows for a categorification : the notion of an algebra is substituted with the one of a monoidal category , while the one of a module over an algebra gets replaced by the one of a module category .
this way one arrives ( * ? ? ?
* thm3.1 ) at the statement that semisimple tensor categories with braided - equivalent drinfeld centers have equivalent bicategories of module categories , i.e. they are related by what is called @xcite a weak monoidal morita equivalence .
it is natural to ask whether the bicategories of module categories over @xmath33-mod and over @xmath142-mod are closely related , possibly even equivalent .
this in turn suggests that the categorified representation theory of a hopf algebra @xmath33 , i.e.the theory of module and bimodule categories over the monoidal category @xmath33-mod , should be regarded as an important aspect of the theory of hopf algebras . since the brauer - picard group of the monoidal category @xmath33-mod is isomorphic to the group of braided equivalences of the drinfeld center @xmath170 @xcite , it is invariant under partial dualization . on the other hand
it is not hard to give conditions @xcite that ensure the existence of a hopf isomorphism @xmath171 .
( there is also a generalization involving a modified partial dualization of @xmath33 . )
every pair @xmath172 then induces an equivalence @xmath173 and thus provides an element in the group of braided autoequivalences of the drinfeld center @xmath170 , and thereby an element of the brauer - picard group .
this result has an interesting application in three - dimensional topological field theories of turaev - viro type .
a turaev - viro theory is a fully extended oriented three - dimensional topological field theory , which assigns a spherical fusion category @xmath139 to the point .
types of topological surface defects in this theory are given by @xmath139-bimodule categories .
now it is a general field - theoretic pattern that topological invertible codimension - one defects are related to symmetries . as a consequence , the brauer - picard groups of the fusion categories in question can be identified @xcite as symmetry groups of three - dimensional topological field theories of turaev - viro type . in the special case
that the fusion category is the category of modules over the group algebra @xmath174 $ ] of an abelian group @xmath112 , the reflections just discussed amount to what in the physics literature are known as electric - magnetic dualities .
this constitutes yet another instance of an enlightening relation between algebraic notions and notions from field theory .
29 n. andruskiewitsch , on finite - dimensional hopf algebras .
proceedings of the icm 2014 , s.y .
jang , y.r .
kim , d.w .
lee , and i. yie , eds .
kyung moon , seoul , vol .
ii , pp .
117 - 141 ( 2014 ) .
arxiv:1403.7838 [ math.qa ] .
p. bantay , characters of crossed modules and premodular categories . in : moonshine : the first quarter century and beyond , j. lepowsky , j. mckay , and m. tuite , eds .
cambridge university press , cambridge , 2010 .
arxiv : math/0512542 [ math.qa ] . | structures in low - dimensional topology and low - dimensional geometry often combined with ideas from ( quantum ) field theory can explain and inspire concepts in algebra and in representation theory and their categorified versions .
we present a personal view on some of these instances which have appeared within the research priority programme spp 1388 `` representation theory '' .
primary : 81t45 ; secondary : 57r56 topological field theory , tensor categories , categorification |
symbiotic stars are spectroscopically characterized by prominent emission lines with tio absorption bands that are typical of a giant .
they are generally known to be wide binary systems consisting of a hot white dwarf and a giant star ( e.g. kenyon 1986 ) .
they are usually divided into s and d type symbiotic stars , where d type systems exhibit infrared excess indicative of warm dust with temperature @xmath10 .
many d type symbiotic stars are believed to contain a mira type giant as cool component .
their orbital separation of the hot and cool components may lie in the range of 10 - 100 au , from which we expect the orbital periods of order @xmath11 ( whitelock 1987 ) .
the giant component may lose mass in the form of slow stellar wind , a significant fraction of which can be photoionized by the hot white dwarf component .
taylor & seaquist ( 1984 ) investigated the ionization structures in order to interpret their radio observations of symbiotic stars .
many prominent emission lines in symbiotic stars can be excellent diagnostic tools to study the physical conditions around the white dwarf component ( e.g. mrset et al .
1991 ) .
it has been proposed by a number of researchers that some fraction of wind material from the giant component may be gravitationally captured by the hot component ( e.g. sokoloski bildsten & ho 2001 ) .
the mass loss rate from the giant component may range from @xmath12 depending on the stellar evolution stage ( e.g. iben & tutukov 1996 , luthardt 1992 ) .
wind accretion in binary star systems was investigated using the smoothed particle hydrodynamics ( sph ) by theuns and jorrissen ( 1993 ) , who concluded that an accretion dis can be formed in their isothermal flow models .
mastrodemos & morris ( 1998 ) also adopted an sph method to show that an accretion disk can be formed around the white dwarf component .
there are only limited studies about the existence of accretion disks and the insufficient numerical resolution prevents one from deducing the physical properties of an accretion disk in symbiotic stars .
symbiotic stars being considered as a candidate for type ia supernova progenitors , the mass transfer process in symbiotic stars can also be a very important subject in other branches of astronomy . about a half of symbiotic stars are known to exhibit very broad and mysterious emission features around 6825 and 7088 that were suspected to originate from highly ionized species ( allen 1980 ) .
these features were finally identified by schmid ( 1989 ) , who proposed the raman scattering origin . according to him ,
the raman scattering process starts with an o vi 1032 line photon incident upon a hydrogen atom in the ground state , which can be converted into an optical 6825 photon as a result of an inelastic scattering off the hydrogen atom that is left in its excited @xmath13 state . from an incident o vi 1038 line photon
, we obtain a raman photon at around 7088
. the raman scattering nature of these features is strongly supported by many observational studies including spectropolarimetric observations by harries & howarth ( 1996 ) and contemporaneous uv and optical observations by birriel et al .
( 2000 ) .
the wavelength @xmath14 of the inelastically scattered radiation is related with the incident wavelength @xmath15 and @xmath16 of hydrogen ly@xmath17 by @xmath18 differentiation of this equation leads to @xmath19 in the case of raman scattering involving o vi 1032 , @xmath20 .
this provides two important characteristics in the profiles of raman scattered o vi 6825 that they are broader than their parent emission lines by a factor of @xmath21 and that they are mainly determined by the relative motion between the neutral scatterers and line photon emitters .
therefore , the profiles are little affected by the direction of the observer s line of sight .
this is due to the inelasticity of raman scattering and noted to be one of the most important spectroscopic properties retained by raman scattered features ( e.g. nussbaumer , schmid & vogel 1989 , schmid 1989 ) . in this respect ,
raman scattered lines provide a very interesting opportunity to study the accretion process involving a slow stellar wind , which is quite unique to symbiotic stars . thus far ,
raman scattered lines are found for o vi 1032 , 1038 resonance doublets and he ii emission lines near h i lyman series in many symbiotic stars and in a few young planetary nebulae ( e.g. van groningen 1993 , pquignot et al .
1997 , groves et al . 2002 , lee et al .
2003 , birriel 2004 , zhang et al .
2005 , lee et al
. 2006b ) .
raman scattered o vi 6825 from o vi 1032 is particularly strong in a number of symbiotic stars , for which a refined profile analysis is feasible .
furthermore , raman scattered features consist of purely scattered photons , not mixed with the direct emission , which makes them characterized by strong linear polarization .
fairly intensive spectropolarimetric studies of raman scattered o vi lines were carried out by harries & howarth ( 1996 ) .
they found that the raman scattered o vi 6825 and 7088 show double or triple peak profiles and that they are strongly polarized with complicated structures .
schmid ( 1996 ) used a monte carlo method to calculate the profiles and polarization of these scattered features formed in a slowly expanding neutral region that mimics the stellar wind around the giant component .
a similar study with a slightly different velocity law of the slow stellar wind was performed by lee & lee ( 1997b ) . in these studies ,
no kinematic structure is assumed for the emission region , which is treated effectively as point - like without any complicated internal structures . instead , the stronger red part of raman scattered o vi 6825
is mainly attributed to scattering at the receding part of the giant stellar wind that has a bigger geometric covering of the o vi emission region than the approaching part does . however , in these models , one has to invoke a rather high velocity scale @xmath22 of the stellar wind from the giant in order to explain the observed raman profiles .
lee & park ( 1999 ) adopted an accretion disk emission model and proposed the origin of the multiple peak profiles with strong linear polarization . in their picture
, the accretion flow can be affected significantly by the binary orbital motion , which results in an asymmetric matter distribution around the white dwarf causing the stronger red part of the raman features . however , their study lacked detailed calculations that can be directly compared with the observed profiles .
both v1016 cyg and hm sge are classified as symbiotic novae or slow novae .
v1016 cyg underwent a slow nova - like outburst in 1965 , which may be interpreted to be a thermonuclear runaway on the surface of the white dwarf ( fitzgerald et al .
a similar nova - like optical outburst occurred in hm sge in 1975 ( dokuchaeva 1976 ) .
these two symbiotic stars are known to exhibit fairly strong raman scattered o vi 6825 with characteristic double peak profiles .
in particular , the appearance of raman scattered 6825 in hm sge was reported by schmid et al .
( 2000 ) in their spectra obtained with the william herschel telescope ( wht ) in 1998 .
both v1016 cyg and hm sge are d - type symbiotic stars , containing a mira type giant as cool component .
thus far , no detailed information is available about the orbital parameters of v1016 cyg and hm sge .
schild & schmid ( 1996 ) measured the rotation of position angle in the polarized raman 6825 with an estimation of the orbital period of @xmath23 in v1016 cyg .
however , schmid ( 1998 ) noted that this estimate may be severely affected by the change of the nebular structure in this system and the orbital period may exceed 100 years ( e.g. schmid 2001 ) . using the _ hubble space telescope ( hst ) _ data , brocksopp et al .
( 2002 ) measured the angular separation of @xmath24 between the white dwarf and the mira of v1016 cyg , proposing the physical binary separation of @xmath25 with their adopted distance of @xmath26 . according to the _ hst _ study of hm sge by eyres et al .
( 2001 ) , the separation of the mira and the white dwarf is @xmath27 with the adopted distance of @xmath28 . in this paper , we present our spectra of these raman features in the symbiotic stars v1016 cyg and hm sge secured with the bohyunsan optical echelle spectrograph ( boes ) , and make quantitative comparisons with the profiles computed by a monte carlo technique adopting an accretion disk emission model .
we observed v1016 cyg and hm sge on the nights of 2005 november 7 through 9 with the bohyunsan optical echelle spectrograph installed on the 1.8 m telescope at mt .
the detector was a 2k@xmath294k e2v ccd with pixel size @xmath30 , in which about 80 spectral orders covering @xmath31 to @xmath32 are recorded in a single exposure .
the 300 @xmath33 optical fiber was used to yield the spectroscopic resolving power of @xmath34 with an aperture field of 6.4 arcseconds .
a more detailed description of boes can be found in lee et al .
( 2006b ) and in kim et al .
the exposure times are from 7000 s and 3600 s for v1016 cyg and hm sge , respectively .
the data have been reduced following standard procedures using the iraf packages .
parts of our spectrum around 6825 of v1016 cyg and hm sge are shown in fig . 1 .
in the figure , the line flux is normalized by the local continuum in order to focus on the profile shape in the current work .
the two profiles are overall similar to each other .
the 6825 feature of hm sge is much weaker and therefore of poorer quality than that of v1016 cyg , which is indicated by the larger scatter in the data . in both systems ,
the red part is relatively stronger than the blue part , which is noted by many previous researchers ( e.g. harries & howarth 1996 , schmid et al .
1999 ) .
there exists a narrow emission line at around 6820 in hm sge .
the presence of this narrow emission feature in hm sge was also noted in the spectropolarimetric studies by schmid et al .
they concluded that this emission line is not related with raman scattering on the ground that it is unpolarized and narrow .
we ignore this feature in the subsequent analysis , by subtracting it by a single gaussian @xmath35 $ ] with @xmath36 and @xmath37 . in the case of v1016 cyg ,
the local maxima are found at 6823.5 and 6830.8 .
the separation of these two peaks corresponds to @xmath38 in the velocity space of o vi 1032 emitters .
the fwzi ( full width at zero intensities ) is @xmath39 .
it is noted that there is an asymmetrically extended red wing that may end at around 6842.7 .
the peaks for hm sge are located at 6823.3 and 6829.4 , for which the separation in velocity space of o vi is @xmath40 .
this is significantly smaller than that for v1016 cyg .
the fwzi is @xmath41 , which is comparable to that of v1016 cyg .
in contrast with v1016 cyg , it appears that there is no ( or significantly weak ) extended red wing in hm sge .
however , this should be taken with the caution that the data quality of hm sge is much poorer than v1016 cyg .
the relative kinematics between the o vi emission region and neutral h i scattering region in both v1016 cyg and hm sge is characterized by velocity scales ranging from @xmath42 to @xmath43 .
if the fwzi of the profiles is attributed to the kinematics of the giant wind as in the models of schmid ( 1996 ) and lee & lee ( 1997b ) , then the wind terminal speed of @xmath43 should be required , which is much faster than the escape velocity of a giant star . for v1016 cyg and @xmath44 for hm sge .
the full widths at zero intensities are @xmath39 and @xmath41 for v1016 cyg and hm sge , respectively . ]
in this work , we assume that the wind accretion occurs in both v1016 cyg and hm sge to form an accretion disk around each white dwarf component . according to model 3 of the sph studies by mastrodemos & morris ( 1998 ) ,
an accretion disk with size ranging 0.4 - 0.8 au is formed in a binary system containing a white dwarf , where the binary orbital period is 18 years .
we consider that the o vi emission region is limited to a part of the whole accretion flow , which is approximated to be keplerian . for simplicity ,
we assume that the o vi emission region is confined to an annular region characterized by the inner radius @xmath45 and the outer radius @xmath46 .
in this work we assume that each white dwarf component has a mass of @xmath47 . noting that the raman scattered o vi 6825 in both v1016 cyg and hm sge extends about @xmath48
, we infer that the o vi emission region extends from @xmath49 au to less than 1 au . with these ingredients we prepare a characteristic double peak profile of o vi 1032
, we first consider the disk component with the emissivity @xmath50 that depends only on @xmath51 through a power law @xmath52 where @xmath53 is a constant .
this functional dependency of emissivity is achieved in our monte carlo code by specifying the location @xmath51 of the initial photon from a uniform random number @xmath54 in the interval @xmath55 with the prescription @xmath56^{1/(\alpha+2)}.\ ] ] with another random number @xmath57 in the interval @xmath58 , we locate the point of generation of the incident photon @xmath59 .
then the photon is assigned a doppler factor @xmath60 along the propagation direction @xmath61 where @xmath62 is the local keplerian velocity at @xmath63 . according to eq .
( [ wavelength ] ) , the observed wavelength of raman scattered o vi is determined by the relative motion between the o vi emitter and the h i scatterer and quite insensitive to the observer s line of sight .
furthermore , the orbital parameters of v1016 cyg and hm sge are only poorly known . in view of these facts , in our monte carlo code
, we collect all the photons scattered inelastically off hydrogen atoms around the giant component .
however , a more refined study should include the effect of the observer s line of sight with full care .
the direct o vi 1032 line photons are scattered in a neutral scattering region , which may be expanding with a velocity @xmath64 .
schmid ( 1996 ) investigated the line profile formation of raman scattered o vi in an expanding slow stellar wind using a monte carlo technique ( see also harries & howarth 1997 ) .
a similar study was also conducted by lee & lee ( 1997b ) using a monte carlo method with a slight different kinematical velocity law for the slow stellar wind .
adopting also a monte carlo technique , we extend the work of lee & lee ( 1997b ) to incorporate the accretion disk emission model considered in the previous section .
the velocity of the slow and spherical stellar wind @xmath65 at the position @xmath66 is given by @xmath67 where @xmath5 is the terminal wind velocity and the origin of the coordinate system coincides with the center of the giant .
for simplicity of the current investigation , we set the power @xmath68 of the velocity law .
we introduce a new dimensionless vector defined by @xmath69 assuming a constant mass loss rate @xmath70 , the density @xmath71 is described by @xmath72 where @xmath73 with @xmath74 and @xmath33 being the mass of a proton and the mean molecular weight , respectively . in the monte carlo calculation , it is essential to locate the next scattering position @xmath75 for a given starting position @xmath76 in the direction @xmath77 , the unit wavevector of the photon under consideration . in order to accomplish this , an optical depth @xmath78 is generated using a uniform random number @xmath79 between 0 and 1 by @xmath80 in terms of the physical distance @xmath81 corresponding to @xmath78 from the starting position @xmath76 along the direction @xmath77 , the new scattering position is expressed as @xmath82 the new scattering position @xmath83 is related with @xmath78 by @xmath84 where @xmath85 is the sum of the cross sections for rayleigh and raman scattering .
lee & lee ( 1997a ) proposed the value @xmath86 with @xmath87 being the thomson scattering cross section , which we adopt in this work .
we introduce the representative optical depth @xmath88 defined by @xmath89 where @xmath90 , @xmath91 and @xmath92 .
the exact choice of @xmath88 and @xmath85 affects the ionization structure and hence the exact profile shape .
however , the ionization structure is also dependent on the uv radiation from the hot component . in view of this , we set @xmath93 in our simulations , in order to reduce the parameter space of investigation .
we also define @xmath94 as the impact parameter of the photon path with respect to the giant center , and the parameter @xmath95 measures the physical distance along the photon path . for the sake of convenience , we set @xmath96 at the foot of the perpendicular from the giant center to the photon path , so that @xmath97 where negative @xmath95 is obtained when @xmath98 . in terms of these parameters ,
the integral ( [ tau12d ] ) can be rewritten as @xmath99 in terms of a new function @xmath100 defined by @xmath101 , \label{urho}\ ] ] the integral in eq.([tau12 ] ) can be explicitly expressed as @xmath102.\ ] ] using the inverse relation of eq.([urho ] ) @xmath103\ ] ] this relation is directly inverted to obtain @xmath104 in terms of @xmath78 and @xmath105 .
a more complicated situation is obtained when @xmath106 , in which case we consider @xmath107 , the optical depth to the point of impact , defined by @xmath108.\ ] ] if @xmath109 , we consider @xmath110 from which we obtain @xmath111 with @xmath112 , @xmath81 is determined , and the location of the next scattering position @xmath113 is then given by eq.([s12 ] ) . with a branching ratio @xmath114 ,
the scattering process is raman , in which case we assume that the raman photon escapes from the region without further interaction . if the scattering process is rayleigh , then the photon propagates along the new direction @xmath115 , which is chosen from the rayleigh phase function ( e.g. schmid 1996 , lee & lee 1997a )
. the part of the stellar wind from the giant facing the hot white dwarf will be ionized by strong far uv radiation .
taylor & seaquist ( 1984 ) investigated the ionization structures that depend the density and the ionizing luminosity .
they introduced a a parameter @xmath116 defined by @xmath117 here , @xmath17 is the recombination coefficient , @xmath118 is the luminosity of ionizing radiation , and @xmath119 is the photoionization cross section .
then , the ionization front is described by the equation @xmath120 where @xmath95 is the distance measured from the white dwarf .
taylor & seaquist ( 1984 ) provided the closed form of @xmath121 using the density law @xmath122 , which is slightly different from the one given by eq .
( [ nr ] ) adopted in this work . although the description of taylor & seaquist ( 1984 ) is also approximately valid for the current work , we present the detailed ionization structure in a closed form in this subsection for the sake of completeness . with the density law given by eq .
( [ nr ] ) the necessary integral to be carried out is @xmath123 where @xmath124 and @xmath95 are related by the impact parameter @xmath94 as in eq .
( [ impact ] ) . for @xmath125 , an explicit expression of the integral @xmath126 is given by @xmath127^{1/2 } \nonumber \\ & \pm & 2\rho_i\cos\phi \left(\ln{\rho\over\rho-1 } -{1\over\rho-1}\right ) .
\label{eqifront1}\end{aligned}\ ] ] a more detailed calculation is given in the appendix . in fig .
[ ifront ] , we show the ionization fronts for various values of @xmath116 .
the horizontal axis coincides with the line connecting the two stars .
the dot at the origin stand for the mass losing giant .
we place the white dwarf at the point @xmath128 , considering the binary separation @xmath129 proposed by brocksopp et al .
the ionization front corresponding to @xmath130 is fitted by a hyperbola given by @xmath131 .
the angle between the horizontal axis and an asymptote is @xmath132 .
this is somewhat larger than the value proposed for v1016 cyg by jung & lee ( 2004 ) .
nevertheless , we perform monte carlo calculations using this value , because the exact value of the half opening angle affects only the overall flux of the raman 6825 feature and the profile shape is less sensitive .
in fig . [ purekep ] , we show the profiles of raman scattered 6825 formed in a stationary h i region relative to a keplerian disk . the h i region is modeled as the hyperboloid given in fig .
[ ifront ] of which the asymptotic cone has an half opening angle @xmath133 . in the figure
, we fix the velocity @xmath1 at the outer edge @xmath134 , and varied the inner radius @xmath45 from @xmath135 to @xmath136 . in the absence of a bright spot ,
the emission profile is symmetric with respect to the line center .
it is quite clear that the main peaks correspond to the velocity @xmath4 at the outer rim of the disk , which dominantly contributes to the whole line flux .
the emission region is almost ring - like in the case @xmath137 , where the profile has a sharp edge at @xmath138 .
with decreasing @xmath45 the profile broadens around the sharp edges at @xmath138 .
however , when @xmath139 , the profiles are effectively similar , because the inner disk is simply physically too small to contribute to the overall profile significantly .
in view of this , we fix the ratio of the inner and outer radii to be @xmath140 from now on . in fig .
[ alpha ] , we show three line profiles for @xmath141 and 1 , where @xmath17 is defined in eq .
[ eqalpha ] . in this figure
, @xmath142 is fixed in order to focus on the dependence of profiles on the functional form of the emissivity . with negative @xmath17 ,
the contribution from the inner disk region becomes larger , which leads to a broader profile with enhanced wings .
however , the locations of the main peaks are fixed at velocities @xmath143 , yielding qualitatively overall similar profiles irrespective of the detailed values of @xmath17 . from this consideration ,
we fix @xmath144 in the remainder of this paper . in fig .
[ giantwind ] , o vi line photons are incident on the stellar wind around the giant that is expanding according to eq .
( [ eqgwind ] ) . in this figure ,
the wind terminal velocity @xmath5 has a range from @xmath145 to @xmath146 . here
, we also fix the inner radius @xmath142 , and other parameters to be the same as those considered in fig .
[ purekep ] . therefore fig .
[ purekep ] is obtained when we set @xmath147 .
as is seen clearly in the figure , the sharp edges near @xmath143 are smoothed by the expanding motion of the h i region .
when @xmath5 exceeds @xmath148 the double peak structure can hardly be noticed . from this , it is concluded that the giant stellar wind with terminal velocity @xmath149 is consistent with the double peak profiles observed in v1016 cyg and hm sge . however , the profile modulation from other sources such as turbulence in the emission region may give rise to similar profiles shown in fig .
[ giantwind ] .
we do not consider this possibility in the current work .
it is clear that the redward asymmetry exhibited in the raman 6825 features of v1016 cyg and hm sge can not be accounted for by a pure keplerian disk emission model .
it is quite uncertain whether we are seeing an excess in red emission or a deficit in the blue part , or possibly both .
an excess red emission may correspond to a hot spot in an accretion disk which will be discussed in the following subsection and in this subsection we consider the first possibility . in order to account for
the stronger red part in the observed profiles , another component need to be added in the previous model considered in fig .
[ giantwind ] .
therefore , we introduce an o vi region between the white dwarf and the giant . this additional o vi region
is assumed to have a very low density so that the o vi emission from this region is negligible compared to the keplerian emission disk . nevertheless , the line center optical depth exceeds unity in a velocity space centered at @xmath150 .
schmid et al .
( 1999 ) investigated o vi 1032 profiles and compared with those of non - resonant o vi lines to conclude that the line transfer effect due to resonance scattering of o vi 1032 is present in symbiotic stars including v1016 cyg .
more specifically , we assume that the scattering optical depth by the hot wind component is given by a gaussian function @xmath151 . \label{eqhotwind}\ ] ] this prescription hinders the incidence of those line photons with the doppler factor near @xmath152 upon the scattering region .
however , the velocity modulation due to the giant wind fills the gap significantly , allowing one to obtain a double - peak profile with some suppresion in the red part .
[ bestfit ] shows the best fitting profiles to out boes spectra of v1016 cyg and hm sge . for v1016 cyg ,
the velocity @xmath4 at the outer rim of the accretion disk is @xmath1 and the terminal velocity of the giant wind @xmath153 .
the corresponding values for hm sge are @xmath154 .
the parameters adopted for the hot wind component in v1016 cyg are @xmath155 and @xmath156 and @xmath157 and @xmath158 for hm sge . in fig .
[ fitacc ] we present the line profiles for various velocities @xmath4 at the disk outer edge ranging from @xmath159 to @xmath160 with @xmath161 and @xmath162 fixed to the best fit values of fig .
( [ bestfit ] ) .
the middle panel shows the best fit line profiles in fig .
[ bestfit ] , whereas the top and bottom panels show line profiles with faster and slower disks , respectively . considering the much poorer
fits shown in the top and bottom panels , we may conclude that the outer rim velocities are constrained to be in the range @xmath163 for v1016 cyg and @xmath164 for hm sge . adopting a typical white dwarf mass of @xmath165 , the main o vi emission region resides from the hot white dwarf component at @xmath166 au for v1016 cyg and @xmath167 au for hm sge .
[ fitwind ] shows the monte carlo line profiles computed for various terminal giant wind velocities @xmath5 from @xmath168 to @xmath148 also with @xmath169 and @xmath162 fixed to the best fit values introduced in fig .
( [ bestfit ] ) .
similarly as in fig . [ fitacc ] , the middle panels show our best fit profiles in fig .
[ bestfit ] for comparison .
as is discussed in the previous section , larger giant wind velocities tend to erase the double peak structure , whereas overall sharper profiles result from smaller wind velocities yielding poor fits at the outer parts of the observed profiles .
it is also interesting to note that the wind terminal velocity @xmath170 for both v1016 cyg and hm sge is quite similar to the escape velocity of a typical giant star .
as is well known from spectroscopic studies of cataclysmic variables , an accretion disk may have a bright spot which is formed by the impact of the accretion stream near the outer edge of the accretion disk ( e.g. warner 1995 ) .
the existence of a similar feature in symbiotic systems is not clear . in this subsection
, we investigate the effect of the introduction of a hot spot in a keplerian disk on the line profile . in our monte carlo code ,
photons generated in the specified region satisfying @xmath171 and @xmath172 is assigned a statistical weight of 4.8 , so that this region is 4.8 times brighter than other part of the emission region .
this mimics the bright spots well known in cataclysmic variables , for which the accretion process takes place via roche lobe overflow . here , we do not consider the hot wind component that provided central absorption part in fig .
[ bestfit ] .
[ hotspot ] shows our profiles .
the solid lines show the profiles from a keplerian disk with a hot spot , whereas the dots show the boes data .
for the sake of comparisons , we also show the best fit profiles given in fig .
[ bestfit ] by dotted lines .
the fitting parameters for v1016 cyg are @xmath173 and @xmath174 . for hm
sge , they are @xmath175 and @xmath176
. it should be noted that the disk outer rims move slightly faster in this model than the ones considered in the case of the best fit model .
the profiles provide good fits to the outer parts @xmath177 and @xmath178 , whereas poor fits are obtained near center .
the introduction of a hot spot enhances the red peak .
however , the conspicuous central dips in the observed data are not obtained in pure keplerian disk models with @xmath179 for @xmath180 .
lowering @xmath5 makes the fit worse by narrowing the widths of the two peaks .
in this work , we present our profile analysis of raman scattered o vi 6825 in the symbiotic stars v1016 cyg and hm sge , assuming that o vi line photons are generated from a thin keplerian disk with a scattering region between the white dwarf and the mira giant .
the neutral scattering region is assumed to be a spherical slow stellar wind with terminal wind velocity @xmath5 with an ionization front approximated by a hyperboloid .
the best fitting parameters are @xmath1 , @xmath181 and @xmath182 for v1016 cyg and @xmath183 , @xmath184 and @xmath185 for hm sge .
the overall fitting quality is quite insensitive to our choice of the inner radius of the o vi emission region and the functional dependence of emissivity on radius .
furthermore , the profile shape is little affected by the exact shape of the ionization structure , which is determined by the mass loss rate of the mira component and the ionizing luminosity from the white dwarf component .
however , the ionization structure mainly affects the total flux of raman scattered 6825 . in this point
, an independent study should be added to current work in order to adequately probe the ionization structure and mass loss and transfer processes in symbiotic stars ( e.g. jung & lee 2004 ) .
although the introduction of a hot spot in our keplerian disk model failed to improve the profile fitting quality in the case of v1016 cyg and hm sge , this should not mean to exclude the possibility of the asymmetrical emission in symbiotic stars .
as harries & howarth ( 1996 ) reported , triple peak structures in raman scattered 6825 in symbiotics are more prevalent than double peak profiles in symbiotic stars .
it is quite probable that more complicated profiles may be fitted using more ingredients in the emission and/or scattering regions .
the resonance scattering o vi component considered in section 4.3 may be envisioned as a part of ionized giant stellar wind , in the vicinity of the inner lagrange point . in order to scatter a significant fraction of slightly blue o vi photons
, the region should move toward the main o vi line emission region around the white dwarf .
we may note that the bestfit speed @xmath186 in fig .
[ bestfit ] is slightly less than the giant wind terminal velocity . because @xmath186 is the velocity component along the direction connecting the giant and the white dwarf
, we may interpret that the resonantly scattering o vi region passes through the inner lagrange point making an angle @xmath187 under the assumption that the speed of the region is coincident with the wind terminal velocity @xmath5 . with the wide binary orbit of @xmath188 au the orbital speed of the giant component in both v1016 cyg and hm
sge is comparable to @xmath186 .
although we only consider the spherical wind around the mira type giant component in the current work , it is also a possibility that the wind around the giant component also possesses an azimuthal component , in which case this inference of @xmath189 should be corrected appropriately .
refined hydrodynamical studies may shed more light on the mass transfer processes in symbiotic stars .
the red wing excess in v1016 cyg that is not well fitted by our model may indicate the existence of another emission component and/or a neutral scattering region .
spectropolarimetric studies by harries & howarth ( 1996 ) show that the far red wing part is polarized in the direction perpendicular to that for the remainder part .
these facts are consistent when the additional emission and/or neutral component is receding with a velocity of several tens of @xmath190 perpendicular to the binary orbital plane .
it is unclear that this component is associated with the bipolar structure of these objects .
solf ( 1983 , 1984 ) detected a bipolar outflow in [ n ii ] lines in hm sge .
the velocity scale of several tens of km s@xmath191 is much smaller than the usual velocity scale associated with the fast hot wind that may emanate from a deep inner region of an accretion disk in the vicinity of the hot star .
schmid et al .
( 1999 ) reported the existence of broad wing components in o vi 1032 , 1038 observed with orfeus . even though the broad wings are consistent with the electron scattering origin , a fast bipolar wind around the hot component can also be a candidate .
it is an interesting possibility that the component responsible for resonance scattering of o vi is a clumpy one driven by the fast hot wind moving in the polar directions .
more interesting results are expected when photoionization calculations and hydrodynamical studies can be combined with spectropolarimetric polarimetric observations of these symbiotic objects .
it is also interesting that these two symbiotic stars exhibit he ii raman scattered features blueward of h balmer lines . in particular , raman scattered he ii 4850 features of v1016 cyg and hm sge exhibit no multiple peak structures ( e.g. jung & lee 2004 , birriel 2004 ) . because the cross section for he ii raman scattering is larger than that for o vi raman scattering by two orders of magnitude , we may expect that the scattering region is much more extended than the raman o vi counterpart . the profiles of raman o vi 6825 in v1016 cyg and hm sge are totally different from other high ionization lines including he ii .
for example , he ii 4686 emission lines of v1016 cyg and hm sge are single - peaked with slight blue asymmetry .
robinson et al . ( 1994 ) considered line profiles from accretion disks in symbiotic stars , and it appears to be difficult to infer the existence of an accretion disk in symbiotic stars from these emission lines .
it may be that the he ii emission region is more extended than the o vi emission region .
it is beyond the scope of this work to model other emission lines and more sophisticated models including photoionization and hydrodynamics are required .
the disparity of the profiles in raman scattered o vi 6825 and raman scattered o vi 7088 can be an important clue to the structure of the accretion disk in symbiotic stars .
a number of researchers pointed out the relative weakness in the blue part of raman o vi 7088 ( e.g. schmid et al .
1999 , harries & howarth 1996 ) .
the profile difference in resonance doublets of @xmath192 was reported in symbiotic stars and young planetary nebulae ( e.g. feibelman 1983 ) .
we thank the staff at the bohyunsan observatory with our particular gratitude to kang min kim and byung cheol lee .
we are also grateful to hwankyung sung for useful discussions on our spectroscopic observation .
we thank an anonymous referee , whose suggestions improved the presentation of the current paper .
this work is a result of research activities of the astrophysical research center for the structure and evolution of the cosmos ( arcsec ) funded by the korea science and engineering foundation .
we describe in detail the photoionization calculation for the ionization fronts , which was briefly discussed in section 2 .
the distance from the giant is measured in units of @xmath193 , the radius of the giant , by @xmath194 .
let @xmath195 be the separation of the giant and the white dwarf . along a path from the white dwarf making an angle @xmath196 with the line connecting the giant ,
the physical distance is measured by the parameter @xmath95 satisfying @xmath197 where @xmath198 is the impact parameter from the giant to the path of the ionization radiation from the white dwarf . the recombination rate along the path of ionization radiation originating from the white dwarf is proportional to the density squared along the path , given by @xmath201 in terms of @xmath124 , this integral can be re - written as @xmath202 we define the three integrals as @xmath203 the integral @xmath204 is elementary and is given by @xmath205 considering the substitution @xmath206 the integrals under consideration become @xmath207 \nonumber \\ i_2&= & \int d\theta b^2(\cosh^2\theta-1 ) \left [ { 1\over b\cosh\theta}-{1\over b\cosh\theta-1 } + { 1\over ( b\cosh\theta-1)^2 } \right ] \label{intj}\end{aligned}\ ] ] noting eq.([rhotheta ] ) , we may immediately see that eq.([strom1 ] ) reduces to eq.([eqifront1 ] ) . in fig .
[ ifront ] , we show ionization fronts for a range of ionizing luminosities characterized by the parameter @xmath116
. the cool giant is located at the origin and the white dwarf is at @xmath227 .
we also approximate the ionization corresponding to @xmath130 by a hyperbola given by @xmath131 . | we present the high resolution spectra of the d type symbiotic stars v1016 cygni and hm sagittae obtained with the bohyunsan optical echelle spectrograph ( boes ) , and investigate the double - peaked asymmetric profiles of the raman scattered o vi 6825 .
this feature is formed through raman scattering of o vi 1032 by atomic hydrogen with small scattering cross section of @xmath0 , requiring a specific condition of coexistence of a highly ionized emission nebula and a thick neutral region . by adopting a wind accretion disk model , we assume that the o vi emission region is described by a keplerian thin disk .
the raman scattering occurs in a neutral region near the giant , taking in the form of a slow stellar wind , part of which is ionized by the strong uv radiation from the hot white dwarf . using a monte carlo technique
, we compute the line profiles that are modulated by the slow spherical stellar wind from the giant component with the ionization front approximated by a hyperboloid . in order to account for the asymmetry and the existence of a central dip in the profiles , we add an o vi resonance scattering region between the hot white dwarf and the giant star which hinders the incidence of slightly blue o vi photons upon the h i region .
overall good fits to the observed data are obtained from our model , which lends support to the accretion disk emission model in these objects .
the best fitting parameters for v1016 cyg are @xmath1 , @xmath2 , and @xmath3 , where @xmath4 , @xmath5 and @xmath6 are the velocity of the outer disk rim , the terminal velocity of the giant wind , and the velocity component of the resonance scattering o vi region along the binary axis , respectively .
similar fitting parameters @xmath7 , @xmath8 and @xmath9 are obtained for hm sge .
we also investigate the effect of a hot spot in a disk that is well known in accretion disks in cataclysmic variables .
however , the introduction of a hot spot in our keplerian disk model failed to improve the overall profile fitting quality significantly .
brief discussions about our profile analysis in relation to bipolar morphology and accretion processes are presented . |
let @xmath18 be a bounded smooth domain and @xmath5 , and consider the problem @xmath19 where @xmath20 is a positive parameter and @xmath21 satisfies @xmath22 here , @xmath23 is the fractional laplacian , defined for @xmath5 by @xmath24 where @xmath25 is a constant .
it is well known see @xcite or the excellent monograph @xcite and references therein that in the classical case @xmath26 there exists a finite extremal parameter @xmath27 such that if @xmath28 then problem ( [ pb ] ) admits a minimal classical solution @xmath29 , while for @xmath30 it has no solution , even in the weak sense . moreover , the family of functions @xmath31 is increasing in @xmath20 , and its pointwise limit @xmath32 is a weak solution of problem ( [ pb ] ) with @xmath33 .
it is called the extremal solution of ( [ pb ] ) . when @xmath34 , we have that @xmath35 if @xmath36 @xcite , while @xmath37 if @xmath38 and @xmath39 @xcite .
an analogous result holds for other nonlinearities such as powers @xmath40 and also for functions @xmath41 satisfying a limit condition at infinity ; see @xcite . in the nineties h. brezis and j.l .
vzquez @xcite raised the question of determining the regularity of @xmath42 , depending on the dimension @xmath43 , for general nonlinearities @xmath41 satisfying ( [ condicions ] ) . the first result in this direction
was proved by g. nedev @xcite , who obtained that the extremal solution is bounded in dimensions @xmath44 whenever @xmath41 is convex .
some years later , x. cabr and a. capella @xcite studied the radial case .
they showed that when @xmath39 the extremal solution is bounded for all nonlinearities @xmath41 whenever @xmath36 .
for general nonlinearities , the best known result at the moment is due to x. cabr @xcite , and states that in dimensions @xmath45 then the extremal solution is bounded for any convex domain @xmath1 .
recently , s. villegas @xcite have proved , using the results in @xcite , the boundedness of the extremal solution in dimension @xmath46 for all domains , not necessarily convex .
the problem is still open in dimensions @xmath47 . the aim of this paper is to study the extremal solution for the fractional laplacian , that is , to study problem for @xmath5 .
the closest result to ours was obtained by capella - dvila - dupaigne - sire @xcite .
they studied the extremal solution in @xmath39 for the spectral fractional laplacian @xmath48 .
the operator @xmath48 , defined via the dirichlet eigenvalues of the laplacian in @xmath1 , is related to ( but different from ) the fractional laplacian .
we will state their result later on in this introduction .
let us start defining weak solutions to problem .
[ def ] we say that @xmath49 is a _ weak solution _ of if @xmath50 where @xmath51 , and @xmath52 for all @xmath53 such that @xmath53 and @xmath54 are bounded in @xmath1 and @xmath55 on @xmath56 .
any bounded weak solution is a _ classical solution _ , in the sense that it is regular in the interior of @xmath1 , continuous up to the boundary , and holds pointwise ; see remark [ solucions ] .
note that for @xmath26 the above notion of weak solution is exactly the one used in @xcite . in the classical case ( that is , when @xmath26 ) , the analysis of singular extremal solutions involves an intermediate class of solutions , those belonging to @xmath57 ;
see @xcite .
these solutions are called @xcite energy solutions . as proved by nedev @xcite ,
when the domain @xmath1 is convex the extremal solution belongs to @xmath57 , and hence it is an energy solution ; see @xcite for the statement and proofs of the results in @xcite .
similarly , here we say that a weak solution @xmath58 is an _ energy solution _ of when @xmath59 .
this is equivalent to saying that @xmath58 is a critical point of the energy functional @xmath60 where @xmath61 and @xmath62 our first result , stated next , concerns the existence of a minimal branch of solutions , @xmath63 , with the same properties as in the case @xmath26 .
these solutions are proved to be positive , bounded , increasing in @xmath20 , and semistable .
recall that a weak solution @xmath58 of is said to be _ semistable _ if @xmath64 for all @xmath65 with @xmath66 in @xmath3 .
when @xmath58 is an energy solution this is equivalent to saying that the second variation of energy @xmath67 at @xmath58 is nonnegative .
[ existence ] let @xmath68 be a bounded smooth domain , @xmath5 , and @xmath41 be a function satisfying .
then , there exists a parameter @xmath69 such that : * if @xmath28 , problem admits a minimal classical solution @xmath29 . *
the family of functions @xmath31 is increasing in @xmath20 , and its pointwise limit @xmath32 is a weak solution of with @xmath33 . * for @xmath30 , problem admits no classical solution . * these solutions @xmath29 , as well as @xmath42 ,
are semistable .
the weak solution @xmath42 is called the extremal solution of problem . as explained above , the main question about the extremal solution @xmath42 is to decide whether it is bounded or not .
once the extremal solution is bounded then it is a classical solution , in the sense that it satisfies equation pointwise .
for example , if @xmath70 then @xmath42 bounded yields @xmath71 .
our main result , stated next , concerns the regularity of the extremal solution for problem . to our knowledge
this is the first result concerning extremal solutions for . in particular , the following are new results even for the unit ball @xmath39 and for the exponential nonlinearity @xmath34 . [ s ]
let @xmath1 be a bounded smooth domain in @xmath72 , @xmath5 , @xmath41 be a function satisfying , and @xmath42 be the extremal solution of .
* assume that @xmath41 is convex .
then , @xmath42 is bounded whenever @xmath6 . *
assume that @xmath41 is @xmath73 and that the following limit exists : @xmath74 then , @xmath42 is bounded whenever @xmath7 .
* assume that @xmath1 is convex .
then , @xmath42 belongs to @xmath10 for all @xmath75 and all @xmath5 .
note that the exponential and power nonlinearities @xmath76 and @xmath77 , with @xmath78 , satisfy the hypothesis in part ( ii ) whenever @xmath7 . in the limit @xmath8
, @xmath9 is optimal , since the extremal solution may be singular for @xmath26 and @xmath79 ( as explained before in this introduction ) .
note that the results in parts ( i ) and ( ii ) of theorem [ s ] do not provide any estimate when @xmath80 is small ( more precisely , when @xmath81 and @xmath82 , respectively ) .
the boundedness of the extremal solution for small @xmath80 seems to require different methods from the ones that we present here .
our computations in section [ sec - exp ] suggest that the extremal solution for the fractional laplacian should be bounded in dimensions @xmath83 for all @xmath5 , at least for the exponential nonlinearity @xmath34 .
as commented above , capella - dvila - dupaigne - sire @xcite studied the extremal solution for the _ spectral _ fractional laplacian @xmath48 in @xmath39 .
they obtained an @xmath17 bound for the extremal solution in a ball in dimensions @xmath84 , and hence they proved the boundedness of the extremal solution in dimensions @xmath85 for all @xmath5 . to prove part ( i ) of theorem [ s ] we borrow the ideas of @xcite , where nedev proved the boundedness of the extremal solution for @xmath26 and @xmath44 . to prove part ( ii ) we follow the approach of m. sanchn in @xcite .
when we try to repeat the same arguments for the fractional laplacian , we find that some identities that in the case @xmath26 come from local integration by parts are no longer available for @xmath86 .
instead , we succeed to replace them by appropriate inequalities .
these inequalities are sharp as @xmath8 , but not for small @xmath80 . finally , part ( iii )
is proved by an argument of nedev @xcite , which for @xmath86 requires the pohozaev identity for the fractional laplacian , recently established by the authors in @xcite .
this argument requires also some boundary estimates , which we prove using the moving planes method ; see proposition [ bdyestimates ] at the end of this introduction . an important tool in the proofs of the results of nedev @xcite and sanchn @xcite is the classical @xmath12 to @xmath87 estimate for the laplace equation .
namely , if @xmath58 is the solution of @xmath88 in @xmath1 , @xmath89 in @xmath56 , with @xmath90 , @xmath91 , then @xmath92 this estimate and the sobolev embeddings lead to @xmath93 or @xmath94 estimates for the solution @xmath58 , depending on whether @xmath95 or @xmath96 , respectively . here , to prove theorem [ s ] we need similar estimates but for the fractional laplacian , in the sense that from @xmath97 we want to deduce @xmath98 or @xmath99
however , @xmath12 to @xmath100 estimates for the fractional laplace equation , in which @xmath101 is replaced by the fractional laplacian @xmath23 , are not available for all @xmath14 , even when @xmath102 ; see remarks [ rem1 ] and [ rem11 ] .
although the @xmath12 to @xmath100 estimate does not hold for all @xmath14 in this fractional framework , what will be indeed true is the following result .
this is a crucial ingredient in the proof of theorem [ s ] .
[ th2 ] let @xmath103 be a bounded @xmath104 domain , @xmath5 , @xmath105 , @xmath106 , and @xmath58 be the solution of @xmath107 * for each @xmath108 there exists a constant @xmath109 , depending only on @xmath43 , @xmath80 , @xmath110 , and @xmath111 , such that @xmath112 * let @xmath113 .
then there exists a constant @xmath109 , depending only on @xmath43 , @xmath80 , and @xmath14 , such that @xmath114 * let @xmath115 .
then , there exists a constant @xmath109 , depending only on @xmath43 , @xmath80 , @xmath14 , and @xmath1 , such that @xmath116 we will use parts ( i ) , ( ii ) , and ( iii ) of proposition [ th2 ] in the proof of theorem [ s ] .
however , we will only use part ( iii ) to obtain an @xmath17 estimate for @xmath58 , we will not need the @xmath13 bound .
still , for completeness we prove the @xmath13 estimate , with the optimal exponent @xmath117 ( depending on @xmath14 ) .
[ n=1 ] proposition [ th2 ] does not provide any estimate for @xmath118 . since @xmath5 , then @xmath119 yields @xmath120 and @xmath121 . in this case ,
any bounded domain is of the form @xmath122 , and the green function @xmath123 for problem is explicit ; see @xcite .
then , by using this expression it is not difficult to show that @xmath124 is @xmath125 in case @xmath126 and @xmath127 for all @xmath128 in case @xmath129 .
hence , in case @xmath130 it follows that @xmath131 , while in case @xmath132 it follows that @xmath133 for all @xmath134 and @xmath135 for @xmath78 .
proposition [ th2 ] follows from theorem [ th1 ] and proposition [ prop : u - is - cbeta ] below .
the first one contains some classical results concerning embeddings for the riesz potential , and reads as follows .
[ th1 ] let @xmath5 , @xmath105 , and @xmath136 and @xmath58 be such that @xmath137 in the sense that @xmath58 is the riesz potential of order @xmath138 of @xmath136 .
assume that @xmath58 and @xmath136 belong to @xmath139 , with @xmath140 .
* if @xmath141 , then there exists a constant @xmath109 , depending only on @xmath43 and @xmath80 , such that @xmath142 * if @xmath113 , then there exists a constant @xmath109 , depending only on @xmath43 , @xmath80 , and @xmath14 , such that @xmath143 * if @xmath115 , then there exists a constant @xmath109 , depending only on @xmath43 , @xmath80 , and @xmath14 , such that @xmath144_{c^{\alpha}(\r^n)}\leq c\|g\|_{l^p(\r^n)},\quad \textrm{where}\quad \alpha=2s-\frac{n}{p},\ ] ] where @xmath145_{c^{\alpha}(\r^n)}$ ] denotes the @xmath146 seminorm . parts ( i ) and ( ii ) of theorem [ th1 ] are proved in the book of stein ( * ? ? ?
* chapter v ) . part ( iii ) is also a classical result , but it seems to be more difficult to find an exact reference for it .
although it is not explicitly stated in @xcite , it follows for example from the inclusions @xmath147 which are commented in @xcite . in the more general framework of spaces with non - doubling @xmath43-dimensional measures , a short proof of this result
can also be found in @xcite .
having theorem [ th1 ] available , to prove proposition [ th2 ] we will argue as follows .
assume @xmath148 and consider the solution @xmath149 of the problem @xmath150 where @xmath136 is extended by zero outside @xmath1 .
on the one hand , the maximum principle yields @xmath151 in @xmath15 , and by theorem [ th1 ] we have that @xmath152 . from this ,
parts ( i ) and ( ii ) of the proposition follow . on the other hand , if @xmath153 we write @xmath154 , where @xmath155 solves @xmath156 in @xmath15 and @xmath157 is the solution of @xmath158 as before , by theorem [ th1 ] we will have that @xmath159 , where @xmath160 . then
, the @xmath13 regularity of @xmath58 will follow from the following new result .
[ prop : u - is - cbeta ] let @xmath1 be a bounded @xmath104 domain , @xmath5 , @xmath161 for some @xmath162 , and @xmath58 be the solution of @xmath163 then , @xmath164 , with @xmath165 , and @xmath166 where @xmath109 is a constant depending only on @xmath1 , @xmath167 , and @xmath80 . to prove proposition [ prop : u - is - cbeta ] we use similar ideas as in @xcite .
namely , since @xmath58 is harmonic then it is smooth inside @xmath1 .
hence , we only have to prove @xmath13 estimates near the boundary .
to do it , we use an appropriate barrier to show that @xmath168 where @xmath169 is the nearest point to @xmath170 on @xmath56 , @xmath51 , and @xmath165 . combining this with the interior estimates , we obtain @xmath13 estimates up to the boundary of @xmath1 . finally , as explained before , to show that when the domain is convex the extremal solution belongs to the energy class @xmath10 which is part ( iii ) of theorem [ s] we need the following boundary estimates .
[ bdyestimates ] let @xmath171 be a bounded convex domain , @xmath5 , @xmath41 be a locally lipschitz function , and @xmath58 be a bounded positive solution of @xmath172 then , there exists constants @xmath173 and @xmath109 , depending only on @xmath1 , such that @xmath174 where @xmath175 .
this estimate follows , as in the classical result of de figueiredo - lions - nussbaum @xcite , from the moving planes method .
there are different versions of the moving planes method for the fractional laplacian ( using the caffarelli - silvestre extension , the riesz potential , the hopf lemma , etc . ) .
a particularly clean version uses the maximum principle in small domains for the fractional laplacian , recently proved by jarohs and weth in @xcite . here
, we follow their approach and we show that this maximum principle holds also for integro - differential operators with decreasing kernels . the paper is organized as follows . in section [ sec - exist ]
we prove proposition [ existence ] . in section [ sec - exp ]
we study the regularity of the extremal solution in the case @xmath34 . in section [ sec - reg ]
we prove theorem [ s ] ( i)-(ii ) . in section [ sec - bdry ]
we show the maximum principle in small domains and use the moving planes method to establish proposition [ bdyestimates ] . in section [ sec - hs ]
we prove theorem [ s ] ( iii ) . finally , in section [ sec3 ] we prove proposition [ th2 ] .
in this section we prove proposition [ existence ] . for it
, we follow the argument from proposition 5.1 in @xcite ; see also @xcite .
_ we first prove that there is no weak solution for large @xmath20 .
let @xmath176 be the first eigenvalue of @xmath23 in @xmath1 and @xmath177 the corresponding eigenfunction , that is , @xmath178 the existence , simplicity , and boundedness of the first eigenfunction is proved in ( * ? ? ?
* proposition 5 ) and ( * ? ? ?
* proposition 4 ) .
assume that @xmath58 is a weak solution of .
then , using @xmath179 as a test function for problem ( see definition [ def ] ) , we obtain @xmath180 but since @xmath41 is superlinear at infinity and positive in @xmath181 , it follows that @xmath182 if @xmath20 is large enough , a contradiction with .
_ next we prove the existence of a classical solution to for small @xmath20 .
since @xmath183 , @xmath184 is a strict subsolution of for every @xmath4 .
the solution @xmath185 of @xmath186 is a bounded supersolution of for small @xmath20 , more precisely whenever @xmath187 . for such values of @xmath20 ,
a classical solution @xmath29 is obtained by monotone iteration starting from zero ; see for example @xcite .
_ we next prove that there exists a finite parameter @xmath27 such that for @xmath188 there is a classical solution while for @xmath30 there does not exist classical solution .
define @xmath27 as the supremum of all @xmath4 for which admits a classical solution . by steps 1 and 2
, it follows that @xmath189 .
now , for each @xmath188 there exists @xmath190 such that admits a classical solution @xmath191 . since @xmath192 , @xmath191 is a bounded supersolution of , and hence the monotone iteration procedure shows that admits a classical solution @xmath29 with @xmath193 .
note that the iteration procedure , and hence the solution that it produces , are independent of the supersolution @xmath191 .
in addition , by the same reason @xmath29 is smaller than any bounded supersolution of .
it follows that @xmath29 is minimal ( i.e. , the smallest solution ) and that @xmath194 .
_ we show now that these minimal solutions @xmath29 , @xmath28 , are semistable .
note that the energy functional for problem in the set @xmath195 admits an absolute minimizer @xmath196 .
then , using that @xmath29 is the minimal solution and that @xmath41 is positive and increasing , it is not difficult to see that @xmath196 must coincide with @xmath29 . considering the second variation of energy ( with respect to nonpositive perturbations ) we see that @xmath196 is a semistable solution of .
but since @xmath196 agrees with @xmath29 , then @xmath29 is semistable .
thus @xmath29 is semistable .
_ we now prove that the pointwise limit @xmath197 is a weak solution of for @xmath33 and that this solution @xmath42 is semistable .
as above , let @xmath176 the first eigenvalue of @xmath23 , and @xmath177 be the corresponding eigenfunction .
since @xmath41 is superlinear at infinity , there exists a constant @xmath198 such that @xmath199 using @xmath179 as a test function in for @xmath29 , we find @xmath200 in the last inequality we have used .
taking @xmath201 , we see that @xmath202 is uniformly bounded in @xmath203 .
in addition , it follows from the results in @xcite that @xmath204 for some positive constants @xmath205 and @xmath206 , where @xmath51 .
hence , we have that @xmath207 for some constant @xmath109 that does not depend on @xmath20 .
use now @xmath185 , the solution of , as a test function .
we obtain that @xmath208 for some constant @xmath109 depending only on @xmath41 and @xmath1 . here
we have used that @xmath209 in @xmath1 for some constant @xmath210 , which also follows from @xcite .
thus , both sequences , @xmath29 and @xmath211 are increasing in @xmath20 and uniformly bounded in @xmath203 for @xmath188 . by monotone convergence , we conclude that @xmath212 is a weak solution of for @xmath33 .
finally , for @xmath188 we have @xmath213 , where @xmath214 is defined by , for all @xmath65 with @xmath66 in @xmath3 .
since @xmath215 , fatou s lemma leads to @xmath216 and hence @xmath42 is semistable . [ solucions ]
as said in the introduction , the study of extremal solutions involves three classes of solutions : classical , energy , and weak solutions ; see definition [ def ] .
it follows from their definitions that any classical solution is an energy solution , and that any energy solution is a weak solution .
moreover , any weak solution @xmath58 which is bounded is a classical solution .
this can be seen as follows . first , by considering @xmath217 and @xmath218 , where @xmath219 is a standard mollifier , it is not difficult to see that @xmath58 is regular in the interior of @xmath1 . moreover , by scaling , we find that @xmath220 , where @xmath221 .
then , if @xmath222 , we can integrate by parts in to obtain @xmath223 for all @xmath222 .
hence , since @xmath224 , by density holds for all @xmath225 such that @xmath55 in @xmath3 , and therefore @xmath58 is an energy solution .
finally , bounded energy solutions are classical solutions ; see remark 2.11 in @xcite and @xcite .
in this section we study the regularity of the extremal solution for the nonlinearity @xmath34 .
although the results of this section follow from theorem [ s ] ( ii ) , we exhibit this case separately because the proofs are much simpler .
furthermore , this exponential case has the advantage that we have an explicit unbounded solution to the equation in the whole @xmath15 , and we can compute the values of @xmath43 and @xmath80 for which this singular solution is semistable . the main result of this section is the following . [ exp ] let @xmath1 be a smooth and bounded domain in @xmath72 , and let @xmath42 the extremal solution of .
assume that @xmath34 and @xmath7 .
then , @xmath42 is bounded .
let @xmath167 be a positive number to be chosen later .
setting @xmath226 in the stability condition ( [ semistable ] ) ( note that @xmath66 in @xmath3 ) , we obtain that @xmath227 next we use that @xmath228 for all real numbers @xmath229 and @xmath230 .
this inequality can be deduced easily from the cauchy - schwarz inequality , as follows @xmath231 using , , and integrating by parts , we deduce @xmath232 thus , using that @xmath233 , we find @xmath234 therefore , combining and , and rearranging terms , we get @xmath235 from this , it follows from hlder s inequality that for each @xmath236 @xmath237 for some constant @xmath109 which depends only on @xmath167 and @xmath111 .
finally , given @xmath7 we can choose @xmath236 such that @xmath238 .
then , taking @xmath239 in proposition [ th2 ] ( iii ) ( see also remark [ n=1 ] ) and using we obtain @xmath240 for some constant @xmath109 that depends only on @xmath43 , @xmath80 , and @xmath1 . letting @xmath241 we find that the extremal solution @xmath42 is bounded , as desired .
the following result concerns the stability of the explicit singular solution @xmath242 to equation @xmath243 in the whole @xmath15 .
[ stability - exp ] let @xmath5 , and let @xmath244 then , @xmath245 is a solution of @xmath246 in all of @xmath15 for some @xmath247 .
moreover , @xmath245 is semistable if and only if @xmath248 as a consequence : * if @xmath249 , then @xmath58 is unstable for all @xmath5 . * if @xmath250 , then @xmath58 is semistable if and only if @xmath251 . * if @xmath252 , then @xmath58 is semistable if and only if @xmath253 . * if @xmath38 , then @xmath58 is semistable for all @xmath5 .
proposition [ stability - exp ] suggests that the extremal solution for the fractional laplacian should be bounded whenever @xmath254 at least for the exponential nonlinearity @xmath34 .
in particular , @xmath42 should be bounded for all @xmath5 whenever @xmath249 .
this is an open problem . when @xmath26 and when @xmath255 , inequality coincides with the expected optimal dimensions for which the extremal solution is bounded for the laplacian @xmath256 and for the bilaplacian @xmath257 , respectively . in the unit ball @xmath39 , it is well known that the extremal solution for @xmath26 is bounded whenever @xmath36 and may be singular if @xmath38 @xcite , while the extremal solution for @xmath255 is bounded whenever @xmath258 and may be singular if @xmath259 @xcite . taking @xmath26 and @xmath255 in
, one can see that the inequality is equivalent to @xmath9 and @xmath260 , respectively .
we next give the first , using the fourier transform , it is not difficult to compute @xmath261 where @xmath262 thus , @xmath245 is a solution of @xmath263 .
now , since @xmath34 , by we have that @xmath245 is semistable in @xmath102 if and only if @xmath264 for all @xmath65 .
the inequality @xmath265 is known as the fractional hardy inequality , and the best constant @xmath266 was obtained by herbst @xcite in 1977 ; see also @xcite .
therefore , it follows that @xmath245 is semistable if and only if @xmath267 which is the same as .
in this section we prove theorem [ s ] ( i)-(ii ) .
we start with a lemma , which is the generalization of inequality .
it will be used in the proof of both parts ( i ) and ( ii ) of theorem [ s ] .
[ po ] let @xmath41 be a @xmath268 function , @xmath269 , @xmath270 , and @xmath271 then , @xmath272 for all nonnegative numbers @xmath229 and @xmath230 .
we can assume @xmath273 .
then , since @xmath274 , the inequality can be written as @xmath275 which follows from the cauchy - schwarz inequality
. the proof of part ( ii ) of theorem [ s ] will be split in two cases .
namely , @xmath276 and @xmath277 , where @xmath278 is given by .
for the case @xmath276 , lemma [ lemamanel ] below will be an important tool . instead , for the case
@xmath277 we will use lemma [ lemapotencia ] .
both lemmas are proved by sanchn in @xcite , where the extremal solution for the @xmath14-laplacian operator is studied .
[ lemamanel ] let @xmath41 be a function satisfying , and assume that the limit in exists . assume in addition that @xmath279 then , any @xmath280 satisfies @xmath281 where @xmath136 is given by .
[ lemapotencia ] let @xmath41 be a function satisfying , and assume that the limit in exists . assume in addition that @xmath282 then , for every @xmath283 there exists a positive constant @xmath109 such that @xmath284 the constant @xmath109 depends only on @xmath278 and @xmath285 .
the first step in the proof of theorem [ s ] ( ii ) in case @xmath276 is the following result .
[ propgamma ] let @xmath41 be a function satisfying .
assume that @xmath286 satisfies , where @xmath136 is given by .
let @xmath29 be the solution of given by proposition [ existence ] ( i ) , where @xmath188 .
then , @xmath287 for some constant @xmath109 which does not depend on @xmath20 .
recall that the seminorm @xmath288 is defined by .
using lemma [ po ] , , and integrating by parts , @xmath289 moreover , the stability condition applied with @xmath290 yields @xmath291 this , combined with , gives @xmath292 finally , by and since @xmath293 as @xmath294 , it follows from that @xmath295 for some constant @xmath109 that does not depend on @xmath20 , and thus the proposition is proved .
we next give the proof of theorem [ s ] ( ii ) .
assume first that @xmath276 , where @xmath296 by lemma [ propgamma ] and lemma [ lemamanel ] , we have that @xmath297 for each @xmath280 .
now , for any such @xmath298 , we have that @xmath299 is increasing and convex ( since @xmath300 ) , and thus @xmath301 therefore , we have that @xmath302 and thus , @xmath303 let now @xmath157 be the solution of the problem @xmath304 where @xmath149 is given by .
then , by and proposition [ th2 ] ( i ) ( see also remark [ n=1 ] ) , @xmath305 since @xmath306 is a subsolution of by , it follows that @xmath307 therefore , @xmath308 for all @xmath309 , where @xmath109 is a constant that does not depend on @xmath20 .
this can be done for any @xmath280 , and thus we find @xmath310 hence , using proposition [ th2 ] ( iii ) and letting @xmath311 it follows that @xmath312 hence , the extremal solution is bounded whenever @xmath7 .
assume now @xmath277 .
in this case , lemma [ lemapotencia ] ensures that for each @xmath283 there exist a constant @xmath109 such that @xmath313 then , by we have that @xmath308 for each @xmath314 .
next we show that if @xmath7 by a bootstrap argument we obtain @xmath35 . indeed , by proposition [ th2 ] ( ii ) and we have @xmath315 where @xmath316 .
now , we define recursively @xmath317 now , since @xmath318 then the bootstrap argument yields @xmath35 in a finite number of steps provided that @xmath319 .
this condition is equivalent to @xmath320 , which is satisfied for @xmath285 small enough whenever @xmath321 , since @xmath322 for @xmath277 .
thus , the result is proved . before proving theorem
[ s ] ( i ) , we need the following lemma , proved by nedev in @xcite .
[ lemah ] let @xmath41 be a convex function satisfying , and let @xmath323 then , @xmath324 where @xmath269 .
as said above , this lemma is proved in @xcite .
more precisely , see equation ( 6 ) in the proof of theorem 1 in @xcite and recall that @xmath325 at infinity
. we can now give the let @xmath136 be given by . using lemma [ po ] with @xmath326 and integrating by parts , we find @xmath327 the stability condition ( [ semistable ] ) applied with @xmath328 yields @xmath329 which combined with gives @xmath330 this inequality can be written as @xmath331 in addition , since @xmath41 is convex we have @xmath332 and thus , @xmath333 hence , by lemma [ lemah ] we obtain @xmath334 now , on the one hand we have that @xmath335 since @xmath41 is increasing and convex .
this yields , as in , @xmath336 on the other hand , let @xmath157 the solution of the problem @xmath337 by and proposition [ th2 ] ( i ) ( see also remark [ n=1 ] ) , @xmath338 since @xmath339 is a subsolution of , then @xmath340 .
therefore , @xmath341 and using proposition [ th2 ] ( iii ) , we find @xmath342 as desired .
in this section we prove proposition [ bdyestimates ] . this will be done with the celebrated moving planes method @xcite , as in the classical boundary estimates for the laplacian of de figueiredo - lions - nussbaum @xcite .
the moving planes method has been applied to problems involving the fractional laplacian by different authors ; see for example @xcite . however
, some of these results use the specific properties of the fractional laplacian such as the extension problem of caffarelli - silvestre @xcite , or the riesz potential expression for @xmath343 , and it is not clear how to apply the method to more general integro - differential operators . here , we follow a different approach that allows more general nonlocal operators . the main tool in the proof is the following maximum principle in small domains . recently ,
jarohs and weth @xcite obtained a parabolic version of the maximum principle in small domains for the fractional laplacian ; see proposition 2.4 in @xcite .
the proof of their result is essentially the same that we present in this section .
still , we think that it may be of interest to write here the proof for integro - differential operators with decreasing kernels .
[ mpsd ] let @xmath171 be a domain satisfying @xmath344 .
let @xmath345 be a nonnegative function in @xmath15 , radially symmetric and decreasing , and satisfying @xmath346 for some positive constants @xmath347 and @xmath348 , and let @xmath349 let @xmath350 be any bounded function , and @xmath351 be a bounded function satisfying @xmath352 where @xmath353 is the symmetric to @xmath170 with respect to the hyperplane @xmath354 .
then , there exists a positive constant @xmath355 such that if @xmath356 then @xmath357 in @xmath1 . when @xmath358 is the fractional laplacian @xmath23 , then the condition can be replaced by @xmath359 .
the identity @xmath360 in @xmath1 written in weak form is @xmath361 for all @xmath362 such that @xmath363 in @xmath3 and @xmath364 .
note that the left hand side of can be written as @xmath365 where @xmath366 denotes the symmetric of @xmath367 with respect to the hyperplane @xmath354 .
choose @xmath368 , where @xmath369 is the negative part of @xmath157 , i.e. , @xmath370 .
then , we claim that @xmath371 indeed , first , we have @xmath372 where we have used that @xmath373 for all @xmath374 . thus , rearranging terms and using that @xmath375 in @xmath376 , @xmath377 we next use that , since @xmath345 is radially symmetric and decreasing , @xmath378 for all @xmath170 and @xmath367 in @xmath379 .
we deduce @xmath380 and since @xmath381 for all @xmath367 in @xmath379 by assumption , we obtain .
now , on the one hand note that from we find @xmath382 moreover , since @xmath383 , then @xmath384 and therefore @xmath385 on the other hand , it is clear that @xmath386 thus , it follows from , , and that @xmath387 finally , by the hlder and the fractional sobolev inequalities , we have @xmath388 where @xmath389 .
thus , taking @xmath355 such that @xmath390 the lemma follows .
now , once we have the nonlocal version of the maximum principle in small domains , the moving planes method can be applied exactly as in the classical case . replacing the classical maximum principle in small domains by lemma [ mpsd ]
, we can apply the moving planes method to deduce @xmath391 for some constants @xmath109 and @xmath173 that depend only on @xmath1 , as in de figueiredo - lions - nussbaum @xcite ; see also @xcite .
let us recall this argument .
assume first that all curvatures of @xmath56 are positive .
let @xmath392 be the unit outward normal to @xmath1 at @xmath367 .
then , there exist positive constants @xmath393 and @xmath167 depending only on the convex domain @xmath1 such that , for every @xmath394 and every @xmath395 with @xmath396 and @xmath397 , @xmath398 is nondecreasing in @xmath399 $ ] .
this fact follows from the moving planes method applied to planes close to those tangent to @xmath1 at @xmath56 . by the convexity of @xmath1
, the reflected caps will be contained in @xmath1 .
the previous monotonicity fact leads to the existence of a set @xmath400 , for each @xmath401 , and a constant @xmath270 that depend only on @xmath1 , such that @xmath402 the set @xmath400 is a truncated open cone with vertex at @xmath170 .
as mentioned in page 45 of de figuereido - lions - nussbaum @xcite , the same can also be proved for general convex domains with a little more of care .
when @xmath39 , proposition [ bdyestimates ] follows from the results in @xcite , where birkner , lpez - mimbela , and wakolbinger used the moving planes method to show that any nonnegative bounded solution of @xmath403 is radially symmetric and decreasing .
when @xmath58 is a bounded semistable solution of , there is an alternative way to show that @xmath58 is radially symmetric .
this alternative proof applies to all solutions ( not necessarily positive ) , but does not give monotonicity .
indeed , one can easily show that , for any @xmath404 , the function @xmath405 is a solution of the linearized problem @xmath406 then , since @xmath407 by assumption , it follows that either @xmath408 or @xmath409 and @xmath157 is a multiple of the first eigenfunction , which is positive see the proof of proposition 9 in ( * ? ? ?
* appendix a ) .
but since @xmath157 is a tangential derivative then it can not have constant sign along a circumference @xmath410 , @xmath411 , and thus it has to be @xmath408 .
therefore , all the tangential derivatives @xmath412 equal zero , and thus @xmath58 is radially symmetric .
in this section we prove theorem [ s ] ( iii ) .
a key tool in this proof is the pohozaev identity for the fractional laplacian , recently obtained by the authors in @xcite .
this identity allows us to compare the interior @xmath16 norm of the extremal solution @xmath42 with a boundary term involving @xmath413 , where @xmath414 is the distance to @xmath56 .
then , this boundary term can be bounded by using the results of the previous section by the @xmath415 norm of @xmath42 , which is finite .
we first prove the boundedness of @xmath413 near the boundary .
[ bdyestimates - quotient ] let @xmath1 be a convex domain , @xmath58 be a bounded solution of , and @xmath221 . assume that @xmath416 for some @xmath417 .
then , there exists constants @xmath173 , @xmath418 , and @xmath109 such that @xmath419)}\right),\ ] ] where @xmath175 .
moreover , the constants @xmath414 , @xmath418 , and @xmath109 depend only on @xmath1 and @xmath205 .
the result can be deduced from the boundary regularity results in @xcite and proposition [ bdyestimates ] , as follows .
let @xmath173 be given by proposition [ bdyestimates ] , and let @xmath420 be a smooth cutoff function satisfying @xmath66 in @xmath421 and @xmath422 in @xmath423 .
then , @xmath424 and @xmath425 in @xmath3 .
moreover , we claim that @xmath426 for some function @xmath427 , with the estimate @xmath428 to prove that holds pointwise we argue separately in @xmath429 , in @xmath430 , and in @xmath431 , as follows : * in @xmath429 , @xmath432 .
since @xmath433 vanishes in @xmath423 and also outside @xmath1 , @xmath136 is bounded and satisfies . * in @xmath430 , @xmath434 .
then , using @xmath435 and that @xmath420 is smooth , we find that @xmath136 is bounded and satisfies . * in @xmath431 , @xmath434 . since @xmath436 vanishes in @xmath421 , @xmath136 is bounded and satisfies .
now , since @xmath58 is a solution of , by classical interior estimates we have @xmath437 see for instance @xcite .
hence , by and theorem 1.2 in @xcite , @xmath438 for some @xmath162 and @xmath439 thus , @xmath440 in the last inequality we have used and .
then , the result follows from proposition [ bdyestimates ]
. we can now give the recall that @xmath29 minimizes the energy @xmath67 in the set @xmath441 ( see step 4 in the proof of proposition [ existence ] in section [ sec - exist ] ) .
hence , @xmath442 now , the pohozaev identity for the fractional laplacian can be written as @xmath443 see @xcite .
therefore , it follows from and that @xmath444 now , by proposition [ bdyestimates - quotient ] , we have that @xmath445 for some constant @xmath109 that depends only on @xmath1 and @xmath446 .
thus , @xmath447 , and since @xmath212 , letting @xmath241 we find @xmath448 as desired .
the aim of this section is to prove propositions [ th2 ] and [ prop : u - is - cbeta ] . we prove first proposition [ th2 ] .
\(i ) it is clear that we can assume @xmath449 .
consider the solution @xmath149 of @xmath450 given by the riesz potential @xmath451 .
here , @xmath136 is extended by 0 outside @xmath1 .
since @xmath452 in @xmath3 , by the maximum principle we have that @xmath453 in @xmath1 . then , it follows from theorem [ th1 ] that @xmath454 and hence we find that @xmath455 for some constant that depends only on @xmath43 , @xmath80 , and @xmath111 .
\(ii ) the proof is analogous to the one of part ( i ) . in this case , the constant does not depend on the domain @xmath1 .
\(iii ) as before , we assume @xmath456 .
write @xmath154 , where @xmath155 and @xmath157 are given by @xmath457 and @xmath458 then , from and theorem [ th1 ] we deduce that @xmath459_{c^\alpha(\r^n)}\leq c,\quad\textrm{where}\quad \alpha=2s-\frac{n}{p}.\ ] ] moreover , since the domain @xmath1 is bounded , then @xmath136 has compact support and hence @xmath155 decays at infinity .
thus , we find @xmath460 for some constant @xmath109 that depends only on @xmath43 , @xmath80 , @xmath14 , and @xmath1 . now , we apply proposition [ prop : u - is - cbeta ] to equation . we find @xmath461 where @xmath462 .
thus , combining , and the result follows .
note that we have only used proposition [ prop : u - is - cbeta ] to obtain the @xmath13 estimate in part ( iii ) .
if one only needs an @xmath17 estimate instead of the @xmath13 one , proposition [ prop : u - is - cbeta ] is not needed , since the @xmath17 bound follows from the maximum principle . as said in the introduction , the @xmath12 to @xmath100 estimates for the fractional laplace equation , in which @xmath101 is replaced by the fractional laplacian @xmath23 , are not true for all @xmath14 , even when @xmath102 .
this is illustrated in the following two remarks .
recall the definition of the fractional sobolev space @xmath463 which , for @xmath464 , consists of all functions @xmath465 such that @xmath466 is finite ; see for example @xcite for more information on these spaces .
[ rem1 ] let @xmath5 .
assume that @xmath58 and @xmath136 belong to @xmath139 , with @xmath91 , and that @xmath467 * if @xmath468 , then @xmath469 .
* if @xmath470 and @xmath471 then @xmath58 may not belong to @xmath472 .
instead , @xmath473 , where @xmath474 is the besov space of order @xmath475 and parameters @xmath14 and @xmath476 . for more details
see the books of stein @xcite and triebel @xcite . by the preceding remark
we see that the @xmath12 to @xmath100 estimate does not hold in @xmath15 whenever @xmath470 and @xmath477 .
the following remark shows that in bounded domains @xmath1 this estimate do not hold even for @xmath468 .
[ rem11 ] let us consider the solution of @xmath478 in @xmath1 , @xmath2 in @xmath3 .
when @xmath39 and @xmath479 , the solution to this problem is @xmath480 see @xcite . for @xmath14 large enough one can see that @xmath245 does not belong to @xmath481 , while @xmath479 belongs to @xmath482 for all @xmath14 .
for example , when @xmath483 by computing @xmath484 we see that @xmath245 does not belong to @xmath485 for @xmath468 .
we next prove proposition [ prop : u - is - cbeta ] . for it
, we will proceed similarly to the @xmath486 estimates obtained in ( * ? ? ?
* section 2 ) for the dirichlet problem for the fractional laplacian with @xmath17 data .
the first step is the following : [ bound - u ] let @xmath1 be a bounded domain satisfying the exterior ball condition , @xmath5 , @xmath487 be a @xmath488 function for some @xmath162 , and @xmath58 be the solution of . then @xmath489 where @xmath169 is the nearest point to @xmath170 on @xmath56 , @xmath165 , and @xmath51 .
the constant @xmath109 depends only on @xmath43 , @xmath80 , and @xmath167 .
lemma [ bound - u ] will be proved using the following supersolution .
next lemma ( and its proof ) is very similar to lemma 2.6 in @xcite .
[ supersolution ] let @xmath5 .
then , there exist constants @xmath285 , @xmath205 , and @xmath206 , and a continuous radial function @xmath362 satisfying @xmath490 the constants @xmath205 and @xmath206 depend only on @xmath43 , @xmath80 , and @xmath117 .
we follow the proof of lemma 2.6 in @xcite .
consider the function @xmath491 it is a classical result ( see @xcite ) that this function satisfies @xmath492 for some positive constant @xmath493 .
thus , the fractional kelvin transform of @xmath245 , that we denote by @xmath494 , satisfies @xmath495 recall that the kelvin transform @xmath494 of @xmath245 is defined by @xmath496 then , it is clear that @xmath497 while @xmath494 is bounded at infinity .
let us consider now a smooth function @xmath420 satisfying @xmath66 in @xmath498 and @xmath499 observe that @xmath500 is bounded in @xmath501 , since @xmath502
. then , the function @xmath503 for some big constant @xmath198 , satisfies @xmath504 indeed , it is clear that @xmath363 in @xmath505 .
moreover , taking @xmath109 big enough it is clear that we have that @xmath506 .
in addition , the condition @xmath507 is satisfied by construction .
thus , @xmath362 satisfies , and the proof is finished .
once we have constructed the supersolution , we can give the first , we can assume that @xmath508 .
then , by the maximum principle we have that @xmath509 .
we can also assume that @xmath510 , since @xmath511 let @xmath512 and @xmath513 be small enough .
let @xmath514 be a ball of radius @xmath515 , exterior to @xmath1 , and touching @xmath56 at @xmath169 .
let us see that @xmath516 is bounded by @xmath517 in @xmath518 . by lemma [ supersolution ]
, we find that there exist constants @xmath205 and @xmath206 , and a radial continuous function @xmath362 satisfying @xmath490 ] let @xmath519 be the center of the ball @xmath514 .
since @xmath508 , it is clear that the function @xmath520 with @xmath521 big enough , satisfies @xmath522 here we have used that @xmath510 . then , since @xmath523 @xmath524 and @xmath525 it follows from the comparison principle that @xmath526 therefore , since @xmath527 in @xmath528 , @xmath529 moreover , since this can be done for each @xmath169 on @xmath56 , @xmath530 , and we have @xmath531 , we find that @xmath532 where @xmath169 is the projection on @xmath56 of @xmath170 .
repeating the same argument with @xmath58 and @xmath487 replaced by @xmath533 and @xmath534 , we obtain the same bound for @xmath535 , and thus the lemma follows .
the following result will be used to obtain @xmath13 estimates for @xmath58 inside @xmath1 . for
a proof of this lemma see for example corollary 2.4 in @xcite .
[ int - est - brick2 ] let @xmath5 , and let @xmath157 be a solution of @xmath536 in @xmath501 .
then , for every @xmath537 @xmath538 where the constant @xmath109 depends only on @xmath43 , @xmath80 , and @xmath298 .
now , we use lemmas [ bound - u ] and [ int - est - brick2 ] to obtain interior @xmath13 estimates for the solution of .
[ cbetabounds ] let @xmath1 be a bounded domain satisfying the exterior ball condition , @xmath161 for some @xmath162 , and @xmath58 be the solution of .
then , for all @xmath539 we have the following estimate in @xmath540 @xmath541 where @xmath462 and @xmath109 is a constant depending only on @xmath1 , @xmath80 , and @xmath167 .
note that @xmath542 .
let @xmath543 .
we have that @xmath544 moreover , using lemma [ bound - u ] we obtain @xmath545 furthermore , observing that @xmath546 in all of @xmath15 , we find @xmath547 with @xmath109 depending only on @xmath1 , @xmath80 , and @xmath167 . now , using lemma [ int - est - brick2 ] with @xmath548 , and taking into account , , and , we deduce @xmath549 where @xmath550 .
finally , we observe that @xmath144_{c^\beta\left(\overline{b_{r/4}(x)}\right)}=r^{-\beta}[\tilde u]_{c^\beta\left(\overline{b_{1/4}}\right)}.\ ] ] hence , by an standard covering argument , we find the estimate for the @xmath13 norm of @xmath58 in @xmath551 . now , proposition [ prop : u - is - cbeta ] follows immediately from lemma [ cbetabounds ] , as in proposition 1.1 in @xcite .
this proof is completely analogous to the proof of proposition 1.1 in @xcite .
one only have to replace the @xmath80 in that proof by @xmath117 , and use the estimate from the present lemma [ cbetabounds ] instead of the one from ( * ? ? ?
* lemma 2.9 ) .
the authors thank xavier cabr for his guidance and useful discussions on the topic of this paper . | we study the extremal solution for the problem @xmath0 in @xmath1 , @xmath2 in @xmath3 , where @xmath4 is a parameter and @xmath5 .
we extend some well known results for the extremal solution when the operator is the laplacian to this nonlocal case . for general convex nonlinearities
we prove that the extremal solution is bounded in dimensions @xmath6 .
we also show that , for exponential and power - like nonlinearities , the extremal solution is bounded whenever @xmath7 . in the limit @xmath8
, @xmath9 is optimal .
in addition , we show that the extremal solution is @xmath10 in any dimension whenever the domain is convex . to obtain some of these results we need @xmath11 estimates for solutions to the linear dirichlet problem for the fractional laplacian with @xmath12 data .
we prove optimal @xmath11 and @xmath13 estimates , depending on the value of @xmath14 .
these estimates follow from classical embedding results for the riesz potential in @xmath15 . finally , to prove the @xmath16 regularity of the extremal solution we need an @xmath17 estimate near the boundary of convex domains , which we obtain via the moving planes method . for it , we use a maximum principle in small domains for integro - differential operators with decreasing kernels . |
a * _ positroid _ * is a matroid that can be represented by a @xmath0 matrix with nonnegative maximal minors .
the classical theory of total positivity concerns matrices in which all minors are nonnegative , and this subject was extended by lusztig @xcite .
lusztig introduced the totally nonnegative variety @xmath1 in an arbitrary reductive group @xmath2 and the totally nonnegative part @xmath3 of a real flag variety @xmath4 .
he also conjectured that @xmath3 is made up of cells , and this was proved by rietsch @xcite . in this paper
, we will restrict our attention to @xmath5 , the * _ totally nonnegative grassmannian_*. then there is a more refined decomposition using matroid strata .
postnikov obtained a relationship between @xmath5 and certain planar bicolored graphs , producing a combinatorially explicit cell decomposition of @xmath5 @xcite .
the cells correspond to positroids .
one of the results of @xcite is that each cell is an intersection of @xmath5 and schubert cells corresponding to a combinatorial object called the grassmann necklace .
this result implies that each positroid is included in an intersection of cyclically shifted schubert matroids .
we extend this result : each positroid is exactly an intersection of certain cyclically shifted schubert matroids .
a more detailed formulation of the main result follows .
let @xmath6:=\{1,\cdots , n\}$ ] and let @xmath7 \choose k}$ ] be the collection of all @xmath8-element subsets in @xmath6 $ ] . fix some @xmath9 $ ] .
we define the ordering @xmath10 on @xmath6 $ ] by the total order @xmath11 . for @xmath12 \choose k}$ ] , where @xmath13 and @xmath14 we set @xmath15 for each @xmath16 \choose k}$ ] and @xmath17 , we define the * _ cyclically shifted schubert matroid _ * as @xmath18 \choose k } | i \leq_t j \}.\ ] ] we will show that a matroid @xmath19 \choose k}$ ] is a positroid if and only if it can be written as @xmath20 for a grassmann necklace @xmath21 , @xmath22 \choose k}$ ] .
our proof is purely combinatorial .
the paper is organized as follows . in section 2
, we go over the basics of matroids and the totally nonnegative grassmannian . in section 3 ,
we review @xmath23-diagrams and @xmath23-graphs . in section 4
, we give the proof of our main result . in section 5
, we introduce the upper grassmann necklace . in section 6
, we view lattice path matroids as special cases of positroids . in section 7
, we define flag positroids .
* acknowledgment * i would like to thank my adviser , alexander postnikov for introducing me to the field and the problem .
i would also like to thank allen knutson and lauren williams for useful discussions .
i would also like to thank anna de mier and tom lenagan for corrections .
we would like to guide the readers unfamiliar with basics in this section to @xcite and @xcite .
an element in the grassmannian @xmath24 can be understood as a collection of @xmath25 vectors @xmath26 spanning the space @xmath27 modulo the simultaneous action of @xmath28 on the vectors .
the vectors @xmath29 are the columns of a @xmath0-matrix @xmath30 that represents the element of the grassmannian . then an element @xmath31 represented by @xmath30 gives the matroid @xmath32 whose bases are the @xmath8-subsets @xmath33 $ ] such that @xmath34 .
here , @xmath35 denotes the determinant of @xmath36 , the @xmath8 by @xmath8 submatrix of @xmath30 with the column set @xmath37 .
then @xmath24 has a subdivision into * _ matroid strata _
* @xmath38 labeled by some matroids @xmath39 : @xmath40 the elements of the stratum @xmath38 are represented by matrices @xmath30 such that @xmath34 if and only if @xmath41 .
let us define the totally nonnegative grassmannian and its cells .
the * _ totally nonnegative grassmannian _ * @xmath42 is the quotient @xmath43 , where @xmath44 is the set of real @xmath45-matrices @xmath30 of rank @xmath8 with nonnegative maximal minors @xmath46 and @xmath47 is the group of @xmath48-matrices with positive determinant . * _ totally nonnegative grassmann cells _ *
@xmath49 in @xmath50 are defined as @xmath51 .
@xmath39 is called a * _ positroid _ * if the cell @xmath49 is nonempty .
note that from above definitions , we get @xmath52 in @xcite , postnikov showed a bijection between each cell and a combinatorial object called the grassmann necklace .
a * _ grassmann necklace _ * is a sequence @xmath53 of subsets @xmath54 $ ] such that : * if @xmath55 then @xmath56 for some @xmath57 $ ] , * if @xmath58 then @xmath59 .
the indices are taken modulo @xmath25 .
in particular , we have @xmath60 .
an example of a grassmann necklace would be @xmath61 .
two of the results in @xcite are the following : [ lem : p ] for a matroid @xmath19 \choose k}$ ] of rank @xmath8 on the set @xmath6 $ ] , let @xmath62 be the sequence of subsets such that @xmath63 is the minimal member of @xmath39 with respect to @xmath64 . then @xmath65 is a grassmann necklace .
[ thm : p ] let @xmath49 be a nonnegative grassmann cell , and let @xmath66 be the grassmann necklace corresponding to @xmath39 .
then @xmath67 where @xmath68 is the cyclically shifted schubert cell , which is the set of elements @xmath31 such that @xmath63 is the lexicographically minimal base of @xmath69 with respect to ordering @xmath70 on @xmath6 $ ] .
these results imply that bases of a positroid are included in an intersection of cyclically shifted schubert matroids .
but it does not imply that they are equal .
postnikov therefore conjectured that each positroid is exactly the intersection of cyclically shifted schubert matroids .
this is what we are going to prove in our paper : [ thm : maino ] @xmath39 is a positroid if and only if for some grassmann necklace @xmath71 , @xmath72 in other words , @xmath39 is a positroid if and only if the following holds : @xmath73 if and only if @xmath74 for all @xmath75 $ ] .
in @xcite , postnikov showed a bijection between positroids and combinatorial objects called @xmath23-diagrams . fix a partition @xmath76 that fits inside the rectangle @xmath77 .
the boundary of the young diagram of @xmath76 gives the lattice path of length @xmath25 from the upper right corner to the lower left corner of the rectangle @xmath77 .
let s denote this path as the * _ boundary path_*. label each edge in the path by @xmath78 as we go downwards and to the left .
define @xmath79 as the set of labels of @xmath8 vertical steps in the path .
each column and row corresponds to exactly one labeled edge .
let s index the columns and rows with those labels .
we will say that a box is at @xmath80 if it is on row @xmath81 and column @xmath82 . a * _ filling _ * of @xmath76
is a diagram of @xmath76 where each box is either empty or filled with a dot . given a filling @xmath83 of shape @xmath76 , we define the sets @xmath84 as : @xmath85 @xmath86 for a partition @xmath76 , let us define a * _ @xmath23-diagram _ * @xmath83 of shape @xmath76 as a filling of boxes of the young diagram of shape @xmath76 such that , for any three boxes indexed @xmath87 , where @xmath88 and @xmath89 , if boxes on position @xmath90 and @xmath91 are filled , then the box on @xmath80 is also filled .
this property is called the @xmath23-property .
we will say that a @xmath23-diagram is * _ full _ * if every box is filled .
-diagram and a @xmath23-graph , title="fig:",scaledwidth=30.0% ] -diagram and a @xmath23-graph , title="fig:",scaledwidth=30.0% ] fix a @xmath23-diagram @xmath83 of shape @xmath76 . for each @xmath92
, we define the * _ nw - region _ * of it as : @xmath93 there is a unique dot @xmath94 that minimizes @xmath95 and @xmath96 at the same time , due to the @xmath23-property .
we will say that @xmath94 covers @xmath80 and write this as @xmath97 .
a * _ @xmath23-graph _ * is obtained from a @xmath23-diagram in the following way .
place a vertex at the middle of each step in the boundary path of the diagram and mark these vertices by @xmath98 .
we will call these vertices the * _ boundary vertices_*. now for each dot inside the @xmath23-diagram , draw a horizontal line to its right , and vertical line to its bottom until it reaches the boundary of the diagram . then orient all vertical edges downward and horizontal edges to the left .
@xmath23-graphs were also used to study tp - basis of positroids in @xcite .
the source set of the @xmath23-graph is given by @xmath79 and the sink set is given by @xmath6 \setminus i(\lambda)$ ] .
a * _ path _ * in a @xmath23-graph is a directed path that starts at some boundary vertex and ends at some boundary vertex .
given a path @xmath99 , we denote its starting point and end point by @xmath100 and @xmath101 .
a * _ vd - family _ * is a family of paths where no pair of paths share a vertex .
a dot at @xmath80 is a * _ nw - corner _ * of a path @xmath99 , if @xmath99 changes direction at @xmath80 . for each @xmath102
, there is a path that starts at a boundary vertex @xmath81 , ends at a boundary vertex @xmath82 and has the dot at @xmath80 as a nw - corner .
we call such path as a * _ hook path of @xmath80_*. given a vd - family of paths @xmath103 , we say that this family represents @xmath104 .
empty family is also a vd - family .
the following proposition follows as a corollary from ( @xcite , theorem 6.5 ) .
fix a positroid @xmath105 that corresponds to a @xmath23-diagram @xmath83 . then @xmath106 if and only if @xmath107 is represented by a vd - family in the @xmath23-graph of @xmath83 .
each @xmath23-diagram corresponds to a positroid , and hence a grassmann necklace . given a @xmath23-diagram @xmath83 ,
let s try to find out its corresponding grassmann necklace @xmath21 directly from the diagram .
it is obvious that @xmath108 . for each @xmath109
, we can get a maximal chain @xmath110 such that @xmath111 is the unique dot in @xmath112 that minimizes @xmath113 and @xmath114 at the same time .
we will call this the * _ chain rooted at @xmath115_*. then the collection of hook paths at @xmath116 for @xmath117 is a vd - family .
so we get @xmath118 . in figure
[ fig : leh ] , chain rooted at @xmath119 is given by @xmath120 .
a chain rooted at @xmath121 is given by @xmath122 .
fix @xmath57 \setminus \{1\}$ ] and a @xmath23-diagram @xmath83 of shape @xmath76 and let @xmath21 be the grassmann necklace of @xmath105 . if @xmath123 , let @xmath115 be the box adjacent to a path labeled @xmath82 in the boundary path of @xmath76 . if @xmath124 , let @xmath115 be the box right above such box . then @xmath125 .
let @xmath126 be a vd - family that represents @xmath127 .
then @xmath126 only contains paths that satisfy @xmath128 . because if not , then @xmath129 represents @xmath107 such that @xmath130 .
so any path @xmath131 has to pass through a dot in the region @xmath112 .
let the chain rooted at @xmath115 be @xmath132 .
for each @xmath117 , denote the hook path at @xmath133 by @xmath134
. then @xmath126 must contain @xmath135 . if not , then @xmath136 because @xmath137 for all @xmath131 .
if @xmath138 , then we also have @xmath139 because if not , we get @xmath140 due to the fact that for any path @xmath141 , we have @xmath142 and @xmath143 . as a result , we get @xmath144 and @xmath125 .
let s look at an example . in the @xmath23-diagram of figure [ fig : leh ] ,
@xmath145 is given by @xmath146 .
chain rooted at @xmath147 is given by @xmath148 .
so @xmath149 .
@xmath150 is given by @xmath151 .
chain rooted at @xmath152 is given by @xmath153 .
so @xmath154 .
in this section , we will prove the main theorem by showing that for each grassmann necklace @xmath21 , we have @xmath155 . to do this , we need to show that each @xmath156 can be expressed as vd - family inside the @xmath23-graph of @xmath157 . in order to accomplish this , we will start from a full-@xmath23-diagram and use induction by increasing the number of empty boxes .
let @xmath83 be a full @xmath23-diagram of shape @xmath76
. then @xmath158 .
we need to show that for all @xmath159 , we have @xmath106 . due to the definition of @xmath160 , there is a unique bijection @xmath161 such that for any @xmath162 , the two intervals @xmath163 $ ] and @xmath164 $ ] do not cross , meaning that they are either disjoint or nested . for each @xmath165 , we associate a hook path at @xmath166 . then we get a vd - family representing @xmath107 . given any @xmath23-diagram @xmath167 with associated grassmann necklace @xmath21 , we want to add a dot , to obtain a new @xmath23-diagram @xmath168 such that for some @xmath169 $ ] , we have @xmath170 and @xmath171 whenever @xmath172 . the * _ boundary strip _ * of @xmath167 is the set of boxes where it shares at least one vertex with the boundary path .
let s first assume that there exists an empty box in the boundary strip of @xmath167 .
consider an empty box in the strip such that there is no empty box to its right or bottom . then adding a dot to this box will change exactly one element of the grassmann necklace .
so we only need to consider the case when all the boxes of the boundary strip are filled .
we define the * _ middle path _ * of @xmath173 to be a lattice path inside the diagram such that : 1 .
all boxes between the middle path and the boundary path are filled with dots , 2 .
the corner boxes of the upper region is empty . *
_ upper region _ * is the diagram obtained by looking at the boxes above or left of the middle path .
a box is a * _ corner box _ * of a diagram if there is are no boxes to its right and below .
then putting a dot into any corner box of the upper region will work , since the newly added dot will affect only one element of the grassmann necklace .
example of a middle path is given as a red line in figure [ fig : leh ] .
given any grassmann necklace @xmath174 , we have @xmath175 we will prove the proposition by induction on @xmath176 , the number of empty boxes inside the @xmath23-diagram @xmath167 of @xmath157 . when @xmath177 , this is the full @xmath23-diagram case .
so assume for the sake of induction that we know the result for @xmath23-diagrams having @xmath178 empty boxes .
use the construction above to obtain @xmath168 , where @xmath179 and there exists @xmath169 $ ] such that @xmath171 for all @xmath172 and @xmath170 .
induction hypothesis tells us that @xmath180 .
it is enough to show @xmath181 .
let @xmath182 be the chain representing @xmath183 in @xmath168 , such that @xmath184 is the newly added dot going from @xmath167 to @xmath168 .
we have @xmath185 for @xmath186 .
any vd - family @xmath187 representing some @xmath188 should contain a path in which @xmath184 is a nw - corner . in @xmath187
, denote the path going through @xmath184 by @xmath189 .
if there is no path in @xmath187 that passes @xmath190 , we can perturb the path @xmath189 to go through the points @xmath191 , @xmath190 , @xmath192 instead of going through @xmath184 .
so there must be a path @xmath193 that passes @xmath190 . since @xmath184 is a nw - corner of @xmath189
, @xmath190 is also a nw - corner of @xmath194 . repeating this argument , we get @xmath195 each having @xmath196 as a nw - corner .
let @xmath110 be the chain rooted at @xmath184 in @xmath167 .
then @xmath197 represents @xmath127 in @xmath198 .
we have @xmath199 due to the @xmath23-property .
we want to show that @xmath200 .
if @xmath201 or @xmath202 , then we have @xmath203 and we are done .
so let s assume @xmath204 and @xmath205
. if there is no path going through @xmath111 in @xmath187 , the path @xmath189 can be slightly changed so it goes through @xmath111 and this path can not have @xmath184 as its nw - corner .
so there must be a path @xmath206 in @xmath187 that passes through @xmath111 .
due to similar reasons , we only need to consider the case when @xmath207 and @xmath208 . if there is no path going through @xmath209 in @xmath187 , the path @xmath206 can be slightly changed so it goes through @xmath209 .
this path can not pass @xmath111 , since we have @xmath210 and @xmath211 .
so there must be a path @xmath212 that passes through @xmath209 . repeating this argument
, we get @xmath213 .
then @xmath214 tells us that @xmath215 .
( the reason we do this separately from the previous paragraph is because one of @xmath216 and @xmath217 might be true . )
so we have shown @xmath218 , and we are finished .
let s look at an example on using the main theorem .
let @xmath39 be a positroid such that its grassmann necklace is given by : @xmath219 our main theorem tells us that : @xmath220 @xmath221
in this section we will show that a positroid is also an intersection of cyclically shifted dual schubert matroids .
a decorated permutation @xmath222 is a permutation @xmath223 together with a coloring function @xmath224 from the set of fixed points @xmath225 to @xmath226 .
that is , a decorated permutation is a permutation with fixed points colored in two colors .
it is easy to see the bijection between necklaces and decorated permutations . to go from a grassmann necklace @xmath227 to a decorated permutation @xmath228 , * if @xmath229 , @xmath230 , then @xmath231 , * if @xmath59 and @xmath58 then @xmath232 , * if @xmath59 and @xmath55 then @xmath233 . to go from a decorated permutation @xmath228 to a grassmann necklace @xmath227 , @xmath234 | j<_i \pi^{-1}(j ) \textbf { or } ( \pi(j)=j \textbf { and } col(j)=-1 ) \}.\ ] ] let s look at an example . for decorated permutation @xmath235 with @xmath236 and @xmath237 , we get @xmath238 . for @xmath239 \choose k}$ ] , the * _ cyclically shifted dual schubert matroid _ * @xmath240 consists of bases @xmath241 such that @xmath242 . fix a decorated permutation @xmath228 .
let @xmath243 be the corresponding grassmann necklace and @xmath244 the corresponding positroid .
[ lem : maxm ] for any @xmath245 , we have @xmath246 for all @xmath247 $ ] .
we may assume that @xmath248 has no fixed point since they correspond to loops or coloops of @xmath244 .
we will prove only for @xmath249 since the proof is similar in other cases .
denote @xmath250 where @xmath251 are labeled in a way that satisfies @xmath252 .
denote elements of @xmath253 by @xmath254 .
let @xmath82 be the biggest element of @xmath255 $ ] such that : 1 .
@xmath256 for all @xmath257 $ ] and 2 .
@xmath258 .
since @xmath259 $ ] , we have @xmath260 .
we get @xmath261 , but this contradicts @xmath262 .
hence there can not be a @xmath263 $ ] such that @xmath264 .
this tells us that @xmath265 . the collection @xmath266 forms a necklace in the sense that @xmath267 except for @xmath81 such that @xmath268 .
we will call this the * _ upper grassmann necklace _ * of @xmath248 . to go from a decorated permutation @xmath228 to an upper grassmann necklace @xmath269 , @xmath270 | \pi(i ) < _ r i \textbf { or } ( \pi(i)=i \textbf { and } col(i)=-1 ) \}.\ ] ] define @xmath271 as : @xmath272 then lemma [ lem : maxm ] tells us that @xmath273 .
the proof of the following lemma is similar to lemma [ lem : maxm ] . for any @xmath274
, we have @xmath275 for all @xmath247 $ ] .
so we obtain the following result : [ thm : duality ] pick a decorated permutation @xmath228 .
let @xmath276 and @xmath277 be the corresponding grassmann necklace and the upper grassmann necklace .
then @xmath278 for all @xmath247 $ ] .
we also have the equality : @xmath279
lattice path matroids were defined in @xcite .
these are simple cases of positroids . in this section
we will show a way to get the decorated permutation of a lattice path matroid .
pick @xmath12 \choose k}$ ] such that @xmath280 .
the * _ lattice path matroid _ * is defined as : @xmath281 \choose k } , i \leq h \leq j \ } = sm_i \cap \tilde{sm}_j\ ] ] since @xmath282 corresponds to two lattice paths in a @xmath283-by-@xmath8 grid , @xmath284 expresses all the lattice paths between them .
[ lem : pathposi ] a lattice path matroid is a positroid . pick @xmath285 such that @xmath280 .
let s construct a @xmath8-by-@xmath25 matrix such that @xmath286 for all @xmath287 \choose k } \setminus lp_{i , j}$ ] and @xmath288 for all @xmath289 .
let @xmath290 be a @xmath8-by-@xmath25 vandermonde matrix .
set @xmath291 for all @xmath292 $ ] .
so @xmath293 would look like : @xmath294 assign values to variables @xmath295 such that @xmath296 and @xmath297 for all @xmath298 $ ] .
let s denote @xmath299,[c_1,\cdots , c_i]}$ ] as a submatrix of @xmath293 by taking rows from @xmath300 to @xmath81 and columns @xmath301 .
we have @xmath302 if and only if @xmath303,h}$ ] has nonzero diagonal entries , which happens if and only if @xmath289 .
let s try to find @xmath235 that corresponds to @xmath284 .
denote @xmath304 and @xmath305 . if @xmath306 for some @xmath307 $ ] , this is a coloop in the matroid .
the permutation @xmath248 we are trying to find should satisfy : * @xmath308 | i \leq
\pi^{-1}(i ) \}$ ] , * @xmath309 | \pi(i ) \leq i ) \}$ ] and * @xmath310 .
if @xmath248 satisfies the above conditions , then @xmath244 is contained in @xmath284 .
so @xmath284 is the biggest positroid under inclusion inside the collection satisfying the above property .
the following lemma is an immediate corollary of ( @xcite , theorem 17.8 ) . if we have @xmath311 , then @xmath312 , where @xmath313 is obtained from @xmath248 by switching @xmath314 and @xmath315 ( i.e. @xmath316 ) . combining this with lemma [ lem : pathposi ]
, we get the following result : choose any @xmath317 and @xmath318 \choose k}$ ] such that @xmath280 .
denote @xmath6 \setminus j = \{d_1 < \cdots < d_{n - k}\}$ ] and @xmath6 \setminus i = \{c_1 < \cdots < c_{n - k}\}$ ]
. then @xmath284 is a positroid and its decorated permutation @xmath228 is given by : @xmath319$},\ ] ] @xmath320$},\ ] ] @xmath321
positroids correspond to the matroid strata of the nonnegative part of the grassmannian .
flag matroids correspond to the matroid strata of a flag variety . a * _ flag _ * @xmath322 is a strictly increasing sequence @xmath323 of finite sets .
denote by @xmath324 the cardinality of the set @xmath325 .
we write @xmath326 .
the set @xmath325 is called the @xmath81-th constituent of @xmath322 . 1 . for all @xmath328 $ ] , @xmath329 the collection of @xmath325 s for each @xmath330 form a matroid .
2 . for every @xmath17 , the @xmath331-minimal bases of each @xmath329 form a flag .
if this holds , we say that @xmath329 s are concordant .
every flag @xmath332 such that @xmath333 is a basis of @xmath329 for @xmath334 belongs to @xmath126 . | postnikov gave a combinatorial description of the cells in a totally - nonnegative grassmannian .
these cells correspond to a special class of matroids called positroids .
we prove his conjecture that a positroid is exactly an intersection of cyclically shifted schubert matroids .
this leads to a combinatorial description of positroids that is easily computable . |
planetary nebulae ( pne ) are the ionised nebulae ejected by low - intermediate mass stars that have undergone extensive mass loss during the agb phase . the central stars of planetary nebulae ( cspn ) constitute a rich resource to study the late stages of binary stellar evolution .
at least 40 close binary cspn are known that have orbital periods less than @xmath21 day ( miszalski et al .
2011a ) and these make up at least @xmath3% of all cspn ( miszalski et al .
2009a ) . with their short - lived nebulae ( @xmath210@xmath4 yrs ) close binary cspn are assured to have just recently passed through the common - envelope ( ce ) phase ( iben & livio 1993 ) . with significantly less time to undergo further angular momentum loss compared to other more evolved post - ce binaries ( schreiber & gnsicke 2003 ) , the orbital periods of close binary cspn reflect the true post - ce distribution .
this makes them an excellent but relatively unexplored population to constrain ce population synthesis models .
miszalski et al .
( 2009a ) showed that there is a drop - off in the post - ce period distribution for periods @xmath5 1 day consistent with more evolved post - ce binaries ( rebassa - mansergas et al .
2008 ) . while the cause behind this sharp drop - off is not yet understood ( davis et al .
2010 ) , even less is known about intermediate period binaries in pne ( @xmath21001500 days ) that are predicted by some models ( yungelson et al .
1993 ; han et al .
1995 ; de marco et al . 2009 ) and their anticipated connection with the period distribution of post - agb binaries that precede pne ( van winckel 2003a ) . at present
there is essentially no hard evidence to support the existence of these binaries in pne , primarily because there have been no substantial long - term radial velocity monitoring campaigns .
this is starting to change with some bright central stars included in the survey described by van winckel et al .
( 2010 ) .
the best way to find intermediate period binaries is to look for giant or sub - giant companions whose orbits must be sufficiently large to accommodate their larger radii . as such cool companions are typically more luminous than their white dwarf ( wd ) companions , these binaries can usually only be found with the aid of uv photometry ( e.g. maxted et al .
the arduous uv selection is already done for pne where the nebula acts as a natural signpost for the presence of a hot wd .
all that is left is to identify a suitable cspn candidate too cool to ionise the pn and prove it has a physical connection to the nebula .
a small number of cspn have suspected cool central stars ( e.g. lutz 1977 ; tab .
4 of de marco 2009 ) , but very few have been studied in sufficient detail to rule out a line - of - sight superposition ( excluding k 1 - 6 , see frew et al .
poor uv sensitivity is a likely explanation for suggestions that some are single cspn caught momentarily during a born - again phase ( bond & pollacco 2002 ) .
perhaps the most studied of the cool central stars are those of the so - called a 35 type ( bond , ciardullo & meakes 1993 ) .
the initial list included a 35 , lotr 1 and lotr 5 which have rapidly rotating sub - giants or giants accompanied by very hot white dwarfs peaking at uv wavelengths ( @xmath6 kk ) .
the most interesting aspect of these binaries is that the secondaries of a 35 and lotr 5 both exhibit enhanced barium abundances ( thvenin & jasniewicz 1997 ) , while this has yet to be demonstrated for lotr 1 . since these initial discoveries bond , pollacco & webbink ( 2003 ) added webo 1 to the list and frew ( 2008 ) raised suspicions that the nebula of a 35 may not be a bona - fide pn .
further additions of hd 330036 and as 201 may be drawn from barium enhanced yellow or d-type symbiotic stars which show extended nebulae ( schmid & nussbaumer 1993 ; pereira et al . 2005 ) , however the classification of such objects as pne is controversial ( corradi 2003 ; jorissen et al . 2005 ) . all these binaries may be classified as barium stars , peculiar red giants characterised by an overabundance of carbon and s - process heavy elements such as barium and strontium ( bidelman & keenan 1951 ) .
they make up @xmath21% of all g k giants , all of them are binaries with apparently wd companions ( mcclure et al .
1980 ) and have orbital periods ranging from 87 up to @xmath22000 days ( jorissen & boffin 1992 ; mcclure & woodsworth 1990 ) . a smaller sample of barium dwarfs are known and are thought to evolved into barium giants , although they are less studied on the whole ( e.g. north , jorissen & mayor 2000 ; gray et al .
2011 ) .
boffin & jorissen ( 1988 ) devised what is now believed to be the canonical model for the formation of these stars . in this scenario the wind of its companion pollutes the future barium - enhanced star , most likely while it is on the main - sequence . this wind transfers carbon and s - process elements dredged up during thermal pulses on the agb .
the agb star then evolved into a wd , which may have produced a pn , while the contaminated star evolved to become a red giant with chemical anomalies : a barium star .
this accounts for the few known pne caught exactly during this phase , but there remains much work to identify and characterise further examples .
none of the three firm examples , a 35 , lotr 5 and webo 1 , have determined orbital periods , while photometric monitoring has revealed the cool components to be rapid rotators with rotation periods of a few days .
the rapid rotation could possibly have been caused by mass accretion ( jeffries & stevens 1996 ; theuns et al .
1996 ) . in this work we present uv and optical observations of the _ _ diamond ring _
_ pn a 70 ( pn [email protected] , abell 1966 ) that prove the existence of a barium star binary cspn . the 2 micron all sky survey ( 2mass , skrutskie et al . 2006 ) recorded an unusually red @xmath7 mag colour for the cspn which suggested a sub - giant or giant classification .
this prompted our investigation of the object even though it may have been possible to reach this conclusion from the original photographic magnitudes on blue ( 18.6 mag ) and red ( 17.4 mag ) plates .
more recent studies of the abell sample lacked colour information for the cspn of a 70 and mostly repeated earlier measurements ( kaler 1983 ; kaler & jacoby 1989 ; kaler , shaw & kwitter 1990 ) .
narrow - band imaging of the nebula was acquired by jewitt , danielson & kupferman ( 1986 ) , schwarz , corradi & melnick ( 1992 ) and most notably hua , dopita & martinis ( 1998 ) .
we obtained spectroscopic observations of a 70 with gemini south under program gs-2009a - q-35 and the vlt under programs 083.d-0654(a ) and 085.d-0629(a ) .
table [ tab : obs ] summarises the 0.7-wide longslit observations made using the gemini multi - object spectrograph ( gmos , hook et al .
2004 ) and the focal reducer and low dispersion spectrograph ( fors2 , appenzeller et al .
1998 ) that had the blue - optimised e2v detector installed .
basic data reduction was performed using the gemini iraf package and the eso fors pipeline after which the iraf task apall was used to trace and extract one - dimensional spectra . in all cases the surrounding nebula emission was subtracted close to the central star .
flux calibration was also applied using spectrophotometric standard stars observed during the respective programs in the usual fashion .
the s / n reached in the continuum near @xmath8 was measured using the splot task in iraf to range between 19 ( fors2 2009 ) and 30 ( gmos b ) .
some narrow - band images were also observed by gmos ( see sect . [
sec : imaging ] ) . [ cols="<,^,<,^,^,^,>,^ " , ] further distances imply a higher @xmath9 , higher @xmath10 , higher stellar luminosities and older @xmath11 , while shorter distances imply a lower @xmath9 , lower @xmath10 , lower stellar luminosities and a younger @xmath11 .
if the ssv08 distance were adopted the stellar luminosities would imply an implausibly massive ionising star for the small associated @xmath11 and @xmath10 would be unusually low for a type i pn ( @xmath10 can reach 0.51 @xmath12 , see frew , parker & russeil 2006 ) .
for these reasons we favour the 5.0 kpc sbr distance , although the accompanying @xmath9 is difficult to reconcile with the only moderately sub - solar oxygen abundance of 8.4 dex ( @xmath13 dex c.f .
solar ) . a shorter distance would alleviate but not solve this problem , although a thick disk progenitor may be one solution to the problem .
the type - i abundance of a 70 also places it against the general trend of decreasing n / o with increasing @xmath9 ( kppen & cuisinier 1997 ) and the empirical finding that most bipolar type - i pne are found at very low @xmath9 ( corradi & schwarz 1995 ) . whether the binary nature of a 70 could explain this incongruity remains to be proven .
irrespective of the adopted distance , we can say that the white dwarf has turned onto the wd cooling track ( e.g. frew et al .
2006 ) and that the luminosity of the companion is less certain .
it could be either g8v ( for shorter distances ) or g8iv ( for larger distances , see sandage et al .
improved central star magnitudes and spectroscopy are required to decide between each luminosity class , so for now we adopt g8iv - v .
the g8iv - v star with ba ii and sr ii enhancement , chromospheric h@xmath1 emission and negligible to low radial velocity amplitude , coupled with the _ galex _ uv detection of the wd , all strongly support the presence of a barium star nucleus in a 70 . having detected a pn around a barium star is quite rare , however a 70 may represent an even more transient stage of stellar evolution if the secondary can be proven to be a sub - giant ( sandage et al .
the evolutionary history of a 70 starts with the slightly more massive component of a binary system with a mass ratio near unity evolving through the agb , and after experiencing thermal pulses , becomes s - process enhanced . through heavy mass loss its
s - process rich wind contaminated its less massive companion and reversed the mass ratio of the system to produce a barium star .
the agb star then reached the pn phase , while its companion left the main sequence to become a subgiant , at which point we observe a 70 the penultimate phase in the making of a barium star ! a 70 thus represents an evolutionary stage just before webo 1 ( bond et al . 2003 ) , which is a carbon and s - process enhanced late - type giant star surrounded by a ring - like planetary nebula . similar to a 70 , webo 1 is chromospherically active and it undergoes rapid rotation every 4.7 days .
such a short period , which is not expected as tidal forces should have synchronised the stars , may be due to spin - up of material accreting from the companion ( theuns et al .
1996 ; jeffries & stevens 1996 ) . given the present state of the system , we can also consider how a 70 will evolve , although the final state will depend on many parameters .
in a few thousand years the nebula will disappear and the hot star will gradually cool off .
soon after this the cool star will ascend the red giant branch ( rgb ) and the system will appear as a common barium star with a peculiar red giant and no visible companion . at the tip of the rgb
the star will experience heavy mass loss by transferring matter to the white dwarf . during this time
it may be observed as a symbiotic system where accreting matter will heat the white dwarf and ionise the wind of the cool giant .
the barium stars have close ties with symbiotic stars as well as with planetary nebulae .
several extrinsic s stars have been found to exhibit symbiotic - like features ( e.g. ake et al .
1991 ) , while several symbiotic stars are also known to present overabundances of s - process elements ( smith et al .
1996 , 1997 ; pereira et al . 2005 ) .
the d-type symbiotic stars are the closest relatives to barium stars ( schmid & nussbaumer 1993 ; pereira et al . 2005 ; jorissen et al . 2005 ) , especially considering some of them are surrounded by apparently _ planetary _ nebulae .
schwarz ( 1991 ) discovered an inner and outer nebula around one of these , as 201 , of which the outer nebula is likely to be a pn ( ejected by the wd dwarf ) .
miszalski et al .
( 2011c ) found symbiotic characteristics in the galactic bulge pn m 2 - 29 , which also exhibits inner and outer nebulae , however the secondary has yet to observed against the glare of the primary .
appendix [ sec : app ] describes the discovery of a bipolar nebula around hd 330036 ( cn 1 - 1 ) , which if considered to be a _ planetary _ nebula , would add further evidence to the link between barium stars with pne and barium enhanced d-type symbiotic stars . observing a system such as a 70 implies that the initial mass ratio must have been close to unity .
the rarity of such a configuration may be used as an argument against the present formation scenario , and indeed an identical approach was taken by corradi ( 2003 ) concerning the potential presence of mira secondaries in pne .
such an a 70-like configuration can in principle occur since lucy ( 2006 ) found an excess of so - called twins , i.e. systems with mass ratios between 0.98 and 1 .
as further binaries similar to a 70 and webo 1 are found , then these probabilities may have to be revised in favour of a greater frequency of twins .
this would foster a greater overlap between symbiotic stars ( at least those of yellow or d-type ) and pne ( jorissen et al . 2005 ) , with the formal difference becoming notional .
wide binaries will interact and produce genuine pne in a wide variety of cases , however only a relatively narrow range is currently observed ( e.g. de marco 2009 ) .
the measurement of nebular s - process abundances in pne has considerable potential to improve agb nucleosynthesis models ( e.g. sterling & dinerstein 2008 , hereafter sd08 ; karakas et al .
2009 ; karakas & lugaro 2010 ) .
these abundances are a valuable constraint upon the number of third dredge - up episodes experienced during the thermally - pulsing agb phase .
of particular interest are type - i pne whose he- and n - rich abundances are well - reproduced in models that require a progenitor mass @xmath14 m@xmath15 to achieve hot bottom burning . quantifying the stellar s - process abundances of a 70 via high - resolution spectroscopy will therefore be of great interest to compare against the type - i nebula abundance .
the high surface brightness of a 70 compared to e.g. webo 1 may also allow nebula s - process abundances to be measured via nir spectroscopy . in principle
the s - process abundances of type - i pne should be relatively straight - forward to understand , however no firm patterns have been found so far ( karakas et al .
sd08 noted that binary interactions may be responsible for reducing their measured s - process abundances , however the paucity of known binaries in their sample meant this remained untested .
new discoveries of binary central stars in the sd08 sample should therefore help resolve the issue .
miszalski et al . ( 2011b ) recently found a close binary in ngc 6778 which is a bipolar type - i pn in the sd08 sample with strong he and n enhancement ( he / h=0.155 and n / [email protected] , perinotto , morbidelli & scatarzi 2004 ) .
it is possible that post - ce binaries may reduce s - process abundances to a greater extent than wider binary systems , making wider binaries a potentially more reliable probe of s - process abundances in pne .
the growing sample of a 70 and webo 1 as barium stars , and perhaps also me 1 - 1 , therefore provides a powerful alternative to analysis of close binaries in the sd08 sample .
the discovery of the cool cspn of a 70 has important implications for further survey work to find similar binaries . to establish a physical connection between a cool cspn candidate and a pn a number of factors
must be considered .
the two most important being : ( i ) a uv excess or spectroscopic features of the wd must be detected to rule out the born - again scenario , and ( ii ) agreement between the radial velocities of the nebula and the cool star ( modulo the expected radial velocity amplitude of a binary ) .
it is also preferable to have a small nebular diameter or equivalently an uncrowded field to help rule out a superposition ( e.g. ciardullo et al .
a favourable spectroscopic distance should also preferably agree well with the approximate statistical nebula distance .
a general lack of deep enough uv observations has made the task of confirming hot components in pne difficult .
sahai et al .
( 2008 ) found 9/21 agb stars to have fuv excesses which they attribute to a hot companion , but there has not been a systematic uv - excess survey of pne .
the only other object studied so far is k 1 - 6 ( frew et al . 2011 ) , however no spectroscopic data were presented in this study .
we also advise against searching too hard for a cool companion , especially in large pne where many candidates may be present .
it is certainly possible that an intrinsically very faint @xmath17 mag single wd nucleus may be beyond the detection limit of typical optical surveys , and this is especially true for the most distant pne .
this emphasises the crucial role of uv photometry to confirm such a 70-like binaries if they were located in more crowded stellar fields .
on the other hand , the unique spectroscopic signature of a barium star is also a sound means to secure a binary confirmation and multi - object spectroscopy may be an efficient tool in this respect .
we have presented 8-m optical spectroscopic and imaging observations of the unique diamond - ring pn a 70 . combined with _
uv photometry , the data prove the binary nature of the central star which is expected to have an expected orbital period of a few hundred days .
our main conclusions are as follows : * optical spectroscopy of the central star revealed a g8iv - v star , slightly enhanced in s - process elements , with chromospheric h@xmath1 emission .
_ galex _ uv photometry detected the signature of a wd coincident with the central star , therefore providing the first clear proof for a barium dwarf binary inside a pn .
this is firm evidence for the standard formation scenario of barium stars since barium dwarfs are generally difficult to observe ( e.g. gray et al .
* radial velocities of the g8iv - v secondary were analysed in comparison to the @xmath18 km / s heliocentric nebula velocity of a 70 .
fors2 observations are consistent with the nebula velocity proving they are physically connected , while one gmos observation presented a difference of 24 km / s .
such a difference is of the same magnitude expected for orbital motion , if indeed this one measurement can be trusted .
further observations are required at higher resolution to measure the orbital period and rv amplitude . *
chemical abundances of the nebula were measured and found to have a type - i composition with strong he and n enrichment .
as the g8iv - v companion is also s - process enhanced , this makes a 70 an exceptional laboratory for further improving our understanding of agb nucleosynthesis and the origin of type - i pne ( e.g. karakas et al . 2009 ) . *
the distance remains uncertain , however it is clear that the wd has evolved onto the cooling track .
a plausible 5 kpc distance would imply a height below the galactic plane of @xmath19 kpc , in contradiction with the type - i composition and the only slightly sub - solar oxygen abundance ( @xmath13 dex c.f .
solar ) . it may be possible
that a thick disk origin or binary stellar evolution could explain the type - i composition , but this remains to be proven with improved observations .
* a 70 is one of very few pne with apparently cool central stars to have the wd detected in the uv .
the observations presented here strongly suggest that insensitive uv observations were responsible for the non - detections of bond & pollacco ( 2002 ) and the conclusion that the companions are single nuclei caught during the ` born - again ' phase .
a 70 also serves as a template to guide future discoveries of similar binary central stars .
we acknowledge the conscientious and helpful support of simon otoole and steve margheim during the course of our gemini program .
bm thanks observatoire astronomique de strasbourg for travel support towards the later stages of this work and r. napiwotzki for assistance in building the sed of a 70 .
hb thanks leo rivas and yuri beletsky for their dedication and excellence at the vlt . their persistence in staying until the last moments of the night in the control room
is especially appreciated .
afjm is grateful to nserc ( canada ) and fqrnt ( quebec ) for financial assistance .
synphot is a product of the space telescope science institute , which is operated by aura for nasa .
based on observations obtained at the gemini observatory , which is operated by the association of universities for research in astronomy , inc .
, under a cooperative agreement with the nsf on behalf of the gemini partnership : the national science foundation ( united states ) , the science and technology facilities council ( united kingdom ) , the national research council ( canada ) , conicyt ( chile ) , the australian research council ( australia ) , ministrio da cincia e tecnologia ( brazil ) and ministerio de ciencia , tecnologa e innovacin productiva ( argentina ) . 99 abell , g. o. 1966 , apj , 144 , 259 acker , a. , & jasniewicz , g. 1990 , a&a , 238 , 325 acker , a. , raytchev , b. , kppen , j. , & stenholm , b. 1991 , a&as , 89 , 237 acker , a. , marcout , j. , ochsenbein , f. , stenholm , b. , & tylenda , r. 1992 , garching : european southern observatory ake , t. b. , iii , johnson , h. r. , & ameen , m. m. 1991 , apj , 383 , 842 appenzeller , i. , et al .
1998 , the messenger , 94 , 1 asplund , m. , grevesse , n. , & sauval , a. j. 2005 , cosmic abundances as records of stellar evolution and nucleosynthesis , 336 , 25 bagnulo , s. , jehin , e. , ledoux , c. , cabanac , r. , melo , c. , gilmozzi , r. , & the eso paranal science operations team 2003 , the messenger , 114 , 10 bidelman , w. p. , & keenan , p. c. 1951 , apj , 114 , 473 boffin , h. m. j. , & jorissen , a. 1988 , a&a , 205 , 155 bond , h. e. , & livio , m. 1990 , apj , 355 , 568 bond , h. e. , ciardullo , r. , & meakes , m. g. 1993 , planetary nebulae , 155 , 397 bond , h. e. , & pollacco , d. l. 2002 , ap&ss , 279 , 31 bond , h. e. , pollacco , d. l. , & webbink , r. f. 2003 , aj , 125 , 260 cardelli , j. a. , clayton , g. c. , & mathis , j. s. 1989 , apj , 345 , 245 ciardullo , r. , bond , h. e. , sipior , m. s. , fullton , l. k. , zhang , c .- y . , &
schaefer , k. g. 1999 , aj , 118 , 488 corradi , r. l. m. 2003 , astronomical society of the pacific conference series , 303 , 393 davis , p. j. , kolb , u. , & willems , b. 2010 , mnras , 403 , 179 de marco , o. 2009 , pasp , 121 , 316 de marco , o. , farihi , j. , & nordhaus , j. 2009 , journal of physics conference series , 172 , 012031 epchtein , n. , et al . 1999 ,
a&a , 349 , 236 fouqu , p. , et al .
2000 , a&as , 141 , 313 frew , d. j. , & parker , q. a. 2006 , planetary nebulae in our galaxy and beyond , 234 , 49 frew , d. j. , parker , q. a. , & russeil , d. 2006 , mnras , 372 , 1081 frew , d. j. 2008 , ph.d .
thesis , macquarie university frew , d. j. , et al .
2011 , pasa , 28 , 83 gawryszczak , a. j. , mikoajewska , j. , & ryczka , m. 2002 , a&a , 385 , 205 girard , p. , kppen , j. , & acker , a. 2007 , a&a , 463 , 265 gonalves , d. r. , corradi , r. l. m. , & mampaso , a. 2001 , apj , 547 , 302 gray , r. o. , mcgahee , c. e. , griffin , r. e. m. , & corbally , c. j. 2011 , aj , 141 , 160 han , z. , podsiadlowski , p. , & eggleton , p. p. 1995 ,
mnras , 272 , 800 hook , i. m. , jrgensen , i. , allington - smith , j. r. , davies , r. l. , metcalfe , n. , murowinski , r. g. , & crampton , d. 2004 , pasp , 116 , 425 howarth , i. d. 1983 , mnras , 203 , 301 hua , c. t. , dopita , m. a. , & martinis , j. 1998 , a&as , 133 , 361 iben , i. , jr . , &
livio , m. 1993 , pasp , 105 , 1373 jeffries , r. d. , & stevens , i. r. 1996 , mnras , 279 , 180 jewitt , d. c. , danielson , g. e. , & kupferman , p. n. 1986 , apj , 302 , 727 jones , d. h. , et al . 2004 ,
mnras , 355 , 747 jorissen , a. , & boffin , h. m. j. 1992 , binaries as tracers of star formation , 110 jorissen , a. , zas , l. , udry , s. , lindgren , h. , & musaev , f. a. 2005 , a&a , 441 , 1135 kaler , j. b. 1983 , apj , 271 , 188 kaler , j. b. , & jacoby , g. h. 1989 , apj , 345 , 871 kaler , j. b. , shaw , r. a. , & kwitter , k. b. 1990 , apj , 359 , 392 karakas , a. i. , tout , c. a. , & lattanzio , j. c. 2000 , mnras , 316 , 689 karakas , a. i. , van raai , m. a. , lugaro , m. , sterling , n. c. , & dinerstein , h. l. 2009 , apj , 690 , 1130 karakas , a. i. , & lugaro , m. 2010 , pasa , 27 , 227 kingsburgh , r. l. , & barlow , m. j. 1994 , mnras , 271 , 257 kppen , j. , acker , a. , & stenholm , b. 1991 , a&a , 248 , 197 kppen , j. , & cuisinier , f. 1997 , a&a , 319 , 98 kohoutek , l. 1997 , a&as , 125 , 445 kurtz , m. j. , & mink , d. j. 1998 , pasp , 110 , 934 l , p. k. , dawson , d. w. , upgren , a. r. , & weis , e. w. 1983 , apjs , 52 , 169 lucy , l. b. 2006 , a&a , 457 , 629 lutz , j. h. 1977 , a&a , 60 , 93 martin , d. c. , et al . 2005 , apjl , 619 , l1 mastrodemos , n. , & morris , m. 1999 , apj , 523 , 357 maxted , p. f. l. , gnsicke , b. t. , burleigh , m. r. , southworth , j. , marsh , t. r. , napiwotzki , r. , nelemans , g. , & wood , p. l. 2009 , mnras , 400 , 2012 mcclure , r. d. , fletcher , j. m. , & nemec , j. m. 1980 , apjl , 238 , l35 mcclure , r. d. , & woodsworth , a. w. 1990 , apj , 352 , 709 meatheringham , s. j. , wood , p. r. , & faulkner , d. j. 1988 , apj , 334 , 862 miranda , l. f. , ramos - larios , g. , & guerrero , m. a. 2010 , pasa , 27 , 180 miszalski , b. , acker , a. , moffat , a. f. j. , parker , q. a. , & udalski , a. 2009a , a&a , 496 , 813 miszalski , b. , acker , a. , parker , q. a. , & moffat , a. f. j. 2009b , a&a , 505 , 249 miszalski , b. , corradi , r. l. m. , jones , d. , santander - garca , m. , rodrguez - gil , p. , & rubio - dez , m. m. 2011a , asymmetric planetary nebulae v , arxiv:1009.2890 miszalski , b. , jones , d. , rodrguez - gil , p. , boffin , h. m. j. , corradi , r. l. m. , & santander - garca , m. 2011b , a&a , 531 , a158 miszalski , b. , et al . 2011c , a&a , 528 , a39 morrissey , p. , et al .
2007 , apjs , 173 , 682 north , p. , jorissen , a. , & mayor , m. 2000 , the carbon star phenomenon , 177 , 269 oke , j. b. 1990 , aj , 99 , 1621 parker et al .
2005 , mnras , 362 , 689 peimbert , m. , & torres - peimbert , s. 1983 , planetary nebulae , 103 , 233 pereira , c. b. , smith , v. v. , & cunha , k. 2005 , a&a , 429 , 993 pereira , c. b. , miranda , l. f. , smith , v. v. , & cunha , k. 2008 , a&a , 477 , 535 perinotto , m. , purgathofer , a. , pasquali , a. , & patriarchi , p. 1994
, a&as , 107 , 481 perinotto , m. , morbidelli , l. , & scatarzi , a. 2004 , mnras , 349 , 793 pickles , a. j. 1998 , pasp , 110 , 863 rebassa - mansergas , a. , et al .
2008 , mnras , 390 , 1635 sahai , r. , findeisen , k. , gil de paz , a. , & snchez contreras , c. 2008 , apj , 689 , 1274 sandage , a. , lubin , l. m. , & vandenberg , d. a. 2003 , pasp , 115 , 1187 schmid , h. m. , & nussbaumer , h. 1993 , a&a , 268 , 159 schreiber , m. r. , gnsicke , b. t. 2003 , a&a , 406 , 305 schwarz , h. e. 1991 , a&a , 243 , 469 schwarz , h. e. , corradi , r. l. m. , & melnick , j. 1992 , a&as , 96 , 23 shen , z .- x . , liu , x .- w . , & danziger , i. j. 2004 , a&a , 422 , 563 skrutskie , m. f. , et al . 2006 ,
aj , 131 , 1163 smith , v. v. , cunha , k. , jorissen , a. , & boffin , h. m. j. 1996 , a&a , 315 , 179 smith , v. v. , cunha , k. , jorissen , a. , & boffin , h. m. j. 1997 , a&a , 324 , 97 sterling , n. c. , & dinerstein , h. l. 2008 , apjs , 174 , 158 ( sd08 ) stanghellini , l. , shaw , r. a. , & villaver , e. 2008 , apj , 689 , 194 ( ssv08 ) theuns , t. , boffin , h. m. j. , & jorissen , a. 1996 , mnras , 280 , 1264 thvenin , f. , & jasniewicz , g. 1997 , a&a , 320 , 913 tylenda , r. , sidmiak , n. , grny , s. k. , corradi , r. l. m. , & schwarz , h. e. 2003 , a&a , 405 , 627 van winckel , h. 2003a , ara&a , 41 , 391 van winckel , h. 2003b , astronomical society of the pacific conference series , 303 , 294 van winckel , h. , et al .
2010 , mmsai , 81 , 1022 yungelson , l. r. , tutukov , a. v. , & livio , m. 1993 , apj , 418 , 794
the aao / ukst supercosmos h@xmath1 survey ( shs , parker et al .
2005 ) is a deep 4000 deg@xmath20 h@xmath1 and broadband red ( short - red ) photographic survey of the southern galactic plane .
figure [ fig : bipolar ] depicts a faint bipolar nebula in the shs data surrounding hd 330036 ( pn [email protected] , also known as cn 1 - 1 ) found during the course of this work .
this is an entirely new discovery since previous studies were concerned with the bright emission - line core only and were not sensitive enough to reveal the very low surface brightness lobes ( e.g. kohoutek 1997 ) .
the flattened @xmath21-shape is uncannily similar to the ` outer lobes ' of hen 2 - 104 ( corradi 2003 ) , suggesting a similar process was responsible for their formation . at the 2.3 kpc distance estimated by pereira et al .
( 2005 ) the lobes would measure 1.2 pc tip - to - tip .
-shaped lobes are seen around hd 330036 ( cn 1 - 1 ) in this @xmath22 arcmin@xmath20 continuum - divided h@xmath1 image ( north is up and east to left ) .
the opposing parts of the lobes align along pa@xmath2@xmath23 and pa@xmath2@xmath24 and measure 105110 tip - to - tip . ] | abell 70 ( pn [email protected] , hereafter a 70 ) is a planetary nebula ( pn ) known for its diamond ring appearance due a superposition with a background galaxy .
the previously unstudied central star is found to be a binary consisting of a g8iv - v secondary at optical wavelengths and a hot white dwarf ( wd ) at uv wavelengths .
the secondary shows ba ii and sr ii features enhanced for its spectral type that , combined with the chromospheric h@xmath1 emission and possible 2030 km / s radial velocity amplitude , firmly classifies the binary as a barium star . the proposed origin of barium stars is intimately linked to pne whereby wind accretion pollutes the companion with dredged - up material rich in carbon and s - process elements when the primary is experiencing thermal pulses on the asymptotic giant branch ( agb ) .
a 70 provides further evidence for this scenario together with the other very few examples of barium central stars .
the nebula is found to have type - i chemical abundances with helium and nitrogen enrichment , which when combined with future abundance studies of the central star , will establish a 70 as a unique laboratory for studying s - process agb nucleosynthesis .
we also discuss guidelines to discover more binary central stars with cool secondaries in large orbits that are required to balance our knowledge of binarity in pne against the currently better studied post common - envelope binary central stars .
[ firstpage ] planetary nebulae : individual : pn g038.1 - 25.4 - planetary nebulae : general - stars : chemically peculiar - stars : agb and post - agb - binaries : general - binaries : symbiotic |
auroral intensification has been vigorously studied with narrow field of view or high resolution optical measurements [ e.g. , donovan et al .
, 2007 ; mende et al . , 2009 ] .
it is known that the time scale of intensification is 12 min , and fading of auroral arcs appears just before the onset .
furthermore , vortex street structures called as `` beads '' emerged on the auroral arc , probably due to a development of strong flow shears .
it has been considered that , on the basis of the above facts , shear alfv@xmath0n waves propagating along the magnetic field line are involved in auroral intensification .
three - dimensional simulations that treated amplification of shear alfv@xmath0n waves , driven by a coupling between the ionospheric and magnetospheric plasmas with field - aligned currents , were performed [ jia and streltsov , 2014 ; hiraki , 2015a , b ] .
hiraki [ 2015b ] presented that a threshold exists for the instability of alfv@xmath0n waves due to an enhancement of plasma convection ; the threshold electric field was @xmath1 mv / m in their situation .
these studies did not explicitly treat the electron acceleration by field - aligned electric fields and the ionization of neutral atmosphere , which are the essence of discrete auroras .
they assumed that the upward field - aligned current carried by waves is identical to the auroral luminosity .
measurements of electron density by eiscat radar [ kirkwood et al . , 1990 ; olsson et al . , 1996 ] and of particles by dmsp and fast satellites [ yago et al . , 2005 ; mende et al . ,
2003 ] revealed that the field - aligned electric field originates from inverted - v type potential drops at a height of @xmath2 km or inertial alfv@xmath0n waves excited at @xmath3 km .
energies of electrons accelerated by each electric field are 210 kev or @xmath4 kev , respectively . although previous simulations treated inertial alfv@xmath0n waves [ lysak and song , 2008 ; jia and streltsov , 2014 ] , evolution of waves with a realistic field - aligned electric field is not still resolved .
this study examined the other topic , i.e. , the effect of ionization by inverted - v type non - thermal electrons on the instability of shear alfv@xmath0n waves .
these electrons with an energy of 210 kev strongly increase the conductivities in a height of 100120 km where horizontal currents that can couple with shear alfv@xmath0n waves maximize [ fang et al .
, 2008 ] .
the ionization rate by the non - thermal electrons was formulated as @xmath5 , @xmath6 : electron yielding coefficient , @xmath7 : field - aligned current by alfv@xmath0n waves [ lysak and song , 2002 ] .
however , its effect on the growth of shear alfv@xmath0n waves was not clarified .
this study analyzed the linear instability of shear alfv@xmath0n waves , including effects of ionization , in the magnetosphere - ionosphere ( m - i ) coupling system .
we especially focused on the possibility that the threshold field value ( @xmath1 mv / m ) presented in our previous study is reduced to be a popular level in observations , e.g. , 1020 mv / m [ provan et al .
, 2004 ; bristow and jensen , 2007 ] .
the target region of our analysis is a magnetic flux tube that involves an auroral arc with a width of @xmath8 km .
shear alfv@xmath0n waves , propagating along the magnetic field right above the arc , flow the field - aligned current @xmath7 into the ionosphere .
it connects with horizontal currents caused by plasma convection .
a situation is considered where the coupling of currents destabilizes shear alfv@xmath0n waves , implying the initial brightening of aurora .
the field line position @xmath9 is defined as @xmath10 at the ionosphere and @xmath11 at the magnetic equator .
we consider a latitude of @xmath12 in the southern hemisphere , with the dipole @xmath13 value of @xmath14 and @xmath15 km .
we take coordinates @xmath16 and @xmath17 orthogonal to each @xmath18 : @xmath19 points southward ( poleward ) and @xmath20 points eastward at @xmath10 .
we set a local flux tube , e.g. , a square of ( @xmath21 ) at @xmath10 and a rectangle of ( @xmath22 km @xmath23 @xmath24 km ) at @xmath11 using @xmath25 km and dipole metrics ; see hiraki and watanabe [ 2011 ] for details .
the dipole magnetic field is written as @xmath26 .
the system has a convective electric field @xmath27 that is applied poleward ( @xmath28 ) and is uniform in every @xmath29-@xmath30 planes . with perturbed electric and magnetic fields of @xmath31 and @xmath32 ,
linearized equations for shear alfv@xmath0n waves are expressed as @xmath33 the domain of definition is @xmath34 . here
, @xmath35 stands for vorticity , @xmath36 field - aligned current , @xmath37 the convection speed , and @xmath38 the alfv@xmath0n velocity . the same as hiraki [ 2015b ] ,
@xmath38 is set to be constant as @xmath39 km / s along the field line .
the alfv@xmath0n transit time is thus @xmath40 s. by integrating the continuity equations of ions and electrons , the linearized equations @xmath41 are obtained for plasma motion at the ionosphere @xmath10 . here
, @xmath42 stands for electron density , @xmath43 its ambient component , @xmath44 pedersen mobility , @xmath45 the recombination coefficient , and @xmath46 the diffusion coefficient . the hall mobility appearing in eqs .
( 3 ) and ( 4 ) is normalized to be unity .
these equations couple with eqs .
( 1 ) and ( 2 ) of shear alfv@xmath0n waves through field - aligned current @xmath7 .
the electron yielding rate was set to be @xmath47 in hiraki [ 2015b ] by assuming continuity of the thermal electron flux . in this study , @xmath48 is given on the supposition that ionization of neutral particles occurs due to the non - thermal electron precipitation . in order to solve eqs .
( 1)(4 ) , field variables @xmath49 , @xmath50 , and @xmath42 are expanded such as @xmath51 with complex frequency @xmath52 and perpendicular wave number @xmath53 .
the boundary condition at the magnetic equator is set to be @xmath54 , or @xmath55 ( zero current ) , and shapes of eigenfunctions @xmath56 and @xmath57 at @xmath10@xmath58 are calculated .
we can redefine the frequency @xmath59 in the frame of @xmath60 to be @xmath52 without loss of generality . from eqs .
( 3 ) and ( 4 ) , we get the dispersion relation as @xmath61 here , two parameters @xmath62 and @xmath63 are introduced .
the former means the convective drift frequency , while the latter corresponds to the pedersen / alfv@xmath0n impedance ratio @xmath64 .
the ionospheric parameters used for the solution of eq .
( 5 ) are set to be @xmath44/@xmath65 , @xmath66 , @xmath67
s@xmath68 , and @xmath69 m@xmath70/s . it is also noted that the ambient density and magnetic field are set to be @xmath71 @xmath72 and @xmath73 t. of finite @xmath74 and @xmath75 modes , normalized by @xmath76 , in the nominal case of the electron yielding coefficient @xmath47 .
the convection electric field is changed to be @xmath77 ( red ) , 15 , 20 , 25 ( purple ) , 30 , 35 , 40 , and 45 ( black ) mv / m . ] -@xmath6 diagram of the growth rate @xmath78 of @xmath79 mode ; open circles show the cases of @xmath80 and crosses show the cases of @xmath81 .
a function @xmath82 is fitted to the open - cross boundary . ] , ( b ) growth rate @xmath78 , and ( c ) eigenfunctions @xmath83 on the electron yielding rate @xmath6 for @xmath79 mode along with @xmath77 mv / m . ]
the simplest case is that there are an auroral arc in the ionosphere and alfv@xmath0n waves right above it .
the @xmath84 mode is distinguished in the arc if its spatial scale is approximated to the gaussian distribution ; hereafter , wave number is normalized by @xmath85 .
the arc without east - westward structures , i.e. , @xmath75 , is considered for simplicity .
figure 1 shows the linear growth rates @xmath86 of ( @xmath87 , @xmath75 ) modes that were obtained by solving eqs .
( 1 ) , ( 2 ) , and ( 5 ) under the conditions in sec . [ sec : 2 ] . here , the electron yielding rate being @xmath47 , the response of @xmath78 to changes in the poleward electric field @xmath88 was found .
we readily found the value of the critical field for @xmath89 to be @xmath90 mv / m .
hiraki [ 2015b ] demonstrated that an arc quickly splits into two arcs due to a density perturbation produced by the poleward pedersen current .
for the higher mode @xmath91 , we also evaluated the critical field for @xmath92 to be @xmath93 mv / m ( fig . 1 ) .
it is , however , noted that a lower electric field , e.g. 10 mv / m , just before auroral intensification was estimated with superdarn radars [ provan et al . , 2004 ] .
we investigate how the critical field for this ( 2 , 0 ) mode responds when @xmath6 ( @xmath94 ) is changed .
figure 2 shows a diagram of @xmath95 in ( @xmath88 , @xmath6 ) space : open circle stands for the case of @xmath92 , while cross for the case of @xmath96 .
it is clearly found that the critical value of @xmath88 for @xmath92 can be reduced if @xmath6 is increased .
note that the term proportional to @xmath97 appears in the left - hand side of eq .
( 5 ) in case of @xmath98 .
it means that a mode is unstable for any set of @xmath6 and @xmath88 that satisfies @xmath99 ; the diffusion term ( @xmath46 ) slightly changes the relation to a fitting function @xmath100 for the open circle - cross boundary in fig . 2 .
because the @xmath6 value has an upper limit in the real situation just before a faint arc starts to intensify , the critical @xmath88 necessarily has a lower limit .
we will discuss in sec .
[ sec : 4 ] on the way to estimate the realistic range of @xmath6 values and the reason why we remove the possibility of higher modes @xmath101 . at the end of this section
, we describe the characteristics of eigenmodes in the moderate case of electric field @xmath77 mv / m .
figure 3 shows the responses of real frequency @xmath102 , growth rate @xmath103 , and eigenfunction @xmath83 of the @xmath91 mode to a change in @xmath6 .
the gradient of @xmath104 steepens in the lower @xmath74 side , owing to @xmath105 in eq .
related to this change as @xmath6 increases , a sharp peak in @xmath78 appears and moves to the lower @xmath74 side .
a high @xmath78 in the case of high @xmath6 is supported by the fact that the amplitude of eigenfunction @xmath83 at @xmath10 increases .
let us first mention why the @xmath79 mode is chosen when the critical field @xmath106 for auroral intensification is proposed at the part of fig .
2 . as pointed in secs .
[ sec : 1 ] and [ sec : 3 ] , the critical value @xmath107 mv / m for the @xmath108 mode in case of @xmath47 is likely to be much higher than the typical value in observations [ bristow and jensen , 2007 ] .
the growth rate increases with the mode number , and the critical field for @xmath109 is @xmath110 mv / m ( fig . 1 ) .
however , we should start a speculation that @xmath91 , @xmath111 modes are produced through successive splits of an arc .
it may require a very long time for the @xmath109 mode to emerge to cause an instability involving alfv@xmath0n waves .
we suppose that an arc with a width of @xmath112 ( @xmath8 km ) , splitting into two arcs , requires to move at least a distance of @xmath112 by a drift in the critical field @xmath113 . producing @xmath114-arcs requires a distance of @xmath115 .
the condition for this @xmath116-splitting is that the time scale @xmath117 is shorter than the time scale @xmath118 representing field changes .
this relation is expressed as @xmath119 here , @xmath120 s and we determined @xmath121 and 17 mv / m ( @xmath122 and 0.34 km / s ) for @xmath91 and @xmath123 modes . substituting these values into eq .
( 6 ) , we calculated @xmath124 and 58 s , respectively , and the latter does not satisfy the above inequality .
thus , we are compelled to remove the possibility of the instability with higher modes @xmath101 .
it is , however , noted that the @xmath123 mode can be preferable rather than the @xmath91 mode in the other specific case of eq .
( 6 ) with a stretched magnetic field . for motivations of future observations of electron energy spectrum just before auroral intensification , we illustrated the way to estimate a realistic @xmath6 value . equating it ,
@xmath125 which meets @xmath126 . here ,
@xmath127 stands for injected electron energy ( ev ) , @xmath128 the mean energy loss per ion pair production , and @xmath129 the ratio of the injected ( non - thermal ) electron flux to the thermal electron flux carrying alfv@xmath0nic @xmath7 . the typical value of @xmath128 was estimated to be 35 ev [ fang et al . , 2008 ; and references therein ] .
generally , @xmath130 since we assumed the ionization term in eq .
( 3 ) as a product @xmath5 of @xmath6 and @xmath7 proportional to the thermal electron flux .
referring to the precipitating electron model by fang et al . [
2008 ] , we pick energies of @xmath131 or 10 kev that produce peaks in the ionization rate at our target region of 100120 km . for these energies , @xmath132 and @xmath133 , respectively . we assume measured energy distributions with eiscat radars [ kirkwood et al . , 1990 ; olsson et al . , 1996 ] and
dmsp satellite [ yago et al . , 2005 ] as a 1 kev maxwellian distribution , and roughly estimate the energy ratio @xmath129 to be up to @xmath134 . in this case , @xmath135 and 2.8 are obtained for 5 and 10 kev electrons , respectively .
the above estimation of @xmath6 values is supported by the assumption that an inverted - v potential right above the arc keeps a robust energy spectrum including accelerated electrons .
however , it has a possibility to change the @xmath6 value ( even a slight change ) if the inertial alfv@xmath0n wave supplies @xmath4 kev accelerated electrons ; here , we refer to the stopping height of 1 kev electrons being @xmath136 km [ fang et al . ,
although we did not consider any dependence of @xmath6 on @xmath7 , @xmath6 could correlate with @xmath7 of shear alfv@xmath0n waves through @xmath137 of inertial alfv@xmath0n waves .
the linear form of @xmath5 may be corrected in our future studies that treat both waves .
our prime suggestion for future observations is that an increase in the ionization rate by non - thermal electrons can reduce the threshold field , to a great extent , of instabilities of alfv@xmath0n waves related to auroral intensification .
however , compared to the average energy component carrying @xmath7 , the amount of the non - thermal component is limited .
the @xmath6 value has an upper limit , and thus @xmath106 value has a lower limit .
superdarn radar observations revealed that the convection speed in the vicinity of arcs exceeds @xmath138 m / s over @xmath139 min before onsets of auroral intensification [ bristow and jensen , 2007 ] ; the time scale ( @xmath140 ) is reasonable for shear alfv@xmath0n waves to be amplified [ see hiraki , 2015b ] . with the magnetic field of @xmath73 t
, the speed corresponds to the electric field of @xmath141 mv / m .
the rate @xmath1424 is sufficient for instabilities of the waves , if this field value is regarded as the threshold .
however , we need the higher rate @xmath1436 , if future precise and statistical observations will present a lower value , e.g. , @xmath144 mv / m . we urge fascinating field / particle observations to determine both the lower limit of threshold fields and the upper limit of @xmath6 values .
this study investigated changes in the threshold electric field @xmath106 of plasma convection , causing a linear instability of alfv@xmath0n waves , due to the ionization ( @xmath6 ) by precipitating auroral electrons .
we clarified that the value of @xmath106 is reduced in the cases of @xmath145 .
a realistic value of @xmath146 , sufficient for a fast growth of waves , was obtained from observed electron energy spectra . in future ,
our results motivate the thorough search for the @xmath6 value with electron energy spectra at auroral breakup and precise measurements of convection electric fields by radars .
bristow , w. a. , and p. jensen ( 2007 ) , a superposed epoch study of superdarn convection observations during substorms , _ j. geophys .
, 112 , _ a06232 ,
donovan , e. , s. mende , b. jackel , m. syrj@xmath147suo , m. meurant , i. voronkov , h. u. frey , v. angelopoulos , and m. connors ( 2007 ) , the azimuthal evolution of the substorm expansive phase onset aurora , _ in proc . of int .
substorms-8 , _ edited by m. syrj@xmath147suo and e. donovan , pp .
55 - 60 , university of calgary , alberta , canada .
fang , x. , c. e. randall , d. lummerzheim , s. c. solomon , m. j. mills , d. r. marsh , c. h. jackman , w. wang , and g. lu ( 2008 ) , electron impact ionization : a new parameterization for 100 ev to 1 mev electrons , _ j. geophys .
, 113 , _ a09311 ,
hiraki , y. , and t .- h .
watanabe ( 2011 ) , feedback instability anal- ysis for dipole configuration with ionospheric and magnetospheric cavities , _ j. geophys .
res . , 116 , _
a11220 , doi:10.1029/2011ja016721 .
hiraki , y. ( 2015a ) vortices at the magnetic equator generated by hybrid alfv@xmath0n resonant waves , _ phys .
plasmas , 22 , _ 012903 .
hiraki , y. ( 2015b ) , auroral vortex street formed by the magnetosphere - ionosphere coupling instability , _ ann .
, 33 , _ 217224 , doi:10.5194/angeo-33 - 217 - 2015 .
jia , n. , and a. v. streltsov ( 2014 ) , ionospheric feedback instability and active discrete auroral forms , _ j. geophys .
res . , 119 , _ doi:10.1002/2013ja019217 . s. kirkwood , and l. eliasson ( 1990 ) energetic particle precipitation in the substorm growth phase measured by eiscat and viking , _ j. geophys .
, 95 , _ 60256037 .
lysak , r. l. , and y. song ( 2002 ) energetics of the ionospheric feedback interaction , _
j. geophys .
res . , 107 , _ 1160 , doi:10.1029/2001ja000308 .
lysak , r. l. , and y. song ( 2008 ) propagation of kinetic alfv@xmath0n waves in the ionospheric alfv@xmath0n resonator in the presence of density cavities , _ geophys .
lett . , 35 , _
l20101 , doi:10.1029/2008gl035728 .
mende , s. b. , c. w. carlson , h. u. frey , l. m. peticolas , and n. @xmath148stgaard , fast and image - fuv observations of a substorm onset , _
j. geophys .
res . , 108(a9 )
, _ 1344 , doi:10.1029/2002ja009787 , 2003 .
mende , s. , v. angelopoulos , h. u. frey , e. donovan , b. jackel , k .- h .
glassmeier , j. p. mcfadden , d. larson , and c. w. carlson ( 2009 ) timing and location of substorm onsets from themis satellite and ground based observations , _ ann .
, 27 , _ 28132830 , doi:10.5194/angeo-27 - 2813 - 2009 .
a. olsson , m.a.l .
persson , h.j .
opgenoorth , and s. kirkwood ( 1996 ) particle precipitation in auroral break - ups and westward traveling surges , _ j. geophys .
res . , 101 , _ 24,66124,673 .
provan , g. , m. lester , s. b. mende , and s. e. milan ( 2004 ) , statistical study of high - latitude plasma flow during magnetospheric substorms , _ ann .
, 22 , _ 36073624 .
yago , k. , k. shiokawa , k. hayashi , and k. yumoto ( 2005 ) , auroral particles associated with a substorm brightening arc , _ geophys .
lett . , 32 , _
l06104 , doi:10.1029/2004gl021894 . | it has been known that discrete aurora suddenly intensifies and deforms from an arc - like to a variety of wavy / vortex structures , especially during a substorm period .
the instability of alfv@xmath0n waves reflected from the ionosphere has been analyzed in order to comprehend the ignition process of auroral intensification .
it was presented that the prime key is an enhancement of plasma convection , and the convection electric field has a threshold .
this study examined effects of auroral electron precipitation , causing the ionization of neutral atmosphere , on the linear instability of alfv@xmath0n waves .
it was found that the threshold of convection electric fields is significantly reduced by increasing the ionization rate , the realistic range of which could be estimated from observed electron energy spectra . |
this year is the centenary of the 1905 einstein _ agnus mirabilis _ @xcite . in that year
he published five important papers for the future of the physics .
one of them @xcite was about the photoelectric effect , and for it einstein won the 1921 nobel prize in physics ; two of them @xcite were about statistic mechanics , and the others two on what was later known as the special theory of relativity .
the first of those papers @xcite , _ on the electrodynamic of moving bodies _ , was published on june 30 , and contains the basic theory of special relativity . in that paper
he proved the relativistic formula of addition of velocities .
this formula shows _
`` that the velocity of the light @xmath1 can not be altered by addition with a velocity less than the light '' _ ( see @xcite , p.51 ) , and therefore he solved the problem motivated by the fact _ `` that light is always propagated in empty space with a definite velocity @xmath1 which is independent of the state of motion of the emitting body '' _ ( see @xcite , p.38 ) .
three months later he published : _
`` does the inertia of a body depend upon its energy - content ? ''
_ @xcite . in that paper
he concluded that _
`` the mass of a body is a measure of its energy content''_. this led to the famous formula @xmath0 . in this paper
we reverse in some sense the logic of the expository order of those papers .
our basic physical assumptions are : \(i ) the mass - energy equivalence relation , ( ii ) the energy conservation law , ( iii ) the momentum conservation law , ( iv ) the special relativity principle . from these assumptions
we show how can be deduced the relativistic formula of addition of velocities , and then the _ lorentz transformations_(lt ) .
to simplify our exposition we choose the _ standard _ coordinates @xmath2 and @xmath3 in two arbitrary _ inertial frames_(if ) @xmath4 and @xmath5 , in the _ standard configuration _ , that is , the @xmath5 origin moves with velocity * v * along the @xmath6-axis of @xmath4 , the @xmath7-axis coincides with the @xmath6-axis , while the @xmath8- and @xmath9-axes remain parallel , as do the @xmath10- and @xmath11-axes ; and all clocks are set to zero when the two origins meet ( see @xcite , p. 43 ) .
we have assumed the _ special relativity principle_(srp ) , and it is well known that from the srp follows the homogeneity and isotropy of all the if , but as it is noted in ref .
@xcite , p. 40
, it is perhaps less well known that , conversely , the homogeneity and isotropy of all the if imply the srp .
now @xmath12 and @xmath13 follow by assuming that the transformation between if is linear and the srp holds .
in fact , linearity follows from newton s first law and temporal and spatial homogeneity ( see ref .
@xcite , p. 43 ) .
our starting point will be the equivalence between mass and energy , that is , @xmath0 , where @xmath14 is the energy content of one _ particle _
, @xmath1 is the vacuum velocity of the light , and @xmath15 is the _ inertial mass _ of the particle , which is given by @xmath16 , where @xmath17 in an if of reference in which the particle moves with velocity @xmath18 ( @xmath19 ) , and @xmath20 is the rest mass of the particle
. the relativistic momentum is given by @xmath21 .
consider a massless box , which contains two particles , 1 and 2 , with equal rest mass @xmath20 , moving along a straight line with speeds @xmath22 and @xmath18 respectively , in the if @xmath4 in which the box is at rest , so @xmath4 is the _ zero - momentum frame _ ( @xcite , p.117 ) .
we will calculate now the total energy and the total momentum of the system of the two particles in the frame @xmath5 by two different ways : one by considering the system as a unique _ box - particle _ @xmath23 , whose rest mass @xmath24 is a measure of the energy content of the box , and the other by considering the system as two separated particles . in the first case
the total energy of @xmath23 in the frame @xmath5 is given by @xmath25 , and by use of the einstein s assumption that @xmath24 is a measure of the energy content of the box @xmath23 , @xmath26 , so @xmath27 .
in the same way the momentum of @xmath23 in @xmath5 is given by @xmath28 .
in the second case @xmath29 , and @xmath30 , where @xmath31 and @xmath32 are respectively the velocities of the particles 1 and 2 in the frame @xmath5 . comparing the energy @xmath14 and the momentum @xmath33 calculated by the two ways we obtain : @xmath34 of the _
box - particle _ @xmath23 is the sum of the energy of the two particles 1 and 2 . to find the addition of the parallel velocities @xmath18 and
@xmath35 , we calculate the total energy and the total momentum in @xmath5 by two ways : as one unique particle @xmath23 and as two separated particles 1 and 2 . ] [ criadofig1 ] we will find now the formula of addition of the velocities @xmath18 and @xmath35 for the cases in which they are parallel or perpendicular . in the following it is assumed that coordinate transformations between if include all rigid motions in the 3-space , in particular rotations and space - reflections which are the transformations used in this paper .
consider that the particles 1 and 2 of the massless box move along the @xmath6-axis ( see fig . 1 ) .
this configuration is invariant under rotations about @xmath6-axis in @xmath4 , and then from the srp this configuration must be invariant under rotations about @xmath7-axis in @xmath5 too .
that follows from the fact that the rotations about @xmath6-axis commute with the transformation between the if @xmath4 and @xmath5 , because in the _ standard configuration _ the @xmath7-axis coincides with the @xmath6-axis . from
that it follows that @xmath31 and @xmath32 have both the direction of the @xmath7-axis and so we can denote them by @xmath36 and @xmath37 .
moreover , if we take @xmath38 and use as new variables the rapidities @xmath39 , @xmath40 , @xmath41 , and @xmath42 defined by @xmath43 , @xmath44 , @xmath45 , and @xmath46 , the eqs .
( [ eq:1 ] ) become : @xmath47 @xmath48 dividing member by member eq .
[ eq:3 ] by eq .
[ eq:2 ] , we get @xmath49 , and then @xmath50 .
comparing with eq .
[ eq:2 ] we get @xmath51 , and then @xmath52
. therefore , @xmath53 and @xmath54 .
finally we obtain the following expression for the velocity of particle 1 : @xmath55 and , similarly , for particle 2 : @xmath56 these last equations are precisely the formulas of addition of velocities when @xmath18 and @xmath35 are parallel . and @xmath35 .
] [ criadofig2 ] we shall now study the case of @xmath18 being perpendicular to @xmath35 . in this case
we consider that the particles 1 and 2 of the massless box move along the @xmath8-axis ( see fig . 2 ) .
in this configuration both particles have coordinate @xmath57 for any time @xmath58 in @xmath4 . then @xmath59 ( because @xmath13 ) for both particles and for any time @xmath60 in @xmath5 .
this implies that @xmath31 and @xmath32 must be in the plane @xmath61 .
since the configuration is invariant under the reflection @xmath62 in @xmath4 , and this reflection commute with the transformation between the if @xmath4 and @xmath5 , then the configuration in @xmath5 must be invariant under the reflection @xmath63 .
this implies that @xmath31 and @xmath32 must be in the plane @xmath61 . on the other hand
, the invariance under @xmath63 implies that @xmath64 and the angle between @xmath31 and the @xmath7-axis , @xmath65 , is the opposite to the angle between @xmath32 and the @xmath7-axis .
then @xmath66 , @xmath67 .
now from eqs .
( [ eq:1 ] ) we get : @xmath68 therefore @xmath69 , and from the first equation of ( [ eq:1 ] ) it follows that @xmath70 .
this last formula in velocity parameters is @xmath71 . using the expression of @xmath72 obtained from this formula
we finally obtain : @xmath73 and @xmath74 these equations correspond to the formulas of addition of the perpendicular velocities @xmath18 and @xmath35 .
[ criadofig2 ] we will now get the general formulas for the transformation of the components of the velocities in a change of reference frame . to this end , consider the collision of two particles 1 and 2 , with the same rest mass @xmath75 , and with orthogonal velocities @xmath76 and @xmath77 .
let @xmath78 be the rest mass of the particle 3 resulting from the collision , and @xmath79 its velocity in an if @xmath4 ( see fig . 3 ) .
we also denote by @xmath80 , @xmath81 and @xmath82 the corresponding velocities in the frame @xmath5 that moves with velocity @xmath35 along the direction of @xmath76 . using the conservation of the energy and the momentum we get : @xmath83 @xmath84 from this we get : @xmath85 , and
a similar expression follows for @xmath82 . from the results of sections [
sec : level3 ] and [ sec : level4 ] , it follows that @xmath86 , and @xmath87 .
finally , an easy calculation using the well - known relations @xmath88 and @xmath89 ( see @xcite , p. 69 ) gives the expected formulas : @xmath90 @xmath91 observe that , equivalently , the above calculation can be done considering the movement of the center of mass of the system .
the lorentz transformations can be obtained from the transformation of velocities if we assume the srp . to see this , choose as above the _ standard _ coordinates @xmath2 and @xmath3 in two arbitrary if @xmath4 and @xmath5 , in the _ standard configuration_. consider a particle moving with uniform velocity from the spacial origin for time zero to the point @xmath92 for time @xmath58 in the frame @xmath4 , so its velocity in @xmath4 is given by : @xmath93 .
let @xmath3 the the associated coordinates to @xmath94 in @xmath5 .
as it has been noted in the first paragraph of section [ sec : level2 ] , from the srp , it follows that @xmath12 and @xmath13 .
the velocity in @xmath5 is given by @xmath95 .
if we now use the formulas ( [ eq11 ] ) and ( [ eq12 ] ) of transformation of velocities we get : @xmath96 @xmath97 now , since @xmath12 and @xmath98 , we get : @xmath99 the homogeneity and isotropy of all the if implies that these transformations are valid for the coordinates @xmath2 associated to any _ event _ in @xmath4 .
thus we have found _ the standard lorentz transformation equations_.
we have proved that the law of addition of velocities can be obtained from @xmath0 , the conservation of energy and momentum , and the srp . moreover , the lorentz transformations are obtained immediately from the law of addition of velocities and the srp .
thus , taking @xmath0 and the conservation of energy and momentum , together with the srp as the starting principles of special relativity , is a possible option .
in fact , the relation between energy and inertial mass was pointed out in some particular cases 24 years before einstein s fundamental paper .
moreover , @xmath0 can be considered as an extension of the first principle of the thermodynamic , and as such , a good principle for the foundation of any theory .
however we have to accept that although the derivation of the lt in this paper is logically interesting , it presupposes ad hoc the strongly counterintuitive assumptions of the mass - energy and mass - velocity relations .
therefore , it makes little sense as a means to conceive the lt ( in other words , the premises from which the lt follows are conceptually equally complicated with their consequence , namely the lt ) .
consequently , it is didactically of limited interest .
we also note that if we take the terms in @xmath100 and @xmath101 in the power - series expansion of the total energy , @xmath27 , of the compound particle b , we get the newtonian total energy of b. then , the same reasoning of this paper leads , as it is expected , to the galileo transformations via the galileo addition of velocities .
this work was partially supported by the spanish grants fqm-192 ( c. criado ) and mtm-2004 - 06262 ( n. alamo ) .
we also thank one referee whose suggestions have improved considerably this paper .
9 stachel j 1988 _ einstein s miraculous year _ ( new york : princeton university press ) . einstein a 1905 on the electrodynamics of moving bodies translated and reprinted in _
principle of relativity _ 1952 ( new york : dover ) pp 35 - 65 . einstein a 1905
does the inertia of a body depend on its energy - content ?
translated and reprinted in _ principle of relativity _ 1952 ( new york : dover ) pp 67 - 71 .
rindler w 2001 _ relativity special , general and cosmological _
( new york : oxford university press ) .
thomson j j 1881 _ phil.mag . _ * 11 * 229 .
pais a 1982 _ `` subtle is the lord ... '' the science and the life of albert einstein _
( new york : oxford university press ) pp 155 - 67 .
hasenhrl f 1904 _ adp _ * 15 * 344 . | in this paper we show how to get the lorentz transformations from @xmath0 , the laws of conservation of energy and momentum , and the special relativity principle . to this end
we first deduce the law of addition of relativistic velocities . |
let @xmath2 $ ] be a bivariate polynomial with complex coefficients and let @xmath3 be the ideal generated by @xmath0 .
the variety @xmath4 of @xmath3 is the set of ordered pairs @xmath5 such that @xmath6 .
we give a combinatorial structure to @xmath4 by taking its elements as arcs of a digraph @xmath1 , and we explore the relationship between the polynomial @xmath0 and the digraph @xmath1 . the digraphs @xmath1 were introduced in @xcite for symmetric polynomials
@xmath0 under the name of _ galois graphs_. lengths of walks , distances and cycles were described in terms of @xmath0 .
also , when the coefficients of @xmath0 belong to a field @xmath7 and @xmath8 is the algebraic closure of @xmath7 , the action of the galois group @xmath9 on @xmath1 was studied ( and this was the motivation for the name of galois graphs ) .
here we give a general overview of the topic for non necessarily symmetric polynomials and some concrete families of polynomials are studied . in @xcite
we adopt an algebraic approach to the problem of deciding which polynomials produce a given graph as a connected component of @xmath1 .
let us fix some notation . in this paper digraphs
are allowed to be infinite , and to have multiple arcs and loops .
a _ graph _ is a digraph without loops nor multiple arcs such that for each arc @xmath10 there exists the arc @xmath11 . the arcs @xmath12 form the _ edge _ @xmath13 of the graph .
let @xmath14 be a vertex of a digraph @xmath15 .
the _ strong ( connected ) component _ of @xmath14 is the subdigraph of @xmath15 induced by @xmath14 and the set of vertices @xmath16 such that there exist a directed path from @xmath14 to @xmath16 and a directed path from @xmath16 to @xmath14 .
the _ underlying graph _ of a digraph @xmath15 is the graph obtained by taking as edges the set of @xmath13 with @xmath10 an arc of @xmath15 and @xmath17 .
the _ ( weakly connected ) component _ of a vertex @xmath14 in a digraph @xmath15 is the subdigraph of @xmath15 induced by @xmath14 and all vertices of the strong component of @xmath14 in the underlying graph of @xmath15 .
note that , in a graph , components and strong components coincide . for undefined concepts about graph theory
, we refer to @xcite .
the component of @xmath14 in @xmath1 is denoted by @xmath18 and the strong component by @xmath19 . in a monomial @xmath20 , @xmath21 , the non negative integer @xmath22
is called the _ partial degree _ respect to @xmath23 ( analogously for @xmath24 and @xmath25 ) .
the _ total degree _ or _ degree _ of the monomial is the integer @xmath26 .
the _ partial degree _
respect to @xmath23 of a bivariate polynomial @xmath0 is the maximum of the partial degrees of its monomials ; and the _ total degree _ or _ degree _ of @xmath0 is the maximum of the degrees of its monomials .
if @xmath27 is the expression of @xmath0 as a product of irreducible polynomials over @xmath28 , the _ radical _ of @xmath0 is the polynomial @xmath29 .
a polynomial is _ radical _ if @xmath30 .
we refer to @xcite for any other undefined concept about polynomials .
the paper is organized as follows . in the rest of this section
we give an informal description of its contents . in the next section
we define the digraph @xmath1 .
we shall see that some natural conditions on @xmath0 ( such as @xmath0 to be radical , @xmath31 for all @xmath32 , etc . ) can be assumed .
polynomials satisfying such conditions are called _ standard polynomials_. if @xmath0 is a standard polynomial , then @xmath1 has only a finite number of loops and multiple arcs .
moreover , all vertices have finite indegree and all of them , but a finite number , have the same indegree ; analogously for outdegrees . the components ( resp
. strong components ) of @xmath1 containing these vertices , multiple arcs and loops are called _ singular components _ ( resp . _ singular strong components _ ) . in section 3
we show that every finite strongly connected @xmath33-regular digraph is isomorphic to a strong component of @xmath1 for an appropriate @xmath0 .
nevertheless , the construction produces an infinite component . a digraph is called _ polynomial _ if it is isomorphic to @xmath1 or to a non singular component of @xmath1 . in section 4 we will see that cayley digraphs on the additive and multiplicative groups of @xmath28 are polynomial and we give the corresponding polynomial @xmath0 .
this implies that directed and undirected cycles , finite complete graphs @xmath34 , finite bipartite complete graphs @xmath35 and , in general , circulant digraphs are polynomial .
in section 5 we study polynomials of partial degree one in each indeterminate .
it is shown that all non - singular components are isomorphic , and a characterization of polynomials of partial degree one which give directed @xmath36-cycles as non - singular components is given , as well as those giving infinite directed paths . by relating these polynomials to the group of linear fractional transformations , we prove that cayley digraphs on dihedral groups and on the groups of symmetries of regular polyhedra are polynomial .
in section 6 we consider symmetric polynomials of total degree 2 . in this case
the structure of @xmath1 is also completely determined . in particular , all non - singular components are isomorphic
. finally , the results obtained here and the discussions in @xcite give support to the conjecture stated in the last section .
let @xmath37 be a polynomial with complex coefficients .
the digraph @xmath1 has @xmath28 as set of vertices and an ordered pair @xmath10 is an arc of multiplicity @xmath38 if @xmath16 is a root of multiplicity @xmath38 of the polynomial @xmath39 .
if @xmath39 is the zero polynomial for some @xmath32 the multiplicity of all arcs @xmath10 , @xmath40 , is taken to be 1 , and the vertex @xmath14 is called a _ source universal vertex_. note that the source universal vertices are the roots of the polynomial @xmath41 .
analogously , a vertex @xmath16 is called a _ sink universal vertex _ if @xmath42 is the zero polynomial ; the sink universal vertices are the roots of the @xmath43 . in the following
we assume that @xmath1 has no universal vertices , or , equivalently that the polynomials @xmath44 and @xmath45 are constant .
if @xmath0 is constant then @xmath1 is the complete or the null digraph on @xmath28 depending on whether the constant is zero or not .
therefore , we can also assume from now on that @xmath46 and that @xmath47 and @xmath48 are non zero polynomials .
the structure of @xmath1 is given by the structure of its components .
to study the components of @xmath1 , it is useful to put aside some special cases . consider the discriminant @xmath49 where @xmath50 denotes the partial derivative of @xmath0 with respect to @xmath25 .
we have that @xmath51 if and only if @xmath0 has a multiple factor of positive degree in @xmath25 , say @xmath52 with @xmath53 . in this case , the digraphs @xmath1 and @xmath54 differ only in the multiplicity of the arcs corresponding to the factor @xmath55 .
therefore , we can assume that @xmath56 is not the zero polynomial .
analogously , we can assume that @xmath57 is not the zero polynomial .
note that all vertices have outdegree at most @xmath33 and a vertex @xmath14 has outdegree @xmath58 if and only if @xmath59 .
analogously , all vertices have indegree at most @xmath60 and a vertex @xmath14 has indegree @xmath61 if and only if @xmath62 .
the roots of @xmath47 are called _ out - defective vertices _ and the roots of @xmath48 are called _ in - defective vertices_. let @xmath63 and @xmath64 .
the leading coefficients of @xmath0 and @xmath50 as polynomials in the indeterminate @xmath25 are @xmath47 and @xmath65 respectively , so @xmath47 is a factor of @xmath56 .
thus , the out - defective vertices are roots of @xmath56 .
if @xmath14 is the origin of a multiple arc , then @xmath39 has a multiple root .
therefore , @xmath66 .
conversely , if @xmath66 , then either @xmath59 or @xmath39 has a multiple root , i.e. @xmath14 is an out - defective vertex or it is the origin of a multiple arc .
we conclude that the roots of @xmath56 are the out - defective vertices and the origins of multiple arcs .
analogously , the roots of @xmath67 are the in - defective vertices and the ends of multiple arcs .
the vertices with a loop are the roots of @xmath68 .
there is a loop at each vertex if and only if @xmath69 is the zero polynomial , which means that @xmath0 admits a factorization @xmath70 with @xmath71 and @xmath55 not divisible by @xmath72 .
the structure of @xmath1 is then completely determined by the structure of @xmath73 , and thus it is not a restriction to assume that @xmath69 is not the zero polynomial .
a polynomial @xmath0 is _ standard _ if it is non constant ( @xmath1 is neither the complete or the null graph ) , @xmath74 is constant ( there are not universal vertices ) , @xmath75 ( the polynomial @xmath0 is a radical polynomial ) , and @xmath69 is not the zero polynomial ( @xmath1 has no loops at every vertex ) . for a standard polynomial @xmath0
, the roots of @xmath76 are called _
singular vertices_. they are the vertices with a loop , vertices which are origin or end of multiple arcs , and defective vertices .
a _ singular component _ ( resp . _ singular strong component _ ) is a component ( resp .
strong component ) which contains some singular vertex .
note that only a finite number of singular components ( strong components ) exists .
we denote by @xmath77 the digraph obtained from @xmath1 by removing all its singular components . note that different polynomials @xmath0 can give isomorphic digraphs @xmath1 , as stated in the following lemma of straightforward proof .
[ iso ] let @xmath2 $ ] and @xmath78 with @xmath79 .
if @xmath80 , then the mapping @xmath81 is an isomorphism from @xmath82 to @xmath1 .
our immediate goal is to show that every finite strongly connected @xmath33-regular digraph can be seen as a strong component of @xmath1 for some appropriate @xmath0 .
we need the following lemma : [ factorization ] every @xmath33-regular digraph admits a 1-factorization .
lemma [ factorization ] can be proved by using that every regular graph of even degree admits a 2-factorization @xcite .
see also @xcite for a detailed proof .
[ strong ] let @xmath83 be a finite strongly connected @xmath33-regular digraph of order @xmath84 with @xmath85 . then there exists a polynomial @xmath0 such that @xmath15 is a strong component of @xmath1 .
lemma [ factorization ] ensures that @xmath15 admits a @xmath86-factorization .
let @xmath87 be the set of arcs of the 1-factors . for each @xmath88=\{1,\ldots , d\}$ ]
let @xmath89 be the interpolation polynomial such that @xmath90 for each @xmath91 .
in this way , the vertices adjacent from @xmath92 are @xmath93 .
define @xmath94 . in @xmath1 ,
the vertex @xmath14 is also adjacent to @xmath95 .
therefore @xmath96 . in the proof of theorem [ strong ] ,
the polynomials @xmath89 can have degree @xmath97 , so @xmath0 can have degree @xmath33 in @xmath25 and degree @xmath98 in @xmath23 . if the given digraph @xmath15 has order @xmath99 , then @xmath100 and the component @xmath18 is infinite , while the strong component @xmath101 is finite .
for instance , take @xmath102 , the complete symmetric digraph of order 3 , and choose @xmath103 and @xmath104 as the vertices of @xmath15 .
the digraph @xmath15 is 2-regular and admits the factorization @xmath105 where @xmath106 and @xmath107 .
figure [ figk3 ] shows the factorization of @xmath108 ; the arcs of @xmath109 are the thick ones .
( 0,0 ) # 1#2#3#4#5 @font ( 2136,2455)(1408,-2411 ) ( 2416,-121)(0,0)[lb ] ( 1471,-2138)(0,0)[lb ] ( 3368,-2145)(0,0)[lb ] the polynomial of degree 2 such that @xmath110 , @xmath111 and @xmath112 is @xmath113 , and the polynomial of degree 2 such that @xmath114 , @xmath115 is @xmath116 .
if @xmath117 , then @xmath118 is @xmath102 .
nevertheless , the component @xmath119 is infinite . in the figure [ k3ext ] vertices adjacent to 1 ,
2 and 3 which are in @xmath119 but not in @xmath120 are shown .
( 0,0 ) # 1#2#3#4#5 @font ( 5397,4663)(369,-4079 ) ( 3061,-631)(0,0)[lb ] ( 4486,-2978)(0,0)[lb ] ( 1711,-2978)(0,0)[lb ] ( 4013,419)(0,0)[lb ] ( 1913,419)(0,0)[lb ] ( 5544,-2587)(0,0)[lb ] ( 5094,-4036)(0,0)[lb ] ( 780,-4029)(0,0)[lb ] ( 369,-2580)(0,0)[lb ] by theorem [ strong ] the condition on a finite strongly connected @xmath33-regular digraph of being isomorphic to a strong component of @xmath1 for some polynomial @xmath0 is not restrictive at all .
a @xmath33-regular digraph @xmath15 is said to be _ polynomial _ if , for some standard polynomial @xmath0 , the digraph @xmath15 is isomorphic to @xmath1 or to a non singular component of @xmath1 .
cayley digraphs are relevant structures in different contexts such as modeling interconnection networks @xcite , tessellations of the sphere and of the euclidean plane @xcite and in combinatorial group theory @xcite .
recall that , given a group @xmath121 and a finite set @xmath122 with @xmath123 , the _ cayley digraph _ @xmath124 is defined by taking the elements in @xmath121 as vertices and an ordered pair @xmath10 is an arc if @xmath125 for some @xmath126 .
a cayley digraph @xmath124 is connected if and only if @xmath127 is a generating set of @xmath121 ( this is the reason why the condition of @xmath127 being a generating system is often included in the definition ) .
if @xmath128 for all @xmath126 , then @xmath124 is a graph .
cayley digraphs are known to be vertex transitive .
this implies that all components of a cayley digraph are isomorphic and also that all strong components are isomorphic .
[ cayley+ ] 1 .
let @xmath0 be a standard polynomial . if @xmath129 for some univariate polynomial @xmath130 $ ] , then @xmath1 is a cayley digraph on @xmath131 .
2 . let @xmath132 be a cayley digraph on @xmath131 .
then there exists a univariate polynomial @xmath130 $ ] such that @xmath15 is isomorphic to @xmath1 where @xmath129 is a standard polynomial .
\(i ) let @xmath133 be the roots of @xmath134 .
then @xmath135 with @xmath21 . in @xmath1
a vertex @xmath14 is adjacent to the vertices @xmath136 . because @xmath0 is standard , @xmath137 for all @xmath22 and @xmath138 for @xmath139 . if @xmath140 , then we get @xmath141 .
\(ii ) given @xmath142 with @xmath143 , consider the polynomial @xmath144 and take @xmath129 .
the polynomial @xmath0 is standard and @xmath145 .
note that for a polynomial @xmath129 as in theorem [ cayley+ ] the components are always infinite .
for instance , if @xmath146 then @xmath1 has no singular vertices and it is isomorphic to @xmath147 . in this example ,
components and strong components coincide and all of them are isomorphic to @xmath148 , the grid of integer coordinates .
next theorem is the corresponding to theorem [ cayley+ ] for cayley digraphs on the multiplicative group of @xmath28 . as usual
, @xmath149 denotes @xmath150 .
[ cayley * ] 1 .
let @xmath0 be an homogeneous standard polynomial .
then @xmath77 is a cayley digraph on @xmath151 .
2 . let @xmath152 be a cayley digraph on @xmath151 .
then , there exists an homogeneous standard polynomial @xmath0 such that @xmath153 .
\(i ) let @xmath0 be an homogeneous standard polynomial of total degree @xmath33 .
note that @xmath0 being standard , it must be also of partial degree @xmath33 in both indeterminates .
we have @xmath154 where @xmath134 is a univariate polynomial in @xmath155 of degree @xmath33 .
let @xmath133 be the roots of @xmath134 .
then , @xmath156 for @xmath157 and @xmath158 for some @xmath21 . as @xmath0 is standard , @xmath159 and @xmath137 for all @xmath22 , and @xmath138 for @xmath139 .
each vertex @xmath160 is adjacent to the @xmath33 vertices @xmath161 .
therefore , if @xmath162 , we have @xmath163 .
\(ii ) given a cayley digraph @xmath152 on the multiplicative group on @xmath151 , where @xmath162 , then @xmath164 is a standard polynomial and @xmath163 .
note that if @xmath0 is an homogeneous standard polynomial of total degree @xmath33 , then @xmath1 has 0 as the unique singular vertex , and @xmath165 is a loop of multiplicity @xmath33 . as a corollary of theorems [ cayley+ ] and [ cayley * ] we have : cayley digraphs on the additive and multiplicative groups of @xmath28 are polynomial .
a _ circulant _ digraph is a strongly connected cayley digraph on a finite cyclic group .
it is not a restriction to take the group @xmath166 of the @xmath36-th roots of the unity as the cyclic group of order @xmath36 .
then , a circulant digraph is a cayley digraph of the form @xmath167 , where @xmath127 is a generating set of @xmath166 .
now , @xmath167 is the component of @xmath86 in @xmath168 , so we conclude circulant digraphs are polynomial .
for instance , if @xmath169 is a primitive @xmath36-root of unity and we define @xmath170 , the components of @xmath77 are complete graphs @xmath171
. if @xmath169 is a primitive @xmath172-root of unity and @xmath173 , then the components of @xmath77 are complete bipartite graphs @xmath35 .
the @xmath36-prisms are a family of cayley digraphs over non cyclic groups that are also polynomial .
the @xmath36-_prism _ is the cayley graph @xmath174 for instance , the @xmath104-dimensional cube is the @xmath175-prism .
the @xmath36-prism can be obtained by the polynomial @xmath176 , where @xmath169 is a @xmath36-th primitive root of the unity .
if @xmath177 is a polynomial of partial degree one in @xmath25 , then the vertices @xmath16 such that there exists a directed path from @xmath14 to @xmath16 are the vertices @xmath178 defined by @xmath179 and @xmath180 for all @xmath181 . thus , the structure of @xmath1 is closely related to the dynamical system defined by the rational function @xmath182 .
the iteration of rational functions was studied by g. julia @xcite as early as 1918 and p. fatou @xcite in 1922 , and there is a huge literature on the topic ( see , for instance @xcite ) . in this section
we give a complete description of the digraphs @xmath1 when @xmath0 is a polynomial of partial degree one in both indeterminates .
note that a component of a digraph of indegree and outdegree equal to one is isomorphic either to a directed @xmath36-cycle @xmath183 for some @xmath36 or to an infinite path @xmath184 .
we shall see that in any case , all components of @xmath77 are isomorphic .
this result is applied to exhibit examples of cayley digraphs on non commutative groups that are polynomial digraphs .
first , we characterize the standard polynomials of partial degree 1 . [ standard ]
a polynomial @xmath185 is standard if and only if @xmath186 and it is not divisible by @xmath72 .
assume that @xmath187 is standard .
then the polynomials @xmath188 and @xmath189 are not the zero polynomials .
first , consider the case @xmath190 . then @xmath191 and @xmath192 , so @xmath193 .
second , assume @xmath194 . then @xmath21 . if @xmath195 , then @xmath196 , a contradiction .
thus , @xmath197 and @xmath198 .
finally , let @xmath199 . by dividing @xmath200 by @xmath201 ,
the remainder is @xmath202 . as @xmath203
, this remainder must be not zero .
therefore @xmath186 .
the condition of not being divisible by @xmath72 ensures that @xmath69 is not the zero polynomial .
conversely , assume that @xmath186
. then @xmath204 and @xmath33 can not be simultaneously zero , so @xmath188 is not the zero polynomial .
analogously , @xmath189 is not the zero polynomial . if @xmath205 is non constant , then @xmath206 and @xmath207 , which implies @xmath208 , a contradiction .
analogously , @xmath209 non constant implies @xmath208 . finally , if @xmath210 is the zero polynomial , then @xmath211 and @xmath212 .
therefore @xmath213 is divisible by @xmath72 . for standard polynomials @xmath185 of partial degree one , the structure of @xmath77 is as observed above closely related to the properties of the maps @xmath214 with @xmath186 ( or , equivalently , with @xmath215 ) .
these maps , called _ linear fractional transformations _ @xcite or _ mebius transformations _ @xcite , apply the complex plane minus @xmath216 to the complex plane minus @xmath217 .
they are examples of conformal maps and form a group @xmath218 under composition .
it is useful to represent the group @xmath218 as a quotient of the group @xmath219 of square matrices of order @xmath220 with determinant 1 as follows .
the mapping @xmath221 is an surjective group homomorphism and its kernel is @xmath222
. therefore @xmath223 . given @xmath224 defined by @xmath214 we denote @xmath225 and @xmath226 $ ] the class of @xmath227 in @xmath228 .
[ bilinear ] if @xmath185 is a standard polynomial , then all components of @xmath77 are isomorphic .
let @xmath229 be the roots of @xmath230 .
note that if @xmath190 all vertices have outdegree @xmath86 ; otherwise , @xmath216 is the unique out - defective vertex .
let @xmath231 if @xmath190 and @xmath232 if @xmath21 .
a vertex @xmath14 in @xmath233 is adjacent to the vertex @xmath234 .
consider the function @xmath235 defined on @xmath233 .
assume that there exist components of @xmath77 which are directed cycles and let @xmath84 be the minimum of the lengths of these cycles .
then @xmath36 is the minimum positive integer such that there exists a vertex @xmath14 in a non - singular component such that @xmath236 , or equivalently , @xmath237 thus , the solutions of @xmath238 are the roots of @xmath239 .
but @xmath240 , @xmath241 and @xmath14 are three roots of @xmath242 ( if @xmath243 , then @xmath240 has multiplicity 2 ) , so @xmath242 is the zero polynomial and @xmath244 for all @xmath16 .
this implies that all components of @xmath77 are isomorphic to @xmath245 .
.[abcd ] conditions for the components of @xmath77 being directed @xmath36-cycles . [ cols=">,<",options="header " , ] from a computational point of view , conditions on @xmath246 and @xmath33 for the components of @xmath77 being directed @xmath36-cycles are easily obtained .
the polynomial @xmath242 in the above proof is of the form @xmath247 .
therefore , @xmath248 .
now , for each divisor @xmath7 of @xmath36 , the polynomial @xmath249 must be a factor of @xmath250 . by dividing @xmath250 by all the factors corresponding to digraphs @xmath77 with directed @xmath7-cycles as components ( @xmath7 divisor of @xmath36 ) , we obtain the condition @xmath251 for the components of @xmath77 to be directed cycles of length @xmath36 . in the table
[ abcd ] the conditions @xmath251 for @xmath36 from 2 to 10 are given .
the above proposition can be interpreted in the sense that if a linear fractional transformation @xmath214 generates a cyclic group of order @xmath36 , then all its orbits but the fixed points ( the loops in the graph ) have @xmath36 points ( and therefore the components of @xmath77 , where @xmath185 , are directed @xmath36-cycles ) . the following theorem is a characterization of polynomials of partial degree one which give directed @xmath36-cycles . by a _ primitive _
@xmath36-root of a complex number @xmath252 we mean a complex @xmath253 such that @xmath36 is the minimum positive integer with @xmath254 .
[ ncycles ] let @xmath185 be a standard polynomial and let @xmath253 be a square root of @xmath255
. then the components of @xmath77 are directed @xmath36-cycles if and only if @xmath256 and @xmath257 are both @xmath36-th roots of @xmath86 or both @xmath36-roots of @xmath258 , and at least one of them primitive .
let @xmath224 defined by @xmath214 with @xmath215 .
we have the following equivalences : + the components of @xmath77 are isomorphic to @xmath245 + @xmath259 @xmath260 is the identity mapping + @xmath259 the order of @xmath226 $ ] is @xmath36 + @xmath259 @xmath227 has two distinct eigenvalues which are @xmath36-th roots of @xmath86 or @xmath258 and at least one of them
the characteristic polynomial of @xmath227 is @xmath261 and the discriminant @xmath255 .
therefore @xmath227 has two distinct eigenvalues if and only if @xmath262 . in this case the eigenvalues are @xmath257 and @xmath226 $ ] is of order @xmath36 if they are @xmath36-th roots of @xmath86 or @xmath258 and at least one primitive . for instance ,
consider the polynomial @xmath263 .
we have the values @xmath264 , @xmath265 and @xmath266 .
they satisfy the 6 cycle - condition @xmath267 of table [ abcd ] . on the other hand
, we can use the theorem [ ncycles ] .
the matrix associated to the corresponding @xmath268 is @xmath269 .
the eigenvalues of @xmath227 are @xmath270 which are primitive @xmath271-roots of @xmath258 .
therefore @xmath226 $ ] has order 6 and all components of @xmath77 are directed 6-cycles . taking into account that @xmath218 is a non commutative group , we can find polynomial cayley digraphs on non abelian groups .
the following theorem opens the way .
let @xmath121 be a subgroup of @xmath218 generated by a set @xmath272 of @xmath33 linear fractional transformations .
then @xmath273 is polynomial .
let @xmath274 be a transformation defined by @xmath214 .
we associate to @xmath268 the polynomial @xmath275 .
the group @xmath121 is generated by a finite set , so it is a countable set .
therefore , the set of @xmath32 such that there exists @xmath276 with @xmath14 belonging to a singular component of @xmath277 is also a countable set .
thus , we can choose @xmath32 such that @xmath278 is not a singular component of @xmath277 for all @xmath274 .
let @xmath279 .
the mapping @xmath280 defined by @xmath281 is a digraph isomorphism . indeed : it is injective , because @xmath282 implies that @xmath268 is the identity ( otherwise @xmath14 would be a loop vertex in @xmath277 ) .
it is surjective , because if @xmath16 is a vertex in @xmath19 , then there exists a path @xmath283 .
now , each arc is of the form @xmath284 for some @xmath285 .
if @xmath286 , we have @xmath287 . finally , it preserves adjacencies : @xmath288 is an arc in @xmath289 @xmath259 @xmath290 for some @xmath291 @xmath259 @xmath292 for some @xmath22 @xmath259 @xmath293 for some @xmath22 @xmath259 @xmath294 .
the finite subgroups of @xmath218 are determined ( @xcite , chapter vi ) .
in particular , dihedral groups and the groups of symmetries of regular polyhedra are finite subgroups of @xmath218 . therefore , we have : cayley digraphs on dihedral groups and on the groups of symmetries of regular polyhedra are polynomial . for instance
, the dihedral group @xmath295 is the subgroup of @xmath218 generated by @xmath296 and @xmath297 , where @xmath169 is a primitive @xmath36-root of unity .
then , the cayley digraph @xmath298 is obtained by the polynomial @xmath299 .
in this section we give a method for analyzing the components of @xmath1 for a standard symmetric polynomial @xmath0 of total degree two .
the method implies long but routine calculations , so we skip them and give only results .
from theorem [ bilinear ] we can assume that the standard polynomial is not of partial degree one , so it is of the form @xmath300 . since @xmath77 is a 2-regular graph ,
a component of @xmath77 is isomorphic to a ( undirected ) cycle @xmath301 or to the double ray graph @xmath302 .
denote @xmath303 the vertices of @xmath304 with @xmath305 and @xmath306 adjacent vertices .
the polynomial of degree two @xmath307 has two roots , namely @xmath308 and @xmath309 . as the sum of the two roots is minus de coefficient of @xmath25
, we have @xmath310 or , equivalently , @xmath311 the two roots of @xmath312 are @xmath313 and @xmath314 .
the solutions of the recurrence ( [ recurrence ] ) with initial values @xmath315 and @xmath314 determine the vertices with positive subscripts and analogously for the negative subscripts taking @xmath315 and @xmath313 as initial values . then , to determine the structure of @xmath1 the method is the following .
first , to solve the recurrence ( [ recurrence ] ) , wich give @xmath309 in terms of two initial values @xmath315 and @xmath314 .
second , to find the singular vertices .
there are not defective vertices , so only vertices with loops and multiple arcs should be calculated . in any case , there exist at most two loops and two origin of multiple arcs . for each loop - vertex @xmath316 , the solution of ( [ recurrence ] ) for the initial values @xmath317 gives the vertices in @xmath318 .
analogously , for each multiple arc @xmath319 , the solution of ( [ recurrence ] ) for the initial conditions @xmath320 and @xmath321 gives the vertices in @xmath322 .
the explicit form of the solutions allows to decide if some of these singular components coincide and if they are finite or not . finally , for @xmath315 in a non singular component , the solutions of ( [ recurrence ] ) for the initial values @xmath315 and @xmath314 and @xmath315 and @xmath313 gives the vertices in the non singular component @xmath304 .
the characteristic equation of the second order linear recurrence ( [ recurrence ] ) is @xmath323 , and the discriminant is @xmath324 . to solve the recurrence three cases have to be considered : @xmath325 , @xmath326 , and @xmath327 .
we sumarize the discussion in each case .
first consider the case @xmath325 , see figure [ menosdos ] .
we have @xmath328 . as @xmath0 is standard , @xmath329 and @xmath204
can not be simultaneously zero . for @xmath195
there are no singular components
. for @xmath197 there exist two singular components , both infinite , one containing the vertex - loop @xmath330 and one containing the origin of a double arc @xmath331 . in both cases ( @xmath195 and @xmath197 ) all non singular components are isomorphic to @xmath332 . for @xmath333
, it follows from lemma [ iso ] that if @xmath334 , then @xmath1 is isomorphic to @xmath82
. therefore we can assume ( and we do ) without loss of generality that the polynomial @xmath0 is of the form @xmath335 .
consider now the case @xmath326 , see figure [ dos ] . now
the polynomial is @xmath336 . as @xmath0 is a standard polynomial ,
necessarily @xmath21 .
there exist two non - singular components , both infinite , one containing a loop at the vertex @xmath337 , where @xmath338 is one of the two square roots of @xmath339 , and the other containing a loop at the vertex @xmath340 .
the non singular components are isomorphic to @xmath332 . finally , consider the case @xmath327 , see figures [ noroot ] and [ root ] . if @xmath190 , there exists only one singular component , which contains only the vertex @xmath341 with a double loop .
if @xmath21 , there exist two loop - vertices @xmath342 and @xmath340 and two origin of double arcs , @xmath343 and @xmath344 .
the number and the finitness of the components depends on the values of @xmath345 or , equivalently , on the values of the two distinct roots @xmath346 and @xmath347 of @xmath348 . if @xmath349 are primitive @xmath36-th roots of the unity for some positive integer @xmath36 , then the components of @xmath1 are finite , otherwise they are infinite .
now , @xmath349 is a primitive @xmath36-th root of the unity if and only if @xmath345 is of the form @xmath350 for some @xmath7 with @xmath351 .
the following theorem sumarizes the form of the non singular components .
[ thm degree 2 ] let @xmath300 be a standard polynomial . if @xmath352 and @xmath350 for some positive integers @xmath36 and @xmath7 with @xmath351 , then all components of @xmath77 are isomorphic to @xmath353 , the cycle of length @xmath36 .
otherwise , they are isomorphic to the double ray graph @xmath332 .
if @xmath0 is an homogeneous standard polynomial , then @xmath77 is a cayley digraph ( theorem [ cayley * ] ) .
therefore if a component of @xmath77 is finite , all of them are finite and isomorphic . in section 5
we have seen that if @xmath0 is a polynomial of partial degree one , then if a component of @xmath77 is a directed @xmath36-cycle , then all of them are directed @xmath36-cycles .
also , for a symmetric polynomials @xmath0 of partial and total degree 2 , if a component of @xmath77 is a @xmath36-cycle , then all of them are @xmath36-cycles .
all these examples suggest the following conjecture : [ cd ] let @xmath0 be a standard polynomial of partial degree @xmath33 and @xmath354 a @xmath33-regular digraph isomorphic to a component of @xmath77 .
then , all components of @xmath77 are isomorphic to @xmath354 . in @xcite more evidence of conjecture [ cd ]
for instance , it is shown that if for a symmetric polynomial @xmath0 of partial degree two @xmath77 has a component which is a @xmath36-cycle ( for small values of @xmath36 ) then all of them are @xmath36-cycles . also , if @xmath0 is a polynomial such that @xmath77 has a component isomorphic to @xmath355 ( @xmath356 ) then all of them are isomorphic to @xmath355 .
we would like to thank pelegr viader for his many helpful comments and his insightful perusal of our first draft , and jos luis ruiz , who first studied @xmath1 for an homogeneous polynomial @xmath0 and whose observations lead to theorem [ cayley * ] .
m. c. heydemann .
cayley graphs and interconnection networks . in g.
hahn and g. sabidussi , editors , _ graph symmetry : algebraic methods and applications _ , kluwer academic publishers , montral , _ nato asi series c _ * 497 * ( 1996 ) 167224 .
mmoire sur litration des fonctions rationelles .
( grand prix des sciences mathematiques 1918 ) .
, @xmath357 srie:1 ( 1918 ) 47245 . also in m.
herv , editor , _ oeuvres de gaston julia _ , gauthier - villars , paris , volume i ( 1968 ) 121319 .
m. e. watkins .
ends and automorphisms of infinite graphs . in g.
hahn and g. sabidussi , editors , _ graph symmetry , algebraic methods and applications _ , kluwer academic publishers , _ nato asi _ * 497 * ( 1997 ) 379414 . | let @xmath0 be a bivariate polynomial with complex coefficients .
the zeroes of @xmath0 are given a combinatorial structure by considering them as arcs of a directed graph @xmath1 .
this paper studies some relationship between the polynomial @xmath0 and the structure of @xmath1 . _
key words : _ polynomial digraph , polynomial graph , galois graph , cayley digraph , algebraic variety . |
-0.2 cm noise is not generally detrimental to biological systems but can be employed to generate genotypic , phenotypic , and behavioral diversity @xcite .
real ecosystems are affected by the presence of noise sources which consist of random variability of environmental parameters , such as temperature , food availability , general conditions which can favour or thwart the increase of some biological species .
this randomly fluctuating behaviour can be modeled by gaussian noise sources , which influence , through a multiplicative interaction , the system dynamics .
multiplicative noise often causes the appearance of fluctuating barriers or processes of anomalous diffusion and has been investigated in the context of population growth and extinction @xcite . in this paper
we study the time evolution of three interacting species , two preys , @xmath3 and @xmath4 , and one predator , @xmath5 .
the interaction between the two preys is symmetric and it is given by the parameter @xmath0 .
we study the ecosystem dynamics , described by generalized lotka - volterra equations , in the presence of two different kinds of noise sources : ( i ) a dichotomous noise acting on the @xmath0 parameter , ( ii ) three external sources , modeled as independent multiplicative gaussian noises , which act directly on the three species .
first we consider the deterministic dynamics of the system in a single site and we get the time behaviour of @xmath3 , @xmath4 and @xmath5 , by analyzing the stability of the ecosystem with different constant values of the interaction parameter @xmath0 , which correspond to a coexistence regime ( @xmath6 ) or to an exclusion regime ( @xmath7 ) .
then we consider the interaction parameter @xmath0 varying dichotomously between these two values . in this condition
we study the time behaviour of the species concentrations @xmath3 , @xmath4 and @xmath5 for different levels of the multiplicative noise intensity .
we find noise - induced oscillations and strong anticorrelations between the preys .
afterwards we take into account the spatial version of our ecosystem , considering a two - dimensional domain formed by @xmath1 sites and adding a diffusion term in the l - v equations . by using a mean field approach
, we obtain the corresponding moment equations in gaussian approximation .
we find that , for @xmath0 varying dichotomously , the @xmath8 order moments of the three species concentrations are independent on the multiplicative noise intensity . on the other hand ,
the behavior of the @xmath9 order moments is strongly affected by the presence of external noise sources .
in particular we find that the time behavior is anticorrelated for the species densities of the two preys , and correlated between the predator and the total density of the two preys .
finally we get the time behavior of the @xmath8 and @xmath9 order moments using a coupled map lattice ( cml ) model @xcite and we compare these results with those previously obtained within the mean field approach . in view of an application on real systems
, the results obtained could be useful to explain experimental data , reproducing the behaviour of natural ecosystems @xcite .
-0.2 cm our system is described by a time evolution model of lotka - volterra equations , within the ito scheme , with diffusive terms in a spatial lattice consisting of @xmath1 sites @xmath10 + z_{i , j}\thinspace\sqrt{\sigma_z}\thinspace\xi^z_{i , j } + d(<z>-z_{i , j}),\label{lotka_eq_3}\end{aligned}\ ] ] where the dot indicates the time derivative . the variables
@xmath11 , @xmath12 and @xmath13 are functions of the time @xmath14 , and denote the densities , respectively , of the two preys and the predator in the lattice site @xmath15 .
@xmath16 and @xmath17 are scale factors , @xmath18 is the growth rate for the two preys , @xmath19 is the diffusion coefficient , and @xmath20 , @xmath21 , @xmath22 indicate the spatial mean , performed on the whole lattice , of the three species densities .
the coefficient @xmath0 is the interaction parameter between the two preys .
the coefficients @xmath23 and @xmath24 account for the interaction between preys and predator .
@xmath25 , @xmath26 , @xmath27 are statistically independent gaussian white noises with zero mean and unit variance , and they model the interaction between species and environment .
finally , @xmath28 , @xmath29 , @xmath30 are the intensities of the three sources of gaussian white noise .
-0.2 cm depending on the value of the interaction parameter , coexistence or exclusion regimes take place . in the absence
both of multiplicative noise ( @xmath31 ) and diffusion terms ( @xmath32 ) , eqs .
( [ lotka_eq_1])-([lotka_eq_3 ] ) describe the deterministic dynamics of a single site ecosystem . in these conditions , for the generic site of lattice the stationary values of the three species densities are given by @xmath33 where the indices @xmath34 where suppressed . from eqs .
( [ lotka_eq_1])-([lotka_eq_3 ] ) one can see that the two prey densities have stationary values that are independent on the interaction parameter @xmath0 .
conversely , the stationary value of the predator density is connected with the value of @xmath0 .
this indicates that the interaction parameter between the two preys determines the coexistence or exclusion regimes for the whole system , affecting the stationary value @xmath35 . from eq .
( [ z_stat ] ) the survivance condition for the predator is @xmath36 , which allows to get the coexistence condition for the three species as a function of @xmath0 @xmath37 the inequality ( [ coex_2 ] ) indicates that the system is characterized by two stationary states , which become stable or unstable depending on the values that @xmath0 , @xmath24 and @xmath18 take on . in particular , when the condition ( [ coex_2 ] ) is satisfied , the stable state is represented by the coexistence of the three species . otherwise , after a transient , the predator tends to disappear ( inequality ( [ coex_2 ] ) does nt hold anymore ) and we get a system formed by two competing species , whose coexistence / exclusion conditions depend directly on the value of the parameter @xmath0 @xcite . in this sense ,
the predator plays a regulatory role for the dynamics of the two preys , whose reciprocal behavior is mediated by the interaction parameter @xmath0 through the presence of the species @xmath5 .
we calculate the numerical solutions for single site dynamics , setting in eqs .
( [ lotka_eq_1 ] ) - ( [ lotka_eq_3 ] ) @xmath38 , @xmath39 , @xmath40 , @xmath41 , @xmath42 , with two different values @xmath43 and @xmath44 of the interaction parameter @xmath0 and initial conditions @xmath45 = @xmath46 = @xmath47 , @xmath48 = @xmath49 .
the values of multiplicative noise intensity are the same for the three species , that is @xmath50 . in fig .
[ beta = const ] we show the time series of the three species in coexistence ( @xmath51 ) and exclusion ( @xmath52 ) regimes , for @xmath53 and @xmath54 .
we note that , when the system is subject to deterministic dynamics , the coexistence regime causes , after a transient , the three species to reach the equilibrium values , @xmath55 = @xmath56 = @xmath57 , @xmath35 = @xmath58 , obtained from eqs .
( [ x_stat ] ) - ( [ z_stat ] ) using @xmath40 , @xmath41 , @xmath42 , @xmath59 ( see fig .
[ beta = const]a ) . in deterministic exclusion regime
the predator tends very slowly to vanish .
however , the two prey densities reach the stationary values , remaining constant ( fig . [ beta = const]b ) . in this case
, the stationary values correspond to an unstable equilibrium point .
in fact , in the presence of a small level of multiplicative noise , the symmetry , due to the parameter values and initial conditions used in our simulations , is broken and one of the two preys prevails , displacing the other one ( fig . [ beta = const]d ) , according to the previously obtained results @xcite .
finally we note that no significative modifications occur , with respect to the deterministic case , when a small level of noise is present in coexistence regime ( see fig .
[ beta = const]c ) .
this obviously depends on the fact that , for @xmath51 , the system occupies a stable equilibrium point , which is maintained also in the presence of low levels of multiplicative noise . in ( c ) coexistence and ( d ) exclusion regime .
values of the parameters and initial conditions are @xmath38 , @xmath39 , @xmath40 , @xmath41 , @xmath42 , @xmath45 = @xmath46 = @xmath47 , @xmath48 = @xmath49.,width=491 ] -0.3 cm however , environmental perturbations , due to the presence both of deterministic and random fluctuations of biological and physical variables , such as the temperature , affect the dynamics of the species
. these external forces can modify the behaviour of the populations , either introducing multiplicative noise sources which act directly on the species or affecting the dynamics of the interaction parameter @xmath0 .
in fact , the environmental variations can cause the system dynamics to change between coexistence ( @xmath60 ) and exclusion ( @xmath61 ) regimes .
this dynamical behavior can be described by considering that the interaction parameter @xmath2 is a stochastic process driven by a dichotomous noise , whose jump rate is given by @xmath62{ll}% 0 , & \quad \delta t \leq \tau_d\\ \chi_0 \left(1 + a \thinspace \vert \cos\omega t \vert \right ) , & \quad \delta t > \tau_d\;. \end{array } \right . \label{jump_rate}\end{aligned}\ ] ] where @xmath63 is the time interval between two consecutive switches , and @xmath64 is the delay between two jumps , that is the time interval after a switch , before another jump can occur . with initial value @xmath65 and
delay time @xmath66 .
the interaction parameter @xmath2 switches quasi - periodically between @xmath43 and @xmath44 .
the values of the other parameters are @xmath67 , @xmath68 , @xmath69.,width=226 ] -0.3 cm in eq .
( [ jump_rate ] ) , @xmath70 and @xmath71 are respectively amplitude and angular frequency of the periodic term , and @xmath72 is the jump rate in the absence of periodic term .
this causes @xmath2 to jump between two values , @xmath73 and @xmath74 @xmath75 .
according to the condition ( [ coex_2 ] ) , these values determine the two possible dynamical regimes ( coexistence or exclusion ) of the deterministic lotka - volterra s model for three interacting species . for given values of the parameters @xmath70 , @xmath76 and @xmath72
the switching time between the two levels of @xmath2 depends on @xmath64 . applying a procedure analogous to that followed for the two - species case @xcite , we set @xmath67 , @xmath68 , @xmath69 , obtaining the time series of @xmath2 for @xmath77 , with @xmath78 and @xmath79 . the results , shown in fig .
[ beta ] , indicate the presence of a synchronization between the jumps and the periodicity of the rate @xmath80 . for a system formed by two competing species this causes a quasi - periodical time behavior of the two populations , which can be considered as a signature of the stochastic resonance phenomenon @xcite in population dynamics @xcite .
therefore we fix the delay at the value @xmath81 , which determines an oscillating dynamical regime . in these conditions , @xmath2 switches quasi - periodically between @xmath82 and @xmath83 ( see fig . [ beta ] ) , causing the system to be alternatively subject to the coexistence and exclusion regimes .
-0.2 cm in this section we analyze the time behaviour of three interacting species in a single site of the lattice . from eqs .
( [ lotka_eq_1])-([lotka_eq_3 ] ) , by setting @xmath32 we get @xmath84 where the indices @xmath34 where suppressed . , ( b ) @xmath85 , ( c ) @xmath86 , ( d ) @xmath87 .
here @xmath38 , @xmath39 , @xmath40 , @xmath41 , @xmath42 .
the values of the other parameters are the same of fig .
the initial values of the species densities are @xmath45 = @xmath46 = @xmath47 , @xmath48 = @xmath49 .
the time series of @xmath88 , @xmath89 ( preys ) and @xmath90 ( predator ) show a correlated behaviour in the absence of noise ( panel a ) . in the presence of the noise ( panels ( b)-(d ) ) an anticorrelated behaviour of @xmath88 and @xmath89 appears.,width=491 ] -0.3 cm
-0.3 cm by choosing @xmath91 and @xmath66 , we obtain for @xmath2 the time behaviour shown in fig . [ beta ] .
we analyze the time evolution of the species densities by numerical simulation of eqs .
( [ lotka_eq_1_bis])-([lotka_eq_3_bis ] ) .
the time series of @xmath3 , @xmath4 and @xmath5 are obtained for different values of the multiplicative noise intensity , namely @xmath92 , @xmath93 , @xmath94 , @xmath95 .
the values of the other parameters are the same used in the previous section , that is @xmath38 , @xmath39 , @xmath40 , @xmath41 , @xmath96 , @xmath43 , @xmath44 .
the initial values of the species densities are @xmath45 = @xmath46 = 0.1 , @xmath48 = @xmath49 . in fig .
[ single_compartment ] , where the results are reported , the time series of @xmath88 , @xmath89 ( preys ) and @xmath90 ( predator ) show correlated behaviour in the absence of noise ( panel a ) . in the presence of noise intensity
an anticorrelated oscillating behaviour of @xmath88 and @xmath89 appears ( see panels ( b)-(d ) ) .
moreover we note that , for all the values of multiplicative noise intensity , the two prey densities oscillate , with the frequency of the external driving force , around the stationary values , @xmath97 . we observe that the predator density show an oscillating behaviour , with the same frequency , around a value smaller than @xmath98 .
however , the oscillations of @xmath90 are characterized by a larger amplitude with respect to @xmath99 and @xmath100 .
this behaviour is connected with the different effect that the alternating regime ( exclusion / coexistence ) produces on preys and predator .
in fact , the quasi - periodical behaviour of @xmath2 affects directly the dynamics of the predator ( see eq .
[ z_stat ] ) , causing a decrease of the mean value of @xmath5 during the exclusion regime .
conversely , in coexistence regime the two preys maintain a constant value ( see eq .
( [ x_stat ] ) ) going towards an anticorrelated regime for @xmath101 . in this last condition
the two preys are subject to a pure competitive dynamics , recovering the behaviour observed in a system of two competing species @xcite .
-0.2 cm in this section we analyze the time behaviour of three interacting species in a spatially extended system by using a mean field approach .
the system dynamics is described by eqs .
( [ lotka_eq_1])-([lotka_eq_3 ] ) in the presence of the diffusive term ( @xmath102 ) . in order to use a mean field approach
we derive the moment equations for this system .
assuming @xmath103 , we write eqs . ( [ lotka_eq_1])-([lotka_eq_3 ] ) in a mean field form @xmath104 where @xmath20 , @xmath21 and @xmath22 are average values on the spatial lattice considered ( ensemble averages in the thermodynamic limit ) and we set @xmath105 , @xmath106 , @xmath107 , @xmath108 , @xmath109 $ ] , @xmath110 . by site
averaging eqs .
( [ mean_eq_1])-([mean_eq_3 ] ) , we obtain @xmath111 by expanding the functions @xmath112 , @xmath113 , @xmath114 , @xmath115 , @xmath116 , @xmath117 around the @xmath8 order moments @xmath99 , @xmath100 and @xmath118 , we get an infinite set of simultaneous ordinary differential equations for all the moments @xcite . to truncate this set
we apply a gaussian approximation , for which the cumulants above the @xmath9 order vanish .
therefore we obtain @xmath119\mu_{110}\nonumber\\ & -&\lambda \beta(<x > \thinspace \mu_{020}+<y > \thinspace \mu_{200 } ) -\lambda\alpha(<x>\thinspace\mu_{011}+<y > \thinspace \mu_{101 } ) -2d\mu_{110}\quad\label{mu110_eq}\\ \dot{\mu}_{101 } & = & \lambda(1 - 2\nu\thinspace < x>-\beta\thinspace < y>-\alpha < z>)\mu_{101 } + \lambda_z(-1+\gamma \thinspace < x>+\gamma \thinspace < y>)\mu_{101}\nonumber\\ & -&\lambda < x > ( \alpha \thinspace \mu_{002}+\beta \thinspace \mu_{011 } ) + \lambda_z\gamma < z>(\mu_{110}+\mu_{200})-2d \mu_{101}\quad\label{mu101_eq}\\ \dot{\mu}_{011 } & = & \lambda(1 - 2\nu < y>-\beta < x>-\alpha < z>)\mu_{011 } + \lambda_z(-1+\gamma \thinspace < x>+\gamma \thinspace < y>)\mu_{011}\nonumber\\ & -&\lambda < y > ( \alpha \thinspace \mu_{002}+\beta \thinspace \mu_{101 } ) + \lambda_z\gamma
< z>(\mu_{110}+\mu_{020})-2d \mu_{011}.\quad\label{mu011_eq}\end{aligned}\ ] ] where @xmath120 , @xmath121 , @xmath122 , @xmath123 , @xmath124 , @xmath125 are the @xmath9 order central moments defined on the lattice @xmath126 in order to get the dynamics of the three species we analyze the time evolution of the @xmath8 and @xmath9 order moments according to eqs .
( [ x_mean_eq])-([mu011_eq ] ) . as initial conditions we consider each species uniformly distributed on the spatial domain , that is we set @xmath127 = @xmath128 = @xmath47 , @xmath129 = @xmath49 , @xmath130 = @xmath131 = @xmath132 = @xmath133 = @xmath134 = @xmath135 = @xmath136 .
therefore , from eqs .
( [ x_mean_eq])-([mu011_eq ] ) we get , in the deterministic case , the stationary values for @xmath20 , @xmath21 and @xmath22 @xmath137 using for the parameters the same values of the single site analysis , we obtain @xmath138 and @xmath139 we also fix the delay time at the same value @xmath66 used in the single site case . finally , by numerical integration of eqs .
( [ x_mean_eq])-([mu011_eq ] ) , setting @xmath140 , we get the time series of the @xmath8 and @xmath9 order moments for the following values of multiplicative noise intensity @xmath92 , @xmath93 , @xmath94 , @xmath95 .
the results are reported in figs .
[ moment_series_1 ] , [ moment_series_2 ] . here
we note that , after a transient , the mean values of the two prey densities ( see panels a and d of figs .
[ moment_series_1 ] , [ moment_series_2 ] ) oscillate around the stationary values .
the oscillations are connected with the presence of two stable equilibrium points . for @xmath141 the stable equilibrium is given by the contemporary presence of the three species ( coexistence regime ) .
conversely , for @xmath142 the system goes towards a new equilibrium point , with the predator tending to disappear ( exclusion regime ) . in the presence of a dynamical regime (
the system switches periodically from coexistence to exclusion ) , we observe the appearance of correlated oscillations in the time series of @xmath99 , @xmath100 and @xmath118 . in particular , we note that @xmath118 is subject to oscillations occurring around a value smaller than the stationary one ( @xmath143 ) and characterized by a larger amplitude with respect to @xmath99 and @xmath100 .
this behaviour is analogous to that observed in the case of single site dynamics . and @xmath9 order moments in the mean field approach .
the values of the multiplicative noise intensity are : @xmath53 , @xmath93 from top to bottom . in the absence of noise the time series of @xmath99 , @xmath100 ( panel a ) ,
@xmath120 , @xmath121 , @xmath122 ( panel b ) and @xmath123 , @xmath124 , @xmath125 ( panel c ) are completely overlapped .
the predator ( mean value of species @xmath5 ) shows a behaviour correlated with those of both preys ( mean values of species @xmath3 and @xmath4 ) . for @xmath85 ,
no changes are observed in the behaviour of the mean values ( panel d ) , the variances of the two preys oscillate overlapping each other and a correlation is observed with the variance of the species @xmath5 ( panel e ) , the covariance of the two preys , @xmath123 , oscillates taking on only negative values ( the two preys are anticorrelated each other ) , while @xmath124 and @xmath125 are always zero ( panel f ) .
the initial values of the moments are @xmath127 = @xmath128 = @xmath47 , @xmath129 = @xmath49 , @xmath130=@xmath131=@xmath132 = @xmath133=@xmath144=@xmath135=@xmath136 .
the diffusion coefficient is @xmath140 .
the values of the other parameters are the same used in fig .
[ single_compartment].,width=566 ] -0.3 cm -0.3 cm and @xmath9 order moments
. the values of the multiplicative noise intensity are : @xmath94 , @xmath95 from top to bottom .
no changes are observed in the time behaviour of @xmath20 = @xmath21 = @xmath22 ( see panels a and d ) for both values of the noise intensity .
an increase in the amplitude of oscillations , as a function of the noise intensity , appears both in the variances of the two predators , @xmath120 , @xmath121 , ( see panels b and e ) and in the covariance of the two preys , @xmath123 , ( see panels c and f ) .
the values both of initial conditions and parameters are the same used in fig .
[ moment_series_2].,width=566 ] -0.6 cm -0.3 cm in the absence of noise ( top of fig .
[ moment_series_1 ] ) , the time series of @xmath99 , @xmath100 and @xmath118 ( panel a ) , @xmath145 , @xmath146 , @xmath147 ( panel b ) and @xmath148 , @xmath149 , @xmath150 ( panel c ) are completely overlapped and each species maintains a homogeneous distribution over the lattice , that is all the @xmath9 order moments remain equal to zero . for @xmath85 ( bottom of fig . [ moment_series_1 ] )
no changes are observed in the behaviour of the mean values ( see panel d ) , and the variances of the three species show correlated oscillations ( panel e ) . in panel f , @xmath123 oscillates taking on only negative values .
this indicates that the spatial distributions in the lattice will be characterized by the presence of regions where species @xmath3 or species @xmath4 prevails .
the two preys will be distributed therefore in non - overlapping spatial patterns .
this picture is in agreement with previous results obtained with a different model @xcite .
conversely , @xmath124 and @xmath125 are always zero ( see panel f of fig .
[ moment_series_1 ] ) .
this behaviour indicates that the predator is uncorrelated with the density of each prey : the species @xmath5 tends to occupy indifferently the sites where @xmath3 or @xmath4 prevails ( see the time behaviour of @xmath122 in panel e of fig .
[ moment_series_1 ] ) , but is correlated with the total prey density ( a global increase of food availability improves the life conditions of the predator ) .
this explains why the variance of the predator shows small oscillations . on the other hand , when exclusion regime takes place ,
the two preys tend to occupy different sites , `` spreading out '' in the spatial domain and causing an increase of their variances ( see panel e of fig .
[ moment_series_1 ] ) with a stronger anticorrelation ( see the behaviour of @xmath123 in panel f of fig . [ moment_series_1 ] ) .
finally we note that the amplitude of the oscillations both of the variances and covariances increases as a function of the noise intensity : in particular they have the same order of magnitude of @xmath151 ( see panels b , c , e , f , in figs .
[ moment_series_1 ] , [ moment_series_2 ] ) .
in fact , for higher levels of multiplicative noise ( @xmath152 ) the amplitude of the oscillations increases and the periodical anticorrelated behaviour between the two preys becomes more evident .
conversely , no modifications appear in the time series of the mean values as a function of the multiplicative noise intensity ( see panels a , d in figs . [ moment_series_1 ] , [ moment_series_2 ] ) . even if it is related to a very different mechanism ,
this behavior is similar to the stochastic resonance effect produced in population dynamics , when the interaction parameter is subjected to an oscillating bistable potential in the presence of additive noise @xcite .
we note that in the absence of external noise ( @xmath92 ) both populations coexist and the species densities oscillate in phase around their stationary value @xcite .
this occurs identically in each site of the spatial lattice ( single site dynamics ) .
the behavior of the mean values reproduces this situation . for @xmath153 ,
in the single site dynamics we observe anticorrelated oscillations of @xmath3 and @xmath4 ( preys ) . by site
averaging these noise - induced oscillations ( see ref .
@xcite ) we recover the average behavior obtained in the absence of noise .
this spatial auto - averaging effect explains why the @xmath8 order moment behavior is independent on the external noise intensity , while the @xmath9 order moments give information on `` spreading '' and anticorrelation of the species densities in the spatial domain .
-0.2 cm in this section we adopt a different approach to analyze the dynamics of the three species on the square lattice defined in section [ sec:1 ] .
we consider the time evolution of our system by using a coupled map lattice ( cml ) model @xcite . in this formalism both correlated and anticorrelated spatial patterns of the three interacting species have been found @xcite . here
we calculate the moments by using the cml model . by this approach
, the dynamics of the spatial distributions of the three species is given by the following equations @xmath154 where @xmath155 , @xmath156 and @xmath157 denote respectively the densities of prey @xmath3 , prey @xmath4 and predator @xmath5 in the site @xmath15 at the time step @xmath158 .
according to the notation used for the mean field approach , @xmath16 , @xmath17 , @xmath18 , @xmath0 , @xmath23 , @xmath24 and @xmath19 represent the same quantities defined in section [ sec:1 ] .
@xmath159 , @xmath160 , @xmath161 are independent gaussian white noise sources with zero mean and unit variance .
the interaction parameter @xmath162 corresponds to the value of @xmath2 taken at the time step @xmath158 , according to eq .
( [ jump_rate ] ) . here
@xmath163 indicates the sum over the four nearest neighbours .
-0.2 cm applying a procedure analogous to that used for eqs .
( [ lotka_eq_1])-([lotka_eq_3 ] ) , we consider eqs .
( [ clm - lotka_1])-([clm - lotka_3 ] ) in the absence both of noise sources and diffusion terms ( @xmath32 ) . in this conditions , for @xmath164 , @xmath165 , @xmath166 , we obtain the stationary values of the three species densities for the generic site @xmath167\label{cml_x_stat}\\ z_{_{cml}}^{stat}&=&\frac{2\gamma\left[\frac{\lambda-1}{\lambda}\right]- ( \beta+\nu)\left[\frac{\lambda_z+1}{\lambda_z}\right]}{2\alpha\gamma},\label{cml_z_stat}\end{aligned}\ ] ] where the indices @xmath34 were suppressed . as in the approach based on the use of differential equations , the stationary values of the two prey densities are independent on the interaction parameter @xmath0 , which is responsible for the two different dynamical regimes , coexistence or exclusion , and affects the dynamics of the whole system through its action on the stationary value @xmath168 . the existence condition for the predator @xmath169- ( \beta+\nu)\left[\frac{\lambda_z+1}{\lambda_z}\right]}{2\alpha\gamma}>0 \label{coex_cml_1}\ ] ] allows to get the following inequality for the interaction parameter @xmath0 @xmath170}{\left[\frac{\lambda_z+1}{\lambda_z}\right]}-\nu.\label{coex_cml_2}\ ] ] the inequality ( [ coex_cml_2 ] ) indicates that , according to the analysis performed in section [ single site dynamics ] , the cml model is characterized by two stationary states that become stable or unstable depending on the values of the parameters . comparing the inequalities ( [ coex_2 ] ) and ( [ coex_cml_2 ] )
, we note that in the cml model the coexistence condition and the regulatory role , played by the predator on the dynamics of the two preys , depend also on the scale factors @xmath16 and @xmath17 . in view of a comparison between mean field approach and cml model ,
we define the @xmath8 and @xmath9 order moments on the discrete lattice , at the time step @xmath158 .
the mean values , @xmath171 , @xmath172 , @xmath173 , given by @xmath174 represent the @xmath8 order moments .
the variances @xmath175 , @xmath176 , @xmath177 defined as @xmath178 and the covariances @xmath179 are the @xmath9 order central moments . in order to get , for the species densities , stationary values close to those obtained in the mean field approach ( see eqs .
( [ z_stat_bis ] ) ) , we choose for all the parameters , except @xmath24 , the same values of section [ sec:1 ] .
therefore , setting @xmath38 , @xmath39 , @xmath40 , @xmath41 , @xmath180 , from eqs .
( [ cml_x_stat ] ) , ( [ cml_z_stat ] ) , we calculate the stationary values for the densities of the two preys and predator in the coexistence regime ( @xmath59 ) @xmath181 the cml model can be considered as a time discrete version of the lotka - volterra system , with time step @xmath182 . for the numerical integration of eqs .
( [ x_mean_eq])-([mu011_eq ] ) we used @xmath183 , which is a suitable value to obtain convergence of the solution . obviously , with these values of @xmath63 and @xmath184 , the dynamics of the cml model results to be faster with respect to that obtained within the moment formalism .
in particular , for @xmath142 , using the same parameter values of the mean field approach , the exclusion regime causes the species @xmath5 to vanish in one time step ( @xmath182 ) .
this implies that , when the system is subject to the dynamical regime discussed in section [ single site dynamics ] , the predator disappears .
this behaviour disagrees with the results found by using the moment equations ( see section [ spatially_extended ] ) . in order to remove this discrepancy between cml model and mean field approach , in the discrete time equations we use a much smaller value for the diffusion constant , namely @xmath185 . by this way
, we obtain a slowdown of the diffusion dynamics and , as a consequence , the survivance of the predator in the coexistence / exclusion dynamical regime . in order to get the time behaviour of the @xmath8 and @xmath9 order moments within the scheme of the cml model , we consider a square lattice with @xmath186 , using for @xmath2 the time behaviour given in fig .
afterwards , at each time step @xmath158 we calculate , from eqs .
( [ clm - lotka_1 ] ) , ( [ clm - lotka_2 ] ) , ( [ clm - lotka_3 ] ) , the new values of @xmath187 , @xmath188 , @xmath189 , and the moments according to eqs . ( [ cmlmeans ] ) , ( [ cmlvariances ] ) , ( [ cmlcovariances ] ) . by iterating this procedure
, we obtain the time series shown in figs .
[ cml_series_1 ] , [ cml_series_2 ] . , @xmath172 , @xmath173 , the variances , @xmath190 , @xmath191 , @xmath192 , and the covariances , @xmath193 , @xmath193 , @xmath193 for @xmath53 .
the same quantities are shown in panels ( d ) , ( e ) and ( f ) for @xmath85 . the time series are obtained within the formalism of the cml model ( see eqs .
( [ clm - lotka_1 ] ) , ( [ clm - lotka_2 ] ) , ( [ clm - lotka_3 ] ) ) .
the diffusion coefficient is @xmath185 , and @xmath180 .
the initial values of the species concentrations are @xmath194 , @xmath195 for all the sites @xmath15 .
the values of the other parameters are the same of fig .
[ moment_series_1 ] : @xmath38 , @xmath39 , @xmath40 , @xmath41.,width=566 ] -0.3 cm -0.4 cm , @xmath172 , @xmath173 , the variances , @xmath190 , @xmath191 , @xmath192 , and the covariances , @xmath193 , @xmath193 , @xmath193 for @xmath86 .
the same quantities are shown in panels ( d ) , ( e ) and ( f ) for @xmath87 .
the time series are obtained within the formalism of the cml model ( see eqs .
( [ clm - lotka_1 ] ) , ( [ clm - lotka_2 ] ) , ( [ clm - lotka_3 ] ) ) .
the values of the other parameters and the initial conditions are the same of fig .
[ cml_series_1].,width=566 ] -0.3 cm -0.4 cm the @xmath8 and @xmath9 order moments calculated within the formalism of the cml model can be compared with the same quantities obtained in the mean field approach ( see figs .
[ moment_series_1 ] , [ moment_series_2 ] ) .
we note that the two set of time series are in a good qualitative agreement .
according to the results obtained in the formalism of the moment equations , the mean values of the three species show time oscillations , whose amplitude is larger for the predator ( panels a , d of figs .
[ cml_series_1 ] , [ cml_series_2 ] ) . in the absence of noise ,
the @xmath9 order moments remain equal to zero ( see panel b of fig . [ cml_series_1 ] ) , recovering the conditions of homogeneous distributions obtained for @xmath53 in the mean field approach ( see panel b of fig . [ moment_series_1 ] ) . in the presence of multiplicative noise ,
no modifications occur in the time series of the @xmath8 order moments ( see left side panels in figs .
[ cml_series_1 ] , [ cml_series_2 ] ) .
however , for @xmath196 a symmetry breaking is introduced , with non - vanishing variances that are responsible for inhomogeneous distributions of the three species . for higher levels of the noise intensity , the amplitude of the oscillations remains constant in the time series of @xmath197 , @xmath198 and @xmath199 ( see panels b , e of figs .
[ cml_series_1 ] , [ cml_series_2 ] ) .
these results show some difference with those obtained in the formalism of moment equations , where higher noise intensities cause the oscillation amplitudes of @xmath120 , @xmath121 and @xmath122 to become larger ( see panels b , e of figs .
[ moment_series_1 ] , [ moment_series_2 ] ) .
finally , we find that for @xmath196 , temporal oscillations also appear in the time series of @xmath123 .
this agrees with the results of the mean field approach , revealing the presence of an anticorrelated dynamics between the two preys . on the other hand , @xmath124 and
@xmath125 remain equal to zero also in the presence of multiplicative noise .
this behaviour , in agreement with that obtained in the mean field formalism , indicates that the spatial distribution of the predator is uncorrelated with those of each prey considered separately , but depends on the total density of preys .
the comparison between the two approaches shows that the mean values @xmath171 , @xmath172 , @xmath173 and those obtained within the formalism of the moment equations oscillate around different values .
moreover , the amplitudes of the oscillations in the @xmath9 order moments appear significantly larger in the cml model .
this discrepancies can be explained recalling that : i ) in the two approaches the stationary values are different ( see eqs .
( [ z_stat_bis ] ) and eqs .
( [ cml_x_stat])-([cml_z_stat ] ) ) ; ii ) in the mean field formalism the interaction between sites is extended to the whole spatial domain , conversely in the cml model the species interaction is restricted to the nearest neighbors ; iii ) the dynamics of the cml model is faster since an unitary time step ( @xmath182 ) is taken , instead of the time step @xmath183 used in the moment equations .
-0.2 cm we report a study on the stochastic dynamics of an ecosystem with three interacting species ( two preys and one predator ) , described by generalized lotka - volterra equations .
after considering the single site dynamics of the ecosystem , we consider a spatially extended domain ( two - dimensional lattice ) by introducing `` long range '' diffusive terms ( diffusion occurs among each site and all the other ones ) .
the study is performed by a mean field approach , in the formalism of the moment equations .
the system is affected by the presence of two noise sources , namely a multiplicative white noise and a correlated dichotomous noise .
the role of the correlated dichotomous noise is to control the dynamical regime of the ecosystem ( see fig . [ beta ] ) , while the multiplicative noise is responsible for the anticorrelated behavior of the species concentrations .
the mean field approach in gaussian approximation enables us to obtain the time series of the @xmath8 and @xmath9 order moments .
we compare the results obtained within the mean field approach with the time series calculated by a coupled map lattice ( cml ) model .
the agreement is quite good , even if some discrepancies are present , due to the discrete nature of the cml model and the limited extension of the diffusive interaction ( nearest neighbors ) among different sites of the coupled map lattice .
our theoretical results could explain the time evolution of populations in real ecosystems whose dynamics is strictly dependent on random fluctuations , always present in natural environment @xcite .
-0.2 cm authors are thankful to prof .
lutz schimansky - geier that inspired this work by fundamental ideas and useful discussions .
authors acknowledge the financial support by esf ( european science foundation ) stochdyn network and partially by miur .
s. ciuchi , f. de pasquale and b. spagnolo , phys .
e * 53 * , 706 ( 1996 ) ; m. scheffer _
et al . _ ,
nature * 413 * , 591 ( 2001 ) ; a. f. rozenfeld et al .
a * 280 * , 45 ( 2001 ) ; b. spagnolo , m. cirone , a. la barbera and f. de pasquale , journal of physics : cond . mat . * 14 * , 2247 ( 2002 ) . j. m. g. vilar and r. v. sol , phys . rev
. lett . * 80 * , 4099 ( 1998 ) ; b. spagnolo and a. la barbera , physica a * 315 * , 114 - 124 ( 2002 ) ; a. la barbera and b. spagnolo , physica a * 315 * , 201 ( 2002 ) ; b. spagnolo , a. fiasconaro , d. valenti , fluc . noise lett . * 3 * , l177 ( 2003 ) .
r. benzi , a. sutera , a. vulpiani , j. phys . : math gen .
* 14 * , l453 ( 1981 ) ; l. gammaitoni , p. hanggi , p. jung , and f. marchesoni , rev .
phys . * 70 * , 223 ( 1998 ) ; v. s. anishchenko , a. b. neiman , f. moss , and l. schimansky - geier , phys .
usp . * 42 * , 7 ( 1999 ) ; t. wellens , v. shatokhin , and a. buchleitner , rep .
prog . phys .
* 67 * , 45 ( 2004 ) . | the spatio - temporal dynamics of three interacting species , two preys and one predator , in the presence of two different kinds of noise sources is studied . to describe the spatial distributions of the species we use a model based on lotka - volterra equations . a correlated dichotomous noise acts on @xmath0 , the interaction parameter between the two preys , and a multiplicative white noise affects directly the dynamics of each one of the three species .
we study the time behaviour of the three species in single site for different values of the multiplicative noise intensity , finding noise - induced oscillations of the three species densities with an anticorrelated behaviour of the two preys .
afterwards , by considering a spatially extended system formed by a two - dimensional lattice with @xmath1 sites and applying a mean field approach , we get the corresponding moment equations in gaussian approximation . within this formalism
we obtain the time behaviour of the first and second order moments for different values of multiplicative noise intensity , with @xmath2 subject to the same dichotomous noise source .
finally , we compare our results with those obtained by using a coupled map lattice model , consisting of a time discrete version of the lotka - volterra equations . |
there is growing evidence that the intracluster medium ( icm ) of galaxy clusters is permeated with magnetic fields , as indicated by the detection of large - scale , diffused radio emission called radio halos and relics ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the radio halos are sometimes extended over @xmath8 mpc , covering the whole cluster . by assuming that the magnetic energy is comparable to the total energy in relativistic electrons , one often deduces that the magnetic fields in the cluster halos can reach @xmath9 @xmath10 g and the total magnetic energy can be as high as @xmath7 ergs @xcite . the faraday rotation measurements ( frm ) are extensively used to study the magnetic fields in galaxy clusters .
by combining frm and density measurements , magnetic field strengths in clusters have been measured as high as a few to ten @xmath10 g level , mostly in the cluster core region@xcite .
more interestingly , frm was used to suggest that magnetic fields can have a kolmogorov - like turbulent spectrum in the cores of clusters @xcite , with an energy spectrum peak at several kpc .
other studies @xcite have suggested that the coherence scales of magnetic fields can range from a few kpc to a few hundred kpc , implying large amounts of magnetic energy and fluxes in the icm .
recently , radial profiles of magnetic fields extended to the outer part of clusters have been estimated with frm from observations of additional radio galaxies behind these clusters @xcite . to accurately constrain the magnetic fields in clusters , knowledge of the three dimensional distribution of the magnetic fields
is required .
this will be especially true when the extended very large array ( evla ) becomes operational , as it will be able to obtain frm over large areas of galaxy clusters .
it is also suggested that the evolution and distribution of magnetic fields in clusters may be important to the cluster formation because they play a crucial role in processes such as heat transport , which consequently affect the applicability of clusters as sensitive probes for cosmological parameters @xcite .
in addition , the distribution of magnetic fields can potentially serve as a tracer of the merger history of a cluster .
magnetic field evolution is complicated and difficult to study analytically in highly non - linear systems , such as galaxy clusters , and therefore cosmological mhd simulations are used to study the properties of magnetic fields in the icm .
although the existence of cluster - wide magnetic fields is clear , their origin is still poorly understood .
consequently , initial magnetic fields are usually added to the simulations by hand .
simulations have been performed with initial magnetic fields either from some random or uniform fields at high redshifts @xcite or from the outflows of normal galaxies @xcite .
these simulations found that fields can be further amplified by cluster mergers @xcite and turbulence .
though it is not clear whether the initial conditions of their simulations are correct , their findings roughly match results from observations .
this suggests that cosmological mhd simulations are a good way to study cluster magnetic fields . on the other hand , very small seed fields from some first principle mechanism , like biermann battery effect
have also been studied @xcite .
it has been suggested that very week seed fields can be amplified by dynamo processes in clusters @xcite , though current simulations have not been able to model such processes self - consistently .
in addition , the exact mechanism for dynamo is still being debated ( e.g. * ? ? ? * ) .
observationally , large scale radio jets and lobes from agns serve as one very intriguing source of cluster magnetic fields , because they could carry large amounts of magnetic energy and flux @xcite .
the magnetization of the icm and the wider inter - galactic medium ( igm ) by agns has been suggested on the energetic grounds @xcite , without details of the exact physical processes of how magnetic fields might be transported and amplified .
recently , using self - consistent cosmological mhd simulations , @xcite showed that magnetic fields from an agn injected in a local region can be sufficient to magnetize the whole cluster to the micro gauss level through the action of cluster mergers and the icm turbulence .
this result , which is consistent with semi - analytic models of the evolution of magnetic fields and turbulence @xcite , shows that the small - scale turbulent dynamo @xcite may operate in the icm . in this paper
, we investigate the influence of injection energy and redshift on the evolution of cluster magnetic fields and their properties at the present epoch . in a following paper
, we will study the magnetic field evolution of galaxy clusters with different masses and at different dynamical stages . here , we present the properties of magnetic fields in the icm and the distribution of synthetic faraday rotation measurement .
we also discuss the importance of high numerical spatial resolution in a large volume to follow the magnetic fields correctly in cluster formation simulations .
the setup of our simulations is described in section [ sec : model ] and results are presented in section [ sec : result ] .
our results are summarized in section 4 .
cosmological mhd simulations were performed using the newly developed enzo+mhd code @xcite , which is an eulerian cosmological mhd code with adaptive mesh refinement ( amr ) .
the initial conditions of our simulations are generated from an @xcite power spectrum .
the simulations use a @xmath11cdm model with parameters @xmath12 , @xmath13 , @xmath14 , @xmath15 , and @xmath16 .
these parameters are different from those used in previous simulations @xcite and are close to the values from the recent wmap observations @xcite . the most important difference between simulations presented here and the previous one is the baryon to dark matter ratio , which is 14.76@xmath17 ( compared to 8.20@xmath17 in @xcite ) .
it is encouraging that the evolution of the magnetic field in this work is similar to the evolution found in @xcite , given the differences in cluster properties and @xmath18 .
the simulation volume is @xmath19 mpc on a side , and it uses a @xmath20 root grid and @xmath21 level nested static grids in the lagrangian region where the cluster forms .
this gives an effective root grid resolution of @xmath22 cells ( @xmath0 0.69 mpc ) and dark matter particles of mass resolution of @xmath23 .
as the simulation proceeds , @xmath24 levels of refinement are allowed beyond the root grid , for a maximum spatial resolution of @xmath25 kpc .
the energy of an agn is initially injected into the icm in the form of magnetic fields ( see description in @xcite ) over the most massive halo locally .
part of the magnetic energy , however , is quickly converted to other energies ( kinetic motion and thermal energy ) during the formation of radio jets and lobes in the icm @xcite . because it is currently not possible to resolve both the galaxy cluster and the agn environment simultaneously ,
we have adopted an approach that mimics the possible magnetic energy injection by an agn @xcite .
the size of the injection region and the associated field strength are not realistic when compared to the real agn jets , but on global scales ( @xmath26 tens of kpc ) , the previous studies by @xcite and @xcite showed that this approach can reproduce the observed x - ray bubbles and shock fronts ( e.g. * ? ? ?
* ; * ? ? ?
the simulations were evolved from redshift @xmath27 to @xmath28 without radiative cooling , star formation , or feedback ( so - called adiabatic ) . in this study , we `` turn on '' the agn with magnetic fields at different redshifts of @xmath29 , @xmath30 , @xmath31 and @xmath32 , centered at the most massive halo in the proto - cluster at the injection time .
we have two sets of simulations .
first , we have six simulations of the agn activated at z=3 with different magnetic energies .
the reason why our study concentrates on z=3 is that the comoving quasar number density peaks at z @xmath0 2.5 ( e.g. * ? ? ?
we label simulations of this set as runs a to f , from the most to the least injected energy .
the second set of simulations inject magnetic energy similar to run c , but with the injection redshift varied .
simulations with injection at z=2 , z=1 , z=0.5 are labeled as z2 , z1 and z05 .
all the parameters of the injection model @xcite except the strength of the magnetic fields are kept the same for all the simulations .
the magnetic energy is injected inside 40 kpc ( proper not comoving ) and the injection lasts for 36.7 myr .
the properties of the halo at the agn injection and the initial and final magnetic energies are summarized in table [ table : halos ] .
[ table : halos ] .halo properties and magnetic energy of simulations . [ cols="^,^,^,^,^,^,^,^ " , ] amr is applied only in a region of ( 50 mpc)@xmath33 where the galaxy cluster forms . during the course of cluster formation ,
before the magnetic fields are injected , the refinement is controlled by baryon and dark matter overdensity , which is similar to other cluster formation simulations ( e.g. * ? ? ?
* ; * ? ? ?
after the magnetic field injection , in addition to refinement by overdensity , all the regions where the magnetic field strength is higher than @xmath34 g are refined to the highest level . using this refinement method ,
the volume that is refined to the highest level contains more than @xmath35 of the total magnetic energy @xcite . for run
a , the simulation is equivalent to @xmath0 600@xmath33 uniform grid ( @xmath0 600@xmath33 zones at the highest level ) mhd runs in the cluster region with full cosmology . for cases with lower injected magnetic energy , the highly refined region will be smaller ( e.g. the volume with the highest refinement of run f is about half of that of run a ) .
the data analysis in this paper is performed using yt for enzo @xcite .
before we discuss the properties of magnetic fields in the icm , we present the history of the galaxy cluster formation by hierarchical mergers .
the merger history determines the properties of the icm dynamics and , in turn , the evolution of magnetic fields . in fig .
[ fig : density_casea ] , we show the projected gas density at different redshifts corresponding to different stages of cluster formation for run a. the halos with agns injected at different redshifts are indicated by an arrow .
the arrow in the z=3 panel applies to runs a to f , while the arrows in the z=2 , 1 , 0.5 panels apply to runs z2 , z1 , z05 , respectively .
as we will show later , the magnetic field injection does not change the halo merger properties , so we can use the same projection to indicate the agn host halo for all simulations .
each figure is a projection of baryon density along the y direction in a box of @xmath36 comoving centered at the center of the simulation domain .
[ fig : virialmass ] plot the evolution of mass and averaged kinetic energy density inside the virial radius ( @xmath37 ) of the cluster .
this shows the mass growth of the cluster and the kinetic energy available to amplify the magnetic fields .
when we compute the kinetic energy , we substract the averaged velocity of the whole cluster ( the bulk motion of the cluster that can not be used to amplify the magnetic fields ) .
galaxy clusters are formed by hierarchical mergers as described by @xcite . during the period we studied from z=3 to z=0 ,
there are two different stages . from z=3 to z=1 , the cluster is still small and increases in mass and size rapidly by frequent mergers with halos of similar sizes . from z=3
to 1 , the mass of the cluster increases from 1.5 @xmath1 10@xmath38 to 5 @xmath1 10@xmath39 @xmath40 . since the mass from major mergers
is added to the cluster slowly ( over a few 100 myrs ) , the virial mass inceases quite smoothly . due to frequent mergers ,
the cluster is active and the kinetic energy density is relatively high .
the peaks of the evolution of kinetic energy density show the impact of major mergers .
by z=1 , the major part of cluster has been formed .
mergers have become rare and the merging halos are relatively small compared to the central cluster , producing smaller disturbance on the icm .
thus the mass grows at a much slower rate ( a factor of 2.5 for the remaining 7 gyr ) and the cluster becomes relaxed , with a lower kinetic energy density that decreases gradually to only about one quarter its value before z=1 .
the cluster has a final virial mass of 1.25 @xmath1 10@xmath5 @xmath40 with a virial radius of 2.16 mpc . to show the influence of the magnetic fields on the cluster formation , fig .
[ fig : densityz0 ] plots the projected gas density at redshift z=0 for runs a , c , f , z2 , z1 , z05 , in smaller projection boxes of @xmath41 , while fig .
[ fig : profile_z0 ] shows radial profiles of their baryon density and temperature .
though the injected magnetic fields and the final field distributions are quite different for various simulations , the gas density and temperature distributions are just moderately affected , mostly in the cluster central region .. the gas density increases and the temperature decreases with both injection energy and injection time , but only at most 10% .
the magnetic energy in the clusters is much smaller than the kinetic and thermal energy ( see @xcite and fig . [
fig : beta_volume_histogram ] in sec .
[ sec : profile ] ) throughout most of the cluster .
so they are dynamically unimportant during the cluster formation .
the impact of magnetic fields is more visible in the inner 400 kpc but these do nt change the properties of the cluster as a whole significantly . to illustrate the spatial evolution of the magnetic fields , we present the projections of magnetic energy density ( @xmath42 ) at different redshifts in fig .
[ fig : med_z3 ] for cases with agn `` turned on '' at z=3 and in fig .
[ fig : med_z2 ] for all other cases .
we set the center of the view at the cluster center at the observed time , since the magnetic fields move with the cluster as shown in fig .
[ fig : density_casea ] . with the strong local injections ,
the magnetic fields first expand by their own pressure to form bubbles .
the mergers and random motions of the icm then destroy the bubbles and spread the magnetic fields away from the vicinity of the injection locations to the whole cluster .
these processes were described in @xcite . for cases with an agn injected at z=3 , the magnetic fields basically evolve in a similar way and are all distributed throughout a large volume of the cluster at low redshifts .
the magnetic fields are spread throughout the cluster very quickly before z=0.5 when the cluster is more active .
the strength of the initial magnetic fields determines the early expansion of the fields , thus the final field distribution .
magnetic fields in run a , with the strongest injected fields , expand the fastest at early times and are distributed in the largest volume at any stage during the cluster formation .
after @xmath32 , the magnetic field distribution in the icm shows signs of saturation . note that the cluster is continuously growing in mass and undergoing minor mergers ( albeit slowly ) . here
we regard that magnetic fields have reached a `` saturated '' state when they have stopped growing ( or have very slow growth ) both in their total energy ( see fig . 7 in the next subsection ) and spatial distribution .
the detailed physical processes of saturation , however , deserve a more careful evaluation , which will be presented in a future publication . though expanding with different rates at early time , magnetic field distributions of runs b , c , and d are quite similar at z=0 .
it seems that they are all saturated at z=0 , and this saturation level is determined more by the icm properties than the initial magnetic field strength . for runs
e and f , the magnetized regions are significantly smaller than the stronger injection energy cases , and the strength of projected magnetic fields is weaker .
their magnetic fields have more difficulty expanding in the icm at the early time due to their weak initial magnetic pressure .
this leads to a much smaller cross section for the mergers , and consequently less chance of being brought to the other parts of the cluster and to be amplified . they continue to expand to a much larger volume from z=0.5 to z=0 , and
do nt clearly show indication of being saturated . for agn injected at z=2 , which injects magnetic fields about 1 gyr later ,
the magnetic fields are also spread throughout the whole cluster . though the initial magnetic fields are injected at a different places , and with larger energy , the field distribution at z=0 is very similar to runs c and e. for the run with injection at z=1 , magnetic fields still spread out of the cluster center , but are distributed in a much smaller area .
this is because the injected magnetic fields miss the most active period of the cluster mergers .
for run z05 , though the magnetic bubbles are still destroyed , the magnetic fields are confined in the central region of the cluster .
the cluster at z=0.5 is already quite large , so it is difficult for the mergering halos to bring the magnetic fields far away from the center . for the case
when the agn injection is much later at z=0.05 @xcite , the magnetic fields could survive in a bubble morphology , which resembles observations of x - ray cavities in clusters .
since we have only one source of magnetic fields in each simulation , it is straight forward to track the history of magnetic energy amplification in the cluster .
the evolution of the total magnetic energy of all the simulations is shown in fig .
[ fig : energy_z3 ] .
the magnetic energy generally decreases rapidly initially for a few hundred million years because of the rapid expansion of the magnetic structure . for agn injected before the main mergers ( runs a - f , z2 and z1 )
, the magnetic energy gradually increases due to the combined effects of major mergers and the icm turbulence until saturation occurs or the simulation ends . at z = 0 ,
the total magnetic energy is ( 14 , 9.6 , 6.5 , 7.3 , 5.1 , 2.9 ) @xmath1 10@xmath3 ergs , for runs a - f with a gain of about 10 , 15 , 40 , 150 , 400 , and 900 times , respectively .
it is interesting that the magnetic energy of run f , which has an injected energy of just 3 @xmath1 10@xmath2 ergs , or one - thousandth of the magnetic energy injected in run a , is one quarter the energy of run a at the final stage .
so a small amount of magnetic fields from agns can still be amplified to magnetize the whole cluster to the observed level , provided that the icm turbulence is strong .
for the later injection cases , the situations can be quite different . for z05 ,
the magnetic energy never increases , and drops to 7 @xmath1 10@xmath2 ergs at z=0 . as shown in the magnetic field distribution ,
since the magnetic fields are put into the system too late , there is not enough time , or enough mergers , to distribute the magnetic fields to large volumes . for the other two runs , z2 and z1 ,
the magnetic fields still have enough time to expand and get amplified in the icm turbulence .
their magnetic energy increases in a very similar way to cases of the injections at z=3 .
run z1 has a final magnetic energy of 10@xmath3 ergs , and the magnetic energy evolution of z2 is almost identical to the run c after 3 gyr , and seems saturated . in fig .
[ fig : bprofile ] , we present the spherically averaged radial profiles of rms magnetic field strength at low redshifts of runs a to f. by z=1 , the strong initial magnetic field of run a allow it to reach the @xmath10 g level , and fill the virial radius of the cluster .
by z=0.5 , all runs a - f have field strengths of @xmath43 at the cluster center . by z=0 , only the run with the weakest initial field strength , run f , has not filled the virial radius . from z=1 to z=0
, the magnetic field gradually saturates with the icm motions from the center of the cluster to the virial radius . magnetic field strength in the cluster center
is maintained at the micro gauss level , while at larger radii , the field strength is continues to be amplified .
there is a clear trend that the slope of the radial profile decreases with time . in run
a , the magnetic field strength at the virial radius remains at a relatively constant value of 0.2 - 0.5 @xmath10 g from z=0.5 to z=0 .
the magnetic field strenth in the other cases continues to be amplified at the virial radius in this reshift interval , and their field strength profiles in these runs approach the distribution in run a. the run with the weakest injected magnetic field strength , run f , has its field amplified fastest , and over the widest range of radii .
it seems that the profiles of magnetic fields are decided by the icm motions , rather than the initial magnetic field strength , provided there is enough time for the magnetic field to be mixed throughout the cluster and amplified by the turbulence .
[ fig : b_profile_z0 ] shows the radial profiles of magnetic field strength at z=0 for simulations with magnetic field injected after z=3 .
the magnetic fields of z2 are also higher than 1 @xmath10 g and have almost the same profile as run c , except in the central 100 kpc . for run z1
, the magnetic field strength is well distributed through most of the cluster and gets amplified , while the final field strength is at the micro gauss level , and spread out to more than 1.5 mpc away from the center .
but since there is less time for the field to be amplified in the outer part of the cluster , the radial profile is steeper .
the magnetic fields of z05 is much weaker and locally distributed .
it shows that the magnetic field injected this late are not likely to be an important sources of cluster magnetic fields . at z = 0 , the magnetic field strengths in all cases injected before z=1 are at the micro gauss level at the cluster center , and extend to large radii .
magnetic field strength profiles in these cases tend to a power law , with best fit slope of -0.6 between r = 0.2 to 2 mpc .
this is flatter than magnetic profiles seen in some other simulations @xcite .
one possible reason for the flatter radial profiles in our simulations is that we maintain a high resolution in regions with lower density but significant magnetic field strength .
we have also performed a simulations with a less agressive set of refinement criteria , the usual refinement criteria of gas and dark matter overdensity .
this causes lower resolution in the outer regions , where the density is low but the magnetic field is significant .
we see a faster drop in magnetic field strength with lower resolutions at large radius ( details will be discussed in sec . [
sec : comparision ] ) .
the difference in profiles could also be caused by the different initial magnetic fields used in our simulations ( vs uniform initial fields , for example ) .
for instance , the gas collapse plays a less important role in the magnetic field amplifications in our simulations , since there is no magnetic field initially in the collapsing gas . in fig .
[ fig : med_volume_histogram ] , we plot the volume histograms of magnetic field strength of the central 1 mpc sphere to demonstrate the spatial distribution of the magnetic fields .
magnetic fields are mixed with the icm plasma very well for all cases with injections before z=2 .
more than 90% of the icm volume is populated with magnetic fields @xmath44 0.1 @xmath10 g .
the peak of magnetic field strength distribution increases with stronger or earlier magnetic field injections . on the other hand ,
magnetic fields of run z05 are very weak and fill less than 10% of the volume .
figure [ fig : beta_volume_histogram ] illustrates the volume histograms of the plasma @xmath45 ( = nk@xmath46t/@xmath47 ) at z=0 for the same volume .
the thermal energy is always much larger than the magnetic energy .
the peak of the @xmath45 distribution is at @xmath0 500 for run a. the fact that @xmath48 for these runs indicates that in most of the region , magnetic fields are dynamically unimportant and follow the gas passively . to show the correlation between the gas density and the strength of magnetic fields , we plot in fig .
[ fig : med_density ] the two - dimensional distribution of the magnetic field strength vs. the gas density at z=0 for runs a , c , f , z2 , z1 , and z05 .
there is no obvious correlation between the field strength and the gas density .
the distribution is similar for the first five cases .
most of the magnetic field strength is between 0.1 to a few micro gauss , and are mixed with plasmas with a wide range of densities .
this is because the magnetic fields are originally injected in a small volume and are randomly spread through the gas and amplified by the icm turbulence .
so the field strength does not scale as @xmath49 , as in it would if the magnetic field was initially distributed in a large volume and mostly amplified by collapse .
this property may be used to distinguish different magnetic field origins .
the magnetic field strength in z05 is very weak and mostly confined to the high density cluster core region . the top panel of fig .
[ fig : power ] plots the power spectra of kinetic and magnetic energy density of a box of 512@xmath33 ( @xmath0 ( 5.5 mpc)@xmath33 comoving ) enclosing the cluster for run a at different redshifts .
the kinetic power spectra grow rapidly from z=2 to z=1 during the active merger period of the cluster formation , then remain relatively constant level for the rest of the simulation .
the kinetic power spectra show a smooth power law similar to a kolmogorov - like spectrum for incompressible turbulence in the range k @xmath0 0.01 - 0.1 kpc@xmath50 .
they are similar to the spectra obtained from pure hydrodynamic simulations @xcite .
the spectrum steepens gradually for k @xmath26 0.1 on small scales due to several effects : the damping of kinetic energy by the shocks @xcite ; the back reaction of magnetic fields , especially at low redshifts when the magnetic energy density is close to the kinetic energy density at those scales ; and numerical diffusion .
the magnetic energy power spectra at large scales follow the @xmath51 kazantsev law . at early times ,
when the magnetic energy density is much smaller than the kinetic energy density at all scales , the magnetic energy grows exponentially in all scales .
the power laws of both the kinetic and magnetic energy , and the exponential growth of magnetic energy clearly show that the small scale dynamo is operating in the icm ( see * ? ? ?
the characteristic timescale for magnetic energy growth is the eddy turnover time at the dissipation scale ( at k @xmath0 0.1 to 0.2 kpc@xmath50 in this simulation ) , which is @xmath0 1 gyr from z=2 to z=1 , and increases gradually to 3 gyr at z=0 . with the rapid growth from z=2 to z=0.5
, the magnetic energy density reaches equipartition with the kinetic energy density at small scales .
then , the operation of small scale dynamo goes into the second stage . during this stage
, the magnetic energy grows linearly with a rate proportional to the turbulence energy cascade rate @xcite and the peak of magnetic energy moves to larger scales @xcite .
the characteristic timescale is again the eddy turnover time at the dissipation scale , which is @xmath0 2 gyr at z=0.5 and is longer than 3 gyr at z=0 .
the longer eddy turnover time is due to the relaxation of the cluster .
the linear growth of magnetic energy happens between z=0.37 and z=0.16 at a rate of @xmath0 1.5 @xmath1 10@xmath3 ergs / gyr .
after that , the magnetic energy stops growing and even decreases slowly .
we consider the magnetic fields saturated at this time , even though the magnetic and kinetic energy have not reached equipartition globally . here
, the saturation of magnetic fields happens when the operation of small scale dynamo is in the linear growth stage and the relaxation of the cluster makes this linear growth of magnetic energy very slow with the long eddy turnover time .
the magnetic power spectra of six runs of the same volume at z=0 are plotted in the bottom panel of fig .
[ fig : power ] .
we do nt show the kinetic power spectra since they are all almost the same as that of run a. however , the magnetic power spectra for different runs are at quite different levels .
all the magnetic energy density power spectra show a power law @xmath51 over a small range , except for run z05 , so the magnetic fields are indeed amplified by the small scale dynamo process in all these runs .
magnetic fields in simulations with weaker initial magnetic fields or later injection need more time to spread throughout the cluster , and occupy relatively smaller volumes , especially for runs f and z1 .
the evolution of magnetic energy shows that the magnetic fields of runs c and z2 are also in the second stage of small scale dynamo with the linear growth . the magnetic energy density of these runs saturates at a subequipatition level with respect to the kinetic energy density for k @xmath26 0.3 kpc@xmath50 , which is likely because their magnetic fields fill smaller volumes of the cluster .
the magnetic energy of run f and z1 continues to exponentially grow for z @xmath52 0.5 with a lower rate due to the longer eddy turnover time and does not clearly show whether or not the increase becomes linear at z @xmath0 0 . for run z05 , there are not enough big mergers to spread the fields to large enough volume to catch the turbulence motions .
even though there has been 6 gyr since the agn input , its power spectrum shows no sign of the operation of the small scale dynamo .
even though we have solved the ideal mhd equations in our simulations , there is numerical diffusion that has allowed the turbulence to damp and the magnetic fields to diffuse in the icm .
the rate of diffusion is determined by the numerical reynolds number and magnetic reynolds number .
we estimated that these numbers in our simulations are on the order of a few hundred .
the real icm could have reynolds number is on the order of 100 or may be bigger ( up to @xmath01000 ) if magnetic fields are present @xcite , which is close to our simulations .
but its magnetic reynolds number is not well determined , especially in a magnetized turbulent medium .
the small scale dynamo theory and simulations @xcite have shown that the dynamo will operate under our simulation conditions .
most of the previous research of the small scale dynamo is in incompressible turbulence limit .
for the icm , the compression by shocks generated by mergers are very frequent and important , and they could modify the turbulence properties
. this in turn could affect the operation of the small scale dynamo .
further study on the small scale dynamo in compressible turbulence is needed .
another important observational description of magnetic fields in the icm is the frm .
the rm maps are used not only to determine the strength of magnetic fields but also the turbulent structure of them @xcite . for a distant polarized source ,
the frm , which is defined as ratio of change of polarization angle to wavelength squared , in unit of rad m@xmath53 , is @xcite : @xmath54 where @xmath55 is free electron number density in @xmath56 , @xmath57 is the line of sign component of magnetic fields in gauss , and @xmath58 is the comoving path increment per unit redshift in parsec .
we compute the synthetic frm maps by integrating 8 mpc along each axis .
the rm maps observed from different directions are quite similar , so we only show the results along the y axis here
[ fig : rm ] shows the spatial distribution of frm at z = 0 for six cases .
the rm maps of runs with different initial field strengths or of different injection times have similar features but are different in details .
interestingly , the frm map shows not only the small scale variations reminiscent of the icm mhd turbulence , but also show long , narrow filaments , which may be relics of recent mergers .
there are some high @xmath59 regions at the outer part of the cluster , which are consistent with the @xmath26 100 rad m@xmath53 rm observed far from the cluster center in abell 2255 @xcite and coma cluster @xcite . for run z05
, the frm not only covers the smallest area , but has much less small - scale structure , which is because the magnetic fields has had little amplification on small scales . in fig .
[ fig : rm_profile ] , we present the circularly averaged radial profiles of the absolute value of rm ( @xmath59 ) , which is @xmath60 , where @xmath61 is the azimuth angle , and the radial profiles of the standard deviations of rm , which is @xmath62 , where @xmath63 = @xmath64 .
all the @xmath59 profiles are self - similar except for z05 .
the @xmath59 of run a are as high as 1000 rad m@xmath53 at the center , then drop gradually with increasing radius to a few rad m@xmath53 at 2 mpc .
the central @xmath65 of runs f and z1 are just above 300 rad m@xmath53 .
the rm values are sensitive to the initial magnetic fields . with much weaker fields ,
the @xmath59 of z05 are very small and only present in the core region .
the rm vary quickly on small scales , so their standard deviations are large .
we also show the histograms of rm and the cumulative histograms of @xmath59 inside a circle of 500 kpc in fig [ fig : rm_histogram ] . for the cases with magnetic fields injected at z=3 ,
significant rm covers a large portion of the area .
only @xmath0 10@xmath17 of the area is covered with rm close to zero for run a , and just above 20% for run f. the positive and negative rm are roughly symmetric . in run a , the average and standard deviation of rm of this area are -4.09 rad m@xmath53 and 307.75 rad m@xmath53 , respectively .
there is a significant fraction of the area covered with @xmath59 larger than 500 rad m@xmath53 .
the general features of rm distributions of our simulations are consistent with the observations of galaxy clusters ( e.g. * ? ? ?
* ; * ? ? ?
but since present observations can only measure frm on small areas in a cluster , due to the limited number of the observed radio sources behind or embedded in a cluster , it is too early to make a detailed comparison between the distributions of rm in our simulations and observations .
we expect the comparison will be available soon , when many more radio sources will be observed by evla .
we find that the additional refinement criterion of magnetic field strength is very important in following the magnetic fields correctly in our simulations . with only refinement by the baryon and dark matter overdensity ,
the spatial resolution in the outer part of the cluster drops quickly with increasing radius .
this leads to numerical reconnection in low resolution regions , which will destroy magnetic fields .
we reran run a without the magnetic field refinement to test the importance of resolution .
results are given in fig .
[ fig : med_densityonly ] and [ fig : comparison_densityonly ] . without high resolution in a large volume
, the simulation can not resolve the turbulent motions , and in turn , can not capture the small - scale dynamo effect .
in addition , the magnetic fields suffer much larger numerical dissipation with the lower resolutions .
the magnetic energy in the low resolution run keeps decreasing after injection until the end of simulation .
the few small increases in magnetic energy that do occur correspond to major mergers . during the early stage , the magnetic energy drops much faster in the lower resolution run , because most of the magnetic fields numerically reconnect in low resolution regions .
the amplification of the magnetic field fails , most likely because the small scale turbulent motions needed by the dynamo process are not captured by this simulation in a large enough volume .
when the simulations end at z = 0 , there are 5000 times more magnetic energy in the case with magnetic refinement .
though the magnetic energy evolution is very different for different refinement criteria , the cluster still forms in a similar way . at z = 0 ,
the radial profiles of the baryon density and temperature are almost the same . for density
, there is only small difference out to 400 kpc due to the decrease of resolutions in the no magnetic refinement run .
the temperature profiles are different between 200 and 600 kpc , which may be related to the differences caused by resolving the merger shocks with different resolutions .
but the magnetic field strength profiles are totally different for these two simulations . even in the central region with the highest resolution
, the magnetic field strength of the lower resolution case is below 0.1 @xmath10 g , which is much weaker than that of the well refined one .
since both simulations have the same resolution in the core region , up to 50 kpc radius , the dramatic difference in the field strength shows that the magnetic fields amplified in the outer regions of the cluster and transported inwards play a large role in the magnetic field strength in the cluster core .
the radial profile of the run without magnetic field refinement is also steeper than that of the magnetic refinement one .
this clearly shows that the simulation resolution is very important to the magnetic field evolution in the case when magnetic fields are initially injected in a local region .
simulations without high resolution in a large enough region will fail to capture the small - scale dynamo process , and lose most of the magnetic field .
whether differnet seeding processes , for example with initial magnetic fields everywhere , will have similar effect is still unclear .
we are performing some simulations to address this issue .
though the current spatial resolution of our simulations is enough to capture the operation of the small scale dynamo in the icm , we do nt know whether the simulation is converged . with current computational capability , we ca nt have simulations with even higher spatial resolutions in a large enough volume to test convergence .
our current results should be taken only as the lower bound of magnetic fields in the icm .
we will study the convergence of these simulations once the computation becomes possible .
in this work , we have described the magnetic field evolution and properties of the icm with simulations in which magnetic energy was injected by an agn outburst .
we are encouraged by the similarities seen in field evolution as compared to our earlier work @xcite , which used a completely different galaxy cluster , and magnetic energy injected at different strengths and at different times .
using magnetic fields from active galaxies and furthing amplifying them by the icm motions is a robust way to populate the cluster with magnetic fields . on the other hand , since we have not verified whether the numerical convergence is reached in our simulations due to the limit of computational power , the magnetic fields obtained in this paper should be treated as the minimal possible magnetic fields in the icm .
we have simulated the agn injection with a wide range of magnetic energy between 10@xmath2 and 10@xmath3 ergs , which covers the typical range of energy output of radio galaxies @xcite .
it is interesting that , with the huge range in injected magnetic field strength , the galaxy cluster is magnetized to a similar level , especially at the cluster central regions .
the magnetic energy is amplified by more than 900 times in the case with the least initial magnetic field injection .
magnetic energy growth of some runs seems to saturate at low redshift .
the saturated magnetic fields in the cluster center are at the micro gauss level , which is consistent with observed magnetic field strengths , as measured by different observational methods .
further study is needed to understand the saturation mechanism and why the magnetic fields saturate with magnetic energy at a few percent of the kinetic energy @xcite .
we also report simulations with magnetic fields injected at different redshifts . we find that magnetic fields injected as late as z=1 can still effectively magnetize the cluster , while the magnetic fields from agns after z=0.5 will not contribute substantially to the cluster wide magnetic fields .
so as long as the magnetic fields are injected into the cluster before the major mergers in its formation history , the cluster will be significantly magnetized . since more injected magnetic energy
does nt necessarily correspond to significantly higher final magnetic energy , additional initial magnetic fields from multiple agns or from other sources ( e.g. normal galaxies or early universe ) may not necessarily change the magnetic field strength in a cluster .
we have performed a simulation with two agn injections ( each injected in different halos with magnetic energy similar to run d ) . its final magnetic energy ( 7.62 @xmath1 10@xmath3 ergs ) and the magnetic field profile are very similar to the single injection result .
these results will be presented in a future publication .
the injected local magnetic fields mix with the icm efficiently by the mergers and the turbulent icm motions . in all cases except z05 ,
the volume filling factors are quite high .
the magnetic fields fill the central region of the cluster up to 1 mpc .
the magnetic field strength is not strongly correlated with the gas density , as the fields are distributed through a large range of gas density .
we have produced frm maps of the whole cluster based on our simulation results .
they have similar features as the observed cluster rm ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
more detailed comparisons need to wait for the additional observational results from the evla .
we thank s. colgate and h. aluie for discussions .
this work was supported by the ldrd and igpp programs at lanl and by doe / office of fusion energy science .
computations were performed using the institutional computing resources at lanl .
enzo@xmath66mhd is developed at the laboratory for computational astrophysics , ucsd with partial support from nsf grants ast-0708960 and ast-0808184 to m.l.n .
, x. , strauss , m. a. , schneider , d. p. , gunn , j. e. , lupton , r. h. , becker , r. h. , davis , m. , newman , j. a. , richards , g. t. , white , r. l. , anderson , jr .
, j. e. , annis , j. , bahcall , n. a. , brunner , r. j. , csabai , i. , hennessy , g. s. , hindsley , r. b. , fukugita , m. , kunszt , p. z. , ivezi , . , knapp , g. r. , mckay , t. a. , munn , j. a. , pier , j. r. , szalay , a. s. , & york , d. g. 2001 , , 121 , 54 , d. n. , bean , r. , dor , o. , nolta , m. r. , bennett , c. l. , dunkley , j. , hinshaw , g. , jarosik , n. , komatsu , e. , page , l. , peiris , h. v. , verde , l. , halpern , m. , hill , r. s. , kogut , a. , limon , m. , meyer , s. s. , odegard , n. , tucker , g. s. , weiland , j. l. , wollack , e. , & wright , e. l. 2007 , , 170 , 377 | we present a series of cosmological magnetohydrodynamic ( mhd ) simulations that simultaneously follow the formation of a galaxy cluster and evolution of magnetic fields ejected by an active galactic nucleus ( agn ) .
specifically , we investigate the influence of both the epoch of agn ( z @xmath0 3 - 0.5 ) and the agn energy ( @xmath0 3 @xmath1 10@xmath2 - 2 @xmath1 10@xmath3 ergs ) on the final magnetic field distribution in a relatively massive cluster ( m@xmath4 @xmath010@xmath5 m@xmath6 ) .
we find that as long as the agn magnetic fields are ejected before the major mergers in the cluster formation history , magnetic fields can be transported throughout the cluster and can be further amplified by the intra - cluster medium ( icm ) turbulence cause by hierarchical mergers during the cluster formation process .
the total magnetic energy in the cluster can reach @xmath0 @xmath7 ergs , with micro gauss fields distributed over @xmath0 mpc scale . the amplification of the total magnetic energy by the icm turbulence can be significant , up to @xmath01000 times in some cases .
therefore even weak magnetic fields from agns can be used to magnetize the cluster to the observed level .
the final magnetic energy in the icm is determined by the icm turbulent energy , with a weak dependence on the agn injection energy .
we discuss the properties of magnetic fields throughout the cluster and the synthetic faraday rotation measure maps they produce .
we also show that high spatial resolution over most of the magnetic regions of the cluster is very important to capture the small scale dynamo process and maintain the magnetic field structure in our simulations . |
an oriented smooth hypersurface @xmath2 is called _ translating soliton _ of the mean curvature flow ( _ translator _ for short ) if its mean curvature vector field @xmath3 satisfies the differential equation @xmath4 where @xmath5 is a fixed unit vector and @xmath6 stands for the orthogonal projection of @xmath7 to the normal bundle of the immersion @xmath8 .
translators are important in the singularity theory of the mean curvature flow since they often occur as type - ii singularities .
the classic examples of translators ( see figure [ fig : examples ] ) are * any hyperplane containing the direction of translation @xmath7 ; * the _ canonical grim reaper cylinder _
@xmath9 , which is the product of the grim reaper curve and @xmath10 .
a parametrization of the grim reaper curve is given by @xmath11 * the _ translating paraboloid _ or _ bowl solution _ @xmath12 , which is an entire rotationally symmetric strictly convex graphical translator ( * ? ? ?
* lemma 2.2 ) ; * a _ translating catenoid _ or _ winglike translator _
@xmath13 , where @xmath14 , which is a non - convex rotationally symmetric graphical translator ( * ? ? ?
* lemma 2.3 ) .
.45 .45 1.0 1.0 for more examples , we refer the reader to ( * ? ? ?
* examples ) . on the other hand
, we will often discuss the hypothesis in our results using pieces of these examples , in which case the following notation will be very useful : for any @xmath15 , we denote the corresponding closed upper and lower half - space in @xmath0 , respectively , by @xmath16 the aim of this paper is to use these classic examples of translators and the tangency principle ( see section [ sec : non - existence ] ) to deduce interesting consequences on translating solitons of the mean curvature flow .
the structure of the paper is as follows . in section [ sec : non - existence ] , we use the tangency principle to derive two non - existence results for translators . in section
[ sec : height_estimate ] we provide a height estimate for compact translators . in section [ sec : graphical_perturbations ] , it is shown that a graphical perturbation of a graph translator of revolution @xmath17 which is asymptotic to @xmath17 , remains a hypersurface of revolution . as an immediate consequence ,
we give an alternative proof of the uniqueness theorem for complete embedded translating solitons with a single end that are asymptotic to a translating paraboloid ( * ? ? ?
* theorem a ) . finally , in section [ sec : symmetries ] , using the alexandrov s reflection principle we prove that if a compact translator lies between two parallel planes @xmath18 and @xmath19 which are orthogonal to @xmath7 , and its boundary consists of two strictly convex curves contained respectively in @xmath18 and @xmath19 , then the translator inherits the symmetries of its boundary .
we begin with the statement of our main tool throughout this paper , the tangency principle .
[ thm : tangency_principle ] let @xmath20 and @xmath21 be embedded connected translators in @xmath0 with boundaries @xmath22 and @xmath23 . 1 .
@xmath24*interior principle*@xmath25 suppose that there exists a common point @xmath26 in the interior of @xmath20 and @xmath21 where the corresponding tangent spaces coincide and such that @xmath27 lies at one side of @xmath28 .
then @xmath20 coincides with @xmath21 .
@xmath24*boundary principle*@xmath25 suppose that the boundaries @xmath22 and @xmath23 lie in the same hyperplane @xmath29 and that the intersection of @xmath27 , @xmath28 with @xmath29 is transversal .
assume that @xmath27 lies at one side of @xmath28 and that there exists a common point of @xmath22 and @xmath23 where the surfaces @xmath20 and @xmath21 have the same tangent space .
then @xmath20 coincides with @xmath21 . roughly speaking
, this maximum principle says that two different translators can not `` touch '' each other at one interior or boundary point .
thanks to the fact that translating solitons are minimal hypersurfaces in a conformally changed riemannian metric @xcite , the proof is based on the well - known tangency principle for minimal hypersurfaces . for more details , please see ( * ? ? ?
* theorem 2.1 ) .
let us prove now our first non - existence result about translators .
[ thm : non_existence_i ] let @xmath30 be a non - compact embedded connected translator with compact boundary ( possiby empty )
. then @xmath17 can not be contained in any cylinder .
we argue by contradiction .
suppose that @xmath31 is contained in a cylinder @xmath32 .
we distinguish two cases : * case 1 * : the axis of @xmath32 is parallel to the direction of translation @xmath7 .
consider first a _ winglike _
translator @xmath33 with center in the axis of @xmath32 and with radius @xmath34 , so that , in particular , @xmath35 . consider next the family of winglike translators @xmath36 .
since @xmath35 , there must be a @xmath37 $ ] such that @xmath38 intersects @xmath17 for the first time . without loss of generality
, we can assume that this first point of contact is an interior point of both surfaces , otherwise it is at the boundary of @xmath17 , in which case it is sufficient to consider the initial winglike translator @xmath33 located at a higher height ( recall that the boundary of @xmath17 is compact by hypothesis ) .
therefore , by the interior tangency principle , @xmath39 , which contradicts that @xmath17 is a non - compact surface contained in @xmath32 .
* case 2 * : the axis of @xmath32 is not parallel to @xmath7 . in this case
the argument is similar but comparing with a grim reaper cylinder .
let us see it in detail .
due to the compactness of the boundary and the non - compactness of the translator , there exists a real number @xmath40 such that @xmath41 , where @xmath42 .
let @xmath43 be the canonical grim reaper cylinder located in this slab @xmath44 at a large height so that it does not intersect @xmath17 .
then translate it down until it `` touches '' @xmath17 for the first time .
observe that this procedure is feasible because @xmath45 is compact , since by hypothesis the cylinder is tilted .
moreover , as @xmath41 , this point of contact must be in the interior of @xmath17 .
hence , by the interior tangency principle , @xmath46 , a contradiction .
let us make here some remarks concerning the previous theorem [ thm : non_existence_i ] .
a. the result is not true if the translator ( with boundary ) is compact .
a counterexample is the piece of translating paraboloid @xmath12 obtained by cutting this surface with a horizontal plane at any arbitrary but fixed height @xmath47 and considering the lower part , that is , @xmath48 .
+ b. the compactness of the boundary is also necessary .
a counterexample is the intersection of the canonical grim reaper cylinder @xmath9 with a cylinder of arbitrary but fixed radius @xmath14 and axis the @xmath49-axis ; this surface is contained , for instance , in the cylinder of radius @xmath50 and axis the @xmath49-axis . in the following result
we prove that there are no translators _ that resemble a handle _
( see figure [ fig : thm2 ] ) .
more precisely , ] [ thm : non_existence_ii ] there do not exist a connected compact embedded translator in @xmath0 whose boundary is contained in a hyperplane orthogonal to the direction of translation @xmath7 and consists of two strictly convex jordan curves located at distance greater or equal than @xmath51 and such that one of them is not contained in the region enclosed by the other one .
we will denote by @xmath52 to an embedding of @xmath17 , and by @xmath53 to the hyperplane that contained the boundary of @xmath54 .
first , note that @xmath17 must be below the plane @xmath53 .
otherwise , by compactness of @xmath17 , the height function of @xmath17 , @xmath55 , would attain a global maximum .
but recall ( * ? ? ?
* lemma 2.1 ( d ) ) that this height function satisfies the equation @xmath56 , so @xmath57 does not admit any local maxima in the interior , a contradiction .
now consider the segment @xmath58 realizing the distance between the two boundary curves of @xmath17 .
the length of this segment is greater or equal than @xmath51 by hypothesis .
let @xmath59 be the straight line in the direction of @xmath7 passing through the middle point of the segment @xmath58 .
place a canonical grim reaper cylinder @xmath43 in such a way that its lower generatix coincides with @xmath59 .
observe that @xmath43 is strictly contained in a slab @xmath44 defined as the cartesian product of the segment @xmath58 times the line @xmath59 .
initially @xmath43 does not intersect @xmath17 because @xmath17 is below the hyperplane @xmath53
. then translate @xmath43 down following the direction of translation @xmath7 until it intersects @xmath17 for the first time , which necessarily occurs in an interior point of @xmath17 because any of these translations of @xmath43 is strictly contained in the slab @xmath44 .
then , by the interior tangency principle , @xmath46 , which is absurd .
a. in theorem [ thm : non_existence_ii ] , it is necessary that the boundary curves lie in the same hyperplane .
otherwise the result is not true , as the following example shows : the piece of the translating paraboloid which is between two horizontal hyperplanes : @xmath60 , where @xmath61 . + b. moreover , if it is allowed that one of the boundary curves is in the region enclosed by the other one , then there exist translators under the rest of the hypothesis of the theorem [ thm : non_existence_ii ] .
for instance , the intersection of a winglike translator with a lower half - space , @xmath62 , where , obviously , @xmath40 is large enough so that this intersection is non - empty .
our aim in this section is to develop a geometric argument for obtaining an upper bound to the maximum height that a compact embedded translator in @xmath1 can achieve .
[ thm : height_estimate ] let @xmath63 be a connected compact embedded translator whose boundary is a connected curve @xmath64 contained in a plane @xmath53 orthogonal to @xmath7 .
assume that the diameter of @xmath64 is @xmath65 .
then , for all @xmath66 , the distance in @xmath1 from @xmath67 to @xmath53 is less or equal than @xmath68 where @xmath69 is the function given by @xmath70 and @xmath71 the idea is to compare @xmath17 with an appropiate grim reaper cylinder .
first suppose that @xmath72 .
without loss of generality , assume that the diameter of lenght @xmath73 coincides with @xmath74 for an arbitrary but fixed @xmath75 .
consider a canonical grim reaper cylinder @xmath9 and observe that , since @xmath76 , the region between the two parallel planes asymptotic to @xmath9 contains @xmath64 .
hence , this grim reaper cylinder can be translated down until it does not intersect @xmath17 .
now traslate it up until their first point of contact occurs . by the tangency principle
, this must happen at a boundary point of @xmath17 .
observe also that for any point @xmath77 of the grim reaper cylinder , the width between its two `` wings '' is precisely @xmath78 , so the width is @xmath73 when the height is @xmath79 . in conclusion
, this argument shows that @xmath17 must be contained in the compact region enclosed by the intersection of the horizontal plane @xmath53 and a canonical grim reaper cylinder whose lowest point is at distance @xmath79 from @xmath53 , which proves the boundedness if @xmath72 .
second , suppose that @xmath80 .
in this case a canonical grim reaper cylinder @xmath9 can not contained @xmath64 .
it is necessary a dilation of factor @xmath81 .
but then the velocity changes , which does not allow us to use the tangency principle anymore . to overcome this problem , a rotation that maps again the new velocity to @xmath7 can be applied to @xmath82 .
that is , with an appropiate dilation and rotation we can proceed similarly as above . + in detail , consider the canonical grim reaper cylinder @xmath83 apply to @xmath9 a dilation of factor @xmath84 such that @xmath85 , @xmath86 so that @xmath64 fits in the slab determined by the dilated grim reaper cylinder @xmath82 . observe that there are infinite factors of dilation with this property .
a way to parametrize them is to consider @xmath87 , where @xmath88 . for brevity
, we will usually omit the parameter @xmath58 .
+ note that with this dilation the translating velocity changes from @xmath89 to @xmath90 .
a unitary velocity can be achieved noticing that the grim reaper cylinder @xmath82 is invariant under translations in the directions of @xmath91 .
hence , @xmath82 can be considered as a translator in the direction of @xmath92 , for any @xmath15 . in particular , for @xmath93 we have that @xmath94 is a unit vector .
finally , a rotation around the @xmath26-axis is performanced in order to transform @xmath95 into @xmath7 , so that @xmath17 and this dilated and tilted grim reaper cylinder can be compared each other , that is , so that they have the same tranlating velocity vector .
specifically , the angle @xmath96 of rotation must be @xmath97 hence , the rotation matrix @xmath98 is @xmath99 therefore , the surface that will be used to compare with @xmath17 is @xmath100 for brevity , we will denote @xmath101 by @xmath102 .
+ now the idea is to translate @xmath102 until it does not intersect @xmath17 and translate it back until they intersect each other for the first time . by the tangency principle , the first point of contact
must be at the boundary of @xmath17 . to make the computations it is convenient to consider the following static situation , which is equivalent : to determine the intersection of @xmath102 with the cylinder @xmath103 of diameter @xmath73 , @xmath104 and compute the global minimum and maximun of the third coordinate function of the parametrization of this intersection .
+ combining and , we obtain that a parametrization of the intersection of @xmath102 and @xmath103 is @xmath105 \to \real{3}$ ] given by @xmath106 the critical points of @xmath107 correspond to @xmath108 : @xmath109 the points on the boundary of @xmath110 are @xmath111 therefore , the global maximum and minimum of the third coordinate function of @xmath107 are @xmath112 and @xmath113 , respectively .
hence , the boundedness is given in this case by their difference , which is precisely @xmath114 , as claimed .
now observe that the function @xmath114 is positive and @xmath115 hence @xmath114 has a global minimum .
the problem is that it can not be computed analytically .
indeed , the critical points of @xmath114 are the zeros of @xmath116 nevertheless , we can determine an @xmath117 such that @xmath114 is increasing in @xmath118 . specifically , for all @xmath88 , @xmath119 since @xmath120 then @xmath121 once we know that @xmath122 is increasing for @xmath123 , we can easily improve the above lower bound of @xmath124 : @xmath125 now , observe that @xmath126 and taking into account that , due to our previous computations , we can restrict our estimation of @xmath124 to the interval @xmath127 $ ] , we have that it is sufficient to find an @xmath58 such that @xmath128 that is , @xmath129 and the proof is complete . the height estimate is valid in a more general setting : in the statement of the theorem [ thm : height_estimate ] , instead of considering the diameter @xmath73 of @xmath64 , we can assume that the curve @xmath64 is strictly contained in a slab of width @xmath65 , and the proof remains exactly the same .
[ def : graph_perturbation ] let @xmath130 be a connected graph hypersurface given by @xmath131 . let @xmath17 be a hypersurface in @xmath0 . + we say that @xmath17 is a graphical perturbation of @xmath130 if there exists a function @xmath132 such that @xmath17 can be represented as the graph of @xmath133 , that is , @xmath134 we will say that @xmath17 is an asymptotic graphical perturbation of @xmath130 if @xmath17 is a graphical perturbation of @xmath130 and , moreover , for every sequence @xmath135 in @xmath136 such that @xmath137 it holds that @xmath138 . finally , we will say that @xmath17 is a ( an asymptotic ) graphical perturbation of @xmath130 outside a compact set @xmath139 if @xmath140 is a ( an asymptotic ) graphical perturbation of @xmath141 .
note that if there exists a function @xmath142 as in the previous definition , then it is smooth because it is the difference of two graph hypersurfaces , which are always assumed to be smooth in this paper .
roughly speaking , the asymptotic behaviour here means that , outside a bounded region in @xmath17 , @xmath17 is arbitrarily close to @xmath130 .
an example is shown in figure [ fig : perturbation ] .
the goal of this section is to show that if we have two hypersurfaces that are graphically asymptotic outside a compact set , then there are some interesting common properties between them , such as being a graph hypersurface or a hypersurface of revolution .
this is the content of our next theorem . before stating it precisely ,
let us present the idea of the proof .
basically it is another consequence of the tangency principle , comparing the translator _ with a transformation of itself _ according to the following scheme : 1 .
consider @xmath143 a `` copy '' of @xmath17 ; 2 .
translate @xmath143 up , @xmath144 for some @xmath145 , until it does not intersect @xmath17 ; 3 .
apply an isometry @xmath146 of the ambient space to @xmath147 so that @xmath148 ; 4 .
move @xmath149 down until it `` touches '' @xmath17 for the first time . then , by the interior tangency principle , @xmath150 .
hence , @xmath17 is invariant under the isometry @xmath151
. we will follow this scheme in the next results .
for example , * to show that @xmath17 is a graph hypersurface , take @xmath152 ; * to show that @xmath17 is a hypersurface of revolution , consider as @xmath151 any arbitrary but fixed rotation around the axis of symmetry . here by a hypersurface ( or set , in general ) of revolution in euclidean space @xmath0 we mean a hypersurface ( set ) of @xmath0 which is invariant by the action of @xmath153 , the subgroup of the special orthogonal group @xmath154 that fixes a given straight line @xmath59 .
we will assume that all the sets of revolution that appear together are sets of revolution with respect to the same axis unless otherwise stated .
[ lem : graphical ] let @xmath130 be a connected graph translator in @xmath0 .
assume that @xmath155 is , outside a compact set @xmath139 , a translator which is a graphical perturbation of @xmath130 .
suppose that the boundary of @xmath17 is graphical ( possibly empty ) .
then @xmath17 is graphical . obviously , by definition of graphical perturbation , @xmath140 is graphical .
we have to prove that @xmath156 also is . to this end
, just apply the scheme presented above , which clearly works because we deal with a compact region . to avoid contact at the boundary of @xmath156 during step 4 , we work from the very beginning with a bigger compact set @xmath157 with @xmath158 sufficiently large so that the boundary created intersecting @xmath17 with @xmath159 is graphical .
therefore , the contact at the boundary can occur only when @xmath160 comes back to its original position , in which case @xmath17 is graphical , as claimed .
[ cor : outside_compact_set ] if the definitions given in [ def : graph_perturbation ] hold outside a compact set , they hold everywhere .
the case of _ graphical perturbation _ is precisely the content of the previous lemma [ lem : graphical ] . with respect to _
asymptotic graphical perturbation _
, observe simply that this definition is indepedent of what happens in compact regions since it deals with the behaviour of the surfaces at infinity .
[ thm : translator_of_revolution ] let @xmath130 be a connected graph translator of revolution in @xmath0 .
suppose that @xmath17 is , outside a compact set @xmath139 , a translator of @xmath0 which is an asymptotic graphical perturbation of @xmath130 .
assume that the boundary of @xmath17 is graphical ( possibly empty ) and a set of revolution .
then @xmath17 is a hypersurface of revolution . under the hypothesis of theorem [ thm :
translator_of_revolution ] , if there exists the boundary of @xmath17 , then it is not necessarily connected .
for instance , the intersection of a winglike translator with two different and parallel horizontal planes .
but , in general , due to the rotational symmetry hypothesis on the boundary of @xmath17 , each connected component of the boundary of @xmath17 must be contained in a horizontal hyperplane , and indeed it must be a circumference
. we will show that the previous scheme works for any arbitrary but fix element @xmath151 of @xmath153 , where @xmath59 is the axis of symmetry .
indeed , @xmath59 must be parallel to the direction of translation @xmath161 , otherwise @xmath17 would not be graphical .
let us consider such an isometry @xmath151 .
first , by corollary [ cor : outside_compact_set ] , @xmath17 is an asymptotic graphical perturbation of @xmath130 everywhere
. then , there exists @xmath162 ( for instance , @xmath163 ) such that latexmath:[\[\label{rel : slab } that @xmath17 is contained in a slab @xmath44 of diameter @xmath73 centered at @xmath130 : @xmath165 and @xmath166 .
+ since @xmath130 is a hypersurface of revolution by hypothesis , then @xmath44 is a set of revolution .
we can easily argue now that our scheme works : * step 2 is trivially possible to do because @xmath17 is graphical ; * step 3 is achievable because + @xmath166 ( by construction of @xmath44 ) , @xmath167 ( by step 2 ) @xmath168 ; * step 4 : + the first point of contact can not be at infinity because @xmath142 tends to zero at infinity .
+ since the boundary is a graphical set of revolution , the first point of contact can not be at the boundary unless the hypersurface returns to its original position , in which case @xmath169 , as claimed .
* theorem a ) let @xmath2 be a complete embedded translating soliton of the mean curvature flow with a single end that is smoothly asymptotic to a translating paraboloid .
then , @xmath170 is a translating paraboloid .
it is a consequence of our previous theorem [ thm : translator_of_revolution ] , taking @xmath130 as the translating paraboloid .
+ observe that the meaning of _ smoothly asymptotic _ in ( * ? ? ?
* theorem a ) is that there exists a sufficiently large @xmath158 such that @xmath171 can be written as the graph of a function @xmath172 such that @xmath173 where @xmath174 denotes the usual euclidean norm in @xmath175 .
+ now , taking into account that the translating paraboloid is a graph hypersurface for a function @xmath176 satisfying the same asymptotic behaviour @xmath177 then being _ smoothly asymptotic _ clearly implies being _ an asymptotic graphical perturbation_. indeed , from the relation @xmath178 and from and , we deduce that @xmath179 that is , it is sufficient to consider as @xmath142 any smooth function such that @xmath180 ( as @xmath181 ) , i.e. , @xmath182 for all @xmath183 and for some constant @xmath184 .
in this section we apply the method of moving planes @xcite to study compact translators with symmetric boundary .
[ thm : translators_and_symmetries ] let @xmath17 be a connected compact embedded translator in @xmath0 whose boundary consists of two strictly convex curves @xmath185 and @xmath186 contained , respectively , in two parallel planes @xmath18 and @xmath19 which are orthogonal to @xmath7 .
assume that @xmath17 lies between the two planes @xmath18 and @xmath19 , and suppose that the curves @xmath185 and @xmath186 are symmetric with respect to a plane @xmath29 containing the direction of translation @xmath7 .
then @xmath17 is symmetric with respect to the plane @xmath29 .
consider the set @xmath199 : m_{+}(t ) \text { is a graph over } \pi \text { and } m_{+}^{\ast}(t)\geq m_{-}(t ) \},\ ] ] where @xmath200 is a positive real number that exists because of the compactness of @xmath17 .
indeed , @xmath201 , the first point of contact between @xmath17 and a vertical plane coming from @xmath202 , must be a boundary point of @xmath17 , otherwise @xmath17 would coincide with @xmath203 by the interior tangency principle , which is absurd . to show that @xmath205 , we prove that there exists an @xmath208 such that @xmath209\subset \mathcal{a}$ ] .
+ first note that @xmath210 is a bi - graph over its plane of symmetry @xmath29 because , by hypothesis , both boundary curves are strictly convex plane curves . then , in a neighbourhood around @xmath211 , @xmath17 is a graph over @xmath29 .
otherwise , as @xmath17 lies between the planes @xmath18 and @xmath19 , a neighbourhood of @xmath17 around @xmath212 would be contain in the plane @xmath213 , for some @xmath214 , that is , @xmath17 would not be locally around @xmath212 a translator in the direction of @xmath7 , which is absurd . in other words , since @xmath212 is in @xmath210 and it is the first point of contact between @xmath17 and @xmath203 , by continuity this implies that there exists a sufficiently small @xmath208 such that @xmath215 is a graph over @xmath198 for every @xmath216 $ ] .
moreover , as @xmath17 is embedded , considering @xmath208 smaller if necessary , it holds that @xmath217 for every @xmath216 $ ] .
+ for the second part of the claim , let @xmath218 be an arbitrary but fixed number in the interval @xmath219 .
our goal is to prove that @xmath220 .
according to the definition of the set @xmath221 , there are two conditions to be checked , so the proof falls naturally into two parts or steps . _
step 1 _ : @xmath222 is a graph over @xmath29 . + as @xmath206 , then @xmath223 is a graph over @xmath29 . therefore , @xmath215 is a graph over @xmath29 for every @xmath224 $ ] .
in particular , @xmath222 is a graph over @xmath29 .
_ step 2 _ : @xmath225 .
+ on the contrary , if @xmath226 , then , by compactness of @xmath17 , there exists a number @xmath227 such that @xmath228 and @xmath229 intersect for the first time .
furthermore , this first point of contact is an interior point of @xmath230 and @xmath231 because the boundary of @xmath17 consists of two strictly convex plane curves symmetric with respect to @xmath29 .
then @xmath232 by the interior tangency principle .
thus , @xmath233 would be a plane of symmetry of @xmath17 , hence , in particular , it would be a plane of symmetry of the curves @xmath185 and @xmath186 , a contradiction . *
claim 2*. @xmath221 is a closed set of the interval @xmath234 $ ] .
+ the argument is identical to the one in ( * ? ? ?
* theorem a ) : it is proved by contradiction , using the sequential characterization of closed sets ; first it is assumed that the graphical condition in @xmath221 is not satisfied , which contradicts claim 1 ; then the graphical condition and the continuity gives the reflection condition in @xmath221 .
* claim 3*. the minimum of the set @xmath221 is @xmath235 .
+ we argue by contradiction .
suppose @xmath236 .
we will show that then there exists @xmath237 such that @xmath238 , contradicting that @xmath239 is the minimum of @xmath221 .
+ again , we divide the proof into two steps .
_ step 1 _ : there exists @xmath240 such that @xmath241 is a graph over @xmath29 .
+ since @xmath242 , @xmath243 is a graph over @xmath29 .
moreover , there is no point in @xmath243 with normal vector included in @xmath29 .
if there were such a point , let us say that its first coordinate is @xmath244 , then by the tangency principle at the boundary , @xmath245 , that is , @xmath246 would be a plane of symmetry of @xmath17 . in particular , @xmath246 would be a plane of symmetry of the curves @xmath185 and @xmath186 , which contradicts that @xmath247 also is .
thus , @xmath248 as @xmath17 is compact , we have that there exists @xmath240 such that @xmath249.\ ] ] from this fact , together with the compactness of @xmath17 , it follows that @xmath215 can be represented as a graph over the plane @xmath29 for every @xmath250 $ ] .
in particular , @xmath241 is a graph over @xmath29 and the proof of this step is complete .
_ step 2 _ : there exists @xmath251 such that @xmath252 .
+ we are going to show that there exists @xmath253 $ ] such that @xmath254 which in particular implies that @xmath252 , and this step will be proved . + we argue by contradiction .
if it were not true , then there would exist an increasing sequence @xmath255 converging to @xmath239 such that @xmath256 for each natural @xmath257 , denote by @xmath258 a point in the above set . at this point
, we make two key observations : @xmath259,\ ] ] @xmath260 follows from step 1 , that is , from the fact that @xmath241 is a graph over @xmath29 for every @xmath261 $ ] .
therefore , @xmath262 where @xmath263 , simply because , in plain language , where orientation ( right and left ) is considered with respect to the plane @xmath29 .
this is a direct consequence of the definitions , in particular from the meaning of being a graph over a plane .
+ on the other hand , follows from the fact that @xmath242 . indeed ,
if @xmath264 were a set bigger than @xmath265 , then , as @xmath266 , there would be a first point of contact between @xmath267 and @xmath268 , which would be in the interior because the boundary of @xmath17 consists of two strictly convex plane curves symmetric with respect to @xmath29 .
then by the interior tangency principle both surfaces would coincide , hence @xmath269 would be a symmetric plane of @xmath17 , a contradiction .
+ let us come back to the sequence @xmath270 . by the compactness of @xmath17
, we can assume without loss of generality that this sequence converges to a point @xmath271 .
indeed , since @xmath272 , @xmath273 , where the last equality is by . on the other hand , from we see that @xmath274 for each @xmath257 .
thus , @xmath275 , which contradicts that @xmath276 .
the assumption that @xmath17 must be between the two parallel planes @xmath18 and @xmath19 can not be dropped , as the following counterexample shows .
consider the intersection of a winglike solution @xmath277 with two horizontal parallel planes @xmath18 and @xmath19 , so that the lower one , @xmath18 , contains the radius of @xmath277 .
observe that the intersection of each of these planes with @xmath277 consists of two concentric circles .
the counterexample is the piece of @xmath277 between these two planes and whose boundary is the inner circle in @xmath18 and the outer circle in @xmath19 .
simply note that , by theorem [ thm : translators_and_symmetries ] , @xmath17 is symmetric with respect to any plane @xmath29 containing the direction of translation @xmath7 . therefore , @xmath17 is a hypersurface of revolution around @xmath7 .
observe that in this case the translator lies below the plane @xmath53 that contains the curve @xmath64 because @xmath17 is compact and , as a translator , its height function can not attain a local maximum ( * ? ? ?
* lemma 2.1 ( d ) ) .
hence , the same argument using the alexandrov s method proves the corollary .
let @xmath17 be a connected compact embedded translator in @xmath0 whose boundary @xmath64 is a @xmath278-sphere contained in a hyperplane @xmath53 orthogonal to @xmath7 .
then @xmath17 is the compact piece of the translating paraboloid whose boundary coincides with @xmath64 .
let @xmath12 be the compact piece of the translating paraboloid whose boundary coincides with @xmath64 .
place @xmath12 above the plane @xmath53 so that its vertex lies on the same line as the center of @xmath64 .
then translate it down until they `` touch '' for the first time .
there are two possibilities : either they intersect for the first time in an interior point or they do it in a boundary point . in any case , the interior or boundary tangency principle tells us they coincide .
* acknowledgments*. the author thanks professor francisco martn for his great assistance and guidance in the whole preparation of this paper .
the author also thanks professor magdalena rodrguez for comments and discussions that improved the manuscript . | in this article we prove two non - existence results for translating solitons of the mean curvature flow ( _ translators _ for short ) in @xmath0 .
we also obtain an upper bound to the maximum height that a compact embedded translator in @xmath1 can achieve . on the other hand , we study graphical perturbations of translators , showing that asymptotic graphical perturbations of a graph translator of revolution remain a hypersurface of revolution .
finally , we prove that compact translators that lie between two parallel planes inherit the symmetries of their boundaries curves . |
classical double radio sources , also known as frii radio sources ( fanaroff & riley , 1974 ) , have been extensively studied .
numerous surveys of radio sources as well as detailed studies of individual sources have provided a wealth of radio data on many sources .
theoretical models for the sources developed and discussed by various authors ( blandford & rees , 1974 ; scheuer , 1974 , 1982 ; begelman , blandford , & rees , 1984 ; begelman & cioffi , 1989 ; eilek & shore 1989 ; gopal - krishna & wiita 1991 ; daly 1990 , 2000 ) lend much understanding to the physics of the sources .
it is believed that frii sources are powered by highly collimated outflows from an active galactic nucleus ( agn ) , and that the radio emission from these sources is the result of an interaction between the beam , or jet , and the ambient medium . as a result , careful study of frii sources
can yield very useful information about the frii sources , their environments , and their central engines ( e.g. daly 2000 ) .
daly ( 1994 , 1995 ) presents a model for powerful extended frii sources and showed how key parameters of an frii source and its environment , such as the beam power and the ambient gas density , can be estimated using multi - frequency radio data .
results for a sample of powerful 3cr radio sources are also presented by the same author ( daly 1994 , 1995 ) .
a larger sample was compiled and studied in detail ; results from these studies have been presented in a series of papers ( wellman & daly 1996a , b ; wellman 1997 ; wellman , daly , & wan 1997a , b [ hereafter wdw97a , b ] ; wan & daly 1998a , b ; guerra & daly 1996 , 1998 ; guerra 1997 ; daly , guerra , & wan 1998 ; guerra , wan , & daly 1998 ; wan 1998 ; daly 2000 ) .
for example , detailed results on the density and temperature of the ambient medium of the frii sources are presented by wdw97a , b , and the redshift evolution of the characteristic size of an frii source and its application for cosmology are presented by guerra & daly ( 1996 , 1998 ) , and guerra , daly , & wan ( 2000 ) .
this paper presents results on several key parameters of the frii sources and their environments that were not presented and discussed in previous papers .
these parameters include the luminosity in directed kinetic energy of the jet , also known as the beam power , @xmath4 , the total time the agn will produce the collimated outflow , @xmath5 , the thermal pressure of the ambient medium , @xmath6 , the effect of the hot gas on the microwave background radiation , and the total gravitational mass of the host cluster ( we find that the radio sources lie at the centers of gas rich clusters of galaxies ) .
the way that these parameters can be estimated from radio data , and the empirical results for the sample , are presented here .
the paper is structured as follows . a brief description of the sample and data for the sample are presented in [ sec : sample ] .
discussions on the jet luminosity and timescale of jet activity are given in
[ sec : lj ] , the thermal pressure of the ambient medium and estimates of the effect of the hot gas on the microwave background radiation are presented in [ sec : pth ] , and estimating the total gravitational mass of the the host cluster of the radio source is discussed in
[ sec : mass ] .
each of these sections are divided into two subsections , with the first subsection presenting theory and the second one presenting empirical results for the sample .
further discussions of the results and a summary of the paper are presented in [ sec : sum ] .
values of hubble s constant @xmath7 and the de / acceleration parameter @xmath8=0 ( an open empty universe with zero cosmological constant ) are used to estimate all the parameters in this study .
different choices of cosmological parameters , within reasonable limits , give results that are very similar to , and consistent with , those presented here .
the sample used in this study has been described in detail in several previous papers ( wdw97a , b ; guerra & daly 1998 ) .
the readers are referred to these papers for a full description of the sample ; a brief description of it follows .
all the frii sources in this sample are very powerful 3cr sources , having 178 mhz radio power greater than @xmath9 ; such powerful frii sources are referred to as friib sources ( daly 2000 ) , or type 1 frii sources ( leahy & williams 1984 ) .
they are drawn from the samples of leahy , muxlow , & stephens ( 1989 ; hereafter lms89 ) , which contains sources with large angular sizes , and liu , pooley , & riley ( 1992 ; hereafter lpr92 ) , which contains sources with smaller angular extent .
the final sample contains 27 radio lobes from 14 friib radio galaxies , and 14 lobes from 8 friib radio loud quasars .
these sources have redshifts ranging from zero to 2 , and projected core - hot spot separations between @xmath10 kpc and @xmath11 kpc .
note that the questions of projection effects and radio power selection effects have been addressed in great detail by wan & daly ( 1998a , b ) .
wdw97a , b used the radio maps from lms89 and lpr92 to estimate the width of the radio bridge behind the radio hot spot , @xmath12 , the non - thermal pressure of the radio lobe that drives the forward shock front , @xmath13 , and the rate at which the bridge is lengthening , referred to as the lobe propagation velocity @xmath14 .
the radio information may also be used to estimate the beam power and total time the source will be active , as described in 3 . in order to estimate the mach number of lobe advance , high quality radio maps which image large portions of the radio bridge
are required ; the radio bridge is defined as the radio emitting region that lies between the radio hot spot and the origin of the host galaxy or quasar .
there are 16 bridges in the sample for which we have a detection of the mach number of lobe advance , which includes 13 radio galaxy bridges and 3 radio loud quasar bridges .
the mach number may be combined with the lobe propagation velocity @xmath14 to estimate the temperature of the ambient gas ( wdw97a ) .
thus , the ambient gas temperature and pressure may be estimated for these sources .
lower bounds on the mach number , and hence upper bounds on the ambient gas temperature and pressure , are available for the 14 other bridges , including 8 galaxy bridges and 6 quasar bridges , as described in wdw97a , b .
the maps of the sources do not have sufficient dynamic range to allow detailed analyses of their bridge structure , and thus do not have estimates on the mach number of lobe advance , nor on the ambient gas pressure and cluster mass . as a result ,
the sample of sources with estimates of the ambient gas pressure and cluster mass is rather small .
results for radio - loud quasars and radio galaxies are analyzed separately when there are enough data points , as is the case in the study of the beam power . as noted above
, the number of sources with thermal pressure and cluster mass estimates is rather small ( 16 lobes with detections and 14 with bounds ) .
thus it is not practical to separate quasars and galaxies in the study of these parameters .
cygnus a ( 3c 405 ) is the only source with a redshift @xmath15 in our sample , and in some cases , results seem to depend on whether or not this source is included . thus , subsamples both with and without cygnus a are analyzed , and the results included here .
note that cygnus a is not used to normalize any of the quantities studied here , with the exception of the total source lifetime .
note , however , that an identical normalization of the total source lifetime is obtained when cygnus a is excluded from the analysis ( see the companion paper by guerra , daly , & wan 2000 ) .
cygnus a is not needed to normalize any of the quantities studied here .
the jet or beam of an frii radio source refers to the collimated outflow which carries energy from the agn . when the jet impinges upon the external medium , a strong shock is formed and the kinetic energy of the jet is deposited in the vicinity of the shock front , which is marked by the radio hot spot . in principle
, one can calculate the beam power , @xmath16 , carried by the jet by studying the propagation of the hot spots .
however , in practice it is difficult to use hot spot properties to estimate the beam power , since the hot spot is generally not resolved , and the position and properties of radio hot spots vary on short time scales ( laing 1989 ; carilli , perley , and dreher 1988 ; black et al . 1992 . ) .
the model presented and discussed by ( daly 1994 , 1995 ) bypasses these difficulties by using properties of the more stable radio lobes to estimate the luminosity in directed kinetic energy , or the beam power , @xmath4 .
this model is a natural progression following the work of blandford & rees ( 1974 ) , scheuer ( 1974 ) , eilek & shore ( 1989 ) , gopal - krishna & wiita ( 1991 ) , rawlings & saunders ( 1991 ) , and many others .
the method and results presented here are complementary to those of eilek & shore ( 1989 ) in the sense that eilek & shore ( 1989 ) considered the case of a non - supersonic expansion of the radio emitting region , while we consider the supersonic expansion of this region , and the shape of the radio emitting region considered here is different from that considered by eilek & shore .
the results presented here are also complementary to those of gopal - krishna & wiita ( 1991 ) , who consider likely explanations for the decrease in frii radio source size with redshift , and who present a consistent picture in which the increase in radio luminosity and decrease in radio source size with redshift can be understood .
the work presented here differs from that of gopal - krishna & wiita ( 1991 ) in that neither an efficiency nor an evolution of the ambient gas density with redshift are adopted .
instead , we use the radio properties of a source to solve directly for the beam power of the source , as well as several other parameters , including the ambient gas density .
the work presented here on the beam power is similar in spirit to that of rawlings & saunders ( 1991 ) .
the key difference is that here the radio bridge is assumed to be cylindrically symmetric , and thus has a volume that depends on the length of the bridge times the cross - sectional area of the bridge , while rawlings & saunders ( 1991 ) seem to take a bridge volume that is proportional to the cube of the bridge length .
observations and numerical simulations indicate that radio lobes of powerful extended radio sources propagate supersonically relative to the ambient medium ( alexander & leahy 1987 ; prestage & peacock 1988 ; cox , gull , & scheuer 1991 ; daly 1994 ; wdw97a , b ) .
leahy ( 1990 ) shows that the properties of the radio lobe can be used to estimate the rate of energy input . in 3.4.4.3 ,
leahy ( 1990 ) shows that the jet power may be written @xmath17 , since the energy density of a relativistic plasma is 3 times the pressure p ; note that the time rate of change of the volume may be written : @xmath18 since the forward region of the source grows with a velocity @xmath14 and has cross - sectional area @xmath19 .
his equation may be rewritten as : l_j = 4a_l^2 v_l p_l .
[ eq : lj1 ] here @xmath12 is the cross - sectional radius of the radio lobe , @xmath14 is the lobe propagation velocity , and the frii source is assumed to be cylindrically symmetric about the radio axis . using typical units ,
@xmath4 can be expressed as l_j = 3.6 10 ^ 44 ^-1 ( a_l ) ^2 ( v_l c ) ( p_l 10 ^ -10 ^-2 ) , [ eq : lj2 ] all three parameters used to estimate @xmath4 in eq .
( [ eq : lj2 ] ) can be estimated from radio data ( see , for example , daly 1995 , or wdw97a , b ) . the lobe radius @xmath12 can be measured from the radio map ( see wdw97a , b for detail ) .
the lobe propagation velocity @xmath14 can be estimated using multi - frequency observations of radio lobes and bridges , by modeling the spectral aging of relativistic electrons due to synchrotron and inverse compton losses ( wdw97a , b ; wan & daly 1998a , b ; myers & spangler 1985 ; alexander & leahy 1987 ; lms89 ; lpr92 ; carilli 1991 ; perley & taylor 1991 ) . the lobe pressure @xmath13 is given by p_l = ( b^-1.5 + b^2 ) b_min^2 24 , [ eq : pl ] where @xmath20 is the magnetic field strength in the radio lobe under the minimum - energy condition ( miley 1980 ; pacholczyk 1970 ) , @xmath21 is the ratio of the true magnetic field strength @xmath22 to the minimum - energy magnetic field strength : @xmath23 , and the equation is in cgs units . wdw97b
estimate @xmath21 to be about 0.25 , with a source - to - source dispersion less than about @xmath24 .
this value of @xmath21 is consistent with values obtained by other authors ( carilli 1991 ; perley & taylor 1991 ; feigelson 1995 ) , and is used throughout this study . the radio spectral index of an frii source is used in the spectral aging analysis to estimate the lobe propagation velocity @xmath14 .
it is shown in wdw97b that there is a clear correlation between the radio spectral indices of the sources in this sample and redshift .
the spectral index @xmath25 can be expressed as @xmath26 .
the best fitted value of @xmath27 is @xmath28 with a reduced @xmath29 of 1.7 .
such a correlation could be caused by intrinsic curvature in the initial electron energy spectrum or effects of inverse compton cooling on the hot spot spectral index ( see wdw97b ) .
thus , the observed spectral index may not be the appropriate index to used in spectral aging analysis .
this lead wdw97b to also consider correcting the observed spectral indices to zero redshift using the empirical correlation between @xmath25 and @xmath30 for the sample .
results obtained with and without this correction are included here .
the beam power @xmath4 can be used to estimate the total time that the agn will be active , and will produce highly collimated outflows . following daly ( 1994 , 1995 ) and guerra & daly ( 1996 , 1998 ) , the total lifetime of an outflow , defined as @xmath31 ,
is related to the energy extraction rate @xmath4 by a power law : t _ = c l_j^-_/3 , [ eq : tstar ] where @xmath32 is a parameter that has implications for models of energy extraction from the central engine . as described in detail in a companion paper
( see guerra , daly , & wan 2000 ) , eq .
( 4 ) is equivalent to assuming that the source will have an average size close to the average size of the full population located at a similar redshift , and that the average rate of growth of the source will be close to the current rate of growth of the source .
thus , we could have written @xmath33 where @xmath34 is the average size of the full population of friib sources at similar redshift , and @xmath14 is the rate of growth of the source under consideration .
since this is precisely the manner in which the value of @xmath35 is obtained , it is an equivalent approach .
note that the value of @xmath35 is , fortuitously , independent of choice of cosmological parameters , and it is independent of the way the relation is normalized ( see the companion paper of guerra , daly , & wan 2000 ) .
the current best estimate of @xmath32 is @xmath36 ( see the compansion paper of guerra , daly , & wan 2000 ) .
( note that the @xmath32 used here is not related to @xmath37 used in
[ sec : pth ] and [ sec : mass ] for the king density profile . )
the characteristic size of an frii source is related to its lifetime as @xmath38 , where @xmath39 is the normalization in eq .
( [ eq : tstar ] ) .
this normalization factor is chosen so that at @xmath40 , the characteristic size of cygnus a matches the observed average lobe - lobe size for sources at this redshift .
that is , @xmath41 for one side of cygnus a is added to @xmath41 on the other side of cygnus a , and this is set equal to @xmath42 = @xmath43 kpc .
this equation is then solved for the normalization factor @xmath39 ; the uncertainty on @xmath39 is indicated in table 1 .
this normalization is then used in equation ( [ eq : tstar ] ) to obtain an estimate of the total time that the agn will be producing large - scale jets .
note that a nearly identical normalization is obtained when cygnus a is excluded from the data analysis , as shown in a compansion paper ( guerra , daly , & wan 2000 ) .
note also that the constraints on @xmath44 and on cosmological parameters are independent of this normalization , as described in guerra , daly , & wan ( 2000 ) .
the value for @xmath5 estimated using independent information from each side of the source should of course be equal , and are generally very close in value . the value adopted for the beam power , @xmath4 , of cygnus
a used above is listed in table 1 , @xmath14 for cygnus a is obtained from table 1 from wdw97b , and a value of @xmath45 = @xmath46 kpc is adopted from table 3 of guerra , daly , & wan ( 2000 ) . the results on @xmath5 do not depend on the normalization adopted here ; as shown in a companion paper , nearly identical results obtain when cygnus a is excluded from the analysis . and ,
cygnus a is not used to normalize any other quantity .
thus , none of the results presented in this paper rely upon a normalization based on cygnus a. the beam power , @xmath4 , has been estimated using eq .
( [ eq : lj2 ] ) for the lobe propagation velocity estimated using the observed radio spectral index , and the redshift - corrected radio spectral index ; both are listed in table 1 .
typical values of @xmath4 are @xmath47 .
no strong correlation is found between the beam power @xmath4 and the linear size of the source , represented by the core - hot spot separation @xmath48 .
figure 1 is a log - log plot of @xmath4 vs. @xmath48 , where the observed radio spectral index @xmath25 is used .
the figure of @xmath4 vs. @xmath48 after applying the redshift - correction on @xmath25 is almost identical to figure 1 , and is thus not shown here . for radio - loud quasars , the best - fitted line in fig .
1 has a slope of about zero .
a spearman rank correlation analysis shows that the correlation coefficient between @xmath4 and @xmath48 is @xmath49 when no redshift - correction on the spectral index is applied , and @xmath50 when the spectral index correction is applied . here
the number in parentheses is the significance of the correlation .
these results suggest that the correlation between @xmath4 and @xmath48 for radio - loud quasars is insignificant , either with or without the @xmath51 correction .
that is , @xmath4 is independent of @xmath48 for the radio - loud quasars . for radio galaxies , @xmath4 appears to increase slightly with @xmath48 in fig . 1 , and results from the spearman rank correlation analysis hint a marginally significant correlation between @xmath4 and @xmath48 .
the correlation coefficient is about 0.3 , with a significance level of about @xmath52 , which holds with or without the @xmath51 correction .
however , note that the best - fitted line for radio galaxies in fig .
1 has a slope only @xmath53 away from zero , and that a correlation coefficient of about 0.3 is rather weak .
thus it appears that @xmath4 is at most weakly dependent on @xmath48 for radio galaxies .
these results for radio - loud quasars and radio galaxies are consistent with the assumption that @xmath4 is roughly constant over a source s lifetime .
the relationship between @xmath4 and redshift and radio power ( @xmath54 ) has been studied in detail in wan & daly ( 1998a ) .
increases of @xmath4 with @xmath55 and with @xmath54 are observed ( see figures 2 and 3 )
. however , the radio power of the sources in the sample also increases with redshift since the sources used in this study all come from the flux - limited 3cr survey . as a result , which of the @xmath56 and @xmath57 correlations is more significant , and the role of radio power selection effects , needs to be considered carefully .
wan & daly ( 1998a ) used two parameter fitting and partial rank analysis to investigate this .
the two parameter fit yields the following @xmath58 values : @xmath59 for all galaxies , @xmath60 for all galaxies except for cygnus a , and @xmath61 for all radio - loud quasars .
the number in parenthesis is the uncertainty on @xmath58 times @xmath62 , which includes the effect of a reduced @xmath29 @xmath63 that is greater than one for the fit .
this notation will be used throughout this paper .
the values of @xmath58 are more that @xmath53 away from zero , for both radio galaxies and radio - loud quasars .
this means that the @xmath56 correlation is unlikely to be caused purely by radio power selection effects .
that is , a real increase of @xmath4 with redshift exists , though the magnitude of this redshift evolution is not well determined .
the @xmath57 correlation , on the other hand , is rather poorly constrained .
results from the one parameter fit suggest @xmath4 goes roughly proportional to @xmath64 ( figure 3 ) .
the two parameter fit , which takes into account the redshift dependence of radio power , gives values of @xmath65 with large uncertainties : @xmath66 for all galaxies , @xmath67 for all galaxies except for cygnus a , and @xmath68 for all radio - loud quasars .
thus , the exact relationship between @xmath4 and radio power is not clear at the present time .
note , however , large amounts of scatter are seen in the @xmath4 vs. @xmath64 plot ( see figure 3 ) , which is also reflected in the large reduced @xmath29 of the one parameter fit .
for the same beam power , the radio power can vary by about a factor of ten .
this suggests that radio power is not an accurate measure of the beam power , and that the apparent correlation between @xmath4 and @xmath54 may result from the fact that both increase with redshift .
the properties of friib radio sources suggest that the total lifetime of a source decreases systematically with redshift .
this follows because the average size of friib sources decreases with redshift , while the rate of growth of the sources increases with redshift .
another way to state this is : as @xmath4 increases with redshift , the total lifetime of an outflow , @xmath31 , estimated using eq .
( [ eq : tstar ] ) , decreases with redshift .
the values of the full lifetime of the source @xmath31 are listed in table 1 .
typical values of @xmath31 range from about @xmath69 to @xmath2 years .
figure 4 plots @xmath31 as a function of redshift .
it is clear that @xmath31 decreases with @xmath55 , which can explain the fact that the average size of powerful extended 3cr sources decreases with redshift for @xmath70 ( see guerra & daly 1998 ) , while the lobe propagation velocity @xmath14 increases with redshift .
clearly if @xmath14 increases with redshift , and the mean source size @xmath71 decreases with redshift , then @xmath5 must decrease with redshift since @xmath72 .
the sources at high redshift produces more powerful jets ( with larger @xmath4 ) for a shorter period of time ( with smaller @xmath5 ) , which results in smaller average sizes than low - redshift sources .
the studies of wdw97a and b show that both the ambient gas density and temperature of an friib source can be estimated from radio data .
the thermal pressure of the ambient gas @xmath6 is obviously proportional to the product of the density @xmath73 and the temperature @xmath74 . interestingly , as shown below , although the lobe propagation velocity enters into @xmath73 and @xmath74 , it cancels out in the product , so the thermal pressure of the ambient gas can be estimated using single frequency radio data if the radio emission from the bridge is mapped over a large enough region of the bridge .
this offers a completely new method to estimate the thermal pressure of the ambient gas surrounding high - redshift radio sources using only single frequency radio data
. it can be complicated and time consuming to estimate the ambient gas density in the vicinity of classical double radio sources using x - ray observations , especially for sources at high redshift .
an attractive alternative method of estimating @xmath73 in the vicinity of a radio source is to use the radio data . given that the radio lobe represents a strong shock front , the lobe pressure @xmath13 , the ambient gas density @xmath75 , and the lobe propagation velocity are related :
the strong shock jump conditions imply that @xmath76 holds , where @xmath13 is the non - thermal pressure inside the radio lobe , and @xmath75 is the mass density of the ambient gas .
the electron number density of the ambient gas @xmath73 can be estimated using n_a , [ eq : na ] where @xmath77 is the mean molecular weight of the gas in amu , with a value of 0.63 when solar abundances are assumed , and @xmath78 is the proton mass . the ambient gas temperature can be estimated by careful studies of the radio bridge , as discussed in detail in wdw97a .
a brief description of the theory is as follows .
when the non - thermal pressure of the radio bridge is much greater than the ambient gas pressure , the bridge undergoes supersonic expansion .
this lateral expansion can be treated as a blast wave and the expansion velocity is determined by ram pressure confinement .
this causes the width of the bridge , @xmath79 , to vary with time approximately as @xmath80 ( begelman & cioffi 1989 ; daly 1990 ; wdw97a ) .
since the lobe propagation velocity @xmath14 is roughly constant during a given source s lifetime ( lpr92 ; wdw97a , b ) , the lobe front will advance a distance of @xmath81 during @xmath82 as the bridge expands .
this means that at a distance @xmath83 from the hot spot , the bridge expansion time is @xmath84 .
thus the shape of the radio bridge roughly follows a square root law , with the bridge width @xmath85 .
this behavior was in fact found by wdw97a .
as the bridge expands , the pressure inside decreases until it becomes comparable to the ambient gas pressure .
when this occurs , the lateral expansion velocity becomes sonic , causing a break in the functional form of @xmath86 .
the width of the bridge starts to deviate from the square root law and becomes roughly constant as equilibrium is being reached . at the point where such a break occurs
, the lateral expansion velocity is approximately sonic .
thus the sound speed of the ambient gas ( @xmath87 ) can be estimated using c_s .
b = .da dt_b = .da dx_b dxdt = .da dx_b v_l , [ eq : cs ] where @xmath88 denotes the lateral expansion velocity at the break .
the mach number of lobe propagation is defined as @xmath89 .
given the expression for @xmath87 in eq .
( [ eq : cs ] ) , the mach number of lobe propagation is : m ( .da dx_b ) ^-1 = 2 x_b a_b , [ eq : mach ] where @xmath90 is the width of the bridge at the break , and @xmath91 is the position of the break relative to the hot spot ( see wdw97a ) .
it can be seen from eq .
( [ eq : mach ] ) that @xmath92 depends only on the geometrical shape of the radio bridge .
the ambient gas temperature t can be expressed as : t = m_pk c_s^2 ( .da d x _ b ) ^2 v_l^2 v_l^2 m^-2 , [ eq : t ] where @xmath93 is the ratio of specific heats of the ambient gas .
multi - frequency radio maps are required in order to use eqs .
( [ eq : na ] ) and ( [ eq : t ] ) to estimate the ambient gas density and temperature of an frii source , since both @xmath73 and @xmath74 depend on @xmath14 . *
the product @xmath94 is independent of @xmath14 , so @xmath14 is not needed to estimate the thermal pressure of the ambient gas . *
the ambient gas pressure in the vicinity of the radio lobe is simply p_th = ( n_a+n_i ) k t ( p_l v_l^2 ) v_l^2 m^-2 , [ eq : pth1 ] with @xmath13 estimated using eq .
( [ eq : pl ] ) and @xmath92 estimated using eq .
( [ eq : mach ] ) . here
@xmath95 is the number densities of ions , and for a gas with solar abundances , the electron density is @xmath96 . substituting in the numerical constants in eqs .
( [ eq : na ] ) and ( [ eq : t ] ) , the thermal pressure is given by p_th=0.8 p_l m^2 , [ eq : pth2 ] for a gas with a specific heat ratio @xmath97 .
the thermal pressure estimated using eq .
( [ eq : pth2 ] ) depends only on the lobe pressure @xmath13 and the geometrically determined mach number @xmath92 , both of which can be obtained from single frequency observations of the radio lobe and bridge . * no spectral aging analysis is needed in order to estimate @xmath6 , since the dependences of @xmath73 and @xmath74 on @xmath14 cancel .
* as a result , @xmath6 is also independent of whether or not the redshift - correction on the radio spectral index is applied . at low - redshift
, x - ray data can often be used to estimate @xmath73 and @xmath74 , and hence @xmath6 .
however , high - resolution x - ray data are often difficult to obtain for sources at high redshift .
this new method of estimating @xmath6 , using single frequency radio data , offers an attractive alternative .
it provides a powerful tool to study the environments of powerful classical double radio sources , especially those at high - redshift .
since it has been shown that these sources are in cluster - like gaseous environments , the sources may be used to study evolution of gas in clusters of galaxies ( as discussed by daly 1995 ; wdw97a , b ; daly 2000 ) .
the value of @xmath6 , the thermal pressure of the ambient gas in the vicinity of the radio lobe , is listed in table 1 .
most sources in the sample have @xmath6 on the order of ( @xmath98 to , which is typical of gas in low - redshift clusters of galaxies .
this is consistent with results obtained by wdw97(a , b ) , who find cluster - like density and temperature for the ambient gas of the frii sources in the sample .
note that the thermal pressures obtained here do not depend on a spectral aging analysis , whereas density and temperature estimates do .
the x - cluster around cygnus has been observed by arnaud et .
1984 ) and carilli , perley , & harris ( 1994 ) .
the thermal pressure of the ambient gas near its radio lobe is estimated to be about ( see carilli , perley , & harris 1994 ) .
this estimate from x - ray data is consistent with the thermal pressure estimated here for cygnus a ( 3c 405 ) using the new method just described ( see table 1 ) .
the composite pressure profile , @xmath6 as a function of the core - hot spot separation @xmath48 , is shown in figure 5 .
it can be seen that @xmath6 decreases with @xmath48 .
a negative pressure gradient is expected for an isothermal gas distribution that follows the king density profile , which is suggested to be the ambient gas distribution for sources in our sample ( wdw97a , b ) . for this gas distribution ,
the thermal pressure decreases with @xmath48 as p_th(r)= p_th , c ^-3 2 _ 0 , [ eq : pthr ] where @xmath99 is the thermal pressure at the center of the cluster , and @xmath100 is cluster core radius .
the pressure gradient seen figure 5 appears to be consistent with that expected for an cluster gas distribution that is isothermal and follows a king density profile , with a cluster core radius @xmath100 of .
the sources in our sample have values of @xmath48 ranging from . within this radius range , the pressure profile given by eq .
( [ eq : pthr ] ) has an average slope of , for @xmath100 values of , assuming @xmath101 .
these expected slopes are consistent with the best - fit slopes of in figure 5 .
there is some hint of a redshift evolution of the cluster core radius from the data .
figure 6 plots @xmath6 vs. @xmath48 in two redshift bins , @xmath102 and @xmath103 .
the division at @xmath104 is chosen so that the two redshift bins cover about the same range in redshift and contain about the same number of data points . *
it appears that @xmath6 decreases less rapidly with @xmath48 at high redshift than at low redshift , suggesting a larger @xmath100 at high redshift . *
the best - fit slope of @xmath105 in the low - redshift bin is consistent with for the pressure profile given by eq .
( [ eq : pthr ] ) , whereas the best - fit slope of about @xmath106 in the high - redshift bin is consistent with .
this result is still preliminary since only sources with detections of @xmath6 are included in the fits .
better estimates of @xmath6 for sources currently with only upper bounds on @xmath6 will help to better determine whether @xmath100 is evolving with redshift .
the results obtained above are consistent with results obtained by wdw97b .
they study the ambient gas density profile and find that @xmath107 kpc when high- and low - redshift source are considered together .
they also find that that data are consistent with a constant core gas mass model where @xmath100 increases with redshift roughly as @xmath108 .
note that results from the study of @xmath6 do not depend on spectral aging analysis , whereas those from the study of @xmath73 do .
it is thus encouraging that consistent results are obtained from the two studies .
for the following analysis , we use both a non - evolving core radius of , and an evolving core radius of , with a focus on the latter .
the thermal pressure at the center of the cluster ( @xmath99 ) can be estimated by scaling the pressure in the vicinity of the radio lobe ( @xmath6 ) to the cluster center .
this follows because the studies of daly ( 1995 ) and wdw97a , b indicate that these radio sources are located at the centers of clusters of galaxies , so the core - hot spot separation can be used as an estimate of distance from the cluster center .
the current data on @xmath74 , @xmath73 , and @xmath6 are all consistent with the gas distribution being isothermal and following a king profile .
thus we use the pressure profile for such a gas distribution , as given by eq .
( [ eq : pthr ] ) , to estimate the central thermal pressure .
the central pressure is simply p_th , c=
p_th(r ) ^32 _ 0 , [ eq : pthc ] the values of @xmath99 for the sources in our sample are listed in table 1 , where a value of @xmath109 has been assumed .
most sources have values of @xmath99 around , which is rather typical of gas pressure in the core regions of low - redshift galaxies clusters .
results on the redshift evolution of the central gas pressure @xmath99 are somewhat uncertain because @xmath99 seems to correlate with both redshift and radio power when either correlation is considered separately ( see figures 7 and 8) , and it is not clear at present which correlation is more significant . for the non - evolving core radius model , a two - parameter fit of @xmath110 yields @xmath111 and @xmath112 when all sources are included , and @xmath113 and @xmath114 when cygnus a is not included . for the evolving core radius model ,
the two - parameter fit gives @xmath115 and @xmath116 when all sources are included , and @xmath117 and @xmath118 when cygnus a is not included .
note that cygnus a appears to lie in a cooling - flow region while most of the other sources are not in cooling - flow regions ( see wdw97a ; [ sec : mass ] ) . thus fits with and without cygnus a are both performed . in any case , the large uncertainties on @xmath58 and @xmath65 make it hard to determine the magnitude of the redshift evolution of @xmath99 .
the thermal pressure can be used to predict the amount of cosmic microwave background ( cmb ) diminution , also known as the sunyaev - zeldovich ( s - z ) effect , that is expected from the cluster .
the reduction in the cosmic radiation ( @xmath119 ) in the direction of a cluster is given by ( sunyaev & zeldovich 1980 ; sarazin 1988 ; rephaeli 1995 ) i_/i _ = g ( ) y. [ eq : dinu1 ] here @xmath120 is the specific intensity of the cmb at the observing frequency @xmath121 , @xmath122 is the comptonization parameter , and the function @xmath123 is g()=e^ e^-1 , [ eq : gomega ] where @xmath124 , and @xmath125 is the cmb temperature ( mather 1990 ) .
the compton @xmath122 parameter is given by y = _
t n_e dl = _
t m_e c^2 ( n_e n_e + n_i ) p_th dl .
[ eq : ycom ] where @xmath126 ^ 2 $ ] is the thompson electron scattering cross section , @xmath74 is the cluster gas temperature , @xmath127 is the electron density , and @xmath95 is the ion density . using the thermal pressure in a cluster with an isothermal - king gas distribution ( eq . [ [ eq : pthr ] ] ) , this gives y 6.310 ^ -5 ( ) ( ) ( 1+x^2)^-3_0/2 + 1/2 , [ eq : ycom1 ] where @xmath128 , and the gas is taken to have solar abundance so that @xmath129 . for a typical value of @xmath101 , y 1.1210 ^ -4 ( p_th , c 10 ^ -10 ^-2 ) ( r_c 0.25 ) ( 1+x^2)^-1/2 .
[ eq : ycom2 ] at low frequency , clusters with strong radio sources are generally avoided for measurements of the sunyaev - zeldovich effect because emission from the radio source masks the microwave diminution .
such contaminations from radio sources are reduced at high frequency since emission from steep - spectrum radio sources , such as the frii sources studied here , decreases rapidly with increasing frequency . the sunyaev - zeldovich infrared experiment ( suzie )
can measure the s - z effect around 140 ghz , where the amount of cmb intensity diminution is near its peak . at this frequency , a cluster with an isothermal - king density profile will cause a cmb intensity diminution of about ll i _
7.010 ^ -19 ( ) ( ) + ( 1+x^2)^-3_0/2 + 1/2 , [ eq : sz140g1 ] and for a typical value of @xmath101 , i _ 1.2410 ^ -18 ( p_th , c 10 ^ -10 ^-2 ) ( r_c 0.25 ) ( 1+x^2)^-1/2 .
[ eq : sz140g2 ] for a detector with a fwhm beam size @xmath130 , the total amount of cmb diminution within the beam , defined as @xmath131 , is roughly f _ f 24 ( _ b ) ^2 ( p_th , c 10 ^ -10 ^-2 ) ( r_c 0.25 ) , [ eq : sz140g3 ] where @xmath132 is the beam dilution factor , defined as the ratio of the average @xmath119 within the beam to the peak value at the cluster center ; and the suzie fwhm beam size at 140 ghz of 1.7 arcmin ( holzapfel 1997 ) is used to calculate the numerical value . for a gaussian beam , when the hwhm beam size corresponds to 1@xmath100 , @xmath133 , and when the hwhm beam size corresponds to 2@xmath100 , @xmath134 ( see rephaeli 1987 ) . at @xmath135 ,
the suzie beam radius is less than @xmath136 for @xmath137 with no cosmological constant .
thus a cluster with @xmath138 and @xmath139 will have a @xmath140 of about ( 15 to 20 ) mjy within the suzie beam .
the clusters surrounding the frii sources in our sample have @xmath99 on the order of @xmath141 and the data also suggest that @xmath100 increases with redshift , reaching about at @xmath142 . the expected cmb diminution within the suzie beam at 140 ghz for these clusters ranges from several to tens of mjy .
these clusters make good candidates for suzie observations if the fluxes from the radio sources at 140 ghz or their uncertainties are small compared with the expected s - z effect signals .
we are currently in the process of searching for high frequency data on the radio sources in our sample in oder to identify possible suzie observation candidates .
one likely candidate that comes from a preliminary search in the published literature is 3c239 .
the expected cmb diminution for the cluster surrounding it is about @xmath143 mjy within the suzie beam at 140 ghz .
in contrast , extrapolation of the observed spectrum of 3c239 as given by observations at 178 mhz , 10.7 ghz , and 14.9 ghz ( kellermann & pauliny - toth 1973 ; genzel 1976 ; laing , riley , and longair 1983 ) , gives a 140 ghz flux of only 6 mjy .
this is likely to be an overestimate since synchrotron aging and inverse compton cooling can both cause the high frequency spectral index to become steeper than that at low frequency .
the effects of inverse compton cooling can be especially important since this source is at high redshift ( @xmath144 ) , where the energy density of the microwave background is much higher than at low redshift .
thus it appears that emission from this radio source is weak compared with the expected cmb diminution .
observations of the radio source at more frequencies above 14.9 ghz can help to better constrain its flux .
the powerful classical double radio sources studied here are in cluster - like gaseous environments ( see wdw97a , b ) .
the studies of these sources provide information on density and temperature of the ambient gas .
thus it is possible to estimate one of the most important parameters of the cluster , the total gravitational mass , including dark matter .
the total mass can be estimated using the density and temperature profile of the intracluster medium ( icm ) if the gas is in hydrostatic equilibrium .
the sound crossing time in the icm is usually short compared to the age of a high - redshift cluster , and the morphology of the x - ray emission from many low - redshift clusters is often smooth .
thus it is generally believed that hydrostatic equilibrium is a good approximation of the state of the icm for many clusters that are not cooling flow clusters .
assuming spherical symmetry , hydrostatic equilibrium requires = - g m_t(r)r^2 , where @xmath145 is the gas density and @xmath146 is the total gravitating mass within @xmath48 .
this means that m_t(r)=-k t(r ) g m_p r , [ eq : mr1 ] where @xmath147 is the boltzman constant , @xmath77 is the mean molecular weight of the gas in amu , @xmath78 is the proton mass , and @xmath148 and @xmath149 are the temperature and density profiles of the cluster gas .
the most commonly used hydrostatic model of the icm is the isothermal @xmath44-model ( cavaliere & fusco - femiano , 1976,1978 ; sarazin & bahcall 1977 ) , where the clusters gas is isothermal and the density profile follows a modified king model , i.e. , . here
@xmath150 is the core density and @xmath100 is core radius of the gas distribution . using this model for the icm , eq .
( [ eq : mr1 ] ) becomes m_t(r)= 3 _ 0 k g m_p t r .
[ eq : mr2 ] results from numerical simulations of cluster formation suggest that departures of the icm from hydrostatic equilibrium are usually small , and the mass estimated using the standard @xmath44-model is rather accurate ( navarro , frenk & white 1995 ; schindler 1996 ; evrard , metzler , & navarro 1996 ) .
note however , significant temperature decline is observed to occur at outer regions of clusters ( markevitch 1998 ) .
thus , we only use eq .
( [ eq : mr2 ] ) to estimate the cluster mass within @xmath48 rather than extrapolating to large radii .
the mass estimated using eq .
( [ eq : mr2 ] ) is not accurate if the cluster is cooling at the point where the cluster temperature is measured . inside a cooling flow region ,
the temperature measured does not indicate the gravitational potential , since hydrostatic equilibrium conditions do not apply .
considering cooling only by thermal bremsstrahlung , which dominates other mechanisms at typical cluster temperatures , the cooling time of the cluster gas can be estimated using t_cool2.110 ^ -2 ( t10 ^ 7 ) ^1/2 ( n_a^-3 ) ^-1 , [ eq : tcool ] where @xmath74 and @xmath73 are the ambient gas temperature and electron density , respectively ( see wdw97a ) .
note this equation does not depend on the hubble s constant , provided that the estimates of @xmath73 and @xmath74 do not depend on hubble s constant .
given the ambient gas temperature and density estimates , the cooling time for the sources in our sample can be estimated ( see wdw97a ) .
a cooling time less than the age of the universe at that redshift means that the frii source is in a cooling flow region . to estimate the age of the universe at a given redshift ,
an open empty universe is assumed , with a current age of 14 gyr , which corresponds to for @xmath137 with zero cosmological constant .
most sources in our sample do not appear to be cooling at the position where the temperature is measured .
thus their mass estimates are likely to be reliable .
a few sources , including the low - redshift source cygnus a , appear to lie in cooling flow regions .
figure 9 shows the total cluster mass , including dark matter , within radius @xmath48 , @xmath146 , as a function of the core - hot spot separation @xmath48 .
since the sources appear to lie at the centers of clusters of galaxies ( daly 1995 ; wdw97a , b ) , the core - hot spot separation is used to estimate the distance from the cluster center .
the radio spectral index is redshift - corrected , and the cluster core radius is taken to be ( see [ ssec : pth - r ] ) .
clusters that are cooling at the position where the temperature is measured are marked with pentagons .
the plots of @xmath146 vs. @xmath48 for other models , either with or without an @xmath51 correction and/or different core radius evolution , are presented in figures 10 through 12 .
the cluster mass increases with @xmath48 for all models .
the increase is more rapid for models with an evolving core radius than models with a fixed core radius .
this merely reflects the fact that with an increasing @xmath100 , more sources lie within the core region , where the increase of mass with radius is rapid , and varies roughly as @xmath151 . to determine the redshift evolution of the cluster mass , a three parameter fit of is performed since the cluster mass within @xmath48 , @xmath146 , is a function of @xmath48 , and may also be affected by radio power selection effects .
results are listed in table [ tab : ma3p ] for all models considered .
a two parameter fit of is also performed , and results are listed in table [ tab : ma2p ] . for completeness , fits including all sources , all sources except cygnus a , and all sources that are not in cooling flow regions , are all listed in the table . note that cygnus a appears to be in a cooling flow region for all the models considered .
results obtained excluding cooling flow clusters are probably the correct ones to consider .
the sample of clusters with mass estimates is rather small , especially when only non - cooling flow clusters are considered .
thus it is not surprising to see that the best - fit values of @xmath152 have rather large uncertainties , which makes it hard to draw definitive conclusions about the redshift evolution of the cluster mass .
note though , the current data do not indicate any negative evolution of the cluster mass out to a redshift of about two .
this is consistent with results obtained by this group ( daly 1994 ; guerra & daly 1996 , 1998 ; guerra , daly , & wan 2000 ; and daly , guerra , & wan 1998 ) , who study the characteristic size of powerful classical double radio sources and find the data suggest a low value of @xmath153 ; a universe with @xmath154 is ruled out at 99 % confidence ( see guerra , daly , & wan 2000 ) . there are some indications from the data that the redshift evolution of the cluster core radius , and the redshift - correction on the radio spectral index are favored .
the gas mass within @xmath48 , defined as @xmath155 , can be estimated using m_g(r ) = 4_c r_c^3 [ x-^-1(x ) ] , [ eq : mg ] where @xmath156 is the central gas density , @xmath100 is the core radius , @xmath157 , and a typical value of @xmath101 is used for the king density profile .
knowing the gas mass and the total mass within @xmath48 , the gas mass fraction within @xmath48 can be estimated .
the gas mass fraction within radius @xmath48 as a function of redshift is shown in figure 13 , where the radio spectral index is redshift - corrected and the core radius is taken to be .
the values of gas fraction shown in the figure are consistent with observed values for inner regions of many clusters ( donahue 1996 ) , whereas those for models without a redshift - correction on the radio spectral index , and without a core radius evolution seem to be low compared with observed values .
several factors may contribute to the slow decrease of gas fraction with redshift that is seen in figure 13 .
first , the increase of cluster core radius with redshift means that the gas fraction estimated for the high - redshift clusters is over a larger fraction of the cluster core than for the low - redshift clusters .
the gas fraction estimated at high redshift mainly represents the gas fraction inside the cluster core , and it is known that gas fraction increases with increasing radius in clusters ( david , jones , & forman 1995 ) .
second , the mass contribution from galaxies is higher at the cluster center than at large radii ( loewenstein & mushotzky 1996 ) .
thus by sampling a larger fraction of the cluster core at higher redshift , a larger fraction of the total baryon mass is not taken into account at higher redshift .
further , the decrease in gas fraction with redshift could be due to the stripping of gas from galaxies in the cluster core over time .
finally , the cluster gas becomes more concentrated toward the cluster center as it cools , which also causes the gas fraction in the cluster center to increase with time .
these results on cluster mass and gas fraction obtained above are still preliminary due to the small size of the sample .
more sources with estimates on the ambient gas temperature , and hence cluster mass , will help to test these results .
several key parameters of an frii source and its gaseous environment are studied in this paper .
direct estimates of the beam power , which measures the energy extraction rate from the agn by the jet , are obtained .
typical beam powers of about @xmath158 are found for the sources in the sample .
no strong correlation is seen between the beam power and the linear size of the source , which is consistent with the beam power being roughly constant throughout the lifetime of a source .
there is a trend for the beam power to increase with redshift , which is significant even after excluding radio power selection effects .
the magnitude of this redshift evolution , however , is not well determined by the current data . it is well known that the quasar luminosity function undergoes strong evolution between @xmath159 and @xmath160 , with the high - redshift quasars being more luminous than their low - redshift counterparts ( schmidt & green 1983 ; yee & ellingson 1993 ; la franca & cristiani 1997 ) .
this suggests that high - redshift agns are more powerful than low redshift ones .
thus it is perhaps not surprising to find that the beam power , or energy per unit time channeled into the jet by the agn , is also higher at high redshift .
the relationship between the beam power and the radio power is not well constrained after their correlations with redshift are taken into account .
the two parameter fit of yields values of @xmath65 with large uncertainties .
thus it is not clear at present how the beam power and the radio power are related .
however , the large amount of scatter seen in the beam power - radio power relation suggests that radio power is not an accurate measure of beam power .
the beam power @xmath4 can be used to estimate the total lifetime of an outflow produced by an agn . following daly ( 1994 ) and guerra & daly ( 1996 , 1998 ) , the total time for which the outflow will occur , @xmath31 ,
is related to @xmath4 : @xmath161 , where @xmath32 is estimated to be about @xmath162 ( guerra , daly , & wan 2000 ) .
typical lifetimes of about @xmath69 to @xmath2 years are obtained for the sources studied here .
this would is almost precisely the same lifetime as would be obtained by dividing the average size of all friib sources at similar redshift by the rate of growth of the source under consideration .
the lifetime of the outflow decreases with redshift , which explains the decrease of the average size of powerful extended 3cr sources with redshift .
the sources at high redshift produces more powerful jets for a shorter period of time , which results in smaller average sizes than low - redshift sources . a new method of estimating the thermal pressure of the ambient gas in the vicinity of powerful classical double radio source using only single frequency radio data is presented .
the pressure is given by the product of the ambient gas density , estimated using ram pressure confinement of the radio lobe , and the ambient gas temperature , estimated using the mach number for the source and the lobe advance speed .
* it turns out that the lobe propagation velocity cancels out of the product @xmath163 , and the ambient gas pressure can be estimated by studying the shape of the radio bridge , and the non - thermal pressure of the radio lobe ( see [ sec : pth ] ) . *
thus , the thermal pressure , the product of the ambient gas density and temperature , depends only on the non - thermal pressure in the radio lobe and the geometrically determined mach number of lobe advance , both of which can be estimated using single frequency radio data .
this new method to estimate the thermal pressure does not require a spectral aging analysis , and provides a powerful tool to probe the environments of frii sources .
the thermal pressure estimated for cygnus a using this new method agrees with that obtained using x - ray data ( see 4.2 . ) .
thermal pressures on the order of @xmath164 , typical of gas in low - redshift clusters of galaxies , are found for the gaseous environments of the frii sources studied here . there are hints from the current data that the gradient of the composite pressure profile is less steep at high redshift than at low redshift , which can be explained by an increase of the cluster core radius with redshift .
the current data are consistent with a core radius evolution from at @xmath165 to at @xmath166 , which agrees with results obtained from studies of the ambient gas density ( wdw97b ) .
wdw97b find that the data can be described by a model where the core gas density decreases and the core radius increases so that the core gas mass remains roughly constant . the thermal pressures obtained here can be used to estimate the amount of cmb diminution expected from the clusters surrounding the frii sources in the sample .
contaminations from the radio sources can be reduced by observing at high frequency , such as 140 ghz , since emission from the radio sources decreases rapidly with increasing frequency .
the cluster surrounding a source in our sample can be detected by suzie observations at 140 ghz if the flux from the radio source at this frequency or its uncertainty is small compared with the expected s - z effect signal .
a search for high - frequency data of the sources in our sample is ongoing in oder to identify possible suzie observation candidates .
preliminary results suggest that 3c239 is a good candidate .
the gravitational or total mass of the surrounding cluster is estimated for the sources in the sample , assuming hydrostatic equilibrium conditions for the gas .
masses of up to @xmath167 are found for the central regions of the clusters , consistent with typical values for low - redshift clusters .
the redshift evolution of the cluster mass is not well determined .
current data do not indicate negative evolution of the cluster mass .
this is consistent with results obtained by members of this group ( guerra & daly 1996 , 1998 ; guerra , daly , & wan 2000 ; daly , guerra , & wan 1998 ) , who study the characteristic size of frii sources and find the data strongly favor a low value for the density parameter @xmath153 ; a universe with @xmath154 is ruled out at 99 % confidence .
note , however , that with the study presented here , we only study the central regions of clusters . the values of gas mass fraction obtained for the clusters surrounding the frii sources studied here are consistent with observed values for central regions of clusters , after the correlation between the radio spectral index and redshift is taken into account , and a redshift evolution of the cluster core radius is considered .
this suggests that the cluster core radius evolution and the effects of the radio spectral index - redshift correlation are important .
the gas mass fraction seems to decrease slowly with redshift .
the increase of cluster core radius with redshift can be one cause .
another possible explanation is that a large fraction of the gas is still in the galaxies at high redshift , and is later stripped from the galaxies into the cluster at low redshift .
cooling of the cluster gas also tends to cause the gas mass fraction in the cluster center to increase with time .
it is a pleasure to thank greg wellman for important discussions .
we would like to thank paddy leahy for numerous helpful discussions .
the referee deserves special thanks for very helpful comments and suggestions , which have significantly improved the paper .
this work was supported in part by the us national science foundation and the college of liberal arts and sciences at rowan university .
daly , r. a. , guerra , e. j. , & wan , l. 1998 , in fundamental parameters in cosmology : the proceedings of the xxxiiird rencontres de moriond , ed .
j. tran thanh van & y. giraud - heraud ( paris : editions frontieres ) , in press .
3c 55 & g & 0.72 & @xmath169 & @xmath170 & @xmath171 & @xmath172 & @xmath173 & @xmath174 & @xmath175 + & & & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 & @xmath181 & @xmath182 + 3c 68.1 & q & 1.238 & @xmath183 & @xmath184 & @xmath185 & @xmath186 & @xmath187 & @xmath188 & @xmath189 + & & & @xmath190 & @xmath191 & @xmath192 & @xmath193 & @xmath194 & @xmath174 & @xmath195 + 3c 68.2 & g & 1.575 & @xmath196 & @xmath197 & & & & @xmath198 & @xmath199 + & & & @xmath200 & @xmath201 & & & & @xmath202 & @xmath203 + 3c 154 & q & 0.58 & @xmath204 & @xmath205 & & & & @xmath206 & @xmath207 + & & & @xmath208 & @xmath209 & & & & @xmath210 & @xmath211 + 3c 175 & q & 0.768 & @xmath212 & @xmath213 & @xmath214 & @xmath215 & @xmath216 & @xmath217 & @xmath218 + & & & @xmath219 & @xmath220 & @xmath221 & @xmath222 & @xmath223 & @xmath224 & @xmath225 + 3c 239 & g & 1.79 & @xmath226 & @xmath227 & @xmath228 & @xmath229 & @xmath230 & @xmath188 & @xmath189 + 3c 247 & g & 0.749 & @xmath231 & @xmath232 & @xmath233 & @xmath234 & @xmath235 & @xmath236 & @xmath237 + & & & @xmath238 & @xmath239 & & & & @xmath240 & @xmath241 + 3c 254 & q & 0.734 & @xmath242 & @xmath243 & @xmath244 & @xmath245 & @xmath246 & @xmath247 & @xmath248 + 3c 265 & g & 0.811 & @xmath249 & @xmath250 & @xmath251 & @xmath252 & @xmath253 & @xmath254 & @xmath255 + & & & @xmath256 & @xmath257 & @xmath258 & @xmath259 & @xmath260 & @xmath261 & @xmath262 + 3c 267 & g & 1.144 & @xmath263 & @xmath264 & @xmath265 & @xmath266 & @xmath267 & @xmath268 & @xmath269 + & & & @xmath270 & @xmath271 & @xmath272 & @xmath273 & @xmath274 & @xmath275 & @xmath175 + 3c 268.1 & g & 0.974 & @xmath276 & @xmath277 & @xmath278 & @xmath279 & @xmath280 & @xmath281 & @xmath282 + & & & @xmath283 & @xmath284 & @xmath285 & @xmath286 & @xmath287 & @xmath288 & @xmath275 + 3c 268.4 & q & 1.4 & @xmath289 & @xmath290 & @xmath252 & @xmath291 & @xmath292 & @xmath293 & @xmath294 + 3c 270.1 & q & 1.519 & @xmath295 & @xmath296 & & & & @xmath293 & @xmath297 + & & & @xmath298 & @xmath299 & @xmath300 & @xmath301 & @xmath302 & @xmath303 & @xmath202 + 3c 275.1 & q & 0.557 & @xmath304 & @xmath305 & @xmath306 & @xmath307 & @xmath308 & @xmath309 & @xmath310 + & & & @xmath311 & @xmath312 & @xmath313 & @xmath314 & @xmath245 & @xmath315 & @xmath316 + 3c 280 & g & 0.996 & @xmath317 & @xmath318 & & & & @xmath195 & @xmath319 + & & & @xmath320 & @xmath321 & & & & @xmath322 & @xmath323 + 3c 289 & g & 0.967 & @xmath324 & @xmath325 & @xmath326 & @xmath327 & @xmath328 & @xmath329 & @xmath322 + & & & @xmath330 & @xmath331 & @xmath332 & @xmath333 & @xmath245 & @xmath217 & @xmath334 + 3c 322 & g & 1.681 & @xmath335 & @xmath336 & & & & @xmath28 & @xmath337 + & & & @xmath338 & @xmath184 & @xmath339 & @xmath340 & @xmath341 & @xmath188 & @xmath189 + 3c 330 & g & 0.549 & @xmath342 & @xmath343 & @xmath344 & @xmath345 & @xmath346 & @xmath347 & @xmath348 + & & & @xmath349 & @xmath350 & @xmath351 & @xmath352 & @xmath353 & @xmath354 & @xmath355 + 3c 334 & q & 0.555 & @xmath356 & @xmath357 & & & & @xmath358 & @xmath359 + & & & @xmath360 & @xmath361 & & & & @xmath362 & @xmath363 + 3c 356 & g & 1.079 & @xmath364 & @xmath365 & @xmath366 & @xmath367 & @xmath368 & @xmath369 & @xmath370 + & & & @xmath371 & @xmath372 & @xmath373 & @xmath374 & @xmath375 & @xmath376 & @xmath294 + 3c 405 & g & 0.056 & @xmath377 & @xmath378 & @xmath379 & @xmath380 & @xmath381 & @xmath382 & @xmath383 + & & & @xmath384 & @xmath385 & @xmath386 & @xmath387 & @xmath388 & @xmath389 & @xmath247 + 3c 427.1 & g & 0.572 & @xmath390 & @xmath391 & @xmath287 & @xmath392 & @xmath393 & @xmath394 & @xmath395 + & & & @xmath396 & @xmath397 & @xmath398 & @xmath399 & @xmath400 & @xmath401 & @xmath402 + lrllrrrr n.c.cl .
& 11 & yes & yes & 2.38 @xmath403 0.70(0.52 ) & 6.86 @xmath403 5.48(4.1 ) & -1.31 @xmath403 1.57(1.17 ) & 0.56 + g+q - c & 14 & yes & yes & 2.95 @xmath403 0.31(0.23 ) & 2.72 @xmath403 3.80(2.84 ) & -0.08 @xmath403 1.01(0.76 ) & 0.56 + g+q & 16 & yes & yes & 2.88 @xmath403 0.30(0.23 ) & -0.73 @xmath403 0.96(0.73 ) & 0.80 @xmath403 0.37(0.28 ) & 0.58 + + + n.c.cl . & 13 & no & yes & 3.05 @xmath403 0.36(0.28 ) & 4.46 @xmath403 4.11(3.18 ) & -0.13 @xmath403 1.06(0.82 ) & 0.60 + g+q - c & 14 & no & yes & 2.95 @xmath403 0.31(0.23 ) & 3.66 @xmath403 3.82(2.88 ) & 0.02 @xmath403 1.02(0.77 ) & 0.57 + g+q & 16 & no & yes & 2.89 @xmath403 0.30(0.23 ) & 0.49 @xmath403 0.96(0.73 ) & 0.83 @xmath403 0.37(0.28 ) & 0.58 + + + n.c.cl . &
11 & yes & no & 1.36 @xmath403 0.63(0.52 ) & 7.40 @xmath403 4.55(3.72 ) & -0.93 @xmath403 1.33(1.09 ) & 0.67 + g+q - c & 14 & yes & no & 2.12 @xmath403 0.27(0.23 ) & 3.18 @xmath403 3.33(2.85 ) & 0.44 @xmath403 0.90(0.77 ) & 0.73 + g+q & 16 & yes & no & 2.11 @xmath403 0.26(0.21 ) & 1.63 @xmath403 0.91(0.75 ) & 0.84 @xmath403 0.33(0.27 ) & 0.68 + + + n.c.cl .
& 13 & no & no & 2.17 @xmath403 0.31(0.28 ) & 4.45 @xmath403 3.40(3.10 ) & 0.47 @xmath403 0.88(0.80 ) & 0.83 + g+q - c & 14 & no & no & 2.12 @xmath403 0.27(0.23 ) & 4.09 @xmath403 3.20(2.77 ) & 0.53 @xmath403 0.86(0.74 ) & 0.75 + g+q & 16 & no & no & 2.11 @xmath403 0.27(0.22 ) & 2.84 @xmath403 0.91(0.76 ) & 0.86 @xmath403 0.33(0.27 ) & 0.69 + n.c.cl . & 11 & yes & yes & 2.70 @xmath403 0.59(0.45 ) & 2.39 @xmath403 1.24(0.94 ) & 0.58 + g+q - c & 14 & yes & yes & 2.95 @xmath403 0.31(0.22 ) & 2.43 @xmath403 1.05(0.75 ) & 0.51 + g+q & 16 & yes & yes & 2.67 @xmath403 0.29(0.28 ) & 0.63 @xmath403 0.73(0.69 ) & 0.90 + + + n.c.cl .
& 13 & no & yes & 3.04 @xmath403 0.35(0.26 ) & 3.99 @xmath403 1.16(0.86 ) & 0.55 + g+q - c & 14 & no & yes & 2.95 @xmath403 0.31(0.22 ) & 3.74 @xmath403 1.06(0.76 ) & 0.52 + g+q & 16 & no & yes & 2.67 @xmath403 0.29(0.28 ) & 1.90 @xmath403 0.73(0.70 ) & 0.93 + + + n.c.cl .
& 11 & yes & no & 1.60 @xmath403 0.52(0.42 ) & 4.30 @xmath403 1.04(0.84 ) & 0.65 + g+q - c & 14 & yes & no & 2.13 @xmath403 0.27(0.22 ) & 4.75 @xmath403 0.83(0.68 ) & 0.68 + g+q & 16 & yes & no & 1.88 @xmath403 0.25(0.27 ) & 3.34 @xmath403 0.62(0.66 ) & 1.13 + + + n.c.cl .
& 13 & no & no & 2.20 @xmath403 0.31(0.27 ) & 6.19 @xmath403 0.94(0.82 ) & 0.77 + g+q - c & 14 & no & no & 2.13 @xmath403 0.27(0.23 ) & 6.00 @xmath403 0.83(0.70 ) & 0.72 + g+q & 16 & no & no & 1.88 @xmath403 0.25(0.27 ) & 4.57 @xmath403 0.62(0.67 ) & 1.16 + | multi - frequency radio observations of the radio bridge of a powerful classical double radio source can be used to determine : the beam power of the jets emanating from the agn ; the total time the source will actively produce jets that power large - scale radio emission ; the thermal pressure of the medium in the vicinity of the radio source ; and the total mass , including dark matter , of the galaxy or cluster of galaxies traced by the ambient gas that surrounds the radio source .
the theoretical constructs that allow a determination of each of these quantities using radio observations are presented and discussed .
empirical determinations of each of these quantities are obtained and analyzed for 22 radio sources ; cygnus a is one of the sources in the sample , but it is not used , or needed , to normalize any of these quantities . a sample of 14 radio galaxies and 8 radio loud quasars with redshifts between zero and two was studied in detail ; each source has enough radio information to be able to determine each of the physical parameters listed above .
the beam power was determined for each beam of each agn ( so there are two numbers for each source ) ; these agn are highly symmetric in terms of beam powers .
typical beam powers are about @xmath0
. no strong correlation is seen between the beam power and the core - hot spot separation , which suggests that the beam power is roughly constant over the lifetime of a source .
the beam power increases with redshift , which is significant after excluding correlations between the radio power and redshift .
the relationship between beam power and radio power is not well constrained by the current data .
the characteristic or total time the agn will actively produce a collimated outflow is estimated .
typical total lifetimes are @xmath1 to @xmath2 ) years .
total source lifetimes decrease with redshift .
this decrease in total lifetime with increasing redshift can explain the decrease in the average source size ( hot spot - hot spot separation ) with redshift .
thus , high - redshift sources are smaller because they have shorter lifetimes ; note that higher redshift sources grow more rapidly than low redshift sources .
a new method of estimating the thermal pressure of the ambient gas in the vicinity of a powerful classical double radio source is presented .
this new estimate is independent of synchrotron and inverse compton aging arguments , and depends only upon the properties of the radio lobe and the shape of the radio bridge .
a detailed radio map of the radio bridge at a single radio frequency can be used to estimate the thermal pressure of the ambient gas .
thermal pressures on the order of @xmath3 , typical of gas in low - redshift clusters of galaxies , are found for the environments of the sources studied here .
it is shown that appreciable amounts of cosmic microwave background diminution ( the sunyaev - zeldovich effect ) are expected from many of these clusters
. this could be detected at high frequency where the emission from the radio sources is weak .
the total gravitational mass of the host cluster of galaxies is estimated using the composite pressure profile and the equation of hydrostatic equilibrium for cluster gas .
total masses , and mass - density profiles , similar to those of low - redshift clusters of galaxies are obtained .
thus , some clusters of galaxies , or cores of clusters , exist at redshifts of one to two .
the redshift evolution of the cluster mass is not well determined at present .
the current data do not indicate any negative evolution of the cluster mass , which may have implications for models of evolution of structure in the universe . |
in 1982 , joyce @xcite and matveev @xcite introduced the notion of quandles .
a _ quandle _ is a non - empty set equipped with a binary operation satisfying three axioms that correspond to the three reidemeister moves on knot diagrams .
alexander quandles form an important class of quandles .
see section [ sec : fundamental properties ] for details of quandles .
the history of colorings of knot diagrams goes back to fox colorings @xcite , that correspond to homomorphisms of knot groups to the dihedral groups .
it is well - known that if the determinant of a given knot is divisible by an integer @xmath1 with @xmath2 , then any diagram of the knot is fox @xmath1-colorable @xcite .
such colorings are generalized to quandle colorings ( see section [ sec : fundamental properties ] ) , and some conditions on the alexander polynomial of a given knot so that its diagrams have non - trivial colorings with a given alexander quandle are shown in @xcite . in this paper
, we study a generalization of fox - colorings which corresponds to a class of alexander quandle colorings .
let @xmath3 be an oriented classical link in the @xmath4-sphere @xmath5 , @xmath6 a link diagram of @xmath3 on a @xmath7-sphere @xmath8 in @xmath5 , and @xmath9 the number of crossings of @xmath6 .
cutting @xmath6 at the @xmath9 undercrossing points , we obtain @xmath9 arcs , which we call _ strands _ of @xmath6 .
let @xmath10 be the set of strands of @xmath6 .
let @xmath1 be an integer larger than @xmath0 , and @xmath11 and @xmath12 be positive integers with @xmath13 and @xmath14 , i.e. , @xmath1 and @xmath11 are coprime and so are @xmath1 and @xmath12 .
let @xmath15 be the cyclic group of order @xmath1 .
a map @xmath16 is called a _
linear @xmath1-coloring _ , or more precisely , a _ @xmath17-coloring
_ if it satisfies the following _ crossing condition _ at each crossing of @xmath6 .
let @xmath18 be a crossing point of @xmath6 , @xmath19 the strand which goes over @xmath18 , @xmath20 the strand which lies on the right hand of @xmath19 , and @xmath21 the strand on the left hand of @xmath19 .
see figure [ fig : crossingcondition ] .
then the crossing condition at @xmath18 is given by the equation @xmath22 in @xmath15 .
this notion coincides with the usual fox @xmath1-coloring @xcite when @xmath23 and @xmath24 . for general @xmath11 and @xmath12 ,
this is no more than a coloring with the alexander quandle @xmath25 , see section [ sec : fundamental properties ] .
a linear @xmath1-coloring is called _ trivial _ if it is a constant map .
if the diagram @xmath6 admits a non - trivial linear @xmath1-coloring , the link @xmath3 is called _ linear @xmath1-colorable _ or _ @xmath25-colorable_. the number of all @xmath17-colorings is called the _
@xmath17-coloring number_. the @xmath17-colorability and @xmath17-coloring number are invariants of an oriented link .
the @xmath26-torus knot ( @xmath27 in rolfsen s table @xcite ) has the determinant @xmath0 , and does not have a non - trivial fox @xmath1-coloring , refer to @xcite . however , it has a non - trivial @xmath28-coloring as in figure [ fig:10_124 ] , which is also a @xmath29-coloring .
-torus knot @xmath27,width=170 ] in this paper , @xmath30 denotes the alexander polynomial of a knot @xmath3 , where @xmath20 is the variable .
we can assume that the constant term of @xmath30 is of the lowest degree by multiplying it by @xmath31 for some integer @xmath32 . let @xmath33 be the coefficient of @xmath31 in @xmath30 .
that is , @xmath34 with @xmath35 . in the case where @xmath1 is a prime number , @xmath23 and @xmath24 ,
( 1 ) in the next proposition is well - known , refer to @xcite .
note that @xmath1 may be non - prime in ( 1 ) .
when @xmath1 is prime , ( 1 ) is a part of corollary 2 in @xcite by a. inoue .
[ lem : fundamentalproperty ] 1 .
let @xmath1 be an integer larger than @xmath0 , and @xmath11 , @xmath12 be positive integers coprime with @xmath1 .
a knot @xmath3 is @xmath36-colorable if and only if there exists an odd prime natural number @xmath37 such that @xmath38 and @xmath39 ( mod @xmath37 ) , where @xmath40 is the element of @xmath15 with @xmath41 .
2 . let @xmath1 be an odd prime natural number .
suppose that a knot @xmath3 is @xmath42-colorable .
then it is also @xmath43-colorable , where @xmath44 is the element of @xmath15 with @xmath45 .
moreover , @xmath3 is also @xmath46-colorable .
if a knot @xmath3 has the alexander polynomial @xmath47 , then it is @xmath48-colorable for some odd integer @xmath1 larger than @xmath7 and some positive integer @xmath11 with @xmath49 .
in fact , this is the case for @xmath11 and @xmath50 , where @xmath11 is a positive integer with ( i ) @xmath51 and ( ii ) @xmath52 .
4 . for an arbitrarily given knot @xmath3 and an arbitrarily given integer @xmath1 larger than @xmath7
, there is a finite algorithm to determine if a given knot @xmath3 is linear @xmath1-colorable or not .
5 . for an arbitrarily given knot @xmath3 with the alexander polynomial @xmath47
, there is a finite algorithm to determine the minimal number @xmath1 for which @xmath3 is linear @xmath1-colorable .
[ rem : colorable ] ( 1 ) in the above proposition implies the three facts below .
let @xmath1 , @xmath11 and @xmath12 be integers as in proposition [ lem : fundamentalproperty ] ( 1 ) . 1 .
a knot @xmath3 is @xmath36-colorable if @xmath53 ( mod @xmath1 ) .
the converse is also true if @xmath1 is an odd prime natural number .
if a knot @xmath3 is @xmath36-colorable , and @xmath54 is a multiple of @xmath1 , then @xmath3 is @xmath55-colorable .
if a knot @xmath3 is @xmath36-colorable , then @xmath3 is @xmath56-colorable for some odd prime natural number @xmath37 by which @xmath1 is divisible . in remark [ rem :
colorable ] ( a ) , the converse is not true if @xmath1 is not prime .
actually , when @xmath3 is the trefoil knot , @xmath3 is @xmath57-colorable , while @xmath58 and @xmath59 ( mod @xmath60 ) . ,
width=188 ] the _ minimal quandle order _ of a knot @xmath3 , denoted by @xmath61 , is the minimal number among the orders of quandles each of which gives a non - trivial quandle coloring to a diagram of the knot , that is , @xmath62 in section [ sect : minimal ] , we study the minimal quandle orders for twist knots .
we have the following theorem : [ theorem : twist ] let @xmath3 be the twist knot with crossing number @xmath9 shown in figure [ fig : twist ] . 1
. we have @xmath63 if and only if @xmath64 @xmath65 .
in fact , @xmath66 gives the minimal quandle order .
we have @xmath67 if and only if @xmath68 for @xmath69 and some non - negative integer @xmath70 .
in fact , the tetrahedron quandle gives the minimal quandle order .
( see the beginnning of section 3 for the definition of the tetrahedron quandle . )
we have @xmath71 if and only if @xmath72 for some non - negative integers @xmath73 with @xmath74 or @xmath75 .
especially when @xmath76 ( resp .
@xmath77 ) , @xmath78 ( resp .
@xmath79 ) gives the minimal quandle order .
we have @xmath80 if and only if @xmath81 for some non - negative integers @xmath73 with @xmath82 , @xmath83 , @xmath84 , @xmath85 or @xmath86 .
especially when @xmath87 ( resp .
@xmath88 or @xmath89 ) , @xmath90 ( resp . @xmath91 or @xmath92 ) gives the minimal quandle order .
we have @xmath93 , otherwise .
a _ quandle _ is defined to be a set @xmath94 with a binary operation @xmath95 satisfying the following properties : ( 1 ) for any @xmath96 , @xmath97 . ( 2 ) for any pair of @xmath98 , there exists a unique @xmath99 such that @xmath100 .
( 3 ) for any triple of @xmath101 , @xmath102 .
let @xmath103 be quandles .
a map @xmath104 is said to be a _ homomorphism _ if @xmath105 holds for all @xmath106 , where @xmath107 and @xmath108 are quandle operations in @xmath109 and @xmath110 respectively .
if such a map @xmath111 is bijective , then it is called an _ isomorphism _ , and we say that @xmath109 and @xmath110 are _
isomorphic_. alexander quandles form an important class of quandles .
let @xmath112 be the laurent polynomial ring @xmath113 $ ] .
then , any @xmath112-module @xmath114 is a quandle with the operation @xmath115 defined by @xmath116 , that is called an _
alexander quandle_. let @xmath94 be a finite quandle .
let @xmath6 be a diagram of a given oriented classical link @xmath3 , and let @xmath10 be the set of strands .
a _ ( quandle ) coloring _ is a map @xmath117 such that at every crossing , the relation @xmath118 holds , where @xmath20 , @xmath19 and @xmath21 are the strands around the crossing located as shown in figure [ fig : crossingcondition ] .
a quandle coloring is called _ trivial _ if it is a constant map .
if @xmath6 admits a non - trivial quandle coloring , the link @xmath3 is called _
@xmath94-colorable _ or _ quandle colorable by @xmath94_. it is well - known that @xmath94-colorability is an invariant of a link .
the crossing condition @xmath22 in @xmath15 in section [ sect : introduction ] can be deformed to @xmath119 , where @xmath44 means the element in @xmath120 with @xmath121 .
if we define the operation @xmath122 in @xmath15 by @xmath123 for any @xmath124 , then this operation gives a quandle operation in @xmath15 .
this implies that linear colorings are a kind of quandle coloring .
in fact , the quandle @xmath125 is isomorphic to the alexander quandle @xmath126/(t+\ell \overline{k})$ ] .
( since @xmath127 , the map @xmath128/(t+\ell \overline{k})$ ] defined by @xmath129 gives a quandle isomorphism . ) hence we have the following lemma .
[ lem : affinecyclicquandle ] let @xmath130 be positive integers each of which is coprime with @xmath1 . if @xmath131 holds , then the two quandles @xmath125 and @xmath132 are isomorphic .
this implies that the notion of @xmath133-coloring is equivalent to that of @xmath134-coloring ( or @xmath135-coloring ) .
[ remark : nelson ] nelson @xcite studied a class of alexander quandles called _ linear _ , which corresponds to our quandles . by his result , we see that for positive integers @xmath12 and @xmath11 each of which is coprime with @xmath1 , the quandle @xmath125 is trivial if and only if gcd@xmath136 , and that the quandle @xmath125 is indecomposable ( or connected ) if and only if gcd@xmath137 .
when @xmath138gcd@xmath139 , the quandle @xmath125 is decomposed into @xmath140 orbits @xmath141 , @xmath142 , since @xmath143 .
see @xcite for details of orbit decompositions . the orbit @xmath144 forms a quandle under the operation @xmath122 isomorphic to @xmath145 by the map @xmath146 .
since an image of a coloring of a knot is contained in an orbit we can obtain a non - trivial linear @xmath54-coloring with an indecomposable quandle @xmath147 from any non - trivial linear @xmath1-coloring by repeating the operations of taking an orbit .
we prove proposition [ lem : fundamentalproperty ] .
note that for the alexander polynomial of a knot with the form given in section [ sect : introduction ] , the following properties are well - known : ( i ) for any @xmath148 , @xmath149 and ( ii ) @xmath150 . additionally , by the properties ( i ) and ( ii ) , we see that @xmath140 is an even number and @xmath151 is an odd number .
we need to prove remark [ rem : colorable ] ( a ) , ( b ) and ( c ) before proving proposition [ lem : fundamentalproperty](1 ) . by lemma [ lem : affinecyclicquandle ] , it is enough to consider the case of @xmath24 .
the proof is almost the same as that for the usual @xmath1-coloring with @xmath23 and @xmath24 , refer to @xcite .
let @xmath6 be a diagram of the knot @xmath3 .
we number the crossings of @xmath6 , and call them @xmath152 . let @xmath153 be the strands of @xmath6 .
then the crossing condition at @xmath154 is of the form @xmath155 , where @xmath156 is the strand going over the crossing @xmath154 , @xmath157 the strand on the right hand of @xmath156 and @xmath158 the strand on the left hand of @xmath156 .
see figure [ fig : crossingcondition2 ] .
let @xmath114 be the coefficient matrix of the system of equalities of crossing condition for all the crossings .
the @xmath159th element is @xmath0 , the @xmath160th element is @xmath161 and the @xmath162th element is @xmath11 in the @xmath32th row of @xmath114 when @xmath163 and @xmath164 . if @xmath165 , then the @xmath159th element is @xmath166 .
the @xmath162th element is @xmath167 if @xmath168 .
note that @xmath114 is a square matrix of order @xmath9 . by inoue @xcite
, it is proved that @xmath114 coincides with the alexander matrix @xmath169 with @xmath20 replaced by @xmath170 , where @xmath169 is calculated from the wirtinger presentation of the fundamental group @xmath171 by fox s derivative .
see @xcite for details of fox s derivative .
the alexander polynomial @xmath30 is the greatest common divisor of all the minor determinant of order @xmath172 of the alexander matrix @xmath169 .
hence , if @xmath173 is a multiple of an integer @xmath1 larger than or equal to @xmath7 , then all the minor determinant of order @xmath172 of @xmath114 is equivalent to @xmath174 modulo @xmath1 , and hence the system of equations @xmath175 with the vector @xmath176 of variables has at least @xmath177 solutions .
this implies that the coloring number is larger than or equal to @xmath177 , which shows that @xmath3 admits a non - trivial @xmath42-coloring .
suppose that @xmath1 is an odd prime natural number , and @xmath3 is @xmath42-colorable .
then @xmath15 is a field . since the system @xmath175 has more than @xmath1 solutions , the rank of the solution space is @xmath7 or larger .
hence the rank of @xmath114 is smaller than or equal to @xmath178 , and all the minor determinant of order @xmath172 of @xmath114 is equivalent to @xmath174 modulo @xmath1 .
thus we have @xmath179 ( mod @xmath1 ) .
a diagram @xmath6 of @xmath3 has a non - trivial @xmath180-coloring , say @xmath181 .
define @xmath182 by @xmath183 , where @xmath184 means a representative element of @xmath185 .
we easily see that the map @xmath186 is a non - trivial @xmath187-coloring of @xmath6 .
thus , @xmath3 is @xmath187-colorable .
the proof proceeds by induction on the number of prime factors of @xmath1 .
let @xmath6 be a diagram of a knot , and @xmath16 a non - trivial @xmath17-coloring of @xmath6 .
we take an arbitrary prime divisor of @xmath1 , say @xmath37 .
let @xmath188 \in { \mathbb z}_p$ ] be the natural projection map with @xmath189 \equiv z$ ] ( mod @xmath37 ) .
if the composition @xmath190 is not a constant map , then it gives a non - trivial @xmath191-coloring . in this case , @xmath192 since any linear @xmath7-coloroing is trivial for any knot .
we consider the case where @xmath190 is a constant map .
set @xmath193 for all strands @xmath194 , where @xmath9 is a constant in @xmath195 .
then @xmath196 is a multiple of @xmath37 for every @xmath194 , and @xmath197 gives a non - trivial coloring of @xmath6 by @xmath198 .
hence there is an odd prime natural number @xmath199 such that @xmath199 is a divisor of @xmath200 and that @xmath6 is linear @xmath199-colorable with @xmath201 by the assumption of induction .
this completes the proof .
if @xmath3 is @xmath17-colorable , then @xmath3 is @xmath191-colorable for some odd prime divisor @xmath37 of @xmath1 by remark [ rem : colorable ] ( c ) .
hence @xmath202 ( mod @xmath37 ) by remark [ rem : colorable ] ( a ) .
if @xmath202 ( mod @xmath37 ) for some odd prime divisor @xmath37 of @xmath1 , then remark [ rem : colorable ] ( a ) assures that @xmath3 is @xmath203-colorable . hence @xmath3 is @xmath180-colorable by remark [ rem : colorable ] ( b ) .
let @xmath34 be the alexander polynomial of @xmath3 .
by the assumption that @xmath3 is @xmath42-colorable and by remark [ rem : colorable ] ( a ) , we have @xmath202 ( mod @xmath1 ) . on the other hand , @xmath204 .
since @xmath149 for any @xmath148 , we have @xmath205 this implies @xmath206 . hence @xmath3 is @xmath43-colorable by remark [ rem : colorable ] ( a ) , and @xmath46-colorable by lemma [ lem : affinecyclicquandle ] .
let @xmath3 be a knot , and @xmath207 the alexander polynomial of @xmath3 .
let @xmath11 be a positive integer with ( i ) @xmath51 and ( ii ) @xmath52 .
note that @xmath208 because @xmath151 is odd and non - zero . we show that @xmath209 is an odd integer larger than @xmath210 , which is larger than @xmath7 , and that @xmath209 is coprime with @xmath11 .
since @xmath211 ( mod @xmath11 ) , we have @xmath212 by euclidean method of mutual division and the condition ( i ) .
next we show that @xmath213 is odd . when @xmath11 is odd , @xmath214 ( mod @xmath7 ) for any integer @xmath32 with @xmath215 .
hence we have @xmath216 ( mod @xmath7 ) .
this and @xmath217 ( mod @xmath7 ) imply that @xmath213 is odd .
when @xmath11 is even , @xmath218 ( mod @xmath7 ) .
note that @xmath219 is odd since gcd@xmath220 .
we also show that @xmath221 . by the condition ( ii )
, we have @xmath222 @xmath223 @xmath224 ( iii ) .
recall that @xmath225 for the alexander polynomial .
we consider the case where @xmath226 and @xmath219 are negative .
( a similar argument will do for the other case where @xmath226 and @xmath219 are positive . )
then , we have @xmath227 @xmath228 @xmath229 @xmath230 @xmath231(iv ) we show that the equality @xmath232 does not hold . if equality holds on ( iii ) , then @xmath233 holds for all @xmath32 with @xmath234 , and hence there is an positive integer @xmath54 with @xmath235 .
if the equality holds at the second inequality of ( iv ) , then @xmath236 holds for any @xmath32 with @xmath234 . if the equality holds at the third inequality of ( iv ) , then @xmath237
thus the condition @xmath232 leads to @xmath238 , which contradict @xmath239 .
put @xmath50 . then , we have proved that @xmath1 is an odd integer with @xmath13 and @xmath240 .
hence @xmath3 is @xmath241-colorable by remark [ rem : colorable ] ( a ) .
if a knot @xmath3 is linear @xmath1-colorable , then it is @xmath42-colorable for some integer @xmath11 with @xmath242 by lemma [ lem : affinecyclicquandle ] and remark [ remark : nelson ] , and hence by proposition [ lem : fundamentalproperty](1 ) , there exists an odd prime divisor @xmath37 of @xmath1 such that @xmath202 ( mod @xmath37 ) .
if there is a multiple of @xmath37 in @xmath243 for some prime divisor @xmath37 of @xmath1 , then @xmath3 is linear @xmath1-colorable . if not
, then @xmath3 is not linear @xmath1-colorable .
practically , we do not need to calculate all @xmath213 .
for example , if @xmath244 ( mod @xmath245 ) , then @xmath246 ( mod @xmath245 ) by proposition [ lem : fundamentalproperty ] ( 2 ) and @xmath247 ( mod @xmath245 ) . note that @xmath248 and @xmath249 are not isomorphic as quandles .
we perform the procedure described in the proof of proposition [ lem : fundamentalproperty](4 ) for @xmath250 .
this sequence of procedures terminates when we find an integer @xmath1 such that the knot is linear @xmath1-colorable .
such an integer @xmath1 exists and is bounded explicitly by proposition [ lem : fundamentalproperty](3 ) .
the alexander quandle @xmath251/(t^2+t+1)$ ] has four elements .
it is isomorphic to the quandle consisting of @xmath252 degree rotations of a regular tetrahedron , and hence the quandle is called the tetrahedron quandle and we denote it by @xmath253 .
a knot @xmath3 is colorable with the tetrahedron quandle if and only if the alexander polynomial @xmath30 of @xmath3 is equivalent to @xmath174 in @xmath254/(t^2+t+1)$ ] , see @xcite . it is known that indecomposable quandles of order at least @xmath7 and at most @xmath245 are @xmath255 , @xmath253 , @xmath256 , @xmath248 , @xmath249 , there are exactly two indecomposable quandles of order @xmath257 , say @xmath258 and @xmath259 , and every indecomposable quandle of order @xmath260 is @xmath261 , @xmath262 , @xmath263 , @xmath264 , or @xmath265 , refer to @xcite .
the following are the quandle operation matrices of @xmath258 and @xmath259 : the quandle @xmath258 is the extension of the quandle @xmath255 , that is , there is a surjective quandle homomorphism @xmath270 such that for any element of @xmath255 , the cardinality of the inverse image by @xmath111 is constant .
for example , define @xmath111 by @xmath271 , @xmath272 , @xmath273 , @xmath274 , @xmath275 and @xmath276 . the inverse image of each element of @xmath255 forms a trivial subquandle with two elements .
for any knot diagram @xmath6 and any non - trivial coloring @xmath277 of @xmath6 by @xmath258 , put the composition @xmath278 .
it is a quandle coloring of @xmath6 by @xmath255 .
we show that the coloring is non - trivial : assume that the coloring @xmath278 is trivial and denote by @xmath279 the element in @xmath280 .
then we have @xmath281 , which implies that the coloring @xmath277 is essentially the quandle coloring by the subquandle @xmath282 forms .
however it leads to a contradiction since the subquandle is a trivial quandle , and hence @xmath278 is a non - trivial coloring .
therefore any @xmath258-colorable knot is @xmath4-colorable .
[ lem : twist1 ] let @xmath1 be a prime natural number , @xmath11 a natural number with gcd@xmath283 and gcd@xmath284 and @xmath9 a natural number with @xmath285 .
a twist knot with crossing number @xmath9 is @xmath42-colorable if and only if either @xmath286 for some non - negative integers @xmath54 and @xmath37 with @xmath287 ( mod @xmath1 ) and @xmath288 , or @xmath289 for some non - negative integers @xmath54 and @xmath37 with @xmath290 ( mod @xmath1 ) and @xmath291 , where @xmath292 is the element of @xmath15 with @xmath293 . even when @xmath1 is not prime , the @xmath294if part " is true .
let @xmath3 be the twist knot with the crossing number @xmath9 . when @xmath9 is even , @xmath295 for some non - negative integers @xmath54 and @xmath37 with @xmath296 .
an easy calculation shows that the alexander polynomial of @xmath3 is given by the formula @xmath297 ( i ) . substituting @xmath298 for @xmath9 , and @xmath170 for @xmath20 in the polynomial function ( i )
, we have @xmath299 @xmath300 ( mod @xmath1 ) .
hence @xmath202 ( mod @xmath1 ) if and only if @xmath287 ( mod @xmath1 ) .
then the desired conclusion follows from remark [ rem : colorable ] ( a ) .
we consider the case where @xmath9 is odd .
then @xmath289 for some non - negative integers @xmath54 and @xmath37 with @xmath296 .
the alexander polynomial is given by @xmath301 .
then @xmath302 @xmath303 ( mod @xmath1 ) . hence @xmath202 ( mod @xmath1 )
if and only if @xmath290 ( mod @xmath1 ) .
[ lem : twist2 ] let @xmath9 be a natural number with @xmath285 . a twist knot with crossing number
@xmath9 is colorable with the tetrahedron quandle if and only if @xmath304 or @xmath305 for some natural number @xmath54 .
when @xmath9 is even , there is a natural number @xmath54 and an integer @xmath37 with @xmath306 or @xmath0 such that @xmath307 .
then @xmath308 in @xmath254 $ ] , which is equivalent to @xmath174 in @xmath254/(t^2+t+1)$ ] if and only if @xmath306 . when @xmath9 is odd , @xmath309 for integers @xmath54 , @xmath37 with @xmath310 and @xmath306 or @xmath0 .
then @xmath311 in @xmath254 $ ] , which is equivalent to @xmath174 in @xmath254/(t^2+t+1)$ ] if and only if @xmath312 . for a knot @xmath3
, we have @xmath313 if and only if it is quandle colorable by @xmath66 , @xmath67 if and only if it is quandle colorable by @xmath253 , but not by @xmath66 , @xmath71 if and only if it is quandle colorable by @xmath78 or @xmath79 , but not by @xmath66 nor @xmath253 , @xmath80 if and only if it is quandle colorable by @xmath90 , @xmath91 or @xmath92 , but not by @xmath66 , @xmath253 , @xmath78 nor @xmath79 , and @xmath93 , otherwise . note that since a knot is @xmath249-colorable ( resp .
@xmath264-colorable or @xmath265-colorable ) if and only if it is @xmath248-colorable ( resp .
@xmath262-colorable or @xmath263-colorable ) by proposition [ lem : fundamentalproperty ] ( 2 ) , we can omit to consider @xmath249-colorings , @xmath264-colorings and @xmath265-colorings . by lemma [ coloringbyqs6 ]
, we can also omit to consider qs6-colorings and qs6-colorings . for twist knots ,
we consider each case in more detail .
we consider only the case where the crossing number @xmath9 is even .
( a similar calculation will do for the case where @xmath9 is odd , and we omit it . )
then the twist knot with crossing number @xmath9 is @xmath66-colorable if and only if @xmath314 for some positive integer @xmath315 , @xmath78-colorable if and only if @xmath316 for some non - negative integer @xmath317 , @xmath79-colorable if and only if @xmath318 for some non - negative integer @xmath319 , @xmath90-colorable if and only if @xmath320 for some non - negative integer @xmath321 , @xmath91-colorable if and only if @xmath322 for some positive integer @xmath323 , @xmath92-colorable if and only if @xmath324 for some non - negative integer @xmath325 by lemma [ lem : twist1 ] .
the knot is colorable with the tetrahedron quandle if and only if @xmath326 for some positive integer @xmath54 by lemma [ lem : twist2 ] . when @xmath3 is @xmath253-colorable and simultaneously @xmath327-colorable , @xmath328 .
this implies that @xmath329 for some positive integer @xmath330 .
thus @xmath253 is a quandle which gives the minimal quandle order if and only if @xmath331 for @xmath332 and some non - negative integer @xmath70 . when @xmath3 is @xmath333-colorable and simultaneously @xmath327-colorable , @xmath334 .
this implies that @xmath335 ( mod @xmath4 ) .
hence @xmath336 for some non - negative integer @xmath330 .
when @xmath3 is @xmath333-colorable and simultaneously colorable with the tetrahedron quandle , @xmath337 .
this implies that @xmath338 ( mod @xmath7 ) .
hence @xmath339 for some non - negative integer @xmath340 .
thus @xmath78 is a quandle which gives the minimal quandle order if and only if @xmath341 for @xmath342 and some non - negative integer @xmath70 . when @xmath3 is @xmath343-colorable and simultaneously @xmath327-colorable , @xmath344 .
this implies that @xmath345 ( mod @xmath4 ) .
hence @xmath346 for some non - negative integer @xmath330 . when @xmath3 is @xmath343-colorable and simultaneously colorable with the tetrahedron quandle , @xmath347 .
this implies that @xmath348 ( mod @xmath7 ) .
hence @xmath349 for some non - negative integer @xmath340 .
thus @xmath79 is a quandle which gives the minimal quandle order if and only if @xmath350 for @xmath351 and some non - negative integer @xmath70 . when @xmath3 is @xmath352-colorable and simultaneously @xmath327-colorable , @xmath353 .
this implies that @xmath354 ( mod @xmath4 ) .
hence @xmath355 for some non - negative integer @xmath330 . when @xmath3 is @xmath352-colorable and simultaneously colorable with the tetrahedron quandle , @xmath356 .
this implies that @xmath357 ( mod @xmath7 ) .
hence @xmath358 for some non - negative integer @xmath340 . when @xmath3 is @xmath352-colorable and simultaneously @xmath333-colorable , @xmath359 .
this implies that @xmath360 ( mod @xmath245 ) .
hence @xmath361 for some non - negative integer @xmath362 .
when @xmath3 is @xmath352-colorable and simultaneously @xmath343-colorable , @xmath363 .
this implies that @xmath364 ( mod @xmath245 ) .
hence @xmath365 for some non - negative integer @xmath366 .
thus @xmath90 is a quandle which gives the minimal quandle order if and only if @xmath367 for @xmath368 and some non - negative integer @xmath70 . when @xmath3 is @xmath369-colorable and simultaneously @xmath327-colorable , @xmath370 .
this implies that @xmath371 ( mod @xmath4 ) .
hence @xmath372 for some positive integer @xmath330 . when @xmath3 is @xmath369-colorable and simultaneously colorable with the tetrahedron quandle , @xmath373 .
this implies that @xmath374 ( mod @xmath7 ) .
hence @xmath375 for some positive integer @xmath340 . when @xmath3 is @xmath369-colorable and simultaneously @xmath333-colorable , @xmath376 .
this implies that @xmath377 ( mod @xmath245 ) .
hence @xmath378 for some non - negative integer @xmath362 .
when @xmath3 is @xmath369-colorable and simultaneously @xmath343-colorable , @xmath379 .
this implies that @xmath380 ( mod @xmath245 ) .
hence @xmath381 for some non - negative integer @xmath366 .
thus @xmath91 is a quandle which gives the minimal quandle order if and only if @xmath382 for @xmath383 and some non - negative integer @xmath70 . when @xmath3 is @xmath384-colorable and simultaneously @xmath327-colorable , @xmath385 .
this implies that @xmath386 ( mod @xmath4 ) .
hence @xmath387 for some non - negative integer @xmath330 . when @xmath3 is @xmath384-colorable and simultaneously colorable with the tetrahedron quandle , @xmath388 .
this implies that @xmath389 ( mod @xmath7 ) .
hence @xmath390 for some non - negative integer @xmath340 .
when @xmath3 is @xmath384-colorable and simultaneously @xmath333-colorable , @xmath391 .
this implies that @xmath392 ( mod @xmath245 ) .
hence @xmath393 for some non - negative integer @xmath362 . when @xmath3 is @xmath384-colorable and simultaneously @xmath343-colorable , @xmath394 .
this implies that @xmath395 ( mod @xmath245 ) .
hence @xmath396 for some non - negative integer @xmath366 .
thus @xmath92 is a quandle which gives the minimal quandle order if and only if @xmath397 for @xmath398 and some non - negative integer @xmath70 .
next we consider the case where @xmath9 is odd .
then the twist knot with crossing number @xmath9 is @xmath66-colorable if and only if @xmath399 for some non - negative integer @xmath315 , @xmath78-colorable if and only if @xmath400 for some non - negative integer @xmath317 , @xmath79-colorable if and only if @xmath401 for some non - negative integer @xmath319 , @xmath90-colorable if and only if @xmath402 for some non - negative integer @xmath321 , @xmath91-colorable if and only if @xmath403 for some positive integer @xmath323 , @xmath92-colorable if and only if @xmath404 for some non - negative integer @xmath325 by lemma [ lem : twist1 ] .
the knot is colorable with the tetrahedron quandle if and only if @xmath304 for some positive integter @xmath54 by lemma [ lem : twist2 ] . when @xmath3 is @xmath253-colorable and simultaneously @xmath327-colorable , @xmath405
this implies that @xmath406 for some non - negative integer @xmath330 .
thus @xmath253 is a quandle which gives the minimal quandle order if and only if @xmath331 for @xmath407 and some non - negative integer @xmath70 . when @xmath3 is @xmath333-colorable and simultaneously @xmath327-colorable , @xmath408 .
this implies that @xmath409 ( mod @xmath4 ) .
hence @xmath410 for some non - negative integer @xmath330 . when @xmath3 is @xmath333-colorable and simultaneously colorable with the tetrahedron quandle , @xmath411 .
this implies that @xmath412 ( mod @xmath7 ) .
hence @xmath413 for some non - negative integer @xmath340 .
thus @xmath78 is a quandle which gives the minimal quandle order if and only if @xmath341 for @xmath351 and some non - negative integer @xmath70 . when @xmath3 is @xmath343-colorable and simultaneously @xmath327-colorable , @xmath414 .
this implies that @xmath415 ( mod @xmath4 ) .
hence @xmath416 for some non - negative integer @xmath330 .
when @xmath3 is @xmath343-colorable and simultaneously colorable with the tetrahedron quandle , @xmath417 .
this implies that @xmath418 ( mod @xmath7 ) .
hence @xmath419 for some non - negative integer @xmath340 .
thus @xmath79 is a quandle which gives the minimal quandle order if and only if @xmath350 for @xmath342 and some non - negative integer @xmath70 . when @xmath3 is @xmath352-colorable and simultaneously @xmath327-colorable , @xmath420 .
this implies that @xmath421 ( mod @xmath4 ) .
hence @xmath422 for some non - negative integer @xmath330 . when @xmath3 is @xmath352-colorable and simultaneously colorable with the tetrahedron quandle , @xmath423 .
this implies that @xmath424 ( mod @xmath7 ) .
hence @xmath425 for some non - negative integer @xmath340 . when @xmath3 is @xmath352-colorable and simultaneously @xmath333-colorable , @xmath426 .
this implies that @xmath427 ( mod @xmath245 ) .
hence @xmath428 for some non - negative integer @xmath362 . when @xmath3 is @xmath352-colorable and simultaneously @xmath343-colorable , @xmath429 .
this implies that @xmath430 ( mod @xmath245 ) .
hence @xmath431 for some non - negative integer @xmath366 .
thus @xmath90 is a quandle which gives the minimal quandle order if and only if @xmath367 for @xmath432 and some non - negative integer @xmath70 . when @xmath3 is @xmath369-colorable and simultaneously @xmath327-colorable , @xmath433 .
this implies that @xmath434 ( mod @xmath4 ) .
hence @xmath355 for some non - negative integer @xmath330 .
when @xmath3 is @xmath369-colorable and simultaneously colorable with the tetrahedron quandle , @xmath435 .
this implies that @xmath436 ( mod @xmath7 ) .
hence @xmath375 for some non - negative integer @xmath340 . when @xmath3 is @xmath369-colorable and simultaneously @xmath333-colorable , @xmath437 .
this implies that @xmath438 ( mod @xmath245 ) .
hence @xmath439 for some non - negative integer @xmath362 . when @xmath3 is @xmath369-colorable and simultaneously @xmath343-colorable , @xmath440 .
this implies that @xmath441 ( mod @xmath245 ) .
hence @xmath442 for some non - negative integer @xmath366 .
thus @xmath91 is a quandle which gives the minimal quandle order if and only if @xmath382 for @xmath443 and some non - negative integer @xmath70 . when @xmath3 is @xmath384-colorable and simultaneously @xmath327-colorable , @xmath444
this implies that @xmath445 ( mod @xmath4 ) .
hence @xmath446 for some non - negative integer @xmath330 . when @xmath3 is @xmath384-colorable and simultaneously colorable with the tetrahedron quandle , @xmath447 .
this implies that @xmath448 ( mod @xmath7 ) .
hence @xmath449 for some non - negative integer @xmath340 . when @xmath3 is @xmath384-colorable and simultaneously @xmath333-colorable , @xmath450 .
this implies that @xmath392 ( mod @xmath245 ) .
hence @xmath393 for some non - negative integer @xmath362 . when @xmath3 is @xmath384-colorable and simultaneously @xmath343-colorable , @xmath451 .
this implies that @xmath452 ( mod @xmath245 ) .
hence @xmath453 for some non - negative integer @xmath366 .
thus @xmath92 is a quandle which gives the minimal quandle order if and only if @xmath397 for @xmath454 and some non - negative integer @xmath70 . | if a knot has the alexander polynomial not equvalent to @xmath0 , then it is linear @xmath1-colorable . by means of such a coloring , such a knot is given an upper bound for the minimal quandle order , i.e. , the minimal order of a quandle with which the knot is quandle colorable . for twist knots , we study the minimal quandle orders in detail . |
today s widespread incidences of cyber - attacks ( e.g. , most recently of the us democratic party and at companies such as sony , verizon , yahoo , target , jp morgan , office of personnel management , ashley madison ) , each more audacious than earlier ones , makes them perhaps the primary threat faced by individuals , organizations , and nations alike . consequences and implications of cyber - attacks include monetary losses , threats to critical infrastructure and national security , disruptions to daily life , a potential to cause loss of life and physical property , and data leaks exposing sensitive personal information about users and their activities .
the largely quasi - static and unadaptable nature of existing cyber - defenses makes them vulnerable to rapidly evolving attack mechanisms , and thus engender not just a ` warning deficit ' but also a ` detection deficit ' , i.e. , an increase in the mean time between the time of an attack and its discovery .
it has been well argued that , because news about an organization s compromise sometimes originates _ outside _ the organization , one could use open source indicators ( e.g. , news and social media ) as indicators of a cyber - attack . social media , in particular , turns users into social sensors empowering them to participate in an online ecosystem of event detection for happenings such as disease outbreaks @xcite , civil unrest @xcite , and earthquakes @xcite . while the use of social media can not fully supplant the need for internal telemetry for certain types of attacks ( e.g. , use of network flow data to detect malicious network behavior @xcite ) , analysis of such online media can provide insight into a broader range of cyber - attacks such as data breaches , account hijacking and newer ones as they emerge . at the same time it is non - trivial to harness social media to identify cyber - attacks . *
our objective is to detect a range of different cyber - attacks as early as possible , determine their characteristics ( e.g. , the target , the type of attack ) , in an unsupervised manner . *
prior work ( e.g. , @xcite ) relies on weak supervision techniques which will be unable to capture the dynamically evolving nature of cyber - attacks over time and are also unable to encode characteristics of detected events , as we aim to do here .
our main contributions are : * * a framework for cybersecurity event detection based on online social media . *
our dynamic event trigger expansion ( dete ) approach uses a limited , fixed , set of general seed event triggers and learns to map them to specific event - related expansions and thus provide situational awareness into cyber - events in an unsupervised manner . * * a novel query expansion strategy based on dependency tree patterns . * to model typical reporting structure in how cyber - attacks are described in social media ,
we propose a dynamic event trigger expansion method based on convolution kernels and dependency parses .
the proposed approach also employs a word embedding strategy to capture similarities between event triggers and candidate event reports . * * extensive empirical evaluation for three kinds of cyber - attacks*. we manually catalog ground truth for three event classes distributed denial of service ( ddos ) attacks , data breaches , and account hijacking and demonstrate that our approach consistently identifies and encodes events outperforming existing methods .
the input to our methodology is a collection of time - ordered tweets @xmath0 organized along @xmath1 time slots .
let @xmath2 denote the tweet space corresponding to a subcollection @xmath3 , let @xmath4 denote the target tweet subspace ( in our case , comprising cyber - attack events ) , and let @xmath5 denote the rest of the tweets in the considered tweet space . * typed dependency query * : a * _ typed dependency query _ * is a linguistic structure that characterizes a semantically coherent event related topic .
different from n - grams , terms contained in a * _ typed dependency query _ * share both syntactic and semantic relationships .
mathematically , a * _ typed dependency query _ * is formulated as a tree structure @xmath6 , where node @xmath7 can be either a unigram , user mention , or a hashtag and @xmath8 represents a syntactic relation between two nodes . * seed query * : a * _ seed query _ * is a manually selected typed dependency query targeted for a certain type of event .
for instance , `` hacked account '' can be defined as a potential * _ seed query _ * for an account hijacking event .
* expanded query * : an * _ expanded query _ * is a typed dependency query which is automatically generated by the dete algorithm based on a set of seed queries and a given tweet collection @xmath2 . * _ expanded query _ * and its seed query can be two different descriptions of the same subject . more commonly , an * _ expanded query _ * can be more specific than its seed query .
for instance , `` prime minister dmitry medvedev twitter account hack '' , an expanded query from `` hacked account '' , denotes the message of an account hijacking event related with dmitry medvedev .
* event representation * : an event @xmath9 is defined as @xmath10 , where @xmath11 is the set of event - related expanded queries , @xmath12 denotes when the event happens , and @xmath13 refers to the category of the cyber - attack event ( i.e. , ddos , account hijacking , or data breach ) .
here @xmath11 is a defined as a set because , in general , a cyber - attack event can be presented and retrieved by multiple query templates .
for instance , among online discussion and report about event `` fashola s account , website hacked '' , the query template most used are `` fashola twitter account hack '' , `` fashola n78 m website twitter account hack '' and `` hack account '' . given the above definitions , the major tasks underlying the cyber - attack event detection problem are defined as follows : * task 1 : * * target domain generation : * given a tweet subcollection @xmath2 , * _ target domain generation _ * is the task of identifying the set of target related tweets @xmath14 .
@xmath14 contains critical target related information based on which the expanded query can be mined .
* task 2 : * * expanded query extraction : * given target domain @xmath14 , the task of * _ expanded query extraction _ * is to generate a set of expanded queries @xmath15 which represents the generic concept delivered by @xmath14 . thus set @xmath16 can be used to retrieve event related information from other collection sets .
* task 3 : * * dynamic typed query expansion : * given a small set of seed queries @xmath17 and a twitter collection @xmath2 , the task of * _ dynamic typed query expansion _ * is to iteratively expand @xmath18 and @xmath19 until all the target related messages are included .
in traditional information extraction ( ie ) , a large corpus of text must first be annotated to train extractors for event triggers , defined as main keywords indicating an event occurrence @xcite . however , in our scenario using online social media , a manually annotated label set is impractical due to the huge volume of online media and the generally noisy characteristics of the text . in this section ,
we propose a novel method to automatically mine query templates over which the event tracking is performed . in this subsection
, we propose the method of target domain generation , which serves as the source of social indicators for the detection of ongoing cyber - attack events . given a query and a collection of tweets @xmath2 , the typical way to retrieve query - related documentation is based on a bag of words model @xcite which comes with its attendant disadvantages . consider the following two tweets : `` has riseups servers been compromised or * data leaked * ? '' and `` @o2 you completely screwed me over ! my phones back on , still * leaking data * and you are so unhelpful # cancellingcontract # bye '' . though the important indicator `` leak data '' for data breach attack is involved in both tweets , the second tweet is complaining about a phone carrier and is unexpected noise in our case . to address this problem ,
syntactically bound information and semantic similarity constraints are jointly considered in our proposed method . more specifically , each tweet in @xmath2 is first converted into its dependency tree form .
thus for a given seed query @xmath20 , the target domain @xmath21 can be generated by collecting all tweets which are both syntactically and semantically similar to the seed query @xmath20 .
mathematically , given two dependency trees @xmath20 and @xmath22 , a convolution tree kernel @xcite is adopted to measure the similarity using shared longest common paths : @xmath23 where @xmath24 and @xmath25 are two nodes from two trees @xmath20 and @xmath22 respectively , @xmath26 represents set of positive real numbers , @xmath27 is the indicator function and @xmath28 counts the number of common paths between the two trees which peak at @xmath24 and @xmath25 , which can be calculated by an efficient algorithm proposed by kate et al .
@xcite , as described in algorithm [ alg : cpp ] .
set @xmath29 set @xmath30 set @xmath31 @xmath32 in algorithm [ alg : cpp ] , @xmath33 is the number of common paths between the two trees which originate from @xmath25 and @xmath24 , and can be recursively defined as : @xmath34 where @xmath35 denotes the set of children node . in both algorithm [ alg : cpp ] and equation [ equ : cdp ] , we use the semantic similarity operator @xmath36 , introduced to consider the semantic similarity of tree structre .
this semantic similarity is computed by considering cosine similarity of word embeddings vector generated from the word2vec algorithm .
this model considers the common paths which are linguistically meaningful , which reduces the noise introduced by coincidentally matched word chains .
in addition , long - range dependencies between words , which decreases the performance , are avoided because functionally related words are always directly linked in a dependency tree . in this subsection
, we propose a way to dynamically mine an expanded query given a small collect of seed query , as shown in table [ tab : seedquery ] . by providing a small set of seed queries ( unigrams ) , zhao et al .
@xcite proposed a dynamic query expansion ( dqe ) method which is able to iteratively expand the seed query set from currently selected target tweet subspace until convergence . looking beyond the simple unigrams based expansion , by introducing dependency - based tree structure extraction
, we build a dynamic expanded query generation model for the cyber - attack detection task . .seed
queries for cyber - attack events . [ cols="^,<",options="header " , ] 0.40 0.58 we comprehensively show in fig .
[ fig : dqestreamgraph2014 ] and fig .
[ fig : dqestreamgraph2016 ] , the wide range of events that our system is able to detect .
notice , the clear burst in twitter activity that our query expansion algorithm is able to pick . through the following case studies we highlight some of the interesting cases for each of three cyber attack types , that our system detected .
* targeted ddos attacks on sony and dyn * : in late , november 2014 , a hacker group calling itself `` the guardians of peace '' hacked their way into sony pictures , leaving the sony network crippled for days , allegedly perpertrated by north korea .
we capture 12 separate events of ddos attacks including four in last week of august 2014 , starting with the first on august 24th . further in 2015 , more ensuing attacks are captured one highlighted by the data breach of their movie production house , on december 12th and then a massively crippling targeted , ddos attack on their playstation network in late december , 2015 .
another noteworthy case of ddos attacks in 2016 , is the multiple distributed denial - of - service attack on dns provider `` dyn '' from october 21st through 31st in 2016 , that almost caused an worldwide internet outage .
our system detects generates several query expansions , shown in fig .
[ fig : dyndns ] which clearly characterizes the nature of these ddos attacks where the hackers turned a large number of internet - connected devices around the world in to botnets executing a distributed attack . *
ashley madison website data breach : * in july 2015 , a group calling itself `` the impact team '' stole the user data of ashley madison , an adult dating website billed as enabling extramarital affairs .
the hackers stole the website all customer data and threatened to release the personally identifying information if the site was not immediately shut down . on 18 and 20
august , the group leaked more than 25 gigabytes of company data , including user details .
we are able to detect this data breach on july 20 , 2015 .
the word clouds in fig .
[ fig : ashleymadison ] clearly show how our method iteratively expands from the seed queries to the expanded queries in the last , iteration 3 capturing a very rich semantic aspect of the breach .
after the initial burst as seen in the figure , we also see a second corresponding burst a month later , on august 20 when the user data is released anot now the top query expansion captured characterized by the mentions of user data leak of the same website .
* twitter account hijackings * : we were also able to detect with very high date accuracy , several high profile cases of account hijackings of social media accounts of known personalities and government institutions including the twitter account for u.s .
central command which was hacked by isis sympathizers on january 12 , 2015 .
we show in fig .
[ fig : uscentcom ] that our method not only identifies the victim ( `` central command twitter account hack '' ) but also the actor who perpetrated the hacking ( `` isis hack twitter account '' ) .
[ sub : case_studies ]
* cyberattack detection and characterization . * detecting and characterizing cyber attacks is highly challenging due to the constant - involving nature of cyber criminals .
recent proposals cover a large range of different methods , and table [ tab : related_work ] lists representative works in this space .
earlier work primarily focuses on mining network traffic data for intrusion detection .
specific techniques range from classifying malicious network flows @xcite to anomaly detection in graphs to detect malicious servers and connections @xcite .
more recently , researchers seek to move ahead to predict cyber attacks before they happened for early notifications @xcite . for example , liu et al . leverage various network data associated to an organization to look for indicators of attacks @xcite . by extracting signals from mis - configured dns and bgp networks as well as spam and phishing activities ,
they build classifiers to predict if an organization is ( or will be ) under attack . similarly , soska et al .
apply supervised classifiers to network traffic data to detect vulnerable websites , and predict their chances of turning malicious in the future @xcite . in recent years
, online media such as blogs and social networks become another promising data source of security intelligence @xcite .
most existing work focuses on technology blogs and tweets from _ security professionals _ to extract useful information @xcite .
for example , liao et al . builds text mining tools to extract key attack identifiers ( ip , md5 hashes ) from security tech blogs @xcite .
sabottke et al .
leverage twitter data to estimate the level of interest in existing cve vulnerabilities , and predict their chance of being exploited in practice @xcite .
our work differs from existing literature since we focus on crowdsourced data from the much broader user populations who are likely the _ victims _ of security attacks .
the most related work to ours is @xcite which uses weakly supervised learning to detect security related tweets .
however , this technique is unable to capture the dynamically evolving nature of attacks and is unable to encode characteristics of detected events .
* event extraction and forecasting on twitter . * another body of related work focuses on twitter to extract various events such as trending news @xcite , natural disasters @xcite , criminal incidents @xcite and population migrations @xcite .
common event extraction methods include simple keyword matching and clustering , and topic modeling with temporal and geolocation constrains @xcite .
event forecasting , on the other hand , aims to predict future evens based on early signals extracted from tweets .
example applications include detecting activity planning @xcite and forecasting future events such as civil unrest @xcite and upcoming threats to national airports @xcite . in our work ,
we follow a similar intuition to detect signals for major security attacks .
the key novelty in our approach , different from these works , is the need for a typed query expansion strategy that provides both focused results and aids in extracting key indicators underlying the cyber - attack .
cp1.2cmp1.9cmp1.70cmp1.70cmcccp1.85 cm & & & [ 0.65em]goal & [ 0.65em]data + & unsupervised & keyword expansion & information extraction & characterize event & type & detection & & + @xcite & & & & & & & cyberattacks & network data + @xcite & & & & & & & cyberattacks & twitter + @xcite & & & & & & & cyberattacks & wine + @xcite & & & & & & & malware & papers + @xcite & & & & & & & malware & wine + @xcite & & & & & & & vulnerability & twitter + @xcite & & & & & & & intrusion & network data + @xcite & & & & & & & intrusion & network data + @xcite & & & & & & & intrusion & network data + @xcite & & & & & & & insider & access log + @xcite & & & & & & & ioc & tech blogs + ours & & & & & & & cyberattacks & twitter +
we have demonstrated an unsupervised approach to extract and encode cyber - attacks reported and discussed in social media .
we have motivated the need for a careful template - driven query expansion strategy , and how the use of dependency parse trees and word embeddings supports event extraction . given
the widespread prevalence of cyber - attacks , tools such as presented here are crucial to providing situational awareness on an ongoing basis .
future work is aimed at broadening the class of attacks that the system is geared to as well as at modeling sequential dependencies in cyber - attacks .
this will aid in capturing characteristics such as increased prevalence of attacks on specific institutions or countries during particular time periods .
r. j. kate .
a dependency - based word subsequence kernel . in _ proceedings of the conference on empirical methods in natural language processing _ , pages 400409 .
association for computational linguistics , 2008 .
q. mei and c. zhai .
discovering evolutionary theme patterns from text : an exploration of temporal text mining . in _ proceedings of the eleventh acm sigkdd international conference on knowledge discovery in data mining _ ,
pages 198207 .
acm , 2005 .
m. ovelgonne , t. dumitras , b. a. prakash , v. s. subrahmanian , and b. wang . understanding the relationship between human behavior and susceptibility to cyber - attacks : a data - driven approach . in _ proc .
tist16 _ , 2016 .
t. sakaki , m. okazaki , and y. matsuo .
earthquake shakes twitter users : real - time event detection by social sensors . in _ proceedings of the 19th international conference on world wide web _ ,
pages 851860 .
acm , 2010 . | social media is often viewed as a sensor into various societal events such as disease outbreaks , protests , and elections .
we describe the use of social media as a crowdsourced sensor to gain insight into ongoing cyber - attacks .
our approach detects a broad range of cyber - attacks ( e.g. , distributed denial of service ( ddos ) attacks , data breaches , and account hijacking ) in an unsupervised manner using just a limited fixed set of seed event triggers .
a new query expansion strategy based on convolutional kernels and dependency parses helps model reporting structure and aids in identifying key event characteristics . through a large - scale analysis over twitter
, we demonstrate that our approach consistently identifies and encodes events , outperforming existing methods . |
in this paper we consider random schrdinger operators @xmath0 in @xmath1 of the form @xmath2 where @xmath3 is a sequence of i.i.d .
( independent , identically distributed ) variables having a common density @xmath4 ( i.e. @xmath5 ) , which is continuous and has support in the finite interval @xmath6 $ ] . in what follows
we always suppose that the single - site potential @xmath7 is piecewise continuous with @xmath8 $ ] .
moreover , we require that @xmath9
. the spectral properties of ( [ 1.1 ] ) were studied in detail in @xcite .
the results are most complete for the case when @xmath7 is the point interaction ( see @xcite ) .
the main idea of our approach is to approximate the operator @xmath0 ( [ 1.1 ] ) by means of the sequence @xmath10 with unchanged @xmath11 , which converges to @xmath0 in the strong resolvent sense .
this approximation gives the opportunity to use scattering theory in order to study the spectral properties of the limiting operator ( [ 1.1 ] ) .
in fact , we show how to recover the spectral characteristics of @xmath0 from the limiting behaviour of the spectral characteristics of @xmath12 in the large support " limit @xmath13 .
one of the important ingredients of our approach is the lifshitz - krein spectral shift function ( see @xcite for a review ) .
recently it has received renewed interest due to its applications to different problems in the theory of schrdinger operators @xcite . in the context of our approach
the spectral shift function naturally replaces the eigenvalue counting function , which is usually used to construct the density of states for the operator ( [ 1.1 ] ) .
the celebrated birman - krein theorem @xcite relates the spectral shift function to scattering theory .
in fact , up to a factor @xmath14 it may be identified with the scattering phase for the pair ( @xmath12 , @xmath11 ) , i.e. @xmath15 when @xmath16 , @xmath17 here @xmath18 and @xmath19 have the meaning of transmission and reflection coefficients , respectively , such that @xmath20 . for @xmath21 @xmath22
equals minus the counting function of @xmath12 .
these two properties of the spectral shift function , namely its relation to scattering theory and its replacement of the counting function in the presence of an absolutely continuous spectrum convinced the authors already some time ago that the spectral shift function could be applied to the theory of random schrdinger operators .
in fact , our previous articles @xcite were in part preparatory investigations aiming at such an application .
in @xcite we proved convexity and subadditivity properties of the spectral shift function with respect to the potential and the coupling constant , respectively .
such properties often show up when considering thermodynamic limits in statistical mechanics . in @xcite
we studied cluster properties when the potential is a sum of two terms and the center of one is moved to infinity .
again such properties play an important role in statistical mechanics as well as in quantum field theory .
possible applications of the theory of the spectral shift function to random schrdinger operators have also been envisaged by simon @xcite .
some other applications of the scattering theory in one dimension to the study of spectral properties of schrdinger operators with periodic or random potentials can be found in @xcite and @xcite respectively .
below we prove some new inequalities for the spectral shift function , which reflect its `` additivity '' properties with respect to the potential being the sum of two terms with disjoint supports .
these inequalities are closely related to the aktosun factorization formula ( [ aktosun.form ] ) @xcite ( see also @xcite ) . combined with the superadditive ergodic theorem they will allow us to prove the almost sure existence of the limit @xmath23 which we call the spectral shift density .
we prove the equality @xmath24 , where @xmath25 and @xmath26^{1/2}$ ] are the integrated density of states of the hamiltonians @xmath0 and @xmath11 , respectively .
also we reconsider the aktosun factorization formula and show that it is a direct consequence of the propagator property of the fundamental ( or transfer ) matrix of the schrdinger equation .
another very important quantity associated with the hamiltonian ( [ 1.1 ] ) is the lyapunov exponent @xmath27 .
in particular , according to the ishii - pastur - kotani theorem @xcite the set @xmath28 is the essential support of the absolute continuous part of the spectral measure for @xmath0 ( [ 1.1 ] )
. we will establish that for @xmath16 the function @xmath29 can be interpreted as the density of the logarithm of the transmission amplitude @xmath30 , @xmath31 the connection between the lyapunov exponent and the transmission coefficient @xmath32 was recognized long ago @xcite , though a rigorous analysis was still absent .
it is well known ( see e.g. @xcite ) that the function @xmath33 can be analytically continued in the complex half - plane @xmath34 as a nevanlinna function ( i.e. an analytic function which maps the upper complex half plane into itself ) .
we will recover this property of @xmath35 directly from the analytic properties of the transmission amplitudes @xmath36 .
moreover , as a direct consequence of these properties we give a new proof of the thouless formula @xmath37 where @xmath38^{1/2}$ ] is the lyapunov exponent for @xmath11 . also we prove a new representation for the lyapunov exponent for positive energies , @xmath39 where @xmath40 and @xmath41 , @xmath42 , @xmath43 are elements of the s - matrix at energy @xmath44 for the pair of hamiltonians ( @xmath45 , @xmath11 ) .
this representation allows us to apply the theory of random matrices to prove that @xmath46 for a.e .
@xmath16 almost surely , which in turn by ishii - pastur - kotani theorem implies that the spectrum of @xmath47 has no absolute continuous component in @xmath48 .
we give also an explicit description of the set of special energies where @xmath49 . to our best knowledge
this set was known explicitly only for the two particular cases when the single - site potential @xmath7 is a @xmath50-potential or the characteristic function of the interval @xmath51 $ ] ( see @xcite ) .
we express the density of states @xmath25 for positive energies in terms of the product of matrices ( [ l.tilde ] ) , @xmath52 with @xmath53 , @xmath54 and @xmath55 being the inner product in @xmath56 .
this representation is similar to the definition of the density of states through the rotation number of the fundamental solution of the schrdinger equation @xcite .
formulae ( [ xi ] ) , ( [ 1.3 ] ) , ( [ xxx ] ) , and ( [ 1.6 ] ) also provide very simple and efficient numerical algorithms to compute the density of states and the lyapunov exponent of disordered systems in one dimension . in this context
we also remark that the representations ( [ xi ] ) and ( [ 1.6 ] ) for finite @xmath57 and @xmath16 give smooth approximations to the spectral shift density and density of states , respectively .
this contrasts with the usual procedure ( see e.g. @xcite ) , where the density of states is approximated by step - like functions .
as an example in section 4 we have calculated the spectral shift density for the deterministic kronig - penney model .
our approach is also applicable to deterministic hamiltonians of the form @xmath58 with a potential @xmath59 , which is supposed to be uniformly in @xmath60 and for technical reasons also bounded , but without any assumption on its decay at infinity .
let @xmath61 be a sequence of real numbers such that @xmath62 as @xmath63 and @xmath64 for all @xmath65 .
in this case we approximate @xmath47 through @xmath66 with @xmath67 being the characteristic function of the interval @xmath68 $ ] .
again we show that the relations analogous to ( [ xi ] ) , ( [ 1.3 ] ) , ( [ xxx ] ) , ( [ 1.6 ] ) hold .
an extension of ideas developed in the present paper to the case of higher dimensions will be given in @xcite .
* acknowledgements : * we are indebted to j.m .
combes and v. enss for valuable remarks .
we start with a short discussion of some important properties of the spectral shift function @xmath69 for a pair of self - adjoint hamiltonians @xmath58 and @xmath70 on @xmath1 with domain of definition being the sobolev space @xmath71 ( see e.g. @xcite for the definition ) . here
@xmath59 denotes the multiplication operator by the real valued function @xmath72 , which is supposed to satisfy @xmath73 the spectral shift function is defined by the trace formula @xmath74 which is valid for a wide class of functions @xmath75 ( see @xcite ) .
for instance , elements in @xmath76 are in this class . relation ( [ one ] ) defines @xmath77 only up to an additive constant .
we fix this ambiguity by the condition that @xmath78 for all @xmath44 below the spectrum @xmath79 of @xmath47 . with this normalization condition
the function @xmath77 for @xmath21 equals minus the number of the eigenvalues of @xmath58 less than @xmath44 , @xmath80 for @xmath16 the relation between @xmath77 and the scattering matrix @xmath81 is given by the celebrated birman - krein theorem @xcite , @xmath82 here the branch of the logarithm is fixed by the condition @xmath83 as @xmath84 ( see lemma 2.1 below ) . since @xmath81 is continuous for all @xmath16 @xcite
, so is @xmath77 .
the relations ( [ two ] ) and ( [ three ] ) define @xmath77 everywhere , i.e. also for @xmath85 except at the finite number of points of discontinuity at the eigenvalues of @xmath47 and possibly at @xmath86 .
we can redefine @xmath77 by requiring that @xmath87 for all @xmath88 such that @xmath77 becomes right semicontinuous .
we recall that the spectral shift function is monotone with respect to the perturbation , i.e. if @xmath89 and @xmath90 are self - adjoint operators such that @xmath91 in the sense of quadratic forms , then @xmath92 ( see e.g. @xcite ) .
the value of @xmath93 depends on the spectral properties of the point @xmath86 .
we call the point @xmath86 regular iff @xmath94 exists and is bounded at @xmath95 . in the opposite case @xmath86 is called an exceptional point . here
we used the notation @xmath96 and @xmath97 and @xmath98 is the resolvent of the free hamiltonian @xmath11 .
other characterizations of the exceptional case can be found in @xcite .
the levinson theorem for hamiltonians on a line @xcite states that @xmath99 if @xmath86 is a regular point for @xmath47 and @xmath100 if @xmath86 is an exceptional point .
an extension of ( [ funf ] ) for potentials with slower decrease can be found in @xcite .
the scattering matrix @xmath81 for the pair of hamiltonians ( @xmath47 , @xmath11 ) at fixed energy @xmath101 is a 2 by 2 unitary matrix ( see @xcite ) @xmath102 below we will use the fact that due to unitarity the s - matrix can be parameterized by the absolute value of the transmission amplitude @xmath103 and two real valued phases @xmath104 and @xmath105 : @xmath106 here @xmath104 is the scattering phase such that @xmath107 . for reflection
symmetric potentials @xmath108 .
below we will need the following auxiliary results .
* lemma 2.1 _ let @xmath72 satisfy the condition ( [ condition ] ) .
then there is a constant @xmath109 such that @xmath110 independent of @xmath59 and @xmath44 such that @xmath111 ^ 2\right\}\ ] ] for all @xmath16 . _
* _ proof .
first we prove the inequality ( [ 2.1.2 ] ) . by monotonicity of the spectral shift function we have @xmath112 the operator
@xmath113 is trace class for all @xmath114 off the positive real semiaxis ( see ( * ? ? ?
* problem 161 ) ) .
this operator has limiting values @xmath115 for all @xmath16 in the hilbert - schmidt norm .
since the resolvent @xmath116 of @xmath11 is an integral operator with kernel @xmath117 , it follows that @xmath118 has the integral kernel @xmath119 and thus has a rank 2 and is obviously positive semidefinite . _ for positive energies
the spectral shift function can be calculated as follows @xcite , @xmath120 where the branch of the logarithm is fixed by the condition @xmath83 as @xmath84 .
obviously , the operator @xmath115 has no real eigenvalues ( otherwise for some value of the coupling constant @xmath121 the number @xmath122 would be an eigenvalue of @xmath123 , which implies that @xmath124 , the positive singular part of the spectrum of @xmath125 ) .
therefore , estimating the r.h.s .
of ( [ toestim ] ) as in @xcite ( lemmas 2.2 , 2.3 , and proof of theorem 3.1 ) for all @xmath16 we obtain @xmath126,\end{aligned}\ ] ] where the constant @xmath127 is independent of @xmath59 and @xmath44 . here
@xmath128 and @xmath129 denote the trace and hilbert - schmidt norms respectively .
since @xmath130 is positive semidefinite it follows that @xmath131 obviously , the r.h.s . is also a bound for @xmath132
now we estimate @xmath93 . according to ( [ four ] ) and ( [ funf ] ) @xmath133 where @xmath134 is a total number of eigenvalues of @xmath47 . by the well - known bargman - type estimate ( see e.g. @xcite ) @xmath135 we obtain latexmath:[\[\label{3/2 } @xmath86 . since @xmath77 is a nonincreasing function of @xmath21 and @xmath78 for @xmath137 , the estimate ( [ 3/2 ] ) is valid for all @xmath138 .
now let us fix some @xmath139 .
for all @xmath140 $ ] the estimate @xmath141 ^ 2\right\}\ ] ] is valid .
the function @xmath77 is continuous on @xmath142 $ ] and thus by the weierstrass theorem attains its maximum and its minimum , which by ( [ 2.1.2 ] ) and ( [ 3/2 ] ) are finite .
therefore , there exists @xmath143 such that @xmath144 for all @xmath145 $ ] .
this inequality combined with ( [ cv1 ] ) gives the estimate ( [ 2.1.1 ] ) and completes the proof of the lemma .
@xmath146 _
remarks : _ 1 . the first term in the estimate ( [ 2.1.2 ] ) with @xmath147
represents the high - energy asymptotics of the spectral shift function ( see e.g. @xcite ) . below ( lemma 6.1 )
we prove that for bounded potentials @xmath59 with compact support the second term in ( [ 2.1.2 ] ) can be omitted ( with @xmath127 independent of @xmath59 ) .
the estimate ( [ 2.1.1 ] ) for @xmath101 also follows from the results in @xcite . * lemma 2.2 _
let @xmath59 be piecewise continuously differentiable , satisfy ( [ condition ] ) and @xmath148 then for each closed interval @xmath149 @xmath150 where @xmath151 and @xmath152 is the spectral resolution for the operator @xmath47 .
* _ remark : _ a similar formula for the case of operators acting in @xmath153 with @xmath154 and @xmath155 was proved by robert and tamura @xcite and by jensen @xcite . _ proof .
our proof closely follows the ideas of @xcite .
let @xmath156 be the solutions of the integral equations @xmath157 these functions are solutions of the schrdinger equation @xmath158 let us also consider the functions @xmath159 where @xmath30 is the transmission amplitude .
they satisfy the lippmann - schwinger equations which are integral equations of the fredholm type @xcite , @xmath160 with @xmath161 the functions @xmath162 are continuously differentiable with respect to @xmath163 and @xmath101 @xcite .
thus , @xmath164 are continuously differentiable with respect to @xmath165 . from ( [ fredholm ] ) and since @xmath166
it follows that @xmath167 satisfy the integral equation @xmath168 this easily gives @xmath169 where @xmath170 is the resolvent of @xmath171 .
_ it is known @xcite that the scattering matrix ( [ smatrix.def ] ) can be calculated as @xmath172 where @xmath173 @xmath174 is the 2 by 2 unit matrix and @xmath175 denotes hermitian conjugation , such that @xmath176 now we calculate @xmath177 by ( [ dsigma ] ) and by the identity @xmath178 we obtain @xmath179 using the transformation property of the scattering matrix @xcite @xmath180 we get @xmath181 by the birman - krein theorem @xmath182 and therefore @xmath183 for all @xmath16 .
now we use the spectral representation for the spectral decomposition of the operator @xmath171 ( see e.g. @xcite ) , @xmath184 where @xmath185 , @xmath186 and @xmath187 are the eigenvectors of @xmath47 with eigenvalues @xmath188 .
since @xmath189 is trace class @xcite , we have @xmath190 it remains to prove that @xmath191 or equivalently @xmath192 the proof of ( [ trace ] ) is quite elementary .
it suffices to consider the case when @xmath193 contains a single eigenvalue @xmath194 of @xmath47 with eigenfunction @xmath195 such that @xmath196 is the projector onto @xmath195 .
let @xmath197 be the generator of dilations , @xmath198 in @xcite it is shown that @xmath199+h\ ] ] as a bounded operator from @xmath71 to @xmath200 , such that @xmath201 is well defined
. then @xmath202\phi_0)=e_0.\end{aligned}\ ] ] @xmath146 * lemma 2.3 * _ for every @xmath203 with @xmath204 the function @xmath205 is analytic and given by @xmath206 moreover , for all real @xmath207 we have @xmath208 where the integral is understood in the riemann - stieltjes sense . _
_ the analyticity of @xmath205 in the open upper complex half - plane @xmath209 is well - known ( see e.g. @xcite ) .
also @xmath205 is bounded in @xmath209 and @xmath210 in all points of continuity of @xmath77 .
therefore , we can apply the schwarz integral formula for the half - plane ( see e.g. @xcite ) to reconstruct @xmath205 from the limiting values of its imaginary part , @xmath211 with @xmath127 being a real constant .
for @xmath212 we have @xmath213 @xcite . therefore , @xmath127 in ( [ schwarz ] ) must be zero .
@xmath146 below we will make use of the aktosun factorizarion formula @xcite , which we formulate in the following form .
let @xmath72 be some real - valued locally integrable bounded function .
let @xmath214 be a sequence of real numbers such that @xmath215 as @xmath216 and @xmath217 for all @xmath218 .
let @xmath219 be the characteristic function of the interval @xmath220 $ ] .
we denote @xmath221 such that @xmath222 tends to @xmath72 as @xmath223 .
let @xmath224 and @xmath225 denote the hamiltonians with domains of definition being the sobolev space @xmath226 , @xmath227 let @xmath228 and @xmath229 be the corresponding s - matrices , @xmath230 @xmath231 we also consider the matrices @xmath232 and @xmath233 from the unitarity of the scattering matrix ( see e.g. @xcite ) it follows that the matrices @xmath234 and @xmath235 are unimodular .
we note that in faddeev s terminology @xcite the elements of @xmath236 are given by the coefficients @xmath237 .
more precisely , @xmath238 the aktosun factorization formula states that @xmath239 here and below we understand the product @xmath240 in the ordered sense , i.e. @xmath241 the theorem below provides an alternative proof of ( [ aktosun.form ] ) .
actually we show that the factorization property of the matrices @xmath242 is directly related to the propagator property of the fundamental solution of the corresponding schrdinger equation . * theorem 2.4 * _ for arbitrary @xmath16 consider the matrix @xmath243 which solves the initial value problem @xmath244 with @xmath245 the matrices @xmath234 are related to @xmath246 such that @xmath247 _ the factorization formula ( [ aktosun.form ] ) follows immediately from ( [ theor.2.3 ] ) and from the propagator property of @xmath248 , @xmath249 .
_ remarks _ : 1 .
the matrix @xmath246 is related to the fundamental solution @xmath250 of the schrdinger equation @xmath251 which satisfies @xmath252 it is easy to see that @xmath253 where @xmath254 such that @xmath255 \2 .
the formula ( [ theor.2.3 ] ) reduces the problem of the study of @xmath256 to the study of the scattering matrix for the corresponding single - site potential .
note that @xmath257 is not a multiple of the identity operator .
therefore , the hilbert - schmidt norms of @xmath234 and of @xmath256 are not equal .
this will become relevant in section 5 .
the connection between the solutions of ( [ u ] ) and scattering characteristics was noted earlier in @xcite . _ proof . _ to prove the theorem it suffices to consider @xmath72 supported on the interval @xmath258 $ ] and to show that @xmath259 , i.e. the solution of the equation @xmath260 satisfies @xmath261 where @xmath30 , @xmath262 , and @xmath263 correspond to the potential @xmath59 .
consider the solutions @xmath162 of the equation ( [ schroed ] ) , such that @xmath264 @xmath265 we introduce the 2 by 2 matrix @xmath266 where @xmath267 and @xmath268 are given by @xmath269,\\ b_\pm(x , e)=\frac{1}{2}e^{-i\sqrt{e}x}\left[\psi_\pm(x , e)-\frac{i}{\sqrt{e } } \frac{d\psi_\pm(x , e)}{dx } \right].\end{aligned}\ ] ] it is easy to see that @xmath270 satisfies the equation @xmath271 and @xmath272 also we have @xmath273 obviously , @xmath274 also satisfies ( [ w ] ) and equals @xmath174 for @xmath275 .
thus , @xmath276 moreover , @xmath277 since @xmath278 finally note that @xmath246 , defined by ( [ u ] ) , is given as @xmath279 .
in this section we establish a cluster property of the spectral shift function for schrdinger operators in @xmath1 of the form @xmath280 where the @xmath281 are in @xmath282 and have compact supports ( see also @xcite for results concerning cluster properties in the higher dimensional case ) .
we study the behavior of the spectral shift function @xmath283 for the pair ( @xmath284 , @xmath11 ) when @xmath285 is sufficiently large .
more precisely let @xmath286 be such that the intersection of the minimal closed intervals containing @xmath287 and @xmath288 is at most a point .
this implies that @xmath289 a.e . for all @xmath290 .
we will henceforth assume that @xmath290 .
denote @xmath291 with @xmath292 ( @xmath293 ) .
also we set @xmath294 such that @xmath295 . by the translation invariance of the spectral shift function
, we have @xmath296 for all @xmath297 . for brevity in
what follows we will write @xmath298 , @xmath299 , @xmath293 . below we will need the following simple result : * lemma 3.1 _ suppose that @xmath300 , @xmath301 and @xmath302 , @xmath303 are semi - bounded self - adjoint operators in the hilbert spaces @xmath304 and @xmath305 respectively , such that @xmath306 , @xmath293 are trace class . then for a.e .
@xmath88 @xmath307 moreover , @xmath308 a.e . on @xmath309 . _ * _ proof .
let @xmath310 .
for every @xmath311 we have @xmath312 from this it follows that ( [ 0.1 ] ) and ( [ 0.2 ] ) are valid up to an additive constant . from our normalization and from the semiboundedness of the operators involved it follows that this constant equals zero .
@xmath146 _ * theorem 3.2 _ for all @xmath313 @xmath314 _ * _ proof .
let us fix some @xmath315 lying between the supports of @xmath316 and @xmath317 .
more precisely we require that @xmath318 is such that @xmath319 $ ] and @xmath320 if @xmath321 and @xmath322 $ ] and @xmath323 if @xmath324 . without loss of generality we may suppose that @xmath321 .
for every potential @xmath59 satisfying ( [ condition ] ) along with the hamiltonian @xmath47 we consider the self - adjoint schrdinger operators on @xmath1 @xmath325 with dirichlet and neumann boundary conditions at @xmath326 .
in accordance with the decomposition @xmath327 one can write @xmath328 . from krein s formula ( see e.g. @xcite )
it follows that @xmath329 and @xmath330 are rank one perturbations of @xmath47 .
also we have @xmath331 in the sense of quadratic forms .
since @xmath316 and @xmath317 have disjoint compact supports , we have @xmath332 _ from lemma 3.1 it follows that @xmath333 now by the chain rule for the spectral shift function @xcite and from ( [ xi.prove ] ) it follows that @xmath334 on the other hand and again by the chain rule one has @xmath335 and analogously with @xmath89 replaced by @xmath336 . from ( [ xi.1 ] ) and ( [ xi.2 ] ) it follows that @xmath337 such that @xmath338 we note that in the terminology of gesztesy and simon @xcite @xmath339,@xmath340 , @xmath293 , and @xmath341 are the xi - functions for the operators @xmath284 , @xmath342 , @xmath293 , and @xmath11 respectively . from ( [ ineq.nd ] ) and from the fact that the absolute value of the spectral shift function for rank one perturbations is not greater than one
, we have that for all real @xmath44 @xmath343 where @xmath47 stands for one of the operators @xmath284 , @xmath89 , @xmath336 , or @xmath11 . also ( see @xcite )
@xmath344 therefore , from ( [ ineq.long ] ) we obtain @xmath345 using dirichlet boundary conditions and @xmath346 using neumann boundary conditions .
this completes the proof of the theorem .
@xmath146 for our further purposes to be elaborated below theorem 3.2 provides a completely sufficient information .
however , we note that when @xmath101 the bound for @xmath347 can be improved : * theorem 3.3 _ for all @xmath313 and all @xmath101 @xmath348 for all @xmath313 and all @xmath88 .
let us define the functions @xmath349 * corollary 3.4 _ for all @xmath88 the functions ( [ 3.6 ] ) satisfy the inequalities @xmath350 i.e. @xmath351 is a subadditive and @xmath352 is a superadditive function with respect to the potentials @xmath316 and @xmath317 .
_ * _ proof . due to ( [ 3.3.neu ] ) and theorems 3.2 , 3.3 we have @xmath353 adding @xmath354 to both sides of this inequality we arrive at ( [ 3.2.1 ] ) .
similarly , we have @xmath355 subtracting @xmath354 from both sides yields ( [ 3.2.2 ] ) .
@xmath146 _ _ proof of theorem 3.3 .
recall ( see @xcite ) that the scattering matrix at energy @xmath101 for the pair of hamiltonians ( @xmath356 , @xmath11 ) is given by @xmath357 we use the fact that for all @xmath313 and all @xmath16 @xmath358 the formula ( [ 3.5 ] ) has appeared earlier in @xcite and follows from the aktosun factorization formula [ aktosun.form ] @xmath359 and the identity ( see e.g. @xcite ) @xmath360 _ according to ( [ smatrix.def ] ) the reflection amplitudes @xmath361 and @xmath362 can be represented in the form @xmath363 with @xmath364 .
moreover @xmath365 only when @xmath366 .
this in turn can happen only if @xmath86 ( see @xcite ) .
therefore @xmath367 due to the fact that @xmath368 for @xmath16 the absolute value of this expression is strictly bounded by @xmath369 .
@xmath146 since the spectral shift function is right continuous the estimate @xmath370 ) for @xmath86 also follows from the levinson theorem @xcite , theorem 3.1 of @xcite and the results of @xcite .
we include this discussion ( see also @xcite ) to show the different aspects of the problem .
for brevity we set @xmath371 , the number of bound states below @xmath86 .
theorem 3.1 of @xcite states in particular that @xmath372 for all @xmath297 such that @xmath313 if @xmath86 is an exceptional point for at least one of @xmath89 or @xmath90 .
if @xmath86 is a regular point for both @xmath89 and @xmath90 then @xmath373 therefore , by ( [ four ] ) and ( [ funf ] ) we have @xmath374 if @xmath86 is a regular point for both @xmath89 and @xmath90 , and @xmath375 if @xmath86 is an exceptional point for at least one of the operators @xmath89 and @xmath90 . then according to theorem 2.3 of @xcite we have the alternatives \(i ) let @xmath86 be an exceptional point for exactly one of the operators @xmath89 and @xmath90 .
then @xmath86 is a regular point for @xmath284 and therefore @xmath376 \(ii ) let @xmath86 be an exceptional point for both operators @xmath89 and @xmath90 .
then @xmath86 is an exceptional point for @xmath284 and therefore @xmath377 let now @xmath86 be a regular point for both @xmath89 and @xmath90 .
the discussion at the end of section 2 in @xcite states that @xmath86 is a regular point for @xmath284 for all except possibly one value @xmath378 . excluding this value @xmath379 ( however , we found no proof that @xmath380 ) we get that @xmath381 for all @xmath313 with @xmath382 .
now we reconsider the case @xmath21 .
we recall that in this case the spectral shift function equals minus the number of eigenvalues less than @xmath44 .
the discussion below shows that the bound @xmath383 for @xmath21 in general can not be improved .
let @xmath384 , @xmath385 , @xmath293 be the eigenvalues of @xmath342 .
let @xmath386 , @xmath387 be the eigenvalues of @xmath284 .
we recall some known properties of @xmath388 @xcite .
the functions @xmath388 are real analytic in @xmath389 .
first let @xmath21 be an eigenvalue of both @xmath89 and @xmath90 .
then @xmath284 has two eigenvalues @xmath390 , @xmath391 , which both converge to @xmath44 as @xmath392 .
moreover , @xmath393 and @xmath394 ( with prime denoting the derivative with respect to @xmath297 ) for all @xmath290 .
secondly , let @xmath44 be an eigenvalue of ( say ) @xmath89 but not of @xmath90 .
then the only eigenvalue @xmath395 of @xmath284 approaches @xmath44 as @xmath392 .
moreover , either @xmath396 , or @xmath397 , or @xmath398 for all @xmath290 .
there are no eigenvalues of @xmath284 other than the ones described above .
we note also that for all @xmath290 either @xmath399 for any @xmath400 and @xmath401 , or @xmath402 for some @xmath400 and @xmath401 .
indeed let us suppose that there is @xmath403 such that @xmath404 for some @xmath400 and @xmath401 . then inspecting the proof of theorem 1.2 in @xcite we see that this implies @xmath405 for all @xmath290 .
as is well known in the one - dimensional case the eigenvalues of @xmath284 are simple and therefore the @xmath388 s can not cross each other .
now fix some @xmath406 .
there are two cases to be considered : ( i ) @xmath194 is an eigenvalue of neither @xmath89 nor @xmath90 , ( ii ) @xmath194 is an eigenvalue of at least one of the hamiltonians @xmath89 , @xmath90 . by the discussion above in case ( i )
there is @xmath407 such that for all @xmath408 one has @xmath409 .
when @xmath290 decreases only one of the curves @xmath388 can pass through @xmath194 ( since the @xmath388 s can not cross each other ) thus decreasing or increasing @xmath410 by one . in case
( ii ) @xmath410 does not depend on @xmath290 and equals @xmath411 or differs from @xmath411 by one .
as stated in the introduction we will consider random schrdinger operators @xmath0 in @xmath1 of the form ( [ 1.1 ] ) with @xmath3 being a sequence of i.i.d .
variables on a probability space @xmath412 having a common density @xmath4 , which is continuous and has support in the finite interval @xmath6\subset{\mathbb{r}}$ ] . also the sequence @xmath3 is supposed to form a stationary , metrically transitive random field , i.e. there are measure preserving , ergodic transformations @xmath413 on @xmath414 such that @xmath415 .
we suppose that the single - site potential @xmath7 is in @xmath416 with @xmath417 $ ] and @xmath9 .
first we introduce the hamiltonians @xmath418 such that @xmath419 .
let @xmath420 be the spectral shift function for the pair @xmath421 .
the operators @xmath422 and @xmath423 are unitarily equivalent .
in fact , consider the unitary shift operator @xmath424 on @xmath1 given as @xmath425 .
then one has @xmath426 since the spectral shift functions for pairs of unitarily equivalent operators are equal , we have @xmath427 this remains true for the functions @xmath428 now let @xmath429 be an arbitrary integer such that @xmath430 . then due to corollary 3.4 we have that @xmath431 and @xmath432 now we show that @xmath433 where @xmath434 denotes the expectation with respect to the probability measure @xmath435 .
first we note that @xmath436 from the inequality ( [ subadditiv ] ) it follows that @xmath437 and then by the monotonicity theorem of the spectral shift function with respect to perturbations we have @xmath438.\ ] ] hence by lemma 2.1 @xmath439 similarly we can prove that @xmath440 .
indeed , @xmath441 thus we have proved * theorem 4.1 _ for every @xmath88 the family @xmath442 is a subadditive and @xmath443 is a superadditive random process . _
* applying now the akcoglu - krengel superadditive ergodic theorem we obtain that for every @xmath88 there is a set @xmath444 of full measure such that @xmath445 exists and is non - random .
we call this limit the spectral shift density .
now the problem is to show that the set @xmath446 of full measure can be chosen to be independent of @xmath44 as long as @xmath44 is a point of continuity of the limit .
we note that a priori the set @xmath447 is not necessarily of full measure .
we recall how this problem is solved for the density of states @xmath25 ( see e.g. @xcite ) .
once one has established the existence of the limit @xmath448 for every fixed @xmath44 and @xmath435-almost all @xmath449 , one can choose @xmath450 as the intersection of all sets @xmath446 when @xmath44 runs through the rationals and redefine the limiting function @xmath25 to make it right continuous .
since @xmath25 is a monotone nondecreasing function of @xmath44 , this could change the values of the limiting function on at most a countable set of discontinuities .
hence ( [ den.states ] ) is valid at every continuity point of @xmath25 for all @xmath451 , which is obviously of full measure . in our case
the limiting function @xmath77 is not monotone .
however intuition says that @xmath77 must be equal to @xmath452 ( this is indeed the case as will be proven below ) .
therefore , @xmath77 is expected to be at least of bounded variation . the simplest way to prove
this is to show that @xmath453 are lipshitz functions with lipshitz constants bounded uniformly in @xmath57 and @xmath454 , which is nothing but a wegner - type estimate @xcite for the spectral shift density .
this guarantees that @xmath77 is lipshitz continuous
. we will need the spectral averaging theorem @xcite : * lemma 4.2 _ let @xmath455 with @xmath456 .
let @xmath457 be nonnegative . @xmath458
the spectral decomposition of unity for @xmath459 .
then for any borel set @xmath149 @xmath460 _ * _ remark : _
@xmath461 is well defined since for every @xmath462 the operator @xmath463 is trace class ( see @xcite ) .
the present formulation of the spectral averaging theorem is a direct consequence of a slightly extended version of theorem 4 in @xcite .
this extension is straightforward and therefore we do not discuss details here .
now we prove * theorem 4.3 _ let @xmath7 be piecewise continuously differentiable with @xmath464 $ ] and assume there is a constant @xmath465 such that @xmath466 for a.e . @xmath467 .
then @xmath468 is lipshitz continuous for all @xmath88 , i.e. for each closed interval @xmath149 there is a constant @xmath469 such that @xmath470 for all @xmath471 . _ * thus it suffices to determine @xmath77 for @xmath44 running over a lebesgue - dense set of @xmath309 , say the rationals .
we choose the set @xmath450 as the intersection @xmath472 .
the limit ( [ limit ] ) then exists for all @xmath88 and all @xmath451 .
now we turn to the proof of theorem 4.3 .
we note that the method of the proof is not restricted to the one - dimensional case and can be extended to higher dimensions ( see @xcite ) .
_ proof of theorem 4.3 .
we apply lemma 2.2 to @xmath473 .
let us set @xmath474\subseteq\delta$ ] .
then we have @xmath475 where @xmath476 and @xmath477 is the spectral resolution for @xmath478 .
therefore @xmath479 ) , @xmath480 since for any @xmath481 the inequality @xmath482 holds , we have @xmath483 let us denote @xmath484 and let @xmath485 be the corresponding resolution of the identity .
then we have @xmath486 now we apply lemma 4.2 , according to which we obtain @xmath487 where @xmath488 stands for the spectral shift function of the pair ( @xmath489 , @xmath490 ) . by the chain rule and corollary 3.4 we have @xmath491 therefore , the second term on the r.h.s . of ( [ 3.terms ] ) can be bounded by @xmath492.\ ] ] by lemma 2.1 and since @xmath493 the maximum of @xmath494 is bounded .
_ now we estimate the first term on the r.h.s . of (
[ 3.terms ] ) .
by the fubini theorem @xmath495 using monotonicity and corollary 3.4 we obtain @xmath496.\ ] ] the expressions in the square brackets in ( [ bra.1 ] ) and ( [ bra.2 ] ) are finite by lemma 2.1 .
thus we have proved that @xmath497 now we note that by the lebesgue dominated convergence theorem @xmath498 for every fixed @xmath88 .
thus , taking the limit @xmath499 in ( [ limit.vor ] ) we arrive at the claim of theorem .
@xmath146 * theorem 4.4 _ the spectral shift density is the difference of the integrated densities of states for the free and the interacting hamiltonians , @xmath24 for all @xmath88 . _ * _ proof .
let @xmath500 and @xmath501 be the self - adjoint schrdinger operators corresponding to the differential expression ( [ hamiltonian ] ) and @xmath70 respectively on @xmath502 with dirichlet boundary conditions at @xmath503 and @xmath504 ( the friedrichs extension of ( [ hamiltonian ] ) on @xmath505 ) .
they have purely discrete spectrum and therefore the spectral shift function @xmath506 is simply the difference of the corresponding counting functions @xmath507 it is well known ( see e.g. @xcite ) that for all @xmath88 @xmath508 almost surely .
hence @xmath509 for almost all @xmath449 .
_ now we prove that the difference @xmath510 is bounded in absolute value by 2 uniformly in @xmath511 and @xmath449 .
therefore , the existence of the limit ( [ n0ne ] ) immediately implies the existence of @xmath77 and the equality @xmath24 .
thus , in fact we do not need theorem 4.3 in this context . by the chain rule for the spectral shift function we have @xmath512 here
@xmath513 denotes the operator ( [ hamiltonian ] ) on @xmath1 with dirichlet boundary conditions at @xmath503 and @xmath504 such that @xmath514 with respect to the direct decomposition of the hilbert space @xmath515 .
similarly there is a decomposition of @xmath516 with the same boundary conditions . from krein
s formula it follows that @xmath513 and @xmath517 are rank two perturbations of @xmath478 and @xmath11 , respectively .
thus , we immediately have @xmath518 for all @xmath519 and @xmath449 .
actually , ( [ improve ] ) can be improved ( * ? ? ?
* remark 5.2 ) such that @xmath520 from ( [ decomposition ] ) by lemma 3.1 it follows that @xmath521 this completes the proof of the theorem .
@xmath146 _
remark : _ we comment on the formula ( [ discuss ] ) .
we note that @xmath522 is continuous in @xmath16 although @xmath523 is a step - like function being the difference of two counting functions .
this means that the difference @xmath524 compensates the jumps of @xmath525 making the r.h.s .
of ( [ discuss ] ) continuous .
this was noted by jensen and kato @xcite . a similar phenomenon was found for obstacle scattering in @xmath526 by eckmann and pillet @xcite .
fig . 1 : spectral shift density for the deterministic kronig - penney model ( see text ) . as an illustration to theorem 4.4 ( see fig .
1 ) we have calculated @xmath527 ( dotted line ) for @xmath528 with @xmath7 taken to be the point interaction and @xmath529 ( the kronig - penney model ) and compared this result with @xmath452 ( solid line ) .
the density of states @xmath25 for this case can be given in closed analytic form ( see e.g. @xcite ) .
now we turn to a discussion of the density of the transmission coefficient .
let @xmath530 be the transmission amplitude at energy @xmath16 for the hamiltonian @xmath478 ( [ hamiltonian ] ) .
we now prove * theorem 5.1 _ for every fixed @xmath16 and almost all @xmath449 the limit @xmath531 exists and is non - random .
_ * _ remark : _ for periodic deterministic potentials the behavior of @xmath532 as @xmath13 was studied numerically in @xcite .
a very similar statement was proven earlier by marchenko and pastur @xcite .
our proof is a slight modification of that given in @xcite .
we start with * lemma 5.2 _ for all @xmath101 the transmission amplitude for the hamiltonian ( [ 2.1 ] ) satisfies the inequality @xmath533 _ * _ proof . by the aktosun factorization formula @xmath534 by the unitarity of the scattering matrix @xmath535 and @xmath536 , @xmath537 for all @xmath101 .
hence @xmath538 and the claim follows .
@xmath146 _ _ proof of theorem 5.1 we set @xmath539 from the fact that @xmath540 which is is an immediate consequence of ( [ transl ] ) we obtain @xmath541 by lemma 5.2 we have @xmath542 for all @xmath429 with @xmath543 . thus @xmath544 is a superadditive random process .
theorem 5.1 now follows by akcoglu - krengel superadditive ergodic theorem .
@xmath146 _
* theorem 5.3 _ for every @xmath16 @xmath545 equals @xmath27 , the upper lyapunov exponent for the fundamental matrix of the schrdinger operator . _
* this observation ( although without a complete proof ) is known ( see @xcite ) .
we start with recalling the definition of the lyapunov exponent ( see e.g. @xcite ) .
let @xmath546 be the fundamental matrix of the schrdinger equation @xmath547 such that @xmath548 for any solution of ( [ schr ] ) .
the lyapunov exponents @xmath549 are defined by @xmath550 actually it can be shown that @xmath551 for fixed @xmath44 and @xmath435-almost all @xmath552 .
moreover , @xmath553 can be replaced by @xmath554 ( see e.g. @xcite ) .
now fix @xmath16 .
we have to show that for @xmath435-almost all @xmath552 @xmath555 first let us redefine the fundamental matrix such that @xmath556 this obviously does not change the lyapunov exponent , which we can calculate as a limit over integer @xmath557 , @xmath558 this implies that we can express @xmath559 in terms of the transmission and reflection amplitudes for the hamiltonian @xmath560 .
more precisely , let us consider the particular solutions @xmath561 of the schrdinger equation with hamiltonian @xmath560 corresponding to the energy @xmath16 and for @xmath562 having the form @xmath563 then it is easy to see ( see @xcite and section 2 ) , that for @xmath564 these solutions have the form @xmath565 where @xmath566 and @xmath567 are transmission and reflection amplitudes for the hamiltonian @xmath560 , respectively .
therefore , the matrix elements of @xmath559 are given by @xmath568_{11 } & = & \frac{1}{2 } \left[\frac{l_\omega^{(1,n)}(e)^\ast}{t_\omega^{(1,n)}(e)^\ast } e^{-i\sqrt{e}(n+1)}+ \frac{l_\omega^{(1,n)}(e)}{t_\omega^{(1,n)}(e)}e^{i\sqrt{e}(n+1)}\right]\\ & & + \frac{1}{2}\left[\frac{e^{i\sqrt{e}n}}{t_\omega^{(1,n)}(e)^\ast } + \frac{e^{-i\sqrt{e}n}}{t_\omega^{(1,n)}(e)}\right],\\ \left[\widetilde{\phi}_\omega(n;e)\right]_{12 } & = & \frac{1}{2i\sqrt{e } } \left[\frac{l_\omega^{(1,n)}(e)^\ast}{t_\omega^{(1,n)}(e)^\ast } e^{-i\sqrt{e}(n+1)}- \frac{l_\omega^{(1,n)}(e)}{t_\omega^{(1,n)}(e)}e^{i\sqrt{e}(n+1)}\right]\\ & & + \frac{1}{2i\sqrt{e}}\left[\frac{e^{i\sqrt{e}n}}{t_\omega^{(1,n)}(e)^\ast } - \frac{e^{-i\sqrt{e}n}}{t_\omega^{(1,n)}(e)}\right],\\ \left[\widetilde{\phi}_\omega(n;e)\right]_{21 } & = & -\frac{i\sqrt{e}}{2 } \left[\frac{l_\omega^{(1,n)}(e)^\ast}{t_\omega^{(1,n)}(e)^\ast } e^{-i\sqrt{e}(n+1)}- \frac{l_\omega^{(1,n)}(e)}{t_\omega^{(1,n)}(e)}e^{i\sqrt{e}(n+1)}\right]\\ & & + \frac{i\sqrt{e}}{2}\left[\frac{e^{i\sqrt{e}n}}{t_\omega^{(1,n)}(e)^\ast } - \frac{e^{-i\sqrt{e}n}}{t_\omega^{(1,n)}(e)}\right],\\ \left[\widetilde{\phi}_\omega(n;e)\right]_{22 } & = & -\frac{1}{2 } \left[\frac{l_\omega^{(1,n)}(e)^\ast}{t_\omega^{(1,n)}(e)^\ast } e^{-i\sqrt{e}(n+1)}+ \frac{l_\omega^{(1,n)}(e)}{t_\omega^{(1,n)}(e)}e^{i\sqrt{e}(n+1)}\right]\\ & & + \frac{1}{2}\left[\frac{e^{i\sqrt{e}n}}{t_\omega^{(1,n)}(e)^\ast } + \frac{e^{-i\sqrt{e}n}}{t_\omega^{(1,n)}(e)}\right].\end{aligned}\ ] ] now using the relation ( see ( [ s.param ] ) ) @xmath569 we calculate the hilbert - schmidt norm of @xmath559 .
after some simple transformations we get that @xmath570 here for brevity we have introduced the notations @xmath571 obviously , the expression in the braces on the r.h.s . of ( [ long.1 ] ) is bounded from above for every @xmath16 uniformly in @xmath572 by @xmath573 . in appendix
a we show that this expression is also bounded from below by the positive constant @xmath574 uniformly in @xmath572 .
thus , @xmath575 now consider @xmath576 where @xmath577 and @xmath578 . by
( [ aux.4 ] ) for @xmath435-almost all @xmath552 @xmath579 independently of @xmath429 .
@xmath146 now we recall that @xmath580 .
then by ( [ limit ] ) and theorems 5.1 , 5.3 we have * corollary 5.4 _ for every @xmath16 and @xmath435-almost all @xmath552 @xmath581 _ * we note that corollary 5.4 can be reformulated in such a way that it permits a generalization to the higher - dimensional case .
we recall that ( see @xcite ) @xmath582 where @xmath583 now we can rewrite ( [ col.4 ] ) as follows @xmath584 in this form ( [ col.4 ] ) can now be generalized to the higher - dimensional case , thus defining @xmath27 , which is something like the multi - dimensional lyapunov exponent ( for details see @xcite ) .
also note that the above determinant equals the determinant of the jost matrix @xcite .
we can not use ( [ col.4 ] ) to calculate @xmath27 for @xmath21 .
however , this can be done by means of analytic continuation ( see section 6 ) .
we turn to the claim @xmath46 for almost all @xmath16 .
we start with some preparations .
let us denote @xmath585 by the identity @xmath586 and by theorem 5.3 one has that for every @xmath16 @xmath587 almost surely .
also ( [ gamma.e.1 ] ) follows directly from the definition of the lyapunov exponent ( [ lyap.def ] ) and theorem 2.4 .
let @xmath588 for some @xmath121 .
the corresponding elements of the scattering matrix at energy @xmath16 we denote by @xmath41 , @xmath42 , and @xmath43 .
let @xmath589 and @xmath590 obviously , @xmath591 where @xmath592 from the aktosun factorization formula ( [ aktosun.form ] ) it follows that @xmath593 since @xmath594 is unitary one obtains @xmath595 with @xmath596 since the @xmath597 form a sequence of random i.i.d .
variables , for every @xmath16 the sequence @xmath598 is a sequence of random i.i.d .
@xmath599-valued variables with corresponding distribution @xmath600 .
recall that the distribution density @xmath4 is supposed to be continuous with compact support .
the matrices @xmath601 have the form @xmath602 the closed subgroup of all matrices from @xmath599 having this form ( which we denote by @xmath603 ) is isomorphic to @xmath604 . indeed ,
@xmath605 with @xmath606 let @xmath607 be the spectral shift function for the pair of hamiltonians ( @xmath608 , @xmath11 ) such that @xmath609 for @xmath16 .
it is well known ( see e.g. @xcite ) that @xmath41 , @xmath42 , and @xmath43 at fixed energy @xmath16 are real analytic functions of @xmath121 .
since @xmath7 has compact support , @xmath41 , @xmath42 , and @xmath43 at fixed @xmath121 are real analytic with respect to @xmath16 @xcite .
moreover , they are jointly real analytic in @xmath121 and @xmath16 . since @xmath610 for all @xmath16 , @xmath611 and @xmath607 are also jointly real analytic in @xmath121 and @xmath16 .
we recall that potentials with compact support can not be reflectionless , i.e. @xmath612 for all @xmath16 implies @xmath613 @xcite . by real analyticity
the set @xmath614 for every fixed @xmath615 is discrete or empty .
let @xmath616 by the assumption that @xmath617 has a positive lebesgue measure and by real analyticity of @xmath42 , the condition @xmath612 for all @xmath618 or even for @xmath619 in a subset of @xmath617 of positive measure implies that @xmath612 for all @xmath121 . obviously , @xmath620 and therefore is also discrete or even empty . _
remark : _ one can easily show that a necessary ( but not sufficient ) condition for @xmath621 is @xmath622 we know , however , no example of a potential with @xmath623 and expect that actually @xmath624 , since , intuitively , it is clear that for the potential @xmath9 at hand @xmath625 as @xmath626 .
now for every fixed @xmath16 , @xmath627 we define the functions @xmath628 let @xmath629 if for some @xmath16 one of the functions
@xmath630 for all @xmath618 , then @xmath631 is real analytic with respect to @xmath121 and hence @xmath632 for all @xmath121 . also , it follows that the zeros of @xmath633 ( if any ) are all of even order .
thus , we have that if @xmath16 belongs to one of the sets @xmath634 , then either @xmath635 or @xmath636 for all @xmath121 .
taking @xmath613 in ( [ case.two ] ) and using @xmath637 and @xmath638 gives that @xmath639 with some @xmath572 . now calculating the limit @xmath640 of @xmath641 and taking into account that @xmath42 is not identically zero
, we obtain that for @xmath642 @xmath643 respectively .
thus , @xmath644 , \\ & & \\ \nonumber s^{(- ) } & \subseteq&\bigcup_{n=0}^\infty[\pi^2 ( 2n+1)^2 , 4\pi^2(n+1)^2].\nonumber\end{aligned}\ ] ] now let us suppose that @xmath44 belongs to the interior of the sets on the r.h.s . of ( [ set ] ) .
this implies that @xmath645 if @xmath646 and @xmath647 if @xmath648 . then using @xmath649
, we calculate @xmath641 ( with the corresponding choice of sign ) for small @xmath619 : @xmath650 thus @xmath651 and therefore @xmath652 for all @xmath121 .
this implies that @xmath653 for all @xmath121 .
hence @xmath612 for all @xmath121 , which contradicts the assumption @xmath627 . for the sake of convenience
we summarize some of the established properties of the sets @xmath634 : * lemma 5.5 * _ the sets @xmath634 are at most discrete .
more precisely , @xmath654 if @xmath642 then either ( [ case.one ] ) or ( [ case.two ] ) holds
. _ now we define @xmath655 which is at most discrete .
* theorem 5.6 * _ for almost all @xmath16 the upper lyapunov exponent @xmath46 almost surely .
more precisely @xmath27 vanishes for @xmath656 and almost surely nowhere else .
_ _ proof_. we split the proof in several steps and start with the case @xmath621 . calculating the eigenvalues of @xmath657
one easily finds that @xmath658 where the norm is understood in the operator sense .
obviously , @xmath659 iff @xmath612 .
therefore , if @xmath621 then @xmath660 and hence @xmath49 . to proceed
further we define the family of auxiliary periodic hamiltonians @xmath661 the spectrum of every @xmath662 is purely absolute continuous and has a band structure .
we recall ( see e.g. @xcite ) that the discriminant @xmath663 of ( [ hill ] ) is defined by @xmath664 , where @xmath665 and @xmath666 are solutions of @xmath667 with the initial data @xmath668 , @xmath669 .
@xmath670 is an entire function of @xmath114 .
the real solutions of the inequality @xmath671 determine the gaps in the spectrum of @xmath662 , whereas @xmath672 implies that @xmath44 belongs to the spectrum .
we say that _
@xmath44 is in a gap of @xmath662 _ for some @xmath121 if there is @xmath673 such that @xmath674 .
keller @xcite proved that for @xmath16 , @xmath675 thus , @xmath16 is in a gap of @xmath662 iff @xmath676 .
_ remark : _ we sketch another proof of this fact based on the ishii - pastur - kotani theorem @xcite .
if for some @xmath16 @xmath677 , then one of the eigenvalues of @xmath601 , @xmath678 is strictly larger than 1 in absolute value .
therefore , for the lyapunov exponent corresponding to @xmath662 we have @xmath679 since the spectrum of @xmath662 is absolutely continuous , by the ishii - pastur - kotani theorem @xmath44 lies in a gap .
conversely , if @xmath680 , then both eigenvalues ( [ eigenvalues ] ) lie on the unit circle and the lyapunov exponent vanishes , @xmath681 hence @xmath44 belongs to the spectrum .
we proceed with the proof of theorem 5.6 .
let us suppose that @xmath682 .
let @xmath683 be the smallest closed subgroup which contains the support of @xmath684 ( recall that @xmath684 is the image of @xmath4 under the map @xmath685 ) . to establish that @xmath46 for almost all @xmath16 almost surely we will also use the sufficient conditions of the furstenberg theorem ( * ? ? ?
* theorem a.ii.4.1 , proposition a.ii.4.3 ) ( in the complex form ) , i.e. for almost all @xmath16 the group @xmath686 is not compact and for any @xmath687 the set @xmath688 has more than two elements .
we start with the second condition in the furstenberg theorem .
first we calculate the ( non normalized ) eigenvectors of @xmath601 , @xmath689 in the case @xmath690 both eigenvectors coincide and the corresponding generalized eigenvector ( i.e. the solution of @xmath691 ) is @xmath692 . by means of the bijection @xmath693 , @xmath694 we can identify the complex projective line @xmath695 and @xmath696 . obviously , @xmath697 for fixed @xmath44 the reflection amplitude @xmath42 as a function of @xmath618 has at most a discrete set of zeros
. actually ( see ( * ? ? ?
* problem 6.4 ) ) for any subgroup @xmath698 one has the alternatives : ( ) @xmath699 is finite , there is @xmath700 such that @xmath701 there is @xmath700 such that @xmath702 for any @xmath703 the set @xmath704 has more than two elements .
let us first suppose that @xmath705 is finite .
since @xmath601 is real analytic with respect to @xmath619 this implies that @xmath601 is constant for all @xmath121 and hence @xmath611 does not depend on @xmath619 .
therefore @xmath706 and thus @xmath621 .
consider the case ( ii ) for @xmath705 .
let us suppose that for some @xmath707 the matrices @xmath708 are diagonal for all @xmath619 in a subset of @xmath617 of positive measure .
the existence of such @xmath707 implies that the eigenvectors of all @xmath601 are constant as elements of @xmath695 and thus @xmath709 with @xmath710 for all such @xmath619 .
but then these relations hold for all @xmath121 by real analyticity .
thus @xmath711 .
conversely , if @xmath712 ( such that @xmath713 ) and if in addition @xmath714 for all @xmath619 in a subset of @xmath617 of positive measure , then @xmath715 does not depend on @xmath619 in this set and @xmath708 is diagonal for all these @xmath619 .
hence for this choice of @xmath707 and by real analyticity @xmath716 is diagonal for all @xmath121 . by the previous discussion ( see lemma 5.5 ) in case @xmath717 with equality for @xmath619 only on a set of measure zero the eigenvalues of @xmath598 lie on the unit circle . therefore
, since @xmath718 we have @xmath719 . thus @xmath49 .
consider the opposite case when @xmath720 with equality for @xmath619 only on a set of measure zero . in this case one of the eigenvalues @xmath721 ( say @xmath722 ) for almost all @xmath618 is larger than 1 in absolute value . since @xmath723 the eigenvalues of @xmath724
are given by @xmath725 and @xmath726 and thus @xmath727 almost surely .
here we used the fact that @xmath728 is metrically transitive and the birkhoff - khintchin theorem .
it remains to consider the case when @xmath711 and @xmath690 for almost all @xmath619 in @xmath617 . by real analyticity
this relation holds for all @xmath121 .
but then @xmath729 and hence @xmath730 for all @xmath121 .
evaluating at @xmath613 gives @xmath731 for all @xmath619 and thus @xmath621 .
we note that from the real analyticity with respect to @xmath619 it follows that if there exists @xmath700 such that @xmath708 is diagonal for @xmath619 on a set of positive measure , then @xmath708 is diagonal for all @xmath121 .
to conclude the discussion of case ( ii ) for @xmath705 suppose now that there is @xmath700 such that for almost all @xmath618 @xmath732 holds for some @xmath733 . by real analyticity in @xmath619 of @xmath601
such a relation holds for all @xmath121 . taking traces
gives @xmath734 for all @xmath121 .
again by real analyticity this implies @xmath735 .
in particular , ( [ q.off ] ) can not hold if @xmath736 . in case @xmath737
observe that for @xmath738 @xmath739 and @xmath740 with @xmath741 hence any matrix which diagonalizes @xmath601 is necessary of the form @xmath742 with @xmath743 .
this implies that @xmath744 for almost all @xmath618 , where @xmath745 and @xmath746 are suitable nonvanishing functions .
writing ( [ bez ] ) explicitly gives @xmath747 from this it follows that for all @xmath618 @xmath748 from both ( [ q.off ] ) and ( [ neu.2 ] ) it follows that @xmath749 is real analytic with respect to @xmath121 , and thus the relations ( [ neu.1 ] ) , ( [ neu.2 ] ) hold for all @xmath121 .
taking @xmath613 gives @xmath750 the existence of a nontrivial solution in @xmath751 and @xmath752 of this system implies that @xmath753 .
thus , we obtain @xmath754 and @xmath755 . inserting this in ( [ neu.1 ] )
we obtain @xmath756 for all @xmath121 .
this together with @xmath757 obviously implies that @xmath758 is constant and thus @xmath612 for all @xmath121 .
thus @xmath621 .
this completes the discussion of the case ( ii ) .
now consider the case ( iii ) . as it is easy to see in this case that all @xmath601 have a common eigenvector , such that only one of the functions @xmath759 is constant or @xmath760 ( and thus @xmath690 ) for all @xmath618 .
in the first case either @xmath646 or @xmath648 but @xmath761 . in the second case the matrices @xmath762 are not diagonalizable and @xmath763 conversely , if one of @xmath759 ( say @xmath764 ) does not depend on @xmath619 , then the matrix @xmath765 is such that @xmath766 let @xmath646 and @xmath680 for all @xmath619 .
then all eigenvalues of all @xmath601 lie on the unit circle .
we have @xmath767 obviously , there is @xmath673 such that @xmath768 for all @xmath618 .
it is easy to see that if the numbers @xmath769 are such that @xmath770 , then @xmath771 with @xmath772 satisfying the inequality @xmath773 .
thus , the norm of the matrix on the r.h.s . of ( [ kreutz ] ) is less or equal @xmath774 .
therefore , @xmath775^{1/2},\ ] ] and thus @xmath49 .
let now @xmath646 and @xmath677 for almost @xmath618 . in this case either @xmath776 for almost all @xmath618 or @xmath777 for almost all @xmath618 .
the eigenvalue of @xmath724 corresponding to @xmath778 is given by @xmath725 .
therefore @xmath779 if @xmath776 , or @xmath780 if @xmath777 .
by the birkhoff - khintchin theorem we have @xmath781 almost surely .
if @xmath776 ( @xmath782 ) for almost all @xmath619 , we have @xmath783 ( @xmath784 ) , and hence @xmath785 almost surely .
similarly , in the case @xmath648 and @xmath677 for almost all @xmath618 we have @xmath46 almost surely .
thus , we have shown that the cases ( i ) , ( ii ) , ( iii ) occur iff @xmath786 .
moreover , @xmath49 for @xmath787 and @xmath46 almost surely for @xmath788 . for those and only those @xmath16 , which do not belong to @xmath789 the case ( iv ) occurs , and thus the second condition of the furstenberg theorem is fulfilled .
we turn to the first condition of the furstenberg theorem .
if @xmath44 is in a gap of @xmath662 for at least one @xmath618 , then one of the eigenvalues ( [ eigenvalues ] ) is strictly larger than 1 in absolute value .
therefore @xmath790 as @xmath13 , and hence @xmath686 is not compact
. now suppose that @xmath44 is not in a gap of @xmath662 for all @xmath618 , i.e. @xmath717 for such @xmath619 . in this case
the eigenvalues @xmath721 lie on the unit circle . since @xmath617 contains at least two @xmath619 s such that the corresponding @xmath791 s have no common eigenvectors , by the assumed continuity of @xmath4 we can select them in such a way that either @xmath792 or @xmath690 for both @xmath619 s . in the first case we can apply the arguments of @xcite to construct a matrix , which belongs to @xmath686 and has an eigenvalue strictly larger than 1 . to treat the second case we consider two matrices @xmath793 with no common eigenvectors , satisfying @xmath794 for both @xmath537 .
since @xmath795 one has @xmath796 .
the matrices @xmath797 are not diagonalizable , because to the eigenvalue @xmath798 corresponds the only eigenvector @xmath799 .
further we consider the matrices @xmath800 , @xmath801 , which are defined as follows : @xmath802 ,
@xmath537 if @xmath803 ; @xmath804 , @xmath805 if @xmath806 .
these matrices have the form @xmath807 such that @xmath808 and @xmath809 .
we can write the matrices @xmath810 in the following form @xmath811 with @xmath812 where @xmath813 . obviously , @xmath814 .
since by assumption the eigenvectors of the matrices @xmath797 are linearly independent , we have that @xmath815 .
now we calculate @xmath816 therefore , one of the eigenvalues of @xmath817 is strictly larger than 1 .
thus , by the furstenberg theorem it follows that @xmath46 almost surely for @xmath818 .
this completes the proof of the theorem .
let @xmath55 denote the inner product in @xmath56 .
we note that @xmath819 with @xmath53 , @xmath54 . from ( [ lambda.u ] )
it follows that @xmath820 obviously , @xmath821 . therefore the r.h.s . of ( [ stern.2 ] )
equals @xmath822 thus we obtain @xmath823 since @xmath824 , it follows that @xmath825 this representation is similar to the definition of the density of states through the rotation number of fundamental solution of the schrdinger equation @xcite .
the representations ( [ gamma.e.1 ] ) , ( [ gamma.e.2 ] ) can also be rewritten in a similar form , @xmath826 * example 1 . * here we consider the hamiltonian ( [ 1.1 ] ) where @xmath7 is ( formally ) replaced by the dirac @xmath50-function . in this case
the transmission and reflection amplitudes are given by @xmath827 therefore @xmath624 and @xmath828 thus @xmath829 .
from theorem 5.6 it follows that @xmath27 is positive on @xmath48 except for the set @xmath830 . that @xmath49 iff @xmath831 , @xmath832 was proved by ishii @xcite ( see also @xcite ) .
kirsch and nitzschner @xcite proved that @xmath833 . here
we reconsider these facts once more .
the matrices @xmath235 can be calculated explicitly : @xmath834 where @xmath835 for @xmath836 , @xmath837 @xmath838 obviously @xmath839 is nilpotent , i.e. @xmath840 . therefore @xmath841 from this it follows that @xmath842 and hence @xmath843 since @xmath844 is metrically transitive , by the birkhoff
khintchin ergodic theorem we have @xmath845 therefore , the sum @xmath846 in the limit @xmath223 increases not faster than @xmath847 . hence @xmath848
in this section we show that the analyticity of @xmath849 in the upper complex @xmath44-plane @xmath209 and the thouless formula ( [ thouless ] ) are a direct consequence of the fact that the functions @xmath850 are analytic in @xmath209 for every @xmath519 . by lemma 2.3
we have @xmath851 * lemma 6.1 * _ let @xmath852 with @xmath853 .
then there is a constant @xmath127 independent of @xmath44 and @xmath519 such that @xmath854 for all @xmath855 . _ _ proof .
_ by the monotonicity of the spectral shift function we have @xmath856 where @xmath857 is the characteristic function of the interval @xmath858 $ ] .
calculating @xmath859 explicitly we find @xmath860 where @xmath861 is the multivalued @xmath862 function such that @xmath863 is continuous and nondecreasing with respect to @xmath557 .
since @xmath864 for all @xmath855 , it follows that @xmath865 therefore @xmath866 obviously , @xmath867 for all @xmath868 , thus proving the lemma .
@xmath146 now to study the limit @xmath499 we can use the theorem on the continuity of the stieltjes transform ( see e.g. ( * ? ? ?
* appendix a ) ) .
the applicability of this theorem is guaranteed by lemma 6.1 , from which it follows that @xmath869 and @xmath870 therefore , since @xmath871 for all @xmath88 and @xmath435-almost all @xmath449 , and @xmath77 is nonrandom , we obtain that @xmath872 converges for all @xmath114 with @xmath204 and @xmath435-almost all @xmath449 to some deterministic limit @xmath873 , which is given by @xmath874 and therefore is analytic for @xmath204 .
since @xmath77 is continuous ( moreover hlder continuous ) by the sokhotski - plemelj formula we have that @xmath875 exists for all @xmath88 and @xmath876 as in the proof of lemma 2.3 , since @xmath77 is continuous and is of bounded variation , we get @xmath877 on the other hand , by lemma 2.3 we have @xmath878 by theorem 5.1 we have that for every fixed @xmath16 @xmath879 almost surely .
now we prove that for lebesgue almost all @xmath88 there are subsequences @xmath880 , @xmath881 tending to infinity and such that for @xmath435-almost all @xmath552 @xmath882 thus , from ( [ rew ] ) and ( [ game ] ) it will follow that @xmath883 and therefore @xmath884 for almost all @xmath16 .
the arguments used below are very similar to those of pastur and figotin ( * ? ? ?
* theorem 11.6 ) .
consider the functions @xmath885 let us fix some interval @xmath886 and consider @xmath887 where @xmath888 ( the set of all borel subsets in @xmath889 ) .
the family of functions @xmath890 is uniformly bounded and equicontinuous on any bounded interval and @xmath891 for some @xmath892 .
also we have @xmath893 since the family @xmath894 is uniformly bounded and equicontinuous and since @xmath895 converges vaguely to @xmath896 almost surely , it follows that @xmath897 on the other hand one has @xmath898 from ( [ stern ] ) it follows that @xmath899 denote by @xmath900 the subset of @xmath414 such that @xmath901 and for which @xmath902 @xmath903 converges to @xmath896 .
taking an arbitrary @xmath904 we can choose subsequences @xmath880 , @xmath881 tending to infinity and such that @xmath905 for lebesgue almost all @xmath906 . since @xmath907 for almost all @xmath449 , relation ( [ stern2 ] ) remains valid for almost every @xmath449 .
thus ( [ 6.5 ] ) is proven .
we have shown that @xmath908 for almost all @xmath16 .
therefore , @xmath873 is an analytic continuation of @xmath909 from the real semiaxis .
let us consider the function @xmath910 , where @xmath911 with @xmath912 being the lyapunov exponent for @xmath70 .
it is obviously analytic in the upper half - plane @xmath913 and @xmath914 from ( [ integ.repr ] ) and from the fact that @xmath77 is given by the difference of two nonnegative functions @xmath915 and @xmath25 , it follows that @xmath873 is the difference of two nevanlinna functions . since @xmath916 is a nevanlinna function ,
so is @xmath917 . also from ( [
6.5 ] ) it follows that the thouless formula ( [ thouless ] ) holds .
let @xmath72 be some real - valued uniformly locally integrable function . for simplicity we suppose that @xmath72 is uniformly bounded , i.e. @xmath918 , @xmath919 .
the last assumption can be weakend but we do not go into details here .
let @xmath61 be a sequence of real numbers such that @xmath62 as @xmath63 and @xmath920 for all @xmath65 .
we suppose that @xmath61 is such that all the differences @xmath921 are finite ( but not necessarily uniformly bounded ) .
let @xmath922 be the characteristic function of the interval @xmath923 $ ] .
we denote @xmath924 such that @xmath222 tends to @xmath72 as @xmath223 . by the above assumption one has @xmath925 .
let @xmath47 , @xmath225 and @xmath224 denote the hamiltonians with domains of definition being the sobolev space @xmath226 , @xmath926 let @xmath927 , @xmath928 and @xmath929 , @xmath930 be the spectral shift function and transmission amplitude for the pairs ( @xmath225 , @xmath11 ) and ( @xmath224 , @xmath11 ) , respectively . also as above we denote @xmath931 by corollary 3.4 and theorem 5.2 we have latexmath:[\ ] ]
therefore @xmath973 thus proving the claim .
t. aktosun , m. klaus , and c. van der mee , factorization of scattering matrices due to partitioning of the potentials in one - dimensional schrdinger - type equations " , _ j. math . phys .
* 37 * , 5897 - 5915 ( 1996 ) . _
d.boll , f. gesztesy , h. grosse , w. schweiger , and b.simon , witten index , axial anomaly and krein s spectral shift function in super - symmetric quantum mechanics " , _ j. math .
_ * 28 * , 1512 - 1525 ( 1987 ) .
a. jensen , high energy asymptotics for the total scattering phase in potential scattering theory " , in h. fujita , t. ikebe , and s.t .
kuroda ( eds . ) _ functional - analytic methods for partial differential equations , lecture notes in mathematics 1450 , berlin , springer , 1990 . _
w. kirsch , s. kotani , and b. simon , absence of absolutely continuous spectrum for some one dimensional random but deterministic operators " , _ ann .
henri poincar , phys . theor . * 42 * , 383 - 406 ( 1985 ) .
_ w. kirsch and f. nitzschner , lifshitz - tails and non - lifshitz - tails for one - dimensional point interactions " , in p. exner and h. neidhardt ( eds . ) _ order , disorder and chaos in quantum systems , operator theory : advances and applications vol .
46 , basel , birkhuser , 1990 , pp .
171 - 178 . _
s. kotani , ljapunov indices determine absolutely continuous spectra of stationary random one - dimensional schrdinger operators " , in k. ito ( ed . ) _ stochastic analysis , north - holland mathematical library , vol .
32 , amsterdam , north holland , 1984 , pp .
225 - 247 . _
papanicolaou ,
random matrices and waves in random media " , in j.e .
cohen , h. kesten , and c.m .
newman ( eds . ) _ random matrices and their applications _ , contemporary mathematics , vol .
soc . , providence , r.i .
, 1986 , pp .
311 - 317 . | methods from scattering theory are introduced to analyze random schrdinger operators in one dimension by applying a volume cutoff to the potential .
the key ingredient is the lifshitz - krein spectral shift function , which is related to the scattering phase by the theorem of birman and krein .
the spectral shift density is defined as the thermodynamic limit " of the spectral shift function per unit length of the interaction region .
this density is shown to be equal to the difference of the densities of states for the free and the interacting hamiltonians . based on this construction
, we give a new proof of the thouless formula .
we provide a prescription how to obtain the lyapunov exponent from the scattering matrix , which suggest a way how to extend this notion to the higher dimensional case .
this prescription also allows a characterization of those energies which have vanishing lyapunov exponent . |
only a small fraction of active glactic nuclei ( agns ) exhibit double - peaked broad - line profiles , for example , the largest sample available so far is the 116 double - peaked balmer line agns , which are found from an initial sample consisting of 5511 broad line agns with @xmath7 observed by the sloan digital sky survey(sdss ) @xcite .
some previous authors mainly focused on the double - peaked broad lines in radio - loud(rl ) agns ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
a complete survey on rl agns finally resulted in 20 double - peaked line emitters @xcite .
it is still a mystery why ( only ) a small fraction of agns exhibit double - peaked broad - line profiles .
some different scenarios were suggested for the origin of double - peaked emission lines , namely , ( 1 ) emission from the accretion disk ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , ( 2 ) emission from a binary broad - line region ( blr ) in a binary massive black hole system ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , ( 3 ) emission from the bipolar outflows ( e.g. , * ? ? ?
* ) , ( 4 ) emission from a spherically symmetric blr illuminated by an anisotropic ionizing radiation source ( e.g. , * ? ? ?
the accretion disk model is the most favorable one among them ( see detailed comparison between these different scenarios in * ? ? ?
* and references therein ) .
the widths of double - peaked lines range from several thousand to nearly 40000 km s@xmath8 ( e.g. * ? ? ?
* ) . in the accretion disk model ,
the double - peaked emission lines are radiated from the disk region between around several hundred gravitational radii to more than two thousand gravitational radii , and their profiles can be well fitted by the accretion disks with a power - law line emissivity ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
however , none of these scenarios can answer why only a small fraction of agns have been detected as double - peaked line emitters .
for example , in the accretion disk model , one may expect to observe double - peaked lines in most agns , as the agns with higher inclination angles have more chances to be observed , until their accretion disks are obscured by the putative tori ( e.g. , * ? ? ? * ) , if the orientations of agns are isotropically distributed in space .
another difficulty for the disk model is the energy budget " problem , i.e. , the local viscously dissipated power in the line - emitting disk portion is usually insufficient for the observed double - peaked broad - line luminosity in most sources , and the temperatures of the line - emitting disk regions are too low to produce the observed h@xmath9 lines @xcite .
the x - rays from a hot ion - supported torus in the inner disk region are assumed to irradiate the outer line - emitting disk region and then to solve the energy budget " problem ( e.g. , * ? ? ? * ) .
we discuss the illumination of the outer disk regions by the inner ion - supported tori in 2 . in this paper , we suggest a physical model for these double - peaked line emitters . in this model ,
a fraction of photons from the accretion disk are scattered by the electrons in the jet / outflow back to photo - ionize the outer line - emitting region , and then to produce the observed emission lines .
we describe our model in 3 .
the cosmological parameters @xmath10 , @xmath11 , and @xmath12 have been adopted in this work .
the putative hot ion supported tori in double - peaked line emitters are well modelled by the advection dominated accretion flows ( adafs , old version of riafs ) @xcite .
the riaf can survive only if its accretion rate @xmath13 ( @xmath14 ) is less than a critical value @xmath15 , which is a function of the viscosity parameter @xmath16 ( e.g. , * ? ? ?
the three - dimensional mhd simulations suggest that the viscosity parameter @xmath16 in the accretion flows is @xmath17 @xcite , or @xmath18 @xcite . for some double - peaked line emitters
, their bolometric luminosities could be as high as one - tenth of the eddington value .
their x - ray luminosities @xmath19 could be as high as 10@xmath20 ergs s@xmath8 for the double - peaked line emitters with black hole masses of 10@xmath21 ( e.g. , table 7 in * ? ? ? * and table 1 in this paper ) .
the maximal x - ray luminosity for a riaf surrounding a massive black hole can be calculated based on the global solution for the riaf .
adopting @xmath22 , corresponding to @xmath23 , which is a conservative choice , we can calculate the global structure of the riaf .
for such a riaf accreting at the critical rate , its x - ray luminosity @xmath24 ergs s@xmath8 for a 10@xmath25 @xmath26 black hole ( see * ? ? ?
* ; * ? ? ?
* for detailed description of the calculation on the flow structure ) .
the x - ray luminosity of the riaf depends almost linearly on the black hole mass while @xmath13 is fixed . for some x - ray luminous double - peaked line emitters ,
a very high viscosity @xmath27 is required for the presence of riafs in these sources , which is in contradiction with the mhd simulations .
it is well known that most gravitational energy of accretion matter is released in a small region close to the black hole , either for a standard thin disk @xcite or a riaf . in fig .
[ fig1 ] , we calculate the ratio @xmath28 as a function of @xmath29 based on the global riaf structure @xcite . it is found that more than 70 per cent radiation is from the region within 0.3 @xmath30 , where @xmath30 is the radius of the riaf connecting to the outer standard thin disk . assuming homogeneous distribution of the gases in @xmath31-direction of the riaf , our calculation shows that less than 2.3 per cent radiation from the riaf in the inner region of the disk can illuminate the line - emitting disk region with @xmath32 , because the solid angle of the outer disk portion subtended to the inner region of the riaf is too small . for more realistic @xmath31-direction density distribution of the riaf ,
the gases are denser in the midplane of the flow and the resulted fraction should even be less than this value .
thus , the hot ion - supported tori are unable to heat the outer line emitting disk regions efficiently for most double - peaked line emitters , even if the observed luminous x - ray emission can be attributed to hot ion - supported tori . 1.0 cm
in this paper , we employ the inhomogeneous conical jet model ( see * ? ? ?
* for details ) , which can successfully explain most observational features of radio - loud agns ( e.g. , * ? ? ?
* ) . for a free jet , its half - opening angle @xmath33 @xcite , where @xmath34 is the lorentz factor of the jet .
the non - thermal electrons in the jet is described by a power - law energy distribution , @xmath35 from @xmath36 to @xmath37 @xcite . in knigl
s jet model , the jet is assumed to move at a constant velocity , which is a good approximation , though the dynamics of electron - positron jets is very complicated ( e.g. , * ? ? ?
* ) . for simplicity
, we adopt a conventional assumption of the jet moving at a constant velocity @xmath38 .
it is still unclear for the jet composition , i.e. , electron - positron or electron - proton ( see * ? ? ?
* for a recent review and references therein ) . for an electron - proton jet , the lower limit of electron lorentz factor @xmath36 is required to be greater than 100 , while @xmath36 could be as low as unity for an electron - positron jet ( e.g. , * ? ? ?
this is consistent with the detection of circular polarization , which strongly suggests that the jets are electron - positron plasmas and have low @xmath39 ( e.g. * ? ? ?
the similar conclusion is arrived from powerful large scale x - ray jets , if they are interpreted as inverse - compton scattering of cosmological microwave background photons in fast jets .
the typical energy of the photons radiated from the standard thin accretion disk surrounding a massive black hole is @xmath40 .
the soft photons from such a standard disk are compton up - scattered to the hard x - ray band ( @xmath41 10 kev , or even @xmath42-ray bands ) for an electron - proton jet , because @xmath43 .
these hard x - ray photons can hardly photo - ionize the outer line - emitting disk region .
if a thermal electron population with similar density co - exists with the non - thermal power - law electrons in the relativistic jet , the soft x - ray photons scattered by the thermal electrons in the jet can ionize the disk .
such thermal plasmas in a relativistic jet may produce significant doppler shifted emission or absorption lines depending on their temperature , which seems inconsistent with observations .
thus , we will not consider this possibility in this work , though the presence of thermal electrons in the jet can not be ruled out .
the situation is different for the electron - positron jet , of which @xmath44 .
most compton scattered photons by the electrons in electron - positron jets are in soft x - ray bands , which can efficiently photo - ionize the line - emitting disk region , so we will focus on the electron - positron jets hereafter in this work .
the kinetic luminosity of a relativistic moving electron - positron jet with lorentz factor @xmath34 is @xmath45^{-1}\ ] ] @xmath46 @xmath47 where @xmath48 is electron rest mass , @xmath49 is assumed , and the mass loss rate of the jet is given by @xmath50 for an accretion disk surrounding a black hole , most gravitational energy is released in the inner region of the disk ( within several schwarzschild radii close to the marginal stable orbit @xmath51 , * ? ? ?
* ) , where @xmath51 is in units of gravitational radius @xmath52 . for simplicity
, we use a ring with radius @xmath53 , where is the main contribution to the integral disk luminosity @xcite , to approximate the accretion disk radiation .
the solid angle of the surface of the jet slice between @xmath54 and @xmath55 subtended to the accretion disk ( ring ) with radius @xmath56 is @xmath57 @xmath58\times[(r_{\rm j}-r_{\rm d}\sin\phi_{\rm j})^2+r_{\rm d}^2\cos^2\phi_{\rm j}]^{1/2 } } } , \label{domega}\ ] ] where @xmath54 is the distance from the apex of the jet .
the jet is compton thin in transverse direction at large radii , while it becomes compton thick at small radii .
we can estimate the transition radius @xmath59 by @xmath60 beyond which the jet becomes compton thin in transverse direction for electron scattering . here the thomson cross - section @xmath61 @xmath62 . in the region of the jet with @xmath63
, the photons from the disk are mostly scattered in a thin layer at the jet surface .
the angle - dependent inverse - compton scattered photons from the slice of the jet @xmath64 measured in the source frame is @xmath65 @xmath66^{1/2 } } } { \frac { { \rm d}\omega(r_{\rm j})}{{\rm d}r_{\rm j}}},\label{dnscatangthick}\ ] ] for the compton thick part of the jet ( @xmath67 ) , where @xmath68 , and @xmath69 is the angle of the scattered photons measured from the axis of the jet in the source frame given by @xmath70^{1/2}}}. \label{mu}\ ] ] as only the scattered photons with energy @xmath71 ev can efficiently photo - ionize the gases in the line - emitting disk region , the correction factor @xmath72 describing the fraction of the scattered photons with energy @xmath73 measured in the source rest frame , is given by @xmath74 .
\label{fic}\ ] ] only the photons scattered by the electrons with @xmath75 are more energetic than 13.6 ev measured in the source frame , so the value of @xmath76 is given by @xmath77 , \label{gammaeminp}\ ] ] where @xmath78^{1/2}}}.\ ] ] assuming a line - emitting ring with radii @xmath79 extending from @xmath80 to @xmath81 , we obtain the total photons scattered in the compton thick jet slice which can photo - ionize the line - emitting disk region , @xmath82 @xmath83^{1/2 } } } { \frac { { \rm d}\omega(r_{\rm j})}{{\rm d}r_{\rm j}}}2\pi{\rm d}\mu_{\rm l}.\ ] ] @xmath84 @xmath85^{3/2}\times[(r_{\rm j}-r_{\rm d}\sin\phi_{\rm j})^2+r_{\rm d}^2\cos^2\phi_{\rm j}]^{1/2}}}.\label{dnscat}\ ] ] for the jet part with @xmath86 , it is compton thin for electron scattering . in this case , we have to include radiative transfer of the incident photons in the jet .
similar to compton thick case , we can derive the inverse - compton scattered photons from the slice of the jet @xmath87 photo - ionizing the line - emitting ring measured in the source frame as @xmath88 @xmath89\ ] ] @xmath90^{3/2}\times[(r^\prime_{\rm j}-r_{\rm d}\sin\phi_{\rm j})^2+r_{\rm d}^2\cos^2\phi_{\rm j}]^{1/2 } } } , \label{dnscatthin}\ ] ] where @xmath91 and @xmath92 here the fraction @xmath72 is also given by eqs .
( [ fic ] ) and ( [ gammaeminp ] ) , while @xmath93 is given by @xmath94 the electron density , @xmath95 , for a relativistic jet as required by the mass conservation along the jet .
the optical depth @xmath96 @xmath97 where @xmath98^{1/2}}}.\ ] ] the occultation by the compton thick part of the jet is considered , so the lower integral limit of eq .
( [ dnscatthin ] ) is given by @xmath99 , and @xmath100^{1/2}\ ] ] @xmath101 where @xmath102^{1/2}}}.\ ] ] the total photons scattered by the jet that can photo - ionize the line - emitting disk region is @xmath103 we can roughly estimate the luminosity of h@xmath16 line emitted from the disk region between @xmath80 and @xmath81 by @xmath104 if the line emitting portion is photo - ionized by the x - ray photons scattered in the jet and the thermalization is not important .
we can derive @xmath105 from eqs .
( [ lkin1 ] ) , and ( [ mdotjet ] ) , if the ratio @xmath106 is specified . using eqs .
( [ lkin1 ] ) , ( [ dnscat ] ) , ( [ dnscatthin ] ) , ( [ dnscatdt ] ) and ( [ lhalpha ] ) , we can calculate @xmath107 as a function of @xmath106 , when @xmath34 , @xmath36 , and @xmath108 are specified .
in this paper , we use the sample of rl agns with double - peaked broad lines given by @xcite to test our model .
there are 20 sources with detected double - peaked h@xmath9 lines in this sample ( see table 7 in their paper , and references therein ) .
the bolometric luminosity is derived from soft x - ray luminosity .
they found that the viscous power output @xmath109 of the line - emitting disk region is insufficient for the observed h@xmath9 luminosity for these sources , of which seven sources even have @xmath110 .
we list the bolometric and h@xmath9 luminosities in table 1 ( converted to the cosmology adopted in this paper ) .
we estimate the jet power of these sources from their low - frequency radio emission .
the relation between jet power and radio luminosity proposed by @xcite is @xmath111 where @xmath112 is the total radio luminosity at 151 mhz in units of 10@xmath113 w hz@xmath8 sr@xmath8 . @xcite
have argued that the normalization is uncertain and introduced the factor @xmath114 to account for these uncertainties .
they use a wide variety of arguments to suggest that @xmath115 .
@xcite argued that @xmath116 is a likely consequence of the evolution of magnetic field strengths as radio sources evolve . in this paper , we adopt @xmath117 in all our calculations .
finally , we need to estimate the masses of the black holes in these double - peaked line emitters .
it is still unclear whether the empirical relation between the blr size and optical continuum luminosity suggested by @xcite still holds for these agns ( wang et al .
2005 ; but also see * ? ? ?
* ) , because the double - peaked lines are assumed to originate from accretion disks , which is different from the conventional cloud models for blrs . in this paper , we use the relation between host galaxy absolute magnitude @xmath118 at r - band and black hole mass @xmath119 proposed by @xcite , @xmath120 to estimate the black hole masses .
there are seven sources in this sample , of which the host galaxy magnitudes are unavailable in literature .
we adopt the total magnitudes ( including the nuclear emission ) , which may over - estimate the black hole masses for these seven sources . in this work
, we adopt @xmath121 for a free jet @xcite , which is a good approximation even for a slow jet with @xmath122 . @xcite
have argued that the lower energy cut for the non - thermal distributed relativistic electrons in the jets can be as low as @xmath44 .
our model calculations are carried out for @xmath123 .
the spectral index for optically thin synchrotron radiation @xmath124 is adopted for all calculations in this work , which is consistent with observations on radio - loud agns ( e.g. , * ? ? ?
the model calculations can be carried out , if the integral limits : @xmath125 , @xmath80 , and @xmath81 are specified . in all our calculations ,
@xmath126 is adopted . in fig .
2 , we show how the transition radius @xmath59 of the jet from compton thick to compton thin varies with jet kinetic luminosity @xmath127 .
the transition radius @xmath59 increases with @xmath127 .
the jet can be compton thin down to @xmath125 , when @xmath128 .
for the cases with high kinetic luminosity @xmath106 , the jet is compton thick even at very a large radius @xmath54 .
we integrate eq .
( [ dnscatdt ] ) over @xmath54 ( from @xmath126 ) to calculate the emissivities @xmath129 as functions of radius in fig .
[ fig3 ] for different values of @xmath34 .
the final results depend insensitively on the value of @xmath130 , because the solid angle of the outer line - emitting region subtended to the jet is very small if @xmath54 is small .
it is found that the emissivities have nearly power - law @xmath1-dependent distributions : @xmath131 , and the values of @xmath132 are @xmath133 .
this implies that the ratio @xmath107 is mainly governed by the inner radius @xmath80 of the line - emitting disk region .
we integrate eq .
( [ dnscatdt ] ) , and the relations between @xmath106 and @xmath134 are plotted in fig .
[ fig4 ] for different values of @xmath34 .
the data of the sample are plotted in the same figure for comparison .
in the ion - supported torus scenario , the torus is truncated to a thin disk at a radius @xmath30 , and the x - ray photons from the torus are thought to illuminate the outer line - emitting disk region .
the most favorable candidate for such an ion - supported hot torus is the riaf @xcite .
however , some double - peaked emitters have very luminous x - ray emission , which can not be reproduced solely by the riafs in these sources if a reasonable viscosity parameter is adopted ( see discussion in 2 ) . it is quite doubtful whether riafs are in these x - ray luminous double - peaked line emitters . in fig . 1
, we find that most energy is radiated in the inner edge of the riaf ( more than 70 per cent of the total radiation of the riaf is emitted from the region @xmath135 ) .
our calculation shows that less than 2.3 per cent radiation from the riaf in the inner region of the disk can illuminate the outer disk region with @xmath136 ( see 2 ) , which implies that the radiation of riafs is unable to solve the energy budget problem for most sources even if all the observed luminous x - ray emission in some double - peaked line emitters can be attributed to riafs .
comparison between jet power estimated from the extended radio emission and kinetic luminosity of the jet requires @xmath44 for electron - positron jets , and @xmath43 for electron - proton jets ( e.g. , * ? ? ?
* ) . for electron - proton jets , the inverse - compton scattered photons by the electrons in the jet
are in hard x - ray/@xmath42-ray bands , which are inefficient for photo - ionizing the line - emitting disk region . in our present model ,
the line - emitting disk region is assumed to be irradiated by the scattered x - ray photons from the electron - positron jet .
most photons compton up - scattered by the electrons in the electron - positron jets are in the soft x - ray band , which can efficiently photo - ionize the outer line - emitting disk region .
our calculations show that the jet can be compton thin down to @xmath137 if the jet kinetic luminosity @xmath106 is low , while it can be compton thick even at very large radii for high values of @xmath106 , because much matter is loaded by the jet with high @xmath106 . in our calculations ,
the minimal radius of the jet @xmath126 is adopted . in principle
, we can calculate the cases for a jet down to a radius lower than @xmath126 , though it is still unclear whether the knigl s conical jet is still valid to a very small radius near the black hole .
fortunately , the photons scattered near the bottom of the jet can not photo - ionize the outer line - emitting disk region efficiently , because of the small solid angle subtended to the line - emitting region , which is similar to the riaf case as discussed in 2 .
thus , the final results have not been affected much even if a value of @xmath138 is adopted . in fig .
[ fig3 ] , we find that the emissivities @xmath139 have similar power - law @xmath1-dependence for different lorentz factors @xmath34 adopted .
we find @xmath139 varies with @xmath1 nearly @xmath140 along the disk radius for all cases .
the model fittings on the observed double - peaked line profile requires a power - law line emissivity with an index of @xmath141 for different sources ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , which is consistent with our model calculations .
it is found that the emissivity decreases rapidly with increasing the lorentz factor @xmath142 . for given jet kinetic luminosity , a higher @xmath34 leads to lower electron density ( see eqs .
[ lkin1 ] and [ mdotjet ] ) , which decreases the inverse - compton scattered photons from the jet , and then the emissivity @xmath143 .
the ratios @xmath144 varying with @xmath145 for different values of @xmath34 are plotted in fig .
we plot the data of the sample in fig .
[ fig4 ] to test our model by comparison with the model calculations .
we find that the lorentz factors of the jets @xmath5 are required to explain the observed data for most sources in this sample .
the apparent angular velocity of the brightest jet component in 3c390.3 measured by vlbi is @xmath1460.54 mas / year @xcite .
the lorentz factor of the jet in this source @xmath147 is therefore derived from its inclination angle @xmath148 , which is estimated from the fitting of its double - peaked line profile by @xcite . in fig .
[ fig4 ] , one can find that a jet with @xmath149 is required by the model calculation for this source .
for this source , @xmath150 @xcite , which means that the blamer line emission may be partly attributed to the local viscous heating of the line - emitting disk region , so our result is roughly consistent with that derived from the vlbi observations .
only one source , pictor a , a jet with relatively larger lorentz factor @xmath151 is required to illuminate the line - emitting region .
this is a powerful radio galaxy ( e.g. , * ? ? ?
* ) , and x - ray jet emission is detected in this source @xcite .
the x - ray emission from the jet is important for this source , and the accretion disk luminosity derived from its x - ray luminosity is obviously over - estimated , so the location of this source in fig .
[ fig4 ] should be shifted upwards .
a mild relativistic jet may probably provide sufficient scattered photons for photo - ionizing the line - emitting disk region in this source , if the x - ray emission is dominantly from the jet . for arp 102b , a very low @xmath152 ( @xmath153 ) is required ( see fig .
[ fig4 ] ) , i.e. , a slowly moving conical outflow is present in this source , which is consistent with its compact radio structure @xcite .
another special source , iras 0236.6 - 3101 , which is difficult to be explained by this model ( see fig . [ fig4 ] ) .
this is a star - forming galaxy with very weak radio emission ( 9.5 mjy at 1.4 ghz ) , and it probably has no jet . the origin of its double - peaked lines may be similar to that of radio - quiet double - peaked line emitters , which will be addressed in the last paragraph of this section . for the seven sources with only upper limits on black hole masses ( lower limits on the ratios @xmath106 ) , they are similar to others and the jets with moderate lorentz factors @xmath5 can solve the energy budget " problem . as the detailed model fittings for individual sources indicate that the line - emitting regions in these sources have different sizes ( see * ? ? ? * for details ) , our calculations have been carried out for different values of @xmath80 and @xmath81(see different line types in fig .
[ fig4 ] ) , which have not altered our main conclusion .
there are 13 sources in this sample with @xmath154 .
although it means the local viscous power in the line - emitting disk regions may be sufficient for the line emission for these 13 sources , the temperatures of the line - emitting disk portions are still too low for all the observed h@xmath9 line emission .
the external illuminating from the jets is still necessary for these sources . in our model
, an electron - positron jet is required to scatter disk photons back to photo - ionize the outer disk surface to produce balmer lines .
we suggest that future vlbi polarization observations on these radio - loud double - peaked line emitters can test this model . in this paper , some quantities , @xmath119 , @xmath155 , and @xmath127 ,
are derived from observed quantities by using the conventional approaches reported in literature , of which the uncertainties might affect our results .
however , the qulititative conclusion that slow jets / outflows are required in double - peaked emitters would not be altered ( see fig .
4 ) , for typical errors for these quantities , i.e. , a factor of @xmath156 for @xmath119 , and the derived kinetic luminosity @xmath127 is accurate at an order of magnitude . in our present calculation of @xmath157
, we have not included the thermalization od the disk .
if the thermalization is considered , the resulted @xmath157 will become lower , so a lower bulk lorentz factor @xmath34 of the jet than our present model calculation is required ( see fig .
[ fig4 ] ) . besides radio - loud agns in this sample
, several ten radio - quiet double - peaked line emitters have been discovered @xcite . although no jet is present in these radio - quiet agns , we speculate that slow moving outflows may be in these sources , which play the same role of the jets in their radio - loud counterparts .
if the electrons in the outflows are non - relativistic , a fraction of disk power radiated from its inner region can be thomson scattered to the outer line - emitting portion of the disk .
if the thomson scattering depth is @xmath158 , the outflows can scatter nearly one - tenth of the disk radiation to illuminate the outer line - emitting disk portion .
if this is the case , the absorption by the outflows in these radio - quiet agns is expected to be detected by x - ray observations , which can be a test on our model .
the lack of detected double - peaked broad emission lines in strong radio - loud quasars with relativistic jets is a natural prediction of our model , i.e. , the compton scattering power is too weak to irradiate the outer line - emitting portion for fast moving jets . for radio - quiet agns , only those having outflows with suitable scattering depth can scatter disk photons back to illuminate the outer line - emitting portion , which leads to double - peaked broad emission lines .
recent x - ray observations showed that a fraction of quasars and seyfert galaxies have x - ray absorption with column density @xmath159 ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , which implies that the agns having outflows with suitable scattering depth @xmath158 are not common among all agns .
however , it does not mean that all agns with suitable scattering depth will exhibit double - peaked lines .
some other factors , such as , the viewing angle and the properties of the accretion disk , may play important roles in producing double - peaked lines . the quantitative model calculations to fit the line profiles for individual agns are beyond the scope of this work , which will be reported in the forthcoming paper . in this work
, we suggest that the soft photons scattered by electrons in electron - position jets for radio - loud double - peaked line emitters ( or in slow outflows with suitable scattering depth for radio - quiet counterparts ) can efficiently ionize the line - emitting disk regions .
another possibility is that slow outflows play the same role in radio - loud double - peaked line emitters as radio - quiet sources , though it is still unclear whether relativitic jets can co - exist with such cold outflows in radio - loud double - peaked line emitters .
however , it is unable to explain the fact that no double - peaked emission line is present in strong radio quasars with relativistic jets , while this is a natural prediction of the electron - positron jet model proposed in this work .
antonucci , r. 1993 , , 31 , 473 armitage p.j .
1998 , , 501 , l189 biermann , p. , preuss , e. , kronberg , p. p. , schilizzi , r. t. , & shaffer , d. b. 1981 , , 250 , l49 blandford , r. d. , & knigl , a. 1979 , , 232 , 34 blundell , k.m .
, & rawlings , s. 2000 , , 119 , 1111 caccianiga , a. , march ? , m. j. m. , thean , a. , & dennett - thorpe , j. 2001 , , 328 , 867 cao , x. , 2005 , , 631 , l101 celotti , a. , & fabian , a. c. 1993 , , 264 , 228 chen , k. , & halpern , j. p. 1989
, , 344 , 115 chen , k. , halpern , j. p. , & filippenko , a. v. , 1989 , , 339 , 742 dumont , a. m. , & collin - souffrin , s. 1990 , , 229 , 313 dunlop , j. s. et al .
2003 , , 340 , 1095 eracleous , m. , & halpern , j. p. 1994
, , 90 , 1 eracleous , m. , & halpern , j. p. 2003
, , 599 , 886 falomo , r. , carangelo , n. , & treves , a. 2003 , , 343 , 505 freeman , k. c. , et al .
1977 , , 55 , 445 fukugita , m. , shimasaku , k. , & ichikawa , t. 1995 , , 107 , 945 gaskell , c. m. 1983 , in quasars and gravitational lenses ; proceedings of the twenty - fourth liege international astrophysical colloquium , belgium , universite de liege , p473 gaskell , c. m. 1988 , in active galactic nuclei .
proceedings of a conference held at the georgia state university , atlanta , georgia , eds .
miller , h. r. , & wiita , p. j. , publisher , springer - verlag , berlin , p61 goad , m. , & wanders , i. 1996 , , 469 , 113 hardcastle , m. j. 2005 , , in press , ( astro - ph/0511511 ) hawley , j.f . , & balbus , s.a .
2002 , , 573 , 738 hutter , d. j. , & mufson , s. l. 1986 , , 301 , 50 jiang , d.r . ,
cao , x. , & hong , x.y .
, 1998 , , 494 , 139 kaspi , s. , smith , p.s . ,
netzer , h. , maoz , d. , jannuzi , b.t .
, & giveon , u. 2000 , , 533 , 631 kellermann , k. i. et al . , 2004 , , 609 , 539 knigl , a. 1981 , , 243 , 700 lauberts , a. , & valentijn , e. a. 1989 , the surface photometry catalogue of the eso - uppsala galaxies , garching bei munchen : european southern observatory manmoto , t. , 2000 , , 534 , 734 mclure , r.j . , & dunlop , j.s .
327 , 199 mclure , r.j . , &
dunlop , j.s .
352 , 1390 narayan , r. , & yi , i. 1994 , , 428 , l13 narayan , r. , & yi , i. 1995 , , 444 , 231 pagani , c. , falomo , r. , & treves , a. 2003 , , 596 , 830 perez , e. , mediavilla , e. , penston , m. v. , tadhunter , c. , & moles , m. 1988 , , 230 , 353 piconcelli , e. et al .
, 2005 , , 432 , 15 puschell , j. j. , moore , r. , cohen , r. d. , owen , f. n. , & phillips , a. c. 1986 , , 91 , 751 reeves , j. n. , obrien , p. t. , & ward , m. j. , 2003 , , 593 , l65 renaud , n. , & henri , g. , 1998 , , 300 , 1047 schneider , d. p. , schmidt , m. , & gunn , j. e. 1994 , , 107 , 1245 shakura , n.i . , & sunyaev , r.a .
1973 , , 24 , 337 smith , e. p. , & heckman , t. m. , 1989 , , 69 , 365 stepanian , j. a. et al . 2001 , , 122 , 3361 strateva , i. v. , et al .
2003 , , 126 , 1720 two micron all sky survey team , 2003 , 2mass extended objects , final release wang , t.g . , et al . , 2005 , , 625 , l35 wardle , j. f. c. , homan , d. c. , ojha , r. , & roberts , d. h. , 1998 , nat , 395 , 457 willott , c.j . , rawlings , s. , blundell , k.m . ,
& lacy , m. 1999 , , 309 , 1017 wilson , a. s. , young , a. j. , & shopbell , p. l. 2001 , , 547 , 740 worrall , d.m . , & birkinshaw , m. , 2004 , astro - ph/0410297 wu , x. -b . , & liu , f. k. 2004 , , 614 , 91 zheng , w. , sulentic , j. w. , & binette , l. 1990 , , 365 , 115 zirbel , e. l. 1996 , , 473 , 713 young , a. j. et al . , 2005 , , 631 , 733 ccccccccc 3c 17 & 0.220 & 8.81@xmath160 & 1 & 43.50 & 45.80 & 45.73 & 5.0@xmath161 & 5.92@xmath162 + 4c 31.06 & 0.373 & 9.66@xmath160 & 2 & 43.08 & 46.66 & 45.17 & 2.65@xmath163 & 2.91@xmath161 + 3c 59 & 0.109 & 8.59 & 3 & 42.82 & 46.30 & 44.89 & 3.28@xmath163 & 1.45@xmath162 + pks 0235 + 023 & 0.209 & @xmath1648.91@xmath160 & 4 & 43.85 & 46.68 & 44.95 & 1.48@xmath161 & 7.87@xmath161 + iras 0236.6 - 3101 & 0.062 & 8.83 & 5 & 42.20 & 45.01 & 42.07 & 1.55@xmath161 & 1.25@xmath165 + pks 0340 - 37 & 0.285 & @xmath1648.87 & 2 & 44.22 & 46.65 & 45.50 & 3.71@xmath161 & 3.16@xmath162 + 3c 93 & 0.357 & 8.99 & 6 & 43.20 & 46.59 & 45.86 & 4.14@xmath163 & 5.43@xmath163 + ms 0450.3 - 1817 & 0.059 & 7.63@xmath166 & 7 & 41.28 & 44.69 & 43.20 & 3.91@xmath163 & 2.68@xmath161 + pictor a & 0.035 & 7.62@xmath160 & 8 & 41.77 & 45.76 & 45.31 & 1.02@xmath163 & 3.53@xmath167 + b2 0742 + 31 & 0.462 & @xmath16411.22@xmath160 & 9 & 44.63 & 47.43 & 45.81 & 1.58@xmath161 & 2.87@xmath163 + cbs 74 & 0.092 & @xmath1648.73@xmath160 & 2 & 42.92 & 46.29 & 44.50 & 4.22@xmath163 & 4.28@xmath161 + pks 0921 - 213 & 0.053 & @xmath1647.87@xmath160 & 2 & 42.42 & 45.70 & 43.80 & 5.21@xmath163 & 6.19@xmath161 + pks 1020 - 103 & 0.197 & 8.85 & 10 & 43.54 & 46.86 & 44.61 & 4.83@xmath163 & 4.23@xmath161 + cso 643 & 0.276 & @xmath1649.77@xmath160 & 11 & 43.25 & 47.05 & 44.62 & 1.59@xmath163 & 5.16@xmath163 + 3c 303 & 0.141 & 8.27@xmath160 & 8 & 42.29 & 45.81 & 44.97 & 3.04@xmath163 & 3.269@xmath162 + 3c 332 & 0.151 & 8.53@xmath160 & 8 & 42.46 & 45.79 & 45.20 & 4.63@xmath163 & 3.40@xmath162 + arp 102b & 0.024167 & 8.34 & 12 & 41.87 & 44.46 & 42.54 & 2.61@xmath161 & 1.14@xmath163 + 3c 382 & 0.05787 & 9.38@xmath160 & 1 & 43.07 & 46.78 & 44.64 & 1.93@xmath163 & 1.32@xmath161 + 3c 390.3 & 0.0561 & 8.59 & 13 & 42.57 & 46.14 & 44.87 & 2.69@xmath163 & 1.38@xmath162 + pks 1914 - 45 & 0.368 & @xmath16410.37@xmath160 & 9 & 43.75 & 46.24 & 45.83 & 3.23@xmath161 & 2.11@xmath161 + | the double - peaked broad emission lines are usually thought to be linked to accretion disks , however , the local viscous heating in the line - emitting disk portion is usually insufficient for the observed double - peaked broad - line luminosity in most sources . it was suggested that the x - ray radiation from an ion - supported torus in the inner region of the disk can photo - ionize the outer line - emitting disk region .
however , our calculations show that only a small fraction ( @xmath0 2.3 per cent ) of the radiation from the radiatively inefficient accretion flow ( riaf ) in the inner region of the disk can photo - ionize the line - emitting disk portion , because the solid angle of the outer disk portion subtended to the inner region of the riaf is too small . we propose a physical model for double - peaked line emitters , in which only those agns with sufficient matter above the disk ( slowly moving jets or outflows ) can scatter enough photons radiated from the inner disk region to the outer line - emitting disk portion extending from several hundred to more than two thousand gravitational radii .
our model predicts a power - law @xmath1-dependent line emissivity @xmath2 , where @xmath3 , which is consistent with @xmath4 required by the model fittings for double - peaked line profiles . using a sample of radio - loud agns with double - peaked emission lines ,
we show that the outer disk regions can be efficiently illuminated by the photons scattered from slow or mild relativistic electron - positron jets with @xmath5 .
it is consistent with the fact that no double - peaked emission line is present in strong radio quasars with relativistic jets . for radio - quiet double - peaked line emitters , slow outflows with thomson scattering depth @xmath6 instead of jets can scatter sufficient photons to ( illuminate ) the line - emitting regions .
this model can therefore solve the energy budget problem for double - peaked line emitters . |
it is well known that the equation satisfied by a world - line @xmath1 of a massive charged particle in a @xmath0-dimensional riemannian space - time in presence of an electromagnetic field is given by @xmath2 in this expression , @xmath3 and @xmath4 are the mass and the charge , respectively , of a test particle , @xmath5 is the @xmath0-dimensional proper time and the electromagnetic field @xmath6 is defined by @xmath7 where @xmath8 is the vector potential .
the @xmath9 are the @xmath0-dimensional christoffel symbols .
the second term of the above equation represents the inertial force whereas the last term is the lorentz force .
the study of geodesics in multidimensional theories of kaluza - klein type has been performed by many authors @xcite - @xcite in quite an exhaustive manner .
it has been known since the very beginning @xcite that , in the simplest version of the theory ( without scalar field ) , the geodesic equation in @xmath10 dimensions coincides with the geodesic equation in @xmath0 dimensions with an extra term which can be identified with the lorentz force .
indeed , in @xmath10 dimensions , the geodesic equation is given by @xmath11 where @xmath12 is the @xmath10-dimensional interval length , and the brackets @xmath13 are the @xmath10-dimensional christoffel symbols .
projecting this equation under the assumption that all the quantities are independent of the fifth coordinate , @xmath14 , it is easy to establish , using the formulae displayed in the appendix , that one obtains @xmath15 the second equation , telling us that the quantity @xmath16 is constant along the @xmath10-dimensional geodesics parametrized by @xmath12 , reflects the fact that the backgrounds have been chosen such that @xmath17 is a cyclic co - ordinate . since we have the following relation between the squares of the intervals in @xmath10 and @xmath0 dimensions : @xmath18 which amounts to @xmath19 , we see that the @xmath0-dimensional equation ( [ geo4 ] ) is recovered provided that we make the following identification : @xmath20 this means that we suppose that @xmath21 , so that @xmath22 is timelike whenever @xmath23 is timelike , and vice versa . a more general situation is discussed in detail in @xcite .
the non - abelian generalization has been considered in @xcite and @xcite .
since the multidimensional theories of kaluza - klein type are constructed as copies of einstein s general relativity theory in more than @xmath0 dimensions , all the usual mathematical corollaries remain valid .
for example , it is possible to cancel the christoffel symbols along a given geodesic curve by an appropriate choice of coordinates , which amounts to the annulation of forces acting on a test particle moving along that geodesic line . in @xmath0 dimensions ,
the cancellation along the worldline of the @xmath10-dimensional christoffel symbols may be interpreted as the simultaneous compensation of gravitational and the lorentz forces by an appropriate acceleration field .
it is also a well known fact that such cancellation can be performed along only one given geodesic line at once , but not necessarily along all its neighbors .
this fact becomes obvious when one looks at the _ geodesic deviation _
equation @xmath24 where @xmath25 is an infinitesimal `` geodesic deviation '' vector , and @xmath26 denotes the pull - back of covariant derivatives along the time - like geodesics . for a massive _ and _ charged test particle in the presence of both gravitational and electromagnetic fields ( @xmath27 and @xmath28 )
, it is not difficult to derive the generalized world - line deviation equation by taking direct variation of the world - line equation ( [ geo4 ] ) : @xmath29,\ ] ] here the riemann tensor appears explicitly , making it automatically impossible to cancel its influence by any local or global coordinate or gauge transformations .
this is why the study of the geodesic deviation in multidimensional theories , which reads @xmath30 is of particular interest . indeed ,
when explicited in the form that splits up the @xmath0-dimensional space - time and the @xmath31-dimensional internal space , _ a priori _ , new terms show up , containing quadratic expressions of the type @xmath32 appearing in the @xmath10-dimensional riemann tensor ( cf .
appendix ) , that can not be foreseen or derived from a purely @xmath0-dimensional point of view , even if one tries to introduce the interaction of charges with gauge and scalar fields .
it will fix in a canonical way the terms describing the purely gravitational influence of those fields , which by their energy density must influence the trajectories of chargeless massive particles , too , provoking tidal effects which should deform the initially parallel geodesic lines .
= n y hence the question arises whether calculation of the geodesic deviation after reduction to 4 dimensions yields the same result as projecting the 5-dimensonal geodesic deviations to 4 dimensions , rendering the diagram of fig . 1 commutative .
this is of course an important issue since any difference could be used to discriminate a purely @xmath0-dimensional theory from a kaluza - klein approach .
the purpose of this short technical note is to address this question . to our knowledge
, the analysis of this problem can not be found in the existing literature , and we believe that the present study will close this gap , and in addition , will shed some new light on the interplay between the gauge fields and gravitation , and on the interpretation of the equivalence principle in multidimensional theories as well .
we consider here the version of the @xmath10-dimensional kaluza - klein theory in which the scalar field @xmath34 is put equal to @xmath35 from the beginning , which makes the field equations arising from the variational principle in @xmath10 dimensions strictly equivalent to the einstein - maxwell system .
we shall perform all the calculations in a holonomous coordinate system in order to make the interpretation of the geodesic equations and the affine parameters as straightforward as possible .
the space - time components obey the equation @xmath36
\frac{{\rm d } x^{\nu}}{{\rm d } s } \frac{{\rm d } x^{\rho } } { { \rm d } s } \delta x^{\kappa } - \frac{q}{2}[\nabla_{\kappa } f_{\rho}{}^{\mu } + \nabla ^{\mu } f_{\rho \kappa}]\frac{{\rm d } x^{\rho } } { { \rm d } s } \delta x^{\kappa } - \frac{1}{2}(a_{\kappa}\delta x^{\kappa } + \delta x^5 ) ( \nabla_{\rho } f^{\mu}{}_{\nu } ) \frac{{\rm d } x^{\rho } } { { \rm d } s } \frac { { \rm d } x^{\nu } } { { \rm d } s } \nonumber \\ & & -\frac{q^2}{4}f^{\mu}{}_{\lambda}f_{\rho } { } ^{\lambda}\delta x^{\rho } + \frac{q}{4}(a_{\kappa}\delta x^{\kappa } + \delta x^5 ) f^{\mu } { } _ { \lambda}f_{\rho}{}^{\lambda } \frac{{\rm d } x^{\rho } } { { \rm d } s } \label{geomu}\end{aligned}\ ] ] the equation ( [ geomu ] ) still is not explicit enough to be solved as function of the space - time variables .
the derivatives with respect to @xmath10-dimensional interval @xmath22 should be replaced by the derivatives with respect to the @xmath0-dimensional proper time @xmath37 .
next , the second - order covariant derivative appearing on the right - hand side and containing the @xmath10-dimensional connection coefficients and their partial derivatives , has to be expressed in terms of ordinary derivatives and @xmath0-dimensional christoffel symbols along with the gauge - invariant quantities containing the faraday tensor @xmath38 .
the relation between the covariant derivations with respect to the parameters @xmath22 ( the @xmath10-dimensional line element ) and @xmath37 ( particle s proper time in @xmath0 space - time dimensions ) is quite complicated when it comes to the second - order covariant derivatives .
this is why we omit all the intermediary calculations , giving here the final result .
the second covariant derivatives are related as follows for the @xmath0-dimensional components of a given @xmath10-dimensional vector @xmath39 : @xmath40.\end{aligned}\ ] ] combining equations ( [ geomu ] ) and ( [ der54 ] ) and using homogeneous maxwell s equations , @xmath41 and the identification of the physical charge - to - mass ratio ( [ identify ] ) , one obtains @xmath42 + f^{\mu } { } _ { \lambda } \frac{{\rm d } x^{\lambda } } { { \rm d } \tau } \biggl[\frac{{\rm d}}{{\rm d}\tau } ( a_{\kappa}\delta x^{\kappa } + \delta x^5)+f_{\kappa \rho } \frac{{\rm d } x^{\kappa } } { { \rm d } \tau } \delta x^{\rho } \biggr].\ ] ] this equation would coincide with the usual @xmath0-dimensional deviation equation ( [ devi4f ] ) if it were not for the last term , which contains the standard lorentz force multiplied by the expression in square brackets , linear in the infinitesimal deviation vector . however , it is easily recognized that the last term just represents the deviation of the 5th component of the momentum @xmath43 , which by equation ( [ eom ] ) is conserved , and which we identified with the charge through equation ( [ identify ] ) : @xmath44.\ ] ] it is noteworthy that the same result can be obtained by using the simpler , but non - covariant form of the geodesic deviation equation ( [ devi5 ] ) : @xmath45 which makes the calculations much less tedious .
the fifth component of the previous equation leads to @xmath46 = 0.\ ] ] which means that @xmath47 is indeed a constant . from a purely mathematical point of view
, this constant can take on any real value , depending on the arbitrary choice of _ initial conditions _
, which include the initial values of _ ten _ variables , @xmath48 and @xmath49 ( 0)$ ] , as it is the case for any system of five ordinary differential equations of second order .
the fact that @xmath47 is a constant means that not all the initial data can be independent . as a matter of fact , the first derivative of the fifth component of the deviation , @xmath50 ( 0)$ ] , is an imposed function of the four - dimensional initial deviations , namely @xmath51(0).\ ] ] requiring that @xmath52 is the condition which must be imposed if we want to maintain a one - to - one correspondence between the geodesic deviation equation in @xmath10-dimensional kaluza - klein space and the usual deviation equation in presence of the electromagnetic field in @xmath0 dimensions .
now , returning to equation ( [ inter ] ) , we get the final result that can be stated very simply as follows : 0.2 cm _ the space - time projection of the five - dimensional kaluza - klein geodesic deviation equation yields for fixed @xmath43 the usual four - dimensional world - line deviation equation in the presence of both gravitational and electromagnetic fields , for particles of the same @xmath53 ; geodesic deviations between five - dimensional worldlines with different values of @xmath43 describe the four - dimensional deviation of world - lines for particles with different values of @xmath53 . _ 0.2 cm
the only influence of the electromagnetic fields on chargeless particles comes through the term linear in the @xmath0-dimensional riemann tensor , which is a solution of the coupled einstein - maxwell equations ( see ref .
although the final answer to the problem of the projection of the geodesic deviation equation from the 5-dimensional kaluza - klein metric space onto its 4-dimensional space - time basis is very simple and does not bring any surprise , it is worth to be checked ( the above calculations have never been published elsewhere , at least to our knowledge ) , and does not seem to be totally trivial . it can be interpreted as a strong _ equivalence principle _
generalized to the @xmath10-dimensional theory incorporating electromagnetism into geometry .
our result can be easily generalized to the non - abelian case @xcite , @xcite , where the conservation of charge @xmath43 is replaced by a condition on the rotation of the charge isovector in the lie algebra space in which it takes its value .
there is no guarantee that the higher - order deviation equations , obtained with a similar technique of new independent variation of einstein - maxwell equations do also project properly , although such a statement seems very plausible .
however , the introduction of the _ dilaton field _ , i.e. supposing that the radius of the compactified 5-th dimension depends on the space - time position @xmath54 , may bring new effects leading to certain anomalies in the deviation equation and its projection onto the usual space - time . one can apply a similar technique of probing the deviations to the equations of motions of the @xmath55 embedded in multi - dimensional spaces , which represent a natural generalization of geodesic curves in kaluza - klein theories .
these developments should become the object of another independent study .
the @xmath10-dimensional metric tensor of the theory reads ( @xmath56 ) : @xmath57 where @xmath58 and @xmath59 , which means that we consider that both @xmath60 and @xmath61 do not depend on the fifth coordinate @xmath62 . here
are the christoffel symbols of the metric ( [ 5metric ] ) : @xmath63 with this in mind , we can proceed with the computation of the components of @xmath10-dimensional riemann tensor in a holonomous system . using the convention : @xmath64 and
after some calculus , using also the bianchi identities satisfied by the tensor @xmath38 , we get : @xmath65 here the greek indices @xmath66 are raised and lowered by means of the @xmath0-dimensional metric tensors @xmath67 and @xmath68 .
a. balakin , j .- w .
van holten , r. kerner , _
`` motions and world - line deviations in einstein - maxwell theory '' _ , preprint lgcr-2000/08/2 and nikhef/00 - 023 , ( submitted to class . and quantum gravity
; also available as electronic preprint gr - qc/0009016 ) | we study in detail the equations of the geodesic deviation in multidimensional theories of kaluza - klein type .
we show that their @xmath0-dimensional space - time projections are identical with the equations obtained by direct variation of the usual geodesic equation in the presence of the lorentz force , provided that the fifth component of the deviation vector satisfies an extra constraint derived here . |
let @xmath2 be a uniformly strongly elliptic differential operator on @xmath3 ( @xmath4 ) @xmath5 with real bounded smooth coefficients with bounded derivatives , satisfying @xmath6 and @xmath7 with @xmath8 .
we denote by @xmath9 the maximal realization in @xmath10 ; it is selfadjoint positive .
let @xmath11 be the exterior of a bounded smooth open set @xmath12 , with boundary denoted @xmath13 ( @xmath14 ) , and let @xmath15 , @xmath16 and @xmath17 be the selfadjoint lower bounded realizations in @xmath18 determined by the dirichlet condition ( @xmath19 ) , the oblique neumann condition ( @xmath20 , see below ) , resp .
a robin condition ( @xmath21 ) with @xmath22 real and smooth .
the coefficient @xmath23 is assumed to be taken so large positive that all four operators have positive lower bound .
it is known that the operators @xmath24 have an unbounded essential spectrum , consisting of an interval @xmath25 if the coefficients converge to a limit for @xmath26 , and more generally being a subset of @xmath25 with possible gaps ( e.g. when the coefficients are periodic ) .
birman showed in @xcite a general principle concerning the stability of the essential spectrum : @xmath27 where @xmath28 denotes the class of compact operators whose characteristic values @xmath29 are @xmath30 for @xmath31 .
( when @xmath32 is a disjoint union of open sets , and @xmath33 acts in @xmath34 , we denote by @xmath35 the operator in @xmath36 that acts like @xmath33 on @xmath34 , naturally injected in @xmath36 . ) in particular , all four operators have the same essential spectrum @xmath37 ; this extends a result of povzner , as referred to in @xcite .
( birman s paper also allowed unbounded coefficients and limited smoothness , but we shall not follow up on those aspects here . )
the result was refined by birman and solomiak in @xcite , where a weyl - type spectral asymptotics formula was obtained ( @xmath38 converges to a limit for @xmath39 ) . in grubb @xcite similar spectral asymptotics formulas were shown by methods of pseudodifferential boundary problems , and refinements with a spectral resolvent parameter were studied in @xcite .
spectral estimates of resolvent differences have been taken up again in recent works of alpay and behrndt @xcite , gesztesy and malamud @xcite .
the present paper extends the results to higher - order operators , but aims in particular for a slightly different question , namely of how much one can perturb the essential spectrum of @xmath17 by replacing the robin condition by a more general _
neumann - type _ boundary condition ( not necessarily elliptic ) @xmath40 let @xmath41 denote the realization of @xmath2 on @xmath42 determined by , i.e. , with domain @xmath43 the outcome is as follows : \1 ) for any nonzero @xmath44 , @xmath45 can be chosen as a pseudodifferential operator of order @xmath46 such that @xmath41 is selfadjoint with @xmath47 more generally , when @xmath48 is an invertible selfadjoint operator in a separable infinite dimensional hilbert space @xmath49 , one can choose an operator @xmath45 such that @xmath41 is selfadjoint and @xmath50 \2 ) for any choice of an operator @xmath45 in defining a selfadjoint invertible realization @xmath41 , @xmath37 remains in the essential spectrum of @xmath41 .
we also reprove the spectral asymptotics formulas , and extend the results to strongly elliptic operators of order @xmath1 for positive integer @xmath51 .
the question of whether points of @xmath52 can be removed by a perturbation of the boundary condition was brought up in a conversation with m. marletta , m. brown and i. wood in cardiff in may 2008 ; the author thanks these colleagues for useful discussions .
let us first recall some well - known facts .
the sobolev space @xmath53 ( @xmath54 ) can be provided with the norm @xmath55 ; here @xmath56 is the fourier transform and @xmath57 .
there is a standard construction from this of sobolev spaces over an open subset and over the boundary manifold .
we denote by @xmath58 resp .
@xmath59 the operators acting like @xmath2 with domains @xmath60 here @xmath59 is closed symmetric , and @xmath61 .
the operators @xmath41 satisfying @xmath62 are called the realizations of @xmath2 . the symmetric sesquilinear forms @xmath63 are bounded on @xmath64 resp .
@xmath65 and satisfy @xmath66 there , with @xmath67 .
moreover , @xmath68 where @xmath69 with @xmath70 denoting the interior unit normal to @xmath42 at @xmath71 .
hence the standard variational construction ( the lax - milgram lemma ) applied to the triples @xmath72 , @xmath73 , resp .
@xmath74 defines the positive selfadjoint operators @xmath9 in @xmath75 , @xmath15 and @xmath16 in @xmath18 mentioned in the introduction .
( the variational construction is known e.g. from lions and magenes @xcite , and is also explained in grubb @xcite . ) in view of elliptic regularity theory and the uniform symbol estimates , the domains are in fact contained in @xmath76 .
moreover , the operators representing the nonhomogeneous boundary value problems ( cf .
e.g. @xcite ) @xmath77 where @xmath78 resp .
@xmath79 , have solution operators , continuous in the opposite direction : @xmath80 in modern terminology , @xmath81 where @xmath82 is the _ pseudodifferential operator _ @xmath83 on @xmath0 and @xmath84 is its truncation to @xmath85 ( here @xmath86 extends to @xmath3 by 0 on @xmath87 , @xmath88 restricts from @xmath3 to @xmath85 ) .
the operators @xmath89 and @xmath90 are _ poisson operators _ solving the respective boundary value problems with nonzero boundary data , zero data in the interior of @xmath42 ; their mapping properties extend to the full scale of sobolev spaces with @xmath91 .
@xmath92 and @xmath93 act in @xmath18 as the inverses of the realizations @xmath15 resp .
@xmath16 of @xmath2 with domains @xmath94 the operator @xmath17 representing the robin condition @xmath95 is defined similarly from the sesquilinear form @xmath96 on @xmath65 and has similar properties as @xmath16 : its domain is @xmath97 , and the operator @xmath98 the above facts have been known for many years , although the emphasis was not always placed on including low values of @xmath99 . instead of accounting for this aspect in detail here , we mention that the results are covered by the construction in the book grubb @xcite , chapter 3 , and that the general @xmath1-order case will be treated below in section [ higher ]
. we shall now regard the realization defined by from the point of view of general nonlocal boundary value problems .
the basic theory was presented in grubb @xcite and was taken up again and further developed in a joint work with brown and wood @xcite ; applications to exterior domains are included in @xcite .
( an introduction is also given in @xcite . )
the fundamental result is that the closed realizations @xmath41 are in a 11 correspondence with the closed , densely defined operators @xmath100 , where @xmath101 and @xmath102 are closed subspaces of @xmath103 , the nullspace of @xmath104 .
many properties are carried along in this correspondence , for example , @xmath41 is invertible if and only if @xmath105 is so , and in the affirmative case one has the kren - type formula @xmath106 where @xmath107 denotes the injection @xmath108 and @xmath109 denotes the orthogonal projection onto @xmath101 , in @xmath110 .
we have here taken the dirichlet realization @xmath15 as the reference operator for the correspondence theorem .
consider in particular a realization @xmath41 corresponding to an operator @xmath111 ( i.e. , with @xmath112 ) .
as shown in the mentioned references , @xmath41 can be interpreted as representing a boundary condition . to describe that boundary condition ,
we first recall that implies the green s formula valid for @xmath113 , @xmath114 we denote by @xmath115 the restriction of @xmath116 to @xmath103 , @xmath117 with adjoint @xmath118 ( recall that for @xmath119 , @xmath120 identifies with the antidual ( conjugate dual ) space @xmath121 of @xmath122 , with a duality consistent with the scalar product in @xmath123 ) .
moreover , we set @xmath124 here @xmath125 is a first - order elliptic pseudodifferential operator over @xmath13 , and @xmath126 is a ( nonlocal ) trace operator .
there holds a generalized green s formula for all @xmath127 , @xmath128 where @xmath129 indicates the ( sesquilinear ) duality pairing between @xmath130 and @xmath131 . the boundary condition that @xmath41 represents
is then found to be @xmath132 where @xmath133 is the closed , densely defined operator from @xmath134 to @xmath135 defined from @xmath105 by @xmath136 since @xmath137 , the condition can also be written @xmath138 so it is of the form with @xmath45 acting like @xmath139 . to sum up : [ proposition2.1 ] when @xmath41 corresponds to @xmath111 , it equals the realization defined by the neumann - type boundary condition , where @xmath140 assume in the following that @xmath141 , equivalently @xmath105 has a bounded , everywhere defined inverse @xmath142 , and @xmath133 has a bounded everywhere defined inverse @xmath143 . then takes the form : @xmath144 here @xmath145 is the poisson operator for the dirichlet problem ( cf . ) , considered as a mapping from @xmath134 to @xmath18 ( also equal to @xmath146 ) ; its adjoint @xmath147 goes from @xmath18 to @xmath148 .
the formula can clearly be used to examine @xmath149 as a perturbation of @xmath150 ; we pursue this fact below in our analysis of essential spectra .
[ remark2.2 ] we are interested in cases where @xmath105 has an essential spectrum outside of 0 . as a specific example , one can think of @xmath151 its essential spectrum is @xmath152 , since @xmath153 . in this case
, @xmath154 is a pseudodifferential operator elliptic of order @xmath155 , and invertible .
( this is in contrast to those boundary conditions that satisfy the shapiro - lopatinski condition ; they have @xmath133 elliptic of order @xmath156 . ) since this @xmath133 is defined on all of @xmath134 , which is mapped by @xmath157 to @xmath158 , @xmath45 maps @xmath159 into @xmath158 ; it is only the difference @xmath160 that is assured to map into @xmath148 .
the realization @xmath41 defined by this choice has @xmath161 , so @xmath162 is not contained in @xmath163 for any @xmath164 .
it is a variant of kren s `` soft extension '' .
for the analysis of the operators on exterior domains we shall need to study cutoffs , by multiplication either by a smooth function or by a `` rough '' characteristic function supported at a distance from the boundary . in @xcite ,
smooth cutoffs were used and the exterior singular green operators estimated by a commutator argument based on a series of nested cutoff functions . we shall here give a simpler argument based on rough cutoffs .
let @xmath165 be a smooth open subset of @xmath42 such that @xmath166 , and denote @xmath167 .
so @xmath168 . we denote by @xmath169 resp .
@xmath170 the restriction operators from @xmath42 to @xmath165 resp .
@xmath171 , and by @xmath172 resp .
@xmath173 the extension operators extending a function given on @xmath165 resp.@xmath171 to a function on @xmath42 by zero on the complement in @xmath42 . in the following we draw on the analysis of singular numbers of compact operators as presented in gohberg and kren @xcite .
the operators lying in the intersection of schatten classes @xmath174 ( also equal to @xmath175 ) will be called _
spectrally negligible_. [ proposition3.1 ] let @xmath89 be the poisson operator entering in , continuous from @xmath176 to @xmath177 for all @xmath54 , and consider the operators @xmath178 and @xmath179 .
then @xmath180 in fact maps continuously @xmath181 moreover , the operators @xmath182 and @xmath183 are compact and spectrally negligible .
similar statements hold for @xmath184 and @xmath185 for @xmath186 .
denote by @xmath187 the operator restricting to @xmath188 .
when @xmath189 , it follows by the interior regularity for solutions of the dirichlet problem for @xmath2 on @xmath42 that @xmath190 . then @xmath191 is a null - solution of the dirichlet problem for @xmath2 on @xmath165 with @xmath192 boundary value .
this will also hold if @xmath193 , any @xmath194 .
we know from the variational theory and regularity theory for the dirichlet problem on @xmath165 that a null - solution with @xmath192 boundary value lies in @xmath195 for any @xmath196 ; hence holds .
it follows by duality that @xmath197 here @xmath198 when @xmath199 ( generally it equals the space @xmath200 of distributions in @xmath201 supported in @xmath202 ) .
taking @xmath203 , we see that @xmath204 so since @xmath13 is compact , this operator is compact ( from @xmath205 to @xmath206 , any @xmath207 ) , and lies in @xmath208 , i.e. is spectrally negligible .
then @xmath182 is compact from @xmath209 to @xmath210 for any @xmath99 , in particular for @xmath211 , so @xmath212 is compact in @xmath213 ) , and hence @xmath214 is compact . in view of the identity @xmath215 , all @xmath216 , all four operators are spectrally negligible . the proofs for @xmath217 and @xmath218 follow the same pattern .
[ corollary3.2 ] let @xmath219 be such that @xmath220 on a neighborhood of @xmath221 .
then the operators @xmath222 from @xmath223 to @xmath224 for @xmath225 , resp
. from @xmath226 to @xmath224 for @xmath186 , map continuously @xmath227 and are spectrally negligible .
we can write @xmath228 , where proposition 3.1 applies to @xmath229 , and @xmath230 is bounded from @xmath195 to @xmath231 , any @xmath196 .
[ corollary3.3 ] consider the singular green operators @xmath232 as in , with @xmath233 for @xmath234 as in corollary 3.2 , the operators @xmath235 are spectrally negligible .
this follows since @xmath236 is bounded from @xmath237 to @xmath238 for @xmath225 , and from @xmath237 to @xmath148 for @xmath186 , and the @xmath239 map into @xmath192 and are spectrally negligible by corollary 3.2 .
[ remark3.4 ] the proofs given above rely on the solvability properties of the exterior problems for @xmath2 .
the properties can also be inferred from a general principle shown in @xcite , lemma 2.4.8 , on cutoffs of poisson operators , prepared for the definition on admissible manifolds ( which include exterior domains ) .
moreover , the lemma deals with a parameter - dependent pseudodifferential boundary operator calculus , including a spectral parameter @xmath240 . in this
setting , when we consider the poisson operator family @xmath241 for @xmath242 , @xmath243 on a ray @xmath244 in @xmath245 , it is of _ regularity _ @xmath246 .
lemma 2.4.8 then implies that @xmath247 is of order @xmath248 and regularity @xmath249 , hence maps @xmath250 for all @xmath251 .
then @xmath252 maps @xmath253 for all @xmath254 .
( the @xmath255-norms are based on the definition of the norm on @xmath256 , namely @xmath257 . ) from this we can conclude both corollary 3.2 and the fact that any schatten norm of @xmath258 is @xmath259 ( any @xmath260 ) for @xmath261 on the ray , as first shown in @xcite .
proposition 3.1 follows from this if we replace @xmath234 by @xmath262 supported in @xmath263 and equal to 1 on a neighborhood of @xmath12 ; then @xmath264 . also here
we get the rapid decrease in @xmath243 of the schatten norms .
we use the results first to reprove the theorems of birman @xcite and birman - solomiak @xcite with a slight elaboration , essentially as in @xcite , @xcite .
[ theorem3.5 ] for @xmath265 , let @xmath266 then @xmath267 moreover , there are spectral asymptotics formulas for @xmath39 : @xmath268 where the constants are determined from the principal symbols . here
@xmath269 is the constant in the spectral asymptotics formula for @xmath270 , namely @xmath271 , defined from the principal symbol of @xmath9 on @xmath12 .
we use the notation in ff . and corollary 3.3 ; in particular , @xmath272 .
it is well - known that @xmath273 is compact , with the asserted spectral asymptotics .
consider first @xmath274 ; in view of it can be written @xmath275 let @xmath234 be as in corollary 3.2 and let @xmath276 , supported in a smooth bounded set @xmath277 and with @xmath278 on a neighborhood of @xmath279 .
we can rewrite @xmath280 as follows : @xmath281 here the first term is a singular green operator on @xmath282 to which the calulus for bounded domains can be applied , and the two other terms are spectrally negligible .
in fact , @xmath283 is so by corollary 3.2 , and for @xmath284 we can use that it maps @xmath53 continuously into @xmath285 for all @xmath99 and @xmath196 , since @xmath286 so that the operator is of order @xmath287 .
then since @xmath288 is bounded , the operator is spectrally negligible , and so are its compositions with bounded operators .
the same arguments apply to @xmath289 , so we find that @xmath290 where @xmath291 is spectrally negligible and the first term is a singular green operator in @xmath292 . to the first term we apply @xcite th . 4.10 , which shows that this term is in @xmath293 and satisfies a spectral asymptotics formula as in ; these facts are preserved when the spectrally negligible term @xmath291 is added on .
this shows the assertions for the @xmath274 .
the treatments of @xmath294 in @xcite ( with misprints ) and @xcite are a bit more complicated in their use of commutators and nested sequences of cutoff functions .
next , consider @xmath295 . here ,
since @xmath296 and @xmath297 , @xmath298 for @xmath299 we proceed as in @xcite , th . 5.1 : consider @xmath300 the second term acts like @xmath301 , where @xmath302 is the `` leftover operator '' for the composition of @xmath273 with @xmath273 ; it is a singular green operator and has a spectral asymptotics formula with exponent @xmath303 , by @xcite th . 4.10 .
( it was in the quoted paper that the analysis of leftover operators in terms of @xmath304 and @xmath305 was first introduced . )
the first term in identifies similarly with a leftover operator on @xmath306 , hence a singular green operator , but since @xmath42 is unbounded , we need more argumentation to show that it is a compact operator with the desired spectral asymptotics . with @xmath234 and @xmath307 as above , we can write it : @xmath308 here @xmath284 is spectrally negligible as noted above , and its adjoint @xmath309 is likewise spectrally negligible .
so @xmath310 is the sum of a spectrally negligible part and @xmath311 , a singular green operator in @xmath312 .
thus @xmath313 plus spectrally negligible terms , so it follows from @xcite th . 4.10
that @xmath314 has a spectral asymptotics behavior @xmath315 , and then @xmath316 satisfies @xmath317 .
we still have to include the term @xmath318 , but the nontrivial part was already treated further above , and is seen to have a similar spectral asymptotics behavior . adding all contributions and using the rules for @xmath99-numbers
, we find that @xmath319 . for @xmath320
, we simply use that @xmath321 where perturbation formulas as in @xcite show that the spectral asymptotics formula for @xmath273 dominates the behavior .
one could moreover give remainder estimates ( as done in @xcite ) .
[ remark3.6 ] the estimates also hold when @xmath22 for @xmath17 is replaced by a first - order differential operator @xmath322 such that the realization is elliptic and invertible .
related results are found for @xmath323 , which is a singular green operator on @xmath42 of the form of a sum of poisson operators composed with trace operators ; this leads to asymptotic estimates for all positive integers @xmath260 : @xmath324 for @xmath39 .
we shall now investigate the question of perturbations of the essential spectrum .
when @xmath325 , we also write @xmath326 , @xmath327 . let us rewrite the action of @xmath149 on @xmath328 in terms of its action on the parts @xmath329 and @xmath330 , with matrix notation .
when @xmath331 , we have that @xmath332 recalling , we shall decompose the operators @xmath150 and @xmath333 in a similar way .
for @xmath150 we have : @xmath334 the entries in the first matrix are compact in @xmath335-norm , since @xmath336 maps @xmath337 into @xmath338 and @xmath339 maps @xmath210 into @xmath340 , where the injection @xmath341 is compact , and @xmath342 is the adjoint of @xmath343 in @xmath18 . since a compact perturbation leaves the essential spectrum invariant , the second matrix has the same essential spectrum as @xmath150 , and we know from theorem [ theorem3.5 ] that this equals @xmath344 . in other words , @xmath345 where @xmath346 is compact in @xmath18 and @xmath347 .
next , we write @xmath348 in the last matrix , every nonzero element is the composition of a bounded operator with either @xmath180 or @xmath349 , hence is spectrally negligible in view of proposition [ proposition3.1 ] .
so this whole matrix is spectrally negligible . in other words , @xmath350 where @xmath351 is spectrally negligible . in particular
, @xmath352 has the same essential spectrum as @xmath353 .
recall furthermore that @xmath354 where @xmath103 is infinite dimensional , hence @xmath355 adding and , setting @xmath356 , and observing that @xmath357 , @xmath358 ( @xmath9 and @xmath41 are unbounded operators ) , we conclude : [ theorem4.1 ] let @xmath41 be as in proposition [ proposition2.1 ] , and assume that @xmath359 . then @xmath149 can be written as the sum of a compact operator @xmath360 in @xmath18 and an operator decomposed into a part acting in @xmath337 and a part acting in @xmath210 : @xmath361 here @xmath362 and hence @xmath363 since the essential spectrum of @xmath41 itself is the reciprocal set of the nonzero essential spectrum of @xmath149 , we also have : [ corollary4.2 ] when @xmath41 is as in theorem 4.1 , @xmath364 in particular , @xmath365 contains all points of @xmath366 , and the points in @xmath367 can not be removed from @xmath365 .
the statements in the introduction follow : in case 1 ) we take @xmath105 as in remark [ remark2.2 ] of section [ description ] in order to add a point @xmath152 ; when @xmath105 acts like @xmath368 , @xmath45 acts like @xmath369 .
a general choice of a selfadjoint invertible @xmath48 in a separable infinite - dimensional hilbert space @xmath49 gives rise to a selfadjoint invertible @xmath105 in the hilbert space @xmath103 with the same essential spectrum , by use of a unitary operator from @xmath49 to @xmath103 .
the statement in 2 ) follows since we have covered all possibilities for @xmath105 in the case of neumann - type boundary conditions .
similar results can be shown for higher order elliptic operators .
the selfadjoint strongly elliptic even - order case is the natural generalization of the case considered in the preceding sections ; here there is a solvable dirichlet problem , and a selfadjoint invertible realization defined by another boundary condition can be related to the dirichlet realization by a kren - type formula generalizing as in @xcite .
invertible realizations exist in greater generality , though , so to save later repetitions , we consider to begin with a more general class assuring existence of resolvents @xmath370 at least when @xmath243 is large , lying in a suitable subset of @xmath371 .
we take for @xmath2 an elliptic operator @xmath372 of order @xmath1 , @xmath51 integer , with complex @xmath192 coefficients on @xmath3 that are bounded with bounded derivatives and with the principal symbol @xmath373 satisfying ( with @xmath374 ) @xmath375 uniform strong ellipticity . here
@xmath376 , @xmath377 .
denote by @xmath9 the maximal realization of @xmath2 in @xmath10 ; the uniform ellipticity implies that @xmath378 .
@xmath2 satisfies a grding inequality , for which we include a quick proof : [ lemma5.1 ] there are constants @xmath379 and @xmath380 such that @xmath381 using the calculus of globally estimated pseudodifferential operators as in @xcite sect .
18.1 , and @xcite , we can write @xmath382 where @xmath383 , @xmath15 is of order 0 with principal symbol @xmath384 satisfying @xmath385 @xmath386 is of order 0 with principal symbol @xmath387 , and @xmath322 is of order @xmath155 .
since @xmath386 is elliptic , it has a parametrix @xmath82 of order 0 so that @xmath388 is of order @xmath155 ; hence @xmath389 for @xmath390 ( the schwartz space of rapidly decreasing functions , dense in any @xmath53 ) .
it follows that ( with @xmath391 ) @xmath392 where we used that @xmath393 , any @xmath394 . since @xmath395 and @xmath396
, we can infer from that @xmath397 hence the numerical range of @xmath9 , @xmath398 , is contained in a sectorial region @xmath101 , @xmath399 with @xmath400 .
the numerical range of the adjoint @xmath401 is likewise contained in @xmath101 , and @xmath101 contains the spectrum of @xmath9 .
( the elementary functional analysis used here is explained e.g. in @xcite , ch .
12 . ) for simplicity we add @xmath402 to @xmath2 , so that we can use the information with @xmath403 in the following , replacing @xmath101 by @xmath404 the dirichlet trace @xmath405 is in the @xmath1-order case defined by @xmath406 with @xmath407 . for the dirichlet problems on smooth exterior or interior subsets of @xmath3 ,
the variational construction gives a realization with numerical range and spectrum likewise contained in @xmath408 .
moreover , there are sobolev space mapping properties of the solution operator ; this is extremely well - known for bounded domains , and for exterior domains it is covered e.g. by cor .
3.3.3 of @xcite ( the differential operator @xmath409 is uniformly parameter - elliptic on all rays @xmath410 outside @xmath408 , and parameter - ellipticity of the boundary problem holds uniformly for @xmath411 in the boundary ) .
let us specify the result for @xmath42 and @xmath13 defined as in section [ introduction ] .
we denote @xmath272 .
then @xmath412 has for @xmath413 the solution operator , continuous in the opposite direction , @xmath414 here @xmath415 is the inverse of the dirichlet realization @xmath416 , which acts like @xmath2 with domain @xmath417 .
the general theory of @xcite is here interpreted by use of the poisson operator @xmath418 ( and variants with @xmath243-dependence ) .
@xmath419 acts as an inverse of @xmath420 @xmath103 denoting the @xmath18 nullspace of @xmath2 .
the formulas are exactly the same as in @xcite section 3.3 .
we shall compare @xmath421 with the realization @xmath422 of a general _ normal _ boundary condition , defined as in @xcite , @xcite ( 3.85 ) .
let @xmath423 the cauchy data .
let @xmath424 be a subset of @xmath425 with @xmath51 elements , and let @xmath322 be a @xmath426-matrix of differential operators @xmath427 on @xmath13 of order @xmath428 : @xmath429 the boundary condition @xcite ( 3.85 ) : @xmath430 for @xmath431 , can then be written @xmath432 it defines the realization @xmath422 with domain @xmath433 special examples are the cases where @xmath434 or @xmath435 , @xmath436 they define dirichlet - type resp . neumann - type conditions .
let us assume that @xmath437 is uniformly parameter - elliptic for @xmath243 on a ray outside @xmath408 ; then for large @xmath243 on the ray , @xmath438 is invertible .
take a @xmath439 where this invertibility holds ( assuming invertibility of @xmath440 and @xmath441 also ) , and denote in the rest of this section @xmath442 by @xmath2 ; then we are in the situation where @xmath443 for @xmath79 has the solution operator , continuous in the opposite direction , @xmath444 here @xmath445 is the inverse of the realization @xmath446 , which acts like @xmath2 with domain @xmath447 the difference between @xmath448 and @xmath449 , and more generally between two solution operators @xmath449 and @xmath450 , can be described spectrally very much like in section [ cutoff ] .
first there is a generalization of proposition [ proposition3.1 ] and its corollaries .
define @xmath451 , @xmath452 and @xmath453 as in section [ cutoff ] .
[ proposition5.2 ] @xmath454 the operators @xmath455 and @xmath456 map continuously @xmath457 and are spectrally negligible .
@xmath458 when @xmath234 is a function in @xmath459 that is @xmath46 on a neighborhood of @xmath221 , the operators @xmath460 and @xmath461 map continuously @xmath462 and are spectrally negligible .
denote by @xmath463 the dirichlet trace operator for @xmath1-order operators on @xmath165 .
when @xmath464 , @xmath465 is @xmath192 on @xmath42 , hence @xmath466 .
then @xmath467 is a null - solution of the dirichlet problem for @xmath2 on @xmath165 with @xmath192 boundary value .
this will also hold if @xmath193 , any @xmath194 .
we now use that the dirichlet problem on @xmath165 has a solution operator with mapping properties similar to the problem for @xmath42 , and the proof is completed in the same way as the proofs of proposition [ proposition3.1 ] and corollary [ corollary3.2 ] .
[ remark5.3 ] it may be observed as in remark [ remark3.4 ] of section [ cutoff ] that the proofs can also be inferred from @xcite , lemma 2.4.8 , and this moreover implies a rapid decrease in @xmath243 of any schatten norm . with proposition [ proposition5.2 ]
it is easy to generalize theorem [ theorem3.5 ] as follows : [ theorem5.4 ] for @xmath468 and @xmath469 as above , defining invertible elliptic realizations , let @xmath470 then @xmath471 moreover , there are spectral asymptotics formulas for @xmath39 : @xmath472 where the constants are determined from the principal symbols . here
@xmath45 is the same constant as for @xmath270 , namely @xmath473 .
we proceed as in the proof of theorem [ theorem3.5 ] .
first , @xmath474 is written as a singular green operator on @xmath312 plus a spectrally negligible term , by a version of applied to both terms .
the assertions for @xmath475 then follow from @xcite th . 4.10 .
next , @xmath476 is treated similarly to @xmath295 in theorem [ theorem3.5 ] , noting that the operators of order 2 have been replaced by operators of order @xmath1 . finally , @xmath477 , where the spectral asymptotics behavior of @xmath273 dominates the sum , in view of the rules for @xmath99-numbers .
it is here allowed to take the set @xmath424 for @xmath468 different from the corresponding set @xmath478 for @xmath469 .
there are similar results for differences between higher powers @xmath479 , as in remark [ remark3.6 ] of section [ cutoff ] . a result of the type @xmath480 has been announced by gesztesy and malamud in @xcite , apparently based on a consideration of @xmath425-functions . in all the calculations
, @xmath2 can be taken to be a @xmath481-system , acting on @xmath482-vectors .
when @xmath2 is scalar , the boundary conditions with are the most general ones for which parameter - ellipticity can hold ( cf .
@xcite , sect .
1.5 ) ; in the systems case there exist more general normal boundary conditions , as studied in @xcite .
the above analysis can be extended to include these , mainly at the cost of a more complicated notational apparatus .
pseudodifferential @xmath427 could be allowed as in @xcite .
[ remark5.2]for bounded domains , the result for @xmath475 has been known since 1984 , since @xmath483 is then itself a singular green operator of order @xmath484 on a bounded domain , to which @xcite th . 4.10 applies . for selfadjoint cases ,
see also @xcite sect .
for the study of perturbations of essential spectra we restrict the attention to selfadjoint realizations .
first of all , this requires that @xmath2 equal its formal adjoint @xmath485 , moreover , it restricts the sets @xmath424 and matrices @xmath322 that can be allowed . with the notation @xmath486 , we have as a necessary condition on @xmath424 is that it should equal its reversed complement : @xmath487 to explain further , we recall some details from @xcite . from the green s formula @xmath488 where @xmath489 is skew - selfadjoint and invertible , it is seen that when we set @xmath490 ( taking @xmath491 of the contribution from @xmath492 along ) , we get the symmetric formula @xmath493 valid for @xmath494 . here
@xmath495 is indexed by @xmath496 , @xmath497 with @xmath498 of order @xmath499 ; it replaces @xmath500 in systematic considerations and maps from @xmath163 to @xmath501 .
green s formula has the extension to @xmath502 , @xmath503 : @xmath504 where @xmath505 denotes the duality between @xmath506 and @xmath507 . with @xmath495 replaced by the `` reduced neumann trace operator '' @xmath126
, one has for @xmath508 : @xmath509 here @xmath510 now when @xmath424 satisfies , the subsets @xmath511 satisfy @xmath512 we set @xmath513 as shown in @xcite and recalled in @xcite , the boundary condition @xmath514 may then be rewritten in the form , with differential operators @xmath515 , @xmath516 , @xmath517 , @xmath518 when we take into account . here the first condition @xmath519 can be viewed as the `` dirichlet part '' , purely concerned with @xmath405 , whereas the second condition @xmath520 can be viewed as the `` neumann - type part '' , where part of the neumann data @xmath521 is given as a function of the other data .
note that @xmath516 links the free dirichlet data @xmath522 to neumann data and has entries of positive order , and @xmath517 has entries of order @xmath523 .
we assume in the following that @xmath437 is uniformly parameter - elliptic on a ray outside @xmath408 as in the preceding section , so that @xmath528
. then @xmath529 is necessary and sufficient for selfadjointness of @xmath422 .
is assumed from now on .
the operator @xmath422 corresponds to a selfadjoint operator @xmath530 by the general theory , where @xmath101 is the @xmath18-closure of @xmath531 ( here @xmath532 ) .
@xmath101 is mapped by @xmath533 onto the closure @xmath534 of @xmath535 in @xmath536 . here
@xmath534 is the graph of @xmath515 , so it is homeomorphic to @xmath537 , by the mappings @xmath538 as shown in @xcite and recalled in @xcite . here
@xmath539 . the operator @xmath105 in @xmath101 is carried over to an operator @xmath543 which is further translated to an operator @xmath544 we now recall from @xcite how the form of @xmath545 is determined ( this detail was not repeated in @xcite ) .
consider the condition defining the correspondence between @xmath422 and @xmath105 ( cf .
@xcite ) : @xmath546 here the right - hand side is rewritten as @xmath547 whereas the left - hand side takes the form , in view of and : @xmath548 since @xmath549 runs in a dense subset of @xmath550 , implies @xmath551 so @xmath545 acts like @xmath552 .
the boundary condition may then be rewritten as @xmath553 since @xmath554 is elliptic , @xmath545 is an elliptic selfadjoint mixed - order pseudodifferential operator ; its domain is @xmath555 .
we replace @xmath530 for @xmath422 by an operator @xmath558 , selfadjoint invertible with a nonempty essential spectrum , and want to see how this effects the realization . as above
, @xmath559 carries over to @xmath560 with @xmath561 , and the boundary condition now takes the form @xmath562 here @xmath563 [ theorem6.1 ] consider the realization @xmath422 of @xmath2 in @xmath18 defined by a normal boundary condition @xmath514 ( cf . ) with @xmath526 , and assume that ellipticity and selfadjointness holds , cf . .
@xmath422 corresponds to an operator @xmath530 , where @xmath539 , cf . also , , .
let @xmath559 be a selfadjoint invertible operator in @xmath101 with nonempty essential spectrum , and let @xmath564 be the realization of @xmath2 corresponding to @xmath558 , i.e. , where the boundary condition , equivalently written , is replaced by .
then @xmath565 the proof goes in exactly the same way as the proof of theorem [ theorem4.1 ] and corollary [ corollary4.2 ] .
we cut up @xmath42 in a bounded part @xmath171 and an exterior part @xmath165 , and use and proposition [ proposition5.2 ] with @xmath566 to see that @xmath564 can be written as @xmath567 where @xmath360 is compact , the operator @xmath568 in @xmath569 has the same essential spectrum as @xmath448 , and the operator @xmath570 in @xmath337 has the same essential spectrum as @xmath571 outside 0 .
briefly expressed , the theorem states that any normal boundary condition ( apart from the dirichlet condition ) defining a selfadjoint invertible elliptic realization , can be perturbed by addition of a suitable operator to @xmath516 ( the map from the free dirichlet data to neumann data ) to provide a selfadjoint invertible realization with a prescribed augmentation of the essential spectrum .
[ example6.2 ] let @xmath572 .
clearly , @xmath2 satisifes the positivity and selfadjointness requirements , and it has the green s formula with @xmath573 as in @xcite example 3.14 . the dirichlet operator @xmath574 where @xmath575 , has an inverse @xmath576 continuous in the opposite direction .
let us take ( as in @xcite , ex .
3.14 ) @xmath577 ; it satisifes , and @xmath578 , @xmath579 . with this choice
, the boundary condition is of the form @xmath580 ( @xmath515 and @xmath517 vanish , being differential operators of negative order . ) @xmath516 is of order 1 .
selfadjointness of @xmath422 requires @xmath581 , and if this holds and the problem is elliptic , then @xmath422 is selfadjoint with domain @xmath582 continuing under this assumption , we find that @xmath583 and @xmath545 is the first - order pseudodifferential operator @xmath584 with @xmath585 in view of the ellipticity . there is a corresponding operator @xmath530 where @xmath586 .
invertibility holds e.g. when @xmath545 has a positive lower bound . replacing @xmath530 by @xmath558 ,
selfadjoint and invertible with a nonempty essential spectrum , corresponds to replacing @xmath516 by @xmath587 the corresponding realization @xmath41 is defined by the boundary condition @xmath588 and satisfies @xmath589 .
m. s. birman _ perturbations of the continuous spectraum of a singular elliptic operator under changes of the boundary and boundary condition _ , vestnik leningrad , 1 , 1962 , 2255 ; translated in amer . math .
( 2 ) vol . 225 , 2008 , 1953 .
m. s. birman and m. z. solomiak , _ asymptotics of the spectrum of variational problems on solutions of elliptic equations in unbounded domains _ funkts .
analiz prilozhen . 14 , 1980 , 2735 ; translated in funct .
appl . 14 , 1981 , 267274 . | for a second - order symmetric strongly elliptic differential operator on an exterior domain in @xmath0 it is known from works of birman and solomiak that a change of the boundary condition from the dirichlet condition to an elliptic neumann or robin condition leaves the essential spectrum unchanged , in such a way that the spectrum of the difference between the inverses satisfies a weyl - type asymptotic formula . we show that one can increase , but not diminish , the essential spectrum by imposition of other neumann - type non - elliptic boundary conditions .
the results are extended to @xmath1-order operators , where it is shown that for any selfadjoint realization defined by an elliptic normal boundary condition ( other than the dirichlet condition ) , one can augment the essential spectrum at will by adding a suitable operator to the mapping from free dirichlet data to neumann data .
we here also show an extension of the spectral asymptotics formula for the difference between inverses of elliptic problems .
the proofs rely on kren - type formulas for differences between inverses , and cutoff techniques , combined with results on singular green operators and their spectral asymptotics .
_ dedicated to vsevolod a. solonnikov on the occasion of his 75 years birthday .
_ exterior domain , essential spectrum , singular green operator , schatten class , krein formula , spectrally negligible cutoffs 35j40 , 35p20 , 35s15 , 47a10 |
let @xmath4 be a simple connected graph .
we use the standard graph notation ( @xcite ) .
in particular , @xmath5 denotes the automorphism group of @xmath0 . for simple connected graph @xmath0 , and @xmath6 ,
the _ neighborhood _ of a vertex @xmath7 is the set @xmath8 .
the _ degree of a vertex _ @xmath7 in a graph @xmath0 , denoted by @xmath9 , is the number of edges of @xmath0 incident with @xmath7 . in particular
, @xmath9 is the number of neighbours of @xmath7 in @xmath0 .
we denote by @xmath10 and @xmath11 the minimum and maximum degrees of the vertices of @xmath0 .
a graph @xmath0 is _ @xmath12-regular _ if @xmath13 for all @xmath6 .
the _ diameter of a graph _
@xmath0 is the greatest distance between two vertices of @xmath0 , and denoted by @xmath14 .
the _ distinguishing index _ @xmath1 of a graph @xmath0 is the least number @xmath15 such that @xmath0 has an edge labeling with @xmath15 labels that is preserved only by the identity automorphism of @xmath0 .
the distinguishing edge labeling was first defined by kalinowski and piniak @xcite for graphs ( was inspired by the well - known distinguishing number @xmath16 which was defined for general vertex labelings by albertson and collins @xcite ) .
the distinguishing index of some examples of graphs was exhibited in @xcite .
for instance , @xmath17 for every @xmath18 , and @xmath19 for @xmath20 , @xmath21 for @xmath22 . also , for complete graphs @xmath23 , we have @xmath24 for @xmath20 , @xmath25 for @xmath22 .
they showed that if @xmath0 is a connected graph of order @xmath18 and maximum degree @xmath26 , then @xmath27 , unless @xmath0 is @xmath28 or @xmath29 .
it follows for connected graphs that @xmath30 if and only if @xmath31 and @xmath0 is a cycle of length at most five .
the equality @xmath32 holds for all paths , for cycles of length at least 6 , for @xmath33 , @xmath34 and for symmetric or bisymmetric trees .
also , pilniak showed that @xmath35 for all other connected graphs .
@xcite[updisindg ] let @xmath0 be a connected graph that is neither a symmetric nor an asymmetric tree .
if the maximum degree of @xmath0 is at least 3 , then @xmath36 unless @xmath0 is @xmath33 or @xmath34 .
pilniak put forward the following conjecture .
@xcite[konj ] if @xmath0 is a @xmath37-connected graph , then @xmath38 . in this paper , we prove the following theorem which proves the conjecture .
let @xmath0 be a connected graph of maximum degree @xmath26 .
if the minimum degree @xmath39 , then @xmath40 . for our purposes ,
we consider graphs with specific construction that are from dutch - windmill graphs . because of this , in section 2 , we compute the distinguishing index of the dutch windmill graphs . in section 3
, we use the results to prove the main result . in the last section we present graphs @xmath0 for which @xmath3 .
to obtain the upper bound for the distinguishing index of connected graphs with minimum degree at least two , we characterize such graphs with minimum number of edges . for this characterization
we need the concept of dutch windmill graphs .
the _ dutch windmill graph _
@xmath41 is the graph obtained by taking @xmath42 , ( @xmath43 ) copies of the cycle graph @xmath44 , ( @xmath45 ) with a vertex in common ( see figure [ fig0 ] ) . if @xmath46 , then we call @xmath47 , a friendship graph . in the following theorem
we compute the distinguishing number of dutch windmill graphs .
examples of dutch windmill graphs.,scaledwidth=75.0% ] [ thmbasic ] for every @xmath43 and @xmath45 , @xmath48 .
* proof .
* we consider two cases : * case 1 ) * if @xmath12 is odd .
there is a natural number @xmath49 such that @xmath50 .
we can consider a blade of @xmath41 as figure [ fig00 ] .
the considered polygon ( or a cycle of size @xmath12 ) in the proof of theorem [ thmbasic].,scaledwidth=90.0% ] let @xmath51 be the label of vertices @xmath52 of the @xmath53th blade where @xmath54 .
suppose that @xmath55 is a labeling of the vertices of @xmath41 except its central vertex . in an @xmath56-distinguishing labeling we must have : *
there exists @xmath57 such that @xmath58 for all @xmath59 . * for @xmath60 we must have @xmath61 and @xmath62 .
there are @xmath63 possible @xmath64-arrays of labels using @xmath56 labels satisfying ( i ) and ( ii ) , and so @xmath65 . * case 2 ) * if @xmath12 is even .
there is a natural number @xmath49 such that @xmath66 .
we can consider a blade of @xmath41 as figure [ fig00 ] .
let @xmath67 be the label of vertices @xmath68 of @xmath53th blade where @xmath54 .
suppose that @xmath69 is a labeling of the vertices of @xmath41 except its central vertex . in an @xmath56-distinguishing labeling we must have : *
there exists @xmath70 such that @xmath58 for all @xmath59 .
* for @xmath60 we must have @xmath71 there are @xmath72 possible @xmath73-arrays of labels using @xmath56 labels satisfying ( i ) and ( ii ) ( @xmath56 choices for @xmath74 and @xmath75 choices for @xmath76 ) .
therefore @xmath77.@xmath78 the following theorem gives the distinguishing index of @xmath41 .
[ indexwindmil ] for any @xmath43 and @xmath45 , @xmath79 .
* proof .
* since the effect of every automorphism of @xmath80 on its non - central vertices is exactly the same as the effect of an automorphism of @xmath41 on its edges and vice versa , so if we consider the non - central vertices of @xmath80 as the edges of @xmath41 , then we have @xmath81 .
therefore the result follows from theorem [ thmbasic].@xmath78
in this section , we shall prove conjecture [ konj ] . to do this ,
first we state some preliminaries . by the result obtained by fisher and isaak @xcite and independently by imrich , jerebic and klavar @xcite
the distinguishing index of complete bipartite graphs is as follows : @xcite[indcombipar ] let @xmath82 be integers such that @xmath83 and @xmath84 .
then @xmath85 if @xmath86 then the distinguishing index @xmath87 is either @xmath15 or @xmath88 and can be computed recursively in @xmath89 time .
@xcite[uppcompbipart ] if @xmath90 , then @xmath91{q } \rceil + 1 $ ] .
also we need the following result : @xcite[hamiltopath ] if @xmath0 is a graph of order @xmath92 such that @xmath0 has a hamiltonian path , then @xmath93 .
now , we state and prove the main theorem of this paper . [ uppinddelta2 ]
let @xmath0 be a connected graph with maximum degree @xmath26 .
if @xmath39 then @xmath94 .
* proof . * if @xmath95 , then the result follows from theorem [ updisindg ] .
so , we suppose that @xmath96 .
let @xmath7 be a vertex of @xmath0 with the maximum degree @xmath26 . by theorem [ indexwindmil ]
, we can label the edges of the dutch windmill graph attached to @xmath0 at vertex @xmath7 ( a subgraph @xmath97 is attached to graph @xmath0 , if it has only one vertex in common with graph @xmath0 ) for which @xmath7 is the central point of the dutch windmill graph , with at most @xmath98 labels from label set @xmath99 , distinguishingly .
if there exists triangle attached to @xmath0 at @xmath7 , then we label the two its incident edges to @xmath7 with @xmath100 and @xmath101 , and another edges of the triangle with label @xmath37 .
let @xmath102 be the vertices of @xmath0 at distance one from @xmath7 , except the vertices of dutch windmill or triangle attached to graph at @xmath7 .
we continue the labeling by the following steps : * step 1 ) * since @xmath103 , so for @xmath104 and @xmath105 , we label the edges @xmath106 with label @xmath53 , and we do not use the label @xmath100 any more . with respect to the number of incident edges to @xmath7 with label 0 ,
we conclude that the vertex @xmath7 is fixed under each automorphism of @xmath0 preserving the labeling .
also , since the dutch windmill or the triangle graph attached to @xmath0 at @xmath7 has been labeled distinguishingly , so the vertices of attached graph are fixed under each automorphism of @xmath0 preserving the labeling .
hence , every automorphism of @xmath0 preserving the labeling must map the set of vertices of @xmath0 at distance @xmath53 from @xmath7 to itself setwise , for any @xmath107 .
we denote the set of vertices of @xmath0 at distance @xmath53 from @xmath7 , by @xmath108 for @xmath109 .
if for any @xmath110 , @xmath111 , then @xmath0 has a hamiltonian path , and since @xmath96 , so the order of @xmath0 is at least @xmath112 , and hence @xmath93 by theorem [ hamiltopath ] .
thus we suppose that @xmath113 , for some @xmath110 .
now we partition the vertices of @xmath114 to two sets @xmath115 and @xmath116 as follows : @xmath117 the sets @xmath115 and @xmath116 are mapped to @xmath115 and @xmath116 , respectively , setwise , under each automorphism of @xmath0 preserving the labeling . for @xmath104 , we set @xmath118 . by this notation , we get that for @xmath104 , the set @xmath119 is mapped to @xmath119 setwise , under each automorphism of @xmath0 preserving the labeling .
let the sets @xmath120 and @xmath121 for @xmath104 are as follows : @xmath122 it is clear that the sets @xmath120 and @xmath121 are mapped to @xmath120 and @xmath121 , respectively , setwise , under each automorphism of @xmath0 preserving the labeling .
since for any @xmath104 , we have @xmath123 , so we can label all incident edges to each element of @xmath120 with labels @xmath124 , such that for any two vertices of @xmath120 , say @xmath125 and @xmath126 , there exists a label @xmath12 , @xmath127 , such that the number of label @xmath12 for the incident edges to vertex @xmath125 is different from the number of label @xmath12 for the incident edges to vertex @xmath126 .
hence , it can be deduce that each vertex of @xmath120 is fixed under each automorphism of @xmath0 preserving the labeling , where @xmath128 .
thus every vertices of @xmath115 is fixed under each automorphism of @xmath0 preserving the labeling . in sequel
, we want to label the edges incident to vertices of @xmath116 such that @xmath116 is fixed under each automorphism of @xmath0 preserving the labeling , pointwise . for this purpose , we partition the vertices of @xmath121 to the sets @xmath129 , ( @xmath130 ) as follows : @xmath131 since the set @xmath108 , for any @xmath53 ,
is mapped to itself , it can be concluded that @xmath129 is mapped to itself , setwise , under each automorphism of @xmath0 preserving the labeling , for any @xmath53 and @xmath132 .
let @xmath133 .
it is clear that @xmath134 .
let @xmath135 , and @xmath136 .
we assign to the @xmath132-ary @xmath137 of edges , a @xmath132-ary of labels such that for every @xmath138 and @xmath139 , @xmath140 , there exists a label @xmath141 in their corresponding @xmath132-arys of labels for which the number of label @xmath141 in the corresponding @xmath132-arys of @xmath138 and @xmath139 is distinct . for constructing @xmath142 numbers of such @xmath132-arys
we need , @xmath143 distinct labels . since for any @xmath130 , we have @xmath144 so we need at most @xmath145 distinct labels from label set @xmath146 for constructing such @xmath132-arys . for instance , let @xmath147 , and @xmath148 . by our method ,
we label the edge @xmath149 with label @xmath12 for @xmath150 where @xmath151 .
hence , the vertices of @xmath129 , for any @xmath130 , are fixed under each automorphism of @xmath0 preserving the labeling .
therefore , the vertices of @xmath121 for any @xmath104 , and so the vertices of @xmath116 are fixed under each automorphism of @xmath0 preserving the labeling .
now , we can get that all vertices of @xmath114 are fixed .
if there exist unlabeled edges of @xmath0 with the two endpoints in @xmath114 , then we assign them an arbitrary label , say 1 .
now we consider @xmath152 .
we partition this set such that the vertices of @xmath152 with the same neighbours in @xmath116 , lie in a set . in other words , we can write @xmath153 , such that @xmath154 contains that elements of @xmath152 having the same neighbours in @xmath116 , for any @xmath53 .
since all vertices in @xmath116 are fixed , so the set @xmath154 is mapped to @xmath154 setwise , under each automorphism of @xmath0 preserving the labeling .
let @xmath155 , and we have @xmath156 we consider two following cases : case 1 ) if for every @xmath157 and @xmath158 in @xmath154 , where @xmath159 , there exists a @xmath12 , @xmath160 , for which the label of edge @xmath161 is different from label of edge @xmath162 , then all vertices of @xmath0 in @xmath154 are fixed under each automorphism of @xmath0 preserving the labeling .
case 2 ) if there exist @xmath157 and @xmath158 in @xmath154 , where @xmath159 , such that for every @xmath12 , @xmath160 , the label of edge @xmath161 and @xmath162 are the same , then we can make a labeling such that the vertices in @xmath154 have the same property as case 1 , and so are fixed under each automorphism of @xmath0 preserving the labeling , by using at least one of the following actions : * by commutating the coordinates of @xmath132-ary of labels assigned to the incident edges to @xmath163 with an end point in @xmath152 .
* by using a new @xmath132-ary of labels , with labels @xmath146 , for incident edges to @xmath163 with an end point in @xmath152 , such that ( by notations in step 1 ) for every @xmath164 and @xmath165 , @xmath166 , there exists a label @xmath141 in their corresponding @xmath132-arys of labels with different number of label @xmath141 in their coordinates , where @xmath167 .
* by labeling the unlabeled edges of @xmath0 with the two end points in @xmath152 which are incident to the vertices in @xmath154 .
* by labeling the unlabeled edges of @xmath0 which are incident to the vertices in @xmath154 , and another their endpoint is @xmath168 . * by labeling the unlabeled edges of @xmath0 with the two end points in @xmath168 for which the end points in @xmath168 are adjacent to some of vertices in @xmath154 . using at least one of above actions
, it can be seen that every two vertices @xmath157 and @xmath158 in @xmath154 have the property as case 1 .
thus we conclude that all vertices in @xmath154 , for any @xmath53 , and so all vertices in @xmath152 , are fixed under each automorphism of @xmath0 preserving the labeling .
if there exist unlabeled edges of @xmath0 with the two endpoints in @xmath152 , then we assign them an arbitrary label , say 1 . by continuing this method , in the next step we partition @xmath168 exactly by the same method as partition of @xmath152 to the sets @xmath154 s in step 2 , and
so we can make a labeling such that @xmath108 is fixed pointwise , under each automorphism of @xmath0 preserving the labeling , for any @xmath169.@xmath78 for a @xmath37-connected planar graph @xmath0 , the distinguishing index may attain @xmath170 .
for example , consider the complete bipartite graph @xmath171 with @xmath172 , where @xmath56 is a positive integer @xmath56 . by theorem [ indcombipar ] , @xmath173 .
in this section , we present graphs @xmath0 with specific construction such that @xmath175 . to do this we state the following definition .
[ minimally ] let @xmath0 be a connected graph with @xmath176 .
the graph @xmath0 is called a @xmath177-minimally graph , if the minimum degree of each spanning subgraph of @xmath0 is less than @xmath10 .
it can be concluded from definition [ minimally ] that if @xmath178 is an edge of a connected @xmath177-minimally graph with end points @xmath179 and @xmath180 , then without loss of generality we can assume that @xmath181 and @xmath182 . in fact the distance between the two vertices of degree greater than @xmath177 is at least two .
the simplest connected @xmath37-minimally graphs are cycles @xmath183 and complete bipartite graphs @xmath184 .
now , we explain more on the structure of a @xmath37-minimally graph .
let to call a path in the graph a _ simple path _
, if all its internal vertices have degree two . let @xmath0 be a connected @xmath37-minimally graph
. * if the degree of all vertices of @xmath0 is two , then @xmath0 is a cycle graph . *
if there exist a vertex @xmath7 of @xmath0 with degree at least three .
we consider two following cases : + case 1 ) if @xmath7 is the only vertex of @xmath0 with degree greater than two , then @xmath0 is a graph which is made by identifying the central points of some dutch windmill graphs @xmath185 where @xmath186 , and hence @xmath187 where @xmath188 is a set of indices . in this case
we denote @xmath0 by @xmath189 ( for instance , see figure [ figwind ] ) .
+ examples of @xmath189.,scaledwidth=80.0% ] + case 2 ) if @xmath0 has other vertex @xmath180 of degree greater than two , then there exists at least a simple path between @xmath7 and @xmath180 of length greater than one . since @xmath0 is a connected graph ,
so if there exists no such simple path , then there exists a vertex of degree at least three on each path between @xmath7 and @xmath180 .
hence we can obtain a vertex @xmath179 of @xmath0 with degree greater than two such that there exists at least a simple path between @xmath7 and @xmath179 of length greater than one ( see figure [ fig1 ] ) .
the state of vertices of degree greater than two in a connected @xmath37-minimally graph.,scaledwidth=100.0% ] the state of vertices of degree greater than two in a connected @xmath37-minimally graph.,scaledwidth=100.0% ] now we characterize graphs @xmath0 with @xmath190 .
[ indexminimallygraphs ] let @xmath0 be a connected @xmath37-minimally graph with maximum degree @xmath26 . if @xmath0 is not a cycle @xmath191 or a complete bipartite graph @xmath192 for some integer @xmath56 , then @xmath190 .
* proof . * if @xmath193 , then @xmath0 is a cycle .
it is known that the distinguishing index of cycle graph of order at least @xmath194 is two .
hence , we suppose that @xmath0 is not a cycle , so @xmath195 .
let @xmath7 be a vertex of @xmath0 of maximum degree @xmath26 .
suppose that @xmath196 are all vertices of @xmath0 which are of degree at least three such that there exists at least a simple path between @xmath7 and @xmath197 , for any @xmath198 ( it is possible that @xmath199 ) .
let there exist @xmath200 disjoint simple paths of length @xmath132 between @xmath7 and @xmath197 , for any @xmath198 and @xmath201 where @xmath200 is a non - negative integer and @xmath202 .
we can label these @xmath200 simple paths of length @xmath132 with at most @xmath203 labels , by using @xmath200 numbers of @xmath132-arys such that the coordinates of each @xmath132-ary are in the set @xmath204 , for any @xmath205 and @xmath201 , and for every two paths of length @xmath132 , say @xmath206 and @xmath207 , there exists a label @xmath141 , @xmath208 , such that the number of label @xmath141 in @xmath132-arys related to @xmath206 and @xmath207 are distinct .
let @xmath209 be a simple path between @xmath7 and @xmath197 for some @xmath205 , such that the label of edge of @xmath209 which is incident to @xmath7 , is different from the label of edge of @xmath209 which is incident to @xmath197 .
we do not use of labeling of the simple path @xmath209 , for any other simple path ( with the same length ) between any two vertices of degree greater than two . since @xmath0 is not a complete bipartite graph @xmath210 for some integer @xmath56 , so we can label these paths distinguishingly with at most @xmath211 labels .
now , we label the induced subgraph @xmath212 , for any vertex @xmath125 of degree greater than two , if there exists , with at most @xmath203 labels distinguishingly by theorem [ indexwindmil ] , such that the distinguishing labeling of @xmath213 is nonisomorphic to the remaining distinguishing labeling of @xmath212 , where @xmath214 .
thus any automorphism of @xmath0 preserving this labeling should be fixed @xmath215 and all vertices of degree two on the simple paths between @xmath7 and @xmath197 for any @xmath205 . since for any @xmath216 where @xmath217 ,
the vertices @xmath197 and @xmath218 are fixed , so all the simple paths between @xmath197 and @xmath218 , if there exist , are mapped to each other under each automorphism of @xmath0 preserving this labeling .
hence we can label all edges of these simple paths with at most @xmath219 labels , by assigning distinct ordered arys of labels of length of the simple paths between @xmath197 and @xmath218 such that all vertices of these paths are fixed under each automorphism of @xmath0 preserving this labeling . for any @xmath205
, we consider @xmath197 , and suppose that @xmath220 are all vertices of @xmath221 with degree at least three such that there exists at least a simple path between @xmath197 and @xmath222 for any @xmath223 .
now we do the same method as labeling of simple paths between @xmath7 and @xmath224 , for all simple paths between @xmath197 and @xmath225 with at most @xmath226 labels . also , we do the same method as labeling of simple paths between @xmath197 and @xmath218 , for all simple paths between @xmath227 and @xmath228 with at most @xmath145 labels , where @xmath229 . note that we do not use labeling of @xmath209 for any simple path with the same length as @xmath209 between @xmath197 and @xmath225 .
thus the vertices @xmath230 and all vertices of the simple paths between them are fixed under each automorphism of @xmath0 preserving this labeling .
after the finite number of steps we can obtain a distinguishing edge labeling of @xmath0 with at most @xmath203 labels.@xmath78
we gave an upper bound for the distinguishing index of graphs @xmath0 with minimum degree at least two .
this result proves a conjecture by pilniak ( 2015 ) .
we also studied graphs @xmath0 with @xmath190 .
we think that the following conjecture is true , but until now all attempts to prove this failed .
so , we end this paper by proposing the following conjecture . * if @xmath0 is a @xmath177-minimally graph with @xmath231 such that @xmath0 is not a complete bipartite or @xmath177-regular graph , then @xmath232{\delta ( g)}\rceil$ ] . * if @xmath0 is a connected graph with @xmath231 , then @xmath233{\delta ( g)}\rceil$ ] .
m. j. fisher and g. isaak , _ distinguishing colorings of cartesian products of complete graphs _ , discrete math .
308 ( 11 ) ( 2008 ) , 2240 - 2246 .
r. hammack , w. imrich and s. klavzar , _ handbook of product graphs ( second edition ) _ , taylor & francis group , 2011 .
w. imrich , j. jerebic and s. klavar , _ the distinguishing number of cartesian products of complete graphs _ , european j. combin .
29 ( 2008 ) , 922 - 929 .
r. kalinowski and m. pilniak , _ distinguishing graphs by edge colourings _ , european j. combin . 45 ( 2015 ) , 124 - 131 . | the distinguishing index of a simple graph @xmath0 , denoted by @xmath1 , is the least number of labels in an edge labeling of @xmath0 not preserved by any non - trivial automorphism .
it was conjectured by pilniak ( 2015 ) that for any 2-connected graph @xmath2 .
we prove a more general result for the distinguishing index of graphs with minimum degree at least two from which the conjecture follows .
also we present graphs @xmath0 for which @xmath3 .
department of mathematics , yazd university , 89195 - 741 , yazd , iran + [email protected] , [email protected] * keywords : * distinguishing index ; edge colourings ; bound * ams subj . class . * : 05c25 , 05c15 |
theoretically , two interacting atoms in hyperfine states @xmath69 and @xmath107 under quasi-2d harmonic confinement are described by the hamiltonian @xmath108+v_{int}(\mathbf{r}_1-\mathbf{r}_2),\end{aligned}\ ] ] where the interaction @xmath109 is the interaction potential and @xmath110 . after separating into the relative coordinate @xmath111 and the center of mass coordinate @xmath112
one has @xmath113 with reduced mass @xmath114 and total mass @xmath115 . for a short
ranged potential @xmath116 with range @xmath117 , we can find an approximate solution to the two - body problem by first noting that the ansatz @xcite @xmath118 where @xmath119 is the green s function of the hamiltonian @xmath120 .
@xmath121 satisfies the schrdinger equation for the relative coordinate hamiltonian away from the origin @xmath122 .
close to the origin ( @xmath123 ) we match this outer solution to logarithmic derivative of the asymptotic free space scattering solution of the interaction potential @xmath116 @xmath124 where @xmath125 , @xmath126 and @xmath127 is the s - wave phase shift at energy @xmath128 . with the usual effective range expansion @xmath129 and the expansion of @xmath121 from refs .
@xcite @xmath130 we arrive at eq . 3 of the main text .
it is instructive to study the problem of molecule dissociation in a strict two - dimensional setting @xcite .
such a description should be valid whenever the binding energy @xmath103 ( @xmath74 ) of the molecule in initial ( final ) state channels and the energy of the rf - photon is much less than the spacing of transverse harmonic oscillator modes @xmath0 .
in general the rf spectrum @xmath131 of molecules can be calculated from fermi s golden rule @xcite @xmath132 the summation over final states includes a bound - state at energy @xmath133 and a continuum of unbound atoms with @xmath134 and relative momentum @xmath135 . for brevity ,
we use units where @xmath136 .
the matrix element for the bound - bound transition amplitude @xmath137 is given by the overlap integral @xmath138^ * \psi_{b}^{13}(\mathbf{r})\end{aligned}\ ] ] where @xmath139 [ @xmath140 is the molecule wavefunction in the initial [ final ] state , respectively . at bound - state energy @xmath141 ,
the normalized wavefunction for short - range attractive s - wave interactions in 2d is @xmath142 which gives @xmath80 we note that this result makes sense from a qualitative point of view : when interactions in final and initial state are equal , one has @xmath143 ( as all weight is concentrated in the bound - bound transition ) , whereas for vanishing final state interactions ( i.e. @xmath144 or @xmath145 ) , @xmath137 tends to zero .
the bound - free continuum can be calculated from the matrix element between a s - wave scattering state @xmath146 and @xmath147 .
such a state satisfies a lippmann - schwinger equation @xmath148 and is guaranteed to have the same normalization as the solutions to the free hamiltonian @xmath149 . for a 2d short range potential
the lippmann - schwinger equation is solved in the s - wave channel by @xcite @xmath150 with the 2d scattering amplitude for the final state channel @xmath151 here we picked the normalization of the scattering states such that @xmath152^*\psi^{(+)}_{k'}(\mathbf{r})=\frac{1}{2 \pi k } \delta(k - k')\end{aligned}\ ] ] in this normalization , the bound - free transition amplitude @xmath153 becomes @xmath154^ * \psi_k^{(+)}(\mathbf{r } ) \nonumber\\&=&\frac{\sqrt{e_b^{12}/\pi}}{k^2+e_b^{12}}\left[1-\frac{f^{13}(e)}{f^{12}(e)}\right].\end{aligned}\ ] ] evaluating the sum over final momentum states we obtain a result for the bound - free spectrum part of the spectrum @xmath155 : @xmath156 where @xmath157 is the rf - spectrum without final state interactions and @xmath158 is a correction term accounting for final state interactions . note that the complete spectrum @xmath159 satisfies the sum - rule @xmath160 we will now turn our attention to the problem of a three - dimensional gas confined into two dimensions by tight one - dimensional harmonic confinement . in such a potential
the relative and center of mass coordinates separate , so we only need to consider the hamiltonian for the relative coordinate . in order to calculate the bound - bound and bound - free transition rates , we first need to know the normalized wavefunction of the two - body bound state in this geometry .
the green s function at energy @xmath86 in the harmonic oscillator potential is given by @xcite @xmath161 where @xmath162 and the transverse harmonic oscillator states have @xmath163 $ ] .
we can use the explicit expression for the bound - state wave function @xmath164 in terms of @xmath121 to obtain the normalization factor @xmath165 where @xmath99 is the two - body binding energy , satisfying @xmath166 .
the normalized wavefunction of the two - body bound - state is thus given by @xmath167 with this wavefunction it is straightforward to calculate the bound - bound transition amplitude @xmath168 @xmath169 where @xmath170 .
we note that for equal interactions in initial and final states ( i.e. @xmath171 ) , @xmath172 .
also note that for small @xmath173 , @xmath174 , eq .
reduces to the 2d result eq . .
analogous to the simpler 2d limit we can obtain the contribution of the bound - free continuum from the overlap between scattering state and bound state wavefunction .
for the final scattering states we use the basis of retarded scattering states given described in ref .
@xcite @xmath175 for the precise expressions of the coefficients @xmath176 see ref .
@xcite . here
the prime indicates that we consider the final state channel . by parity symmetry ,
the overlap matrix element @xmath177 vanishes for odd @xmath178 . for even @xmath178
one obtains @xmath179}{(e - e_b^{12})\sqrt{\pi l_z g'(e_b^{12}/\omega_z ) } } \right|_{e = q^2+\nu}.\end{aligned}\ ] ] the transition rate into the continuum is given by @xmath180\\ & & \times \left |1-\frac{f^{13}(\omega - e_b^{12})}{f^{12}(\omega - e_b^{12})}\right|^2.\end{aligned}\ ] ] note that this result is again of the form @xmath181 , where @xmath182 is spectrum without final state interactions and the factor @xmath183 we finally note that it is convenient to use a closed form expression for the sum @xmath184 as given in ref .
@xcite . with this expression
one can directly confirm that the tails of rf - spectrum without final state interactions obeys the 3d tan relations .
since the expression for @xmath185 via the summation of ref .
@xcite converges very slowly we will give a numerically more efficient scheme to evaluate differences of the form @xmath186 where @xmath187 . using the integral representation for @xmath185 of ref .
@xcite one obtains @xmath188 after exchanging integration and differentiation .
we can then numerically compute @xmath189 where @xmath190 is a contour in the lower complex halfplane ( e.g. we used straight lines along @xmath191 ) . | the line shape of radio frequency spectra of tightly bound feshbach molecules in strong transverse confinement can be described by a simple analytic formula that includes final state interactions . by direct comparison to experimental data
, we clarify the role of effective range corrections to two - body bound - state energies in lower dimensions .
recent experimental progress in the creation of quasi - two dimensional ( 2d ) atomic fermi gases @xcite has provided important insights into the physics of fermionic pairing . pairing phenomena in two dimensions
@xcite are intriguing because of the presence of a confinement - induced bound state for a zero - range interaction potential of arbitrary strength and sign @xcite .
this confinement - induced bound state significantly affects many - body pairing in the bec - bcs crossover in two dimensions and leads to the appearance of a pairing pseudogap phase for weak pairing @xcite . in the investigation of fermionic pairing
, the single - particle excitation spectrum @xcite probed by momentum - resolved @xcite or momentum - integrated @xcite radio - frequency ( rf ) spectroscopy plays a crucial role . up to now , theoretical work on rf spectroscopy has focused on the strict two - dimensional case whereas experiments have been conducted in quasi - two dimensions .
the latter refers to the situation in which tight confinement along one axis , usually with a harmonic frequency @xmath0 , establishes a kinematically two - dimensional gas with @xmath1 . here
, @xmath2 is the fermi energy of a two - dimensional fermi gas , @xmath3 is boltzmann s constant and @xmath4 is temperature .
however , if the three - dimensional @xmath5-wave scattering length @xmath6 , parameterizing the zero - range interaction potential , is smaller than the axial ground state @xmath7 , the relative wave function of the molecule explores higher transverse harmonic oscillator modes . here
, @xmath8 denotes the atomic mass . in this rapid communication
, we investigate the effects of quasi-2d confinement on rf spectra with a particular focus on the regime of strong pairing with @xmath9 and @xmath10 , both experimentally and theoretically .
specifically , we focus on three different aspects : first , the effect of finite range corrections , which are required to match experiment and theory in the limit of large binding energies , second the bound - bound transitions between confinement - induced states , and third the effects of final state interactions on rf spectra , which display significant differences between the strict-2d and the quasi-2d scenario .
our analytical theoretical results for binding energies and line shape functions compare well with the experimental data . in our experiment , we prepare a 50/50 spin mixture of @xmath11k atoms in the @xmath12 and @xmath13 states of the hyperfine ground state manifold @xcite .
we achieve quantum degeneracy of two - dimensional fermi gases in an optical lattice potential formed by a horizontally propagating , retro - reflected laser beam of wavelength @xmath14 nm .
the trapping frequency along the strongly confined direction is @xmath15khz , which is calibrated by intensity modulation spectroscopy .
the radial trapping frequency of the two - dimensional gases is @xmath16hz , and we confine on the order of @xmath17 atoms per two - dimensional gas at the center of the trap . along the axial direction
we populate approximately 30 - 40 layers of the optical lattice potential with an inhomogeneous peak density distribution . starting from a weakly interacting gas ( @xmath18 ) at a magnetic field near 209 g
, we adiabatically increase the interaction strength by lowering the magnetic field at a rate of up to 0.25g / ms to a value near the feshbach resonance at 202.1 g .
we apply an rf pulse with frequency @xmath19 near 47mhz with a gaussian amplitude envelope with a full width at half maximum of 230@xmath20s to transfer a fraction of atoms from the @xmath21 state to the @xmath22 state .
atoms in the @xmath23 state have a two - body s - wave scattering length of 250 bohr radii with the @xmath24 state .
we turn off the optical lattice 100@xmath20s after the rf pulse , switch off the magnetic field and apply a magnetic field gradient to achieve spatial splitting of the three spin components in a stern - gerlach experiment during time - of - flight .
finally , we detect the atom numbers in each of the atomic states by absorption imaging .
[ sec : boundstate ] we determine the binding energies of the molecules by recording the number of atoms transferred into the state @xmath23 and determine the threshold of the spectrum corrected for our spectral resolution of 1.5khz @xcite . in the following ,
we subtract the ( trivial ) contribution of the atomic zeeman energy @xmath25 from the rf frequency to obtain @xmath26 . in figure
[ fig : bindingenergy ] we display the measured binding energies and compare with theoretical models .
a pair of atoms interacting via zero range contact interactions under quasi-2d confinement has a binding energy given by the solution to the equation @xcite @xmath27 where @xmath28^{-1/2 } \right)$ ] @xcite .
this prediction is shown as the dashed line in fig .
[ fig : bindingenergy ] . in the limit of large binding energy @xmath29
we find significant deviations from the quasi-2d binding energy .
it has been argued that under tight confinement and for large binding energies , effective - range corrections to the standard zero - range model are of relevance @xcite .
these corrections can be viewed as an energy dependence of scattering length @xmath30 here @xmath31 is the energy in the center of mass frame , @xmath32 and @xmath33 is the effective range . by replacing @xmath34 with @xmath35 , the effective range corrections to eq .
take the form ( see also the supplemental material ) @xmath36 for open channel dominated feshbach resonances , the effective range is given by @xcite @xmath37 ^ 4}{6 \pi^2 } \bar{a } \left [ 1 - 2 \frac{\bar{a}}{a_s}+2 \left(\frac{\bar{a}}{a_s}\right)^2\right],\end{aligned}\ ] ] where @xmath38 ^ 2 $ ] is the mean scattering length .
this implies that , in the vicinity of a feshbach resonance , @xmath11k atoms have an effective range of @xmath39 nm ( where we used @xmath40 in atomic units ) @xcite .
we plot the binding energy including effective range corrections as the solid line fig .
[ fig : bindingenergy ] .
the excellent agreement with our measured binding energies demonstrates the importance of effective range corrections for quasi-2d confinement in the limit of tightly bound feshbach molecules of the @xmath24 and @xmath21 states of @xmath11k atoms .
the agreement between and theory and experiment can presumably be further improved if the finite width of the feshbach resonance is taken into account @xcite .
this is expected to give a contribution of order @xmath41 to the effective range .
@xcite . ]
[ sec : rf ] next , we turn our attention to the effects of the quasi-2d nature on the rf spectra .
first , we study the case without final state interactions .
this situation is particularly simple as it contains only a bound - free transition .
the rf transition is driven with a rabi frequency @xmath42 , described by the operator @xmath43 . the transition rate
is then given by fermi s golden rule @xmath44 which involves computation of the franck - condon overlaps @xmath45 between the normalized bound - state relative wave function @xmath46^{-1/2 } \sum_{n } \phi_{n}(z ) \phi_{n}(0 ) k_0(\kappa_{n } \rho)$ ] , with @xmath47 , and free outgoing plane wave states @xmath48 : @xmath49^ * \psi_b^{12}(\mathbf{r})\nonumber \\ & = & \frac{\hbar^2 \phi_{n}(0)}{m(e_{n\mathbf{q}}-e_b^{12})\sqrt{\pi l_z g'(e_b^{12}/\omega_z)}}.\end{aligned}\ ] ] here , @xmath50 is the wave function of the @xmath51-th transverse harmonic oscillator mode , @xmath52 is the in - plane relative momentum and @xmath53 . after summation over final states ,
we find the resulting line shape to be @xmath54 here , @xmath55 denotes the derivative of @xmath56 , @xmath57 is the heaviside function . for rf frequencies @xmath58 near the binding energy , the line shape shows a @xmath59 scaling which is found also in strict 2d @xcite . in this case , the rf spectrum takes the form @xmath60 in which we have introduced the two - dimensional contact parameter @xmath61 @xcite , which satisfies the 2d tan relation
. however , as the detuning from the binding energy becomes comparable to twice the axial oscillation frequency @xmath62 , additional local maxima appear in the spectrum , signaling the importance of higher axial modes .
the tail of @xmath63 at very large frequencies @xmath64 behaves as @xmath65 using the three - dimensional contact @xcite @xmath66 .
this illustrates the fundamentally three - dimensional character of quasi-2d feshbach molecules for short interatomic distances .
therefore , the strict-2d contact @xcite will only be revealed from rf spectra when confinement effects are negligible , i.e. when frequencies and binding energies much less than @xmath67 are considered .
finally , we include final state effects .
they have to be considered if the final state @xmath68 of the rf spin flip is still interacting with one of the initial states . in our experimental situation
this will be the interaction between @xmath69 and @xmath68 which is parameterized by the scattering length @xmath70 .
the strict-2d problem of this scenario was studied by langmack _
et al_. @xcite and in a three dimensional setting by chin and julienne @xcite .
the effect of the final state interactions is two - fold : firstly , a two - body confinement - induced bound - state appears and hence we expect the spectrum to exhibit always a bound - bound transition in addition to the usual continuum @xcite .
second , the line shape of the continuum spectrum changes as compared to the non - interacting case .
applying eq .
[ eq : fermigoldenrule ] ( fermi s golden rule ) , we find that the rf spectrum of a @xmath71 bound - state takes the form @xmath72 here , @xmath73 denotes the bound - bound transition amplitude and @xmath74 denotes the binding energy in the final state channel as a measure of the strength of the final state interactions . for the bound - bound transition amplitude
, we obtain the analytic expression @xmath75^ * \psi_b^{12}(\mathbf{r})\nonumber \\ & = & \frac{g(\xi_b^{12})-g(\xi_b^{13})}{(\xi_b^{12}-\xi_{b}^{13})\sqrt{g'(\xi_b^{12 } ) g'(\xi_{b}^{13})}}\end{aligned}\ ] ] by considering the overlap between the normalized bound - state wave functions @xmath76 , where @xmath77 refers to the @xmath71 and @xmath78 channels ( see also the supplemental material ) @xcite . here
, we introduced the dimensionless binding energies @xmath79 .
note that this result completely includes the effect of the confinement and in particular applies to molecules with a characteristic size of the order of the transverse confinement or smaller .
recently , this amplitude was also calculated numerically @xcite .
the strict 2d expression @xmath80 is obtained from eq . in the limit of small binding energies
( see also @xcite ) .
we find that eq .
provides a reasonable estimate for the magnitude of the bound - bound transition amplitude even beyond the formal limit of its validity @xmath81 . the transition rate into the continuum involves again the overlap between the @xmath71 bound - state and the unbound scattering states of the interacting final state channel .
with the analytic expression of ref .
@xcite for these eigenstates @xmath82 in the presence of short range interactions between @xmath78 atoms , we obtain the expression ( see also the supplemental material ) @xmath83^ * \psi^{12}_b(\mathbf{r})\ ; \nonumber \\ & = & \frac{\hbar^2 \phi_{n}(0 ) \left [ 1- \frac{f^{13}(e_{n \mathbf{q}})}{f^{12}(e_{n \mathbf{q}})}\right]}{m(e_{n \mathbf{q}}-e_b^{12})\sqrt{\pi l_z g'(e_b^{12}/\omega_z)}}\end{aligned}\ ] ] which contains the scattering amplitude between atoms in states @xmath84 and @xmath85 at relative energy @xmath86
@xcite : @xmath87 after summation over final states in fermi s golden rule , the transition rate into the continuum takes the form @xmath88 where effect of final state interactions on the continuum spectrum is described by the factor @xmath89 when considering dissociation into the lowest axial mode at sufficiently low energies @xmath90 , the scattering amplitudes @xmath91 may be approximated as @xmath92 where @xmath93 @xcite . in this approximation
, @xmath94 becomes @xmath95+i \pi}{\ln[\epsilon_{13}/(\hbar\omega - e_b^{12})]+i \pi } \right|^2\end{aligned}\ ] ] with @xmath96 .
the energy scales @xmath97 appearing in eq .
are analogous to the scattering length @xmath98 in three dimensions in the sense that they are the single parameter describing low energy collisions between atoms in the hyperfine states @xmath84 and @xmath85
. however , note that @xmath97 is different from the binding energy @xmath99 for deeply bound molecules under quasi-2d confinement .
we emphasize that eq . holds for any interaction strength at low energies ( i.e. as long as @xmath90 ) .
this line shape is in general different from what would be obtained by using the strict two - dimensional formula of refs . @xcite and replacing the 2d binding energies with the actual quasi two - dimensional binding energies @xmath100 . in figure
[ fig : rfexperiment ] we display our measured rf spectra and compare with three different theoretical models : strict-2d and quasi-2d with no final state interactions , and quasi-2d with final state interactions .
generally , we find that the quasi-2d theories are more strongly peaked near the binding energy which appears to fit the data better .
this difference in line - shape is particularly striking when the initial state is a tightly bound molecule or if the interactions in the final state are strongly repulsive ( i.e. @xmath101 ) .
in contrast , for our experimental data for @xmath11k with weak repulsive final state interactions ( @xmath102 over the range magnetic fields considered here ) , a quasi-2d line shape that does not take into account final state interactions still provides a good fit , whereas the strict-2d theory suggests a lesser - peaked line shape .
the recent duke experiment @xcite , on the other hand , dissociates a weakly bound molecule into an outgoing channel with weak attractive interactions . in their case ,
@xmath103 ( @xmath74 ) approximately coincide with @xmath104 ( @xmath105 ) , explaining why their strict-2d approach gives a good result for the line shape of a quasi-2d system . in order to gain deeper insight into the more complicated and fundamentally important many - particle properties of confined quantum gases @xcite , a firm understanding of the interplay of confinement and few particle physics is crucial . in this rapid communication we provided analytic results for rf spectra of dilute quasi-2d paired gases in the two - body limit and compared those results to experimental data .
additionally , we demonstrated how effective range corrections contribute to the binding energy of confined quantum gases by explicitly comparing theory to experimental data . beyond quasi-2d quantum gases
, we believe our results can be generalized to quasi - one dimensional paired fermi gases as experimentally studied in refs .
@xcite .
we gratefully acknowledge useful discussions with nigel cooper and jesper levinsen . the work has been supported by epsrc through grants no .
ep / g029547/1 ( m. k. ) and no .
ep / i010580/1 ( s. b. ) , the daimler - benz foundation ( b.f . ) , studienstiftung , and daad ( m.f . ) . 10 k. martiyanov , v. makhalov , and a. turlapov , phys .
rev .
lett . * 105 * , 030404 ( 2010 ) .
b. frhlich , m. feld , e. vogt , m. koschorreck , w. zwerger , and m. khl , phys .
rev .
lett . * 106 * , 105301 ( 2011 ) .
p. dyke , e. d. kuhnle , s. whitlock , h. hu , m. mark , s. hoinka , m. lingham , p. hannaford , and c. j. vale , phys .
rev .
lett . * 106 * , 105304 ( 2011 ) .
a. t. sommer , l. w. cheuk , m. jen - hao ku , w. s. bakr , and m. w. zwierlein , phys .
rev .
lett . * 108 * , 045302 ( 2012 ) .
m. feld , b. frhlich , e. vogt , m. koschorreck , and m. khl , nature * 480 * , 75 ( 2011 ) .
y. zhang , w. ong , i. arakelyan , and j. e. thomas , phys .
rev .
lett .
* 108 * , 235302 ( 2012 ) .
m. randeria , j .-
m . duan , and l .- y .
shieh , phys .
rev .
lett .
* 62 * , 981 ( 1989 ) .
d. s. petrov and g. v. shlyapnikov , phys .
rev .
a * 64 * , 012706 ( 2001 ) .
m. randeria , n. trivedi , a. moreo , and r. t. scalettar , phys .
rev .
lett . * 69 * , 2001 ( 1992 ) .
n. trivedi and m. randeria , phys .
rev .
lett . * 75 * , 312 ( 1995 ) . c. langmack , m. barth , w. zwerger , and e. braaten , phys .
rev .
lett . * 108 * , 060402 ( 2012 ) .
r. schmidt , t. enss , v. pietil , and e. demler , phys .
rev .
a * 85 * , 021602 ( 2012 ) .
vudtiwat ngampruetikorn , jesper levinsen , and meera m. parish , europhys . lett . * 98 * , 30005 ( 2012 ) .
v. pietil , arxiv:1203.6054 ( 2012 ) .
marco koschorreck , daniel pertot , enrico vogt , bernd frohlich , michael feld , and michael kohl , nature * 485 * , 619 ( 2012 ) .
i. bloch , j. dalibard , and w. zwerger , rev .
mod .
phys . * 80 * , 885 ( 2008 ) .
d. blume and chris h. greene , phys .
rev .
a * 65 * , 043613 ( 2002 ) .
e. l. bolda , e. tiesinga , and p. s. julienne , phys .
rev .
a * 66 * , 013403 ( 2002 ) .
z. idziaszek and t. calarco , phys .
rev .
a * 71 * , 050701 ( 2005 ) . d. b. m. dickerscheid and h. t. c. stoof , phys . rev .
a * 72 * , 053625 ( 2005 ) .
p. naidon , e. tiesinga , w. f. mitchell , and p. s. julienne , new journal of physics * 9 * , 19 ( 2007 ) .
n t zinner , journal of physics a : mathematical and theoretical * 45 * , 205302 ( 2012 ) .
v. v. flambaum , g. f. gribakin , and c. harabati , phys .
rev .
a * 59 * , 1998 ( 1999 ) .
b. gao , phys .
rev .
a * 58 * , 4222 ( 1998 ) .
g. f. gribakin and v. v. flambaum , phys . rev .
a * 48 * , 546 ( 1993 ) . c. chin , r. grimm , p. julienne , and e. tiesinga , rev .
mod .
phys .
* 82 * , 1225 ( 2010 ) .
paul s. julienne and bo gao , aip conference proceedings * 869 * , 261 ( 2006 ) . for the feshbach resonance considered here ,
the width parameter is @xmath106 @xcite .
g. bertaina and s. giorgini , phys .
rev .
lett . * 106 * , 110403 ( 2011 ) .
e. braaten , d. kang , and l. platter , phys .
rev .
lett . * 104 * , 223004 ( 2010 ) . c. chin and p. s. julienne , physical review a * 71 * , ( 2005 )
. meera m. parish , phys .
rev .
a * 83 * , 051603 ( 2011 ) .
ville pietil , david pekker , yusuke nishida , and eugene demler , phys .
rev .
a * 85 * , 023621 ( 2012 ) . j. levinsen and s. k. baur , arxiv:1202.6564 ( 2012 ) .
h. moritz , t. stferle , k. gnter , m. khl , and t. esslinger , phys .
rev .
lett . * 94 * , 210401 ( 2005 ) .
y. liao , a. s. c. rittner , t. paprotta , w. li , g. b. partridge , r. g. hulet , s. k. baur , and e. j. mueller , nature * 467 * , 567 ( 2010 ) . |
the term _ magnetic dot array _ ( mda ) @xcite refers to the family of nano - scaled monolayer structures consisting of the identical magnetic nanoparticles , called dots , which are periodically ordered on a non - magnetic substrate .
mda concept is compelling , partly because of qualitatively new properties , that essentially differ from those of the bulk materials .
typical for mda physics is an intricate collective behavior .
its understanding may be valuable for condensed matter physics , material science and nanoscience .
the special properties of mda systems follow from the interplay between intra - dot and inter - dot interactions as well as from the interplay of anomalously large surface compared to the bulk magnetic energy contributions @xcite .
the mda properties are already utilized in technological applications concerning magnetic field sensors @xcite and reading heads of magnetic - disk data - storage devices @xcite .
the technology of fabrication of mdas @xcite has been perfected to an excellent degree in the past few years . in practice , however , an occurrence of technological defects and local irregularities has still a great influence on all of the magnetic properties @xcite . in this paper
we study how the imperfection in form of the single - dot vacancy affects remagnetization of mda under the assistance of thermal activation . in the series of papers @xcite the authors focus on remagnetization of small segments of arrays by simulating elementary models where dots are treated as interacting point dipoles . in ref.@xcite
we draw the effect of uniaxial anisotropy induced by the eccentrically placed node defect .
later @xcite we have investigated how the central defect affects the quasi - static zero - temperature remagnetization of mda . from this study we know that : ( i ) the square lattice seems to be more sensitive to defect occurrence than the triangular one ; ( ii ) zero - temperature differences of hysteresis loops are not very pronounced despite of the remarkable changes in the local arrangement of dots .
in addition , the study has opened question of defect influence on the relaxation modes . in the present study
the effect of nonmagnetic central - node defect , in further referred as _ defect _ only , is reconsidered . however , the additional realistic factor included here is the thermal activation . in such case
the statistical treatment of results is necessary .
therefore , in order to discern defect consequences , two distinct mda arrangements -_defect - free _ ( df ) and _ defect - including _ ( di ) are compared .
because the statistical simulations need computational effort we preferred use of elementary model where each dot is described by the point magnetic dipole .
this simplification is justified for monodomain isotropic nearly spherical ferromagnetic particles separated by a sufficient lattice spacing several times exceeding a dot diameter . additionally , we focus on the small mda samples where dots are placed on square @xmath1 lattice .
the magnetic state of @xmath2-th dot is described by the effective rescaled 3d magnetic moment @xmath3 normalized as @xmath4 .
the inter - dot interactions are assumed to be dipolar and described by the effective field @xmath5 where @xmath6 is the distance between @xmath2-th and @xmath7-th dot in lattice - spacing units @xmath8 .
the field is measured in the @xmath9 units including the dot volume @xmath10 and saturated magnetization @xmath11 . to study the di arrangement we assume that defect is represented by the zero magnetic moment .
the dynamics of magnetic moments is described by the stochastic landau - lifshitz - gilbert equation @xcite @xmath12 where @xmath13 is the dimensionless damping parameter , @xmath14 is the time in @xmath15^{-1}$ ] units and @xmath16 is the gyromagnetic ratio .
the effective field @xmath17 includes the dipolar field @xmath18 , the external field @xmath19 , and the langevin thermal field @xmath20 .
finally @xmath21 .
the random thermal field @xmath20 is defined by averages @xcite [ sprop ] @xmath22 where @xmath23 and @xmath2 , @xmath7 are the site indexes ; @xmath24 is the noise amplitude . according to fluctuation - dissipation relation @xcite , the factor @xmath24 is linked to the temperature @xmath25 @xmath26 where @xmath27 is the characteristic temperature scale .
in all of our numerical experiments we set @xmath28 for mda . the main empirical argument supporting our choice @xmath29 is that ferrite nanoparticles have the mean value of @xmath13 of such order @xcite .
let us first study the remagnetization in df and di systems in the time varying external magnetic field applied in parallel to one of the main mda axes @xmath30 . for simulated geometry
similarly as in the hypothetical experimental setup the quantity of interest is the magnetization projection @xmath31 .
the component @xmath32 has been cycled within the bounds @xmath33 .
the simulation starts from the nearly saturated state with @xmath34 at @xmath35 , where @xmath36 is the bound chosen to keep the system of moments saturated . in the remagnetization regime
each time - integration step @xmath37 is accompanied by the unique change of the external field @xmath38 . both quantities
define the sweeping rate @xmath39 , where @xmath40 .
we observed that numerical results depend on the sweeping rate .
numerical integration of eq .
( [ sllg ] ) has been performed using stochastic predictor - corrector heun scheme @xcite .
this choice is justified by the fact that in general , the statistical error of the scheme can be made arbitrarily small by averaging over the number of stochastic paths @xcite .
we tested numerical stability of scheme for two different integration steps . in order to fix @xmath41 ,
both @xmath42 and @xmath43 have been rescaled by @xmath44 .
the test confirmed the invariance of statistical results with respect to the rescaling . under the conditions of thermal activation we recorded and treated assembly of @xmath45 independent loops .
the treatment assumes averaging of the magnetization data conditioned by @xmath46 and @xmath47 , respectively . in fig.[mloops ] we show averaged loops constructed for @xmath48 , @xmath49 and @xmath50 .
we see that loops differ for di mda and df mda variants .
the difference confirms some anomalous impact of the local assymetry of the couplings broken by defect .
clearly , by increasing the temperature the hysteresis vanishes due to reduction of the impact of irreversible processes due to configurations separated by energy barriers . ,
( b ) @xmath51 , ( c ) @xmath52 . ]
when assume standardly that the amount of free energy responsible for the irreversible processes is proportional to the area @xmath53 of hysteresis loop . according fig.[areas ] the thermally induced hysteresis reveals maximum of the ratio @xmath54 at @xmath55 .
it is clear that for sufficiently high temperatures the differences between di and df systems vanish .
since the remagnetization process of mdas is very inhomogeneous we analyzed the local mean magnetic hysteresis for nearest neighborhood of central node , corners and middle nodes of mda edges .
the local loops differ among each other . from their comparison
it follows that defect affects not only hysteresis of its nearest neighbor dots , but also hysteresis at corners and edges , that clearly follows from the long - range nature of magnetostatic couplings .
one can expect that averaging eliminates some kind of the relevant defect - sensitive information .
therefore , it seems valuable to pay attention to the fluctuations around the averaged loops . to do this we analyzed the noise formed by returns
@xmath56 accumulated during reversals .
also this data have been treated separately for @xmath57 and @xmath58 sweeps .
evidently , the probability density functions of returns exhibit defect sensitivity .
they has been quantitatively characterized by the parameters of leptocurticity @xmath59 and skewness @xmath60 .
their numerical values are listed in table.[tab : ks ] .
.statistical characteristics of the magnetization returns .
the leptocurticity and skewness obtained from fluctuations of hysteresis at different temperatures . [ cols="^,^,^,^,^ " , ] the values clearly indicate that magnetization noise is strongly non - gaussian .
we see that temperature affects the characteristics in common manner , they decrease .
however , the remarkable differences between df and di systems exist again .
the reversal magnetization paths uncovered complex collective inner behavior of dot moments . in addition the thermal fluctuations cause that single remagnetization events qualitatively differ between each other .
we observed that intervals with smoothly varying magnetization , where the quasi - coherent rotation of moments prevails , are broken by the irreversible jumps .
to understand redundant but essential attributes of the reversal statistics we exploited the abilities of artificial neural networks .
the most efficient for our purposes seems to be the usage of unsupervised self - organizing maps ( som ) @xcite that allow us to extract several representative paths from the largely redundant assembly . in fig.[koh ] we show the representative paths ( @xmath48 ) extracted by som . for certain inner parameters of som the classification yields two representative groups of loops ( classification is done for df and di cases separately ) .
the quantitative output of this analysis is that the averaged loops from fig .
[ mloops](a ) can be understood as a mixture of 89% of ( a1 ) constituent with small addition of 11% of ( a2 ) in df case , and 65% of ( b1 ) with 35% contribution of ( b2 ) in di case .
this classification explains why the patterns ( a1 ) , ( b1 ) are much more similar to the averaged loops than less probable types ( a2 ) , ( b2 ) . revealed by som network for df and di array .
the boldface branches belong to @xmath61 . ] by analyzing the revealed loops we conclude that two massive magnetization jumps contribute to the reversal .
deeper insight to the mechanism of their nucleation is offered by the configuration of moments .
the sequence of snapshots that correspond to @xmath58 regime is shown in fig.[conf ] .
we see that the defect evidently supports the nucleation of the intermediate inter - dot _ leaf_. without defect the noncolinear antiferromagnetic order is prefered .
equal to ( 1 ) 1.5 , ( 2 ) 0.0 , ( 3 ) -1.0 , ( 4 ) -3.0 . ]
the additional principle of defect detection is provided by relaxation process in zero external field starting from @xmath62 state .
the relaxation curves for df and di systems are plotted in fig.[rel ] . for @xmath63
the qualitative differences in the relaxation of magnetization become clear .
the most anomalous aspect of relaxation of di is the pronounced peak .
however , df mda does not exhibit this feature .
the configurations show that in df case the pair of antiparallel vortices is formed in contrast to inter - dot flower observed in di case . as for the hysteresis , when the energy barriers are over bridged by thermal fluctuations ,
the transition to relaxation mode without peak is expected .
this scenario is confirmed by the inset of fig .
[ rel ] , where the nearly exponential relaxation occurs without substantial marks of defect sensitivity . .
inset corresponds to @xmath52 .
the final curves are averages obtained from 1000 stochastic relaxation events . ]
the simulation statistical study discusses the theoretical grounds for defect detection in mda .
it has been shown , that that the presence of non - magnetic central node defect in a a small segment of mda has non - negligible influence on dynamical properties of mda , concretely magnetic hysteresis and relaxation .
further , it shows that the including the thermal activation is not only the additional factor of realistic simulation , but also the factor that can substantially enhance the chance to identify defect . | the influence of nonmagnetic central node defect on dynamical properties of regular square - shaped @xmath0 segment of magnetic dot array under the thermal activation is investigated _ via _ computer simulations .
using stochastic landau - lifshitz - gilbert equation we simulate hysteresis and relaxation processes .
the remarkable quantitative and qualitative differences between magnetic dot arrays with nonmagnetic central node defect and magnetic dot arrays without defects have been found .
magnetic dot arrays , thermal activation , stochastic landau - lifshitz - gilbert equation , magnetic hysteresis , magnetic relaxation 61.80.az , 74.25.fy , 74.72.bk , 02.70.bf |
emulsions are mixtures of two immiscible fluids consisting of droplets of one phase dispersed into the other , stabilized against coalescence by surfactant . at low droplets volume fraction ,
the emulsions have basically a newtonian behavior @xcite .
when the volume fraction of the dispersed phase is increased , droplets come into contact .
if a small stress is applied to a dense emulsion , the interfaces between the droplets can be strained to store surface energy ; it results in an elastic response @xcite . above a yield stress , they flow as a result of droplets rearrangements @xcite .
the flow behavior of dense emulsions can be measured in classical rheological experiments , and seems to be well represented by a herschel - bulkley model @xmath1 ; in monodisperse emulsions , the exponent @xmath2 has been found to vary between @xmath3 and @xmath4 for volume fractions varying between 0.58 and 0.65 @xcite . at first glance ,
dense emulsions behavior thus seems to be well understood and modelled .
however , it has been shown recently that the behavior of dense emulsions @xcite , and more generally of pasty materials @xcite , may be much more complex than what can be inferred from simple rheometric experiments , particularly at the approach of the yield stress .
the reason is that the flow behavior of materials is usually studied in viscosimetric flows @xcite , in which one measures macroscopic quantities ( a torque @xmath5 vs. a rotational velocity @xmath6 ) . a constitutive law relating the shear stress @xmath7 to the shear rate @xmath8 in any flow
can then be derived easily from these macroscopic measurements provided that the flow is homogeneous . however , in complex fluids , this last requirement may not be fulfilled as the flow may be localized , as observed in many pasty materials @xcite , even in the cone - and - plate geometry @xcite .
another problem of importance , that can complicate the analysis of the rheological data , is the existence of slippage of dense emulsions at the walls @xcite even with slightly roughened walls @xcite .
these problems can be got round by coupling macroscopic torque measurements and local measurements of the shear rate , e.g. through mri techniques @xcite , optical methods @xcite or sound methods @xcite , in order to account for these heterogeneities and to get a proper constitutive law .
local measurements in dense emulsions have recently yielded surprising results @xcite : in some emulsions , it is impossible to find a single constitutive law compatible with all flows .
the constitutive law of the material then seems to depend on the velocity at the inner cylinder in a couette cell @xcite or the pressure gradient in a poiseuille cell @xcite .
three main reasons have been evoked to explain the apparent absence of a constitutive law : thixotropy , non local effects , and shear - induced migration of the droplets .
thixotropy may be a cause from the apparent absence of a constitutive law .
long times may indeed be needed to reach a steady state . when the measurements are not performed in this steady state
, it may then leave the idea that different laws are required to describe the flow of the sample . such long times to reach steady state have been evidenced in some experiments dealing with highly adhesive emulsion @xcite .
importantly , @xcite , although they claim to have reached a steady state , were unable to account for the flows of their adhesive emulsion with a single constitutive law .
note that we will study again the emulsion of @xcite in this paper .
the apparent absence of a single local constitutive law may also be the signature of non - local phenomena .
@xcite have actually proposed a non local model to describe the flows of jammed materials ( and thus dense emulsions ) , based upon the cooperativity of the flow . in these systems
, flow occurs via a succession of elastic deformation and irreversible plastic events .
these localized plastic events induce a relaxation of the stress on the whole system .
their zone of influence is quantified by a length @xmath9 . in presence of high stress gradient or in the vicinity of a surface , this process creates non local effects : indeed , in these situations , the rate of rearrangements of the neighboring fluid differs from what would be the bulk rearrangement rate . as a consequence , these effects affect drastically the rheological behavior when @xmath9 is comparable to the size of the confinement .
such behavior has been recently pointed out by @xcite . from experimental data ,
the authors show that the cooperativity length @xmath9 is zero below the jamming concentration @xmath10 and is typically a few oil droplet diameters above @xmath10 . using a wide gap
should then prevent us from being sensitive to these non local effects .
we will check this feature by studying the flows of the @xcite emulsion in this paper .
particle migration may also be a cause for the apparent lack of a local constitutive law @xcite .
in multiphase materials , wide gap couette flows are indeed known to lead to concentration heterogeneities due to shear - induced migration of the dispersed elements toward the low shear zones .
the migration of deformable particles , and particularly drops , has been much less studied than the one of rigid particles .
single droplets have been found to migrate away from the rigid walls in any shear flow , due to asymmetric flows around deformed droplets @xcite .
this leads to an equilibrium position somewhere between the walls , exactly at the center of the gap in the case of a narrow gap couette geometry , nearer from the inner cylinder in the case of a wide gap couette geometry @xcite .
when dealing with a dilute emulsion , it has been observed , in contrast with dilute suspensions of rigid particles , that the equilibrium distribution of droplets is parabolic around a position that is near the equilibrium position of single droplets @xcite .
the theoretical explanation is that all particles would tend to accumulate at the same equilibrium position between the walls , but their distribution is broadened by their binary collisions ( this second process may be modelled as it is for rigid particles ) .
only rather dilute emulsions ( up to 10% of droplets ) have been studied so far , and nothing is known about what happens for higher droplet concentrations . in this case , one would expect the wall effect on the droplet migration to be negligible compared to the effect of the interactions between droplets ; then , one would expect migration to product the same effects as in suspensions , with the same kinetics as it has basically the same physical origin .
we can thus expect that relevant information for migration in dense emulsions can be found in the literature dealing with suspensions of rigid particles . in suspensions of noncolloidal rigid particles , migration has been observed in many situations : wide - gap couette flows @xcite , parallel - plate flows @xcite , pipe flows @xcite . in wide - gap couette flows ,
the consequence of migration is an excess of particles near the outer cylinder .
migration is related to the shear - induced diffusion of particles @xcite : the gradients in shear rate that exist in all but the cone and plate geometry generate a particle flux towards the low shear zones ( the outer cylinder in the case of the couette geometry ) , which is counterbalanced by a particle flux due to viscosity gradients . in an alternative model
@xcite , particle fluxes counterbalance the gradients in normal stresses .
most experiments @xcite observe that migration in suspensions of volume fraction up to 55% is rather slow , in accordance with its diffusive origin .
however , it was recently found in a couette geometry that in the case of very dense suspensions ( up to 60% ) , which exhibit an apparent yield stress , migration is almost instantaneous @xcite ( it lasts for a few revolutions ) so that it may be unavoidable .
a reason for this very fast kinetics may be that particles are closely packed together : any shear may then push the particles towards the outer cylinder with an instantaneous long range effect on the particle concentration .
migration may then be expected to be very fast for dense emulsions exhibiting a yield stress , i.e. in which droplets are in contact as the particles in the very dense suspensions of @xcite . in this paper , we address the questions of the existence and the determination of a local constitutive law for the flows of dense emulsions . in this aim , we use a wide gap couette cell to avoid non local effects and we study in detail the question of shear - induced droplet migration in these systems .
we also take care of performing our experiments in a steady state .
we propose to couple macroscopic measurements and local measurements of concentration and velocity through mri techniques during the wide gap couette flows of dense emulsions .
a new method designed to measure the local droplet concentration in emulsions is developed and we seek for the occurrence of migration on many formulations .
local measurements of the constitutive law are then compared to purely macroscopic measurements . in sec .
[ section_display ] , we present the materials and the experimental setup .
we present the concentration profiles obtained on all materials after long time experiments in sec .
[ section_concentration ] .
the velocity profiles are shown in sec .
[ section_velocity ] .
the macroscopic rheometrical measurements are displayed in sec .
[ section_torque ] , and a purely macroscopic determination of the constitutive law is presented .
the results are analyzed in sec .
[ section_local ] : we determine locally the constitutive law of the material from the velocity profiles and we compare macroscopic and local measurements of the constitutive law .
as the timescale for the occurrence of migration is _ a priori _ very sensitive to the droplet size @xmath11 ( if a diffusive process is involved , it should scale as @xmath12 ) , we study dense emulsions of 4 different sizes : 0.3@xmath0 m , 1@xmath0 m , 6.5@xmath0 m , and 40@xmath0 m . moreover , as problems in defining a local constitutive law seem to be linked to the adhesion properties of the droplets @xcite , we took care of formulating adhesive and non adhesive emulsions .
note in particular that we study the emulsions of @xcite and @xcite which posed problems in defining a single local constitutive law accounting for all their flows .
the 0.3@xmath0 m emulsion is the adhesive emulsion of @xcite ; nearly monodisperse oil in water emulsions are prepared by shearing a crude polydisperse emulsion , composed of castor oil droplets in water stabilized by sodium dodecyl sulfate ( sds ) , within a narrow gap of 100 @xmath0 m @xcite .
the surfactant concentration within the aqueous phase is set to 8% wt . which leads to an adhesive emulsion @xcite .
the oil volume fraction is 73% and the droplet
mean diameter is 0.3 @xmath0 m with a polydispersity of about 20% which is enough to prevent crystallization .
the 1@xmath0 m emulsion is a non adhesive emulsion .
the preparation of the sample is the same as the one used to get the adhesive 0.3@xmath0 m emulsion .
it is composed of silicone oil in water stabilized by sodium dodecyl sulfate ( sds ) .
the surfactant concentration within the aqueous phase is set to 8.5% .
the viscosity of the oil is 1 pa.s and the oil volume fraction is 75% .
we prepared adhesive and non - adhesive oil in water emulsions , of 6.5 @xmath0 m mean diameter with a polydispersity of about 20 @xmath13 .
these emulsions are composed of silicone droplets in a mixture of glycerine and water ( 50% w glycerine , 50% w water ) stabilized by 1% wt of brij and trimethyl tetradecyl ammonium bromide , sheared in a narrow gap couette cell .
the trimethyl tetradecyl concentration within the aqueous phase is set to 1 or 6.5 wt @xmath13 , which leads to respectively adhesive and non - adhesive emulsions .
the oil viscosity is 1pa.s .
the oil volume fraction is 75% in all cases .
note that the non - adhesive 6.5@xmath0 m emulsion is the same as in @xcite .
note that the use of glycerol in the continuous phase prevents from obtaining accurate concentration measurements with the method developed below .
only qualitative results can be inferred from the measurements : in particular , if the intensity profiles are unchanged after a flow , it means that there is no migration .
we finally studied a large droplet brine in oil non - adhesive emulsion , of 40 @xmath0 m mean diameter with a polydispersity of about 50 @xmath13 , at a 88% concentration .
the emulsion is prepared by progressively adding the brine ( solution of cacl@xmath14 at 300g / l in water ) in an oil - surfactant solution ( hdf 2000 ( total solvent ) + sorbitan monooleate at 2% ) under a controlled high shear with a silverson l4rt mixer .
we carried out experiments with a velocity controlled magnetic resonance imaging ( mri ) rheometer " .
experiments are performed within a wide - gap couette geometry ( inner cylinder radius @xmath15 cm , outer cylinder radius @xmath16 cm , height @xmath1711 cm ) . in order to avoid slip at the walls ,
sandpaper is glued on the walls .
rheometric measurements are performed with a commercial rheometer ( bohlin c - vor 200 ) that imposes either the torque or the rotational velocity ( with a torque feedback ) .
the whole geometry is inserted in a 0.5 t vertical mri spectrometer ( 24/80 dbx by bruker ) .
the mri measurements allow us to get the local orthoradial velocity and the local droplet concentration of the material anywhere in the gap , with a radial resolution of 0.5 mm . for details on the mri sequence devoted to measure
the local velocity and its application to rheology , see @xcite .
a method devoted to obtain the local droplet concentration was developed specifically for this study and is presented hereafter .
proton nmr relies on the properties of hydrogen nuclei , which bear a magnetic dipole . in a nmr experiment
, the sample is put inside a strong , static , and ( as far as possible ) homogeneous field @xmath18 .
the magnetization of hydrogen nuclei then tend to align along the static field .
the sample develops a non zero magnetization density @xmath19 which follows the curie law : m_o(r)=_h(r)b_o[eq_magnetization]where @xmath20 is the hydrogen density at position @xmath21 , @xmath22 is the gyromagnetic ratio of hydrogen nuclei , @xmath23 is the planck constant , @xmath24 is the boltzmann constant , and @xmath25 is the sample temperature .
@xmath20 can be easily computed for any sample from its chemical formula and its density . when put out of equilibrium , magnetization of liquids
often develops a threefold dynamic behavior combining a precession motion around the permanent field of the magnet , and two relaxation processes : a mono - exponential decay of transverse components of @xmath26 with a characteristic time @xmath27 , and a mono - exponential relaxation of the longitudinal component of @xmath26 towards its equilibrium value , with a characteristic time @xmath28 .
nuclear magnetic resonance technique consists in series of magnetization manipulation and measurements by means of externally applied additional magnetic field , aiming at getting information about physico - chemical properties of the sample . the magnetic resonance imaging extension of nmr permits to get , as far a possible , a
space resolved view of the local magnetization density , and can make such analysis local and non - perturbative .
because of signal to noise ratio limitations , the space resolution of mri facilities for imaging purposes can never be much better than about one hundredth of the size of the experimental setup , that is 1 mm in our working conditions , which is much too big to resolve independent droplets .
measuring the water / oil ratio in a sample then requires to have a way to distinguish oil and water magnetization inside a single pixel of a mri image . up to now
, three main discrimination routes were already reported in emulsion literature .
the first one relies on the so - called chemical shift difference between water and oil .
actually , the exact precession frequency around the permanent field is closely related to the direct chemical neighborhood of hydrogen atoms . because all hydrogen nuclei of water are linked to oxygen atoms , and most of hydrogens in oily material are linked either to carbon or silicium atoms , water and oil magnetization always exhibit two different precession frequency .
this chemical shift is quite small ( about 4 - 4.5 ppm relative difference , whenever carbon or silicon oils are used ) .
nevertheless this shift can be accurately observed provided that @xmath29 inhomogeneities across the sample have a much smaller amplitude than 4.5ppm of the mean @xmath30 value ( e.g. , see spectrum by @xcite).indeed , the precession frequency is also proportional to local @xmath31 intensity , and one should avoid confusion between the original chemical shift and any spurious one originating from field heterogeneities .
homogeneity constraints can be met when the fluid is embedded in a container with a simple - shape .
reported studies making use of the chemical shift contrast then often deal with emulsions that are contained in a simple beaker , or even a cylindrical pipe displayed parallel to @xmath29 .
quite early results deal with the use of chemical shift imaging ( csi ) for a quantitative assessment of oil / fat content in food emulsions @xcite , or the separation of oil and water components of an emulsion during a filtration process @xcite .
more recently , @xcite and @xcite managed to carry out quite well resolved local measurements of water / oil concentrations and/or velocities in model suspensions flowing through a pipe in a laminar regime .
unfortunately , the chemical shift contrast can hardly be transposed in the case of a couette cell .
the geometry is indeed more complex and involves edges between materials with different magnetic susceptibilities , prone to generate strong local field distortions .
csi applications to couette flow are thus more difficult and actually quite rare . for instance , @xcite , who studied the mixing process of oil and water in an horizontal couette like device reported quite long measuring times and only poorly resolved pictures .
a second way to discriminate between water and oil in an emulsion relies on @xmath28 relaxation time .
@xmath28 ( as well as @xmath27 ) indeed strongly depend on a wide set of parameters , such as fluid composition , diffusion coefficient , temperature , or the presence of dissolved species . in water and oil
, they usually turn out to be quite different .
for instance , in our working conditions ( 0.5 t magnetic field ) @xmath28 and @xmath27 usually range between 2 and 3s in bulk water ( depending on water origin and dissolved minerals ) , while many carbonated and silicon oils exhibit @xmath28 about few 100ms , and @xmath27 even smaller . although the use of @xmath27 contrast was already reported as a promising nmr tool of characterizing droplet size in an emulsion @xcite , previous works involving imaging possibilities of nmr prefer @xmath28 contrast . as compared with chemical shift or @xmath27 , @xmath28
physical value and @xmath28 measurement techniques are indeed fairly insensitive to field inhomogeneities .
@xcite demonstrated this contrast to give quantitative information on the creaming process of an oil - water emulsion .
more recently , few works showed its ability to work in more severe conditions .
@xcite used @xmath28 to measure the fat fraction in an emulsion contained inside the stomach of a human volunteer . @xcite
also reported the use of a @xmath28 based method to follow the droplet migration in a low concentrated emulsion under shearing in a couette cell .
a third way also used to study the creaming of an emulsion @xcite , consists in preparing a water - oil emulsion based on deuterium oxide instead of normal water . because proton nmr is only sensitive to 1h nuclei , nmr or mri measurements
then bring exclusive information on the oil phase . in our work , we wanted to get independent information on oil and water phase .
field inhomogeneities inside our couette cell were also in many places of the same order of magnitude as the chemical shift .
this was mainly due to the covering of the surfaces of the cell with sand - paper , that was used so as to avoid slip problems .
@xmath28 contrast was then regarded as the best contrast source . although details of our measurement method was quite different than that of @xcite it was also based on the so - called inversion - recovery technique . starting from a magnetization in equilibrium state ,
it consists in putting magnetization upside down by means of an appropriate external field , and then to let it relax towards its equilibrium value . in a homogeneous sample with relaxation time @xmath28 and equilibrium magnetization density @xmath32
, the magnetization about the main magnetic field should then read m(t)=m_o(1 - 2e^-t / t_1)where @xmath33 at the time of inversion . to perform an accurate analysis
, we found necessary to consider additionally that the initial inversion can never be perfect : the magnetization about the main magnetic field then actually reads m(t)=m_o(1-ie^-t
/ t_1)where @xmath34 reflects the actual inversion conditions : @xmath34 would be equal to 2 in perfect experimental conditions , but was actually always found to be slightly less than 2 .
its exact value and space homogeneity depends on both the hardware and the sample . in a water - oil emulsion
, it is admitted that water and oil behave as two independent phases : the resultant magnetization is then the sum of the individual oil and water magnetization that relax independently .
the mean magnetization inside a water oil - emulsion then reads m(t)=^w m_o^w(1-i e^-t / t_1^w)+^o m_o^o(1-i e^-t / t_1^o)[eq_relaxation ] where @xmath35 and @xmath36 are the equilibrium magnetization density of bulk water and oil respectively , @xmath37 and @xmath38 are the water and oil volume fractions ( @xmath39 ) , and @xmath40 and @xmath41 are the longitudinal relaxation time of water and oil respectively .
@xmath34 is supposed to be the same in oil and water . from eq .
[ eq_relaxation ] , in order to observe only water droplets in a water in oil emulsion , @xcite used the fact that there is always a time during this relaxation process when the oil magnetization is zero .
they made a mri picture of @xmath42 at this time , based on the assumption that @xmath43 everywhere in the sample .
however , we think that the accuracy of such technique may be limited , because a wrong estimation of the actual @xmath34 factor , if not exactly 2 , or a wrong estimation of relaxation times will always lead to a wrong estimation of the time for which the oil magnetization is zero , and then , to an incomplete oil cancelling . using in situ measurements of water and oil relaxation times , we developed another approach which consists in performing 3 mri measurements of the magnetization relaxation , for 3 carefully chosen waiting times , and then in solving an equation system , letting @xmath34 as a free parameter .
we then obtain space resolved measurements of @xmath44 , @xmath45 and @xmath46 as averaged at the pixel scale .
the three waiting times are optimized for each sample in order to minimize the uncertainty on the magnetization values .
as compared with the phase cancelling technique , we calculated that this method is much more accurate , and especially less sensitive to @xmath28 uncertainties when @xmath28 values in water and oil are clearly separated . in order to get the material volume fraction from eq .
[ eq_relaxation ] , in situ measurement of relaxation times were found to be a necessary step of the experimental procedure .
indeed , as far as relaxation phenomena are considered , although water and oil mainly behave as if they would be independent phases , they interact anyway together with the surfactant at the surface of each droplet . relaxation times in water may then be influenced by the presence of the oil phase .
relaxation times in oil may also differ from that of pure oil because aliphatic chains of surfactant molecules mix with oil in the vicinity of the droplet surface .
the sensitivity to the droplet size was reported to be quite dramatic for @xmath27 value in water / oil emulsion containing a large amount of additional impurities @xcite .
we believe in such system that @xmath28 and @xmath27 should qualitatively behave the same . in our samples , which are made of quite pure
water and oil phases , the @xmath28 difference between the pure phases and the droplets was found to be of order 10% .
this difference is probably too small to provide a quantitative sensitivity of our measurements to droplet size changes in our systems .
anyway , @xmath28 changes , as small as they may be , have to be taken into account for the accuracy of local fraction measurements .
eventually , @xmath28 may also evolve with any small change in the temperature of the sample .
that is why before each concentration measurement , the actual relaxation times were measured . as @xmath28 is sensitive to the details of the local structure of the material , @xmath28 measurements should be ideally space resolved .
however , we worked here another way : we used an average @xmath28 measurement in the whole gap of the cell , discriminating water and oil on the basis of their chemical shift ( field inhomogeneities were too strong for a csi - based imaging , but were still ok for @xmath28 measurements ) .
the average @xmath28 values for oil and water were then directly used to process the mri data .
granted that processed oil and water volume fractions were then always found to be homogeneous within the gap of the cell , we had then an a posteriori confirmation that our global @xmath28 measurement were meaningful
. one should be aware anyway that if any heterogeneity would have been detected in measured volume fractions , then it would have been necessary to shift to local @xmath28 measurements to ensure very accurate quantitative results ; measurements based on the global @xmath28 measurement would nevertheless provide a relevant estimation of the volume fractions . finally , in order to obtain the local volume fraction @xmath47 and @xmath48 of each phases , it is sufficient to remember that from eq .
[ eq_magnetization ] : @xmath49 and @xmath50 .
the absolute uncertainty on the concentration measurement performed with this method was found to be of 0.3% from measurements performed on a homogeneous emulsion .
note that when the proton density of the oil phase is not known , the comparison of the @xmath45 and @xmath46 measurements performed on the homogeneous emulsion ( of known volume fraction ) allows to recover its value .
note also that a simple way to check for changes in the local concentration , and to evaluate the amplitude of these changes , consists in comparing the water and oil intensity profiles within the gap after shear to the profiles measured on the emulsion just after loading .
these last profiles are performed on a homogeneous emulsion , any change in these profiles induced by shear would then mean that there is migration . in the experiments presented hereafter , we control the rotational velocity of the inner cylinder . for all materials ,
we first apply a constant rotational velocity of 20rpm to 100rpm until 10000 to 90000 revolutions have been completed .
we record the evolution of the velocity profiles and the concentration profiles in time in order to check for the occurrence of migration ( see sec .
[ section_concentration ] ) .
afterwards , we measure the stationary velocity profiles and the stationary torque exerted by the material on the inner cylinder for various velocities ranging from 0.1rpm to 100rpm ( see sec .
[ section_velocity ] and sec .
[ section_torque ] ) .
in order to search for the occurrence of shear - induced migration in dense emulsions , we sheared all materials for long times .
we chose to apply velocities ranging between 20 and 100 rpm ; this ensures that in all cases the whole gap is sheared ( see sec .
[ section_velocity ] ) .
we also checked on the 6.5@xmath0 m emulsions that the same results are obtained when the flow is localized during the experiment ( i.e. when the material is sheared at 1 rpm ) .
the 0.3@xmath0 m emulsion was sheared for 10000 revs , the 1@xmath0 m emulsion was sheared for 90000 revs , both 6.5@xmath0 m emulsions were sheared for 35000 revs , and the 40@xmath0 m emulsion was sheared for 20000 revs .
the corresponding local strains , extracted from the velocity profiles ( see sec .
[ section_velocity ] ) range between 5 times the number of revolutions near the inner cylinder , and 0.25 times near the outer cylinder . m adhesive emulsion for 10000 revs ( squares ) , and a 40@xmath0 m non - adhesive emulsion for 20000 revs ( open circles).,width=340 ] in all the experiments , within the experimental uncertainty , we observe that the materials remain homogeneous after shear : there is no observable shear - induced migration .
results for the 0.3@xmath0 m and the 40@xmath0 m emulsion are depicted in fig .
[ fig_emul_concentration ] .
note also that the relaxation times @xmath28 of the oil and water phases were not observed to change during all the experiments : as pointed out in sec .
[ section_display ] , this is consistent with the fact that the droplet size is unchanged by shear and that all the measurements deal with the same homogeneous suspension .
consistently , optical observations of the droplets after shear did not reveal size changes . as noted in sec .
[ section_introduction ] , up to now only rather dilute emulsions have been studied : we can not compare our results with any result from the literature dealing with emulsions .
however , the relevant mechanism involved in migration of droplets in dense emulsions should be basically the same as in suspensions of rigid particles , namely a shear - induced diffusion of particles . from this model and results from the literature of migration in dense suspension
, we can thus infer whether or not migration would have been expected in our systems . in the diffusive model of migration @xcite
, the particles undergo a shear - induced diffusion characterized by a self - diffusion coefficient @xmath51 @xcite , where @xmath52 is the particle volume fraction , @xmath53 the shear rate , @xmath54 the particle size , and @xmath55 is a dimensionless coefficient whose dependence on @xmath52 may theoretically be @xmath56 for small @xmath52 .
the gradients in shear rate that exist in a couette geometry then generate a particle flux towards the outer cylinder , which is counterbalanced by a particle flux due to viscosity gradients . as qualitatively confirmed experimentally @xcite
, one would then expect the migration phenomenon to last for a number of revolutions @xmath57 until a stationary heterogeneous profile is established , where @xmath58 and @xmath59 are respectively the outer and inner radius , and @xmath60 . within the frame of this model
, we can now evaluate the expected @xmath61 for complete migration in our experiment from experimental results from literature .
actually , as shown by @xcite two inconsistent set of data exist , depending on the concentration : migration has been shown to occur much more rapidly for concentrations above 55% .
e.g , for moderately dense suspensions , @xcite find @xmath62 revs for 140@xmath0 m particles , at @xmath63 , with @xmath64 cm and @xmath65 cm . in the case of our materials , this would imply a value of @xmath66 revs for the 0.3@xmath0 m emulsion , @xmath67 revs for the 1@xmath0 m emulsion , @xmath68 revs for the 6.5@xmath0 m emulsions , and @xmath69 revs for the 40@xmath0 m emulsion . on the other hand for very dense suspensions
, @xcite found that migration occurs during the first 50 revs within the same couette geometry as our .
this would imply an expected value of @xmath70 revs for the 0.3@xmath0 m emulsion , @xmath71 revs for the 1@xmath0 m emulsion , @xmath72 revs for the 6.5@xmath0 m emulsions , and @xmath73 revs for the 40@xmath0 m emulsion in our experiments .
let us recall that @xmath61 is the number of revolutions expected for _ complete _ migration , and that heterogeneities should be observable for a value of @xmath74 revolutions ( see @xcite and @xcite for measurements of transient concentration profiles in suspensions ) .
if a shear - induced migration mechanism similar to the one observed in suspensions of rigid particles was acting in dense emulsions , we should therefore have observed migration in the 40@xmath0 m emulsion and in the 6.5@xmath0 m emulsions . on the other hand ,
our experiments can not be used to conclude directly for the failure of a diffusive model of migration in the case of the smaller droplets ( 0.3 and 1@xmath0 m ) , but we can expect to predict what happens in this case from the @xmath75 scaling of the diffusion process if this process is relevant . even if it impossible to conclude definitely for the absence of migration for any strain , we can conclude that if a diffusive process acts in dense emulsions , it is much slower than the process involved in dense suspensions of rigid particles .
importantly , our results provide reliable lower bounds for the occurrence of migration in dense emulsions . for practical purposes , in most flows and particularly in most rheometrical studies of dense emulsions
, it can therefore be claimed that the emulsions remain homogeneous .
surprisingly , our results contrast with what was observed by @xcite in dense suspensions of hard spheres where very fast migration was found to occur .
a first reason could be that jamming prevents from migration in dense emulsions .
however , it was found that migration in dense suspensions can lead to regions having a concentration higher than the maximum packing fraction @xmath10 above which there is no more flow @xcite : this shows that jamming does not necessarily prevents from migration .
another reason could be that deformability of the particles play a central role .
the shear stresses involved in the flow are of the order of 100 to 200 pa in all materials , whereas the typical stress due to surface tension range between 1000pa for the 40@xmath0 m droplets and 120000pa for the 0.3@xmath0 m particles .
the droplets may then be poorly deformable ( except for the 40@xmath0 m emulsion ) under shear , but they must be deformed due to confinement and to their high concentration .
in particular they should have flat contacts which may allow the droplets to slip easily past each other ; this is a major difference with the suspensions of hard spheres in which any shear may push the particles towards the outer cylinder with an instantaneous long range effect on the particle concentration thanks to the force chains originating from direct frictional contact forces .
note that consistency of our observations with the modelling of shear - induced migration based on normal stresses is difficult to check as normal stresses in dense emulsions are poorly known .
it should be noted however that the normal stresses due to surface tension should be of the order of the yield stress @xcite in dense emulsions , while normal stresses tend to diverge at the approach of the maximum packing fraction in dense suspensions of rigid particles ( this difference is linked to the deformability of the droplets leading to flat contacts as pointed out above ) .
it is thus not unexpected from this point of view that the kinetics of shear induced migration in a couette flow , which depends on the spatial variation of the first normal stress difference and of the radial normal stress @xcite , is very fast near the maximum packing fraction in suspensions of hard spheres while it would have to remain slow in dense emulsions .
anyway , it is not possible to be more quantitative at this stage and to state if migration would have been expected to occur within this picture in our experimental conditions . in these experiments , after a 5 min preshear at 100 rpm , we measured the velocity profiles at various rotational velocities . in all systems ,
the stationary velocity profiles were found to develop within a few seconds and to remain stable for hours .
this means that these systems are not thixotropic @xcite . in the case of the 0.3 @xmath0 m adhesive emulsion
, this contrasts with the long time evolution of the velocity profiles observed by @xcite ; we will comment on this point below .
the stability of the velocity profiles is also consistent with the absence of shear - induced droplet migration in all systems : as the behavior of dense emulsions depends on their concentration , changes in the concentration profiles in time would have led to changes in the velocity profiles in time .
this feature is important : in the absence of a method to measure the concentration profiles , the stability of the velocity profiles can be used as a proof of the absence of migration as long as the migration process is not nearly instantaneous . in fig .
[ fig_emul_velocity1 ] , we plot the dimensionless velocity profiles for the steady flows of all the emulsions for various rotational velocities ranging from 0.3 to 100rpm . m adhesive emulsion , at various rotational velocities ranging from 5 to 100rpm ; the line is the theoretical dimensionless velocity profile for a newtonian fluid .
b ) same plot for a 1@xmath0 m non - adhesive emulsion .
c ) same plot for a 6.5@xmath0 m adhesive emulsion .
d ) same plot for a 6.5@xmath0 m non - adhesive emulsion .
e ) same plot for a 40@xmath0 m non - adhesive emulsion.,title="fig:",width=302 ] m adhesive emulsion , at various rotational velocities ranging from 5 to 100rpm ; the line is the theoretical dimensionless velocity profile for a newtonian fluid .
b ) same plot for a 1@xmath0 m non - adhesive emulsion .
c ) same plot for a 6.5@xmath0 m adhesive emulsion .
d ) same plot for a 6.5@xmath0 m non - adhesive emulsion .
e ) same plot for a 40@xmath0 m non - adhesive emulsion.,title="fig:",width=302 ] m adhesive emulsion , at various rotational velocities ranging from 5 to 100rpm ; the line is the theoretical dimensionless velocity profile for a newtonian fluid .
b ) same plot for a 1@xmath0 m non - adhesive emulsion .
c ) same plot for a 6.5@xmath0 m adhesive emulsion .
d ) same plot for a 6.5@xmath0 m non - adhesive emulsion .
e ) same plot for a 40@xmath0 m non - adhesive emulsion.,title="fig:",width=302 ] m adhesive emulsion , at various rotational velocities ranging from 5 to 100rpm ; the line is the theoretical dimensionless velocity profile for a newtonian fluid .
b ) same plot for a 1@xmath0 m non - adhesive emulsion .
c ) same plot for a 6.5@xmath0 m adhesive emulsion .
d ) same plot for a 6.5@xmath0 m non - adhesive emulsion .
e ) same plot for a 40@xmath0 m non - adhesive emulsion.,title="fig:",width=302 ] m adhesive emulsion , at various rotational velocities ranging from 5 to 100rpm ; the line is the theoretical dimensionless velocity profile for a newtonian fluid .
b ) same plot for a 1@xmath0 m non - adhesive emulsion .
c ) same plot for a 6.5@xmath0 m adhesive emulsion .
d ) same plot for a 6.5@xmath0 m non - adhesive emulsion .
e ) same plot for a 40@xmath0 m non - adhesive emulsion.,title="fig:",width=302 ] the 1@xmath0 m non adhesive emulsion , the 6.5@xmath0 m adhesive and non adhesive emulsions , and the 40 @xmath0 m non adhesive emulsion share a common behavior .
mri measurements show that the velocity profiles are curved , and that they occupy only a small fraction of the gap at low rotational velocities ( fig .
[ fig_emul_velocity1 ] ) : in this case , the velocity tends to zero within the measurement uncertainty at some radius before the outer cylinder .
the fraction of the material that is sheared increases with the rotational velocity . beyond a critical velocity @xmath76 , that depends on the emulsion and is of order 3rpm in the case of the 6.5@xmath0 m adhesive emulsion ( fig . [ fig_emul_velocity1]c ) , or of the order 50 rpm in the case of the 1@xmath0 m non - adhesive emulsion ( fig .
[ fig_emul_velocity1]b ) , the whole sample is sheared .
this shear localization is a classical feature of yield stress fluids flows in couette geometry where the shear stress is a decreasing function of the radius : the flow must stop at a radius @xmath11 such that the shear stress equals the material yield stress at this place . note that this behavior is not evidenced for the non adhesive 6.5@xmath0 m emulsion and for the non adhesive 40@xmath0 m emulsion : this is simply due to the fact that the applied rotational velocities are too high to probe this behavior .
when all the gap is sheared , the velocity profiles are clearly different from those of a newtonian fluid ( fig .
[ fig_emul_velocity1 ] ) .
the curvature we observe is due to the use of a wide - gap couette geometry and is actually typical of shear - thinning fluid : the shear rate decreases more rapidly within the gap ( i.e when the shear stress decreases ) than for a newtonian fluid .
the shear rate may vary here of a factor 20 within the gap while it would vary of a factor 2.1 for a newtonian fluid in this geometry ( this factor 2.1 is simply the ratio between the shear stress at the inner cylinder and the shear stress at the outer cylinder , see eq . [ eq_localstress ] in sec .
[ section_torque ] ) .
this implies that in the analysis of macroscopic experiments , one can not simply compute a mean shear rate within the gap to characterize the flow . finally , note that no data can be recorded close to the inner cylinder . as a consequence ,
the expected absence of wall slip ( due to the use of rough surfaces of roughness larger than the droplet size ) can not be proved at this stage .
it can only be observed that the extrapolation of the dimensionless velocity profiles to a value of 1 at the inner cylinder is reasonable , and that wall slip if any should be of a limited amount of order 5% or less .
we will show in sec .
[ section_local ] that it is possible to infer the emulsion velocity at the walls from its local constitutive law , and that wall slip if any is actually less than 1% of the inner cylinder velocity . in log scale
( for readability , not all profiles are plotted ) .
the solid lines of fig .
[ fig_emul_velocity1log]a are the velocity profiles predicted by a herschel - bulkley law.,title="fig:",width=302 ] in log scale ( for readability , not all profiles are plotted ) .
the solid lines of fig .
[ fig_emul_velocity1log]a are the velocity profiles predicted by a herschel - bulkley law.,title="fig:",width=302 ] in log scale ( for readability , not all profiles are plotted ) . the solid lines of fig .
[ fig_emul_velocity1log]a are the velocity profiles predicted by a herschel - bulkley law.,title="fig:",width=302 ] in log scale ( for readability , not all profiles are plotted ) .
the solid lines of fig .
[ fig_emul_velocity1log]a are the velocity profiles predicted by a herschel - bulkley law.,title="fig:",width=302 ] in log scale ( for readability , not all profiles are plotted ) . the solid lines of fig .
[ fig_emul_velocity1log]a are the velocity profiles predicted by a herschel - bulkley law.,title="fig:",width=302 ] at first sight , the 0.3@xmath0 m adhesive emulsion presents the same behavior as the other samples .
the velocity profiles are curved , and they occupy only a small fraction of the gap at low rotational velocities . however , focusing on the low velocities part and thus very low shear rates part of the profiles , a striking behavior is evidenced in the semi - logarithmic scale figure ( fig . [ fig_emul_velocity1log]a ) . a slope break in the profile
is evidenced at a given radius inside the gap .
the velocity then starts to decrease very slowly with the radius .
there is then a large zone of the material in which there remains a very slow flow .
this behavior is reminiscent of the one observed previously by @xcite and by @xcite in adhesive emulsions ( this behavior will be shortly commented in sec . [ section_local ] ) .
these slowly varying slow flows occur here for velocities lower than a few [email protected]@xmath77 .
this behavior seems to be a specific feature of the strongly adhesive 0.3 @xmath0 m emulsion .
we measured velocity as low as [email protected]@xmath77 at the approach of flow stoppage in the 1@xmath0 m non adhesive and 6.5@xmath0 m adhesive emulsion ( fig .
[ fig_emul_velocity1log]b , c ) but did not observe any slope break in these cases : there is no evidence of any slow flow in the apparently jammed zones in these emulsions . in the following ,
we build the constitutive laws of the emulsions accounting for their flow properties . to get this information , we use two different methods .
first , focusing on the macroscopic measurements , we present an experimental procedure allowing us to get a relationship between local values of the shear stress and the shear rate by measuring only macroscopic quantities ( sec . [ section_torque ] ) .
second , we analyze in detail the velocity profiles and extract a local constitutive law by differentiating the velocity field to get the local shear rate ( sec . [ section_local ] ) . in this section
, we present the macroscopic rheometric measurements ( torque vs. rotational velocity ) .
we then show that it is possible to extract the constitutive law of the material from a proper analysis of this set of purely macroscopic data only provided that the material is homogeneous ( absence of migration ) and that there is no wall slip , which is what we have ensured . for each rotational velocity @xmath6
, we have measured the torque exerted on the inner cylinder vs. time until a stationary state is reached ( see inset of fig .
[ fig_emul_rheomacro]a for the 0.3 @xmath0 m adhesive emulsion ) .
we observe that for each rotational velocity , a stationary state is reached for a strain of a few unities , consistently with the absence of evolution of the velocity profiles in time . in fig .
[ fig_emul_rheomacro]a we plot the stationary torque vs. the rotational velocity for the steady flows of the 0.3 @xmath0 m adhesive emulsion .
m adhesive emulsion ; inset : torque vs. angular displacement for various rotational velocities ranging between 0.01 and 50 rpm ( bottom to top ) .
b ) constitutive law inferred from the purely macroscopic measurements thanks to eq.[eq_localstress ] and eq .
[ eq_localrate ] ; the solid line is a herschel - bulkley fit @xmath78 of the data with @xmath79pa , @xmath80pa.s and @xmath81.,title="fig:",width=302 ] m adhesive emulsion ; inset : torque vs. angular displacement for various rotational velocities ranging between 0.01 and 50 rpm ( bottom to top ) .
b ) constitutive law inferred from the purely macroscopic measurements thanks to eq.[eq_localstress ] and eq .
[ eq_localrate ] ; the solid line is a herschel - bulkley fit @xmath78 of the data with @xmath79pa , @xmath80pa.s and @xmath81.,title="fig:",width=302 ] we observe that the apparent flow curve of the material is that of a yield stress shear - thinning fluid . as we have shown in sec .
[ section_concentration ] that the material is homogeneous , we are allowed to use these macroscopic data to infer the constitutive law of the material . because of the strong stress and shear rate heterogeneities , in particular shear localization , standard formula relating simply the torque and rotational velocity values to a mean shear stress and a mean shear rate can not be used . however , it is possible to analyze the macroscopic data to account for these heterogeneities and to obtain the unique constitutive law consistent with this set of purely macroscopic data .
the method used to analyze the wide gap couette rheometric data in order to obtain the constitutive law of the material is detailed e.g. in @xcite , we give here the main steps .
first , it has to be noted that the shear stress distribution @xmath82 within the gap at a radius @xmath11 is known whatever the constitutive law of the material ; it reads ( r)=t/(2h r^2)[eq_localstress]where @xmath25 is the torque and h is the height of the couette geometry .
then , the local shear rate @xmath83 is related to the rotational velocity @xmath6 through = ^r_o_r_i ( r)/r dr[eq_omega0]where @xmath59 is the inner radius and @xmath58 is the outer radius . from eqs .
[ eq_localstress ] and [ eq_omega0 ] , we see that from the knowledge of the torque @xmath25 and the rotational velocity @xmath6 , we obtain a non straightforward relationship between the local shear stress and the local shear rate . in order to go one step further , eqs . [ eq_localstress ] and [ eq_omega0 ]
may be combined into = -^(r_o)_(r_i ) ( ) / d[eq_omega]note that the possibility of shear localization is naturally taken into account in this equation , where @xmath84 when @xmath85 with @xmath86 the material yield stress .
the constitutive law may then be derived from the whole macroscopic rheometric curve @xmath5 thanks to the differentiation of eq .
[ eq_omega ] relative to @xmath25 : 2t / t=((r_i))-((r_o))[eq_localrate]finally , in order to get the shear rate @xmath87 at the inner cylinder as a function of the shear stress at the inner cylinder , one needs to eliminate the shear rate at the outer cylinder @xmath88 from this equation .
this can be done by summing this relationship with a series of successive decreasing torques ( that differ by a factor @xmath89 at each step ) chosen to ensure that the shear rate at the outer cylinder is eliminated in the summation at each step , until a zero shear rate is reached at the outer cylinder : the summation stops when the shear stress computed at the outer cylinder at a given step falls below the yield stress @xcite .
while this summation is in principle infinite for simple fluids , only a few steps are needed in the case of yield stress fluids , and the number of steps is more reduced for wider gaps . in the case of our couette geometry , as the ratio between the shear stress at the inner cylinder and the shear stress at the outer cylinder ( @xmath89 ) is equal to 2.1 , we see that the data of fig .
[ fig_emul_rheomacro]a can be easily processed thanks to a maximum number of two summations of eq .
[ eq_localrate ] . from eq .
[ eq_omega ] , note that such analysis is based on the assumption that there is a local constitutive law characterizing the material and that there is negligible wall slip .
the constitutive law obtained with this method is depicted in fig .
[ fig_emul_rheomacro]b .
we observe that the behavior of the material obtained with this procedure is basically well fitted to a herschel - bulkley law . in sec .
[ section_local ] we compare this law to the constitutive law built from the local shear rates extracted from the velocity profiles .
the constitutive laws of the materials accounting for their velocity profiles can be built from our experimental data , using both the velocity profiles and the torque measurements .
the stress distribution @xmath82 within the gap is obtained from the macroscopic torque measurements thanks to eq .
[ eq_localstress ] .
the local shear rate @xmath83 in the gap is inferred from the velocity profiles @xmath90 through ( r)=v / r - v / rboth measurements performed at a given radius @xmath11 for a given rotational velocity @xmath6 thus provide a local data point of the constitutive law @xmath91 .
this analysis provides a fair local measurement of the constitutive law since we measure the true local shear rate within the bulk of the material and the shear stress distribution is known from the momentum balance independently of any hypothesis ; in particular , it is independent of what happens at the interface so that a possible wall slip does no affect this analysis . we are finally allowed to combine the data measured at various radiuses because the materials are homogeneous as shown in sec .
[ section_concentration ] .
m emulsion , a non - adhesive 6.5@xmath0 m emulsion and an adhesive 6.5@xmath0 m emulsion ; the solid line is a herschel - bulkley fit @xmath78 of the data with @xmath81 for all materials , @xmath92pa and @xmath93pa.s for the non - adhesive 1@xmath0 m emulsion , @xmath94pa and @xmath93pa.s for the non - adhesive 6.5@xmath0 m emulsion , and @xmath95pa and @xmath96pa.s for the adhesive 6.5@xmath0 m emulsion .
b)constitutive law measured locally in the gap of a couette cell for the flows of a 0.3@xmath0 m adhesive emulsion ( empty symbols ) ; the solid line is a herschel - bulkley fit @xmath78 of the data with @xmath79pa , @xmath80pa.s and @xmath81 .
the crosses are the data extracted from the analysis of the purely macroscopic measurements ( see fig .
[ fig_emul_rheomacro]b ) , title="fig:",width=302 ] m emulsion , a non - adhesive 6.5@xmath0 m emulsion and an adhesive 6.5@xmath0 m emulsion ; the solid line is a herschel - bulkley fit @xmath78 of the data with @xmath81 for all materials , @xmath92pa and @xmath93pa.s for the non - adhesive 1@xmath0 m emulsion , @xmath94pa and @xmath93pa.s for the non - adhesive 6.5@xmath0 m emulsion , and @xmath95pa and @xmath96pa.s for the adhesive 6.5@xmath0 m emulsion .
b)constitutive law measured locally in the gap of a couette cell for the flows of a 0.3@xmath0 m adhesive emulsion ( empty symbols ) ; the solid line is a herschel - bulkley fit @xmath78 of the data with @xmath79pa , @xmath80pa.s and @xmath81 .
the crosses are the data extracted from the analysis of the purely macroscopic measurements ( see fig .
[ fig_emul_rheomacro]b ) , title="fig:",width=302 ] the local constitutive law extracted from the experiments performed at various rotational velocities @xmath6 on all the emulsions are plotted in fig .
[ fig_emul_local1 ] ( we could not analyze the data for the 40@xmath0 m emulsion as torque data could not be measured in this case ) .
it should first be noted that for each emulsion , all the shear stress vs. shear rate data fall along a single curve .
this means that for a given material , our different data obtained under different boundary conditions effectively reflect the behavior of a single material with a given intrinsic constitutive law in simple shear . note also
that this is consistent with the absence of migration : any heterogeneity would have apparently yielded different constitutive laws for different rotational velocities @xcite . our observation of a single local constitutive law accounting for all flows of the non - adhesive 6.5@xmath0 m emulsion ( fig .
[ fig_emul_local1]a ) contrasts with the observations of @xcite on the same system in microchannel flows of up to 250@xmath0 m width : the velocity profiles observed in the channel can not be accounted for by a single constitutive law .
this apparent paradox has been solved by @xcite : they showed that a rather simple non - local flow rule accounts for all the velocity profiles .
they concluded that non - local effects should be observable in a zone involving up to 100 particles ; beyond this zone , this non - local law reduces to a local law .
this explains why our measurements are unsensitive to such effects and why we measure a single local constitutive law : using a wide gap geometry prevents from being sensitive to these non - local effects .
our observation of a single local constitutive law accounting for all flows of the adhesive 0.3@xmath0 m emulsion ( fig .
[ fig_emul_local1]b ) also contrasts with the observations of @xcite . on exactly the same 0.3@xmath0 m adhesive emulsion sheared in a thin gap couette geometry ,
@xcite were actually unable to fit all their velocity profiles with a single constitutive law . in the @xcite study , the gap being 1 mm for 0.3@xmath0 m droplets
, the non - local effects evidenced by @xcite can not be the reason for the apparent absence of a single constitutive law accounting for all flows .
an important point should be noted that may explain the differences between both studies of the same material : the experimental procedures are slightly different , and the time evolutions of the emulsion properties are very different . in our experiments , the emulsion was first presheared with a mixer and then at 100 rpm for 5 minutes in a couette cell with rough surfaces .
the stationary velocity profiles then measured at various rotational velocities were found to develop within a few seconds and to remain stable for hours . in the experimental procedure of @xcite ,
each experiment is conducted on a fresh sample which is directly sheared at the studied shear rate during at least 3 hours . a key point
may then be that @xcite found their system to evolve for at least two hours , in contrast with our observations : whatever the rotational velocity , they found the apparent viscosity of the system to decrease slowly in time .
this would mean that their material was in an initially structured state , and was destructured by shear , whatever the magnitude of the shear rate .
then , while the authors claim that the velocity profiles no longer change significantly after 2 h , one may think that some slow relaxation still occurs and that the steady state is not really reached .
this would mean that the measurements performed at different rotational velocities have been performed on different structural states of the emulsion ( in contrast with our measurements ) , naturally leading to different constitutive laws for the different velocities studied .
this explanation based on the influence of the initial state of the material would be consistent with the observations of @xcite who showed that simple emulsions whose flow properties show no apparent thixotropic effects when first presheared at high shear rate are actually subject to significant aging at rest : this imply in particular that such systems have to be strongly presheared before any study .
this preshear was performed in our study , it was not in the @xcite study .
two other points may be noted : first , although the composition of the two emulsions is the same , the samples are different and chemical impurities may play an important role in the adhesion process ; second , the couette cell used by @xcite is smooth and huge wall slip is evidenced .
focusing on the non adhesive 1@xmath0 m , the adhesive 6.5 @xmath0 m , and the non adhesive 6.5 @xmath0 m emulsions , we observe that the behavior of all materials is well fitted to a herschel - bulkley behavior @xmath78 of index @xmath81 in all the range of measured shear rates ; this is consistent with previous observations of the behavior of dense emulsions @xcite .
coming back to the 0.3@xmath0 m adhesive emulsion , at moderate and high shear rates ( above 0.1s@xmath77 ) , the behavior is the same as the one encountered for the large droplets emulsions : the constitutive law is well fitted to an herschel - bulkley law on index n=0.5 . at low shear rate ( in a 0.01 - 0.1s@xmath77 range )
, the behavior of the 0.3@xmath0 m adhesive emulsion differs from the previous ones .
all the points still fall on the same curve but a slight slope break is noticed in the flow curve . while a stress plateau should be reached for a simple yield stress fluid , the shear stress continues decreasing when the shear rate decreases .
this is also observed on the law inferred from the macroscopic measurement .
as far as we can say , this behavior is specific to the 0.3@xmath0 m adhesive emulsion : we took care of measuring the local behavior at low shear rates in the 6.5@xmath0 m adhesive emulsion but observed no slope break for shear rates as low as 10@xmath97s@xmath77 .
a probable reason for the differences observed between these systems is that the 0.3 micron emulsion is more adhesive than the others , in particular because of their larger surface / volume ratio .
this low shear rate behavior corresponds to the slow flows observed below a first apparent yield stress in fig .
[ fig_emul_velocity1log]a . at low rotational velocity ,
when there is apparently shear localization , the velocity profiles are well fitted to the herschel - bulkley law predictions only for velocities larger than a few [email protected]@xmath77 . for lower velocities , at the approach of the radius where the yield stress should be reached and the velocity should tend to zero , the velocity starts to decrease very slowly with the radius ; on the other hand , at high rotational velocity , when all the gap is sheared , the velocity profile is very well fitted to the herschel - bulkley law prediction .
the analysis of this behavior is difficult .
one may think that the emulsion flows homogeneously and really follows a single continuous local constitutive law that corresponds to a herschel - bulkley equation at high shear rate and differs from this equation at low shear rate .
one may also suggest that the law at low shear rates does not correspond to steady flows .
indeed , it has to be noted that the measurements of the very low velocities in fig .
[ fig_emul_velocity1log]a are averages over 1min : we can not know if they correspond to steady - flows ( the measurements may e.g. reflect a stick - slip behavior )
. these flows may thus reflect the existence of a jammed phase perturbated by some unsteady local rearrangements . in this case , this behavior would be reminiscent of shear banding , which has already been observed for strongly adhesive emulsions @xcite and is explained by a competition between structuration due to the adhesion process and destructuration by shear @xcite .
the local measurements obtained for the 0.3@xmath0 m adhesive emulsion are compared in fig.[fig_emul_local1]a with the law inferred from the purely macroscopic measurements ( see sec.[section_torque ] ) .
we observe that both laws are very well matched .
in particular , the macroscopic measurements also allow us to evidence the slope break in the constitutive law at low shear rate .
this validate the use of a wide gap couette geometry as a tool to obtain the constitutive law of dense emulsions from purely macroscopic rheometrical measurements , provided that a proper analysis of the macroscopic data is performed .
we recall that this agreement is due to four important features ( i ) we have shown that all dense emulsions seem to be free from migration and thus remain homogeneous during the experiments , ( ii ) the use of a wide gap prevents from the nonlocal effects observed by @xcite , ( iii ) the material is initially destructured and a steady state is studied ( this may have not be the case in the @xcite study ) , ( iv ) wall slip has to be negligible to allow for a proper analysis of the macroscopic data : this is ensured here by the use of rough walls of roughness larger than the droplet size , as shown below . as pointed out in sec .
[ section_velocity ] , our velocity measurement method does not provide the velocity at the walls : the first reliable data is obtained at around 0.5 to 1 mm from the inner cylinder . however , the local measurements of the constitutive law we have performed make it possible to study in more detail a possible wall slip effect . for a given rotational velocity of the inner cylinder ,
the velocity of the emulsion expected at the walls can actually be computed thanks to the knowledge of the shear rate in the emulsion corresponding to the shear stress value at the wall .
the same shear stress level was actually reached somewhere in the bulk of the material for a higher rotational velocity of the inner cylinder , and the shear rate corresponding to this shear stress was then measured locally thanks to the velocity profile measurement .
the velocity profile expected in the zone near the inner cylinder can then be reconstructed from the velocity profile measured at some distance from the walls and from the expected shear rate near the wall .
of course , this analysis is based on the hypothesis that the material is the same as in the bulk in this 0.5 to 1 mm zone near the walls , which may be no more true in a zone where there are nonlocal effects @xcite .
note also that this method can be used only when the reproducibility of the experiments is very good ( allowing data measured at a given rotational velocity to be used to predict what happens at another velocity ) .
an example of this method applied to the flows of the 6.5@xmath0 m adhesive emulsion is depicted in fig .
[ fig_emul_velocity2_slippage ] .
m adhesive emulsion , for a 10 and a 20 rpm rotational velocity .
b ) dimensionless velocity profiles expected near the walls at 10 and 20 rpm ; the velocities are computed thanks to the local shear rate measurements performed at 50 rpm.,title="fig:",width=302 ] m adhesive emulsion , for a 10 and a 20 rpm rotational velocity .
b ) dimensionless velocity profiles expected near the walls at 10 and 20 rpm ; the velocities are computed thanks to the local shear rate measurements performed at 50 rpm.,title="fig:",width=302 ] we observe in fig .
[ fig_emul_velocity2_slippage]b that the dimensionless velocity profiles expected near the walls have a value that is very close to 1 .
this analysis was performed on several velocity profiles , and in all cases the dimensionless velocity values predicted at the inner cylinder were found to be equal to 1 within 1% .
we finally conclude that , thanks to the use of rough surfaces of roughness larger than the droplet size , there is probably no wall slip in our experiments ; if any , it is of order 1% or less .
this feature shows that the macroscopic measurements performed with a wide gap couette rheometer with rough walls can be trusted : this explains the very good agreement observed in fig .
[ fig_emul_local1]a between the constitutive law measured locally and the constitutive law inferred from the purely macroscopic measurements .
we have studied the flows of several dense emulsions in a wide gap couette geometry .
we have coupled macroscopic rheometric experiments and local velocity and concentration measurements through mri techniques .
the method devoted to measure the local droplet concentration was developed specifically for this study .
in contrast with dense suspensions of rigid particles where very fast migration occurs under shear , we showed for the first time that no migration takes place in dense emulsions even for strain as large as 100000 in our systems .
this may imply that another mechanism is involved in dense emulsions than in dense suspensions .
the homogeneity of our materials under shear allows to infer their constitutive law from purely macroscopic measurements .
this constitutive law is consistent with the one inferred from the velocity profiles .
this contrasts with previous results obtained by @xcite within a microchannel where nonlocal finite size effects are likely to have prevented from obtaining a constitutive law .
it also differs from the results of @xcite , where the loading procedure probably prevented the authors to reach a steady state .
we thus suggest that properly analyzed purely macroscopic measurements in a wide gap couette geometry can be used as a tool to study dense emulsions .
all behaviors we observed are basically consistent with herschel - bulkley laws of index 0.5 .
however , we also evidence the existence of discrepancies with this law at the approach of the yield stress due to slow shear flows in the case of strongly adhesive emulsions , whose physical origin is unclear .
finally , we have shown that there is probably no wall slip . | flows of dense emulsions show many complex features among which long range nonlocal effects pose problem for macroscopic characterization . in order to get round this problem
, we study the flows of several dense emulsions , with droplet size ranging from 0.3 to 40@xmath0 m , in a wide gap couette geometry .
we couple macroscopic rheometric experiments and local velocity measurements through mri techniques . as concentration heterogeneities
are expected in the wide gap couette flows of multiphase materials , we also designed a new method to measure the local droplet concentration in emulsions with a mri device .
in contrast with dense suspensions of rigid particles where very fast migration occurs under shear in wide gap couette flows , we show for the first time that no migration takes place in dense emulsions even for strain as large as 100000 in our systems . as a result of the absence of migration and of finite size effect
, we are able to determine very precisely the local rheological behavior of several dense emulsions . as the materials are homogeneous , this behavior can also be inferred from purely macroscopic measurements .
we thus suggest that properly analyzed purely macroscopic measurements in a wide gap couette geometry can be used as a tool to study the local constitutive laws of dense emulsions .
all behaviors are basically consistent with herschel - bulkley laws of index 0.5 .
the existence of a constitutive law accounting for all flows contrasts with previous results obtained within a microchannel by goyon _
et al . _
[ nature * 454 * , 84 ( 2008 ) ] : the use of a wide gap couette geometry is likely to prevent here from nonlocal finite size effects ; it also contrasts with the observations of bcu _ et al . _
[ phys .
rev .
lett . * 96 * , 138302 ( 2006 ) ] .
we also evidence the existence of discrepancies between a perfect herschel - bulkley behavior and the observed local behavior at the approach of the yield stress due to slow shear flows below the apparent yield stress in the case of a strongly adhesive emulsion . |
it is well known that molecular clouds are the sites of present - day star formation . however , our understanding of the physical conditions of molecular clouds is very limited .
investigations of the formation of molecular clouds should be one of the most promising way to reveal the physical conditions of molecular clouds .
recent progress in numerical simulations of the interstellar medium has revealed many vital aspects of interstellar cloud formation .
using one - dimensional hydrodynamics simulations , hennebelle & prault ( 1999 ) and koyama & inutsuka ( 2000 ) showed that shocked warm neutral medium ( wnm or diffuse intercloud medium ; @xmath8 k ) evolves into cold neutral medium ( cnm or interstellar cloud ; @xmath9 k ) owing to radiative cooling . in
the follow - up two - dimensional simulation , koyama & inutsuka ( 2002 ) found that the thermal instability ( field 1965 , field et al .
1969 ) that is caused by runaway cooling in shocked wnm generates highly clumpy small - scale clouds with velocity dispersion that are supersonic with respect to the cnm .
since supersonic turbulence is a universal feature of interstellar clouds ( larson 1981 , heiles & troland 2005 ) , the formation of turbulent clouds without ad - hoc external forcing is very advantageous to the study of the physical nature of clouds .
hennebelle & audit ( 2007 ) and hennebelle et al . ( 2007 ) showed that hi clouds formed by thermal instability can explain various observational characteristics .
in addition to the turbulence , magnetic fields are also an important ingredient of the interstellar clouds ( heiles & crutcher 2005 ) .
using magnetohydrodynamic ( mhd ) simulations , inoue & inutsuka ( 2008 , 2009 ) showed that molecular clouds can be formed behind a shock wave if the shock propagates almost parallel to the mean magnetic field , otherwise only hi clouds are formed since magnetic pressure prevents the accumulation of the gas ( see also hennebelle & prault 2000 , hartmann et al .
2001 , inoue et al .
recently , employing three - dimensional mhd simulations , hennebelle et al .
( 2008 ) , heitsch et al .
( 2009 ) , banerjee et al .
( 2009 ) , and vzquez - semadeni et al.(2011 ) studied the formation of molecular clouds due to converging flows of wnm along magnetic field lines .
note that in the above - mentioned studies the gas is assumed to be atomic , and optically thin radiative cooling / heating are employed in the simulations .
thus , the effects of chemical evolution including the effects of ultraviolet ( uv ) shielding and the resulting modifications to radiative cooling / heating mechanisms should be taken into account ( e.g. glover et al .
2007 , 2010 ) .
recent observations of nearby galaxies , molecular clouds are suggested to be formed from hi clouds with @xmath10 @xmath11 ( blitz et al .
2006 , fukui et al .
a detailed statistical study of molecular clouds in the large magellanic cloud ( lmc ) showed that the timescale of molecular cloud evolution due to hi cloud accretion is approximately @xmath12 myr ( fukui et al .
2008 , kawamura et al .
2009 ) . as discussed by hartmann et al .
( 2001 ) and inoue & inutsuka ( 2009 ) , since the ism can only be accumulated efficiently parallel to the mean magnetic field , forming a cloud with visual extinction @xmath13 ( @xmath14 @xmath11 ) by piling up the wnm ( @xmath15 @xmath11 ) takes @xmath16
@xmath17 @xmath18 @xmath19 , even if the wnm is accumulated at the sound speed @xmath20 km s@xmath6 .
this is an order of magnitude larger than the observational estimation of @xmath12 myr .
thus , in this paper , we examine the formation of molecular clouds by the accretion of hi clouds denser than the wnm , as suggested by observations .
the plan of this paper is as follows : in 2 , we provide a possible scenario for molecular cloud formation and the numerical setup of our simulations ; the results of the simulations are shown in 3 ; finally , in 4 , we summarize our findings and discuss their implications .
in inoue & inutsuka ( 2008 , 2009 ) , we showed for the case in which the angle between the shock normal and the mean magnetic field is not parallel that the wnm compressed by an interstellar shock wave ( such as a supernova blast wave , super - bubble , or galactic spiral shock ) is cooled roughly isochorically and generates thermally unstable gas .
the resulting thermally unstable medium evolves into a complex of sheet - like hi clouds with @xmath21 @xmath11 embedded in the diffuse wnm .
since the total mass of hi clouds formed behind a super - bubble can be huge , they could be a mass reservoir of the building blocks of molecular clouds . according to inoue & inutsuka ( 2009 ) ,
the average number density of the hi medium composed of hi clouds and wnm generated behind a shock is @xmath12 @xmath11 , which agrees well with the observed hi mass reservoir of molecular clouds ( fukui et al .
2009 , dawson et al .
2011 ) .
if the hi reservoir is swept up along the mean magnetic field by shock waves , e.g. due to a galactic spiral shock or super - bubble , molecular clouds can be formed . in this case , because the mean density of the medium is an order of magnitude larger than the wnm , the timescale of mass accumulation to form a cloud of @xmath13 becomes @xmath22 @xmath17 @xmath18 @xmath23 , consistent with the evolutionary timescale of observed molecular clouds ( kawamura et al .
in addition , if the hi reservoir is massive enough to be gravitationally unstable , the formation of molecular clouds due to gravitational contraction along mean magnetic field lines would also be possible .
thus , in this paper , we examine the formation of molecular cloud due to the transonic accumulation of hi clouds along magnetic field lines . in this section ,
we provide the computational setup for forming the hi reservoir for molecular cloud formation .
the detailed modeling of molecular cloud formation from this hi reservoir is presented in the next section .
the basic equations to be solved are the ideal mhd equations including the effects of chemical reactions , radiative cooling / heating , and thermal conduction .
@xmath24 @xmath25 @xmath26 @xmath27 @xmath28 @xmath29 where the subscript @xmath30 denotes the chemical species p , h , h@xmath31 , he , he@xmath32 , c , c@xmath32 , and co , @xmath33 is the chemical reaction term for species @xmath30 , and @xmath34 is the net cooling rate per unit volume .
the other symbols have their usual meaning .
the chemical reaction coefficients and the net cooling rate depend on temperature ( @xmath35 ) , the strength of the external radiation field ( @xmath36 ) , and the column densities of chemical species and dust ( @xmath37 ) . in this paper
, we consider a requisite minimum set of chemical reactions and cooling / heating processes in order to follow the formation of an opaque molecular cloud from an optically thin hi medium .
we summarize the chemical reactions and cooling / heating processes in tables 1 and 2 , respectively .
the basic mhd equations are solved using a second - order godunov - type finite volume scheme ( van leer 1979 ) that employs an approximate riemann solver developed by sano et al .
the consistent method of characteristics with constrained transport is used for solving the induction equation ( clarke 1996 , see also evans & hawley 1988 , stone & norman 1992 ) .
the integration of the ideal mhd part of the basic equations is done in a conservative fashion .
the cooling , heating , and thermal conduction terms are solved using the second - order explicit method .
we use the piecewise exact solution method developed by inoue & inutsuka ( 2008 ) , an unconditionally stable second - order method originally adopted for solving friction terms in two - fluid mhd equations , to calculate the chemical reaction terms .
we confirmed the accuracy of this method by solving the chemical reaction equations in one - zone test problems .
we find that the piecewise exact solution method is as accurate as the crank - nicolson scheme , even though it does not require iterative procedures or matrix inversion ( kamata 2012 ) .
the timestep for the numerical integration is determined by the minimum timestep due to the cfl condition for the mhd fast mode , one - fifth of the local cooling time , and the stability condition for the parabolic conduction equation .
lll [ t1 ] cosmic ray ionization & @xmath38 s@xmath6 , @xmath39 & 1 + h@xmath40 formation on grains & dust temperature @xmath41 k is used & 2 + h@xmath40 photo - dissociation & depends on uv strength @xmath42 , column density @xmath43 , & + & and visual extinction av & 3 + h@xmath40 and co dissociation & collisions with e , p , and h & 4 + h@xmath40 and co dissociation & recombination of he@xmath32 & 1 + h , he , and c ionization & collisions with e , p , h , and h@xmath40 & 1,4 + h@xmath44 , he@xmath44 , and c@xmath44 recombination & & 4,5 + co formation & depends on @xmath36 and av & 6 + co photo - dissociation & depends on @xmath36 , @xmath43 , @xmath45 and av & 6,7 + c photo - ionization & depends on @xmath36 , @xmath46 , @xmath43 and av & 2 + lll [ t2 ] photoelectric heating by pahs & depends on @xmath36 and av & 1,2 + cosmic ray heating & & 3 + h@xmath40 photo - dissociative heating & depends on @xmath36 , @xmath43 , and av & 4 + ly-@xmath47 cooling & & 5 + c@xmath44 cooling ( 158@xmath48 m ) & level population and escape probability & + & as a function of @xmath49 are taken into account & 6 + o cooling ( 63@xmath48 m ) & escape probability as a function of @xmath50 & + & is taken into account ( @xmath51 ) . &
2,6 + co ro - vibrational cooling & escape probability as a function of @xmath52 & + & is taken into account & 7,8,9 + cooling due to rec .
of + electrons with grains & pahs & & 1 + thermal conduction & collisions of h and h@xmath31 & 10 + we prepare an hi medium composed of hi clouds and wnm as building blocks of a molecular cloud . in this stage
, we assume optically thin cooling , heating and chemical reactions , i.e. , the effects of uv shielding in the chemical reaction and heating / cooling terms are switched off .
the background uv field strength is set to the habing flux @xmath53 ( @xmath54 erg @xmath55 s@xmath6 ) .
the thermal and chemical equilibrium state of this optically thin medium is shown as a solid line in the number density - pressure plane in panel ( a ) of figure [ f1 ] .
we setup an initial state of thermally unstable gas with density fluctuations , the spectrum of which is a power law with the kolmogorov spectral index ( @xmath56 ) . thermal pressure and average number density are set to @xmath57 k @xmath11 and @xmath58 @xmath11 , where the initial abundances are @xmath59 and @xmath60 . the initial mean molar weight and the total mass in the numerical domain are @xmath61 and @xmath62 , respectively . the mean density and pressure of this initial state
is shown as a cross in figure [ f1](a ) .
we use a cubic numerical domain of ( 20 pc)@xmath63 in volume which we divide into @xmath64 uniform cells , such that the numerical resolution is @xmath65 pc .
the initial medium is uniformly magnetized by a field of strength 5.0 @xmath48 g oriented in the @xmath66-direction .
as shown by inoue & inutsuka ( 2009 ) , such a thermally unstable medium can be formed behind a shock wave that propagates in turbulent wnm .
periodic boundary conditions are used for this stage of hi cloud formation .
the three - dimensional structure of the density resulting from 8 myr of integration ( several cooling times for the initial thermally unstable gas ) is shown in the top panel of figure [ f2 ] .
regions in blue are hi clouds with @xmath67 @xmath11 and @xmath68 k that are formed as a consequence of the thermal instability .
magnetic field lines are shown as black lines .
a two - dimensional slice is also shown , in which regions in red represent @xmath69 @xmath11 , regions in yellow and green represent @xmath70 @xmath71 @xmath72 , and regions in blue represent @xmath73 @xmath11 . the probability distribution function ( pdf ) of the number density
is plotted in the bottom panel of figure [ f2 ] .
condensation driven by the thermal instability arises along the magnetic field lines , since condensation perpendicular to the field is suppressed due to the enhancement of magnetic pressure ( hennebelle & prault 2000 , inoue et al .
2007 , inoue & inutsuka 2008 , 2009 ) .
this results in the formation of sheet - like clouds .
note that the gas in the intercloud medium is also thermally unstable due to runaway heating that causes it to evolve towards the wnm phase .
linear stability analyses of the thermal transition layer between an hi cloud and wnm have shown that the transition layer is always unstable with respect to corrugational perturbations ( inoue et al .
2006 , stone & zweibel 2009 ) .
thus , the resultant hi clouds are highly corrugated .
the sheet - like morphology of the hi clouds , their magnetization ( @xmath74 g ) , typical density ( @xmath75 @xmath11 ) , and temperature ( @xmath68 k ) agree well with observed hi clouds in the ism ( heiles & troland 2003 ) .
in addition , the mean density of the hi reservoir composed of hi clouds and the wnm is also consistent with observed hi reservoirs ( fukui et al .
as we will show in 3 , the mass spectrum of the hi clouds is fitted by a power - law of @xmath76 that shows good agreement with the theoretical mass spectrum of hi clouds formed as a result of the thermal instability ( hennebelle & audit 2007 , hennebelle & chabrier 2008 ) . in this section
, we provide our numerical setup for molecular cloud formation .
we use the hi medium obtained in the previous section as an initial condition , and consider the converging flows of the hi reservoir along the mean magnetic field ( @xmath77-direction ) . for this purpose
, we set the initial velocity to @xmath78 , where the center of the numerical domain is defined as @xmath79 . in this paper
, we examine the case of a converging flow velocity of @xmath80 km
s@xmath6 , comparable to the case where hi medium is swept by a galactic spiral shock or by collisions of hi shells in super - bubbles .
we use periodic boundary condition at the @xmath81- and @xmath82-boundary planes .
the initial setting is also periodic in the @xmath83-direction except for @xmath84 given above . in effect , we impose continuous flows of the initial two - phase medium at the @xmath83-boundary planes at @xmath85 pc and @xmath86 pc .
thus , there is a continuous mass input into the numerical domain from the @xmath83-boundaries . in this paper ,
we omit the dynamical effects of self - gravity , but we estimate it using a posteriori calculations in 3.2 and 3.5 .
the neglect of self - gravity for the duration of this simulation ( @xmath87 myr ) is justified , because the gravitational energy does not become the dominant component of the energy budget ( see 3.2 ) and most of the dense clumps in the molecular cloud stay in a gravitationally unbound state ( see 3.5 ) . since the converging flow velocity is a few times larger than the sound speed in the wnm ( @xmath88 km s@xmath6 ) and much larger than that in the cnm ( @xmath89 km s@xmath6 ) , two shock waves that bound the shocked slab are formed . in order to consider the effects of ultra - violet ( uv ) photon shielding and the resulting formation of molecules and the transition of the cooling / heating processes , we take into account the dependence of the local column densities ( @xmath37 ) on the chemical reaction coefficients and cooling / heating rates ( see tables 1 and 2 ) . in this paper
, we assume that the sources of the uv field , which dissociates the molecules and causes photoelectric heating by pahs , are located far from the region of molecular cloud formation .
because of the periodic conditions for @xmath81/@xmath82 boundaries , the global morphology of the generating molecular cloud should be regarded as sheet - like structure that lies approximately in the @xmath81-@xmath82 plane .
thus , the uv photons propagating along the @xmath81-@xmath82 plane are well shielded in contrast to those propagating along the @xmath83 direction
. hence , we use a two ray approximation for uv field that irradiates the numerical domain from the @xmath90 pc planes toward the @xmath91 directions with strength @xmath92 in both rays .
note that the shielding factor of the uv flux is larger for an isotropic background uv radiation field , since the shielding increases with increased column density in the pathway of uv photons if the angle between a uv ray and the x - axis becomes larger .
this means that the formation of molecules is easier for an isotropic background uv radiation field with the same mean intensity .
thus , the two - ray approximation considered in this analysis gives the upper bound for the formation timescale of molecules .
we also calculate the approximate escape probabilities for the emission of c@xmath93/o@xmath94 fine structure lines and co ro - vibrational lines using column densities along @xmath77 directions , because the medium is opaque in the @xmath81- and @xmath82-directions .
the local chemical reaction coefficients ( cooling / heating rates ) are modified from the optically thin case by multiplying them by the shielding factors ( escape probabilities ) that are calculated by averaging the two shielding factors for the right and left rays ( two escape probabilities for the left and right directions ) .
their specific forms , which were originally formulated in a slab geometry , are found in the references in tables 1 and 2 . in figure [ f1](a ) , we plot the thermal and chemical equilibrium state under the assumed visual extinction of av=1 @xmath95 as a dashed line . in panel
( b ) of figure [ f1 ] , we also plot the cooling and heating rates of the thermal and chemical equilibrium states for the av=0 ( optically thin ) and av=1 cases as solid and dashed lines , respectively .
these panels shows that our numerical code can adequately treat media of warm atomic gas of @xmath8 k to cold molecular gas of @xmath96 k.
the converging hi flows induce two accretion shock waves that bound a highly inhomogeneous cloud . owing to continuous mass accretion , the mean column density of the cloud along the @xmath83-axis increases linearly with time and a molecular cloud is formed . in the top panel of figure [ f3 ] , we show the three - dimensional density map of the numerical domain at @xmath87 myr after the onset of the converging flows .
regions in yellow and green represent 3 @xmath11 @xmath97 @xmath11 that correspond to the preshock hi clouds and shocked wnm , regions in blue represent @xmath98 @xmath11 @xmath99 @xmath11 that correspond to the shocked hi clouds and molecular clumps , and regions in magenta represent dense clumps of @xmath100 @xmath11 that are formed , e.g. , by collisions of clumps .
the pdfs of the number density ( _ solid _ ) and the initial hi medium ( _ dotted _ ) are plotted in the bottom panel .
the pdf at @xmath101 myr is calculated only for gas in the region bound by the two accretion shocks .
we show cross - sectional maps of the number density ( @xmath102 ) and temperature in the @xmath103 and @xmath104 planes in figure [ f4 ] , where arrows represent the projected velocity on each plane . as shown by koyama & inutsuka ( 2002 ) ,
even if the preshock medium is composed of wnm alone , nonlinear evolution of the thermal instability leads to the formation of turbulent cloud composed of small - scale cloudlets . in the present case , because the accreting two - phase medium is much more inhomogeneous than the single - phase wnm , the generated structure of the cloud is even more turbulent and inhomogeneous .
this is because the interaction of the accretion shock waves with upstream large density inhomogeneities triggers the richtmyer - meshkov instability ( see , e.g. , nishihara et al .
2010 for review ) that induce rippling in the shock front . as a result
, turbulent vorticity is created behind the curved shock as a consequence of crocco s theorem ( kida & orszag 1990 ; kevlahan & pudritz 2009 ; see also inoue et al .
2009 , 2010 , 2012 for recent simulations of the shock - cloud interaction ) . in the panel ( b ) of figure [ f4 ]
, it is evident that the two accretion shock fronts have became curved due to the richtmyer - meshkov instability .
note that the growth timescale of the richtmyer - meshkov instability is given by the shock crossing time of the accreting cnm @xmath89 pc@xmath105 km s@xmath6 @xmath106 myr , so it is influential from the very early stage of cloud formation .
in the top panel of figure [ f5 ] , we plot the evolution of the average column densities along the @xmath83-axis for h and he atoms ( @xmath107 dx\rangle_{y , z}$ ] ; _ solid _ ) , for h@xmath40 molecules ( @xmath108 ; _ dashed _ ) , and for the co - fraction - weighted density ( @xmath109 dx\rangle_{y , z}$ ] ; _ dotted _ ) , where @xmath110 . in the bottom panel ,
we plot the evolution of total mass ( @xmath111 ; _ solid _ ) , h@xmath31 mass ( @xmath112 ; _ dashed _ ) , and co - fraction - weighted total mass ( @xmath113 ; _ dotted _ ) in the numerical domain . because self - shielding of uv is effective for h@xmath31 , whereas uv shielding by dust is necessary for co formation
, the h@xmath31 mass and column density is always larger than the co - fraction - weighted mass and column density . at @xmath114 myr , the total mass and the molecular mass measured by co becomes of the same order of magnitude .
thus , in the present case , the timescale at which most of c nuclei in the cloud become molecule is @xmath115-@xmath86 myr , which is consistent with the estimated timescale of molecular cloud formation associated with super - shells ( dawson et al .
2011 ) .
it has been pointed out that , in a cloud formed under the influence of the thermal instability , the cloud is not composed only of cold gas , but also the fragmented clumps of clouds embedded in warm diffuse gas ( @xmath116 k and @xmath117 @xmath11 ; e.g. , koyama & inutsuka 2002 , hennebelle et al .
note that the diffuse warm gas is the thermally unstable shocked wnm that eventually evolve into cold clumps due to the cooling ( banerjee et al .
2009 ) . in previous studies of molecular cloud formation by converging wnm flows ,
the simulations are performed assuming optically thin radiative cooling and heating .
thus , the warm gas was irradiated by an overestimated uv field that led to excessive photoelectric heating and thus lengthened the cooling time and the lifetime of shocked warm gas . in the present case , however , even after the formation of molecules , i.e. , even in the medium where the uv field is shielded , the shocked warm gas between the cold clumps survives as seen in figure [ f4 ] .
the volume fraction of diffuse gas with @xmath118 k ( @xmath119 @xmath11 ) in the region bounded by the two accretion shock waves at @xmath87 myr is 51.5% !
the reason for the survival of this diffuse gas is explained as follows : the diffuse gas ( more precisely the shocked wnm ) is continuously supplied by the two accretion shocks .
since the typical cooling time for the shocked diffuse gas is only a few myr , it seems very hard for the diffuse gas to exist deep inside the molecular cloud .
however , as we will show in more depth in 3.3 , the shock bounded region is highly turbulent and the velocity dispersion for the diffuse component of @xmath118 k is @xmath120 10 km s@xmath6 .
thus , the turbulent mixing length is evaluated as @xmath121 pc , which makes the supply of diffuse gas to deep inside the molecular cloud possible before the gas is cools down .
note that , as discussed in hennebelle & inutsuka ( 2007 ) , heating due to ambipolar diffusion makes the net cooling time longer than in the present case and would further increase the lifetime of diffuse gas in realistic molecular clouds ( see also , inoue et al .
2007 and vasquez - semadeni et al .
2011 for the effects of ambipolar diffusion on cloud formation ) . in figure [ f6 ]
, we plot the evolution of the kinetic , thermal , magnetic , and gravitational energies as solid lines in red , green , blue , and magenta , respectively .
we also plot the kinetic and gravitational energies using the co - fraction - weighted density field as dashed red and dashed magenta lines , respectively .
these energies are calculated only for gas in the region bounded by the two accretion shock waves , i.e. , the energies in the unshocked converging flows are excluded .
the positions of the two accretion shocks ( @xmath122 $ ] ) are identified by searching for the maximum and minimum @xmath83-coordinate values where the thermal pressure is larger than @xmath123 erg @xmath11 , since the converging hi medium is almost isobaric with @xmath124 erg @xmath11 .
the gravitational energy ( @xmath125 ) is calculated using the numerical solution of the poisson equation ( @xmath126 , and @xmath127 for co - fraction - weighted one ) by assuming vacuum state outside the boundaries in @xmath83 direction and periodic boundary conditions for the @xmath81- and @xmath82-directions .
the gravitational energies calculated for the raw density field and for the co - fraction - weighted field become comparable ( same order ) at @xmath128 myr , which is consistent with the result shown in figure [ f5 ] that the total mass and the co - fraction - weighted mass become the same order at @xmath129-@xmath86 myr .
the epochs at which the gravitational energy and turbulent energy measured by the co - fraction - weighted data become comparable are approximately the same ( @xmath130 myr ) .
this indicates that the scenario of molecular cloud formation examined in this paper is consistent with the fact that most observed molecular clouds are in an approximately virial equilibrium state . as discussed in the previous section ,
the resultant molecular cloud is composed of dense cold clumps embedded in diffuse warm gas .
figure [ f6 ] shows that the total kinetic energy in the shock bounded region ( _ solid red _ ) is an order of magnitude larger than that contained in the molecular clumps ( _ dotted red _ ) .
this is because the velocity dispersion of the diffuse gas component is an order of magnitude larger than that of the molecular gas , while the density of the molecular gas is only an order of magnitude larger than the atomic gas and the volume filling fractions of the atomic and molecular components are comparable .
for instance , in the shock bounded region at @xmath87 myr , the average density and velocity dispersion in gas with co fraction @xmath131 is larger than 0.1 are @xmath132 @xmath11 and @xmath133 km s@xmath6 , and in atomic gas with @xmath134 they are @xmath135 @xmath11 and @xmath136 km s@xmath6 .
this suggests that the true turbulent kinetic energy in the cloud as a whole can be much larger than the kinetic energy of turbulence estimated using line - widths of molecular emissions .
in figure [ f7 ] we plot the evolution of the density - weighted , co - density - weighted , and volume averaged velocity dispersions in the shock bounded region as red , blue , and green lines , respectively .
solid lines represent the @xmath83-component velocity dispersion ( @xmath137 ) , dashed lines represent those of the @xmath81- and @xmath82-component average ( @xmath138 ) . a typical sound speed in molecular gas , @xmath139 km s@xmath6 ,
is also plotted as a dashed line , where the temperature @xmath140 k is the co - density - weighted average temperature at @xmath101 myr . in previous studies
, it was shown that turbulence driven by the thermal instability ( and the nonlinear thin shell instability ) can induce supersonic turbulence " as a superposition of translative motions of clumps in the diffuse warm gas ( koyama & inutsuka 2002 ) , but the cold clumps have subsonic internal velocity dispersions ( koyama & inutsuka 2006 , heitsch et al .
however , in the present simulation , as we will show in 3.5.1 , the velocity dispersion can be supersonic in cold clumps depending on their size ( see figure [ f9 ] ) .
this qualitative difference can be attributed to the additional generation of turbulence triggered by the interaction between the accreting cnm and the accretion shock , i.e. , the richtmyer - meshkov instability , which was absent in previous studies of cloud formation .
figure [ f7 ] clearly shows that the turbulence in the molecular cloud is anisotropic .
this is quite reasonable because the orientation of the converging flows in the simulation is fixed that leads to the selective input of the source of the turbulence ( see also vzquez - semadeni et al .
it is widely known that most observed molecular clouds are filamentary , indicating that the building blocks of molecular clouds are also accumulated anisotropically .
thus , we can reasonably expect that turbulence in molecular clouds is anisotropic as well .
recently , hansen et al . (
2011 ) reported that the dissipation rate of anisotropic turbulence can be smaller than that of isotropic turbulence , because the timescale of the cascade is determined by the crossing timescale in the direction of smaller velocity dispersion . as we have discussed in the previous sections ,
the cold clumps are embedded in diffuse warm gas with typical sound speed @xmath141 km s@xmath6 , because the volume averaged temperature of the shock bounded region is @xmath142 k at @xmath87 myr . in isothermal gas , which is assumed in the conventional picture of molecular clouds , any supersonic motion of a cold clump creates a shock wave that decelerates the clump motion . however , since the sound speed in the warm gas is larger than the velocity of the cold clumps ( @xmath143 velocity dispersion of the turbulence ) , the molecular cloud formed in this simulation is a system with less shock waves than the conventional isothermal model in which the dissipation rate of turbulence is diminished .
thus , owing to the effects of anisotropy and the diffuse warm gas , we can expect supersonic turbulence to be sustained longer , even after the accretion ceases . in the present simulation
, we set up accretion flows along the mean magnetic field , because a molecular cloud can only be formed when the raw materials are accumulated along the mean magnetic field ( inoue & inutsuka 2009 ) .
this leads to a constant mean magnetic field strength ( @xmath144 g ) throughout the evolution of the molecular cloud .
thus , the mass - to - flux ratio ( @xmath145 ) becomes larger than the critical value ( @xmath146 ; mouschovias & spitzer 1976 ) well before the formation of molecular cloud at @xmath147 .
note that the mass - to - flux ratio of the initial two - phase medium in the domain is @xmath148 .
even though the mean field is constant , magnetic field lines are locally stretched and compressed due to turbulence which results in a local amplification of the magnetic field . in panel
( a ) of figure [ f8 ] , we plot the average magnetic field strength ( @xmath149 ) as a function of the local density at @xmath150 myr ( _ red _ ) and @xmath87 myr ( _ blue _ ) , where bars indicate the dispersion of the magnetic field strength at that density . clearly , the average magnetic field strength @xmath149 in the dense gas is much larger than the mean magnetic field strength @xmath151 g , indicating that the magnetic energy of the molecular cloud shown in figure [ f6 ] is dominated by the energy contained in magnetosonic waves rather than the energy of the global mean field . because the turbulence in the formed molecular cloud is not only supersonic but also super - alfvnic , it can easily bend and stretch the magnetic field lines .
if the turbulence was isotropic , the passive nature of the magnetic field would result in an isotropic distribution of local magnetic field orientations .
however , as shown in the previous section , the turbulence is anisotropic and enhanced in the direction of the accretion flows .
this leads to an anisotropic distribution of local magnetic field orientations . in panel ( b ) of figure
[ f8 ] , we show the probability distribution function of the angle @xmath152 between the local direction of the magnetic field and the initial direction ( @xmath66 direction ) in the shock bounded region at @xmath101 myr . the red line represents the density - weighted distribution function , and lines in green , blue , magenta , and light blue represent the distribution functions in gas with @xmath153 , @xmath154 , @xmath155 , and @xmath156 , respectively . in panel
( b ) , the plotted functions are normalized by dividing the calculated distribution by the isotropic distribution @xmath157 , i.e. , the flat distribution corresponds to the isotropic distribution .
the direction of the local magnetic field is biased towards the direction of the converging flows , except in high - density regions with @xmath100 @xmath11 .
the degree of anisotropy decreases with increasing density .
this is because as the gas evolves towards increasing densities due to the condensation of thermally unstable gas and/or the collision of clumps , the local flow direction , which affects the orientation of the local magnetic field , is deflected from the initial direction of the converging flows .
note that the nearly isotropic distribution in dense gas with @xmath100 @xmath11 does not indicate the random orientation of magnetic field in each dense clumps .
it is noteworthy that , even in the low density regions with @xmath153 , the orientation of the magnetic field is flipped from the original direction ( @xmath158 deg . ) with certain probability .
the flipped orientation of the magnetic field even in the low - density gas reflects the fact that the postshock diffuse gas ( shocked wnm ) is not in a state of laminar postshock flow " , but in a state of anisotropic turbulence " . in the formed cloud ,
molecular clumps evolve towards larger masses due to condensational accretion of the thermally unstable diffuse gas and collisional coalescence of clumps . in this section
, we present the statistics and evolution of physical quantities of the clumps . in the following ,
we define a clump as a connected region with density is larger than a threshold value @xmath159 . in order to minimize the effects of numerical noise
, we use the data for clumps that are resolved by more than 100 numerical cells .
figure [ f9 ] shows scatter plots of the plasma @xmath3 ( @xmath162 ) as a function of clump mass , and plots of the velocity dispersion as a function of clump scale .
points in red , green , blue , and magenta represent the clumps identified using threshold densities of @xmath163 @xmath11 , 500 @xmath11 , 2,000 @xmath11 , and 10,000 @xmath11 , respectively .
panels ( a ) and ( b ) show the plasma @xmath3 at @xmath150 myr and @xmath101 myr and panels ( c ) and ( d ) are the size - velocity dispersion relation at @xmath150 myr and @xmath101 myr , respectively .
the size of the clumps is evaluated as @xmath164 , where @xmath165 is the largest eigenvalue of the inertial matrix and @xmath166 is the mass of the clump .
we also used the geometric mean of the three eigenvalues and the cubic root of the clump volume , but they did not change the results significantly . in panels ( c ) and
( d ) , the typical sound speed in molecular gas , @xmath139 km s@xmath6 , is plotted as a thin dotted line , where the average temperatures in the clumps with @xmath167 and @xmath168 @xmath11 are @xmath169 and @xmath170 k , respectively .
the panels ( a ) and ( b ) indicate that the clumps evolve with time towards larger mass and that the number of high - density clumps increases with time , while the typical plasma @xmath3 stays at @xmath171 almost independent of clump density and time .
the size - velocity dispersion relation ( panels [ e ] and [ f ] ) also shows the universal law of @xmath172 km s@xmath6 @xmath7 , irrespective of both the mean density and time .
this relation is compatible with the well known size - line width relation of molecular clouds ( larson 1981 , heyer & brunt 2004 ) , and consistent with the density independent relation observed in molecular cloud l1551 by yoshida et al .
( 2010 ) . as clumps evolve , their mass , density , and size increase towards becoming a gravitationally unstable object .
panels ( a ) and ( b ) in figure [ f10 ] show scatter plots of the clumps in a simple virial diagram in which the horizontal and vertical axes represent @xmath173 and @xmath174 , respectively , where @xmath175 , @xmath176 , @xmath177 , and @xmath178 , where the post - process gravitational potential @xmath179 is used in the evaluation of the gravitational energy term @xmath180 ( see 3.2 ) .
panels ( a ) and ( b ) show the results at @xmath150 myr and @xmath87 myr , respectively . a clump that enters the region of @xmath181 can evolve into star(s ) due to gravitational contraction . because clumps grow for @xmath182 or @xmath183 irrespective of their density , the clumps lie approximately along the line @xmath184 for @xmath185 . for @xmath186 , the kinetic energy term
@xmath187 due to turbulence becomes comparable to or larger than the thermal energy term @xmath188 .
thus , the clumps become gravitationally unstable cores with @xmath189 , indicating that magnetically super - critical cores ( whose thermal and turbulent kinetic energies are larger than the magnetic energy ) are formed as a consequence of the clump evolution .
scatter plots of the mass - to - flux ratio ( @xmath190 ) as a function of clump mass at @xmath150 myr and @xmath87 myr are also shown in panels ( c ) and ( d ) of figure [ f10 ] , respectively .
clumps identified with the same @xmath159 follow @xmath191 , which is consistent with the result of banerjee et al .
( 2009 ) , while a higher @xmath159 leads to a larger mass - to - flux ratio for the same mass .
when collisional coalescence of clumps and accretion of thermally unstable gas onto the clumps occur along the magnetic field , the mass - to - flux ratio of clumps is enhanced .
typical high density clumps with @xmath192 @xmath11 can be super - critical when they grow to @xmath193 m@xmath194 .
it is widely known that the effects of ambipolar diffusion can enhance the mass - to - flux ratio even by condensational accretion of thermally unstable gas onto clumps perpendicular to the magnetic field ( inoue et al .
2007 , stone & zweibel 2010 ) and when clumps are shocked by collisional coalescence ( li & nakamura 2004 , kudoh & basu 2011 ) .
thus , if we include the effects of ambipolar diffusion , the clumps can be super - critical even for @xmath195 m@xmath194 .
note that clumps far from the unstable state @xmath196 ( and @xmath197 ) are small - scale , low - density , low - mass objects , and hence they should not be observed as dense magnetically subcritical cores with column density large enough to enable zeeman effect observations ( e.g. , crutcher et al .
2009 ) .
recent numerical simulations of the gravitational collapse of molecular cloud cores have revealed that cores can become fragmented after the firstcore is formed , if the initial condition of the core satisfies @xmath198 where @xmath199 , and @xmath200 are the rotational , magnetic , thermal , and gravitational energies of the clump ( machida et al . 2004 , 2008 ) .
figure [ f11 ] shows a scatter plot in the @xmath201-@xmath202 plane , where the solid line is the critical line for the above fragmentation condition .
panels ( a ) and ( b ) show the results at @xmath150 myr and @xmath87 myr respectively . since even high - density threshold clumps with @xmath192 @xmath11 that are close to the state of gravitationally bound cores are located in the fragmentation region ( above the dashed line ) , most clumps seem to evolve into cores that satisfy the fragmentation condition .
the ratio of the rotational energy to the gravitational energy of observed molecular cloud cores is typically @xmath203 or @xmath204 ( e.g. , goodman et al .
1993 ) that suggests non - fragmentation for cores of @xmath205
. however , as shown in dib et al .
( 2010 ) , observations tend to underestimate the angular momentum by typically an order of magnitude which leads to comparable rotational energies between the observations and our simulation .
note that , since neither the parameter @xmath206 nor @xmath207 is a conserved variable , these parameters at the epoch of the beginning of the gravitational collapse are necessary to clarify whether the clumps truly evolve into cores that eventually fragment to form binaries , which requires a longer time simulation including the effect of self - gravity . finally , we show the mass spectrum of the clumps .
we plot the mass spectrum of the clumps at @xmath150 myr , and 10.0 myr in figure [ f12 ] .
panel ( a ) and ( b ) represent the spectra for @xmath163 @xmath11 and 1000 @xmath11 , respectively .
the mass spectrum of initial hi clouds are plotted as dotted lines in both panels , where the threshold density is set to be @xmath208 @xmath11 ( a few times smaller than the typical density of the initial hi clouds ) .
the mass spectrum of the initial hi clouds ( dotted line ) can be represented by a power law of the form @xmath209 with @xmath210 .
this value of the spectral index @xmath211 can be understood from the theoretical mass spectrum of clumps formed by the thermal instability ( hennebelle & audit 2007 ) in which the spectral index is calculated as @xmath212 , where @xmath213 is the spectral index of the seed fluctuation grown by the thermal instability that is used in the initial condition of hi cloud formation ( see , 2.2 ) .
the spectra at later times show qualitatively different aspects depending on the threshold density .
the spectrum with a lower threshold density @xmath163 @xmath11 ( panel [ a ] ) is similar to the spectrum of the initial hi clouds , which is consistent with observations of molecular clumps ( kramer et al .
1996 , schneider 2002 ) . at this lower threshold density ,
the spectra represent clumps generated by the compression of raw hi clouds by the two accretion shocks .
thus , it is reasonable for the spectra in panel ( a ) to show similar shape to the initial hi cloud spectrum . on the other hand ,
the higher density threshold clumps in panel ( b ) show very different spectra from the initial one at both @xmath150 and @xmath214 myr .
because the clumps identified using the high - density threshold would be more evolved , the deviation from the initial spectrum is not surprising .
the shallower slope of the spectra for @xmath215 m@xmath216 seems to suggest the growth of mass due to the coalescence of clumps by clump - clump collisions .
it is noteworthy that , for @xmath193 m@xmath216 , the spectra can be fitted by a power law with @xmath217 , which resembles the core mass function in molecular clouds ( e.g. , ikeda et al .
2007 , ikeda & kitamura 2009 ) and the high - mass tail of the initial mass function ( salpeter 1955 ) , although the number of sample clumps is limited .
using three - dimensional magnetohydrodynamic simulations including the effects of radiative cooling / heating , chemical reactions , and thermal conductivity , we have investigated the formation of molecular clouds in the ism .
we considered the formation of a molecular cloud due to the accretion of hi clouds formed by the thermal instability .
such a mode of molecular cloud formation is compatible with the results of recent observations ( blitz et al .
2006 , fukui et al .
since the mean density of the initial multiphase hi medium is an order of magnitude larger than the typical diffuse wnm density , the formation timescale is shorter than in the previous studies of molecular cloud formation that assumed the accumulation of only diffuse wnm .
the resulting timescale of molecular cloud formation of @xmath12 myr is consistent with the evolutionary timescale of the molecular clouds in the lmc ( kawamura et al .
2009 ) .
* the initial multiphase hi medium is compressed and piled up behind shock waves induced by accretion flows . because the initial medium is highly inhomogeneous , the accretion shocks are highly deformed .
the deformed shock waves generate vortical motion behind the shock waves as a consequence of crocco s theorem that results in the formation of a very turbulent molecular cloud . *
the structure of the post shock region is roughly composed of dense cold gas ( @xmath0 k ) and diffuse warm gas ( @xmath1 k ) , and these two components are spatially well mixed by the turbulence ( 3.1 ) . because the source of the turbulence is the accretion flows
, the turbulence is highly anisotropic whose strength is biased toward the orientations of the accretion flows ( 3.3 ) . *
the kinetic energy of the turbulence dominates the thermal , magnetic , and gravitational energies in the total 10 myr evolution .
however , the kinetic energy measured by using the co - fraction - weighted density is comparable to the other energies once the co molecules are sufficiently formed owing to the uv shielding at @xmath115-@xmath86 myr after the passage of shock wave , consistent with the fact that most of the molecular clouds are roughly virialized .
this suggests that the true kinetic energy of the turbulence in the cloud as a hole can be much larger than the kinetic energy of the turbulence estimated by using line - width of molecular emissions ( 3.2 ) .
* since the turbulence is highly super - alfvnic , local magnetic field strength in dense gas is amplified due to the so - called turbulent dynamo in addition to the compression , while the mean magnetic field strength in the cloud as a hole stays at the initial value .
the local orientation of the magnetic field is also biased toward the direction of the accretion flows due to the effect of anisotropic turbulence , except the dense gas with @xmath100 @xmath11 ( 3.4 ) . *
the clumps in the molecular cloud show statistically homogeneous evolution as follows : ( i ) the typical plasma @xmath3 of the clumps is roughly constant @xmath4 ; ( ii ) the size - velocity dispersion relation show @xmath5 km s@xmath6 @xmath7 irrespective of density ; ( iii ) the clumps evolve towards magnetically supercritical cores ( see , figure [ f10 ] ) ; ( iv ) the clumps seem to evolve into gravitationally unstable cores that satisfy the fragmentation condition provided by machida et al .
( 2005 ) ; ( v ) the mass function of the low density clumps can be fitted by a power law @xmath218 that reflects the initial spectrum of the hi clouds formed by the thermal instability , while that of the high density clumps suggests steeper slope of @xmath219 for the high mass tail . in a subsequent paper , using synthetic observations of polarized dust emission , we will show that the anisotropic orientation of magnetic field leads to an overestimation of the mean magnetic field strength when employing the chandrasekhar - fermi effect ( chandrasekhar & fermi 1953 ) . in this paper
, we have omitted the dynamical effects of self - gravity . for the duration of our simulation ( @xmath101 myr ) ,
the neglect of self - gravity can be justified , because the gravitational energy has not yet became the dominant component of the energy budget in the molecular cloud ( see figure [ f6 ] ) and almost all dense clumps in the cloud are still in the gravitationally unbound state ( see figure [ f10 ] ) . however , as clearly seen in figure [ f6 ] , the gravitational energy increases continuously , while other energies plateau .
thus , in order to understand the later time evolution , the inclusion of self - gravity is crucial ( see hennebelle et al . 2008 , heitsch et al .
2008 , banerjee et al .
2009 , vazquez - semadeni et al .
2011 for the cases of colliding wnm flows ) , which we will study in our next paper .
we would like to thank kazuyuki omukai , takashi hosokawa , kohji tomisaka , and jennifer m. stone for fruitful discussions .
numerical computations were carried out on xt4 at the center for computational astrophysics ( cfca ) of national astronomical observatory of japan .
this work is supported by grant - in - aids from the ministry of education , culture , sports , science , and technology ( mext ) of japan , no .
21740146 , no.22@xmath2203369 and no .
23740154 ( t. i. ) , and no .
16077202 and no . 18540238 ( s. i. ) .
bakes , e. l. o. , & tielens , a. g. g. m. 1994 , , 427 , 822 banerjee , r. , vzquez - semadeni , e. , hennebelle , p. , & klessen , r. s. 2009 , mnras , 398 , 1082 beck , r. 2001 , space sci .
, 99 , 243 black , j. h. , & dalgarno , a. 1977 , , 34 , 405 blitz , l. , fukui , y. , kawamura , a. , leroy , a. , mizuno , n. , & rosolowsky , e. 2007 , protostars and planets v , ed .
b. reipurth , d. jewitt , & k. keil ( university of arizona press ) , 951 , 81 chandrasekhar , s. , & fermi , e. 1953 , 118 , 113 clarke , d. a. 1996 , , 457 , 291 crutcher , r. m. , hakobian , n. , & troland , t. h. 2009 , , 692 , 844 de jong , t. , dalgarno , a. , & boland , w. 1980 , a&a , 91 , 68 dawson , j. r. et al .
2011 , , 728 , 127 dib , s. et al .
2010 , , 723 , 425 draine , b. t. , & bertoldi , f. 1996 , , 468 , 269 evans , c. & hawley , j. f. 1988 , , 33 , 659 field , g. b. 1965 , , 142 , 531 field , g. b. , goldsmith , d. w. , & harbing , h. j. 1969 , , 155 , l149 fukui , y. et al .
2008 , , 178 , 56 fukui , y. et al .
2009 , , 705 , 144 glover , s. c. o. & mac low , m - m .
2007 , , 169 , 239 glover , s. c. o. , federrath , c. , mac low , m .- m . , & klessen , r. s. 2010 , , 404 , 2 goldsmith , p. f. , & langer , w. d. 1978 , , 222 , 881 goodman , a. a. , benson , p. j. , fuller , g. a. , & myers , p. c. 1993 , , 406 , 528 hansen , c. e. , mckee , c. f. , & klein , r. i. 2011 , , 738 , 88 hartmann , l. , ballesteros - paredes , j. , & bergin , e. a. 2001 , , 562 , 852 heiles , c. , & troland , t. h. 2003 , , 586 , 1067 heiles , c. , & troland , t. h. 2005 , , 624 , 773 heiles , c. , & crutcher , r. 2005 , lecture notes in physics , 664 , 137 heitsch , f. , hartmann , l. w. , slyz , a. d. , devriendt , j. e. g. , burkert , a. 2008 , 674 , 316 heitsch , f. , stone , j. m. , & hartmann , l. w. 2009 , , 695 , 248 hennebelle , p. , & audit , e. 2007 , , 465 , 431 hennebelle , p. , audit , e. , & miville - desch?nes , m .- a .
2007 , , 465 , 445 hennebelle , p. , & chabrier , g. 2008 , , 684 , 395 hennebelle , p. , & inutsuka , s. 2006 , , 647 , 404 hennebelle , p. , & prault , m. 1999 , , 351 , 309 hennebelle , p. , & prault , m. 2000 , , 359 , 1124 hennebelle , p. , banerjee , r. , vzquez - semadeni , e. , klessen , r. s. , & audit , e. 2008 , , 486 , l43 heyer , m. h. , & brunt , c. m. 2004 , 615 , l45 hollenbach , d. , & mckee , c. f. 1979 , , 41 , 555 hollenbach , d. , & mckee , c. f. 1989 , , 342 , 306 hosokawa , t. , & inutsuka , s. 2006 , , 646 , 240 ikeda , n. , sunada , k. , & kitamura , y. 2007 , , 665 , 1194 ikeda , n. , & kitamura , y. 2009 , , 705 , l95 inoue , t. , inutsuka , s. , & koyama , h. 2006 , , 652 , 1331 inoue , t. , inutsuka , s. , & koyama , h. 2007 , , 658 , l99 inoue , t. & inutsuka , s. 2008 , , 687 , 303 inoue , t. & inutsuka , s. 2009 , , 704 , 161 inoue , t. , yamazaki , r. , & inutsuka , s. 2009 , , 695 , 825 inoue , t. , yamazaki , r. , & inutsuka , s. 2010 , , 723 , l108 inoue , t. , yamazaki , r. , inutsuka , s. , & fukui , y. 2012 , , 744 , 71 kamata , d. 2012 , _ tests of numerical schemes for solving chemical reaction equations _ ( bachelor thesis in aoyama gakuin university ) kawamura , a. et al .
2009 , , 184 , 1 kevlahan , n. , & pudritz , r. e. 2009 , , 702 , 39 kida , s. , & orszag , s. a. 1990 , j. sci .
comput . , 5 , 1 koyama , h. , & inutsuka , s. 2000 , , 532 , 980 koyama , h. , & inutsuka , s. 2002 , , 564 , l97 koyama , h. , & inutsuka , s. 2006 , arxiv : astro - ph/0605528 kramer , c. , stutzki , j. , & winnewisser , g. 1996 , , 307 , 915 kudoh , t. & basu , s. 2011 , , 728 , 123 larson , r. b. 1981 , , 194 , 809 lee , h .- h . ,
herbst , e. , pineau des forets , g. , roueff , e. , & le bourlot , j. 1996 , a&a .
311 , 690 li , z - y . , & nakamura , f. 2004 , , 609 , 83 machida , m. n. , matsumoto , t. , hanawa , t. , & tomisaka , k. 2005 , , 362 , 382 machida , m. n. , tomisaka , k. , matsumoto , t. , & inutsuka , s. 2008 , , 677 , 327 millar , t. j. , farquhar , p. r. a. , & willacy , k. 1997 , a&a suppl .
ser . , 121 , 139 nelson , r. p. , & langer , w. d. 1997 , , 482 , 796 nishihara , k. , wounchuk , j. g. , matsuoka , c. , ishizaki , r. , & zhakhovsky , v. v. 2010 , phil .
r. soc . a , 368 , 1769 parker , e. n. 1953 , , 117 , 431 salpeter , e. e. 1955 , , 121 , 161 sano , t. , inutsuka , s. , & miyama , s. m. 1999 , numerical astrophysics , 240 , 383 schneider , n. , simon , r. , kramer , c. , stutzki , j. , & , bontemps , s. 2002 , , 384 , 225 shapiro , p. r. , & kang , h. 1987 , , 318 , 32 spitzer , l. 1978 , physical processes in the interstellar medium ( new york : wiley ) stone , j. m. , & norman , m. l. 1992 , , 80 , 791 stone , j. m. , & zweibel , e. g. 2009 , , 696 , 233 stone , j. m. , & zweibel , e. g. 2010 , , 724 , 131 tielens , a. g. g. m. , & hollenbach , d. 1985 , , 291,722 van leer , b. 1979 , , 32 , 101 vzquez - semadeni , e. et al .
2007 , , 657 , 870 vzquez - semadeni , e. et al .
2011 , , 414 , 2511 wolfire , m. g. , mckee , c. f. , hollenback , d. , & tielens , a. g. g. m. 2003 , , 587 , 278 yoshida , a. , kitamura , y. , shimajiri , y. , & kawabe , r. 2010 , , 718 , 1019 | using three - dimensional magnetohydrodynamic simulations including the effects of radiative cooling / heating , chemical reactions , and thermal conduction , we investigate the formation of molecular clouds in the multi - phase interstellar medium .
we consider the formation of molecular clouds due to accretion of hi clouds as suggested by recent observations .
our simulations show that the initial hi medium is compressed and piled up behind the shock waves induced by accretion flows .
since the initial medium is highly inhomogeneous as a consequence of the thermal instability , a newly - formed formed molecular cloud becomes very turbulent owing to the development of the richtmyer - meshkov instability .
the structure of the postshock region is composed of dense cold gas ( @xmath0 k ) and diffuse warm gas ( @xmath1 k ) , which are spatially well mixed due to turbulence .
because the energy source of the turbulence is the accretion flows , the turbulence becomes highly anisotropic biased towards the direction of the accretion flows .
the kinetic energy of the turbulence dominates the thermal , magnetic , and gravitational energies throughout the entire 10 myr evolution
. however , the kinetic energy measured using co - fraction - weighted densities is comparable to the other energies at @xmath2 myr , once the co molecules are sufficiently formed owing to uv shielding .
this suggests that the true kinetic energy of turbulence in molecular clouds as a whole can be much larger than the kinetic energy of turbulence estimated using line - widths of molecular emission .
we find that clumps in a molecular cloud show a statistically homogeneous evolution as follows : the typical plasma @xmath3 of the clumps is roughly constant @xmath4 ; the size - velocity dispersion relation is @xmath5 km s@xmath6 @xmath7 , irrespective of the density ; the clumps evolve toward magnetically supercritical , gravitationally unstable cores ; the clumps seem to evolve into cores that satisfy the condition for fragmentation into binaries .
these statistical properties may represent the initial conditions of star formation . |
pwn g54.1 + 0.3 is a crab pulsar wind nebula ( pwn ) analog , with a young central pulsar psr j1930 + 1852 spinning down to energize its pwn ( velusamy & becker , 1988 ) .
the crab nebula , so far , despite intensive searches ( frail et al .
1995 ) shows no evidence for an outer shock ( i.e. supernova remnant ( snr ) shell ) from the supernova explosion which produced the pulsar .
no satisfactory explanation has been proposed for this lack of an outer snr shell . like the crab , g54.1 + 0.3 also , as yet , has no identified outer snr shell . here
we search for such a shell for g54.1 + 0.3 .
we find an snr - like shell around g54.1 + 0.3 and then study its properties to determine if it is associated with g54.1 + 0.3 .
the distance of pwn g54.1 + 0.3 is important for determination of the luminosity and energetics of the pulsar and its pwn and for comparison with other pwn .
the distance has been estimated by several times previously . based on chandra x - ray observations , lu et al .
( 2002 ) estimated an absorption column density of about half the total galactic absorption in the direction to g54.1 + 0.3 , and therefore suggested a distance of @xmath35 kpc . from an hi absorption spectrum , weisberg et al .
( 2008 ) found psr j1930 + 1852 , which is the central pulsar of pwn g54.1 + 0.3 , is at a distance of @xmath45 kpc . using an improved free - electron density / distance model ( cordes & lazio 2002 ) , camilo et al .
( 2002 ) gave a dm distance of @xmath58 kpc to psr j1930 + 1852 .
the distance uncertainty on the system is thus still large , and improved distance determination to the pwn system is needed .
recently , we have revised distances to several snr / pwn or snr / hess source or snr / co cloud interaction systems with good precision by analyzing hi+co spectra of these system ( leahy & tian 2008 ; tian , leahy & wang 2007 ; tian et al . 2008 ) . here ,
we use these effective methods to constrain the distance to the pwn g54.1 + 0.3/ psr j1930 + 1852 system , and to revise its x - ray luminosity .
the radio continuum at 1420 mhz and hi - line emission data sets come from the very large array ( vla ) galactic plane survey ( vgps ) , described in detail by stil et al .
the spatial resolution of the continuum images of snr g54.1 + 0.3 is 1@xmath6 ( fwhm ) at 1420 mhz . the synthesized beam for the hi line images is 1@xmath6 , the radial velocity resolution is 1.56 km@xmath7s , and the rms noise is 2 k per channel . the short - spacing information for the hi spectral line images in the vgps is from additional observations with the 100 m green bank telescope of the nrao .
the co - line ( @xmath8 ) data set is from the galactic ring survey ( jackson et al .
2006 ) made with the five college radio astronomy observatory 14 m telescope .
the co - line data in this paper have velocity coverage of -5 to 135 km / s , an angular sampling of 22@xmath9 , radial velocity resolution of 0.21 km / s , and rms noise of @xmath30.13 k. figure 1 ( left panel ) gives the 1420 mhz continuum image of the area around pwn g54.1 + 0.3 .
the first panel of figure 2 provides a finding chart for the objects in figure 1 . a large elliptical snr - like shell ( centered at @xmath10 , @xmath11 ) with angular size of 30@xmath1226@xmath6 , with major axis tilted at about 45@xmath13 clockwise with respect to the @xmath14 direction ,
is clearly seen .
several bright small compact sources are situated along the low - latitude part of the shell .
a few filamentary structures are seen in the high - latitude part of the shell .
the brightest extended source in the image is a known hii region g54.09 - 0.06 ( lockman 1989 ) , and is just outside the low - latitude part of the shell .
another known hii region g53.64 + 0.24 is detected within the large radio shell .
figure 1 ( right ) shows a new high resolution image at 8@xmath0 m from the galactic legacy infrared midplane survey extraordinaire ( glimpse ) legacy project ( benjamin et al .
2003 ) , with the 1420 mhz radio contours overlaid .
the image , centred on the radio shell g53.9 + 0.2 , shows that bright thermal emission ( an ir shell including the hii regions g53.64 + 0.24 and g54.09 - 0.06 ) just surrounds the large - scale 1420 mhz shell ( down to a limit of 0.4 mjy ( 5 @xmath15 ) at 8@xmath0 m ) , and that little thermal emission comes from the radio shell or from its interior .
we have searched the vgps radial velocity range from -110 to 160 km / s for features in the hi emission which might be related to the morphologies of the large snr - like shell or pwn g54.1 + 0.3 .
we have found no convincing hi emissions coincident with either object .
hi emission and absorption spectra were constructed for the extended sources pwn g54.1 + 0.3 and hii region g54.09 - 0.06 , and for the compact sources g53.83 - 0.06 and g54.1 + 0.1 .
the filamentary emission of the radio shell is too faint to yield an hi absorption spectrum .
the source regions were taken as rectangular areas covering the region of bright 1420 mhz continuum emission , and the background regions were larger rectangular regions ( with the source rectangle excluded ) ( see tian et al . 2007 and leahy & tian 2008 for a detailed description of the methods for hi spectrum analysis ) .
the resulting spectra are given in fig .
2 . also given in figure 2 is the distance vs. radial velocity relation in the direction @xmath14=54.1@xmath13 for our adopted rotation curve ( see the section on kinematic distances below ) . for g54.1
+ 0.3 , absorption features are seen all the way up to the tangent velocity , at @xmath362 km / s , but no significant features at negative velocities are seen .
we conclude that g54.1 + 0.3 lies beyond the tangent point but within the solar circle in that direction .
g54.09 - 0.06 has a similar absorption spectrum to that of g54.1 + 0.3 , i.e absorption features between sun and the tangent point and no absorption at negative velocities .
thus it is located at the far side of its recombination line at velocity ( @xmath342 km / s , lockman 1989 ) .
both g53.83 - 0.06 and g54.1 + 0.1 are outside the solar circle , since prominent absorption features appear at negative velocities , so they have nothing with pwn g54.1 + 0.3 or the large radio and ir shells .
@xmath1co - line data was extracted from the survey of jackson et al.(2006 ) .
the @xmath1co images show that a several clumps of molecular gas with radial velocity of @xmath337 km / s are located just outside the large snr - like shell ( fig.3a ) .
we also find a molecular cloud with radial velocity of @xmath353 km / s is morphologically associated with pwn g54.1 + 0.3 ( fig .
3b ) and thus should be at the same distance os g54.1 + 0.3 .
the @xmath1co spectrum in the direction of pwn g54.1 + 0.3 shows high - brightness emission with a width of @xmath33 km / s centered at a radial velocity of 53 km / s .
this may be explained as evidence of interaction between the pwn shock and this co cloud .
> from the co emission spectrum in this direction , several co clouds have hi absorption identified at the same velocity ( see third panel of fig .
this means these clouds are in front of g54.1 + 0.3 ( except in the case of the 53 km / s co cloud which is located adjacent to g54.1 + 0.3 , and only needs to be partially in front ) .
the flat galactic circular rotation curve model is widely used in order to estimate kinematic distances to galactic objects . taking the commonly - used parameter of v@xmath16 as 220 km / s ( the circular velocity of the sun ) ,
one finds a tangent point velocity , @xmath17 of 42 km / s in the direction of @xmath14=54.1@xmath18 .
this is not consistent with the observed value of @xmath1962 km / s obtained from the hi emission spectra in fig .
2 . to obtain @xmath2062 km / s
, one finds v@xmath21240 km / s .
alternately , the high @xmath17 could be due to peculiar motions , such as spiral arm velocity perturbations , near the tangent point in the direction of @xmath14=54.1@xmath18 . a better knowledge of the rotation curve and non - circular motions is therefore needed in order to determine better distances to the galactic objects
however , this is beyond the scope of this paper . for simplicity ,
hereafter , we calculate kinematic distance to an object by using a rotation curve model with r@xmath16= 7.6 kpc ( the improved distance to the galactic center from the sun , eisenhauer et al .
2005 ) , v@xmath16=220 km / s and v@xmath22 linearly increasing to 240 km / s as r decreases to its value at the tangent point ( @xmath23kpc ) .
we note that reid et al.(2007 ) has given an updated value for @xmath24 which results in v@xmath16=224 km / s and a similar value of v@xmath22(243 km / s ) at the tangent point .
levine et al.(2008 ) also find a high value of rotation velocity ( @xmath3236 km / s ) at longitudes hear 53 degrees .
we estimate an uncertainty in distance determination caused by streaming and random gas motions at @xmath37 km / s ( shaver et al .
the most distant reliable absorption feature seen against g54.1 + 0.3 is at @xmath362 km / s , which is at the tangent point , thus giving a lower limit distance of @xmath24.5 kpc .
no hi absorption feature at negative velocities means that g54.1 + 0.3 is within the solar circle , i.e. an upper limit distance of @xmath29.0 kpc .
the morphological association between g54.1 + 0.3 and the co cloud at @xmath353 km / s suggests that g54.1 + 0.3 is at the far kinematic distance for the velocity 53@xmath2512 km / s , i.e. at a distance of 6.2@xmath26 kpc .
two extended hii regions g54.09 - 0.06 and g53.64 + 0.24 in the ir shell have been detected in a previous recombination line survey at velocities of [email protected] km / s and [email protected] km / s , respectively .
g54.09 - 0.06 is at the far side of its recombination line velocity due to the hi absorption feature at tangent point , i.e. 7.1@xmath27 kpc .
absorption in the spectrum of g53.64 + 0.24 indicates that it is also at the far side of its recombination line velocity , i.e. at 7.5@xmath27 kpc .
if the hii regions are actually part of the ir shell , then the ir shell is likely at the same distance of @xmath37.3 kpc .
infrared images here show the large ir shell just surrounds the large 1420 mhz shell and there is no ir emission within the ir shell interior .
the coincidence of the ir shell morphology and the radio shell morphology indicates that the radio shell is inside the ir shell and at same distance ( @xmath3 7.3 kpc ) .
we next discuss whether the radio shell is thermal or non - thermal .
the radio shell had been described as a bright ring by previous low - resolution wsrt observations at 609 mhz ( velusamy et al .
the wsrt data has no zero - spacing data , so that the fluxes are lower limits .
> from the wsrt image of the snr shell and our 1420 mhz image , we find a lower limit to spectral index of @xmath28 - 0.5 ( with @xmath29 ) , which is not useful to distinguish thermal and non - thermal emission .
we also obtained the gb6 survey ( condon et al .
1991 ) 6 cm image to determine 1420 - 4850mhz spectral index .
the gb6 maps were made using a differencing technique , so they miss large scale flux .
we calculate the spectral index of the bright filament at ( l , b)=(53.75,0.30 ) and the fainter filament at ( l , b)=(54.10,0.35 ) and find in both cases upper limits of @xmath30 1.7 . to check this result we calculate spectral index for the nearby supernova remnant g54.4 - 0.3 , which has a spectral index of 0.5 ( caswell , 1985 ) and find the same upper limit on @xmath31 .
thus the existing radio data does not give definitive evidence on whether the radio shell has a thermal or non - thermal spectrum .
velusamy et al .
( 1986 ) roughly classify the radio shell as an hii region g53.9 + 0.3 due to possible thermal ir emission from the shell .
however our present images , which have higher spatial resolution , clearly show the radio and ir shells are physically separated , invalidating the previous conclusion .
the radio shell might have a thermal radio spectrum and still show no ir emission if the dust in the inner side of the ir shell has mainly been destroyed , e.g. , by intense uv radiation from young stars in the interior .
then the radio shell could be a stellar wind bubble produced by a group of young stars .
pwn g54.1 + 0.3 could be associated with the same star cluster , but since the the radial velocity of g54.1 + 0.3 is so different than than of the ir shell ( 53 km / s vs. 40 km / s ) we do nt think that they are associated . ignoring the velocity difference , and assuming g54.1 + 0.3 and the radio / ir shells are at the same distance
, then we can explain the lack of an outer shock for pwn g54.1 + 0.3 as being caused by the low density interior of the stellar wind bubble inside the radio and ir shells .
if the radio shell is non - thermal , then it is likely a snr .
g54.1 + 0.3 is @xmath30.25@xmath13 offset from the radio shell center , has a young age of @xmath32000 yrs ( lu et al 2002 ) and appears to be significantly nearer ( 6.1 kpc compared to 7.3 kpc ) .
if we assume the pulsar in pwn g54.1 + 0.3 is born at the center of the radio shell and at same distance as the radio ( and ir ) shell , the offset corresponds to a distance of @xmath332 pc , which gives an implausible high proper motion velocity of @xmath316000 km / s for the pulsar .
so , we conclude pwn g54.1 + 0.3 is highly unlikely to be the compact remnant of the same explosion that caused the snr - like shell .
considering that the radio shell may be a new snr g53.9 + 0.2 , it has a distance of @xmath27.3 kpc and mean radius @xmath230 pc .
the sedov age is then 80,000@xmath32yr , with @xmath33 the ambient density and @xmath34 the explosion energy in units of 0.75@xmath35 erg .
for such a large snr to still be in the sedov phase requires a low ambient density or large explosion energy : @xmath36 ( e.g. using the limit that @xmath37 , the time that the dynamics are affected by cooling , cox , 1972 ) .
if the snr is in the shell phase , one can take the current radius of 30 pc as the cooling radius , to estimate a lower limit to the age as the time when it enters the shell phase , @xmath38 yr . in either case
it is clear that g53.9 + 0.2 is an old snr , and is not expected to be detected in x - rays .
we have verified there is no soft x - ray emission detected in the rosat all - sky - survey .
the distance we find for pwn g54.1 + 0.3 is 6.2@xmath26 kpc , consistent but better constrained that previous distance estimates .
camillo et al .
( 2002 ) discuss the energetics of the crab - like psr j1930 + 1852 in g54.1 + 0.3 using a distance of 5 kpc .
the larger distance here implies that the 2 - 10 kev x - ray luminosity is larger by a factor of 1.54 , thus significantly reducing the difference in luminosity with pwn / psr j1124 - 5916 , which has nearly identical spin parameters to psr j1930 + 1852 .
here we have searched for a snr shell counterpart to the pwn g54.1 + 0.3 using radio continuum , hi and co line emission , and infrared observation .
in addition we have used hi absorption and co emission data to constrain distances to pwn g54.1 + 0.3 and other nearby sources .
we determine a distance of @xmath26.2 kpc to g54.1 + 0.3 using an associated co cloud , consistent with the hi absorption spectrum which limits g54.1 + 0.3 to be between the tangent point and the far side of the solar circle .
in addition , we find a snr - like shell in 1420 mhz continuum emission , g53.9 + 0.2 , which is offset from g54.1 + 0.3 .
the radio shell has no thermal ir emission from its shell but is surrounded by an ir shell with an embedded hii region and associated co clumps , both of which are located at 7.3 kpc distance .
the radio shell may be thermal and lack ir emission due to dust destruction , or it may be nonthermal and part of a newly found old snr . in either case
it is located at a different distance than pwn g54.1 + 0.3 .
dal and tww acknowledge support from the natural sciences and engineering research council of canada .
tww thanks the natural science foundation of china for support .
this publication makes use of molecular line data from the boston university - fcrao galactic ring survey ( grs ) .
the nrao is a facility of the national science foundation operated under cooperative agreement by associated universities , inc .
benjamin , r.a . ,
churchwell , e. , babler , b.l .
et al . 2003 , pasp , 115 , 953 camilo , f. , lorimer , d.r . ,
bhat , n.d.r . , et al .
2002 , apj , 574 , l71 caswell , j. 1985 aj 90 1224 condon , j. j. , broderick , j. j. , & seielstad , g. a. 1991 , aj , 102 , 2041 cordes , j.m . & lazio , t.j.w .
2002 , astroph/0207156 cox , d.p . 1972 ,
apj , 178 , 159 eisenhauer , f. , genzel , r. , alexander , t. et al .
2005 , apj , 628 , 246 frail , d. a. , kassim , n.e . , cornwell , t.j .
, & goss , w.m .
1995 , apj , 454 , l129 jackson , j.m . ,
rathborne , j.m . ,
shah , r.h .
2006 , apj suppl . , 163 , 145 leahy , d.a . , tian , w.w .
2008 , aj , 135 , 167 levine , e. s. , heiles , c. , & blitz , l. 2008 , apj , 679 , 1288 lockman , f , j. 1989 , apjs , 71 , 469 lu , f.j .
wang , q.d . , aschenbach , b. et al .
2002 , apj , 568 , l49 reid , m. j. , brunthaler , a. , menten , k. m. , ye , x. , xing - wu , z. , & moscadelli , l. 2007 , iau symposium , 242 , 348 shaver , p.a . , radhakrishnan , v. , anantharamaiah , k.r . , et al .
1982 , a&a , 106 , 105 stil , j.m . ,
taylor , a.r . ,
dickey , j.m .
2006 , aj , 132 , 1158 tian , w.w . , leahy , d.a .
, wang , q.d .
2007 , a & a , 474 , 541 tian , w.w . , leahy , d.a . , haverkorn , m. jiang , b. 2008 , apj , 679l , 85 velusamy , t. & becker , r.h . 1988 ,
aj , 95 , 1162 velusamy , t. goss , w.m . &
arnal , e.m .
1986 , a & a , 7 , 105 weisberg , j.m . , stanimirovi , s. , xilouris , k. , et al .
2008 , , 674 , 286 | we discover a large - scale shell g53.9 + 0.2 around the crab - like pulsar wind nebula ( pwn ) g54.1 + 0.3 with 1420 mhz continuum vla observations .
this is confirmed by a new infrared image at 8 @xmath0 m from the glimpse legacy project , which reveals an intriguing infrared shell just surrounding the large radio shell .
we analyze the 21 cm hi absorption spectra and @xmath1co emission spectra towards pwn g54.1 + 0.3 and bright sources on both radio and ir shells .
continuous hi absorption up to the tangent point and absence of negative hi absorption features imply that pwn g54.1 + 0.3 has a distance beyond the tangent point but within the solar circle , i.e. 4.5 to 9 kpc .
g54.1 + 0.3 is likely at distance of @xmath26.2 kpc due to the morphological association between the pwn and a co molecular cloud at velocity of @xmath253 km / s , as revealed by high - resolution @xmath1co images .
based on the hi absorption spectrum and recombination line velocity ( @xmath240 km / s ) of the bright hii region g54.09 - 0.06 , which is on the ir shell , the ir shell is likely located at a distance of @xmath27.3 kpc , which is also the distance of the associated large - scale radio shell . at this distance
, the radio shell has a radius of @xmath330 pc .
the radio shell may be thermal and lack ir emission due to dust destruction , or it may be nonthermal and part of a newly found old snr . in either case
it is located at a different distance than pwn g54.1 + 0.3 . |
the roots of quantum metrology has a history that dates back to heisenberg s work in the early years of the previous century @xcite .
more recently , the phase sensitivity of interferometers is of interest because quantum interferometers may be used as sensors of classical fields @xcite .
fluctuations limit the precision of phase measurements @xcite and the phase sensitivity of interferometers has been explored for different types of input states , such as squeezed states and number states , see references contained in @xcite .
recently , there has been experimental progress in making efficient single - photon and entangled - photon sources , and photon number resolving detectors .
therefore , we concentrate on estimating the phase ( and thereby the classical field ) from photon number counts in the output ports of an interferometer . the phase sensitivity @xmath0 is often discussed according to how it scales in two limits , known as the standard limit ( shot noise ) , @xmath1 @xmath2 , and the heisenberg limit , @xmath3 @xmath4 , where @xmath5 is the number of particles that enter the interferometer during each measurement cycle .
these results are based on standard estimation theory@xcite which connects a continuous valued measurement outcome , @xmath6 , with a corresponding phase , @xmath7 , through a theoretical relation @xmath8 .
the derivation of the heisenberg limit of phase sensitivity implicitly assumes that the phase probability distribution has a single peak . using a bayesian analysis
we have shown that this assumption is not generally true @xcite , unless we know some prior information about the phase , for example that the phase is in some small region .
we have shown that the phase probability distributions have two peaks when fock states are input into a mach - zehnder ( mz ) interferometer and they have four peaks when maximally entangled states ( n00n ) are input @xcite , when there is no prior information about the phase .
consequently , it is not clear whether the phase sensitivity , described in terms of the width of the phase probability distribution for measurement @xmath6 , @xmath9 , should refer to the width of a single peak or to the width of the whole distribution ( containing all the peaks ) . as a result of this ambiguity
, we have proposed a new metric of interferometer phase sensitivity @xcite , @xmath10 \label{mutualinformationeq}%\ ] ] which we call the fidelity .
this is the shannon mutual information between the phase ( a continuous alphabet ) random variable , @xmath11 , and the photon - number measurement outcomes ( a discrete alphabet ) with random variable @xmath12 .
here @xmath13 is the conditional probability of a measurement outcome @xmath6 given the phase was @xmath7 , and @xmath14 is the _ a priori _ probability distribution that characterizes our prior knowledge of the phase .
the fidelity @xmath15 is a statistical characterization of the phase sensitivity of the interferometer that is valid when multiple peaks exist in @xmath16 .
the fidelity is a measure of the quality of a sensor for given input state , and it applies to both classical and quantum interferometers . for a standard quantum mz interferometer with a phase shift @xmath7 in one arm @xcite , with photon input into ports
a " and b " , the measurement outcomes are @xmath17 , where @xmath18 and @xmath19 are the photon numbers output in port c " and d " .
the phase probability distribution is obtained from bayes rule as @xmath20 where . in this report
, we investigate the dependence of the fidelity of a mz interferometer on the _ a priori _ information available , @xmath14 , comparing fock state and n00n state input .
the four phase probability distribution functions are shown for the measurement outcomes @xmath21 , @xmath22 , @xmath23 , and @xmath24 , for 3-photon n00n state input , @xmath25 , for @xmath26 . ]
the four phase probability distribution functions are shown for the measurement outcomes @xmath21 , @xmath22 , @xmath23 , and @xmath24 , for 3-photon n00n state input , @xmath25 , for @xmath26 . ]
the choice of input states is not intended to be optimum , but simply to compare two possible input states that are of current interest .
for n - photon fock state input into arm a " and vacuum input into port b " , there are generally two peaks in the phase probability distribution @xmath27 , when @xmath26 .
for the same @xmath26 with n - photon n00n ( maximally entangled ) state input , @xmath28 , there are generally four peaks in @xmath27 @xcite . for a single measurement ,
when a given measurement outcome is obtained , eq.([phaseprobabiliydensity ] ) gives the probability distribution for the phase @xmath7 .
when a 1-photon fock state is input into port a " , @xmath29 , the probabilities for photon output into ports c " and d " are @xmath30 and @xmath31 , respectively . when a 1-photon n00n state is input , @xmath32 , the probabilities for photon output in ports
c " and d " are @xmath33 ^ 2 /2 $ ] and @xmath34 ^ 2 / 2 $ ] , respectively .
( it is curious to note that these probabilities , in this extreme quantum limit of 1-photon input , are equal to the classical probabilities for input of energy into only port a " , and input of equal energy into both ports a " and b " ( classical analog of n00n state input ) ) . for 2-photon input ,
there are three possible measurement outcomes . for 2-photon fock state input into port
a " and vacuum in port b " , there are three probability density functions @xmath16 for the phase : @xmath35 , @xmath36 , and @xmath37 , for the measurement outcomes @xmath38 , @xmath39 , and @xmath40 , respectively , for @xmath26 . however , for the 2-photon n00n state input , the mz interferometer behaves as a beam splitter , where there is 1/2 probability that both photons come out port c " , 1/2 probability that both photons come out port d " , and zero probability that one photon comes out each port . for this case ,
the probability of a given measurement outcome is independent of @xmath7 .
consequently , the 2-photon n00n state can not be used for determining the phase .
this is not the case with higher photon - number n00n state input . in figures 1 and 2
, we show the phase probability distributions for @xmath41 photon fock state and n00n state inputs , respectively , for @xmath26 .
note that for n00n state input , the phase probability distributions are not symmetrical about @xmath42 and do not have simple peaks .
as the number of input photons @xmath5 increase , the phase probability distributions get narrower .
a gaussian fit to one peak of the probability distributions for fock state inputs of @xmath45 to @xmath46 photons indicates that the phase sensitivity is given by @xmath47 , where @xmath5 is the number of photons input into port a " .
however , the n00n state phase probability distributions do not have localized peaks , see figure 2 , and hence it is not possible to estimate the phase sensitivity from such widths .
consequently , we use the fidelity of the interferometer , given by eq .
( [ mutualinformationeq ] ) , to compare the statistical phase sensitivity on average , for fock state input and for n00n state input . in previous work @xcite , we computed the fidelity of a mz interferometer for the case of no prior information ( taking @xmath26 ) for fock state input and for n00n state input , and we found that the fidelity for fock state input is greater than or equal to that of n00n state input , for all input photon numbers @xcite . in the opposite regime , where we have very precise prior information , so @xmath14 has one narrow peak about some value @xmath48 , the fidelity can be approximated as @xmath49 ^ 2 } } { { p(m|\phi _ 0 ) } } } \right ] } \label{shannontightapriori}\ ] ] where @xmath50 , @xmath51 , @xmath52_{\phi=\phi_0}$ ] , and where we assumed that @xmath53 . in the limit of narrow distribution @xmath14
, we find that the fidelity for fock state input and n00n state input is equal , @xmath54 , for given total number of input photons @xmath5 .
( the exception to this statement is for n00n state input , with @xmath55 , when @xmath18 is odd , because then @xmath56 . ) from these results we see that in both limiting cases , where we have no prior information , and in the limit of narrow distribution for the prior information , the fidelities are equal , so the use of n00n states does not provide an advantage over fock states .
finally , we compare the fidelity in the intermediate regime , for fock state and n00n state input , for the case when the prior information is given by a gaussian function @xmath57 $ ] , where the normalization @xmath58 is given by @xmath59 .
see figures 3 and 4 .
while there are small variations in fidelity for small photon numbers ( notably for @xmath45 and @xmath60 ) , we find that generally fock state input produces a higher fidelity than n00n state input . in figures 3 and 4 , we have chosen a prior information represented by a gaussian centered at @xmath61 . the result for @xmath62
must be still be determined .
finally , we note that in actual applications the prior information may have a complicated functional form with multiple maxima . | we investigate the dependence of the fidelity of a mach - zehnder quantum interferometer on the prior information about the phase , for fock state input and for maximally entangled ( n00n ) state input . for no prior information ,
the fidelity for fock state input is greater than for n00n state input . in the limit of a narrow distribution describing the prior information
, we find that both fock and n00n state inputs lead to nearly equal fidelity . |
heavy flavor production in heavy ions collisions is an ideal probe to study the early time dynamics of these nuclear collisions .
several theoretical studies predict @xcite a substantial enhancement of open charm production associated to deconfined parton matter relative to the case of a purely hadronic scenario without quark - gluon plasma formation .
recent studies point out that the dynamics of heavy quarks is dominated by partonic interactions in a strongly coupled plasma modeled neither by hadronic interactions nor by color screening alone @xcite .
therefore , these quarks are very relevant in the study of the initial state of the collision . owing to difficulties in reconstruct the @xmath0-mesons decay vertex , rhic experiments have measured open charmed quarks indirectly , via the semileptonic decay to nonphotonic electrons or muons @xcite @xcite . in the standard picture charm quarks
are produced by initial gluon fusion and their production rates are expected to be well described by perturbative quantum chromodynamics ( pqcd ) at fixed order plus next - to - leading logarithms ( fonll ) @xcite .
the suppression of single , nonphotonic electrons or muons at rhic is usually attributed to heavy - quark energy loss . as a charmed quark of energy @xmath1
can not radiate gluons forming an angle below @xmath2 ( dead cone effect ) , it is expected that heavy quarks lose less energy than light quarks @xcite , but the suppression experimentally observed is similar .
in fact , many calculations based on energy loss via hard scattering @xcite or via multiple soft collisions @xcite obtained less suppression than the experimental data when the beauty contribution is taken into account .
similar results are obtained in evaluations based on medium interactions or collisional dissociation @xcite .
however , it has been argued @xcite@xcite that under the assumption of an enhancement of the heavy - quark baryon - to - meson ratio , analogous to the case of the proton - to - pion and the @xmath3-to - kaon ratios measured in au - au collisions at rhic , it is possible to achieve a larger suppression of the nuclear modification factor .
this is possible , because the heavy - quark mesons have a larger branching ratio to decay inclusively into electrons as compared to heavy - quark baryons and therefore , when the former are less copiously produced in a heavy - ion environment , the nuclear modified factor decreases , even in the absence of heavy - quark energy loss .
indeed the single nonphotonic nuclear modified factor , @xmath4 can be expressed as @xcite @xmath5 where @xmath6 is the nuclear modified factor for d and @xmath7 , i.e. @xmath8 with @xmath9 and @xmath10 the produced @xmath0 and @xmath3 in @xmath11 or @xmath12 collisions and the @xmath13 is the number of collisions at a given centrality .
the factor @xmath14 is given by the expression @xmath15 where , @xmath16 and @xmath17 are the charmed baryon - to - meson ratios in proton - proton and @xmath11 collisions respectively . therefore @xmath18 represents the enhancement factor for the ratio of charm baryons to mesons in @xmath11 as compared to @xmath12 collisions , @xmath19 is the ratio between the branching ratios for the inclusive decay of @xmath3 and @xmath0 into electrons : @xmath20 in @xcite @xmath19 has been estimated to be 0.14 . as long as @xmath18 is larger than 1 the factor @xmath14 becomes less than 1 and @xmath21 .
the main question to solve is whether the expected charmed baryon - to - meson expected enhancement is large enough to explain the difference with the experimental data . in a high - energy heavy - ion collision ,
strong color fields are expected to be produced between the partons of the projectile and target @xcite@xcite @xcite .
these color fields are similar to those that appear in the glasma @xcite produced in the color glass condensate ( cgc ) . in a string
heavy - quark pairs are produced via the schwinger mechanism with a rate @xmath22}$ ] where @xmath23 , is the effective string tension , proportional to the strength of the field ( for a single string @xmath24 1 gev / fm ) .
longitudinal string models predict for heavy flavor a very suppressed production rate , since @xmath25 for @xmath26 and @xmath27 .
the color in these strings is confined to a small area in the transverse space , @xmath28 , with @xmath29fm . in a central heavy - ion collision
many strings are formed between the partons of the projectile and target in a limited collision area , starting to overlap each other , forming clusters .
the field strength of the cluster is proportional to the square root of the number of strings .
so , for a cluster of nine strings , the string tension increases more than eight orders of magnitude becoming comparable to the initial fonllpqcd .
the effect of strong color fields for open charm has been investigated before @xcite showing that a three- fold increase of the effective string tension results in a sizable enhancement of the total charm cross section and the nuclear modified factor shows a suppression at moderate @xmath30 consistent with the rhic data . in this paper
, we study the effects of strong color fields in the framework of percolation of strings @xcite .
in this framework , a strong color field is obtained inside the clusters formed by the overlapping of individual strings .
the clusters behave like individual strings with a higher string tension owing to the higher color field , and their energy momentum is the sum of the energy - momenta of individual strings .
the color field of the string is stretched between a quark and an antiquark or between a diquark and a quark located at the extremes of the string . in the case of a cluster , instead of quarks or antiquarks we have complexes @xmath31 and @xmath32 formed from the different quarks and their antiquarks or diquarks and quarks at the extremes of the individual strings @xcite@xcite .
the clusters behave like strings with complexes @xmath33 , located at the end , decaying into new pairs @xmath33 , @xmath34 , until they come to objects with mass comparable to hadron masses , which are identified with the known hadrons by combining the produced quarks or antiquarks with the appropriate statistical weights . in this way ,
the production of baryons and antibaryons is enhanced with the number of strings in the cluster .
the cluster not only has a stronger color field than the individual string giving rise to a mass - enhancement effect but also enhances the production of baryons relative to mesons owing to the increasing probability of getting three quarks or three antiquarks from the complex @xmath33 @xcite .
this second effect is similar to what happens in coalescence or recombination models @xcite@xcite .
the percolation of strings incorporates to some extent the recombination of flavors in a dynamical way .
indeed a dynamical quark recombination model has shown a sizable suppression factor for the nonphotonic electrons nuclear modification factor @xcite .
we evaluate the nuclear modification factor for @xmath35 , @xmath7 and b at rhic energies , computing also the baryon - to - meson ratio in @xmath11 and @xmath12 collisions .
we observe also in @xmath12 a moderate enhancement of the ratio as a function of the transverse momentum which has consequences concerning the value of @xmath14 and therefore the rate of the nonphotonic electron suppression .
the plan of the paper is as follows : in the next section , we introduce briefly the percolation of the strings , and then we present our results and conclusions .
in the string percolation model @xcite@xcite@xcite @xcite @xcite , multiparticle production is described in terms of color strings stretched between the partons of the projectile and the target . with increasing energy and/or atomic number of the colliding particles , the number of strings @xmath36 , grows and they start to overlap forming clusters , very much like disks in two - dimensional percolation theory . at a certain critical density , a macroscopical cluster appears , which marks the percolation phase transition .
this density corresponds to the value @xmath37 ( depending on the profile function of the colliding nuclei ) where @xmath38 and @xmath39 stands for the overlapping area of the colliding objects .
a cluster of @xmath40 strings behaves like a single string with the energy momentum corresponding to the sum of individual ones and with a higher color field corresponding to the vectorial sum in color space of the color fields of the individual strings . in this way ,
the mean multiplicity @xmath41 and the mean transverse momentum squared @xmath42 of the particles produced by a cluster are given by @xmath43 and @xmath44 , where @xmath45 and @xmath46 are the corresponding quantities in a single string .
in the limit of high density of strings , equations ( 5 ) transforms into @xcite @xmath47 , @xmath48 with @xmath49 . for a specific kind of particle @xmath50
, we will use : @xmath51 , @xmath52 , @xmath53 , and @xmath54 for the corresponding quantities . to compute the multiplicities ,
we must know @xmath36 and @xmath55 ( so for a fixed centrality , knowing @xmath36 we deduce the density @xmath56 ) .
up to rhic energies , in the central rapidity region @xmath36 is approximately twice the number of collisions , @xmath13 .
however @xmath36 is larger than @xmath57 at rhic and lhc energies , in the same way as in nucleon - nucleon collisions . according to color exchange models , such as dual parton model or the quark gluon string model @xcite@xcite , the number of produced strings @xmath36 is larger than two , starting at the rhic energies .
indeed , at high enough energy the strings are stretched not only between the diquarks(quarks ) and quarks(diquarks ) of the projectile and target respectively , but also between quarks(antiquarks ) and antiquarks(quarks ) of the sea . as the energy increases , more @xmath58 or @xmath59 are formed and @xmath36 becomes larger than two .
for the same reason in @xmath11 collisions , @xmath36 at high energy is larger than @xmath57 . in this work
we take the values of @xmath36 from a monte - carlo based on the quark - gluon string model @xcite .
concerning the transverse momentum distribution , one needs the distribution @xmath60 for each cluster , and the mean square transverse momentum distribution of the clusters @xmath61 , where @xmath19 is the inverse of the mean of the squared transverse momentum of each cluster which is related to the cluster size by equation ( 5 ) .
we take @xmath62 as it is used for fragmentation of the lund string . for the weight function
@xmath61 we take the gamma distribution .
the generalized gamma distributions are unique distributions stable under the cluster - size transformations @xcite@xcite@xcite ; for simplicity we choose gamma distribution simplicity @xcite .
@xmath63 with @xmath64 and @xmath65 the function @xmath23 measures the width of the distribution ( 7 ) and is the inverse of the normalized dispersion of the transverse momentum squared .
the function @xmath23 depends on the density of strings @xmath56 .
the transverse momentum distribution @xmath66 of particle @xmath50 is @xmath67 @xmath68 the formula ( 10 ) is valid for all types of collisions , energies and also all kind of flavors .
later we will extend ( 10 ) for baryons .
the function @xmath69 was determined by comparing ( 10 ) to rhic data .
the function @xmath23 decreases with @xmath56 up to values @xmath70 ( peripheral au - au collisions at rhic energies ) from there it increases slowly .
this behavior was expected .
in fact , at low density there is no overlapping of strings and there are isolated strings ; therefore @xmath71 . when the density and therefore the numerator of eq . ( 9 ) increases then k decreases
. the minimum of k will be reached where the fluctuations in the cluster - size reach its maximum . above this point , increasing @xmath56 these fluctuations decrease and @xmath23 increases .
the agreement with data for @xmath30 up to 5 gev / c is very good @xcite@xcite . in percolation of strings
the fragmentation of a cluster of many strings is via the schwinger mechanism , producing successive pairs @xmath33 , where @xmath31 represents the complexes of quarks , diquarks and antiquarks at the extremes of the original string .
it is clear that formula ( 10 ) only contains the effect of the stronger color field of the cluster , which enhances heavy particles production , irrespective of their being mesons or baryons , but it does not contain the breaking via flavor complexes @xmath33 and therefore can not describe baryons correctly . in previous papers
@xcite@xcite@xcite monte carlo codes were presented where this mechanism was built up , but with the approximation of fusion of only two strings @xcite@xcite or using an effective color field @xcite .
to to keep a closed analytical formula , incorporating the antibaryon and baryon enhancement from the mechanism depicted here , we observe that this enhancement is similar to using the formula ( 10 ) with a larger density , or equivalently with a larger @xmath36 .
this means that for antibayons or baryons if we want to continue with formula ( 10 ) we must replace @xmath56 by @xmath72 , @xmath73 and instead of the first equation ( 6 ) we must use @xmath74 where the parameter @xmath75 is obtained from a fit to the experimental dependence of the @xmath30-integrated @xmath76 spectra with centrality @xcite .
the obtained value is @xmath77 . in same sense
we can say that the antibaryons ( baryons ) probe a higher density than mesons for a fixed energy and type of collisions .
the equations ( 10 ) , ( 11 ) and ( 12 ) allow us to compute the antibaryon ( baryon ) spectra .
the equations ( 11 ) and ( 12 ) replace the recombination process described in this section and it should be considered as an approximation to keep the analytical formula ( 10 )
. the formulae ( 10 ) , ( 11 ) and ( 12 ) are valid for all kind of particles and not only for heavy flavor .
we will show some results concerning light flavor .
the equation ( 10 ) is limited to low and moderate @xmath30 not higher than @xmath78 gev / c .
in fact , we consider a gaussian @xmath30 distribution for the particles produced from the fragmentation of a string , without any power - like tail .
this excludes the high-@xmath30 behavior , although our formula ( 10 ) allows for an interpolation from low to high @xmath30 . by continuity ,
the high @xmath30 suppression observed at rhic should give rise to a suppression at moderate @xmath30 , say @xmath78 gev / c , which is the limit where our equations apply . to know the @xmath30 distributions given by formula ( 6 )
we need the values of @xmath79 and @xmath80 , i. e. the mean @xmath30 of @xmath0 and @xmath7 particles produced from one string .
we use @xmath81 gev / c and @xmath82gev / c . the difference between these two values is close to the difference between the masses of @xmath35 and @xmath7 and also agrees with the difference between the values commonly used of primordial transverse momentum of pions and protons , @xmath83 gev / c , @xmath84 gev / c . for b
we use @xmath85 gev / c . in formula ( 10 )
the normalization is established by the values of @xmath86 at @xmath87 , which are computed using the formulae ( 6 ) for d and ( 9 ) for @xmath7 . to do this
, we use the values @xmath88 and @xmath89 , with @xmath90 @xcite .
we use these functions for @xmath91 and @xmath92 because for heavy particles , @xmath93 is very different from @xmath94 . concerning the function @xmath69 , we take the shape and values from the studies done in reference @xcite for @xmath11 collisions . in the case of @xmath12
we take @xmath95 at @xmath96 gev and @xmath97 at @xmath98 tev .
we discuss later the sensibility of the obtained result for the ratio @xmath99 to different k values . in fig .
1 we present our results for the nuclear modified factor @xmath100 for au - au collisions at rhic for @xmath101 , @xmath7 , @xmath102 and @xmath4 using formulae ( 1 ) as a function of @xmath94 compared with the experimental data on nonphotonic electrons . for au+au central collisions , bars is data taken from phenix @xcite . ]
the overall normalization is given by the value of @xmath100 at @xmath103 which has to do with the factor @xmath104 . since we know the number of strings produced in @xmath12 and @xmath11 collisions , we know @xmath105 , @xmath106 , @xmath107 and @xmath108 and the only free parameter is @xmath109 . from the data we obtain @xmath110 gev / c .
the experimental errors allow us a @xmath111 freedom in the value of @xmath112 , however a higher value than 1.5 gev / c would not be realistic an a lower value will give rise to a higher normalization and therefore @xmath100 for @xmath113gev / c will exceed the experimental data even more than with the used value .
the value of @xmath100 for @xmath101 at low @xmath30 agrees with the results in @xcite@xcite , for @xmath114gev / c we obtain an @xmath100 larger than the non - photonic leptonic data . in fig .
2 , we present our results on @xmath100 at @xmath98 tev for a @xmath101 , @xmath7 and b. we see that , as expected , as energy increases the low @xmath30 @xmath100 increases , although the suppression at intermediate @xmath30 is similar . for pb+pb central collisions at @xmath98 tev . ] in fig 3 , we present the ratio @xmath115 for au - au at @xmath96gev .
we observe that the ratio increases up to a maximum of 1.45 around @xmath116 gev / c .
a very similar enhancement has been obtained in the dynamical recombination model @xcite .
, starts are used for @xmath117 , and errorbars are used for data from phenix for au - au central collisions at @xmath96 ( gev ) . ] for comparison we include also our results for @xmath117 at central au - au collisions together with experimental data @xcite . in fig .
4 we show the ratio @xmath115 for @xmath12 at @xmath96 gev . we observe a very smooth enhancement . for p+p collisions at @xmath96 gev .
] in fig 5 and fig .
6 we show our results for the ratio @xmath115 at @xmath98tev for pb - pb collisions and @xmath12 collisions respectively .
we observe in both of them larger enhancement than at rhic energies , particularly in the nuclear case . for pb - pb central collisions at @xmath98 tev .
] @xmath115 for p+p collisions at @xmath98 tev . ] in fig . 7 we plot the factor f at @xmath96 gev ( blue ) and at @xmath98 tev ( red line ) .
we observe that at rhic energies the factor @xmath14 is only slightly below one , and for @xmath118 gev / c it is clearly over 0.5 , which means that the @xmath115 enhancement in @xmath11 is not able to explain all the difference between experimentally observed suppression of @xmath100 for non - photonic electrons and the pqcd expectations .
we have studied the effects due to the uncertainties in the @xmath23 values for @xmath12 at this energy . for reasonable alternative @xmath23
values the enhancement of @xmath119 in pp with @xmath30 is larger giving rise to lower c factor in equation ( 3 ) and therefore the factor f is near to one , consistent with our main conclusion , namely that the @xmath115 enhancement is not able to explain all the difference between the experimentally observed values and the perturbative expectations .
the overlapping of the strings formed in the collision of heavy nuclei particles produces strong color fields which give rise to an enhancement of heavy flavor .
we have computed the nuclear modification factor of @xmath101 , @xmath7 and @xmath120 at rhic and lhc energies for @xmath11 collisions . referring to @xmath101
, we obtain a good agreement at low @xmath30 with the experimental data for the nuclear modification factor of non - photonic electrons . for @xmath30 values between 2 and 6
gev / c our results obtained are over the experimental data as pqcd @xcite .
the ratio @xmath115 as a function of @xmath30 for au - au collisions at @xmath96 gev is enhanced showing a maximum around 5 gev / c . such as enhancement is much larger at lhc energies .
however , the enhancement @xmath115 can explain only half of factor 2 difference between the experimental data and the pqcd expectations at rhic energies . in p - p collisions the ratio @xmath115 also rises as a function of @xmath30 but very smoothly at rhic energies . at lhc
this increase is a factor of 2 between @xmath87 and @xmath121 gev / c .
the enhancements of @xmath115 in aa and pp collisions are larger at lhc than at rhic as it was expected due to the stronger color fields produced .
we thank n. armesto and a. ayala and m. a. braun for discussions .
this research was supported by micinn of spain , under grant fpa2008 - 01177 and by xunta de galicia .
b. muller , and x. n. wang , _ phys .
lett . _ * 68 * , 2437 ( 1992 ) ; e. v. shuryak , _ phys .
lett . _ * 68 * , 3270 ( 1992 ) .
o. linnyk , e. l. bratkovskaya and w. cassing , _ int . j. mod
_ e * 17 * , 1367 ( 2008 ) .
j. adams _ et al . _ [ star collaboration ] , phys . rev
. lett . * 94 * , 062301 ( 2005 ) ; b. i. abelev _ et al . _ [ star collaboration ] , phys .
lett . * 98 * , 192301 ( 2007 ) .
s. s. adler _ et al .
_ [ phenix collaboration ] , phys .
lett . * 94 * , 082301 ( 2005 ) ; a. adare _ et al . _
[ phenix collaboration ] , phys .
lett . * 98 * , 172301 ( 2007 ) .
m. cacciari , m. greco and p. nason , jhep * 9805 * , 007 ( 1998 ) .
y. l. dokshitzer and d. e. kharzeev , phys .
b * 519 * , 199 ( 2001 ) .
m. djordjevic , m. gyulassy , r. vogt and s. wicks , phys .
b * 632 * , 81 ( 2006 ) .
n. armesto , m. cacciari , a. dainese , c. a. salgado and u. a. wiedemann , phys . lett .
b * 637 * , 362 ( 2006 ) .
a. adil and i. vitev , phys .
b * 649 * , 139 ( 2007 ) .
g. martinez - garcia , s. gadrat and p. crochet , phys .
b * 663 * , 55 ( 2008 ) ; g. martinez - garcia , s. gadrat and p. crochet , j. phys .
g * 35 * , 044023 ( 2008 ) .
d. c. zhou , z. conesa del valle , a. dainese , h. t. ding and g. martinez garcia , j. phys .
g * 36 * , 064055 ( 2009 ) .
a. ayala , j. magnin , l. m. montano , g. toledosanchez , phys .
c * 80 * , 064905 ( 2009 ) .
v. toporpop , m. gyulassy , j. barrette , c. gale , s. jeon and r. bellwied , phys .
c * 75 * , 014904 ( 2007 ) .
v. t. pop , j. barrette and m. gyulassy , phys .
lett . * 102 * , 232302 ( 2009 ) .
n. armesto , m. a. braun , e. g. ferreiro and c. pajares , phys .
b * 344 * , 301 ( 1995 ) ; j. dias de deus , e. g. ferreiro , c. pajares and r. ugoccioni , phys . lett .
b * 581 * , 156 ( 2004 ) .
d. kharzeev and k. tuchin , nucl .
phys . a * 735 * , 248 ( 2004 ) ; l. mclerran , j. phys .
g * 35 * , 104001 ( 2008 ) . n. armesto , m. a. braun , e. g. ferreiro and c. pajares , phys .
* 77 * , 3736 ( 1996 ) n. s. amelin , m. a. braun and c. pajares , z. phys .
c * 63 * , 507 ( 1994 ) .
n. s. amelin , m. a. braun and c. pajares , phys .
b * 306 * , 312 ( 1993 ) .
n. s. amelin , h. stocker , w. greiner , n. armesto , m. a. braun and c. pajares , phys .
c * 52 * , 362 ( 1995 ) .
r. c. hwa and c. b. yang , phys .
c * 65 * , 034905 ( 2002 ) [ erratum - ibid .
c * 67 * , 059902 ( 2003 ) ] .
r. j. fries , b. muller , c. nonaka and s. a. bass , phys .
* 90 * , 202303 ( 2003 ) .
j. dias de deus , e. g. ferreiro , c. pajares and r. ugoccioni , eur .
j. c * 40 * , 229 ( 2005 ) . c. pajares , eur .
j. c * 43 * , 9 ( 2005 ) .
m. a. braun and c. pajares , phys .
lett . * 85 * , 4864 ( 2000 ) ; m. a. braun and c. pajares , eur .
j. c * 16 * , 349 ( 2000 ) l. cunqueiro , j. dias de deus , e. g. ferreiro and c. pajares , eur .
j. c * 53 * , 585 ( 2008 ) a. capella , u. sukhatme , c. i. tan and j. tran thanh van , phys .
b * 81 * ( 1979 ) 68 .
a. capella , u. sukhatme , c. i. tan and j. tran thanh van , phys .
* 236 * ( 1994 ) 225 .
a. capella , c. pajares and a. v. ramallo , nucl .
b * 241 * ( 1984 ) 75 .
a. b. kaidalov and k. a. ter - martirosian , phys .
b * 117 * , 247 ( 1982 ) .
n. s. amelin , n. armesto , c. pajares and d. sousa , eur .
j. c * 22 * , 149 ( 2001 ) ; n. armesto , c. pajares and d. sousa , phys .
b * 527 * , 92 ( 2002 ) .
g. jona - lasinio , nuovo cim .
b * 26 * ( 1975 ) 99 .
j. dias de deus , c. pajares and c. a. salgado , phys .
b * 407 * , 335 ( 1997 ) a. capella , j. tran thanh van , r. blankenbecler , c. pajares and a. v. ramallo , phys .
b * 107 * , 106 ( 1981 ) phenix collaboration , phys .
88 * , 192303 ( 2002 ) .
m. g. ryskin , a. g. shuvaev and yu .
m. shabelski , phys . atom .
nucl . * 64 * , 120 ( 2001 ) ; [ yad . fiz . * 64 * , 123 ( 2001 ) ] ; c. merino , c. pajares , m. m. ryzhinskiy , yu .
m. shabelski , a. g. shuvaev , arxiv:0910.2364 [ hep - ph ] .
s. s. adler _ et al .
_ [ phenix collaboration ] , phys .
c * 69 * ( 2004 ) 034909 | the clustering of color sources provides a natural framework for soft partonic interactions producing strong color fields .
we study the consequences of these color fields in the production of heavy flavor and the behavior of the nuclear modification factor . |
in this paper we give a new characterization of the @xmath0-polynomial property for distance - regular graphs . in order to motivate and describe our result , we first recall some notions .
let @xmath1 denote a distance - regular graph with diameter @xmath13 and vertex set @xmath14 ( see section 2 for formal definitions ) .
recall that there exists a minimal idempotent @xmath4 of @xmath1 such that @xmath15 , where @xmath16 is the all 1 s matrix .
we call @xmath4 _ trivial_. let @xmath17 denote an ordering of the nontrivial minimal idempotents of @xmath1 .
it is known that for @xmath18 the entry - wise product @xmath19 is a linear combination of the minimal idempotents of @xmath1 , so that @xmath20 the coefficients @xmath21 are called the _ krein parameters _ of @xmath1 .
they are real and nonnegative ; see for example @xcite .
now consider when is a krein parameter zero .
note that @xmath22 for @xmath23 , so @xmath24 for @xmath25 .
the ordering @xmath17 is called _
@xmath0-polynomial _ whenever for all @xmath18 the krein parameter @xmath26 is zero if @xmath27 and nonzero if @xmath28 .
let @xmath3 denote a nontrivial minimial idempotent of @xmath1 .
we say that @xmath3 is _ @xmath0-polynomial _ whenever there exists a @xmath0-polynomial ordering @xmath17 of the nontrivial minimal idempotents of @xmath1 such that @xmath29 .
we now explain the @xmath0-polynomial property in terms of representation diagrams .
let @xmath3 denote a nontrivial minimal idempotent of @xmath1 , and for notational convenience write @xmath29 .
the _ representation diagram _
@xmath30 is the undirected graph with vertex set @xmath31 such that vertices @xmath32 are adjacent whenever @xmath33 and @xmath34 . by our earlier comments , @xmath35 for @xmath23
therefore , in @xmath30 the vertex @xmath36 is adjacent to the vertex @xmath37 and no other vertex , see figure 1(a ) .
observe that @xmath3 is @xmath0-polynomial if and only if @xmath30 is a path , and in this case the natural ordering @xmath38 of the vertices in @xmath30 agrees with the @xmath0-polynomial ordering associated with @xmath3 .
see figure 1(c ) .
+ ( a ) ( b ) ( c ) figure 1 : for @xmath29 the representation diagram @xmath30 around vertices 0 and 1 : ( a ) in general ; ( b ) when @xmath29 is a tail ; ( c ) when @xmath3 is @xmath0-polynomial .
= let @xmath3 denote a nontrivial minimal idempotent of @xmath1 . in this paper
we give a condition that is necessary and sufficient for @xmath3 to be @xmath0-polynomial .
we now describe the condition , which has three parts .
the first part has to do with the representation diagram @xmath30 . according to lang @xcite ,
@xmath3 is a _ tail _ whenever @xmath6 is a linear combination of @xmath4 , @xmath3 , and at most one other minimal idempotent of @xmath1 . in terms of the diagram @xmath30 , and writing @xmath29 for notational convenience ,
@xmath3 is a tail if and only if vertex @xmath37 is adjacent to at most one vertex besides vertex @xmath36 , see figure 1(b ) .
note that if @xmath3 is @xmath0-polynomial , then @xmath3 is a tail .
the next part of our condition involves the dual eigenvalue sequence @xmath39 for e. this sequence satisfies @xmath40 , where @xmath41 are the distance matrices of @xmath1 .
following @xcite , we say that @xmath3 is _ three - term recurrent _ ( in short ttr ) whenever there exists a complex scalar @xmath7 such that @xmath42 is independent of @xmath9 for @xmath10 .
if @xmath3 is @xmath0-polynomial , then @xmath3 is ttr by ( * ? ?
* theorem 8.1.2 ) , cf .
@xcite .
the third part of our condition involves both the diagram @xmath30 and the dual eigenvalues @xmath39 . by (
* proposition 2.11.1 ) , @xmath30 is connected if and only if @xmath43 for @xmath44 .
these equivalent statements hold if @xmath3 is @xmath0-polynomial , since in this case @xmath30 is a path .
we now state our main result .
[ main ] let @xmath1 denote a distance - regular graph with diameter @xmath13 .
let @xmath3 denote a nontrivial minimal idempotent for @xmath1 and let @xmath39 denote the corresponding dual eigenvalue sequence .
then @xmath3 is @xmath0-polynomial if and only if ( i ) : : @xmath3 is a tail , ( ii ) : : @xmath3 is ttr , ( iii ) : : @xmath11 for @xmath45 . in section [ cx ]
we discuss the minimality of assumptions ( i)(iii ) of theorem [ main ] .
we show that in general , no proper subset of ( i)(iii ) is sufficient to imply that @xmath3 is @xmath0-polynomial .
theorem [ main ] gives a characterization of the @xmath0-polynomial distance - regular graphs .
a similar characterization , where assumption ( i ) is replaced by some equations involving the dual eigenvalues and intersection numbers , is given by pascasio @xcite .
in this section we review some definitions and basic concepts .
see brouwer , cohen and neumaier @xcite and terwilliger @xcite for more background information .
let @xmath46 denote the complex number field and @xmath14 a nonempty finite set .
let @xmath47 denote the @xmath46-algebra consisting of all matrices whose rows and columns are indexed by @xmath14 and whose entries are in @xmath46 .
let @xmath48 denote the vector space over @xmath46 consisting of column vectors whose coordinates are indexed by @xmath14 and whose entries are in @xmath46 .
observe that @xmath47 acts on @xmath49 by left multiplication .
for all @xmath50 , let @xmath51 denote the element of @xmath49 with @xmath37 in the @xmath52-th coordinate and @xmath36 in all other coordinates
. from now on @xmath1 denotes a finite , undirected , connected graph , without loops or multiple edges , with vertex set @xmath14 , the shortest path - length distance function @xmath53 and diameter @xmath54 . for a vertex @xmath55 and integer @xmath56 define @xmath57 for notational convenience abbreviate @xmath58 . for an integer @xmath59
, the graph @xmath1 is said to be _ regular with valency k _ whenever @xmath60 for all @xmath61 .
the graph @xmath1 is said to be _ _ distance - regular whenever for all integers @xmath62 @xmath63 and vertices @xmath64 with @xmath65 , the number @xmath66 is independent of @xmath67 .
the constants @xmath68 are called the _ _ intersection numbers of @xmath1 . from now on assume @xmath1 is distance - regular with diameter @xmath13 .
note that @xmath1 is regular with valency @xmath69 .
we recall the bose - mesner algebra of @xmath1 . for @xmath70
let @xmath71 denote the matrix in @xmath47 with @xmath72-entry equal to @xmath37 if @xmath73 and @xmath36 otherwise .
we call @xmath71 the _
_ @xmath9-th distance matrix of @xmath1 .
note that @xmath71 is real and symmetric .
we observe that @xmath74 , where @xmath75 is the identity matrix , and abbreviate @xmath76 .
we observe that @xmath77 and @xmath78 for @xmath79 .
let @xmath80 denote the subalgebra of @xmath47 generated by @xmath81 .
by @xcite the matrices @xmath41 form a basis for @xmath80 .
we call @xmath80 the _ _ bose - mesner algebra of @xmath1 . by @xcite , @xmath80 has a basis @xmath82 such that ( i ) @xmath15 ; ( ii ) @xmath83 ; ( iii ) @xmath84 for @xmath85 .
by @xcite the matrices @xmath82 are real and symmetric .
we call @xmath86 the _ _ minimal idempotents of @xmath1 .
we call @xmath4 _
trivial_. since @xmath86 form a basis for @xmath80 , there exist complex scalars @xmath87 such that @xmath88 by ( [ prodaei0 ] ) and since @xmath89 we have @xmath90 we call the scalar @xmath91 the _ _ eigenvalue of @xmath1 corresponding to @xmath92
. note that the eigenvalues @xmath87 are mutually distinct since @xmath81 generates @xmath80 .
moreover @xmath93 are real , since @xmath81 and @xmath86 are real .
let @xmath3 denote a minimal idempotent of @xmath1 .
since @xmath41 form a basis for @xmath80 , there exist complex scalars @xmath39 such that @xmath94 we call @xmath95 the _ _ @xmath9-th dual eigenvalue of @xmath1 corresponding to @xmath3 .
note that @xmath96 are real , since @xmath3 and @xmath41 are real .
let @xmath97 denote the entry - wise product in @xmath47 .
observe @xmath98 for @xmath18 , so @xmath80 is closed under @xmath97 . therefore there
exist complex scalars @xmath21 such that @xmath99 we call the @xmath21 the _ _ krein parameters of @xmath1 .
these parameters are real and nonnegative @xcite . for the moment
fix integers @xmath62 @xmath63 . by construction @xmath100 . by
* lemma 2.3.1 ) we have @xmath101
. therefore @xmath102 we recall the dual bose - mesner algebra of @xmath1 @xcite . for the rest of this section ,
fix a vertex @xmath55 . for @xmath103
let @xmath104 denote the diagonal matrix in @xmath47 with @xmath105-entry @xmath106 we call @xmath107 the _ _ @xmath9-th dual idempotent of @xmath1 with respect to @xmath108 .
we observe that @xmath109 and @xmath110 for @xmath18 .
therefore the matrices @xmath111 form a basis for a commutative subalgebra @xmath112 of @xmath47 .
we call @xmath113 the _ _ dual bose - mesner algebra of @xmath1 with respect to @xmath108 . for @xmath103
let @xmath114 denote the diagonal matrix in @xmath47 with @xmath105-entry @xmath115 we call @xmath116 the _ dual distance matrix _ corresponding to @xmath92 .
by @xcite the matrices @xmath117 form a basis for @xmath113 .
select an integer @xmath118 @xmath119 and set @xmath120 , @xmath121 .
let @xmath39 denote the dual eigenvalues corresponding to @xmath3 . then using ( [ dualeigenvaluesdef ] ) we find @xmath122
moreover @xmath123 we now recall how @xmath80 and @xmath113 are related . by the definition of the distance matrices and dual idempotents
, we have @xmath124 by ( * ? ? ? *
lemma 3.2 ) , @xmath125 let @xmath126 denote the subalgebra of @xmath47 generated by @xmath80 and @xmath113 .
we call @xmath127 the _ _ subconstituent algebra or
_ _ terwilliger algebra of @xmath1 with respect to @xmath108 @xcite . by a
_ _ @xmath127-module we mean a subspace @xmath128 such that @xmath129 for all @xmath130 .
let @xmath131 denote a @xmath127-module .
then @xmath131 is said to be _ _
irreducible whenever @xmath131 is nonzero and contains no @xmath127-modules other than @xmath36 and @xmath131 .
we mention a special irreducible @xmath127-module .
let @xmath132 denote the all 1 s vector in @xmath49 .
observe that @xmath133 and @xmath134 for @xmath135 .
therefore @xmath136 . denote this common space by @xmath137 and note that @xmath137 is a @xmath127-module .
this @xmath127-module is irreducible by ( * ? ? ?
* lemma 3.6 ) .
for @xmath135 the vector @xmath138 is a basis for @xmath139 and @xmath140 is a basis for @xmath141 .
we call @xmath137 the _ primary _ @xmath127-module .
in this section we prove our characterization of @xmath0-polynomial distance - regular graphs .
[ claim1 ] let @xmath1 denote a distance - regular graph with diameter @xmath13 .
let @xmath3 denote a nontrivial minimal idempotent of @xmath1 that is ttr and let @xmath39 be the corresponding dual eigenvalues .
let @xmath7 and @xmath142 denote complex scalars such that @xmath143 for @xmath144 .
then the expression @xmath145 is independent of @xmath9 for @xmath45 .
let @xmath146 denote the expression in ( [ delta ] ) . for @xmath10
the difference @xmath147 is equal to @xmath148 and is therefore 0 .
it follows that @xmath146 is independent of @xmath9 for @xmath44 .
[ claim2 ] let @xmath1 denote a distance - regular graph with diameter @xmath13 .
let @xmath3 denote a nontrivial minimal idempotent of @xmath1 that is ttr . fix a vertex @xmath108 of @xmath1 and let @xmath149 denote the dual distance matrix corresponding to @xmath3
. then @xmath150,\ ] ] where @xmath7 and @xmath142 are from lemma [ claim1 ] and @xmath151 it the common value of ( [ delta ] ) . here
@xmath152 $ ] means @xmath153 .
let @xmath154 denote the expression on the right in ( [ commutator ] ) .
we show that @xmath155 .
observe @xmath156 to show @xmath155 it suffices to show @xmath157 for @xmath158 .
for notational convenience define a polynomial @xmath159 in two variables @xmath160 for @xmath161 we have @xmath162 by ( [ productdual1 ] ) , where @xmath96 are the dual eigenvalues corresponding to @xmath3 . by ( [ combcoeff ] )
we find @xmath163 if @xmath27 . by lemma [ claim1 ]
we find @xmath164 if @xmath28 .
of course @xmath165 if @xmath166 .
therefore @xmath167 as desired .
we have now shown @xmath155 . in ( [ propertiesdualbma ] )
we gave a characterization of a vanishing krein parameter .
we will need a variation on that result .
we will obtain this variation after a few lemmas .
the first lemma follows directly from the definitions of @xmath168 and the dual adjacency matrices . [ claim ]
let @xmath1 denote a distance - regular graph with diameter @xmath13 .
let @xmath3 denote a nontrivial minimal idempotent of @xmath1 .
fix a vertex @xmath108 of @xmath1 and let @xmath149 denote the dual distance matrix corresponding to @xmath3 .
then @xmath169 for all @xmath170 .
[ claimne00 ] let @xmath1 denote a distance - regular graph with diameter @xmath13 and let @xmath86 denote the minimal idempotents of @xmath1 .
fix @xmath55 .
then for @xmath171 , @xmath172 where @xmath173 . from lemma [ claim ]
we find @xmath174 and observe that this equals @xmath175 . now @xmath176 [ posledi ] let @xmath1 denote a distance - regular graph with diameter @xmath13 and let @xmath86 denote the minimal idempotents of @xmath1 .
fix @xmath55 and let @xmath137 denote the primary module for @xmath177 . then the following ( i ) , ( ii ) are equivalent for @xmath171 .
( i ) : : @xmath178 .
( ii ) : : @xmath179 vanishes on @xmath137 , where @xmath173 .
suppose ( i ) , ( ii ) fail
. then @xmath180 .
( i)@xmath181(ii ) @xmath179 is zero by ( [ propertiesdualbma ] ) and hence vanishes on @xmath137 .
( ii)@xmath181(i ) observe @xmath182 so @xmath183 .
therefore @xmath184 in view of lemma [ claimne00 ] .
the vector @xmath185 is a basis for @xmath186 so @xmath187 .
. suppose ( i ) , ( ii ) fail . then @xmath189 is a nonzero subspace of the one - dimensional space @xmath186 and is therefore equal to @xmath186 .
we are now ready to prove our main result .
_ proof of theorem _ [ main ] .
first suppose that @xmath3 is @xmath0-polynomial . then condition ( i ) holds by definition of a tail and the definition of the @xmath0-polynomial property , condition ( ii )
holds by ( * ? ? ?
* theorem 8.1.2 ) , cf .
leonard @xcite , and condition ( iii ) holds by ( * ? ? ?
* proposition 4.1.8 ) . to obtain the converse , assume that @xmath3 satisfies ( i)(iii ) .
we show @xmath3 is @xmath0-polynomial .
to do this we consider the representation diagram @xmath30 from the introduction .
we will show that @xmath30 is a path .
let @xmath17 denote an ordering of the nontrivial minimal idempotents of @xmath1 such that @xmath29 .
let @xmath14 denote the vertex set of @xmath1 .
fix @xmath55 and let @xmath190 .
we first show that @xmath30 is connected . to do this
we follow an argument given in ( * ? ? ?
* theorem 3.3 ) .
suppose that @xmath30 is not connected .
then there exists a nonempty proper subset @xmath191 of @xmath192 such that @xmath9 , @xmath193 are not adjacent in @xmath30 for all @xmath194 and @xmath195 . invoking ( [ propertiesdualbma ] ) we find @xmath196 and @xmath197 for @xmath198 and @xmath199 .
define @xmath200 and observe @xmath201 by a similar argument @xmath202 , so @xmath203 commutes with @xmath204 . since @xmath205 there exist complex scalars @xmath206 such that @xmath207 .
we have @xmath208 we claim that the matrices @xmath209 are linearly independent .
to prove the claim , for @xmath210 define @xmath211 , and observe @xmath212 .
the vectors @xmath213 are linearly independent and @xmath214 for @xmath44 so the vectors @xmath215 are linearly independent .
therefore the matrices @xmath216 are linearly independent and the claim is proved . by the claim and ( [ com ] ) we find @xmath217 for @xmath44 .
now @xmath218 .
but @xmath219 so @xmath220
. thus @xmath221 , in which case @xmath222 , or @xmath223 , in which case @xmath224 . in either case
we have a contradiction so @xmath30 is connected .
as we mentioned in the introduction , in @xmath30 the vertex @xmath36 is adjacent to vertex @xmath37 and no other vertex of @xmath30 . by the definition of a tail and since @xmath30 is connected , vertex @xmath37 is adjacent to vertex @xmath36 and exactly one other vertex in @xmath30 . to show that @xmath30 is a path
, it suffices to show that each vertex in @xmath30 is adjacent to at most two other vertices in @xmath30 .
we assume this is not the case and obtain a contradiction .
let @xmath225 denote a vertex in @xmath30 that is adjacent to more than two vertices of @xmath30 . of all such vertices ,
we pick @xmath225 such that the distance to @xmath36 in @xmath30 is minimal .
call this distance @xmath9 .
note that @xmath226 by our above comments and the construction .
for notational convenience and without loss of generality we may assume that the vertices of @xmath30 are labelled such that for @xmath227 the vertex @xmath193 is adjacent to @xmath228 and @xmath229 and no other vertex in @xmath30 . by construction
the chosen vertex @xmath225 is labelled @xmath9 .
this vertex is adjacent to @xmath230 and at least two other vertices in @xmath30 .
let @xmath231 denote a vertex in @xmath30 other than @xmath2321 that is adjacent to vertex @xmath9 . since @xmath3 is ttr there exists @xmath233 such that @xmath234 is independent of @xmath193 for @xmath235 .
we claim that @xmath236 to prove the claim we consider the equation ( [ commutator ] ) . in that equation
we expand the right - hand side to get @xmath237 in this equation we multiply each term on the left by @xmath238 and on the right by @xmath239 . to help simplify the results we make some comments . using ( [ prodaei ] )
we find @xmath240 . using ( [ propertiesdualbma ] ) we find @xmath241 similarly we calculate @xmath242 and @xmath243 e_{i-2 } a^ * a e_t&=&0 , & & \phantom{012345 } & e_{i-2 } a a^ * e_t & = & 0 .
\end{array}\ ] ] from these comments we find @xmath244 we show that @xmath245 . by the last assertion of corollary [ posledi ] and
since the sequence @xmath246 is a path in @xmath30 , we find @xmath247 and @xmath248 and @xmath249 .
therefore @xmath250 observe @xmath251 so @xmath245 , as desired . by this and ( [ calculateh ] )
we obtain ( [ unique ] ) . by ( [ unique ] )
the scalar @xmath252 is uniquely determined .
the scalars @xmath253 are mutually distinct so @xmath231 is uniquely determined , for a contradiction .
we have shown that @xmath30 is a path and therefore @xmath3 is @xmath0-polynomial .
in this section we make some remarks concerning the three conditions in theorem [ main ] . throughout the section
assume @xmath1 is a distance - regular graph with diameter @xmath254 and eigenvalues @xmath255 .
pick a nontrivial minimal idempotent @xmath256 of @xmath1 and let @xmath257 denote the corresponding dual eigenvalue sequence .
. we have @xmath268 by ( * ? ? ?
* lemma 2.2 ) , so @xmath269 .
define @xmath233 such that @xmath270 . by the construction @xmath271
is independent of @xmath9 for @xmath272 .
we are done in case ( i ) , so assume we are in cases ( ii ) or ( iii ) .
by @xcite we have @xmath273 for @xmath274 .
therefore @xmath271 is independent of @xmath9 for @xmath144 . in other words
@xmath3 is ttr .
* conditions ( ii ) , ( iii ) are not sufficient .
assume @xmath1 is the generalized hexagon of order @xmath275 @xcite .
it is primitive with diameter @xmath265 and eigenvalues @xmath276 .
pick @xmath277 . then @xmath3 satisfies condition ( ii ) by lemma [ lem:3 ] and @xmath3 satisfies condition ( iii ) by note [ note:2 ] .
but @xmath3 does not satisfy ( i ) by @xcite .
in particular @xmath3 is not @xmath0-polynomial .
* conditions ( i ) , ( iii ) are not sufficient .
assume @xmath1 is the dodecahedron , which is antipodal with diameter @xmath267 and eigenvalues @xmath278 , see @xcite .
pick @xmath277 . then @xmath3 satisfies condition ( i ) by @xcite and @xmath3 satisfies condition ( iii ) by note [ note:1 ] . by (
* lemma 2.2 ) we find @xmath279 and using this one verifies that @xmath3 does not satisfy condition ( ii ) . in particular @xmath3 is not @xmath0-polynomial .
* conditions ( i ) , ( ii ) are not sufficient .
assume @xmath1 is the wells graph @xcite , which is antipodal with diameter @xmath266 and eigenvalues @xmath280 .
pick @xmath281 .
then @xmath3 satisfies condition ( i ) by @xcite and @xmath3 satisfies condition ( ii ) by lemma [ lem:3 ] , but @xmath3 does not satisfy condition ( iii ) by note [ note:1 ] .
in particular @xmath3 is not @xmath0-polynomial .
[ note:5 ] assume @xmath1 is bipartite , and consider the conditions ( i)(iii ) of theorem [ main ] .
if @xmath3 satisfies conditions ( i ) , ( iii ) then @xmath3 is @xmath0-polynomial by ( * ? ? ?
* proof of theorem 5.4 ) .
if @xmath3 satisfies conditions ( ii ) , ( iii ) then @xmath3 is @xmath0-polynomial by ( * ? ? ?
* theorem 10.5 ) . | we obtain the following characterization of @xmath0-polynomial distance - regular graphs .
let @xmath1 denote a distance - regular graph with diameter @xmath2 .
let @xmath3 denote a minimal idempotent of @xmath1 which is not the trivial idempotent @xmath4 .
let @xmath5 denote the dual eigenvalue sequence for @xmath3 .
we show that @xmath3 is @xmath0-polynomial if and only if ( i ) the entry - wise product @xmath6 is a linear combination of @xmath4 , @xmath3 , and at most one other minimal idempotent of @xmath1 ; ( ii ) there exists a complex scalar @xmath7 such that @xmath8 is independent of @xmath9 for @xmath10 ; ( iii ) @xmath11 for @xmath12 . = |
in a series of papers @xcite , we have recently introduced and studied a model of three - dimensional gravity that has proven to be very useful to understand some of the fundamental features of canonical loop quantum gravity and spin foam models .
this new formulation of gravity in three dimensions is particularly interesting because it admits a barbero - immirzi parameter , and can be written as a holst or a plebanski action . as such
, it can be used to study the imposition of the simplicity constraints in the construction of spin foam models , the role of @xmath1 and of the @xmath0 ashtekar - barbero formulation , and the relation with the self - dual variables . in @xcite
, it has been shown that this lorentzian three - dimensional model can be written at the hamiltonian level in terms of @xmath0 canonical variables ( just like in the four - dimensional case ) , which leads to discrete and @xmath1-dependent geometrical operators . however , it turns out to be possible to simplify the hamiltonian constraint ( and write it as a flatness constraint ) by working with complex canonical variables , provided that one imposes simplicity - like conditions ensuring that the corresponding connection be @xmath2-valued . in terms of these new variables ,
the spectra of the geometric operators become continuous and @xmath1-independent . in four - dimensional loop quantum gravity , the discreteness of quantum geometry at the planck scale , together with the existence of a non - zero minimal eigenvalue for the area operator , play a crucial role in the loop quantum cosmology scenario of singularity resolution @xcite and are essential for the recovery of the semi - classical thermodynamical properties of black holes @xcite . independently of the question of wether or
not the geometric operators will remain discrete and @xmath1-dependent at the dynamical level , it has been argued that already at the kinematical level a different choice of connection could potentially lead to continuous and @xmath1-independent spectra @xcite .
this observation has always remained at the formal level , since no hilbert space is known for the lorentz - covariant theory in terms of the shifted connection of alexandrov .
it is however known that this approach is closely related to the quantization in terms of the complex self - dual ashtekar connection , a route which has been recently reconsidered in the light of results on black hole entropy @xcite and the asymptotics of spin foam models @xcite . in the absence of a known way of constructing a full quantum theory with the self - dual variables ,
one is forced to investigate toy models and to focus on particular calculations such as the derivation of black hole entropy .
the hope is that this will provide informations as to how the self - dual quantum theory should be built .
in particular , a key open question concerns the role of the barbero - immirzi parameter and of the @xmath0 variables .
are these fundamental variables of quantum gravity with @xmath1 playing a true physical role , or simply a necessary regularization procedure needed in order to deal with the non - compactness of the gauge group of lorentzian gravity ?
it seems to us that the latter point of view should be adopted . in this short paper , we give an example of the above - mentioned duality between the nature of the geometric spectra ( discrete or continuous ) and the choice of connection . for this , we use the holst formulation of three - dimensional gravity to compare the quantum theories built with the @xmath3 ashtebar - barbero and the @xmath4 self - dual connection .
more precisely , our starting point for this study is the observation that the three - dimensional model of interest can be obtained from a symmetry reduction of the four - dimensional holst action , defined by imposing a symmetry along an arbitrary spatial direction .
fortunately , this reduction preserves the presence of @xmath1 , and the resulting theory shares most of the properties of its four - dimensional counterpart .
in particular , in the time gauge , the classical dynamical variables are given by an @xmath0 ashtekar - barbero connection @xmath5 canonically conjugated to an @xmath3-valued electric field @xmath6 .
furthermore , the six first class constraints that generate the isometries of the locally flat space - time can be decomposed , as in four dimensions , in terms of a gauss constraint , a vectorial constraint , and a scalar constraint .
since the theory is topological in three dimensions , the physical degrees of freedom are captured by the holonomies @xmath7 of the connection along the links @xmath8 of a unique graph @xmath9 ( sufficiently refined to resolve the topology of the spatial surface ) , and the `` one - dimensional fluxes '' of the electric field @xmath10 along links @xmath11 ( dual to @xmath8 ) of the dual graph @xmath12 . at the quantum level , these ( non - local ) variables are promoted to non - commutative operators which satisfy the three - dimensional analogue of the holonomy - flux algebra . as usual ,
kinematical states are @xmath0 spin networks associated to the graph @xmath9 , and the geometric operators are constructed in terms of the flux operators @xmath10 .
the spectrum of the length operator is given by a sum of fundamental contributions @xmath13 , where @xmath14 labels unitary irreducible representations of @xmath0 , and @xmath15 is the three - dimensional planck length .
motivated by the fact that this @xmath0 hamiltonian theory should describe lorentzian three - dimensional gravity , i.e. @xmath16 bf theory , the hamiltonian constraint can be recast into the form of a flatness constraint .
the canonical connection then becomes complex and @xmath4-valued , although it is defined with @xmath17 . on top of this complex formulation , it is possible to impose two types of linear simplicity conditions similar to that used in spin foam models .
the first form of the constraint ( which relates the @xmath3 generators and their complement in @xmath4 ) gives back the @xmath3-valued ashtekar - barbero connection , while the second one ( which relates the @xmath2 generators and their complement in @xmath4 ) leads to an @xmath2 connection . constructing the quantum theory with the latter , we observe that the length operator is given by a sum of fundamental contributions @xmath18 ( with @xmath19 or @xmath20 depending on the regularization of the length operator ) , where @xmath21 is a real parameter labeling the continuous series of representations of @xmath16 .
this spectrum is therefore continuous and @xmath1-independent .
this paper is organized as follows . in the next section ,
we briefly review the classical formulation that serves as the starting point for this analysis . in section [ sec:3 ] , we present the two different choices of connection that can be made at the classical level , and discuss the resulting quantum theories .
section [ sec:4 ] is devoted to the introduction of the classical geometric operators , while their spectra are studied in section [ sec:5 ] .
let us start with the four - dimensional holst action s_[e,]=__4(_ijkle^ie^j f^kl+_ijkle^ie^jf^kl ) , and perform a spacetime symmetry reduction that keeps the internal gauge group @xmath22 unbroken . for this , we assume that the four - dimensional spacetime manifold has the topology @xmath23 , where @xmath24 is a three - dimensional spacetime manifold , and @xmath25 is a spatial direction along which the theory is invariant .
denoting by @xmath26 the coordinate along @xmath25 , the invariance of the basic fields can be expressed as [ symmetry conditions ] _ 3=0,_3^ij=0 . as a consequence
, the resulting reduced action is given by [ 3d action ] s[e ,
x,]=__3 ^ 3x^(_ijklx^ie_^jf_^kl+_ijklx^ie_^jf_^kl ) , where @xmath27 is an @xmath4-valued connection over @xmath24 , @xmath28 is an @xmath4-valued one form over @xmath24 , and @xmath29 is an @xmath4-valued scalar on @xmath24 . as usual , @xmath30 is the totally antisymmetric spacetime tensor , @xmath31 denotes the totally antisymmetric internal tensor , and @xmath32 is defined in term of the flat minkowski metric @xmath33 .
we assume that @xmath34 , where @xmath35 is a two - dimensional surface with no boundaries . in order to mimic the construction of four - dimensional loop quantum gravity
, we performed in @xcite the canonical analysis of the three - dimensional action ( [ 3d action ] ) in the time gauge .
this gauge fixing is defined by the conditions @xmath36 , and breaks the internal gauge group @xmath22 into its maximal compact subgroup @xmath0 . just like in the four - dimensional theory
, the resulting phase space is parametrized by the canonical pair @xmath37 , where @xmath38 is the three - dimensional analogue of the @xmath3 ashtekar - barbero connection , and @xmath39 its conjugate electric field .
these variables satisfy the poisson bracket [ poisson bracket ] \{e^a_i(x),a_b^j(y)}=^a_b^j_i^2(x - y ) , and are subject to the following gauss , vectorial , and scalar constraints : [ constraints ] g=_ae^a+a_ae^a , h_a=_abe^bf_12,h_0=x(f_12-(1+^-2)k_1k_2 ) , where @xmath40 is the curvature of @xmath5 .
the variable @xmath41 is the three - dimensional analogue of the extrinsic curvature , and can be written as @xmath42 , where @xmath43 is the reduced version of the four - dimensional levi - civita connection @xmath44 .
its explicit expression is given by _
a(e)=u_au+_abu , where @xmath45 .
to finish with this brief overview of the phase space structure , let us mention that @xmath46 is colinear to @xmath47 , which makes the above expression for @xmath48 strictly similar to the scalar constraint of the four - dimensional full theory . furthermore , because the vectors @xmath49 are orthogonal to @xmath50 @xcite , the quantity @xmath51 is in the direction of @xmath50 , and @xmath52 and @xmath48 can be viewed as the component of the same vector @xmath53 defined by [ f=0 in time gauge ] h = f_12-(1+^-2)k_1k_2 .
this property is in some sense responsible for the fact that this three - dimensional model is exactly soluble .
we now have all the ingredients to perform the loop quantization of the theory .
kinematical states are cylindrical functions of the connection associated with graphs embedded in the spatial surface @xmath35 . because of the topological nature of three - dimensional gravity , a single graph @xmath9 is sufficient to capture all the physical content of the theory . on this graph ,
let us introduce the holonomies of the connection along the links @xmath57 , and the fluxes of the electric field along links @xmath58 dual to @xmath8 .
these variables , i.e. @xmath59 are the building blocks of the three - dimensional ( classical and quantum ) holonomy - flux algebra
. the action of the flux operator on the holonomies evaluated in the spin-@xmath14 representation of @xmath3 is given by [ actionofl ] x^i_^*^(j)(h_(a))=_,^(j)(h_c(a ) ) , where @xmath60 is the index between @xmath8 and @xmath11 , the point @xmath61 denotes the intersection @xmath62 , and @xmath63 denotes the spin-@xmath14 representation matrix associated to the @xmath64-dimensional vector space @xmath65 .
the elements @xmath66 generate the lie algebra @xmath3 and satisfy , by convention , the lie algebra commutation relations @xmath67=\eps_{ij}^{~~k } j_k$ ] , where @xmath68 is the totally antisymmetric tensor with @xmath69 , and indices are lowered and raised with the flat euclidean metric @xmath70 . as usual ,
cylindrical functions associated to the graph @xmath9 form the kinematical hilbert space where the scalar product is constructed from the @xmath0 haar measure . in principle , physical states and observables should be constructed by solving the quantum constraint @xmath71 . in this set of constraints ,
only the imposition of the scalar constraint @xmath72 is problematic since the vectorial constraint @xmath73 imposing spatial diffeomorphism invariance has already been solved implicitly by fixing the graph @xmath9 .
the difficulty in solving @xmath72 is essentially the same as in the four - dimensional theory , and requires an appropriate regularization of the term @xmath51 when written in terms of the non - linear expression @xmath43 .
this difficulty can however be bypassed by noticing that in this three - dimensional model the constraints @xmath71 given by ( [ f=0 in time gauge ] ) can be written in the form of the flatness constraint of a bf theory .
more precisely , one can look for a connection @xmath74 for which @xmath71 is equivalent to @xmath75 , where @xmath76 is the curvature of @xmath74 .
it was shown in @xcite that , up to gauge transformations , there are only two connections that satisfy this requirement .
these are the self - dual and anti self - dual components of the initial @xmath22 connection @xmath77 .
the connection @xmath74 is therefore complex and has to be interpreted as an @xmath4-valued connection .
therefore , we have somehow traded the problem of dealing with the complicated hamiltonian constraint ( [ f=0 in time gauge ] ) for that of imposing the reality conditions . the same fact is obviously true in four dimensions , where one can take @xmath78 to obtain the self - dual scalar constraint , at the expense of working with the complex ashtekar connection .
we arrive at the conclusion that the quantum dynamics can in principle be solved in terms of the connection @xmath74 , since with this variable the hamiltonian constraint is manageable .
this means that one should consider @xmath22 spin network states instead of the standard @xmath0 ones , which introduces numerous technical problems due to the non - compactness of the lorentz group .
a priori , such spin network states require a regularization in order to be well - defined . fortunately ,
as we are about to see , this problem does drastically simplify due to the constraints satisfied by the complex connection @xmath74 .
as explained in @xcite , the connection @xmath74 can be written in the form [ decomposition of a ] = + , where @xmath66 are the @xmath3 generators introduces above , and @xmath79 are the boost generators satisfying the standard @xmath4 commutation relations .
one can see from this expression that @xmath74 possesses two different parts , which are the two terms between the square brackets .
interestingly , while the first term has no direct algebraic interpretation , the second term defines an @xmath2 component .
indeed , it can easily be shown that the elements @xmath80 and the two independent components of @xmath81 generate the lie algebra @xmath2 .
now , one can use the fact that there exists another legitimate choice of gauge for the action ( [ 3d action ] ) . indeed , in the gauge @xmath82 , the action ( [ 3d action ] ) reduces to that of @xmath16 bf theory , which is in complete agreement with the fact that it describes lorentzian three - dimensional gravity .
for this reason , consistency of the theory in the time gauge requires that we impose that the connection @xmath74 be @xmath2-valued as well .
this requirement can be met if the first term in ( [ decomposition of a ] ) is vanishing , which leads to the constraint @xmath83 .
this simplicity - like constraint is equivalent to a reality condition , in the sense that it selects a real section of @xmath22 corresponding to its non - compact subgroup @xmath16 .
once it is imposed at the classical level , the linear simplicity constraint reduces the connection ( [ decomposition of a ] ) to the following @xmath2 connection : [ cal a ] = ^-1(a^1f_1+a^2f_2-a^3f_0 ) , where @xmath84 are the components of the initial ashtekar - barbero connection @xmath5 expressed in the basis @xmath85 , and the family @xmath86 generates @xmath2 with the commutation relations = f_0,=f_1,=-f_2 .
these generators are related by @xmath87 .
now that the @xmath2 connection ( [ cal a ] ) is defined , one can in principle construct the physical hilbert space by imposing the quantum @xmath16 flatness constraint .
physical states are @xmath16 spin networks with support on the graph @xmath9 .
this defines the vector space structure of the physical hilbert space , and it remains to define the scalar product between the states and eventually to eliminate the zero - norm states .
in order to do so , it is much simpler to choose @xmath9 to be a minimal graph . when the spatial slice @xmath35 is a riemann surface of genus @xmath88 with no boundaries , the minimal graph consists in only one vertex @xmath89 and @xmath90 loops @xmath91 starting and ending at @xmath89 .
the loops @xmath92 , with @xmath93 , are the non - contractible loops around the handles of @xmath35 , and they can be identified with the standard generators of the fundamental group @xmath94 of the surface . for obvious reasons , such a minimal graph is called a flower graph .
physical states are then totally defined by complex - valued functions @xmath95 on @xmath96 , where the coset by @xmath16 traduces the gauge - invariance at the unique vertex @xmath89 of @xmath9 , and the physical inner product between two such states is formally defined by [ physical product ] _ 1,_2=(a , b)([a , b])_2(a , b ) , where @xmath97 is a `` regularized '' measure on @xmath98 , @xmath99 is the dirac distribution on @xmath16 , and the commutator is @xmath100=\prod_{i=1}^g a_ib_ia_i^{-1}b_i^{-1}$ ] . for simplicity ,
we have identified here the states with the corresponding funtions on @xmath98 . at this point ,
the physical scalar product and the physical hilbert space are defined only at the formal level , but we will make this construction concrete and explicit later on in the case of the torus .
to finish , let us recall that the physical scalar product ( [ physical product ] ) should reproduce the lorentzian ponzano - regge amplitudes @xcite , just like it is the case for the euclidean signature @xcite .
we now introduce the three - dimensional length and area operators . for this
, we first define the four - dimensional area and volume operators , and then reduce them with the symmetry ( [ symmetry conditions ] ) .
recall that in four spacetime dimensions the hamiltonian theory is written assuming that @xmath101 .
the area @xmath102 and volume @xmath103 operators are then defined on the spatial slice @xmath104 , and measure the area and the volume of a surface @xmath105 and a region @xmath106 , respectively .
the classical expressions for the area @xmath102 of a surface @xmath107 of normal @xmath108 , and the volume @xmath103 of a bounded region @xmath109 , depend only on the electric field @xmath110 , with @xmath111 , and are given by [ a and v in 4d ] _ 4(s)=_s^2x,_4(r)=_r^3x .
notice that for @xmath112 , @xmath113 coincides with the electric field ( [ poisson bracket ] ) of the three - dimensional action ( [ 3d action ] ) , and is given by @xmath114 , whereas for @xmath115 we have @xmath116 .
the compactification that we have used to obtain the three - dimensional holst action is of the type @xmath23 .
since we have further assumed that @xmath34 in order to perform the canonical analysis , we can write that _
4=_3=_3 ^ 1=_2 ^ 1 .
in other words , the spatial slice @xmath104 on which the operators ( [ a and v in 4d ] ) are defined is given by @xmath117 .
therefore , the surface @xmath105 measured by @xmath102 can be of two types , as represented in figure [ operators ]
. it can either be of the form @xmath118 , where @xmath119 is a path ( i.e. a one - dimensional submanifold ) in @xmath35 and @xmath120 a segment in @xmath25 , or it can be a two - dimensional submanifold of @xmath35 . and representation of the two types of surfaces that can be embedded in @xmath104 . the surface
@xmath121 is a two - dimensional submanifold of @xmath35 , while the surface @xmath122 is of the type @xmath123 , where @xmath119 is a one - dimensional submanifold of @xmath35 and @xmath120 a segment in the direction @xmath115 .
the length operator on @xmath35 can be derived from the area operator @xmath124 on @xmath104 , while the area operator on @xmath35 can be derived from the volume operator @xmath103 on @xmath104 . ]
let us start with the case @xmath118 , where @xmath119 is a path in @xmath35 and @xmath120 a segment of length @xmath125 in the third direction @xmath115 .
then , the normal @xmath126 is orthogonal to the third direction and has only components @xmath127 in the directions @xmath128 and @xmath129 ( i.e. in @xmath35 ) .
the area reduces to @xmath130 , where @xmath131 is the length of @xmath119 given by [ 3d length ] ( l)=_4(s)=_lx . in this way
, we obtain immediately the length of any curve @xmath119 in @xmath35 in terms of the electric field and the unit vector @xmath108 normal to this curve . when the surface @xmath107 is a two - dimensional submanifold of @xmath35 , then it is immediate to see that @xmath102 gives the area of the surface @xmath107 . in this case
, @xmath108 is necessarily in the third direction , i.e. @xmath132 and @xmath133 , and the expression @xmath102 for the area reduces to ( s)=_s^2x , which can be expressed in terms of the triad field since @xmath116 . to write this expression in terms of the electric field
@xmath134 only , we use a symmetry of the action ( [ 3d action ] ) described in @xcite , which states that @xmath46 can be chosen to be of unit norm ( @xmath135 ) , and that it is orthogonal to @xmath136 ( i.e. @xmath137 ) .
if we make this choice , then the area takes the simple form [ 3d area ] ( s)=_s^2x=_s^2x .
the operators @xmath131 and @xmath138 defined in ( [ 3d length ] ) and ( [ 3d area ] ) measure respectively the length of a curve @xmath139 and the area of a surface @xmath121 in our three - dimensional model .
notice that we could have derived the expression for the area operator @xmath138 of the three - dimensional theory by considering the volume operator @xmath103 defined for a region @xmath109 of the type @xmath140 , where @xmath107 is a surface in @xmath35 and @xmath120 a segment of length @xmath125 in @xmath25 ( see figure [ operators ] ) .
indeed , for such a region we have @xmath141 , and one can then directly obtain ( [ 3d area ] ) . now that we have derived the classical expressions for the geometric operators of the three - dimensional theory ( that are defined on the two - dimensional surface @xmath35 ) , we can quantize them and study their action on the above - defined @xmath0 and @xmath16 states . as it is the case for the four - dimensional volume operator @xmath103 , the three - dimensional area operator @xmath138 requires a regularization in order to be well - defined in terms of the quantum flux operators @xmath142 @xcite .
these regularizations have been studied and well - understood in the literature , and they will not affect our result .
the length operator is associated to links @xmath11 of the dual graph @xmath12 , and the length @xmath143 of the link @xmath11 dual to @xmath8 can be expressed in terms of the fluxes ( [ holflux ] ) according to [ length operator ] ( ^*)=. the length @xmath144 of any path @xmath145 defined by a composition @xmath146 of a sequence of @xmath108 elementary links around non - contractible handles of the surface @xmath35 .
an open link @xmath119 would not correspond to a physical ( dirac ) observable apart if particles are coupled to gravity and located at the end points of @xmath119 . ] is simply given by the sum @xmath147 .
therefore , computing the action of @xmath143 on the holonomy @xmath148 or @xmath149 allows to compute the action of any length operator @xmath144 on the states of the @xmath0 or @xmath16 quantum theories .
this action is non - trivial only when @xmath150 , it is diagonal and given by [ lengthspectrum ] ( ^*)^(j)(h_(a))&=&^(j)(h_(a ) ) , + ( ^*)^(s)(g_())&=&^(s)(g _ ( ) ) , where @xmath151 and @xmath152 are the quadratic casimir operators of the lie algebras @xmath3 and @xmath2 , respectively .
we have used the notation @xmath14 to denote unitary irreducible representations of @xmath3 and generically @xmath21 for those of @xmath2 ( whose class has not been fixed yet ) . while @xmath153 is always positive and discrete because @xmath14 is a non - negative half - integer , the sign of @xmath154 depends on the class of @xmath2 representations @xmath21 under consideration . in order to have a self - adjoint length operator ( with real and positive eigenvalues ) ,
@xmath21 must belong to the continuous series . in this case
, the spectrum @xmath155 of the length operator is continuous . before studying the area operator ,
let us make some important remarks . 1 .
first , we would like to emphasize that , in addition to being continuous , the length spectrum becomes also independent of the barbero - immirzi parameter @xmath1 .
the reason is simple and comes from the overall factor of @xmath156 in the expression ( [ cal a ] ) for @xmath74 , which cancels with the factor of @xmath1 coming from the poisson bracket ( [ poisson bracket ] ) when computing the action of @xmath142 on @xmath157 , i.e. as to be understood with @xmath158 , while for @xmath66 we have @xmath159 . ]
x^i_^*^(s)(g_ ( ) ) = ( -)(^-1)_,^(s)(g_c ( ) ) , where @xmath160 is equal to @xmath161 when one acts with the third component @xmath162 of the flux .
the factor of @xmath163 is responsible for the fact that the action of the @xmath3-invariant operator @xmath164 produces the @xmath2 quadratic casimir when it acts on the holonomies @xmath157 .
of course , one could think of writing the action of the length operator on the @xmath0 holonomies @xmath7 in the basis @xmath165 instead of @xmath66 , but then the spectrum would come with an incorrect minus sign .
3 . even if the spectrum ( [ lengthspectrum ] ) is continuous when using the connection @xmath74 instead of @xmath5 , there exists a length gap between `` no length '' and the first non - trivial possible length .
another regularization , discussed in @xcite , would give @xmath21 instead of @xmath166 for the action of the length operator on holonomies colored with a representation @xmath21 , and no length gap would exist in this case .
we refer the reader to @xcite for a discussion of this ambiguity .
the area operator in three dimensions is the analogue of the volume operator in four dimensions , and it acts on the vertices of the spin networks . for obvious reasons , we will be interested only in its action on three - valent vertices . denoting by @xmath167 , @xmath168 and
@xmath169 the three links meeting at the vertex @xmath89 , we will use the notation @xmath170 for the area operator at @xmath89 .
it is straightforward to obtain the following expression for the regularized area operator acting on the three - valent vertex @xmath89 @xcite : ( _
1,_2,_3)^2=16_i < j(x_^*_i^2x_^*_j^2-(x_^*_ix_^*_j)^2 ) , where @xmath171 label the edges meeting at the vertex .
one can then show that the action of the area operator is diagonal , with eigenvalues given by ( _ 1,_2,_3)^2=16_(i , j)(_i^2_j^2-(_i_j)^2 ) , where the sum contains only the three terms @xmath172 , @xmath173 and @xmath174 .
the lie algebra elements @xmath175 denote the @xmath3 or @xmath2 generators depending on the states that we act on ( i.e. the @xmath0 or @xmath16 spin networks ) , and the scalar product in the internal space is defined with the flat euclidean metric @xmath70 or the flat minkowskian metric @xmath176 , depending on the choice of connection .
finally , @xmath177 denotes the eigenvalue of the length operator @xmath178 ( [ length operator ] ) . using the gauss law at the vertex @xmath89 to write @xmath179
, an immediate computation leads to [ areaproduct ] ( _ 1,_2,_3)^2&=&16_(i , j)(_i^2_j^2 -14(_i^2+_j^2-_k^2)^2 ) + & = & 116(_1+_2+_3)(-_1+_2+_3 ) ( _ 1-_2+_3)(_1+_2-_3 ) , where @xmath180 . we obtain the classical expression for the area of a triangle with edges of length @xmath177 , and the geometrical interpretation of this formula
is therefore immediate .
however , an extra condition has to be imposed in order for the area eigenvalues to be real .
indeed , the product ( [ areaproduct ] ) is positive only if the three lengths @xmath177 satisfy the triangular inequalities .
contrary to the @xmath0 kinematical case where the triangular inequalities are automatically satisfied due to the standard properties of tensor products between unitary irreducible representations , here this condition has to be imposed by hand because the representations are in the continuous series of @xmath16 .
to finish , let us study more precisely the case in which @xmath35 is a torus .
the torus can be represented by the flower graph with two petals as shown in figure [ torus ] . and two loops , @xmath167 and @xmath168 , corresponding to the non - contractible cycles of the torus .
a spin network state associated to the flower graph is an assignment of unitary irreducible representations of the suitable lie algebra to the two loops and to the vertex .
the vertex being four - valent , it can be decomposed into two three - valent vertices , with the spin previously attached to @xmath89 now coloring the link @xmath169 .
as shown on the right , this graph with three links can be seen as dual to the rectangle with identified opposite edges that defines the torus .
this rectangle can be triangulated into two triangles sharing the side of length @xmath181 and dual to the vertices @xmath182 and @xmath183 .
the spins @xmath184 and @xmath185 coloring the loops @xmath167 and @xmath168 represent the edge lengths of the rectangle . ] in this case , one can construct the kinematical spin network states of the @xmath0 theory based on the three - dimensional ashtekar - barbero connection @xmath38 .
these states , denoted by @xmath186 , are totally characterized by the three spin labels @xmath187 coloring the two loops @xmath167 and @xmath168 ( for the first two labels ) and the vertex @xmath89 .
note that the choice of a `` shape '' for the intertwiner is not important ( i.e. the choice of a triangulation for the rectangle in figure [ torus ] is arbitrary ) .
our choice , which is depicted in figure [ torus ] , corresponds to a triangulation of the parallelogram into two triangles .
since the graph is minimal , the kinematical scalar product is given by _
( j_1,j_2,j_3),_(j_1,j_2,j_3)=_i=1 ^ 3_j_i , j_i , where @xmath188 denotes the kronecker delta .
the length and area operators of the @xmath0 kinematical theory are unitary with respect to this scalar product since their action is diagonal with positive eigenvalues .
furthermore , the geometric interpretation of the spin labels is immediate : they are associated with the length of the three edge of the ( similar ) triangles that discretize the parallelogram whose opposite sides are identified to construct the torus . now , the construction of the physical states would a priori require to solve the ashtekar - barbero hamiltonian constraint ( [ f=0 in time gauge ] ) .
just like in four - dimensional loop quantum gravity , it is not known wether or not at the physical level the @xmath1 dependency will disappear and the nature of the geometric operators change . however , in the present three - dimensional model , the route to constructing the physical states is known .
one can go back to the complex connection ( [ decomposition of a ] ) , impose the simplicity constraint that turns it into an @xmath2 connection , and then the hamiltonian constraint becomes simply the flatness constraint of @xmath16 bf theory . in this case , the scalar product between physical states is given by _
( s_1,s_2,s_3),_(s_1,s_2,s_3)=_i=1 ^ 3(s_i - s_i ) , where @xmath189 is the dirac delta distribution , and any other quantization scheme ( like the @xmath0 formulation in the time gauge ) should lead to the same result in order for it to be anomaly - free .
in this short paper , we have studied the loop quantization of three - dimensional gravity based on two different classical formulations of the hamiltonian theory . using the results of @xcite
, we have first written the canonical theory in terms of @xmath0 ashtekar - barbero variables with a real barbero - immirzi parameter .
then , we have argued that there is a simpler hamiltonian formulation ( in terms of a flatness constraint ) , in which the connection becomes complex and has to be supplemented by appropriate reality conditions .
when these are implemented classically in the form of a linear simplicity constraint , one recovers an @xmath2 connection which does still depend on real values of the barbero - immirzi parameter @xmath1 .
however , as we have shown , this connection leads to a quantum theory with @xmath1-independent and continuous spectra for the geometric operators .
it is important to stress out that what we did was not to compare the new @xmath0 ashtekar - barbero formulation of three - dimensional gravity with usual first order lorentzian gravity ( i.e. @xmath16 bf theory ) . instead ,
starting from the @xmath0 theory , we have naturally ended up with a new choice of connection that enables to work out the hamiltonian constraint ( just like the choice @xmath190 in four dimensions ) , with the additional requirement that this connection should satisfy a reality condition similar to the spin foam linear simplicity constraint .
which lessons can be learned for four - dimensional loop quantum gravity ?
the present model shows that already at the kinematical level one can have two very different quantum theories depending on the choice of connection .
the same observation seems to be true in four dimensions , although it has never been made rigorous because of the absence of a kinematical arena for the self - dual connection of ashtekar or the non - commutative shifted connection of alexandrov .
moreover , it is clear in this three - dimensional model that physical states are @xmath16 spin networks , therefore raising the question of wether the implementation of the ashtekar - barbero hamiltonian constraint ( [ f=0 in time gauge ] ) will also lead to these physical states , and to the disappearance of the barbero - immirzi parameter .
even if we have not addressed this question here , our model could be taken as a starting point to investigate this direction .
to summarize , if one tries to support the point of view that @xmath1 should not play a role at the physical level , there are essentially three possible scenarios . 1 .
the first one is a quantum theory based on the @xmath0 kinematics , and in which the barbero - immirzi parameter disappears in the quantum theory at the level of physical states ( with possibly the spectra becoming continuous ) . from what is known today about the quantum scalar constraint
, this does not seem to be the case .
moreover , the eprl and fk@xmath191 spin foam models are designed in such a way that @xmath1 plays a central role in the definition of the dynamics of the theory .
2 . another possibility would be to go back to the self - dual connection already at the classical level , and to find a way of dealing with the reality conditions .
this is what has been done in the three - dimensional model of @xcite and in the present paper . as we have seen ,
the reason for which this is possible in three - dimensions is that there is a natural notion of reality for the complex connection , which corresponds to considering its non - compact subgroup @xmath16 by means of the linear simplicity constraint .
then , the non - compactness can be handled due to the topological nature of the theory , and the resulting possibility of working on a fixed minimal graph .
3 . it seems to us that the most promising direction would be to consider @xmath1 as a regulator needed in order to deal with the non - compactness of the gauge group .
because the four - dimensional theory has local degrees of freedom , this non - compactness can not be handled like in the three - dimensional case by fixing a minimal graph .
the introduction of the barbero - immirzi parameter could therefore be seen as a generalized wick rotation , which has to be unfolded at the end of the day in order to obtain physical results .
this procedure of going through a compact gauge group in order to define the theory , and then using tools of analytic continuation to return to the non - compact case , was put forward in the computation of black hole entropy in four @xcite and three dimensions @xcite ( although in this last reference the regulator is the chern - simons level rather than the barbero - immirzi parameter ) .
if this last interpretation of the barbero - immirzi parameter turns out to be the correct one , then one should reconsider the results derived from the kinematical structure of the @xmath0 theory , since these rely heavily on the compactness of the gauge group and the value @xmath17 .
the key open question is in fact that of knowing at which step to perform the analytic continuation back to the self - dual value @xmath78 , if this can be given any well - defined meaning in the ( canonical or covariant ) full theory at all .
mg is supported by the nsf grant phy-1205388 and the eberly research funds of the pennsylvania state university .
i. agullo , j. f. barbero , j. diaz - polo , e. fernandez - borja and e. j. s. villaseor , black hole state counting in loop quantum gravity : a number theoretical approach , phys .
* 100 * 211301 ( 2008 ) , ` arxiv : gr - qc/0005126 ` . | we study and compare the spectra of geometric operators ( length and area ) in the quantum kinematics of two formulations of three - dimensional lorentzian loop quantum gravity . in the @xmath0 ashtekar - barbero framework ,
the spectra are discrete and depend on the barbero - immirzi parameter @xmath1 exactly like in the four - dimensional case .
however , we show that when working with the self - dual variables and imposing the reality conditions the spectra become continuous and @xmath1-independent . |
let @xmath8 be a sequence of independent and identically distributed random variables @xmath9 ya .
bernoulli proved @xcite that @xmath10 where @xmath11 @xmath12 @xmath13 and @xmath14 are arbitrary natural numbers . as far as @xmath15
we can make conclusion that @xmath16 therefore , we can formulate the law of large numbers @xmath17 where @xmath18 and @xmath19 are assumed to be fixed and @xmath20 markov @xcite , @xcite considered case of arbitrary @xmath21 and @xmath1 .
uspensky @xcite extended results of markov further and derived the first exponential bound @xmath22 with coefficients @xmath23 additional and more detailed historical notes may be found in the section `` existing inequalities '' @xcite . hoeffding @xcite developed methods of @xcite , @xcite and proved generally that @xmath24 note that similar exponential bounds for the empirical distribution functions were presented in @xcite and @xcite . in the section [ sec : bernoulli ]
we prove that the following values of the constants @xmath25 and @xmath26 may be used in the bound ( [ eq : uspen ] ) in the discreet case formulated and considered by ya .
it is demonstrated in the section [ ssec : optval ] that value of @xmath3 can not be smaller than one .
the following section [ ssec : oneside ] proves one - sided inequalities using methods and results of the theorem [ th : teorem ] .
section [ sec : discreet ] introduces a new structure of the exponential bound in the general discreet case with arbitrary rational @xmath0 and @xmath7 the best bound in asymptotical sense corresponds to the bigger value of the exponential coefficient .
however , as it is discussed in the section [ ssec : relbetbnds ] , the value of linear coefficient is also very important because of the practical reasons .
section [ sec : asymptot ] represents an extension of the methods and results of the section [ sec : discreet ] to the continuous case .
section [ ssec : normal ] considers particular application of the bounds to the normalized sum of random variables .
it is demonstrated that all 3 types of bounds are asymptotically equivalent . in this way ,
advantage of the propose bounds comparing with hoeffding s bound is absolute .
note that using results of @xcite and @xcite we can extend bounds for bernoulli random variables to the case of arbitrary bounded random variables . also , we note paper @xcite where similar exponential bounds were constructed for markov chains . generally , exponential bounds proved to be very effective in order to define required size of the sample in order to ensure proper quality of estimation , see , for example , @xcite .
we will use essentially different approach comparing with @xcite and @xcite .
this technique is based on the properties of convex ( concave ) functions applied to the binomial coefficients .
the following lemma formulates the core of the methods which are employed in the sections [ sec : bernoulli ] , [ sec : discreet ] and [ sec : asymptot ] .
[ lm : main ] suppose that @xmath27 is a function with continuous second derivative where @xmath28 is any natural number .
then , @xmath29 if @xmath30 ( convex case ) ; @xmath31 if @xmath32 ( concave case ) . _
proof : _ the following representations are valid @xmath33 \hspace{0.08 in } if \hspace{0.05 in } n=2 m ; \\
\sum_{j=0}^{m-1 } \left [ \varphi(1+j ) + \varphi(n - j ) \right ] + \varphi(m+1 ) \hspace{0.08 in } if \hspace{0.05 in } n=2m+1 , \end{array } \right.\ ] ] where @xmath34 is a natural number .
we obtain required bounds combining above equation with @xmath35 in the convex case , and with @xmath36 in the concave case .
@xmath37 the following notations will be used below @xmath38
@xmath39 @xmath40 where @xmath41 assuming that @xmath42 we form groups of binomial probabilities of the equal size @xmath43 @xmath44 @xmath45 then , we consider relations of the corresponding binomial coefficients from the neighbor groups @xmath46 where @xmath47 by definition @xmath48 note that @xmath49 if @xmath50 , and @xmath51 if @xmath52
we exclude from consideration the trivial cases : @xmath53 and @xmath54 , and assume that @xmath55 as it will be demonstrated below the task of estimation of @xmath56 is easier comparing with estimation of @xmath57 if @xmath58 . on the other hand ,
the task of estimation of @xmath57 is easier comparing with estimation of @xmath56 if @xmath59 .
as far as the problem is symmetrical , we assume that @xmath60 [ th : teorem ] suppose that @xmath13 and @xmath14 are arbitrary natural numbers
. then , @xmath61 where @xmath62 .
therefore , @xmath63 _ proof : _ by ( [ eq : step1 ] ) , @xmath64 is an increasing function of @xmath65 .
therefore , @xmath66 or @xmath67 where @xmath68 is a convex function of @xmath69 because @xmath70 or @xmath71 , and according to ( [ eq : convf ] ) @xmath72 @xmath73 where @xmath74 using general inequality @xmath75 we obtain @xmath76 @xmath77 furthermore , based on the following properties @xmath78>0 , \hspace{0.05 in } 2kr \geq n,\ ] ] we have @xmath79 then , using relations @xmath80 we transform above inequality to the required form @xmath81 above equation ( [ eq : bnd4 ] ) completes proof of the theorem .
[ th : teorem2 ] suppose that @xmath13 and @xmath14 are arbitrary natural numbers .
then , @xmath82 therefore , @xmath83 _ proof : _ by ( [ eq : step2 ] ) , the coefficient @xmath84 is an increasing function of @xmath65 .
therefore , @xmath85 or @xmath86 where @xmath87 the following condition @xmath88 give us a guarantee that the function @xmath89 is concave .
suppose that @xmath90 and @xmath91 or @xmath92 and @xmath93 then , by ( [ eq : concf ] ) we have @xmath94 @xmath95 @xmath96 if @xmath97 or @xmath98 thus , @xmath99 or @xmath100 the case @xmath101 is easy to consider : @xmath102 or @xmath103 above inequality is valid @xmath104 if @xmath105 .
now , we consider remaining case @xmath106 ( the case @xmath107 was considered already in the first part of the proof because the problem is symmetrical ) .
then , @xmath108 is convex as a function of @xmath69 , and , by ( [ eq : convf ] ) , @xmath109 where @xmath110 using inequality ( [ eq : gbound ] ) , we obtain @xmath111 or @xmath112 @xmath113 above inequality is valid for @xmath114 assuming that @xmath115 @xmath37 [ cor : corr ] the following upper bound is valid @xmath116 where @xmath13 and @xmath14 are arbitrary natural numbers . _
proof : _ it follows from the theorems [ th : teorem ] and [ th : teorem2 ] that @xmath117 the following representation is valid @xmath118 it easy to see that @xmath119 is an increasing function of @xmath120 where @xmath121 is a positive constant .
therefore , @xmath122
@xmath37 c ) function @xmath123 which is defined in ( [ eq : bnd63 ] ) ; bounds ( [ eq : bnd56 ] ) and ( [ eq : hff5 ] ) as a function of @xmath1 where b ) @xmath124 , d ) @xmath125 . ]
clearly , value of the linear coefficient @xmath126 in the exponential bound may not be regarded as an optimal .
the following proposition [ pr : prop ] demonstrates that @xmath127 [ pr : prop ] suppose that @xmath55 then , we can construct sequence of deviations @xmath128 so that @xmath129 and @xmath130 _ proof : _ by definition @xmath131 the following relation is valid @xmath132 by lemma [ lm : main ] , @xmath133 respectively , @xmath134 where @xmath135 figure [ fig : figure1](a ) illustrates graph of the function @xmath136 suppose that @xmath137
. then , @xmath138 and @xmath139 will be empty sets , and @xmath140 as far as @xmath141 at the same time , by definition ,
@xmath142 @xmath37 coefficient @xmath3 in the bound ( [ eq : uspen ] ) may not be smaller than 1 for the arbitrary @xmath143 and @xmath144 in this section we will use again the property that left groups of binomial probabilities may be estimated more effectively if @xmath145 analogously , right groups of binomial probabilities may be estimated more effectively if @xmath146 [ pr : prop5 ] suppose that @xmath147 where @xmath14 and @xmath148 are arbitrary natural numbers .
then , @xmath149 if @xmath150 ( @xmath151 ) ; @xmath152 if @xmath153 ( @xmath154 ) . _
proof : _ combining the equality @xmath155 and ( [ eq : bnd4 ] ) we derive that @xmath156 and @xmath157 next , we use the same bound ( [ eq : bnd4 ] ) in the opposite direction @xmath158 and @xmath159 again , we make an assumption @xmath70 or @xmath151 without loss of generality , and construct upper bound for the conditional probability @xmath160 @xmath161 where the last inequality may be regarded as a consequence of ( [ eq : one1 ] ) and ( [ eq : one2 ] ) .
then , we simplify ( [ eq : one3 ] ) using properties @xmath162 @xmath163 as a next step we prove that @xmath164 let us consider relations of the binomial probabilities which are symmetrical against central point @xmath165 @xmath166 @xmath167 where we used ( [ eq : convf ] ) . therefore , @xmath168 by the way , using identical method , one can prove @xmath169 the final upper bound follows from ( [ eq : one7 ] ) and ( [ eq : one8 ] ) applied to the bayesian formulae @xmath170
in this section we assume that @xmath0 and @xmath1 may be represented by positive rational numbers with denominator as a number of observations in the sample . respectively
, we will cover all possible empirical values of the sample mean as an estimator of the probability @xmath0 .
note that the role of the parameter @xmath43 will be different here comparing with previous section .
[ th : teorem3 ] suppose that @xmath172 and @xmath173 are arbitrary natural numbers with condition @xmath174 then , @xmath61 where @xmath62 .
therefore , @xmath175 _ proof : _ the following upper bound may be obtained similarly to ( [ eq : bnd2 ] ) @xmath176 where @xmath177 and @xmath178 ( else @xmath179 ) .
it follows from ( [ eq : bnd21 ] ) that @xmath180 where @xmath181 by ( [ eq : gbound ] ) , we obtain @xmath182 or @xmath183 + 2 \gamma \varepsilon ( 1-p).\ ] ] above inequality is valid @xmath184 and @xmath185 if @xmath186 @xmath37 the following theorem [ th : teorem4 ] introduces a new structure of the exponential bound as a main result of this paper .
[ th : teorem4 ] suppose that @xmath172 and @xmath173 are arbitrary natural numbers with conditions @xmath187 then , @xmath188 therefore , @xmath189 ( the case @xmath190 is covered by the theorem [ th : teorem2 ] with stronger result ) . _ proof : _ similar to ( [ eq : bnd6a ] ) and assuming that @xmath191 we derive @xmath192 where @xmath193 ( else @xmath194 ) .
the following relations are valid @xmath195 where @xmath196 by ( [ eq : gbound ] ) , we obtain @xmath197 or @xmath198\ ] ] or @xmath199= \gamma k
\left [ 2p(1-p ) + \varepsilon ( 2p-1 ) \right ] + \gamma \varepsilon p.\ ] ] above inequality take place if @xmath200 let us consider remaining case @xmath201 similar to ( [ eq : bnd2 ] ) we have @xmath202 where @xmath203 by ( [ eq : gbound ] ) , we obtain @xmath204 or @xmath205 above inequality take place if @xmath206 note that condition ( [ eq : bnd33 ] ) is less restrictive comparing with ( [ eq : bnd32 ] ) .
@xmath37 the following upper bound is valid @xmath207 where @xmath208 and @xmath173 are arbitrary natural numbers with conditions @xmath209 proof is similar to the corollary [ cor : corr ] : we obtain required value of the parameter @xmath4 as @xmath210 where first value follows from theorem [ th : teorem3 ] , second and third values follow from ( [ eq : bnd32 ] ) and ( [ eq : bnd33 ] ) .
the bound ( [ eq : bnd56 ] ) improves hoeffding s bound @xmath211 if @xmath212 therefore , @xmath213 figure [ fig : figure1](c ) illustrates behavior of the function @xmath214 .
.the following parameters were used in this example : @xmath215 second column represents real probability of absolute deviation of the sample mean from @xmath216 next two columns represent corresponding bound ( [ eq : bnd56 ] ) and hoeffding s bound ( [ eq : hff5 ] ) . [ cols= " < , < ,
< , < " , ] let us denote by @xmath217 distribution function of the normalized sum of bernoulli random variables @xmath218 according to the central limit theorem , @xmath219 where @xmath220 is a standard normal distribution function . respectively
, it appears to be reasonable to use random variable ( [ eq : norm ] ) as a test case in order to compare different bounds .
[ pr : prop3 ] the bounds ( [ eq : bnd1 ] ) , ( [ eq : bnd56 ] ) and ( [ eq : funcbound ] ) , as an upper bounds for the following probability @xmath221 are equivalent asymptotically and equal to @xmath222 where @xmath223 and @xmath224 _ proof : _ we have @xmath225 where @xmath226 . then , we insert @xmath227 into ( [ eq : bnd1 ] ) , ( [ eq : bnd56 ] ) or ( [ eq : funcbound ] ) , and take @xmath228
if @xmath229 as a result , we obtain required formulae ( [ eq : clt7 ] ) .
@xmath37 using hoeffding s bound we will obtain the same asymptotical structure ( [ eq : clt7 ] ) , but with @xmath230
the proposition [ pr : prop ] proves that value of the linear coefficient @xmath25 can not be improved .
taking this fact as a base point we established a new structure of the exponential bound ( [ eq : bnd56 ] ) .
figures [ fig : figure1](a),(c ) , ( d ) and table [ tb : table1 ] demonstrate advantages of the bound ( [ eq : bnd56 ] ) against hoeffding s bound if value of @xmath1 is small enough .
the above section [ ssec : normal ] demonstrates additional arguments in support of the proposed bounds , and these arguments cover not to only bernoulli random variables .
the area of applications of the proposed methods may be extended further , for example , we can consider arbitrary bounded random variables or uniform metric for the empirical distributions .
the paper represents a fresh look at the ideas and methods which ja .
bernoulli proposed in the 17th century . as it was demonstrated in the section [ sec : bernoulli ] the probability of deviation in the classical bernoulli case may be bounded using standard structure of the exponential bound with optimal linear coefficient @xmath25 .
it was the first step of this research which was completed in 1987 shortly after the 1st world congress of the bernoulli society where the author purchased book @xcite .
a. dvoretzky , j. kiefer and j. wolfowitz .
asymptotic minimax character of the sample distribution function and of the classical multinomial estimator . "
_ annals of mathematical statistics _ , vol .
642 - 669 , 1956 .
f. yeh and m. gallagher . an empirical study of hoeffding racing for model selection in k - nearest neighbor classification . " _
ideal 2005 , lncs 3578 , springer - verlag , m. gallagher and j. hogan and f. maire _ ,
220 - 227 , 2005 . | we consider upper exponential bounds for the probability of the event that an absolute deviation of sample mean from mathematical expectation @xmath0 is bigger comparing with some ordered level @xmath1 .
these bounds include 2 coefficients @xmath2 . in order to optimize the bound
we are interested to minimize linear coefficient @xmath3 and to maximize exponential coefficient @xmath4 . generally , the value of linear coefficient @xmath3 may not be smaller than one .
the following 2 settings were proved : 1 ) @xmath5 in the case of classical discreet problem as it was formulated by bernoulli in the 17th century , and 2 ) @xmath6 in the general discreet case with arbitrary rational @xmath0 and @xmath7 the second setting represents a new structure of the exponential bound which may be extended to continuous case . |
considering the fragmentation process @xmath5 we define the collins function , which depends on the longitudinal momentum fraction @xmath6 of the pion and the transverse momentum @xmath7 of the quark , as @xcite @xmath8 \big |_{k^-=p^-/z } \;,\ ] ] with the correlation function @xmath9 + to get an estimate of the collins function at a low energy scale we make use of the chiral invariant model of manohar and georgi @xcite .
the model contains massive constituent quarks and goldstone bosons as effective degrees of freedom .
the lagrangian of the model reads @xmath10 the manohar - georgi model describes the valence part of the normal unpolarized fragmentation function @xmath11 fairly well @xcite . +
while at tree level the collins function is zero , the situation changes if rescattering corrections are included in @xmath12 . in a consistent one - loop calculation
the diagrams in fig .
[ f : loop ] contribute to @xmath2 . because of the presence of nonvanishing imaginary parts in these diagrams , we obtain a nonvanishing collins function .
+ the @xmath6 behaviour of asymmetries containing @xmath2 is typically governed by the ratio @xmath13 in our model , it turns out that this ratio is approximately equal to @xmath14 for convenience of use , the result of our model can be roughly parametrized by means of a simple analytic form , i.e.@xmath15 fig .
[ f : ratio ] shows the result of our model together with the parametrization we suggested .
the ratio @xmath16 is clearly increasing with increasing @xmath6 ; this feature is largely independent of the parameter choice in our approach @xcite .
in semi - inclusive dis , three important single spin and azimuthal asymmetries containing the collins function exist . neglecting quark mass terms one has @xcite @xmath17 , \hphantom{m } \label{e : ul } \\
\langle \sin \phi_h \rangle_{lu } & \propto & \frac{1}{q } \ ; e(x ) \ , h_1^{\perp ( 1/2)}(z ) \ , , \label{e : lu } \\ \langle \sin ( \phi_h + \phi_s )
\rangle_{ut } & \propto & h_1(x ) \ , h_1^{\perp ( 1/2)}(z ) \ , .
\label{e : ut}\end{aligned}\ ] ] where only the asymmetry numerators are written down . @xmath18
( @xmath19 ) is the azimuthal angle of the produced hadron ( target spin ) ; the subscript @xmath20 , e.g. , indicates an unpolarized beam and a longitudinally polarized target ; @xmath21 denotes the transversity distribution and @xmath22,@xmath23 denote two twist-3 distributions ; @xmath24 and @xmath25 are kinematical factors .
+ from the asymmetry ( [ e : ul ] ) it is difficul to draw any conclusion on the collins function
. however , if we assume that the _ other terms _ in ( [ e : ul ] ) are small , then the @xmath6 dependence of the asymmetry should be almost entirely due to the collins function . in fig .
[ f : fit ] we compare the parametrization ( [ e : param ] ) of our results with hermes data on @xmath26 @xcite and preliminary data on the same asymmetry from clas @xcite .
the agreement of the @xmath6 shape is remarkable .
note that we arbitrarily normalized our curve to take into account the unknown distribution functions and prefactors .
+ the transverse spin asymmetry in eq .
( [ e : ut ] ) is one of the most promising observables to measure @xmath21 .
in fact , it is one of the ambitions of the hermes collaboration to extract the transversity distribution from this observable .
however , without any knowledge on the collins function it is impossible to evaluate the magnitude of the asymmetry . using our model and two extreme assumptions on the transversity distribution , we obtain values up to about @xmath3 for the asymmetry ( left panel of fig . [
f : asymm ] ) , which therefore should be observable .
our results support the intention of extracting the transversity in this way .
+ at present , data on semi - inclusive dis allow us to extract in an assumption - free way only the shape of @xmath2
. the azimuthal asymmetry @xcite @xmath27 accessible in @xmath0 annihilation into two hadrons , can provide information on the magnitude of @xmath2 as well . as shown in the right panel of fig .
[ f : asymm ] , our result for this asymmetry is of the order of @xmath4 .
possible accurate measurements of this observable at babar or belle would be very useful to better pin down the collins function from the experimental side . | we estimate the collins function at a low energy scale by calculating the fragmentation of a quark into a pion at the one - loop level in the chiral invariant model of manohar and georgi . we give a useful parametrization of our results and
we briefly discuss different spin and/or azimuthal asymmetries containing the collins function and measurable in semi - inclusive dis and @xmath0 annihilation .
the collins function @xmath1 , describing the fragmentation of a transversely polarized quark into an unpolarized hadron @xcite , plays an important role in studies of the nucleon spin structure .
although @xmath2 is t - odd , it can be nonzero due to final state interactions .
in fact , we were able to show for the first time that a nonvanishing collins function can be obtained in a field theoretical approach through a consistent one - loop calculation of the fragmentation process @xcite .
+ to obtain a reasonable estimate of the collins function , we calculate it in a chiral invariant approach at a low energy scale @xcite .
we use the model of manohar and georgi @xcite , which incorporates chiral symmetry and its spontaneous breaking , two important aspects of qcd at low energies .
+ the collins function enters several spin and azimuthal asymmetries in one - particle inclusive dis .
of particular interest is the transverse single spin asymmetry , where @xmath2 appears in connection with the transversity distribution , for which we predict effects of the order of @xmath3 . in principle in semi - inclusive dis only the shape of the collins function can be studied , while a purely azimuthal asymmetry in @xmath0 annihilation allows one to measure also its magnitude . for this asymmetry
we obtain effects of the order of @xmath4 . |
for convenience , we record some known results that provide hamiltonian cycles in various cayley graphs , after fixing some notation . for any group @xmath0 , we use : 1 .
@xmath7 to denote the _ commutator subgroup _
@xmath8 $ ] of @xmath0 , 2 .
@xmath9 to denote the _ center _ of @xmath0 , and 3 .
@xmath10 to denote the _ frattini subgroup _ of @xmath0 . for @xmath11
, we use @xmath12 to denote the _ conjugate _ @xmath13 .
if @xmath14 is any sequence , we use @xmath15 to denote the sequence @xmath16 that is obtained by deleting the last term .
[ keatingwitte ] if @xmath7 is a cyclic group of prime - power order , then every connected cayley graph on @xmath0 has a hamiltonian cycle .
[ normaleasy ] let @xmath3 generate the finite group @xmath0 , and let @xmath17 .
if * @xmath18 , * @xmath19 has a hamiltonian cycle , and * either 1 .
[ normaleasy - z ] @xmath20 , or 2 .
[ normaleasy - p ] @xmath21 is prime , then @xmath4 has a hamiltonian cycle . [ cyclicnormal2p ]
let @xmath3 generate the finite group @xmath0 , and let @xmath17 .
if * @xmath18 , * @xmath21 is a divisor of @xmath22 , where @xmath2 and @xmath23 are distinct primes , * @xmath24 , * @xmath25 is divisible by @xmath23 , and * @xmath19 has a hamiltonian cycle , then there is a hamiltonian cycle in @xmath4 .
the following results are well known ( and easy to prove ) : [ fgl ] suppose * @xmath3 is a generating set of @xmath0 , * @xmath26 is a cyclic , normal subgroup of @xmath0 , * @xmath27 is a hamiltonian cycle in @xmath28 , and * the product @xmath29 generates @xmath26 .
then @xmath30 is a hamiltonian cycle in @xmath4 .
[ doubleedge ] suppose * @xmath3 is a generating set of @xmath0 , * @xmath26 is a normal subgroup of @xmath0 , such that @xmath31 is prime , * @xmath32 for some @xmath33 with @xmath34 , and * there is a hamiltonian cycle in @xmath28 that uses at least one edge labelled @xmath35 .
then there is a hamiltonian cycle in @xmath4 .
if @xmath36 is any subgroup of @xmath0 , then @xmath37 denotes the multigraph in which : * the vertices are the right cosets of @xmath36 , and * there is an edge joining @xmath38 and @xmath39 for each @xmath40 , such that @xmath41 .
thus , if there are two different elements @xmath42 and @xmath43 of @xmath44 , such that @xmath45 and @xmath46 are both in @xmath47 , then the vertices @xmath38 and @xmath39 are joined by a double edge .
[ multidouble ] suppose * @xmath3 is a generating set of @xmath0 , * @xmath36 is a subgroup of @xmath0 , such that @xmath48 is prime , * the quotient multigraph @xmath37 has a hamiltonian cycle @xmath49 , and * @xmath49 uses some double - edge of @xmath37 .
then there is a hamiltonian cycle in @xmath4 .
[ pksubgrp ] suppose * @xmath3 is a generating set of @xmath0 , * @xmath26 is a normal @xmath2-subgroup of @xmath0 , and * @xmath50 , for all @xmath51 .
then @xmath4 has a hamiltonian cycle .
[ > 3 ] in the proof of our main result , we may assume @xmath52 , for otherwise either : * @xmath53 is of the form @xmath54 , where @xmath23 is prime , so ( * ? ? ?
9.1 ) applies , or * @xmath55 is a prime power , so the main theorem of @xcite applies .
let : * @xmath0 be a group of order @xmath56 , where @xmath2 is prime , and @xmath52 ( see [ > 3 ] ) , * @xmath3 be a minimal generating set for @xmath0 , * @xmath57 be a sylow @xmath2-subgroup of @xmath0 , * @xmath58 be a generator of @xmath59 , and * @xmath60 be a sylow @xmath61-subgroup of @xmath0 in this section , we assume that @xmath59 is a normal subgroup of @xmath0 .
therefore @xmath0 is a semidirect product : @xmath62 we may assume @xmath7 is not cyclic of prime order ( for otherwise [ keatingwitte ] applies ) .
this implies @xmath60 is nonabelian , and acts nontrivially on @xmath59 , so @xmath63 since @xmath60 is a @xmath61-group and acts nontrivially on @xmath64 , we must have @xmath65 .
thus , we may choose @xmath66 , such that @xmath67 dividing @xmath68 by @xmath69 , we see that @xmath70 there are only @xmath71 nonabelian groups of order @xmath72 , and we will consider them as separate cases , but , first , we cover some common ground . since @xmath60 is a nonabelian group of order @xmath72 , and @xmath73 ,
it is easy to see that @xmath74 [ s=2choices ] assume * @xmath75 , such that @xmath35 does not centralize @xmath59 , and * @xmath76 .
then we may assume @xmath3 is either @xmath77 or @xmath78 or @xmath79 or @xmath80 .
since @xmath81 is a @xmath71-generated group of prime - power order , there must be an element @xmath82 of @xmath3 , such that @xmath83 generates @xmath84 .
we may write @xmath85 note that : * by replacing @xmath82 with its inverse if necessary , we may assume @xmath86 .
* by applying an automorphism of @xmath0 that fixes @xmath35 and maps @xmath87 to @xmath88 , we may assume @xmath89 is trivial ( since @xmath90 ) .
* by replacing @xmath58 with @xmath91 if @xmath92 , we may assume @xmath93 .
thus , @xmath94 assume @xmath95 .
then @xmath96 , so @xmath97 .
this yields the four listed generating sets .
. then @xmath99 , and there must be a third element @xmath100 of @xmath3 , with @xmath101 ; after replacing @xmath58 with an appropriate power , we may write @xmath102 with @xmath103
. we must have @xmath104 , for otherwise @xmath105 ( which contradicts the minimality of @xmath3 ) .
therefore @xmath106 we may assume : * @xmath107 , for otherwise @xmath108 , so [ cyclicnormal2p ] applies .
* @xmath109 , by replacing @xmath100 with its inverse if necessary .
* @xmath110 , for otherwise @xmath35 and @xmath100 provide a double edge in @xmath111 , so [ doubleedge ] applies
. then @xmath112 generates @xmath113 .
consider the hamiltonian cycles @xmath114 in @xmath115 .
letting @xmath116 , we see that their endpoints in @xmath0 are ( respectively ) : @xmath117 the final two endpoints both have a nontrivial projection to @xmath59 ( since @xmath35 , being a @xmath61-element , can not invert @xmath58 ) , and at least one of these two endpoints also has a nontrivial projection to @xmath9 .
such an endpoint generates @xmath118 , so [ fgl ] provides a hamiltonian cycle in @xmath4 .
[ exp3s ] assume @xmath60 is of exponent @xmath61 , so @xmath119 = z , \ [ x , z ] = [ y , z ] = e \,\rangle .\ ] ] then we may assume : 1 .
[ exp3s - g ] @xmath120 , but @xmath121 and @xmath89 centralize @xmath59 , and 2 .
[ exp3s - s ] either : 1 .
@xmath122 , or 2 .
@xmath123 .
( [ exp3s - g ] ) since @xmath60 acts nontrivially on @xmath59 , and @xmath124 is cyclic , but @xmath125 is not cyclic , there must be elements @xmath82 and @xmath100 of @xmath126 , such that @xmath82 centralizes @xmath59 , but @xmath100 does not .
( and @xmath89 must centralize @xmath59 , because it is in @xmath127 . ) by applying an automorphism of @xmath60 , we may assume @xmath128 and @xmath129 .
furthermore , we may assume @xmath120 by replacing @xmath130 with its inverse if necessary .
( [ exp3s - s ] ) @xmath3 must contain an element that does not centralize @xmath59 , so we may assume @xmath131 . by applying [ s=2choices ] with @xmath132 and @xmath133 , we see that we may assume @xmath3 is : @xmath134 but there is an automorphism of @xmath0 that fixes @xmath130 and @xmath58 , and sends @xmath121 to @xmath135 , so we need only consider @xmath71 of these possibilities .
assume , as usual , that @xmath1 , where @xmath2 is prime , and that @xmath0 has a normal sylow @xmath2-subgroup .
if the sylow @xmath61-subgroup @xmath60 is of exponent @xmath61 , then @xmath4 has a hamiltonian cycle .
we write @xmath136 for the natural homomorphism from @xmath0 to @xmath137 . from [ exp3s]([exp3s - s ] ) , we see that we need only consider two possibilities for @xmath3 .
assume @xmath122 . for @xmath138 and @xmath139
, we have the following hamiltonian cycle in @xmath140 : @xmath141 its endpoint in @xmath0 is @xmath142 since the walk is a hamiltonian cycle in @xmath84 , we know that this endpoint is in @xmath143 .
so all terms except powers of @xmath58 must cancel .
thus , we need only calculate the contribution from each appearance of @xmath58 in this expression . to do this , note that if a term @xmath144 is followed by a net total of @xmath145 appearances of @xmath130 , then the term contributes a factor of @xmath146 to the product .
so the endpoint in @xmath0 is : @xmath147 since @xmath148 , this simplifies to @xmath149 since @xmath150 , this endpoint generates @xmath59 , so [ fgl ] provides a hamiltonian cycle in @xmath4 .
assume @xmath123 . for @xmath138 and @xmath151
, we have the hamiltonian cycle @xmath152 in @xmath140 .
its endpoint in @xmath0 is @xmath153 since we are free to choose @xmath154 to be either of the two primitive cube roots of @xmath155 in @xmath156 , and the equation @xmath157 has only one solution in @xmath156 , we may assume @xmath154 has been selected to make the exponent nonzero .
then [ fgl ] provides a hamiltonian cycle in @xmath4 .
[ gensexp9 ] assume @xmath60 is of exponent @xmath158 , so @xmath159 = x^3 \,\rangle .\ ] ] there are two possibilities for @xmath0 , depending on whether @xmath160 contains an element of order @xmath158 or not . 1 .
[ exp9-no9 ] assume @xmath160 does not contain an element of order @xmath158 .
then we may assume @xmath121 centralizes @xmath59 , but @xmath120 .
furthermore , we may assume : 1 .
[ exp9-no9-x+yw ] @xmath161 , or 2 .
[ exp9-no9-x+xyw ] @xmath123 .
[ exp9 - 9cent ] assume @xmath160 contains an element of order @xmath158 .
then we may assume @xmath130 centralizes @xmath59 , but @xmath162 .
furthermore , we may assume : 1 .
[ exp9 - 9cent - xw+y ] @xmath163 , 2 .
[ exp9 - 9cent - xyw+y ] @xmath164 , 3 .
[ exp9 - 9cent - xy+xw ] @xmath165 , or 4 .
[ exp9 - 9cent - xy+x2yw ] @xmath166 .
( [ exp9-no9 ] ) since @xmath130 has order @xmath158 , we know that it does not centralize @xmath59 .
but @xmath167 must centralize @xmath59 ( since @xmath167 is in @xmath7 ) .
therefore , we may assume @xmath168 ( by replacing @xmath130 with its inverse if necessary ) . also , since @xmath169 must be cyclic ( because @xmath124 is cyclic ) , but @xmath170 does not contain an element of order @xmath158 , we see that @xmath160 contains every element of order @xmath61 , so @xmath121 must be in @xmath160 . since @xmath3 must contain an element that does not centralize @xmath59 , we may assume @xmath131 . by applying [ s=2choices ] with @xmath132 and @xmath133 , we see that we may assume @xmath3 is : @xmath134 the second generating set need not be considered , because @xmath171 , so it is equivalent to the first .
also , the fourth generating set can be converted into the third , since there is an automorphism of @xmath0 that fixes @xmath121 , but takes @xmath130 to @xmath172 and @xmath58 to @xmath173 .
( [ exp9 - 9cent ] ) we may assume @xmath174 , so @xmath175 .
we know that @xmath3 must contain an element @xmath35 that does not centralize @xmath59 , and there are two possibilities : either 1 .
[ gensexp9pf-9cent-3 ] @xmath35 has order @xmath61 , or 2 .
[ gensexp9pf-9cent-9 ] @xmath35 has order @xmath158 .
we consider these two possibilities as separate cases .
assume @xmath35 has order @xmath61 .
we may assume @xmath176 .
letting @xmath177 , we see from [ s=2choices ] that we may assume @xmath3 is either @xmath178 the second and fourth generating sets need not be considered , because there is an automorphism of @xmath0 that fixes @xmath121 and @xmath58 , but takes @xmath130 to @xmath179 . also , the third generating set may be replaced with @xmath180 , since there is an automorphism of @xmath0 that fixes @xmath121 and @xmath58 , but takes @xmath130 to @xmath181 .
assume @xmath35 has order @xmath158 .
we may assume @xmath182 .
letting @xmath177 , we see from [ s=2choices ] that we may assume @xmath3 is either @xmath183 the second generating set is equivalent to @xmath184 , since the automorphism of @xmath0 that sends @xmath130 to @xmath185 , @xmath121 to @xmath186 , and @xmath58 to @xmath173 maps it to @xmath187 .
the third generating set is mapped to @xmath188 by the automorphism that sends @xmath130 to @xmath189 $ ] and @xmath121 to @xmath190^{-1}y$ ] .
the fourth generating set need not be considered , because @xmath191 is an element of order @xmath61 that does not centralize @xmath59 , which puts it in the previous case .
assume , as usual , that @xmath1 , where @xmath2 is prime , and that @xmath0 has a normal sylow @xmath2-subgroup .
if the sylow @xmath61-subgroup @xmath60 is of exponent @xmath158 , then @xmath4 has a hamiltonian cycle .
we will show that , for an appropriate choice of @xmath82 and @xmath100 in @xmath44 , the walk @xmath192 provides a hamiltonian cycle in @xmath140 whose endpoint in @xmath0 generates @xmath59 ( so [ fgl ] applies ) .
we begin by verifying two situations in which ( [ hc ] ) is a hamiltonian cycle : 1 .
[ hc3 ] if @xmath193 , @xmath194 , and @xmath195 in @xmath196 , then we have the hamiltonian cycle @xmath197 2 .
[ hc9 ] if @xmath193 , @xmath198 , @xmath199 , and @xmath200 in @xmath196 , then we have the hamiltonian cycle @xmath201 to calculate the endpoint in @xmath0 , fix @xmath202 , with @xmath203 and write @xmath204 note that if an occurrence of @xmath205 in the product is followed by a net total of @xmath206 appearances of @xmath207 and a net total of @xmath208 appearances of @xmath209 , then it contributes a factor of @xmath210 to the product .
( a similar occurrence of @xmath211 contributes a factor of @xmath212 to the product . ) furthermore , since @xmath213 , there is no harm in reducing @xmath206 and @xmath208 modulo @xmath61 .
we will apply these considerations only in a few particular situations : 1 .
[ endpt - b = w ] assume @xmath214 ( so @xmath215 and @xmath216 ) .
then the endpoint of the path in @xmath0 is @xmath217 by the above considerations , this simplifies to @xmath218 , where @xmath219 note that : 1 .
[ endpt - b = w - r2=1 ] if @xmath220 and @xmath221 , then @xmath222 simplifies to @xmath223 , because @xmath224 in this case .
[ endpt - b = w - not1 ] if @xmath220 and @xmath225 , then @xmath222 simplifies to @xmath226 , because @xmath227 in this case .
[ endpt - a = w ] assume @xmath228 ( so @xmath229 and @xmath230 ) .
then the endpoint of the path in @xmath0 is @xmath231 by the above considerations , this simplifies to @xmath232 , where @xmath233 note that : 1 .
[ endpt - a = w - r1=1 ] if @xmath234 and @xmath225 , then @xmath222 simplifies to @xmath235 , because @xmath236 in this case .
[ endpt - a = w - not1 ] if @xmath220 and @xmath225 , then @xmath222 simplifies to @xmath237 , because @xmath227 in this case .
now we provide a hamiltonian cycle for each of the generating sets listed in [ gensexp9 ] : 1 .
if @xmath160 has exponent @xmath61 , and @xmath122 , we let @xmath138 and @xmath139 in ( [ hc3 ] ) . in this case , we have @xmath214 , @xmath238 , and @xmath221 , so ( [ endpt - b = w - r2=1 ] ) tells us that the endpoint in @xmath0 is @xmath239 .
if @xmath160 has exponent @xmath61 , and @xmath123 , we let @xmath138 and @xmath240 in ( [ hc9 ] ) . in this case
, we have @xmath214 , @xmath238 and @xmath241 , so ( [ endpt - b = w - not1 ] ) tells us that the endpoint in @xmath0 is @xmath218 , where @xmath242 3 . if @xmath160 has exponent @xmath158 , and @xmath243 , we let @xmath244 and @xmath245 in ( [ hc3 ] ) . in this case , we have @xmath228 , @xmath234 and @xmath246 , so ( [ endpt - a = w - r1=1 ] ) tells us that the endpoint in @xmath0 is @xmath247 .
if @xmath160 has exponent @xmath158 , and @xmath164 , we let @xmath248 and @xmath245 in ( [ hc3 ] ) . in this case
, we have @xmath228 and @xmath249 , so ( [ endpt - a = w - not1 ] ) tells us that the endpoint in @xmath0 is @xmath218 , where @xmath250 5 .
if @xmath160 has exponent @xmath158 , and @xmath165 , we let @xmath244 and @xmath251 in ( [ hc9 ] ) . in this case , we have @xmath228 , @xmath234 , and @xmath241 , so ( [ endpt - a = w - r1=1 ] ) tells us that the endpoint in @xmath0 is @xmath232 , where @xmath252 6 .
if @xmath160 has exponent @xmath158 , and @xmath166 , we let @xmath253 and @xmath254 in ( [ hc9 ] ) . in this case
, we have @xmath214 and @xmath249 , so ( [ endpt - b = w - not1 ] ) tells us that the endpoint in @xmath0 is @xmath218 , where @xmath255 in all cases , there is at most one nonzero value of @xmath154 ( modulo @xmath2 ) for which the exponent of @xmath205 is @xmath256 .
since we are free to choose @xmath154 to be either of the two primitive cube roots of @xmath155 in @xmath156 , we may assume @xmath154 has been selected to make the exponent nonzero .
then [ fgl ] provides a hamiltonian cycle in @xmath4 .
[ pnotnormalis13 ] assume then @xmath257 , and @xmath258 , where a generator @xmath58 of @xmath259 acts on @xmath260 via multiplication on the right by the matrix @xmath261 furthermore , we may assume @xmath262 where @xmath263 , and @xmath264 let @xmath59 be a sylow @xmath2-subgroup of @xmath0 , and @xmath60 be a sylow @xmath61-subgroup of @xmath0 .
since no odd prime divides @xmath265 or @xmath266 , and @xmath267 is the only odd prime that divides @xmath268 , sylow s theorem ( * ? ? ?
15.7 , p. 230 ) implies that @xmath257 , and that @xmath269 , so @xmath0 must have a normal @xmath2-complement ( * ? ? ?
7.4.3 ) ; i.e. , @xmath270 . since @xmath59 must act nontrivially on @xmath60 ( since @xmath59 is not normal ) , we know that it must act nontrivially on @xmath125 ( * ? ? ?
* thm . 5.3.5 , p. 180 ) .
however , @xmath59 can not act nontrivially on an elementary abelian group of order @xmath61 or @xmath271 , because @xmath272 is not a divisor of @xmath265 or @xmath266 . therefore , we must have @xmath273 , so @xmath60 must be elementary abelian ( and the action of @xmath59 is irreducible ) .
let @xmath274 be the matrix representing the action of @xmath58 on @xmath275 ( with respect to some basis that will be specified later ) . in the polynomial ring
@xmath276 $ ] , we have the factorization : @xmath277 since @xmath278 , the minimal polynomial of @xmath274 must be one of the factors on the right - hand side . by replacing @xmath58 with an appropriate power , we may assume it is the first factor . then , choosing any nonzero @xmath279 , the matrix representation of @xmath58 with respect to the basis @xmath280 is @xmath274 ( the rational canonical form ) .
now , let @xmath281 be a primitive @xmath267th root of unity in the finite field @xmath282 .
then any galois automorphism of @xmath282 over @xmath283 must raise @xmath281 to a power .
since the subgroup of order @xmath61 in @xmath284 is generated by the number @xmath61 , we conclude that the orbit of @xmath281 under the galois group is @xmath285
. these must be the @xmath61 roots of one of the irreducible factors on the right - hand side of ( [ factor(x13 - 1 ) ] ) .
thus , for any @xmath286 , the matrices @xmath287 , @xmath288 , and @xmath289 all have the same minimal polynomial , so they are conjugate under @xmath290 .
that is : @xmath291 there is an element @xmath82 of @xmath3 that generates @xmath292 .
then @xmath82 has order @xmath2 , so , replacing it by a conjugate , we may assume @xmath293 , so @xmath294 for some @xmath295 . from ( [ minpolysofz13xq ] ) , we see that we may assume @xmath296 ( perhaps after replacing @xmath82 by its inverse ) .
now let @xmath100 be the second element of @xmath3 , so we may assume @xmath297 for some @xmath145 .
we may assume @xmath298 ( by replacing @xmath100 with its inverse , if necessary ) .
we may also assume @xmath299 , for otherwise @xmath300 , so [ pksubgrp ] applies .
note that , since @xmath308 is conjugate to @xmath274 under @xmath290 ( since they are in the same row of ( [ minpolysofz13xq ] ) ) , we know that the pair @xmath309 is isomorphic to the pair @xmath310 . by replacing @xmath100 with its inverse , and then interchanging @xmath82 and @xmath100 , this is transformed to @xmath311 .
so we may assume @xmath312 .
we may assume @xmath314 is in the second or fourth row of the table ( for otherwise we could interchange @xmath82 with @xmath100 to enter the previous case .
so @xmath315 .
since @xmath299 , this implies @xmath316 .
however , since @xmath317 is conjugate to @xmath318 ( since they are in the same row of ( [ minpolysofz13xq ] ) ) , and we have @xmath319 and @xmath320 , we see that the pair @xmath321 is isomorphic to @xmath322 .
so we may assume @xmath323 . from [ pnotnormalis13 ] ( and [ > 3 ] )
, we may assume @xmath258 .
for each of the generating sets listed in [ pnotnormalis13 ] , we provide an explicit hamiltonian cycle in the quotient multigraph @xmath324 that uses at least one double edge .
so [ multidouble ] applies .
k.kutnar , d.marui , j.morris , d.w.morris , and p.parl : hamiltonian cycles in cayley graphs whose order has few prime factors , _ ars mathematica contemporanea _
( to appear ) .
http://arxiv.org/abs/1009.5795 [ ] | suppose @xmath0 is a finite group , such that @xmath1 , where @xmath2 is prime .
we show that if @xmath3 is any generating set of @xmath0 , then there is a hamiltonian cycle in the corresponding cayley graph @xmath4 .
[ 27p ] if @xmath5 , where @xmath2 is prime , then every connected cayley graph on @xmath0 has a hamiltonian cycle . combining this with results in @xcite
establishes that : @xmath6 the remainder of the paper provides a proof of the theorem .
here is an outline : |
while the physical origins of the surface properties of stiff solids are well understood , a systematic investigation of the surface properties of soft solids remains to be undertaken .
the aims are not only to reveal the appropriate continuum - scale description of the mechanical properties of the interface , but also to understand their origins to allow for effective manipulation . _ key question : what is the appropriate description of the mechanics of a soft solid interface ? _ for a simple liquid , the energy gained from stretching the liquid - vapour interface is equal to the product of the surface energy @xmath0 @xmath1 $ ] and the gain of interfacial area .
moreover , the surface stress , @xmath2 @xmath3 $ ] , is equal to the surface energy , and both are often referred to as the surface tension . for an elastic solid ,
the stress state of the interface is quite different .
as above , stretching the interface increases the energy in proportion to the surface energy .
however , a second contribution arises when the surface energy is strain - dependent , leading to the so - called _ shuttleworth effect _ @xcite . in this case , the surface stress @xmath2 not only depends upon the surface energy , @xmath0 , but also its derivatives with strain , as @xmath4 .
in contrast to liquids , the stress state of solids can be both anisotropic and strain - dependent . for stiff solid materials ,
some of the implications of this have been worked out @xcite . for soft solids ,
the shuttleworth effect is predicted to qualitatively change wetting phenomena , as discussed below
. an alternative perspective on the state of stress at surfaces comes from the study of complex fluid interfaces .
when a fluid interface is laden with surfactants or particles , it can develop a stress state that is anisotropic , strain - dependent , and rate - dependent @xcite .
this surface rheology can have a significant impact on the bulk flow of fluids @xcite . to understand the surface mechanical properties of soft solids
, it will be essential to integrate insights from traditional solid mechanics ( incorporating shuttleworth s results ) with emerging insights from the surface rheology of complex fluid interfaces .
whatever the results , new experimental approaches are needed to quantify the surface stresses of soft solids .
_ key questions : what is the origin of the surface mechanical properties
? how can they be manipulated ? _
polymeric materials have been the dominant choice for experiments in the mechanics of soft solids .
these materials feature a system - spanning cross - linked polymer network ; this network may be swollen either by un - crosslinked polymer , or by a chemically - distinct solvent .
this general description encompasses a broad class of soft materials , including polymeric elastomers and gels , with a wide range of properties arising from the particular chemistry . in all cases ,
the surface properties will be impacted by the structure of the polymer network near the interface . for gels
, the solvent and additional surface - active components may significantly impact the surface properties .
we can learn a lot about possible contributions of the polymer network from our emerging understanding of a closely - related class of materials : polymer melts .
polymer melts are composed of long entangled polymer chains , without cross - linkers , and are known to have significantly - modified mechanical properties near their surfaces .
the presence of interfaces and confinement may alter the network of entanglements of thin polymer films , thus reducing their resistance to deformation @xcite .
surfaces can also modify the thermodynamic properties of polymeric materials .
the glass - transition temperature of a polymer melt can be significantly reduced in confinement , in a manner that sensitively depends on the boundary conditions @xcite .
furthermore , experiments recording the evolution of surface perturbations have shown that the surface mobility of a glassy polymer can be significantly enhanced relative to the bulk @xcite .
are these phenomena manifest in softer materials such as gels ? how do the structure and properties of these materials vary near the surface ? moving forward , we need to quantify changes in the structure and composition of soft solids in the vicinity of interfaces , and correlate these findings with measurements of surface mechanical properties . with these insights , we will be able to rationally design the mechanical properties of soft interfaces .
surface stresses can dramatically impact the way soft solids deform @xcite .
these solid capillary effects become significant below a critical _ elastocapillary length _
@xmath5 , where @xmath6 is the elastic modulus .
the basic physics is highlighted by the following argument : consider a surface with a sinusoidal corrugation of wavelength , @xmath7 , and amplitude , @xmath8 .
surface stresses act to flatten the surface , with a stress given by the laplace law that scales like @xmath9 . at the same time
, elastic forces will resist this deformation with a restoring stress that scales like @xmath10 .
when @xmath11 , surface stresses overpower elastic restoring forces , thereby flattening the surface .
the elastocapillary length is on the micron - scale for gels , nanometre - scale for elastomers , and is irrelevant for structural materials . here , we discuss two applications of the competition between surface and bulk effects in soft solids : wetting and adhesion . _
key questions : how do strain - dependent surface energies impact wetting and other elastocapillary phenomena ? what controls dynamic elastocapillary phenomena ?
_ classically , wetting is a pure surface effect , governed by the surface energies of the three phases @xcite .
recently , it has been shown that deformation of soft substrates by liquid surface tension @xcite can have a dramatic impact on the essential phenomena of wetting , including departures from the young - dupr law @xcite .
the majority of the experimental work on soft wetting has presumed that the surface stress of a soft solid is liquid - like ( isotropic and strain independent ) . as discussed above ,
this is not a robust assumption for soft solids or complex fluids .
recent theoretical work has shown that the state of stress near a three - phase contact line should be significantly changed when a surface is subject to the shuttleworth effect , _
i.e. _ the substrate has a strain - dependent surface energy , @xmath12 @xcite .
the equilibrium state minimizes the sum of the surface and strain - induced elastic energies .
here , the strain - dependence of the surface energy induces an additional elasto - capillary coupling , which for a drop on a membrane can lead to very large strains near the contact line @xcite . for a thin rod that is partially immersed , the shuttleworth effect leads to a discontinuity in strain @xcite , suggesting that great care must be taken when applying the classical capillary `` force '' picture near the contact line .
these ideas are consistent with an experiment reported in @xcite , but further experimental investigation is needed .
while some progress has been made on the importance of simple departures from ideal fluid - like surface tensions in the case of wetting , the issue is completely unexplored in other elastocapillary phenomena .
for instance , what are the implications of complex surface rheology in adhesion , fracture mechanics , and composites ? so far , the study of elastocapillary phenomena has focused on static situations .
however , the dynamics of these phenomena are ripe for investigation . in the context of wetting
, the deformation of the substrate due to liquid surface tension creates a new source of dissipation .
this dissipation , called _ viscoelastic braking _ , can dominate the motion of a contact line , causing droplets to slide more slowly on a soft surface than a stiff one @xcite .
experiments and modelling have directly related the speed of an advancing contact line to substrate viscoelasticity @xcite . on the other hand ,
coupling of surface stresses to bulk deformation can also _ drive _ contact - line motion . for example , droplets spontaneously translate down gradients in the stiffness of soft substrates @xcite . in this case , however , we do not have a physical model to predict the speed of droplet motion .
more generally , the dynamics of contact lines on soft surfaces could be modified by a variety of dissipative phenomena including poroelastic flow and substrate plasticity .
a broad range of exploratory experiments are needed .
_ key question : how does solid - vapour surface tension modify the adhesive contact of elastic solids ? _ the standard model of adhesive solid contacts was developed by johnson , kendall , and roberts ( jkr ) @xcite .
it features a competition of surface adhesive and bulk elastic energies , captured by the dimensionless parameter , @xmath13 . here , @xmath14 is a length scale associated with a contact and @xmath15 is the work of adhesion .
however , it does not include contributions from the surface stress of the solid - vapour interface , captured by the elastocapillary parameter , @xmath16 . in the limit of small contacts on soft substrates ,
theoretical analyses suggest that the jkr theory will be replaced by the young - dupr condition with the soft substrate in the place of the wetting liquid @xcite .
recently , a jkr - yd crossover was observed using silicone gels @xcite at a length scale around @xmath17 m .
however , the generality of these effects needs to be established with further experiments .
first , the jkr - yd crossover should be explored for a wider range of materials including hydrogels and recently - developed monophasic ultrasoft materials @xcite .
the effect of solid - vapour surface tension should also impact nanoscale contacts on stiffer polymeric materials , like elastomers @xcite . at this scale , long - range forces , embodied by the derjaguin - muller - toporov theory @xcite ,
may also compete with surface tension .
the role of solid - vapour surface tension in solid - solid contact mechanics could become manifest at larger length scales by adjusting another parameter : the thickness , @xmath18 , of the elastic medium . if @xmath19 , a separation of length scales leads to the dominance of much softer elastic modes , associated with the stretching modulus or bending rigidity .
this geometric effect is already being exploited in studies of wetting , where liquid surface tension creates macroscopic deformations of partially - wetted rods and sheets @xcite .
we anticipate that adhesion between slender solid objects will be a productive avenue for further investigation .
studies of the impact of solid - vapour surface tension on adhesive contacts have focused on the special case of zero load in equilibrium .
however , it is essential to understand the effects of external loading and the dynamical response .
for instance , the pull - off force and peeling dynamics are expected to depart dramatically from the pure elastic case @xcite . in the same spirit , following recent developments @xcite , the singularity of fracture mechanics near a soft crack tip should be regularized by surface tension .
once again , there is a direct call for experiments to test those new fundamental ideas with numerous practical implications .
the study of solid capillarity is in its early stages .
fundamental challenges remain in measuring and manipulating the surface mechanical properties of soft solids .
the coupling of surface stresses to bulk deformation presents a diverse range of problems , in wetting , adhesion , and other interfacial phenomena involving soft solids .
this nascent field will benefit from cross - fertilization with solid mechanics , rheology , and polymer science .
coupling between surface stress and bulk deformation also plays an important role in many biological phenomena @xcite , and robust dialog between the physical and biological communities promises to be very fruitful .
we thank all the participants at the workshop and the staff of the lorentz center for a productive and thought - provoking week .
the workshop was sponsored by the 3tu center of competence fsm ( fluid and solid mechanics ) and erc ( the european research council ) consolidator grant no .
f.b . acknowledges financial support from the people programme ( marie curie actions ) of the european union s seventh framework programme ( fp7/2007 - 2013 ) under rea grant agreement 623541 .
o.b . acknowledges support from the espci joliot chair 56ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop _
( , ) link:\doibase 10.1021/la201490 m [ * * , ( ) ] link:\doibase 10.1146/annurev.fluid.36.050802.122052 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1039/c4sm00200h [ * * , ( ) ] @noop * * , ( ) http://stacks.iop.org/0370-1298/63/i=5/a=302 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1016/s0001 - 8686(00)00045 - 2 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.97.266105 [ * * , ( ) ] @noop * * @noop * * , ( ) link:\doibase 10.1103/physrevlett.103.247801 [ * * , ( ) ] @noop * * @noop * * link:\doibase 10.1103/physrevlett.109.055701 [ * * , ( ) ] link:\doibase 10.1126/science.1151205 [ * * , ( ) ] @noop * * link:\doibase 10.1126/science.1244845 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1021/la801862 m [ * * , ( ) ] link:\doibase 10.1103/physrevlett.106.186103 [ * * , ( ) ] link:\doibase 10.1017/jfm.2014.152 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1039/c4sm00891j [ * * , ( ) ] link:\doibase 10.1140/epje / i2012 - 12134 - 6 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) http://arxiv.org/abs/1507.06325 [ ( ) ] , link:\doibase 10.1038/nmat4508 [ ( ) , 10.1038/nmat4508 ] @noop * * , ( ) @noop _ _ ( , , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) | soft solids differ from stiff solids in an important way : their surface stresses can drive large deformations . based on a topical workshop held in the lorentz center in leiden
, this opinion highlights some recent advances in the growing field of _ solid capillarity _ and poses key questions for its advancement .
surface tension plays a dominant role in the mechanics of liquids at small length scales . in figure
[ fig : poster ] , the water droplet curves into a spherical shape to minimize the energy due to the liquid - vapour interface . at the contact line , where the liquid , vapour , and solid meet
, the orientation of the liquid interface minimizes the energy associated with all three interfaces , as described by the famous young - dupr ( yd ) equation @xcite .
surface tension can also drive the motion of droplets , for example pulling them out of a tube @xcite or propelling them along a surface @xcite .
several recent studies have highlighted a new role for surface tension in the mechanics of deformable solids .
for instance , liquid droplets can bend slender solid objects @xcite like the flower petal in figure [ fig : poster ] .
additionally , a solid s own surface tension can strongly couple to its deformation .
for instance , soft cylinders can develop rayleigh - plateau pearling instabilities @xcite and soft composites can be stiffened by the surface tension of their inclusions @xcite .
the coupling of surface tension and deformation also plays an important role in wetting and adhesion , challengeing classic results like the young - dupr equation and the jkr theory of contact mechanics @xcite . to foster the burgeoning field of solid capillarity and identify the most important new areas for investigation , the lorentz center at the university of leiden hosted a workshop on https://www.lorentzcenter.nl/lc/web/2015/714/info.php3?wsid=714[capillarity of soft interfaces ] in november 2015 . in this opinion , we highlight some of the key questions identified by the 49 workshop participants , including both academic and industrial researchers
. these issues center on the mechanical properties of soft solid interfaces and their implications for bulk deformation . |
the spectroscopic investigation of hydrogen - like atomic systems has developed into an important test bed for research on atomic structure and bound - state quantum electrodynamics ( qed ) .
already high - precision measurements on atomic hydrogen confirm qed theory and its predictions to a high degree .
even more accurate tests of the predictive power of qed can be achieved by increasing either the precision of the measurement or by increasing the contribution of the effect to be measured @xcite . the use of highly - charged heavy ions shifts the magnitude of qed effects from the perturbative regime to a region where higher order terms become important and perturbation theory is no longer valid @xcite .
in addition , the energy scale of the @xmath8 ground state hyperfine structure is shifted to the laser accessible regime above a nuclear charge number of @xmath9 and the upper state lifetime enters a regime where acceptable fluorescence rates are expected @xcite .
+ based on these considerations , the spectroscopic investigation of the m1 transition of the ground state hyperfine structure of hydrogen - like bismuth ( @xmath0bi@xmath1 ) at @xmath2 nm has become one of the prime targets for further research along these lines .
theoretical investigations of the hyperfine transition energy , the lifetime and their respective qed contributions have been performed @xcite .
measurements of the transition energy have been realized on bunches of @xmath0bi@xmath1-ions travelling at a velocity of @xmath10 c in the experimental storage ring esr at the gesellschaft fr schwerionenforschung ( gsi ) @xcite .
the laboratory value of the transition wavelength was determined to @xmath11 nm @xcite .
the uncertainty was dominated by the limitation in the velocity calibration of the fast moving ions .
this initial uncertainty of approximately @xmath12 ghz has been reduced to approximately @xmath13 ghz in a later analysis @xcite . to overcome this limitation ,
the deceleration of highly - charged ions and a spectroscopic investigation of trapped ions have been proposed @xcite .
spectroscopy of highly - charged ions stored in penning traps is expected to reach accuracies which make tests of bound - state qed calculations significantly more stringent and give access to improved measurement of fundamental quantities @xcite .
+ at the spectrap experiment at the hitrap facility @xcite at gsi , systematic measurements on several species of highly - charged hydrogen- and lithium - like ions will improve the spectroscopic resolution by up to three orders of magnitude over previous experiments .
the assembly consists of a cryogenic penning trap , a frequency stabilized laser for excitation of the hyperfine transition under investigation , and a detection system for emitted fluorescence light @xcite .
the trap will be loaded with highly - charged ions produced at the gsi accelerator facility which are decelerated and delivered by the hitrap beam - lines and cooled to liquid helium temperatures inside the penning trap . for efficient spectroscopy of @xmath0bi@xmath1
the laser system should produce laser light with a power of several mw at @xmath2 nm and a tuning range significantly larger then the standard deviation of @xmath13 ghz of the previous measurement .
in addition , the laser frequency should be stable to a precision and accuracy comparable to or below the expected doppler - width of the transition of the ions inside the penning trap which is expected to be approximately @xmath14 mhz .
we have developed , built , and characterized a laser systems which fulfils these requirements .
the system consists of four major components ( fig .
[ optics - schematics ] ) : a commercial frequency - quadrupled diode laser system ( target laser , toptica ta - fhg 110 ) generating light at a wavelength of @xmath15 nm , a pair of offset - locked diode lasers operating at @xmath16 nm ( master and transfer laser ) , a confocal transfer cavity locking the target laser to the transfer laser at a large wavelength offset , and a set of spectroscopic diagnostics at @xmath17 nm ( rubidium ( rb ) absorption cell ) and @xmath18 nm ( tellurium ( @xmath19te@xmath3 ) absorption cell ) for characterization of the different components of the frequency stabilisation chain . + the target laser produces light around @xmath5 nm and an output power of up to @xmath20 mw .
this is achieved by fourth harmonic generation of the light of a tapered amplifier with an output power of @xmath21 mw seeded by an external cavity diode laser operating at a wavelength of @xmath22 nm . the manufacturer specifies the linewidth of the system to be smaller than @xmath23 mhz at @xmath24 nm ( @xmath25s measurement interval ) .
the modular setup with two successive frequency doubling stages gives access to three laser fields : light at @xmath26 nm is used for frequency stabilization , light at @xmath4 nm is used for frequency diagnostics on @xmath19te@xmath3 vapour , and light at @xmath5 nm is used for high - resolution spectroscopy of @xmath0bi@xmath1 . in order to be able to implement the required large tuning range of several @xmath13 ghz ,
the direct stabilization of one of these wavelengths to an atomic or molecular transition is not possible .
for that reason , we implemented a stabilization scheme where the fundamental laser output at @xmath22 nm is locked to a length - stabilized transfer cavity .
the transfer cavity serves as a stable reference similar to the one used in @xcite , where a cavity is actively stabilized to a polarization stabilized helium - neon laser .
in contrast to this approach , our transfer cavity is stabilized to a tunable diode laser ( transfer laser ) . by tuning the frequency of the transfer laser
, we can tune the resonance frequency of the transfer cavity and in this manner also the output frequency of the target laser .
an overview of the involved laser wavelengths is provided by fig . [ frequencies ] : the anchor of the stabilization chain is the master laser stabilized to a @xmath27rb transition at @xmath28 nm by doppler - free saturation spectroscopy .
the transfer laser is offset - locked to the master laser with a computer - controlled frequency offset in the range of @xmath29 to @xmath30 ghz with respect to @xmath16 nm .
the transfer cavity is stabilized to the transfer laser , bridges the wavelength gap to @xmath26 nm , and serves for locking of the fundamental output of the target laser system . using a sample - and - hold technique ( see sec .
[ mopastab ] ) a scan range of several @xmath31 ghz can be achieved .
finally , two successive frequency doubling stages generate light at @xmath18 and @xmath24 nm . + the following sections give a more detailed discussion of the stabilization chain . * * the master and transfer lasers are home - built external - cavity diode lasers operating at a wavelength of @xmath16 nm . the master laser is stabilized to the @xmath32 cross - over resonance of @xmath27rb by doppler - free absorption spectroscopy in a mu - metal shielded vapour cell . the stabilization is based on phase - sensitive detection of the saturated absorption signal using a lock - in amplifier with direct modulation of the diode laser s injection current at a frequency of @xmath33 khz .
the feedback bandwidth of @xmath34 khz of the pi servo loop is defined by the time constant of the lock - in amplifier .
typical short - term line - widths of the @xmath16 nm lasers are in the range of @xmath35 khz . +
a heterodyne beat - signal between the outputs of master and transfer laser is mixed with the output of an rf - synthesizer with a double balanced mixer .
a frequency - to - voltage converter produces an error signal proportional to the frequency offset between the beat and the synthesizer frequency .
this error signal is fed to a pi controller for stabilization of the transfer laser frequency .
the transfer laser frequency can be tuned with respect to the fixed master laser frequency with sub - khz resolution by tuning the synthesizer output .
this allows us to scan the frequency of the transfer laser with high accuracy and in an arbitrary frequency sequence .
+ the transfer laser stabilization and tuneability are enhanced using the intra - cavity stabilization scheme described in @xcite . here ,
the internal cavity of the laser diode consisting of its front and rear surface is mode - matched to an external cavity formed by the rear surface of the laser diode and an external cavity mirror .
mode - matching is controlled by the injection current of the laser diode . with this technique
we reach a mode - hop free tuning range of @xmath36 ghz ( at @xmath17 nm ) , limited by the bandwidth of the beat - note detection . + we use an independent doppler - free absorption spectroscopy on rubidium to characterize the performance of the frequency stabilization and to determine the uncertainty of the transfer laser frequency . by repeated scans over the doppler - free resonances of the @xmath37 transitions of @xmath38rb and by fitting lorentzian profiles to the occurring resonance lines ( fig .
[ hiss_rb ] ( a ) ) we can extract the centre frequencies of the observed lines .
the resulting histograms of the centre frequencies of all lines and of the @xmath39 cross - over transition are presented in figs .
[ hiss_rb ] ( b ) and ( c ) , respectively . a gaussian fit ( dashed curve in fig .
[ hiss_rb ] ( c ) ) gives a statistical uncertainty of @xmath40 khz ( one standard deviation ) for data accumulated over several hours and @xmath41 khz for data compiled over several days .
we use a transfer cavity for stabilization of the target laser to the transfer laser with large wavelength difference .
the length of the transfer cavity is locked to the transfer laser wavelength using a piezo transducer in one of the cavity mirror mounts .
our locking scheme , which includes a sample - and - hold circuit for a controlled change of the cavity mode number ( see sec .
[ mopastab ] ) is applicable to a wide range of target wavelengths and in addition allows to extend the tuning range of the target laser well beyond the transfer laser tuning range .
the transfer cavity consists of a temperature - stabilized aluminium spacer and low - reflective silver mirrors giving a free spectral range ( @xmath42 ) of about @xmath29 ghz .
an aluminium spacer is sufficient for our purpose due to the active length stabilization to the transfer laser . in a later stage switching to invar , zerodur , or ule
may be considered . by changing the cavity length to the next mode number
, the @xmath42 is changed by approximately @xmath43 khz .
the free spectral range is measured by scanning the laser frequency between absolutely calibrated features of tellurium ( sec .
[ tespec ] ) and recording the number of transmission fringes .
these measurements resulted in a @xmath44 mhz .
+ the reflectivity of the cavity mirrors with coatings of @xmath45 nm ag and @xmath46 nm mgf@xmath3 was chosen to be @xmath47 % at @xmath26 nm and @xmath48 % at @xmath16 nm , respectively , in order to give a large capture range for the transfer cavity locks .
the theoretical value for the finesse of @xmath36 is confirmed experimentally by a measured value of @xmath49 ( @xmath50 mhz ) . though the large fringe width makes the system sensitive to offset drifts in the electronics , monitoring and correcting these drifts resulted in insignificant contributions to the frequency drifts .
the temperature stabilization of the cavity spacer has small residual thermal variations resulting in piezo transducer stabilization voltage variations on long time - scales .
we determined that these variations of approximately @xmath51 v translates to a remaining temperature fluctuation of about @xmath52c which corresponds to the expected precision of the temperature stabilization used .
+ the length stabilization of the transfer cavity as well as the stabilization of the target laser to the transfer cavity are done by top - of - fringe locking since it is insensitive to intensity fluctuations . by modulating the position of one of the cavity mirrors at @xmath53 khz we produce the error signals for both stabilization circuits using phase - sensitive demodulation with lock - in amplifiers .
we stabilize the fundamental wavelength of the target laser system at @xmath26 nm to the transfer cavity .
a sample - and - hold technique allows us to extend the tuning range of the target laser beyond the tuning range of the transfer laser : at the limit of the tuning range , the feedback controller is muted by external control .
this causes the control voltage to stay constant ( `` hold operation '' ) on the last value before muting and the target laser to hold its set - point .
the wavelength is then held without feedback , but small wavelength drifts may occur .
next we scan the transfer cavity back by an integer multiple of the @xmath54 and relock the target laser to another cavity fringe .
the large capture range due to the small finesse of the cavity ( see inset in fig . [
optics - schematics ] ) guarantees relocking although the target laser wavelength may have a small deviation from the expected wavelength .
the fringe shifting by sample - and - hold ensures the cavity length on average to stay the same and the @xmath54 to stay on a well defined value .
in addition , the mode - hop free tuning range of the target laser system of approximately @xmath55 ghz at @xmath24 nm ( limited by the finite tuning range of the doubling and quadrupling cavities ) can be fully exploited .
realigning the doubling and quadrupling stages results in a tuning range of the system of more then @xmath13 ghz in the uv as required .
the fundamental frequency of the target laser system is frequency doubled to @xmath4 nm and then again frequency doubled to @xmath5 nm by two separate external doubling cavities . a small fraction of the blue laser light at @xmath56 is diverted for doppler - free saturation spectroscopy in a vapour cell of tellurium ( @xmath19te@xmath3 ) . similar to iodine,@xmath19te@xmath3
has many resonance lines that cover most of the visible spectrum , so that the second harmonic frequency of the target laser can be used to determine the frequency uncertainty of the stabilized laser system .
in addition , high - precision reference lines in @xmath19te@xmath3 can be used to give absolute accuracy to our laser system . a @xmath57 @xmath19te@xmath3 cell similar to the one employed in @xcite ,
is heated to @xmath58 in a tube furnace . by scanning the laser system , we record the doppler - broadened and the doppler - free transition lines of @xmath19te@xmath3 over a range of several 100 ghz .
we can identify numerous doppler - broadened lines of the `` tellurium atlas '' @xcite and the doppler - free lines published in @xcite . in specific , we scan the known d@xmath59 reference line @xcite numerous times ( fig .
[ hiss - euvar ] ( a ) ) and extract the centre frequency of this line , by fitting a lorentz function to each measurement .
the extracted centre frequencies are plotted in a histogram ( fig .
[ hiss - euvar ] ( b ) ) .
a gaussian fit function describes the distribution well .
the standard deviation of the gaussian is @xmath60 mhz ( at @xmath18 nm ) for the compiled data from over @xmath61 measurements recorded in six consecutive days including a @xmath35 hour period of continuous laser operation and data taking . for the final target wavelength of @xmath24 nm
this corresponds to an uncertainty of @xmath62 mhz which is well below the required uncertainty of @xmath14 mhz as given by the doppler broadening of the targeted @xmath0bi@xmath1 transition . during the six day measurement run resulting in the data of fig .
[ hiss - euvar ] ( b ) , we observed a drift in the measured centre frequency of a few mhz , which we attribute to a change in air pressure . using this additional information it is possible to further reduce the uncertainty of the target laser output frequency in our open cavity configuration .
ultimately , housing the transfer cavity in a constant - pressure or evacuated environment should eliminate this effect .
+ to gain further insight , we analysed our data in the light of potential temperature and pressure dependences of the target laser wavelength caused by a wavelength dependent variation of the transfer cavity @xmath63 at @xmath16 nm and @xmath26 nm . a good description of the influence of pressure and temperature on an open transfer cavity is provided by @xcite .
the resonance frequency of the transfer cavity for the @xmath64 mode is given by @xmath65 where @xmath66 is the resonance frequency , @xmath67 is the order of the longitudinal mode , @xmath68 is the vacuum speed of light , @xmath69 is the separation of the cavity mirrors and @xmath70 is the index of refraction of the medium between the mirrors . at the transfer laser wavelength ,
changes in the refractive index are compensated by keeping the optical path length between the mirrors constant through the feedback loop .
the stabilized frequency of the target laser , on the other hand is affected by the dispersion of air , especially when large frequency spans are bridged by the transfer cavity .
the refractive index of air under standard conditions @xmath71 is calculated for @xmath17 nm and @xmath22 nm following the procedure given in @xcite relative to the refractive index of vacuum which is @xmath72 : @xmath73 & \cdot 10^{-8}\end{aligned}\ ] ] with the wavelength @xmath74 ( in @xmath75 m ) .
the difference in the refractive indices results to @xmath76 .
the dependence of the refractive index of air @xmath77 on pressure @xmath78 ( in pa ) and temperature @xmath79 ( in @xmath80c ) is given by @xmath81 according to @xcite and can be approximated by a taylor expansion up to linear order in t around t@xmath82c through @xmath83 with @xmath84 . according to equation ( [ cavglg ] ) ,
the resulting deviation of the target laser frequency is given by : @xmath85 for a calculation of the uncertainty in @xmath77 , the linearised taylor expansion ( [ taylor ] ) is applied to realistic parameters ( @xmath86 ) . only the difference in the index of refraction @xmath87 which is not compensated by the cavity length stabilization has to be taken into account resulting in : @xmath88 using @xmath89 hpa .
for the assumed temperature and pressure fluctuations of @xmath90 and @xmath91 , respectively , the contribution of temperature fluctuations can be neglected .
pressure fluctuations result in a change of the stabilized target laser frequency of @xmath92 mhz per hpa at @xmath22 mhz which scales to @xmath93 mhz and @xmath94 mhz per hpa at @xmath18 nm and @xmath24 nm , respectively . for the largest assumed pressure variations of @xmath34 hpa ,
the resulting change in the stabilized target frequency of @xmath95 mhz at @xmath24 nm is still well below the targeted absolute accuracy of @xmath14 mhz .
+ we use our data taken over a six day period with significantly varying air pressure to confirm this dependence and to show that knowledge of the air pressure can be used to further reduce the frequency uncertainty of our laser system .
figure [ pressureplot ] presents the frequency to resonantly excite the d@xmath59-line averaged for each day of data taking as a function of the average air pressure reported for the darmstadt area .
we see a clear linear dependence of the d@xmath59 frequency on air pressure . fitting a linear function to the data
gives a measured slope of @xmath96 = @xmath97 mhz / hpa at @xmath18 nm .
this slope agrees well with the calculated slope of @xmath98 mhz / hpa when taking to account that an increase in the target laser frequency for increasing air pressure has to be compensated by a decrease in the transfer laser frequency to be still in resonance with the selected tellurium line .
in this article , we have presented a laser system for the high - precision measurement of the hyperfine structure of hydrogen - like @xmath0bi@xmath1 .
we have discussed a generally applicable approach for stabilization of a laser system to almost any target wavelength by making use of a rubidium - stabilized laser and a broadband transfer cavity .
the performance that can be reached depends on the frequency difference between the laser used to stabilize the transfer cavity and the laser to be stabilized to the cavity due to the pressure dependence of the refractive index of air .
this influence can be further reduced by monitoring the air pressure or by placing the transfer cavity inside a constant - pressure or evacuated environment .
doppler - free saturation spectroscopy on tellurium can be used to determine the frequency uncertainty of the laser system and to provide reference lines for an absolute frequency calibration of the laser system .
we thank m. schlosser , t. lauber , j. kber , j. schtz , s. tichelmann and the spectrap collaboration for numerous helpful discussions . this work has been supported in part by the bmbf ( contract numbers 06da9019i and 06da9020i ) and the dfg ( contract number bi 647/4 - 1 ) .
t. sthlker et al . :
notes phys . *
745 * , ( 2008 ) 157163 .
d. f. a. winters et al . :
* 85 * , ( 2007 ) 403408 ; arxiv:0704.0560 ( 2007 ) .
m. vogel et al . :
instr . * 76 * , ( 2005 ) 103102 .
s. borneis et al . :
. int . * 127 * , ( 2000 ) 305310 .
shabaev et al .
* 86 * , ( 2001 ) 3959 .
m. finkenbeiner , b. fricke , and t. khl : phys .
a * 176 * , ( 1993 ) 113117 .
i. klaft et al .
* 73 * , ( 1994 ) 2425 .
w. quint et al .
: hyperfine interactions * 132 * , ( 2001 ) 457461 .
m. vogel and w. quint : phys . rep . *
490 * , ( 2010 ) 147 .
e. riedle , s. h. ashworth , j. t. farrell , and d. j. nesbitt : rev .
instr . * 65 * , ( 1994 ) 4248 .
t. fhrer , d. stang , and t. walther : optics express * 17 * , ( 2009 ) 49914996 . | we present and characterize a laser system for the spectroscopy on highly - charged @xmath0bi@xmath1 ions at a wavelength of @xmath2 nm . for absolute frequency stabilization , the laser system is locked to a near - infra - red laser stabilized to a rubidium transition line using a transfer cavity based locking scheme .
tuning of the output frequency with high precision is achieved via a tunable rf offset lock .
a sample - and - hold technique gives an extended tuning range of several thz in the uv .
this scheme is universally applicable to the stabilization of laser systems at wavelengths not directly accessible to atomic or molecular resonances .
we determine the frequency accuracy of the laser system using doppler - free absorption spectroscopy of te@xmath3 vapour at @xmath4 nm . scaled to the target wavelength of @xmath5 nm
, we achieve a frequency uncertainty of @xmath6 = @xmath7 mhz ( one standard deviation ) over six days of operation . |
the problem of decomposing a vector as a linear combination of simple vectors is central in many areas of applied mathematics , machine learning , and engineering .
the length of the shortest decomposition is usually called the * rank * of the vector .
we consider the situation where the simple vectors form an * algebraic variety*. this includes well - studied and important cases such as tensor rank ( real or complex ) , which is the rank with respect to the segre variety , symmetric tensor rank ( or waring rank ) , which is the rank with respect to the veronese variety , and anti - symmetric tensor rank , which is the rank with respect to the grassmannian in its plcker embedding
. we will be mostly interested in the rank over the real numbers .
however , we make the definition over an arbitrary field @xmath1 .
let @xmath2 be a projective variety defined over a field @xmath1 and let @xmath3 be the affine cone over @xmath0 .
the variety @xmath0 is called nondegenerate if @xmath0 ( or equivalently @xmath4 ) is not contained in any hyperplane . in this case , for any @xmath5 , @xmath6 , we can define the * rank * of @xmath7 * with respect to * @xmath0 ( @xmath0-rank for short ) as follows : @xmath8 i.e. the rank of @xmath7 with respect to @xmath0 is the smallest length of an additive decomposition of @xmath7 into points of @xmath4 with coordinates in @xmath1 .
a rank @xmath9 is called * generic * if the vectors of @xmath0-rank @xmath9 contain a zariski open subset of @xmath10 . over any algebraically closed field
, there is a unique generic @xmath0-rank for a nondegenerate variety @xmath0 . over the real numbers ,
a rank @xmath9 is called * typical * if the set of vectors of @xmath0-rank @xmath9 contains an open subset of @xmath11 with respect to the euclidean topology
. there can be many typical ranks for a given variety @xmath0 .
in this paper , we study upper bounds for the maximal real rank and the range between the minimal typical rank and the maximal real rank in terms of the variety @xmath0 which we allow to vary .
we now outline our results : let @xmath12 and @xmath13 denote the minimal typical rank and the maximal @xmath0-rank respectively .
a trivial upper bound on the rank of a point is simply the dimension of the ambient vector space , which works over any field @xmath1 : @xmath14 in @xcite , landsberg and teitler showed that over the complex numbers ( or any algebraically closed field ) this bound can be improved : @xmath15 in @xcite , blekherman and teitler showed that for real varieties @xmath16 and also @xmath17 where @xmath18 is the generic rank with respect to @xmath0 over @xmath19 .
the real analogue of the landsberg - teitler bound is @xmath20 this bound can be established by adapting their proof to the real case .
the proof is by a short inductive argument , so we included it below , see theorem [ thm : upperbound ] .
this was already observed by ballico , see @xcite , although his proof is terse .
we give a full argument for completeness . in case that @xmath21 is odd ,
ballico improves upon this bound and shows that @xmath22 .
note that in most cases , the bound will be stronger than , unless the secant varieties of @xmath0 are strongly defective .
however , when @xmath0 is a hypersurface , the blekherman - teitler bound is @xmath23 , while the bound is @xmath24 .
when @xmath0 is a curve , its secant varieties are known to be non - defective , see e.g. lange @xcite , and the bound in is @xmath25 when @xmath26 is even , and @xmath27 when @xmath26 is odd .
this is worse than the bound , which agrees with the trivial bound .
we establish the tightness of the bound by constructing examples of irreducible varieties @xmath2 of any dimension and points of maximal real @xmath0-rank @xmath28 .
we give two constructions of such varieties , one in section [ sec : bounds ] and one in section [ sec : waring ] .
we also find examples , where this maximal real rank is typical , see theorem [ thm : tight ] and corollary [ cor : tight ] .
the rational normal curve establishes the tightness of ballico s improvement in the case of @xmath21 being odd mentioned above , see remark [ rem : ballicobound ] .
our second contribution is a proof of a conjecture of reznick ( * ? ? ?
* conjecture 4.12 ) which states that a bivariate form of degree @xmath29 has maximal real rank @xmath29 if and only if it has only real roots , see theorem [ thm : hyperbolic ] . under the additional assumption that the form has distinct roots ,
this conjecture was established in previous work of causa - re , comon - ottaviani , and reznick , see @xcite .
we provide an inductive argument which does not rely on the distinctness of the roots .
we use this characterization of binary forms of maximal real rank to study the gap between minimal typical rank and maximal real rank . in theorem
[ thm : maxtypicalrank ] , we find curves @xmath0 with the property that the maximal real rank is twice the generic @xmath0-rank , which is the maximal range by the blekherman - teitler bound .
our fourth contribution is that the gap between the maximal real rank and the maximal complex rank with respect to a variety @xmath0 can be arbitrarily large , see corollary [ cor : gapcomplexrank ] .
we are not aware of any instances where this difference had been observed to be larger than @xmath30 .
the proof of reznick s conjecture gives us one of the few instances where the set of points of maximal real rank is fully understood .
our final contribution is a study of the set of points of maximal real rank for the case of plane curves .
we find that its geometry is governed by tangent lines with odd intersection multiplicity , see theorem [ thm : planecurves ] .
we would like to thank alessandra bernardi , jarosaw buczyski , hwangrae lee and giorgio ottaviani for their comments on a draft of this paper .
both authors were partially supported by nsf grant dms-1352073 .
the first author was partially supported by the simons institute for the theory of computing during the algorithms and complexity in algebraic geometry semester .
in this section , let @xmath31 be an irreducible nondegenerate variety with a smooth real point .
this implies that the set of real points of @xmath0 is zariski - dense in @xmath0 , see ( * ? ? ?
* section 2.8 ) .
we give an upper bound on the real rank of any point in @xmath32 in terms of the codimension of @xmath0 and establish its tightness via an explicit construction .
the following theorem is proved by adapting the landsberg - teitler proof to the real case .
this was also observed by ballico , see ( * ? ? ?
* theorem 1(i ) ) .
[ thm : upperbound ] the real rank of any point @xmath33 with respect to @xmath0 is at most @xmath34 the proof is by induction on the dimension of @xmath0 . if @xmath35 , we have @xmath36 , which is a trivial bound on the @xmath0-rank of any point in @xmath37 .
now suppose @xmath38 and let @xmath39 be a point , @xmath40 .
let @xmath41 be a generic hyperplane containing @xmath7 .
we first show that @xmath42 is irreducible and nondegenerate in @xmath43 . to show that @xmath42 is irreducible
, we follow the argument in griffiths - harris @xcite on page 174 : let @xmath44 be a general point and @xmath45 be a linear space of dimension @xmath46 which contains @xmath7 and intersects @xmath0 transversally in @xmath47 .
let @xmath48 be the unique irreducible component of @xmath49 containing @xmath47 .
we consider the pencil @xmath50 of hyperplanes that contain @xmath51 .
then @xmath52 for all @xmath53 and @xmath47 is contained in a unique irreducible component @xmath54 of @xmath55 , because @xmath56 and @xmath0 also intersect transversally in @xmath47 .
now set @xmath57 .
then @xmath58 and therefore @xmath59 is zariski - dense in @xmath0 . since @xmath60 for generic @xmath53 , we conclude @xmath61 is irreducible for generic @xmath53 . to show that @xmath42 is nondegenerate
, we follow landsberg - teitler ( * ? ? ?
* proposition 5.1 ) : suppose @xmath43 intersects @xmath0 transversally and let @xmath62 . if @xmath42 were contained in a hyperplane in @xmath43 , say @xmath63 , then @xmath64 is a regular function on @xmath0 because it has no poles ( whenever @xmath65 for a point @xmath44 , we have @xmath66 ) . since @xmath0 is projective
, @xmath64 is constant on @xmath0 , which means @xmath67 on @xmath0 .
this implies that @xmath68 is degenerate , which contradicts our assumptions on @xmath0 .
so , given a real point @xmath69 , we can find a real hyperplane @xmath70 containing @xmath7 such that @xmath42 is irreducible and nondegenerate .
we can further assume that @xmath42 has a regular real point , because this is true for hyperplanes in at least one connected component of the complement of @xmath71 , the intersection of the dual variety of @xmath0 with the hyperplane of all hyperplanes containing @xmath7 .
so by induction , the real rank of @xmath7 with respect to @xmath42 is bounded by @xmath72 the above theorem remains true if we substitute the assumption that @xmath0 is irreducible by the assumption that for every irreducible component of @xmath0 , the real points are dense in it .
our next goal is to establish the tightness of this bound by an explicit construction .
we make some observations of on the behavior of real rank under projections and joins of varieties : [ prop : projections ] let @xmath73 be a linear projection with center @xmath51 defined over @xmath74 .
let @xmath75 be an irreducible nondegenerate variety and suppose @xmath76 is a proper subvariety .
let @xmath77 be a real point .
then @xmath78 if for every real point @xmath79 there is a real point @xmath80 such that @xmath81 , then the above inequality is an equality .
let @xmath82 be a point and @xmath83 with @xmath84 . then @xmath85 and @xmath86 , where we assume , after reordering , that @xmath87 and @xmath88 .
note that @xmath89 .
so the rank of @xmath90 is at most @xmath91 , which shows the claimed inequality .
now suppose that every real point @xmath79 has a real preimage .
then , given @xmath92 with @xmath93 , we set @xmath94 , where the @xmath95 are real preimages of @xmath96
. then the real @xmath0-rank of @xmath7 is at most @xmath9 and @xmath97 .
this gives the reverse inequality .
[ prop : joins ] given two irreducible , nondegenerate real varieties @xmath98 and @xmath99 , we form the join @xmath100 .
the real @xmath101-rank of any point @xmath102 is @xmath103 the maximum of the real ranks of the projections of @xmath7 onto @xmath104 and @xmath105 with respect to @xmath0 and @xmath106 , respectively .
the inequality @xmath107 follows from proposition [ prop : projections ] because the projection of @xmath101 onto @xmath104 is @xmath0 and the projection of @xmath101 onto @xmath105 is @xmath106 .
conversely , given decompositions of @xmath108 and @xmath109 with @xmath84 and @xmath110 , we get a decomposition of @xmath111 by writing @xmath112 where we assumed without loss of generality that @xmath113 and set @xmath114 for @xmath115 .
note that @xmath116 is a real point of @xmath101 because it lies on the line between @xmath117 and @xmath118 .
so the real @xmath101-rank of @xmath7 is at most the maximum of the real @xmath0-rank of @xmath47 and the real @xmath106-rank of @xmath119 .
we also need the following criterion certifying that a point has maximal real rank with respect to @xmath0 .
[ prop : rankchar ] let @xmath69 and @xmath120 .
assume that every real linear space @xmath121 of dimension @xmath122 and containing @xmath7 intersects @xmath0 in at most @xmath122 real points .
then the real @xmath0-rank of @xmath7 is @xmath123 , i.e. maximal .
we prove the contrapositive : let @xmath124 with @xmath125 and set @xmath120
. then @xmath7 lies in the linear span of @xmath9 real points @xmath126 on @xmath0 , i.e. @xmath127 by adding general real points @xmath128 , if necessary , we find a linear space @xmath129 containing @xmath7 such that @xmath130 and @xmath131 .
the preceding proposition reads as follows in case that @xmath0 is a curve : if every hyperplane through @xmath7 intersects @xmath0 in at most @xmath132 real points , the real @xmath0-rank of @xmath7 is @xmath27 .
our general idea is to prove that a real point @xmath133 , @xmath134 , has maximal real rank @xmath28 with respect to @xmath0 by considering the projection @xmath135 of @xmath0 from @xmath7 and using proposition [ prop : rankchar ] .
the image of the variety @xmath98 under @xmath135 is closed because @xmath40 .
we denote it by @xmath136 .
the assumption of proposition [ prop : rankchar ] that every linear space @xmath137 containing @xmath7 intersects @xmath0 in at most @xmath138 many real points translates by projection to the assumption that every linear space @xmath139 of dimension @xmath140 intersects @xmath141 in at most @xmath142 many real points .
this condition can be realized by curves of minimal degree , which are exactly the rational normal curves .
the condition that the intersection with a linear subspace of complimentary dimension has at most @xmath143 points is closely related to the variety @xmath141 being _ of minimal degree_. however , varieties of minimal degree of higher dimension , except for irreducible quadric hypersurfaces , do not satisfy the condition that the intersection with any linear subspaces of complimentary dimension is @xmath144-dimensional .
see ( * ? ? ?
* theorem 1 ) for a classification of varieties of minimal degree .
we now prove that the upper bound in theorem [ thm : upperbound ] is tight , first in the case of curves .
[ thm : tight ] let @xmath145 .
there is an irreducible nondegenerate curve @xmath2 with a regular real point and a real point @xmath69 such that @xmath146 .
furthermore , the maximal real rank can be typical . for @xmath147 , we can take @xmath148 , see also section [ sec : planecurves ] .
so let @xmath149 .
we choose coordinates in such a way that @xmath150 and set @xmath151 to be @xmath152 .
let @xmath153 be a curve of minimal degree , i.e. @xmath154 .
let @xmath155 be the cone over @xmath156 with cone point @xmath7 .
consider the linear system @xmath157 of cubic hypersurfaces @xmath158 , where @xmath159 is a plane cubic curve and @xmath160 .
note that @xmath157 is the set of all cubic hypersurfaces in @xmath37 whose equations involve only the variables @xmath161 .
the intersection of a general element of this linear system with @xmath155 is irreducible by bertini s theorem , see ( * ? ? ?
* thorme 6.3(4 ) ) ( use @xmath162 , @xmath163 ) . in @xmath157 , the set of pseudohyperplanes , i.e. hypersurfaces @xmath164 such that the set of real points @xmath165 has only one connected component and @xmath166
is connected , has non - empty interior in the euclidean topology , because it is true for plane cubics ( e.g. @xmath167 is an interior point ) .
so set @xmath168 for a general cubic @xmath169 .
note that @xmath0 is nondegenerate . by genericity , the projection @xmath170 induces a @xmath171 correspondence of real points of @xmath0 and @xmath172 .
indeed , this is true for @xmath173 ( see example [ exm : curvep3 ] ) , so also for a general cubic in @xmath157 . now @xmath174 and , by construction , every hyperplane containing @xmath7 intersects @xmath0 in at most @xmath175 real points . by proposition
[ prop : rankchar ] , @xmath7 has real rank @xmath176 with respect to @xmath0 , which proves the first claim . note that in this example , points in a neighbourhood of @xmath7 have the same real rank with respect to @xmath0 : by construction , every real hyperplane @xmath70 containing @xmath7 intersects @xmath0 in at most @xmath132 many real points , counting multiplicities . by upper semi - continuity of the intersection multiplicity ( see the following lemma [ lem : intersectionmult ] ) , the same is true for real points in a neighbourhood of @xmath7 .
so by proposition [ prop : rankchar ] , they also have real @xmath0-rank @xmath27 .
so the maximal real rank is typical in these examples .
we used upper semi - continuity of the intersection multiplicity in the following form in the above proof .
[ lem : intersectionmult ] let @xmath2 be a nondegenerate irreducible curve and let @xmath177 be a sequence of hyperplanes in @xmath37 , converging in the euclidean topology to @xmath178 .
suppose the length of the schemes @xmath179 is at least equal to @xmath180 .
then the length of @xmath42 is at least @xmath181 .
in particular , the number of real intersection points is upper semi - continuous ( counting with multiplicities ) .
first note that the intersection of @xmath0 with any hyperplane is @xmath144-dimensional .
the length of a @xmath144-dimensional scheme is equal to the constant coefficient of the hilbert polynomial of its defining ideal .
so the length of the scheme @xmath179 is the dimension of @xmath182/(i_d + \langle \ell_j\rangle_d)$ ] , where @xmath183 is the homogeneous ideal defining @xmath0 and @xmath184 , for large enough @xmath185 .
choose the scaling of the linear functionals @xmath186 such that the sequence @xmath187 converges to @xmath188 .
then @xmath189 and the codimension of @xmath190 is at least @xmath181 , which is a lower bound on the codimension of @xmath191 , by semi - continuity of the dimension in linear algebra . +
this implies that the number of real intersection points is upper semi - continuous because the condition of having a given number of complex roots is open .
[ exm : curvep3 ] we construct a curve @xmath192 such that the point @xmath193 has real rank @xmath23 with respect to @xmath0 following our proof of theorem [ thm : tight ] .
we start with a curve @xmath141 of minimal degree in @xmath194 , say @xmath195 .
then @xmath155 is the cone over @xmath141 with cone point @xmath196 , i.e. @xmath197 .
we now intersect with the pseudohyperplane @xmath198 and set @xmath168
. then @xmath0 is irreducible and the projection @xmath199 from @xmath196 induces a @xmath171 correspondence on real points .
indeed , the discriminant of the cubic @xmath200 in @xmath201 is @xmath202 so whenever @xmath203 and @xmath204 are real , the discriminant is a negative number as certified by the sum of squares representation .
this means that there is a single real root @xmath201 , given real values for @xmath203 and @xmath204 .
if @xmath203 and @xmath204 are real , then so is @xmath205 because @xmath206 .
this shows that @xmath199 induces a @xmath171 correspondence on real points of @xmath0 and @xmath141 . by construction , every plane @xmath51 in @xmath207 through @xmath196 intersects @xmath0 in at most @xmath208 real points because the line @xmath209 intersects @xmath141 in at most two real points . by proposition [ prop : rankchar
] , the point @xmath196 has real rank @xmath23 with respect to @xmath0 .
we can also get rid of the genericity assumptions in the proof of theorem [ thm : tight ] at the cost of taking a zariski closure of real intersection points , i.e. computing a real radical ideal . as in the proof of theorem [ thm : tight ] , let @xmath210 be a curve of minimal degree .
set @xmath211 and @xmath212 .
let @xmath213 and suppose @xmath214 , which can be achieved by choosing suitable coordinates on @xmath215 .
then the zariski closure @xmath0 of @xmath216 is irreducible and the real rank of @xmath217 with respect to @xmath0 is @xmath27 .
this is true , because @xmath155 and @xmath164 intersect transversally in real points .
note that every intersection point of @xmath155 and @xmath164 is regular on both varieties .
so this claim is true because every real line through @xmath7 intersects @xmath164 in a simple real point ( every line through @xmath218 intersects @xmath219 transversally in one real point . because @xmath220 is connected and the projection @xmath221 gives a @xmath171 correspondence on real points , @xmath216 is also connected .
these points are regular points on @xmath222 , so the zariski closure is irreducible .
[ cor : tight ] for every pair of integers @xmath223 , there is an irreducible , nondegenerate variety @xmath224 of dimension @xmath29 with regular real point and a real point @xmath225 with maximal real rank @xmath123 with respect to @xmath0 .
furthermore , the maximal real rank can also be typical .
we obtain these varieties by constructing a curve @xmath226 of codimension @xmath122 with the required properties .
we then take cones over @xmath141 iteratively until we arrive at the desired dimension @xmath29 . in other words ,
we take the join of @xmath141 and a complimentary linear subspace of dimension @xmath227 as in proposition [ prop : joins ] .
the maximal real @xmath0-rank is at most twice the minimal typical real rank , which is equal to the generic complex rank , see ( * ? ? ?
* theorem 1 ) .
our goal is to construct explicit examples of curves where this gap is achieved .
these examples will be projections of the rational normal curve from suitably chosen points
. we will talk about the rank with respect to the rational normal curve in terms of the waring rank of binary forms ( equivalently symmetric tensor rank of bivariate tensors ) : we identify @xmath228 with the space @xmath229_d$ ] of binary forms of degree @xmath29 by associating the form @xmath230 to the point @xmath231 . in this identification , the rational normal curve in @xmath228
is sent to the set of powers of linear forms .
more precisely , the point @xmath232 is identified with the binary form @xmath233 . via this identification ,
the real rank of a point @xmath234 with respect to the rational normal curve is equal to the real waring rank of the associated binary form , i.e. the smallest @xmath9 such that the form is a sum of @xmath9 powers of order @xmath29 of real linear forms .
the advantage is that we can characterize the forms of maximal real rank in terms of their zeros .
we need some preparations .
we call a binary form @xmath235_d$ ] hyperbolic if all its roots are real , i.e. it splits into linear factors over @xmath74 .
reznick has shown ( * ? ? ?
* corollary 4.11 ) that a hyperbolic binary form has real waring rank @xmath29 , which is maximal . an almost complete converse of this statement was established by @xcite , who showed that any form of rank @xmath29 with distinct roots has all real roots .
we will prove the full converse , see ( * ? ? ?
* conjecture 4.12 ) via a simple inductive argument : [ thm : hyperbolic ] let @xmath235_d$ ] be a binary form of degree @xmath236 and suppose that @xmath237 is not a @xmath29-th power of a real linear form .
the real waring rank of @xmath237 is @xmath29 if and only if @xmath237 is hyperbolic . for the proof of this theorem
, we establish several useful facts about hyperbolic binary forms and state applications of the apolarity lemma .
before we do so , let us observe the following fact .
[ rem : ballicobound ] let @xmath238 be the rational normal curve of degree @xmath29 . then @xmath239 is odd and the maximal real rank @xmath29 with respect to @xmath0 is achieved by every hyperbolic polynomial .
this maximal rank agrees with ballico s upper bound @xmath240 in ( * ? ? ?
* theorem 1(ii ) ) .
let @xmath237 and @xmath241 be two real binary hyperbolic forms of degree @xmath29 .
we say that @xmath237 and @xmath241 interlace if there is a root of @xmath241 in every arc between two roots of @xmath237 in @xmath242 .
if @xmath237 and @xmath241 have common roots , then this condition is understood to mean that after dividing by the greatest common divisor the quotients @xmath243 and @xmath244 interlace in the sense above .
[ prop : interlace ] let @xmath237 and @xmath241 be two real binary hyperbolic forms of degree @xmath29
. then @xmath237 and @xmath241 interlace if and only if @xmath245 is hyperbolic for every @xmath246 .
we need the following characterization of hyperbolicity for binary forms . the same statement for forms with distinct roots
is proved in ( * ? ? ?
* theorem 1 ) .
[ prop : hyperbolic ] a binary form @xmath235_d$ ] of degree @xmath236 is hyperbolic if and only if all its directional derivatives @xmath247 , @xmath248 , are hyperbolic
. the proof will rely on a homotopy argument .
we first observe the following .
[ rem : zeros ] let @xmath237 be a real binary form of degree @xmath29 .
a. if @xmath249 and @xmath250 , then @xmath251 because @xmath252 .
b. suppose @xmath251 , @xmath249 , @xmath253 divides @xmath254 , and @xmath255 .
then @xmath256 divides @xmath237 , which follows from taylor expansion of @xmath257 around @xmath258 .
let @xmath237 be a hyperbolic form and @xmath248 .
after a change of coordinates and dehomogenizing , we can assume that @xmath237 is a polynomial in one variable of degree @xmath29 with only real roots and that the directional derivative @xmath259 is the usual derivative @xmath260 .
then the derivative has @xmath261 real roots by rolle s theorem ; so it is hyperbolic .
conversely , suppose all directional derivatives @xmath259 have only real roots and @xmath237 has at least one complex root .
first note that all complex roots of @xmath237 must be simple because a double root of @xmath237 is a root of every directional derivative @xmath259 .
we will perturb @xmath237 to arrive at a contradiction : first , we show that @xmath237 can have at most one complex conjugate pair of complex roots . after a change of coordinates , if necessary ,
let @xmath262 and @xmath263 be two distinct complex roots of @xmath237 with @xmath264 and @xmath265 . then @xmath266 .
assume that roots @xmath267 of @xmath268 satisfy @xmath269 .
let @xmath270 be the binary form with roots @xmath271 and @xmath272 and set @xmath273 .
then @xmath274 and @xmath275 has a double root at @xmath276 and @xmath277 .
all directional derivatives of @xmath278 have only real roots , which means that @xmath254 and @xmath279 interlace .
the same is true for the directional derivatives of @xmath280 for all @xmath281 $ ] : the partial derivatives @xmath282 and @xmath283 interlace because their zeros can not come together .
if they did , this would force a multiple real root of @xmath237 by remark [ rem : zeros ] because all roots @xmath267 of @xmath280 satisfy @xmath269 .
now @xmath275 has a double complex root and @xmath284 and @xmath285 have only real roots , a contradiction .
next we show that @xmath237 can not have a pair of complex conjugate roots by the same idea : after a change of coordinates , we can factor @xmath237 as @xmath286 where @xmath287 and @xmath288 .
we use the homotopy @xmath289 then @xmath274 and @xmath290 . since the partial derivatives of @xmath237 interlace , the same is true for the partial derivatives of @xmath275 by remark [ rem : zeros ] by the same argument as before .
in particular , all their roots have to be real , but this is false , because @xmath291 has a pair of complex conjugate roots ( @xmath236 ) .
so we have shown that all roots of @xmath237 are real , given that all its directional derivatives have only real roots .
we need the following application of the apolarity lemma , see ( * ? ? ?
* chapter i ) .
[ lem : apolar ] let @xmath292_{d-1}$ ] be a binary form of degree @xmath293 .
suppose that the real waring rank of @xmath294 is at most @xmath227 and that every decomposition of @xmath294 into the sum of at most @xmath227 powers of linear forms uses the power @xmath295 .
then @xmath294 is equal to @xmath295 up to scaling or @xmath296 for some @xmath297 and a linear form @xmath298 $ ] .
the apolar ideal @xmath299 $ ] of @xmath294 is generated by two polynomials @xmath300 $ ] such that @xmath301 , see ( * ? ? ? * theorem 1.44(iv ) ) .
we assume @xmath302 .
the fact that every decomposition of @xmath294 as the sum of at most @xmath227 powers of real linear forms uses @xmath295 is equivalent to the fact that every polynomial in @xmath303 with only real roots is divisible by @xmath47 by the apolarity lemma , ( * ? ? ? * lemma 1.31 ) .
the set of hyperbolic polynomials has non - empty interior in @xmath304_{d-2}$ ] ( see @xcite ) , which implies that every polynomial in @xmath303 is divisible by @xmath47 .
in particular , @xmath305 , because @xmath306 and @xmath307 have no common zeros .
this leaves the two cases @xmath308 , so that @xmath309 , or @xmath310 , so that @xmath311 . in case @xmath312
, @xmath306 must be equal to @xmath47 up to scaling . by the apolarity lemma
, @xmath294 has real waring rank @xmath30 in this case and is equal to @xmath295 up to scaling . in case @xmath313
, there is a linear form @xmath314 such that @xmath315 . in this case ,
@xmath316 , which shows that @xmath294 is equal to @xmath317 , where @xmath318 , for some @xmath319 .
consider the binary form @xmath320 , @xmath321 , for a linear form @xmath298 $ ] and @xmath322 .
then the real waring rank of @xmath237 is at most @xmath261 or @xmath237 is hyperbolic : by the structure theorem , the apolar ideal of @xmath237 is generated by two polynomials @xmath323 with @xmath324 , see ( * ? ? ?
* theorem 1.44(iv ) ) .
clearly , the apolar ideal contains the form @xmath325 , where @xmath326 is the linear form that is apolar to @xmath188 .
now there are three cases : if @xmath327 contains a linear form , then @xmath328 or @xmath329 , i.e. the waring rank of @xmath237 is @xmath30 .
suppose @xmath327 contains a form @xmath330 of degree @xmath208 , then @xmath330 divides @xmath331 , because @xmath332 .
therefore , @xmath333 or @xmath334 . if @xmath334 , then @xmath335 is hyperbolic .
if @xmath333 , then the real waring rank of @xmath237 is @xmath208 by the apolarity lemma .
thirdly , if @xmath327 does not contain a quadric , then @xmath336 and @xmath337 . now choose generic linear forms @xmath338 such that @xmath339 has only distinct roots except for @xmath340 , which is a double root of @xmath306 . by perturbing @xmath9 with a small multiple of @xmath307
, we obtain a polynomial in @xmath327 with distinct real roots of degree @xmath261 . by the apolarity lemma
, this implies that the real waring rank of @xmath237 is at most @xmath261 .
we are now ready to prove that all binary forms of maximal real waring rank and degree at least @xmath24 are hyperbolic .
we prove the statement by induction on the degree of @xmath237 : the base case @xmath341 was done by reznick , see ( * ? ? ?
* theorem 5.2 ) .
so let @xmath235_d$ ] be a hyperbolic binary form of degree @xmath342 .
then every directional derivative of @xmath237 is a hyperbolic form of degree @xmath261 . by induction , they all have real waring rank @xmath261 .
if the waring rank of @xmath237 was @xmath261 , then there would be a directional derivative with real waring rank at most @xmath227 .
so @xmath237 must have real waring rank @xmath29 .
for a different proof of this direction , see @xcite .
on the other hand , let @xmath237 be a binary form of degree @xmath342 and suppose that the real waring rank of @xmath237 is @xmath29 .
if , after a change of coordinates , @xmath343 or @xmath344 for some linear form @xmath298 $ ] and @xmath322 , then @xmath237 is hyperbolic ( see the preceding example for the second case ) . otherwise , every directional derivative of @xmath237 has real waring rank @xmath261 . indeed , if that were false , we can assume after a change of coordinates that @xmath254 has waring rank at most @xmath227 .
by lemma [ lem : apolar ] , there is a decomposition @xmath345 where the coefficient @xmath346 of @xmath47 in all linear forms @xmath347 is nonzero , because @xmath254 is not equal to @xmath348 . by taking antiderivatives ,
we get @xmath349 for some @xmath350 .
this equation implies that the real waring rank of @xmath237 is at most @xmath261 .
so every directional derivative of @xmath237 , which has degree @xmath261 , has real waring rank @xmath261 .
therefore , by induction , every directional derivative of @xmath237 is hyperbolic .
and then so is @xmath237 , by proposition [ prop : hyperbolic ] .
we can use this result on the real waring rank of binary forms to construct a curve such that the gap between the minimal typical real rank and the maximal typical rank is as large as possible .
we construct this curve by projecting the rational normal curve .
[ thm : maxtypicalrank ] let @xmath351_d$ ] be a binary hyperbolic form with distinct real roots .
let @xmath352 be the projection of the rational normal curve ( embedded into @xmath353_d)$ ] as the @xmath29-th powers of linear forms ) from the point @xmath7 .
then the minimal typical real rank with respect to @xmath0 is @xmath354 , whereas the maximal typical real rank with respect to @xmath0 is @xmath29 .
the minimal typical real rank with respect to @xmath352 is equal to the generic complex rank , which is @xmath355 by counting the dimension of the secant varieties of @xmath0 . indeed , @xmath356 the fact that the maximal typical real rank is @xmath29 follows from our above results on the waring rank : let @xmath357_d$ ] be any hyperbolic binary form that interlaces @xmath7 .
then every form in the span of @xmath7 and @xmath330 is hyperbolic by proposition [ prop : interlace ] .
so the image of @xmath330 under the projection from @xmath7 is a real point of rank @xmath29 with respect to @xmath0 by proposition [ prop : projections ] and theorem [ thm : hyperbolic ] .
as the set of interlacers of @xmath7 is open in the euclidean topology , this gives an open set of real points in @xmath358 of rank @xmath29 .
in fact , in our construction in the proof of theorem [ thm : maxtypicalrank ] , the set of all points of real @xmath0-rank @xmath29 is the projection of the set of all interlacing polynomials of the fixed strictly hyperbolic polynomial @xmath7 , which is a convex cone , cf .
* corollary 2.7 ) .
in particular , the set of points of real @xmath0-rank @xmath29 is connected . by choosing a form @xmath7 with repeated roots
, we can make the dimension of the set of points of real @xmath0-rank @xmath29 smaller ; e.g. if we set @xmath359 , then the dimension of bivariate forms interlacing @xmath7 is @xmath208 ( projectively ) .
[ cor : gapcomplexrank ] let @xmath360_d$ ] be a generic binary hyperbolic form with distinct real roots and let @xmath361 be the projection of the rational normal curve as above .
then the maximal complex @xmath0-rank is at most @xmath362 , whereas the maximal real rank is @xmath29 . the generic complex rank with respect to the rational normal curve @xmath363 is @xmath362 by count of dimensions of the secant varieties .
so if we project from a point @xmath7 with this rank , then the maximal rank in the image @xmath358 is at most this rank @xmath362 by proposition [ prop : projections ] .
the condition on the polynomial to be generic in the above corollary is well understood .
it is different for odd and even degrees .
if @xmath29 is even , then the condition is that the hankel matrix ( middle catalecticant ) of @xmath7 is non - singular . if @xmath29 is odd , the condition is slightly more complicated .
see comas - seiguer @xcite for details .
in case of plane curves , we have @xmath35 and @xmath147 , so the maximal real rank of any point is @xmath24 , the trivial bound .
this bound is tight .
our goal is to show that regions of points of rank @xmath24 are bounded by real flex lines to the curve , given that they have non - empty interior . given an irreducible nondegenerate plane curve @xmath364 with regular real point and a point @xmath365 with @xmath366 , the degree of @xmath0 must be odd and @xmath367 has no ovals : if @xmath367 has an oval , then the set of lines through the interior of that oval cover @xmath368 and intersect the oval in two real points .
[ thm : planecurves ] let @xmath364 be an irreducible nondegenerate curve and assume that @xmath367 is non - empty and contains only regular points of @xmath0 .
let @xmath370 be the union of all tangents to a real point of @xmath0 that meet @xmath0 at odd order @xmath371 .
then the set of points in @xmath372 of real @xmath0-rank @xmath24 is a union of connected components of @xmath372 ( possibly empty ) . in other words ,
the regions of points of real @xmath0-rank @xmath24 with non - empty interior are bounded by odd order tangents .
a point @xmath374 has real @xmath0-rank @xmath24 if and only if every line through @xmath7 intersects @xmath367 in at most @xmath30 point . as a condition in the line ,
this is open because @xmath367 is compact , as long as we stay away from odd order tangents . by compactness of the dual projective space @xmath375
, this also translates into an open condition on the point @xmath7 by duality . to see that the regions of points of real @xmath0-rank @xmath24 are closed in @xmath376
, we show that the complement is open .
a point @xmath377 of real @xmath0-rank less than @xmath24 either has rank @xmath208 or @xmath378 .
suppose @xmath379 , then there is a line through @xmath330 that intersects @xmath367 in @xmath208 distinct points .
we can assume that this line is not tangent to @xmath0 at any real point ( by choosing a generic line through @xmath330 ) .
since @xmath367 has no singular points , we cover an open neighbourhood of @xmath330 by varying one of these @xmath208 points in @xmath367 . if @xmath378 , then we can find a line intersecting @xmath367 transversally in @xmath330 and in at least @xmath208 more real points because @xmath0 has odd degree .
using the same argument as for the rank @xmath208 case , we see that all points in a neighbourhood of @xmath330 ( in @xmath376 ) have real @xmath0-rank at most @xmath208 .
a. for most curves , the only odd order tangents with an order of tangency at least @xmath24 are flex lines , i.e. tangents to @xmath0 at inflection points .
this is the case for all plcker curves . in this case , the only singularities of the dual plane curve are nodes and cusps .
b. in particular , we have seen in the proof of the above theorem , that all regions in @xmath380 that meet @xmath367 are regions of rank @xmath208 points .
a. consider the plane cubic @xmath381 .
this cubic has only one connected component in @xmath382 and three real flex lines .
they bound a triangle .
the points in the interior of this triangle have real rank @xmath24 with respect to @xmath0 , see figure [ fig : deg3 ] .
b. consider the plane curve @xmath383 of degree @xmath384 .
it has @xmath384 real flex lines .
the region of points of rank @xmath24 is connected with non - empty interior and is bounded by four flex lines .
it is shown in figure [ fig : deg5 ] .
+ , its @xmath384 real flex lines , and the region of points of real rank @xmath24 . ]
jacek bochnak , michel coste , and marie - franoise roy . , volume 36 of _ ergebnisse der mathematik und ihrer grenzgebiete ( 3 ) [ results in mathematics and related areas ( 3)]_. springer - verlag , berlin , 1998 . translated from the 1987 french original , revised by the authors .
david eisenbud and joe harris . on varieties of minimal degree ( a centennial account ) . in _ algebraic geometry , bowdoin , 1985 ( brunswick , maine , 1985 ) _ , volume 46 of _ proc .
sympos . pure math .
_ , pages 313 .
soc . , providence , ri , 1987 . | we study the real rank of points with respect to a real variety @xmath0 .
this is a generalization of various tensor ranks , where @xmath0 is in a specific family of real varieties like veronese or segre varieties .
the maximal real rank can be bounded in terms of the codimension of @xmath0 only .
we show constructively that there exist varieties @xmath0 for which this bound is tight .
the same varieties provide examples where a previous bound of blekherman - teitler on the maximal @xmath0-rank is tight .
we also give examples of varieties @xmath0 for which the gap between maximal complex and the maximal real rank is arbitrarily large .
to facilitate our constructions we prove a conjecture of reznick on the maximal real symmetric rank of symmetric bivariate tensors .
finally we study the geometry of the set of points of maximal real rank in the case of real plane curves . |
in systems with a single degree of freedom , constancy of the energy allows the momentum variable @xmath1 to be written in terms of the coordinate variable @xmath2 as @xmath3 , and the dependence of both variables on time follows immediately from hamilton s equations . in general systems with @xmath4 degrees of freedom ( dof ) ,
such a solution is generally not possible unless the hamilton - jacobi equation is separable , in which case the separation constants are isolating integrals of the motion .
an isolating integral is a conserved quantity that in some transformed coordinate system makes @xmath5 , thus allowing the motion in @xmath6 to be reduced to quadratures .
each isolating integral restricts the dimensionality of the phase space region accessible to an orbit by one ; if there are @xmath7 such integrals , the orbit moves in a phase space of dimension @xmath8 , and the motion is regular .
the @xmath7-dimensional phase space region to which a regular orbit is confined is topologically a torus ( figure 1 ) .
orbits in time - independent potentials may be either regular or chaotic , respecting a smaller number of integrals typically only the energy integral @xmath9 .
although chaotic orbits are not , strictly speaking , confined to tori , numerical integrations suggest that many chaotic trajectories are effectively regular , remaining confined for long periods of time to regions of phase space much more restricted than the full energy hypersurface .
the most compact representation of a regular orbit is in terms of the coordinates on the torus ( figure 1 ) the action - angle variables @xmath10 . the process of determining the map @xmath11 is referred to as _
torus construction_. there are a number of contexts in which it is useful to know the @xmath0 .
one example is the response of orbits to slow changes in the potential , which leave the actions ( @xmath12 ) unchanged .
another is the behavior of weakly chaotic orbits , which may be approximated as regular orbits that slowly diffuse from one torus to another .
a third example is galaxy modeling , where regular orbits are most efficiently represented and stored via the coordinates that define their tori .
this article reviews techniques for mapping cartesian coordinates into action - angle variables in non - integrable potentials .
two general approaches to this problem have been developed .
trajectory - following algorithms are based on the quasi - periodicity of regular motion : fourier decomposition of the trajectory yields the fundamental frequencies on the torus as well as the spectral amplitudes , which allow immediate construction of the map @xmath13 .
iterative approaches begin from some initial guess for @xmath14 , which is then refined via hamilton s equations with the requirement that the @xmath15 increase linearly with time .
the two approaches are often complementary , as discussed below .
in certain special potentials , every orbit is regular ; examples are the kepler and stckel potentials .
motion in such potentials can be expressed most simply by finding a canonical transformation to coordinates @xmath16 for which the hamiltonian is independent of @xmath17 , @xmath18 ; among all such coordinates , one particularly simple choice is the action - angle variables @xmath19 , in terms of which the equations of motion are @xmath20 ( landau & lifshitz 1976 ; goldstein 1980 ) .
the trajectory @xmath21 is periodic in each of the angle variables @xmath15 , which may be restricted to the range @xmath22 .
the @xmath23 define the cross - sectional areas of the torus while the @xmath15 define position on the torus ( figure 1 ) .
these tori are sometimes called `` invariant '' since a phase point that lies on a torus at any time will remain on it forever .
most potentials are not integrable , but regular orbits may still exist ; indeed these are the orbits for which torus construction machinery is designed .
one expects that for a regular orbit in a non - integrable potential , a canonical transformation @xmath24 can be found such that @xmath25 however there is no guarantee that the full hamiltonian will be expressible as a continuous function of the @xmath23 . in general , the map @xmath26 will be different for each orbit and will not exist for those trajectories that do not respect @xmath7 isolating integrals ( although approximate maps , valid for some limited span of time , may be derived for weakly chaotic trajectories ) . the uniform translation of a regular orbit on its torus implies that the motion in any canonical coordinates @xmath27 is quasi - periodic : @xmath28 , \nonumber \\
{ \bf v}(t ) & = & \sum_k{\bf v}_k({\bf j } ) \exp\left [ i\left(l_k\omega_1 + m_k\omega_2 + n_k\omega_3\right)t\right],\end{aligned}\ ] ] with @xmath29 integers .
the fourier transform of @xmath30 or @xmath31 will therefore consist of a set of spikes at discrete frequencies @xmath32 that are linear combinations of the @xmath7 fundamental frequencies @xmath33 , with spectral amplitudes @xmath34 and @xmath35 .
the most straightforward , and probably the most robust , approach to torus construction is via fourier analysis of the numerically - integrated trajectories ( percival 1974 ; boozer 1982 ; binney & spergel 1982 , 1984 ; kuo - petravic et al . 1983 ; eaker et al .
1984 ; martens & ezra 1985 ) .
the fourier decomposition of a quasiperiodic orbit ( equation 3 ) yields a discrete frequency spectrum .
the precise form of this spectrum depends on the coordinates in which the orbit is integrated , but certain of its properties are invariant , including the @xmath7 fundamental frequencies @xmath33 from which every line is made up , @xmath36 .
typically the strongest line in a spectrum lies at one of the fundamental frequencies ; once the @xmath33 have been identified , the integer vectors @xmath29 corresponding to every line @xmath37 are uniquely defined , to within computational uncertainties .
approximations to the actions may then be computed using percival s ( 1974 ) formulae ; e.g. the action associated with @xmath38 in a 3 dof system is @xmath39 and similarly for @xmath40 and @xmath41 , upon replacing the first factor in the summation by @xmath42 and @xmath43 respectively .
finally , the maps @xmath44 are obtained by making the substitution @xmath45 in the spectrum , e.g. @xmath46 \nonumber \\ & = & \sum_kx_k(j ) \exp\left [ i\left(l_k\theta_1 + m_k\theta_2 + n_k\theta_3\right)\right ] \nonumber \\ & = & x(\theta_1 , \theta_2 , \theta_3).\end{aligned}\ ] ] trajectory following algorithms are easily automated ; for instance , integer programming may be used to recover the vectors @xmath29 ( valluri & merritt 1998 ) .
binney & spergel ( 1982 ) pioneered the use of trajectory - following algorithms for galactic potentials .
they integrated orbits for a time @xmath47 and computed discrete fourier transforms , yielding spectra in which each frequency spike was represented by a peak with finite width @xmath48 centered on @xmath37 .
they then fitted these peaks to the expected functional form @xmath49/(\omega-\omega_k)$ ] using a least - squares algorithm .
they were able to recover the fundamental frequencies in a 2 dof potential with an accuracy of @xmath50 after @xmath51 orbital periods .
binney & spergel ( 1984 ) used equation ( 4 ) to construct the `` action map '' for orbits in a principal plane of the triaxial logarithmic potential .
carpintero & aguilar ( 1998 ) and copin , zhao & de zeeuw ( this volume ) applied similar algorithms to motion in 2- and 3 dof potentials .
the accuracy of fourier transform methods can be greatly improved by multiplying the time series with a windowing function before transforming . the result is a reduction in the amplitude of the side lobes of each frequency peak at the expense of a broadening of the peaks ; the amplitude measurements are then effectively decoupled from any errors in the determination of the frequencies .
laskar ( 1988 , 1990 ) developed this idea into a set of tools , the `` numerical analysis of fundamental frequencies '' ( naff ) , which he applied to the analysis of weakly chaotic motion in the solar system .
laskar s algorithm recovers the fundamental frequencies with an error that falls off as @xmath52 ( laskar 1996 ) , compared with @xmath53 in algorithms like binney & spergel s ( 1982 ) . even for modest integration times of @xmath54 orbital periods ,
the naff algorithm is able to recover fundamental frequencies with accuracies of @xmath55 or better in many potentials .
the result is a very precise representation of the torus ( figure 2 ) .
one drawback of trajectory - following algorithms is the need to extract a large number of terms in the frequency spectrum in cases where the time dependence of the integration variables is very different from that of the angles .
this problem may be dealt with by expressing the numerically - integrated orbit in terms of a set of coordinates that are closer to the angle variables before computing the fourier transform ; for instance , tube orbits are most efficiently expressed in the canonically - conjugate poincar variables ( related to cylindrical coordinates , e.g. papaphilippou & laskar 1996 ) .
trajectory - following algorithms also suffer from the fundamental limitation that they must follow the trajectory sufficiently far to see the longest periodicities of the orbit .
in other words , the trajectory must adequately sample the surface of its invariant torus .
near a resonant torus , i.e. a torus for which the @xmath33 satisfy a relation @xmath56 with the @xmath57 integers , trajectories fill their tori very slowly , necessitating long integration intervals . however even for near - resonant orbits
, one can still efficiently recover the terms in the spectrum associated with the `` faster '' angles , as well as a reasonable approximation to the `` slow '' frequency @xmath58 associated with libration around the resonant orbit ( merritt & valluri 1999 ) , and for many applications these are sufficient . if the precise dependence of the map on @xmath59 is also needed , one possible approach ( papaphilippou & laskar 1998 ) is to integrate the equations of motion using as a time step the period associated with one of the fast angles , thus eliminating it from the spectrum . since fourier techniques focus on the frequency domain , they are particularly well suited to identifying regions of phase space occupied by resonances .
they are also ideal for studying the effect of resonances on the structure of phase space , even in cases where the full tori are difficult to reconstruct .
resonant tori are places where perturbation expansions of integrable systems break down , due to the `` problem of small denominators '' . in perturbed ( non - integrable ) potentials ,
one expects stable resonant tori to generate regions of regular motion and unstable resonant tori to give rise to chaotic regions .
algorithms like naff allow one to construct a `` frequency map '' of the phase space : a plot of the ratios of the fundamental frequencies @xmath60 for a large a set of orbits selected from a uniform grid in initial condition space .
resonances appear on the frequency map as lines , either densely filled lines in the case of stable resonances , or gaps in the case of unstable resonances ; the frequency map is effectively a representation of the arnold web ( laskar 1993 ) .
papaphilippou & laskar ( 1996 , 1998 ) , wachlin & ferraz - mello ( 1998 ) and valluri & merritt ( 1998 ) used frequency maps to study the effect of resonances on the structure of phase space in triaxial potentials .
precisely resonant orbits can be reconstructed using trajectory - following algorithms in a particularly straightforward way .
a resonance has the effect of restricting an orbit to a subset of its torus , reducing the number of independent angle variables by one ; thus a 2 dof trajectory is reduced to a closed curve and a 3 dof trajectory becomes a thin sheet ( born 1960 ; goldstein 1980 ; figure 3 ) .
the two frequencies defining motion on a resonant three - torus may be taken to be @xmath61 in terms of which @xmath62 this definition is not unique since the orbit is not closed .
the motion in cartesian coordinates then becomes @xmath63 t \nonumber \\ &
= & \sum_k { \bf x}_k\exp i \left({l_k}'\omega_0^{(1 ) } + { m_k}'\omega_0^{(2)}\right)t \nonumber \\ & = & \sum_k { \bf x}_k\exp i \left({l_k}'\theta^{(1 ) } + { m_k}'\theta^{(2)}\right ) \nonumber \\ & = & { \bf x}(\theta^{(1 ) } , \theta^{(2)}).\end{aligned}\ ] ] the result is a set of parametric expressions for the cartesian coordinates @xmath64 in terms of the angles @xmath65 that define position on the two - torus ( figure 4 ) .
iterative approaches to torus construction consist of finding successively better approximations to the map @xmath66 given some initial guess @xmath14 ; canonical perturbation theory is a special case , and in fact iterative schemes often reduce to perturbative methods in appropriate limits .
iterative algorithms were first developed in the context of semi - classical quantization for computing energy levels of bound molecular systems , and they are still best suited to assigning energies to actions , @xmath67 .
most of the other quantities of interest to galactic dynamicists
e.g. the fundamental frequencies @xmath33 are not recovered with high accuracy by these algorithms .
iterative schemes also tend to be numerically unstable unless the initial guess is close to the true solution . on the other hand
, iterative algorithms can be more efficient than trajectory - following methods for orbits that are near resonance .
ratcliff , chang & schwarzschild ( 1984 ) pioneered iterative schemes in galactic dynamics .
they noted that the equations of motion of a 2 dof regular orbit , @xmath68 can be written in the form @xmath69 if one specifies @xmath70 and @xmath71 and treats @xmath72 and @xmath73 as functions of the @xmath15 , equations ( 11 ) can be viewed as nonlinear differential equations for @xmath74 and @xmath75 .
ratcliff et al .
expressed the coordinates as fourier series in the angle variables , @xmath76 substituting ( 12 ) into ( 11 ) gives @xmath77 where the right hand side is again understood to be a function of the angles .
ratcliff et al .
truncated the fourier series after a finite number of terms and required equations ( 13 ) to be satisfied on a grid of points around the torus .
they then solved for the @xmath78 by iterating from an initial guess .
convergence was found to be possible if the initial guess was close to the exact solution .
a similar algorithm was developed for recovering tori in the case that the actions , rather than the frequencies , are specified a priori .
guerra & ratcliff ( 1990 ) applied these algorithms to motion in the plane of rotation of a nonaxisymmetric potential .
another iterative approach to torus construction was developed by chapman , garrett & miller ( 1976 ) in the context of semiclassical quantum theory .
one begins by dividing the hamiltonian @xmath79 into separable and non - separable parts @xmath80 and @xmath81 , then seeks a generating function @xmath82 that maps the known tori of @xmath80 into tori of @xmath79 . for a generating function of the @xmath83-type ( goldstein 1980 )
, one has @xmath84 where @xmath10 and @xmath85 are the action - angle variables of @xmath80 and @xmath79 respectively .
the generator @xmath82 is determined , for a specified @xmath86 , by substituting the first of equations ( 14 ) into the hamiltonian and requiring the result to be independent of @xmath87 .
one then arrives at @xmath88 .
chapman et al . showed that a sufficiently general form for @xmath82 is @xmath89 where the first term is the identity transformation , and they evaluated a number of iterative schemes for finding the @xmath90 .
one such scheme was found to recover the results of first - order perturbation theory after a single iteration .
mcgill & binney ( 1990 ) refined the chapman et al .
algorithm and applied it to 2 dof motion in the axisymmetric logarithmic potential .
the generating function approach is not naturally suited to deriving the other quantities of interest to galactic dynamicists .
for instance , equation ( 14 ) gives @xmath91 as a derivative of @xmath82 , but since @xmath82 must be computed separately for every @xmath86 its derivative is likely to be ill - conditioned .
binney & kumar ( 1993 ) and kaasalainen & binney ( 1994a ) discussed two schemes for finding @xmath91 ; the first requires the solution of a formally infinite set of equations , while the latter requires multiple integrations of the equations of motion for each torus effectively a trajectory - following scheme .
kaasalainen & binney ( 1994a ) noted that the success of the generating function method depends strongly on the choice of @xmath80 . for box orbits , which are most naturally described as coupled rectilinear oscillators
, they found that a harmonic - oscillator @xmath80 gave poor results unless an additional point transformation was used to deform the rectangular orbits of @xmath80 into narrow - waisted boxes like those in typical galactic potentials .
kaasalainen ( 1995a ) considered orbits belonging to higher - order resonant families and found that it was generally necessary to define a new coordinate transformation for each family . warnock ( 1991 ) presented a hybrid scheme in which the generating function @xmath82 was derived by numerically integrating an orbit from appropriate initial conditions , transforming the coordinates to @xmath10 of @xmath80 and interpolating @xmath12 on a regular grid in @xmath87 .
the values of the @xmath90 then follow from the first equation of ( 14 ) after a discrete fourier transform .
kaasalainen & binney ( 1994b ) found that warnock s scheme could be used to substantially refine the solutions found via their iterative algorithm .
another hybrid scheme was discussed by reiman & pomphrey ( 1991 ) .
having computed the energy on a grid of @xmath86 values , one can interpolate to obtain the full hamiltonian @xmath88 . if the system is not in fact completely integrable , this @xmath79 may be rigorously interpreted as smooth approximation to the true @xmath79 ( warnock & ruth 1991 , 1992 ) and can be taken as the starting point for secular perturbation theory . kaasalainen ( 1994 ) developed this idea and showed how to recover accurate surfaces of section in the neighborhood of low - order resonances in the planar logarithmic potential .
percival ( 1977 ) described a variational principle for constructing tori .
his technique has apparently not yet been implemented in the context of galactic dynamics .
torus - construction machinery may be applied to orbits that are approximately , but not precisely , regular ( laskar 1993 ) .
the frequency spectrum of a weakly chaotic orbit will typically be close to that of a regular orbit , with most of the lines well approximated as linear combinations of three `` fundamental frequencies '' @xmath33 .
however these frequencies will change with time as the orbit diffuses from one `` torus '' to another .
the diffusion rate can be measured via quantities like @xmath92 , the change in a `` fundamental frequency '' over two consecutive integration intervals .
papaphilippou & laskar ( 1996 , 1998 ) , valluri & merritt ( 1998 ) and wachlin & ferraz - mello ( 1998 ) used this technique to study chaos and diffusion in triaxial galactic potentials .
kaasalainen ( 1995b ) showed that approximate tori could be constructed even in chaotic phase space via the hybrid scheme of warnock ( 1991 ) .
while such tori clearly do not describe the motion of chaotic orbits over long times , they are useful for understanding the onset of chaos and its relationship to resonances , as well as for studying evolution of the phase - space distribution function in action space via the fokker - plank equation ( lichtenberg & leiberman 1992 ) .
trajectory - following schemes for torus construction are robust and easily automated
. they can recover the fundamental frequencies with great precision and are well suited to studies of weak chaos and for mapping resonances .
however they are inefficient for constructing the full torus of an orbit that lies close to , but slightly off of , a resonance .
iterative techniques are efficient for assigning energies to actions but less suited to recovering the other quantities of interest to galactic dynamicists , such as the fundamental frequencies
. however they can be more efficient than trajectory - following algorithms for constructing nearly - resonant tori .
hybrid schemes that combine features of both approaches show considerable promise . | the maximally compact representation of a regular orbit is in terms of its action - angle variables @xmath0 .
computing the map between a trajectory s cartesian coordinates and its action - angle variables is called torus construction .
this article reviews various approaches to torus construction and their application to galactic dynamics . |
during the last decades , the interplay between string theory , the topological string and supersymmetric gauge theories has led to fascinating developments in mathematics as well as in physics .
for instance , from the mathematical point of view , the observables in topological string theory can usually be understood in terms of topological invariants which acquire a physical meaning in string theory .
indeed , considered as a twisted version of the latter , the topological string captures properties of string theory that depend on its bps sector only and can thus be viewed as a subsector of string theory . in order to make this connection more precise , consider a calabi - yau compactification of type ii string theory and perform the topological twist on the underlying superconformal algebra ( sca ) , _ i.e. _ shifting the energy - momentum tensor with the @xmath4 current as @xmath5 ( and similarly for the right - movers ) .
the resulting theory can be shown to be topological @xcite .
on the other hand , the twisting of the sca can be implemented at genus @xmath6 in type ii through a deformation of its @xmath7-model by the operator @xmath8 , see @xcite , where @xmath9 is the two - dimensional scalar curvature and @xmath10 is the free scalar bosonising @xmath11 , _
i.e _ @xmath12 . by choosing a metric such that @xmath13 , it is easy to see that the partition function in the twisted , topological theory can be calculated in the untwisted , type ii theory as an amplitude involving @xmath14 insertions of vertex operators whose internal part is given by @xmath15 .
the latter is identified as the internal part of the graviphoton vertex operator .
hence , it turns out that the genus @xmath6 topological string partition function @xmath16 computes the coupling coefficients of a class of higher - derivative f - terms in the string effective action at genus @xmath6 @xcite , given by @xmath17 with @xmath18 being the anti - self - dual riemann tensor and @xmath19 the anti - self - dual part of the graviphoton field strength .
they originate from the supersymmetric f - term involving the weyl multiplet @xmath20 , @xmath21 where @xmath22 are chiral vector multiplets with scalar component @xmath23 . from the low - energy perspective ,
one of the uses of topological string theory is to geometrically engineer supersymmetric gauge theories in four dimensions @xcite .
for example , consider the local @xmath24 model arising from a type iia compactification on an elliptically fibered @xmath25 over @xmath26 , which is used to engineer the @xmath27 , @xmath28 gauge theory of seiberg and witten . the genus zero partition function of this model , @xmath29 , coincides , in the field theory limit , with the prepotential of the supersymmetric gauge theory .
in addition , @xmath30 can be interpreted as gravitational corrections to the seiberg - witten prepotential encoded in the f - terms .
in fact , one can understand these corrections gauge - theoretically through the @xmath27 gauge theory in the @xmath0-background @xcite .
more specifically , it was shown that the seiberg - witten solution can be obtained by a direct path integral evaluation if one introduces an @xmath31-fibration over @xmath32 such that going around the circle is accompanied by rotations in the two @xmath33-planes of space - time and an @xmath34-symmetry rotation parametrised by @xmath35 and @xmath36 respectively .
the latter is crucial to preserve a fraction of supersymmetry and evaluate the path integral using localisation .
similarly to the seiberg - witten case , the @xmath0-deformed prepotential receives a one - loop correction and a series of instanton contributions only : @xmath37 and its logarithm is known as the nekrasov partition function .
it was shown that in the limit @xmath38 , the leading term in matches the seiberg - witten solution and , for @xmath39 , is the field theory limit of the topological string partition function , with the topological string coupling being identified with the parameter @xmath39 .
this opened up the possibility of ` refining ' the topological string through a one - parameter extension of its partition function such that its field theory limit matches the gauge theory partition function in a general @xmath0-background . from the worldsheet point of view of the untwisted theory
, this would correspond in to additional insertions of self - dual field strenghts .
one of the main difficulties in achieving this is to simultaneously realise the @xmath34-symmetry rotation at the worldsheet level using physical operators .
even though the existence of such objects is _ a priori _ not guaranteed , in what follows we show that this is indeed the case by reviewing the results of @xcite . to date
, most descriptions of the refinement do not follow a worldsheet approach .
for instance , the refined topological partition function can be defined as a bps index in m - theory counting spinning m2-branes @xcite and evaluated using the refined topological vertex techniques @xcite .
similarly to the instanton calculus in gauge theory , the latter is applicable in the asymptotic region of the moduli space only .
a worldsheet realisation of the refined topological string is therefore essential in order to define it at any point in the moduli space . in this perspective
, we consider a ( compact ) calabi - yau compactification of type ii theory admitting a heterotic dual and identify the additional self - dual insertions .
even though we consider the compact calabi - yau case , it is worth mentioning that our primary goal is to identify the full @xmath0-background in a physical string theory setup .
furthermore , we expect that the couplings we calculate define a bps index in type ii theory once the heterotic / type ii duality map is applied and the appropriate non - compact / local limit is taken . in this note
, we review the class of generalised f - terms extending introduced in @xcite and motivate the identification of the additional insertions .
then , in section [ deformedtopoamp ] , we discuss the evaluation of these couplings and their connection to the gauge theory partition function .
section [ sec : conclusions ] highlights some of the open questions and directions for future investigation .
consider a five - dimensional @xmath1 gauge theory on @xmath40 and define its partition function as the trace over the four - dimensional hilbert space of the theory@xcite @xmath41 where @xmath42 is the fermion number , @xmath43 are the cartan generators of the @xmath44 subgroups the euclidean rotation group @xmath45 of @xmath46 , @xmath47 is the cartan generator of the @xmath28 @xmath34-symmetry and the ` inverse temperature ' @xmath48 is identified with the radius of @xmath31 .
this trace may be equivalently described as a vacuum amplitude on a deformed melvin - like background , @xmath49 , with @xmath22 being the @xmath31 coordinate .
this space , known as the @xmath0-background , is a fibration of @xmath46 over @xmath31 , with the @xmath46-fibre being twisted by isometry operations encoded into two constant rotation parameters @xmath35 , through @xmath50 , with the identification @xmath51 . in order to motivate the definition of the generalised f - term couplings of @xcite , it is instructive to obtain a deeper understanding of the relation between the string amplitudes @xmath52 and the nekrasov partition function for @xmath53 .
the connections are summarised in the table below . @xmath54
consider the scattering of two gravitons and @xmath14 graviphotons in type ii string theory on a calabi - yau threefold .
this amplitude first appears at genus @xmath6 and , being topological , receives no perturbative or non - perturbative corrections , consistently with @xmath1 super - renormalisation theorems .
let us focus on the perturbative part of the @xmath16 . using type ii - heterotic duality ,
it may be obtained by evaluating a one - loop amplitude of the same coupling in the dual heterotic theory on @xmath55 in the weak coupling limit . complexifying the space - time directions as @xmath56
, we write the relevant vertex operators in a kinematic configuration consistent with the projection on the anti - self - dual components of the corresponding tensors : @xmath57 let us briefly outline the non - technical aspects of the argument .
the effect of the graviton insertions is to soak up the zero modes of the worldsheet fermions in the space - time directions @xmath58 .
the compexified worldsheet fermions in the @xmath59 direction @xmath60 do not contract and their one - loop contribution cancels against that of the superghost . working in the orbifold limit of @xmath61 , the twisted one - loop contribution of the @xmath25 fermions , cancels against the ( left - moving ) twisted bosonic @xmath25 contribution and the full correlator has the form of a bps amplitude involving a non - trivial bosonic correlator , weighted by the partition function of the theory which produces the elliptic genus .
since only the complexified @xmath59 coordinate @xmath22 appears in the correlator , but never its complex conjugate , it leads to no contractions but , rather , contributes only through left - moving zero modes @xmath62 . in the hamiltonian representation of the lattice , this is replaced by left - moving lattice momentum insertions @xmath63\tau_2 $ ] , with @xmath64 being the khler and complex structure moduli of @xmath59 , respectively , and @xmath65 being the complex structure of the worldsheet torus .
the bosonic correlator then becomes @xmath66 where we suppress the integrals over the positions @xmath67 on the worldsheet torus .
multiplying this correlator by @xmath68 , stripping off its kinematical factors and summing over @xmath6 expresses the bosonic correlator in terms of a generating function @xmath69\bigr\rangle \equiv \langle e^{-\tilde\epsilon_{-}j_{-}}\rangle\,,\end{aligned}\ ] ] where @xmath70 and @xmath71 is the string generator of @xmath72 rotations , defining a map from the worldsheet to target space . in the field theory limit @xmath73 ,
the dependence on @xmath74 disappears and @xmath75 counts the @xmath72 charges of the bps states becoming massless as @xmath76 .
indeed , at the self - dual @xmath77 point in moduli space , the gauge group arising from @xmath59 acquires an @xmath28 enhancement and the modular integral over @xmath74 is dominated by the field theory contribution .
thus , the bosonic correlator only contributes through its zero mode : @xmath78 and one obtains the result in the form of a one - loop schwinger integral @xmath79 with @xmath80 being the bps mass parameter .
therefore , the physical picture is that the scattering of graviphotons , upon exponentiation as in , generates a deformed background which precisely mimics the @xmath0-background in for @xmath53 . inspired by the gopakumar - vafa reformulation of the topological string , the additional insertions should correspond to self - dual field strength vertices @xmath81 in the anti - chiral vector multiplet @xmath82 such that they couple to the spins of the @xmath83 component of @xmath45 , as required by .
the known way of introducing self - dual field strength insertions to the graviphoton couplings is by means of a generalised f - term@xmath84 with @xmath85 being an @xmath1 chiral projection may be pulled in front of @xmath86 in order to modify the grassmann integration measure and recast the coupling in the form of a d - term .
] of an arbitrary function @xmath87 of both chiral and anti - chiral vector multiplets .
we now exploit the relation between the deformed ( exponentiated ) background and the trace representation of the gauge theory partition function in the @xmath0 background in order to motivate the specific choice of species @xmath82 of the self - dual vector insertions . indeed , for the bosonic fields
, we would like to obtain a coupling of @xmath88 to @xmath75 and @xmath89 to the analogous @xmath83 generator . _ a priori _ , there are two possible choices : @xmath90 depending on whether the @xmath83 current arises from the left- or the right- moving part of the self - dual vertex operator insertions . in the corresponding correlators ,
@xmath91 is accompanied by @xmath92 , whereas @xmath93 is accompanied by @xmath94 .
notice that we do not allow @xmath95 or @xmath96 insertions , since that would spoil the requirement of exact solvability of our amplitude .
indeed , the factorisation of the full correlator into its bosonic and fermionic parts and the exact evaluation of the exponentiated background as a gaussian path integral , yielding a correlator with good modular properties , is only possible in the absence of contractions in the @xmath59 coordinates .
consider first the @xmath91 current .
since it is accompanied by a left - moving @xmath97 , the corresponding vertex operator is of the form @xmath98 .
hence , the left - moving part of the vertex is very similar to the graviphoton one and the resulting amplitude is bps at the one - loop level in the heterotic theory .
one can show that the self - dual projection of this vertex operator sits in the anti - chiral dilaton multiplet @xmath99 .
this amplitude was first computed in @xcite and its type ii interpretation was further studied in @xcite . however , as was the case with the graviphoton vertices , there is no non - trivial contribution from the worldsheet fermions . as a result , it is impossible to correctly mimic the @xmath34-symmetry twist and the resulting coupling function , in the field theory limit , fails to reproduce the correct ( perturbative ) gauge theory partition function . on the other hand , the situation is drastically different if one considers instead the @xmath93 current .
since this is left - moving , it is accompanied by the right moving @xmath100 current and the structure of the self - dual vertex operator insertions takes the form @xmath101 it can be checked that the corresponding bosonic correlator indeed exponentiates to @xmath102\rangle$ ] , as required .
however , notice that the self - dual vertex operators may also contribute through their fermionic bilinears parts , @xmath103 and @xmath104 .
this opens up the possibility of generating the correct @xmath34-symmetry twist , but only in an effective way at the string level since the internal @xmath55 space is compact .
indeed , the operator rotating the supercharges is written in terms of the complex @xmath25 fermions @xmath105 and , in a compact space , rotations in the @xmath34-symmetry space may only be defined for quantised angles ( the corresponding bosonic operator is not a well - defined conformal field ) . in the present case , it turns out that there is a way to obtain an effective @xmath34-symmetry twist after summing over the spin structures . in order to show this ,
suppose that @xmath106 self - dual vertices at positions @xmath107 contribute their fermion bilinear piece : @xmath108 where @xmath109 is a short - hand notation for @xmath110 and @xmath111 , respectively , and @xmath112 are the positions of the graviton vertices .
the r.h.s . displays the one - loop jacobi @xmath113-functions at spin structure @xmath114 .
the first two correspond to the space - time directions , while the last two correspond to the @xmath25 fermionic directions , twisted by the orbifold action @xmath115 .
summing over the spin structures with the conventional riemann summation identity , the r.h.s . becomes a product of correlators in the odd spin structure : @xmath116 the fermions in the space - time directions soak up zero modes and yield an @xmath117-independent contribution .
however , the @xmath25 fermion correlator is non - trivial and exponentiates to @xmath118 , with @xmath119 .
this is precisely the operator rotating the supercharges and is identified with the @xmath34-symmetry current in the field theory limit .
hence , the choice of self - dual vertices correctly mimics the @xmath0-background in the trace representation and is guaranteed to produce the perturbative part of the nekrasov partition function in the field theory limit .
this limit is obtained by expanding the amplitude around a wilson line enhancement point , where both @xmath120 , as a consequence of the fact that @xmath121 are dressed with @xmath62 and @xmath122 .
it was shown in @xcite that the vertices correspond to the self - dual field - strengths in the vector multiplet of the khler modulus @xmath123 of the heterotic @xmath59 .
in this section , we review the results of @xcite for the calculation of the effective couplings with insertions of the self - dual @xmath124-vectors . they receive contributions from perturbative and non - perturbative states as depicted in fig .
[ fig - pertinst ] .
the perturbative calculation is presented in heterotic on @xmath55 in which @xmath25 is realised , for convenience , as a @xmath125 orbifold , even though the results are expected to hold for a generic @xmath25 compactification . on the other hand ,
the evaluation of the instanton effective action is shown in the dual type i setup since the latter is more suitable for realising instantons in terms of d - brane bound states @xcite .
receive contributions from both perturbative and instanton sectors as presented below.,scaledwidth=75.0% ] [ [ perturbative - results ] ] perturbative results + + + + + + + + + + + + + + + + + + + + as explained in the previous section , due to the fact that the @xmath59 coordinates @xmath126 appear holomorphically in the vertex operators , the full correlation function splits into a bosonic and a fermionic one .
they can both be exponentiated and thus calculated as deformed partition functions .
first of all , the bosonic generating function takes the form @xmath127 \right>~,\end{aligned}\ ] ] such that the relevant bosonic correlators contributing to @xmath128 are obtained by an appropriate number of derivatives with respect to @xmath121 . notice that has the structure of the @xmath0-deformation in space - time .
it can be explicitely evaluated as a gaussian path integral to yield @xcite @xmath129\,,\label{fullregbos}\end{aligned}\ ] ] where @xmath130 is the non - holomorphic eisenstein series of weight @xmath131 defined as @xmath132 similarly , the fermionic correlation functions can be determined , after the spin structure sum , via the generating function @xmath133(\epsilon_+)=\left\langle e^{-\epsilon_+\int ( \chi^4\chi^5-\bar{\chi}^4\bar{\chi}^5)\bar{\partial}x}\right\rangle_{h , g}\end{aligned}\ ] ] parametrised by @xmath134 , with @xmath115 labelling the orbifold sectors and the sum over @xmath6 enforcing the orbifold projections .
it carries the structure of an @xmath34-symmetry twist as in the gauge theory partition function . the gaussian integral yields @xmath135(\check\epsilon_+)=\frac{\theta[^{1+h}_{1+g}](\check\epsilon_+ ; \tau)\theta[^{1-h}_{1-g}](\check\epsilon_+ ; \tau ) } { \eta^2}~ e^{\frac{\pi}{\tau_2}\check\epsilon_+^2}~.\ ] ] the full amplitude
can then be written by including also the internal and gauge degrees of freedom @xcite .
in particular , the coupling function can be obtained as the @xmath136-term in the @xmath137-expansion of @xmath138 .
we now expand the full amplitude around an @xmath28 wilson line enhancement point characterised by @xmath139 , with @xmath140 being the complexified wilson line of @xmath59 along the @xmath141 directions .
as expected , the field theory limit of the generating function for the refined couplings is @xmath142 in terms of the bps mass parameter @xmath143 .
in particular , it exhibits the correct singularity behaviour as @xmath144 .
hence , we recover the perturbative part of the partition function of the @xmath1 gauge theory in the @xmath0-background , for an @xmath28 gauge group without flavours .
more general gauge groups can be probed by studying different enhancement points as explained in @xcite .
[ [ deformed - adhm - action - from - string - theory ] ] deformed adhm action from string theory + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we now turn to the non - perturbative part of the refined couplings @xmath145 as displayed in the second diagram of fig .
[ fig - pertinst ] and review the results of @xcite .
the difference with the perturbative calculation lies in the choice of the states running in the ` loop ' . in other words
, we are interested in the instanton effective action in the presence of the background of anti - self - dual graviphotons and self - dual @xmath124-vectors . integrating out the instanton moduli in the latter leads to the non - perturbative contributions to @xmath145 .
this extends the results of @xcite where it was shown that the background of graviphotons calculates the unrefined adhm partition function , _
i.e. _ for @xmath146 . as explained above , we work in the type i dual of the heterotic setup .
the graviphoton is mapped to itself and the @xmath124-vectors to the @xmath147-vectors , @xmath148 is identified with the d5-brane coupling . ] . here
, we focus on the leading order in @xmath149 of the non - perturbative corrections .
for this , we realise yang - mills instantons as d5-d9 bound states with the d5-instantons wrapping @xmath55 . in this context , the massless states arising from open strings with at least one endpoint lying on a d5-instanton are in one - to - one correspondence with the adhm moduli vector multiplet . ] .
therefore , the strategy is to calculate the tree - level couplings between the adhm moduli and the background of graviphotons and @xmath147-vectors to leading order in @xmath149 .
some of these also involve the @xmath1 vector multiplet playing the role of the coulomb branch parameter . ] . from the string theory perspective , these couplings are calculated by disc diagrams and are three- and four - point functions ( higher point functions are suppressed in the field theory limit ) . compared to the heterotic theory , the main difference in the closed string sector lies in the form of the vertex operators of interest .
more specifically , they have an ns - ns and a r - r part : @xmath150\,,\label{graviphotonvertex}\\ v^{f^{\bar{s}'}}(y,\bar y)&=\frac{1}{8\pi\,\sqrt{2}}f_{\mu\nu}^{\bar{s } ' } \bigg[\,\psi^{\mu}\psi^{\nu}(y)e^{-\varphi(\bar y)}{\bar\psi}(\bar y)\,+e^{-\varphi(y)}\bar\psi(y)\psi^{\mu}\psi^{\nu}(\bar y)\nonumber\\ & + \frac{i}{2}\ , e^{-\tfrac{1}{2}(\varphi(y)+\varphi(\bar y ) ) } s_{\dot{\alpha}}(y ) ( \bar{\sigma}^{\mu\nu})^{\dot{\alpha}\dot{\beta } } s_{\dot{\beta}}(\bar y)\,\epsilon_{\hat a\hat b}\,s^{\hat a}(y)s^{\hat b}(\bar y)\bigg]\,.\label{tbarvector}\end{aligned}\ ] ] here , @xmath151 are space - time spin - fields whereas @xmath152 are internal ones .
it turns out that both parts of the vertex operators lead to the same coupling to the adhm action for the @xmath147-vector whereas the r - r part of the graviphoton couples with an additional minus sign @xmath153 with respect to its ns - ns part , and this is due to its anti - self - duality .
as dictated by the perturbative interpretation of the couplings , we identify the constant anti - self - dual field strength of the graviphotons with @xmath88 and the self - dual field strength of @xmath154 with @xmath89 .
therefore , both field strenghts lead to non - trivial couplings to the adhm action which we now recall : @xmath155\left([\chi , a^{\dot\beta\alpha}]+\epsilon_{-}(a\tau_3)^{\dot\beta\alpha}\right)+\epsilon_+\left[\chi^{\dag},a_{\alpha\dot{\beta}}\right](\tau_3 a)^{\dot\beta\alpha}\right.\nonumber\\ & \left.-\chi^{\dag}\,\bar\omega_{\dot\alpha}\left(\omega^{\dot\alpha}\chi-\tilde a\,\omega^{\dot\alpha}\right)-\left(\chi\bar\omega_{\dot\alpha}-\bar\omega_{\dot\alpha}\,\tilde a\right)\omega^{\dot\alpha}\,\chi^{\dag}-\epsilon_+\,\bar{\omega}_{\dot{\alpha}}\,{(\tau_3)^{\dot{\alpha}}}_{\dot{\beta}}\,\chi^{\dag}\,\omega^{\dot{\beta}}\right\}\,,\end{aligned}\ ] ] with @xmath156 , @xmath157 , @xmath158 and @xmath159 being the bosonic adhm moduli .
this is precisely the @xmath0-deformed part of the adhm action used to calculate the ( non - perturbative ) nekrasov partition function using localisation - independent parts of the yang - mills and adhm actions can be recovered in a similar fashion from the same string theory setup , see _
e.g. _ @xcite . for a treatment without auxiliary fields
, we refer to @xcite . ] .
consequently , the instanton corrections to the refined couplings involving the @xmath147-vectors also reproduce the non - perturbative part of the @xmath27 gauge theory partition function in the @xmath0-background . it is interesting to perform the same analysis for the coupling involving the vector partner of the dilaton . indeed
, its vertex operator can be obtained from by changing the relative sign between the ns - ns and the r - r parts .
hence , all the couplings to the adhm moduli involving the self - dual @xmath160-vector are trivial and can not lead to the correct gauge theory partition function upon evaluation of the instanton path integral .
we have seen that string amplitudes corresponding to the higher derivative @xmath161 terms in the string effective action , expanded around an appropriate wilson line enhancement point in moduli space , correctly reproduce the partition function of the @xmath1 gauge theory @xmath162 , perturbatively as well as non - perturbatively .
these couplings provide the desired one - parameter extension of the well - known @xmath1 topological couplings @xmath163 , correctly realising the full @xmath0-background .
one may eventually ask what the criterion determining this particular choice of self - dual vertices is .
a partial answer is based on the arguments presented in section [ motivation ] .
indeed , in order to obtain the correct perturbative nekrasov partition function in the field theory limit , the scattering background must properly exponentiate to correctly mimic the @xmath0-background .
recall that a very drastic constraint to our considerations was the requirement of exact solvability of our amplitude .
this essentially boils down to imposing a holomorphic dependence on the @xmath59 coordinate in choosing the self - dual vertex operators .
nevertheless , the present amplitude is only a single element in a family of amplitudes that lie in the same t - duality orbit .
in fact , t - duality transformations that preserve the left - moving structure of the vertex operators also change the enhancement point .
they typically lead to combinations of gauge fields involving also the @xmath141 gauge bundle , but all share the same field theory limit .
we close this discussion with a conjecture .
namely , that it is unlikely for an exactly solvable ( modular ) , truly bps amplitude to give rise to the correct nekrasov partition function , unless it is defined through the present non - bps amplitudes by taking an appropriate limit in moduli space , projecting to the bps sector .
the reason is that it seems impossible to implement the fermionic @xmath34-symmetry twist in a modular invariant fashion , without spoiling the bps nature of the amplitude .
the same conclusion can be reached from the very nature of the higher derivative couplings , _
i.e. _ that the generalised f - terms are not super - protected against non - bps corrections . in this sense , we conjecture that the ansatz is unique , modulo the above mentioned images under t - duality .
as pointed out above , from the string perspective , the couplings we considered also receive contributions from non - bps states and are not truly topological . on the other hand , in some appropriate decompactification limit , we expect the geometric details of the theory to be washed away and the resulting couplings , when properly mapped into the type ii side using the duality dictionary , to be expressed in terms of topological invariants . this expectation is consistent with the fact that preliminary results indicate the presence of differential equations satisfied by our couplings which generalise the standard holomorphic anomaly equations of the @xmath16 s @xcite .
we believe that future work in this direction will provide interesting insights into establishing this connection .
it is a great pleasure to thank i. antoniadis , s. hohenegger and k.s .
narain for the enjoyable collaboration on which these results are based .
we would also like to thank the organisers of the corfu summer institute 2013 and the workshop on noncommutative field theory and gravity " for the opportunity to present this work .
this work was partly supported by the european commission under the erc advanced grant 226371 . | we review recent work @xcite on obtaining a realisation of the @xmath0-background in terms of a special series of higher - derivative generalised f - terms in the effective @xmath1 supergravity action .
we discuss the motivation behind the identification of these couplings , their relation to the nekrasov partition function and connect them to a worldsheet approach towards the refined topological string .
@xmath2}$ ] max - planck - institut fr physik , werner - heisenberg - institut , 80805 mnchen , germany + @xmath3}$ ] department of physics , cern - theory division , ch-1211 geneva 23 , switzerland + _ presented by the authors at the corfu summer institute 2013 workshop on noncommutative field theory and gravity " , september 8 - 15 , 2013 + corfu , greece _ |
in recent lattice calculations , there have been several indications that the presence of sea quark increases the @xmath0 meson decay constant @xmath4 significantly , _ i.e. _ by about 1015% @xcite .
then , the study of the sea quark effect on the @xmath0 parameters has a great phenomenological importance , because the relevant physical quantity to the @xmath5 mixing is the combination @xmath6 . in this work
we perform unquenched lattice calculations of both @xmath4 and @xmath7 using a consistent formulation .
we use the nonperturbatively @xmath1-improved wilson quark action for both sea and valence light quarks , and the nrqcd action for heavy quark . using the same dynamical qcd configurations , a search for sea quark effects in the light hadron spectrum
is also being performed @xcite .
quenched calculations of @xmath4 and @xmath7 using the nrqcd action have been presented previously @xcite . in full qcd , the effect of virtual pion loops leads to chiral logarithms , a specific dependence of physical quantities on the sea quark mass .
hence , it provides a non - trivial test of the sea quark effect in lattice simulations . for @xmath4 and @xmath7 chiral perturbation theory ( chpt )
is extensively studied in @xcite including its extension to the quenched and the partially quenched cases , where sea and valence quark masses can be different .
we test the chiral logarithm using our numerical data .
we generate a set of unquenched gauge configurations with the plaquette gauge action and two - flavors of nonperturbatively @xmath1-improved wilson fermions @xcite .
simulations are performed on a @xmath3 lattice at @xmath8=5.2 , which roughly corresponds to the lattice spacing @xmath9 0.1 fm .
we take five different sea quark masses covering @xmath10=0.60.8 . for comparison
we also perform a quenched simulation at @xmath8=6.0 on the same lattice size .
further details of our simulations are given in @xcite .
we use the nrqcd action for heavy quark including all @xmath11 corrections consistently in the action and operators . in the calculations of @xmath4 and @xmath7
we follow the methods adopted in the previous works @xcite . in particular , the @xmath0 parameters are computed with four different prescriptions , all of which have equivalent accuracy up to higher order contributions we neglect .
we may estimate the systematic uncertainty taking the differences among them @xcite .
.,width=283 ] figure [ fig : chi_phi_fb ] shows the chiral behavior of the heavy - light decay constant @xmath12 with @xmath13 .
both horizontal and vertical axes are normalized with a proper power of @xmath14 and converted to the physical unit using the lattice result for @xmath15 combined with physical input for @xmath16 . if we compare the unquenched result ( filled triangles ) with the quenched data ( open diamonds ) , we clearly find the increase of about 15% with the unquenching as found previously @xcite we consider the following three forms for the chiral extrapolation @xmath17 with @xmath18 .
the `` log '' form is motivated by the chpt result for heavy - light mesons @xcite . for two - flavor qcd , it predicts @xmath19 with @xmath20 the pion decay constant .
the coupling @xmath21 describes @xmath22 interaction , which may be evaluated from @xmath23 decays using heavy quark symmetry .
if we take a range @xmath21=0.20.7 rather conservatively covering @xmath24 decay @xcite and the recent @xmath23 width experiment @xcite , then we obtain @[email protected](2 ) gev@xmath27 .
we find that both `` qua '' and `` log '' fits describe our unquenched data well , while the linear fit gives somewhat poorer @xmath28 than the other two .
for the `` log '' fit we obtain @[email protected](3 ) gev@xmath27 , which is consistent with the above phenomenological estimate . on the other hand , in quenched qcd ,
no deviation from linearity is observed in our data , and the three fit forms lead to almost identical result in the chiral limit .
this behavior contradicts the prediction of quenched chpt ( qchpt ) @xcite , which gives even stronger logarithmic dependence .
there are , however , many unknown parameters in qchpt , so it is still possible to find a parameter region which mimics the linear behavior in our light quark mass regime .
dependence of @xmath29.,width=283 ] [ fig : mas_phi_fb ] since we are not able to decide the fit form with our current unquenched data , the uncertainty in the chiral limit is quite large .
figure [ fig : mas_phi_fb ] shows the @xmath30 dependence of @xmath29 at the physical light quark mass . at the physical @xmath0 meson mass ,
we observe that the two - flavor result ( filled symbols ) is 1425% larger than the quenched result ( open circle ) , depending on the fit form .
we also examine the chiral behavior of @xmath7 by applying the fit forms in ( [ eq : chi_phi_r ] ) , but , because the light quark mass dependence itself is quite small , we do not observe any significant dependence on the fit functions . a weak dependence on sea quark mass is also suggested by chpt , which gives @xmath31 for @xmath7 .
dependence of @xmath32.,height=283 ] figure [ fig : mas_bb ] shows the @xmath30 dependence of @xmath33 for both quenched ( open symbols ) and unquenched qcd ( filled symbols ) .
four different symbols correspond to different prescriptions defined in @xcite . within the uncertainty indicated by the different prescriptions
, we find no significant effect of unquenching .
@xmath34 is also analyzed in a similar manner and we arrive at the same conclusion .
our preliminary results in two - flavor qcd are summarized as follows , @xmath35 where errors are from statistics , chiral extrapolation , and systematics .
we discard the linear fit from the final results because of its poor @xmath28 value .
the central value is obtained by taking an average over `` qua '' and `` log '' results , and their difference is taken as a systematic error from the chiral extrapolation .
some of the systematic errors in the @xmath0 parameters are estimated by the scatter with different prescriptions as in @xcite .
other uncertainties are estimated by a naive dimensional counting , assuming @xmath36=500 mev and @xmath37=@xmath38=0.267 .
our data for @xmath4 and @xmath7 are consistent with the prediction of chpt .
this is different from the light hadron sector where we find that chpt fails to reproduce the lattice data for pion mass and decay constant @xcite . to investigate these issues we are currently accumulating further statistics , which also helps to reduce the statistical error in our results .
this work is supported by the supercomputer project no.66 ( fy2001 ) of high energy accelerator research organization ( kek ) , and also in part by the grant - in - aid of the ministry of education ( nos .
10640246 , 11640294 , 12014202 , 12640253 , 12640279 , 12740133 , 13640260 and 13740169 ) .
k - i.i and n.y are supported by the jsps research fellowship .
ishikawa _ et al .
_ [ jlqcd collaboration ] , phys . rev .
d * 61 * ( 2000 ) 074501 .
n. yamada _
[ jlqcd collaboration ] , nucl .
b ( proc . suppl . )
* 94 * ( 2001 ) 379 .
m. j. booth , phys .
d * 51 * ( 1995 ) 2338 .
s. r. sharpe and y. zhang , phys .
d * 53 * ( 1996 ) 5125 .
k. jansen and r. sommer [ alpha collaboration ] , nucl .
b * 530 * ( 1998 ) 185 .
i. w. stewart , nucl .
b * 529 * ( 1998 ) 62 .
a. anastassov _ et al .
_ [ cleo collaboration ] , hep - ex/0108043 . | we present a two - flavor dynamical qcd calculation of the @xmath0 meson @xmath0 parameters and decay constant .
we use nrqcd for heavy quark and the nonperturbatively @xmath1-improved wilson action for light quark at @xmath2 on a @xmath3 lattice .
we confirm that the sea quark effect increases the heavy - light decay constant , while estimate of its magnitude depends significantly on the fitting form in the chiral extrapolation . for the @xmath0 parameters , on the other hand , we do not find a significant sea quark effect .
the chiral extrapolation with logarithmic term is examined for both quantities and compared with the prediction of chpt .
kek - cp-113 |
although tunnel fets ( tfets ) were originally proposed for low power applications @xcite , the low on - current ( i@xmath9 ) challenge in tfets has concealed their scaling problem @xcite .
the low i@xmath9 challenge can be solved by increasing the electric field at the tunnel junction ; e.g. by using dielectric engineering @xcite , atomistically thin channels @xcite , or internal polarization @xcite .
however , the scaling challenge is more tricky since the tunneling currents i@xmath9 and i@xmath10 depend on the same device parameters . hence an attempt to decrease i@xmath10 would reduce i@xmath9 and vice versa . in contrast ,
i@xmath9 and i@xmath10 in mosfets are more independent of each other and a channel material with a large bandgap ( or optimized effective mass ) can be used for sub-12 nm channels to suppress the direct source - to - drain tunneling @xcite .
[ fig : fig1]a shows the device structure of an inas nanowire ( nw ) tfet with a diameter of 3.4 nm .
the transfer characteristics of the device simulated by the nemo5 tool @xcite are shown in fig .
[ fig : fig1]b with i@xmath10 fixed at 1@xmath11 . in the simulations , we scale @xmath0 down with the channel length ( @xmath1 ) .
the results indicate that the inas nw - tfet exhibits a promising performance with long channel lengths ( i.e. @xmath12 9 nm ) , however it completely fails to switch from off- to on - state for the case of @xmath1=6 nm and @xmath0=0.2v ( i.e. i@xmath9/i@xmath13 ) .
roughly , the transmission in the on - state ( @xmath14 ) and off - state ( @xmath15 ) of tfets depends on @xcite : @xmath16 @xmath17 where @xmath18 and @xmath1 are the tunneling distances in the on- and off - state ( fig .
[ fig : fig1]a ) respectively . @xmath19 and @xmath4 are the reduced effective mass and the bandgap of the channel material .
the scaling of the channel below 10 nm brings @xmath1 close to @xmath18 which reduces i@xmath9/i@xmath10 significantly .
one apparent solution can be a heterostructure channel where the term @xmath20 is different in ( [ eq : tun1 ] ) and ( [ eq : tun2 ] ) due to different materials used in those regions @xcite .
however , it has been shown that the presence of band discontinuity and interface states in heterostructures can deteriorate the off - state performance of tfets @xcite .
hence , in this work the homojunction tfets have been considered as a more practical steep subthreshold swing ( @xmath21 ) device . 0.27 for channel lengths from 15 nm to 6 nm which translates into corresponding @xmath0 from 0.5v to 0.2v ( from negf ) .
a 3 nm thick hfo@xmath22 with @xmath23=25 has been used as the gate oxide .
c ) impact of @xmath5 and @xmath4 variations on i@xmath9/i@xmath10 for a tfet with @xmath1=6 nm and @xmath0=0.2v predicted from an analytical model ( see sec . [
sec : sim]).,title="fig : " ] [ fig : struct1 ] 0.24 for channel lengths from 15 nm to 6 nm which translates into corresponding @xmath0 from 0.5v to 0.2v ( from negf ) .
a 3 nm thick hfo@xmath22 with @xmath23=25 has been used as the gate oxide .
c ) impact of @xmath5 and @xmath4 variations on i@xmath9/i@xmath10 for a tfet with @xmath1=6 nm and @xmath0=0.2v predicted from an analytical model ( see sec . [
sec : sim]).,title="fig : " ] [ fig : same_eot ] on top of the length scaling problem which increases i@xmath10 significantly , the voltage scaling reduces i@xmath9 .
the maximum tunneling window in tfets approximately equals @xmath24 .
thus a short channel tfet with a small @xmath0 is expected to have a small i@xmath9/i@xmath10 . in this work
, it is shown that by using a channel material with optimized @xmath5 and @xmath4 , it is still feasible to obtain an acceptable i@xmath9/i@xmath10 for ultra - scaled tfets ( i.e. i@xmath9/i@xmath25 for @xmath1=6 nm and @xmath0=0.2v ) .
the solution to the scaling problem of tfet is to _ scale down _ @xmath4 of channel material to the smallest possible value to achieve a high i@xmath9 .
of course @xmath4 can not be smaller than @xmath24 , otherwise the channel can not cover and block the tunneling energy window in the off - state .
on the other hand , @xmath5 should _ scale up _ with scaling down the dimensions to decrease i@xmath10 .
[ fig : fig1]c shows that the performance of 6 nm long gate - all - around tfet can be improved more than 4 orders of magnitude by _ scaling down _ @xmath4 and _ scaling up _ @xmath5 .
the favorable design space for @xmath5 and @xmath4 is discussed in sec .
[ sec : opt ] .
the self - consistent 3d poisson - negf ( non - equilibrium green s function ) method is used in the nemo5 software for the simulation of inas tfets @xcite .
the inas channel material is described by a 10 band nearest neighbor tight - binding model @xcite . to find the impact of @xmath5 and @xmath4 on the performance of tfets , a model is needed where
@xmath5 and @xmath4 can be set as free input parameters , in contrast to the atomistic approach where @xmath5 and @xmath4 are the output of the simulation through material composition and geometry induced confinement effects . to reduce the number of free parameters
, it is assumed that the electron and hole effective masses are equal ( @xmath26=@xmath27=@xmath5 ) .
recently , an analytical model was developed which produces results in excellent agreement with negf simulations @xcite . to show the validity of this analytical model for ultra - scaled tfets , the simulation results of scaled inas tfets with the scaling rule of @xmath28 obtained from the analytical model are benchmarked against the negf results first .
[ fig : fig2 ] compares the results of analytical model and negf simulations .
notice that in fig .
[ fig : fig2 ] , the off - state is not fixed unlike fig .
[ fig : fig1]b .
the accuracy and speed of the analytical model and tuneability of @xmath5 and @xmath4 makes this model an ideal tool for optimizing the tfet design .
0.35 -@xmath29 of the inas nw - tfet obtained from analytical model ( dashed lines ) and negf simulations ( solid lines ) with the same parameters as fig .
[ fig : fig1 ] .
[ fig : fig2],title="fig : " ] [ fig : same_eot ]
to analyze different tfet designs , the tunneling transmission path at the top of the tunneling window ( @xmath30 ) is indicated as a function of source - channel tunneling window ( @xmath31 ) in fig .
[ fig : fig3]a .
knowledge of the tunneling transmission probability as a function of @xmath31 ( i.e. @xmath32 ) provides information about the transfer characteristics @xcite .
[ fig : fig3]b shows an example of @xmath32 with the corresponding tfet operational regimes ( e.g. on- and off - states , and n- and p- branches ) .
notice that @xmath33 is the on - off transition point .
for a small drain - source voltage , the i - v can be calculated by integrating the @xmath32 in the tunneling energy window ( energies between @xmath34 and @xmath35 ) .
the tunneling transmission shows how far the tfet is from its ideal performance ( i.e. @xmath36 and @xmath37 at off- and on - state , respectively ) .
accordingly , i@xmath9 , i@xmath10 , and @xmath21 can be estimated from the maximum and minimum values of @xmath32 and its slope at subthreshold region .
the impact of @xmath1 scaling on the transmission profile of inas nw - tfet is shown in fig .
[ fig : fig3]c . reducing the channel length increases
@xmath38 significantly while @xmath39 remains intact which was expected from equations ( [ eq : tun1 ] ) and ( [ eq : tun2 ] ) . 0.25 parameter and b ) the transmission profile @xmath32 .
the impact of c ) channel length @xmath1 , d ) bandgap @xmath4 , e ) effective mass @xmath5 , f ) @xmath40 ratio on the transmission profile @xmath32 .
all these results are obtained using the analytical model.,title="fig : " ] 0.25 parameter and b ) the transmission profile @xmath32 . the impact of c ) channel length @xmath1 , d ) bandgap @xmath4 , e ) effective mass @xmath5 , f ) @xmath40 ratio on the transmission profile @xmath32 .
all these results are obtained using the analytical model.,title="fig : " ] 0.25 parameter and b ) the transmission profile @xmath32 . the impact of c ) channel length @xmath1 , d ) bandgap @xmath4 , e ) effective mass @xmath5 , f ) @xmath40 ratio on the transmission profile @xmath32 .
all these results are obtained using the analytical model.,title="fig : " ] 0.25 parameter and b ) the transmission profile @xmath32 . the impact of c ) channel length @xmath1 , d ) bandgap @xmath4 , e ) effective mass @xmath5 , f ) @xmath40 ratio on the transmission profile @xmath32 .
all these results are obtained using the analytical model.,title="fig : " ] 0.25 parameter and b ) the transmission profile @xmath32 . the impact of c ) channel length @xmath1 , d ) bandgap @xmath4 , e ) effective mass @xmath5 , f ) @xmath40 ratio on the transmission profile @xmath32 .
all these results are obtained using the analytical model.,title="fig : " ] 0.25 parameter and b ) the transmission profile @xmath32 . the impact of c ) channel length @xmath1 , d ) bandgap @xmath4 , e ) effective mass @xmath5 , f ) @xmath40 ratio on the transmission profile @xmath32 .
all these results are obtained using the analytical model.,title="fig : " ] fig .
[ fig : fig3]d shows the effect of bandgap on @xmath32 ; obviously , a larger bandgap decreases both @xmath38 and @xmath39 . notice that changing @xmath4 does not improve the subthreshold slope of @xmath32 ( black lines in fig .
[ fig : fig3]d ) . increasing @xmath4 decreases @xmath38 more than @xmath39 since the prefactor of @xmath41 is larger for @xmath38 ( note that @xmath42 in equations ( [ eq : tun1 ] ) and ( [ eq : tun2 ] ) ) . on the other hand , to reach this lower @xmath38 a larger gate voltage change is needed for larger band gaps ( i.e. @xmath43 ) .
thus , there is no noticeable improvement in @xmath21 with larger @xmath4 . on the other hand ,
increasing @xmath5 improves @xmath21 as shown in fig .
[ fig : fig3]e . since a larger @xmath5 does not require a larger gate voltage change , contrary to a larger @xmath4 .
[ fig : fig3]f compares tfets with a constant @xmath41 but different @xmath44 ratios . notice that not only @xmath21 improves with increasing @xmath44 ratio , but also @xmath39 .
the reason for improved on - state performance is that reducing @xmath4 decreases the depletion width at the source - channel interface and @xmath18 decreases in equation ( [ eq : tun1 ] ) @xcite .
[ fig : fig4]a shows the i@xmath9/i@xmath10 ratio of nw - tfets with @xmath1=6 nm and @xmath0=0.2v and a channel material with different @xmath5 and @xmath4 . to suppress the p - branch of tfets , the drain doping level is chosen to be much smaller than source doping level ( @xmath45 ) and a gate leakage of 1@xmath11
is assumed ( i@xmath46 ) @xcite .
the maximum i@xmath9/i@xmath10 ratio is obtained with an @xmath4 of about @xmath47 . moreover , with increasing @xmath4 , the optimum @xmath48 reduces and for @xmath49 the product @xmath50 ( circle symbols ) saturates ( dashed line ) .
[ fig : fig4]b shows that tfets with @xmath4 between @xmath51 and @xmath52 have acceptable i@xmath9/i@xmath10 ratios according to itrs requirements ( i@xmath9/i@xmath53 ) .
[ fig : fig5]a illustrates the favorable design space for @xmath5 as a function of @xmath0 for tfets with the scaling rule of @xmath54 . the shaded area in fig .
[ fig : fig5]a shows higher and lower bounds on @xmath5 and @xmath4 of the channel material for a high performance ultra - scaled nw - tfet .
0.25 /i@xmath10 ratio for a nw - tfet with @xmath1=6 nm and @xmath0=0.2v for different @xmath4 and their optimized values @xmath55 ( circle symbols ) .
b ) @xmath55 as a function of @xmath4 for @xmath0=0.2v and 0.3v.,title="fig : " ] 0.24 /i@xmath10 ratio for a nw - tfet with @xmath1=6 nm and @xmath0=0.2v for different @xmath4 and their optimized values @xmath55 ( circle symbols ) .
b ) @xmath55 as a function of @xmath4 for @xmath0=0.2v and 0.3v.,title="fig : " ] fig .
[ fig : fig5]b shows the transfer characteristics of nw - tfets with optimized @xmath4 and @xmath5 from equations ( [ eq : opt_vdd ] ) and ( [ eq : opt_m ] ) .
i@xmath9/i@xmath10 ratio of larger than @xmath56 and @xmath21 below 15@xmath57 are obtained for all the cases including the 6 nm long channel .
0.25 for tfets with the scaling rule of @xmath54 for @xmath1 from 15 nm to 6 nm .
b ) @xmath58-@xmath29 of nw - tfets with optimized @xmath4 and @xmath5 from ( [ eq : opt_vdd ] ) and ( [ eq : opt_m ] ) .
, title="fig : " ] 0.25 for tfets with the scaling rule of @xmath54 for @xmath1 from 15 nm to 6 nm .
b ) @xmath58-@xmath29 of nw - tfets with optimized @xmath4 and @xmath5 from ( [ eq : opt_vdd ] ) and ( [ eq : opt_m ] ) .
, title="fig : " ]
in summary , the scaling of tfets pushes the semiconductor industry to look for channel materials with higher @xmath5 , similar to ultra - scaled mosfets @xcite .
however , in tfets channel material should have both @xmath5 and @xmath4 optimized .
more accurately , the scaling of high performance nw - tfets below 10 nm requires : 1 . a channel material with
_ scaled down _ band gap @xmath59\ ] ] 2 . a channel material with
_ scaled up _ effective mass @xmath60\ ] ] 3 .
higher doping level in the source ( @xmath61 ) than drain ( @xmath62 ) .
@xmath63 4 . a channel material with low dielectric constant ( @xmath64 ) and a high - k oxide .
@xmath65 j. appenzeller , y .- m .
lin , j. knoch , z. chen , and ph .
avouris , _ `` comparing carbon nanotube transistors - the ideal choice : a novel tunneling device design , '' _ ieee trans . on electron dev . 52 , 2568 - 2576 ( 2005 ) .
h. ilatikhameneh , y. tan , b. novakovic , g. klimeck , r. rahman , j. appenzeller , _
`` tunnel field - effect transistors in 2d transition metal dichalcogenide materials , '' _ ieee exploratory solid - state computational devices and circuits , vol . 1 , no . 1 , pp .
12 - 18 ( 2015 ) , doi : 10.1109/jxcdc.2015.2423096 .
g. fiori , f. bonaccorso , g. iannaccone , t. palacios , d. neumaier , a. seabaugh , s. k. banerjee , and l. colombo , _
`` electronics based on two - dimensional materials , '' _ nature nanotechnology , vol . 9 , no . 10 , pp .
768 - 779 ( 2014 ) .
h. ilatikhameneh , g. klimeck , j. appenzeller , and r. rahman , _
`` scaling theory of electrically doped 2d transistors , '' _ ieee electron device letters , vol .
36 , no . 7 , pp .
726 - 728 ( 2015 ) , doi:10.1109/led.2015.2436356 .
w. li , s. sharmin , h. ilatikhameneh , r. rahman , y. lu , j. wang , x. yan , a. seabaugh , g. klimeck , d. jena , p. fay , _ `` polarization - engineered iii - nitride heterojunction tunnel field - effect transistors , '' _ ieee exploratory solid - state computational devices and circuits , vol . 1 , no . 1 , pp . 28 - 34 ( 2015 ) ,
doi:10.1109/jxcdc.2015.2426433 .
h. ilatikhameneh , r. b. salazar , r. rahman , g. klimeck , j. appenzeller , _ from fowler - nordheim to non - equilibrium green s function modeling of tunneling _ , ( 2015 ) [ online . ] http://arxiv.org/abs/1509.08170 t. krishnamohan , d. kim , s. raghunathan , and k. saraswat .
_ `` double - gate strained - ge heterostructure tunneling fet ( tfet ) with record high drive currents and subthreshold slope < < 60mv / dec , '' _ in ieee international electron devices meeting iedm , pp
. 1 - 3 ( 2008 ) .
koswatta , s.j .
koester , and w. haensch , _
`` on the possibility of obtaining mosfet - like performance and sub-60-mv / dec swing in 1-d broken - gap tunnel transistors , '' _ ieee trans .
electron devices 57 , 3222 ( 2010 ) .
s. agarwal , and e. yablonovitch , _
`` band - edge steepness obtained from esaki / backward diode current voltage characteristics , '' _ ieee transaction on electron devices , vol .
5 , pp . 1488 - 1493 ( 2014 )
. j. e. fonseca , t. kubis , m. povolotskyi , b. novakovic , a. ajoy , g. hegde , h. ilatikhameneh , z. jiang , p. sengupta , y. tan , g. klimeck , _ `` efficient and realistic device modeling from atomic detail to the nanoscale , '' _ journal of computational electronics , vol .
592 - 600 ( 2013 ) .
s. steiger , m. povolotskyi , h. h. park , t. kubis , and g. klimeck , _
`` nemo5 : a parallel multiscale nanoelectronics modeling tool , '' _ ieee transaction on nanotechnology , vol . 10 , no . 6 , pp .
1464 - 1474 ( 2011 ) . | the main promise of tunnel fets ( tfets ) is to enable supply voltage ( @xmath0 ) scaling in conjunction with dimension scaling of transistors to reduce power consumption .
however , reducing @xmath0 and channel length ( @xmath1 ) typically deteriorates the on- and off - state performance of tfets , respectively .
accordingly , there is not yet any report of a high performance tfet with both low v@xmath2 ( @xmath30.2v ) and small @xmath1 ( @xmath36 nm ) . in this work
, it is shown that scaling tfets in general requires _ scaling down _ the bandgap @xmath4 and _ scaling up _
the effective mass @xmath5 for high performance .
quantitatively , a channel material with an optimized bandgap ( @xmath6 $ ] ) and an engineered effective mass ( @xmath7 $ ] ) makes both @xmath0 and @xmath1 scaling feasible with the scaling rule of @xmath8 for @xmath1 from 15 nm to 6 nm and corresponding @xmath0 from 0.5v to 0.2v .
tfets , nanowire , scaling , sub-10 nm , direct tunneling , negf . |
since complete genomes of many organisms are available from web - based databases , a full and systematic search of genome structures , functions and dynamics becomes an essential part of the study for both biologists and physicists .
for the large amount of genomes , developing quantitative methods to extract meaningful information is a major challenge with respect to applications of statistical mechanics and nonlinear dynamics to biological systems@xcite . to understand the complete genomes ,
some statistical and geometrical methods were developed@xcite .
the studies of the complete genomes of many organisms came up with the determinations of the nontrivial statistical characteristics , such as the long - range correlations , the short - range correlations and the fractal features or genomic signatures .
in particular , it was found that the transposable elements , as the mobile dna sequences , have the ability to move from one place to another and make many replicas within the genome via the transposition@xcite .
their origin , evolution , and tremendous effects on the genome structure and the gene function are issues of fundamental importance in biology@xcite . in general , the symbolic dynamics and the recurrence plots are basic methods of nonlinear dynamics for analyzing complex systems@xcite .
although the conventional methods have made great strides in understanding genetic patterns , they are required to analyze the so - called junk dna with complex functions governing mutations@xcite .
recently , a one - to - one metric representation of a genome borrowed from the symbolic dynamics was proposed to form a fractal pattern in a plane@xcite . by using the metric representation method , the recurrence plot technique of the genome
was established to analyze the correlation structures of nucleotide strings@xcite .
the transference of nucleotide strings appears at many positions of a complete genome and makes a regular and irregular correlation structures , but the periodic correlation structures in the complete genome are the most interesting in view of the dynamics . in this paper , using the metric representation and the recurrence plot method , we identify periodic correlation structures in bacterial and archaeal complete genomes and analyze the mechanism of the periodic correlation structures .
since the nucleotide strings include transposable elements , the mechanism is conducible to understanding the genome structures in terms of nucleotide strings transferring in the genomes and exploring relations between transference of nucleotide strings and the transposable elements .
in what follows , we give a brief presentation of the metric representation and the recurrence plot method , which are detailed in @xcite . for a given symbolic sequence @xmath0 ( @xmath1 ) ,
a metric representation for its subsequences @xmath2 ( @xmath3 ) is defined as @xmath4 where @xmath5 is 0 if @xmath6 or 1 if @xmath7 and @xmath8 is 0 if @xmath9 or 1 if @xmath10 .
it maps the one - dimensional symbolic sequence to the two - dimensional plane ( @xmath11 ) .
the subsequences with the same ending @xmath12-nucleotide string are labeled with @xmath13 .
they correspond to points in the zone encoded by the @xmath12-nucleotide string . with two subsequences @xmath14 and
@xmath15 ( @xmath16 ) , we calculate @xmath17 where @xmath18 and @xmath19 is the heaviside function [ @xmath20 , if @xmath21 ; @xmath22 , if @xmath23 . when @xmath24 , i.e. , @xmath25 , a point @xmath26 is plotted on a plane . repeating the above process for @xmath27 $ ] and @xmath28 $ ] , we obtain a recurrence plot of the symbolic sequence . to present the correlation structure in the recurrence plot plane
, we define a correlation intensity at a given correlation distance @xmath29 as @xmath30 which displays the transference of @xmath12-nucleotide strings in the symbolic sequence . on the recurrent plot plane , since @xmath31 and @xmath15 @xmath32 , the transferring element has a length @xmath12 at least .
we calculate the maximal value of @xmath33 to satisfy @xmath34 i.e. , @xmath35 and @xmath36 .
the transferring element has a length @xmath37 and is placed at the positions @xmath38 and @xmath39 , which implies the correction distance @xmath40 . to understand the transferring characteristics of a complex genome
, we investigate the correlation structures of simple periodic and random sequences . by randomly combining the four letters a , c , g and t , we firstly generate two random nucleotide sequences :
one has the length of 67 and another has the length of 5000 .
then , a periodic nucleotide sequence with the total length of 5000 is formed by repeating the short nucleotide string . using the metric representation and the recurrence plot method
, we may determine the correlation intensities at different correlation distances with @xmath41 for the periodic and random sequences in fig .
it is evident that there exist equidistant parallel lines with a basic correlation distance in fig .
1(a ) , to form the periodic correlation structure for the periodic sequence .
the basic correlation distance hereinafter called the basic periodic length is determined as @xmath42 .
the correlation intensity @xmath43 decreases linearly with the increase of the correlation distance ( @xmath44 , @xmath45 , @xmath46 ) . however , in fig .
1(b ) , the correlation intensity @xmath43 is very small , so there are almost no correlation structures for the random sequence .
therefore , the periodic and random sequences exhibit two very different transferring characteristics : with the periodic correlation structure with a linearly decreasing intensity and without a clear correlation structure .
at the end of 1999 , complete genomes including more of 20 bacteria were in the genbank@xcite . by using the string composition and the metric representation method , the suppressions of all short strings in 23 bacterial and archaeal complete genomes were determined@xcite . in this section , using the metric representation and the recurrence plot method , we determine all long periodic nucleotide strings ( @xmath47 bases ) in the 23 bacterial and archaeal genomes . for the 23 genomes , only 13 have long periodic nucleotide strings .
all basic strings and their lengths of the long periodic nucleotide strings in the 13 bacterial and archaeal genomes are presented in table i in the order of decreasing suppressions of nucleotide strings@xcite .
several periods and different basic strings can be seen depending on the genomes , but not necessarily on the lengths of genomes .
the genomes of helicobacter pylori 26695 ( @xmath48 ) , helicobacter pylori j99 ( @xmath49 ) , haemophilus influenzae rd kw20 ( @xmath50 ) , mycobacterium tuberculosis h37rv ( @xmath51 ) , synechocystis sp .
pcc6803 ( @xmath52 ) have more periods ( @xmath53 ) and basic strings ( @xmath54 ) than others , which have only fewer periods ( @xmath55 ) and basic strings ( @xmath56 ) . in each period
, the number of the basic strings generally depends on the length of the period . the longer / shorter period the basic strings have ,
the smaller / greater their number will be . in the next section , we will investigate the periodic transference of nucleotide strings in the bacterial and archaeal complete genoms and analyze the effects of periodic nucleotide strings on the correlation structures .
the periodic correlation structures of a complete genome contain several basic periodic and/or quasi - periodic lengths , which are determined by using the metric representation and the recurrence plot method as follows . from the relationship between the correlation intensity and the correlation distance obtained by using eq .
( 3 ) , the basic periodic lengths and their integer multiples with strong correlation intensities can be calculated . moreover , in the transference of nucleotide strings obtained by using eq .
( 4 ) , the correlation distance with basic periodic lengths and their integer multiples can also be found . by using both methods , the basic periodic lengths of the periodic correlation structures
are determined , as shown in table ii , where the 23 complete genomes with official genbank accession numbers are arranged in the order of decreasing suppressions of nucleotide strings@xcite .
when the periodic correlation structures have only a few peaks of the correlation intensity within the correlation distance , the basic periodic lengths are put in parentheses . to see the characteristics of the periodic correlation structures
, we also present all basic string lengths in long periodic nucleotide strings ( @xmath57 bases ) in table ii . when a periodic correlation structure is identified based on a long periodic nucleotide string , the transference of nucleotide strings composed of the basic strings appears at some positions where the correlation distance is integer multiples of the period and monotonically increases . at the same time , the lengths of transferred nucleotide strings monotonically decrease .
there exists a `` cascade '' arrangement of nucleotide strings related to the basic periodic length .
however , when a periodic correlation structure is identified based on non - periodic nucleotide strings , the transference of nucleotide strings appears at several positions where the correlation distance is almost integer multiples of the basic periodic length .
there are no `` cascade '' arrangements of nucleotide strings related to the basic periodic length .
according to the characteristics of the periodic correlation structures , the results can be summarized as follows : ( 1)the correlation distance contains a single increasing period .
the most of the complete genomes with a single increasing period have a basic periodic length of 67 .
they include methanococcus jannaschii dsm 2661 ( _ mjan _ ) , methanobacterium thermoautotrophicum str .
delta h ( _ mthe _ ) , pyrococcus horikoshii ot3 ( _ pyro _ ) , archaeoglobus fulgidus dsm 4304 ( _ aful _ ) , pyrococcus abyssi ( _ pabyssi _ ) and thermotoga maritima msb8 ( _ tmar _ ) genomes . consider the _
genome as an example .
2 displays the correlation intensity at different correlation distances with @xmath58 for the _ mjan _ genome .
it is evident that there exist some equidistant parallel lines with a basic periodic length , to form a periodic correlation structure .
the basic periodic length is determined as @xmath42 .
generally , if the genome has a periodic nucleotide string with the basic string length @xmath59 , it would tend to form a periodic correlation structure . in table
ii , the _ mjan _ genome has the correspondent basic string length @xmath59 for periodic nucleotide strings .
for example , the nucleotide string @xmath60 ( 237122 - 237620 ) with @xmath61 is formed by repeating the basic string @xmath62 with the length @xmath59 , where @xmath63 . in other words ,
the basic string is duplicated to the positions with the correlation distances @xmath64 , @xmath65 , @xmath66 , @xmath67 , @xmath68 , @xmath69 and @xmath70 . despite possible contribution from such periodic nucleotide strings , the periodic correlation structure is mainly formed by the transference of non - periodic nucleotide strings , which has approximately the same increasing period .
for example , the nucleotide string @xmath71 ( 447 - 476 ) with @xmath72 is transferred to the places ( 514 - 543 ) , ( 581 - 610 ) , ( 651 - 680 ) , ( 718 - 747 ) , ( 785 - 814 ) , ( 855 - 884 ) , ( 922 - 951 ) , ( 994 - 1023 ) , ( 1064 - 1093 ) , ( 1132 - 1161 ) and @xmath46 with the correlation distances @xmath44 , @xmath45 , @xmath73 , @xmath74 , @xmath75 , @xmath76 , @xmath77 , @xmath78 , @xmath79 , @xmath80 and @xmath46 , respectively . since the nucleotide string @xmath81 is neither periodic nor a part of a periodic nucleotide string , its periodic transference is not a repetition of basic periodic nucleotide strings
. moreover , fig .
2 shows that there also exists a cluster of basic periodic lengths close to @xmath44 .
their integer multiples are distributed near the periodic correlation structure .
table ii shows that there also exists another basic string length @xmath82 for periodic nucleotide strings , which is conducible to form the cluster distribution near the periodic correlation structure .
so both the repetition of basic periodic nucleotide strings and the transference of non - periodic nucleotide strings would form the periodic correlation structure with approximately the same increasing period .
besides the _ mjan _ genome , the other genomes ( _ mthe _ , _ pyro _ , _ aful _ , _ pabyssi _ and _ tmar _ ) have no periodic nucleotide strings with the basic string length @xmath59 to make contributions to the periodic correlation structure .
so the periodic correlation structure is formed by the transference of non - periodic nucleotide strings .
furthermore , the genomes of mycoplasma genitalium g37 ( _ mgen _ ) , _ hinf _ , mycoplasma pneumoniae m129 ( _ mpneu _ ) , treponema pallidum subsp .
pallidum str .
nichols ( _ tpal _ ) , aeropyrum pernix k1 ( _ aero _ ) , rickettsia prowazekii str .
madrid e ( _ rpxx _ ) and borrelia burgdorferi b31 ( _ bbur _ ) have basic periodic lengths @xmath83 , 4 , 12 , 24 , 65 , 84 and 162 , respectively . in table
ii , they correspond to periodic nucleotide strings with the basic length @xmath59 except the _ aero _ genome .
so both the repetition of basic periodic nucleotide strings and the transference of non - periodic nucleotide strings would form the periodic correlation structure with approximately the same increasing period .
\(2 ) the correlation distance contains several increasing periods . the escherichia coli k-12 mg1655 ( _ ecoli _ ) genome has two basic periodic lengths 100 and 113 .
the _ hpyl99 _ genome has three basic periodic lengths 8 , 15 and 21 .
the _ mtub _
genome has three basic periodic lengths 9 , 15 and 57 .
consider the _
hpyl99 _ genome as an example .
3 displays the correlation intensity at different correlation distances with @xmath58 for the _ hpyl99 _ genome .
it is evident that there exist some equidistant parallel lines with basic periodic lengths , to form periodic correlation structures .
three basic periodic lengths are determined as @xmath84 , @xmath85 and @xmath86 .
although there are some peaks of the correlation intensity in the correlation distance as shown in fig .
3 , they do not form any periodic correlation structures and are not accounted .
table ii also shows some periodic nucleotide strings with basic string lengths @xmath87 , @xmath88 , @xmath89 and their integer multiples , which contribute to the periodic correlation structures .
for example , the nucleotide string @xmath90 ( 1061079 - 1061153 ) with @xmath91 is formed by repeating the basic string @xmath92 with the length @xmath93 , where @xmath94 .
the nucleotide string @xmath95 ( 5153 - 5280 ) with @xmath96 is formed by repeating the basic string @xmath97 with the length @xmath98 , where @xmath99 . the nucleotide string @xmath100 ( 659300 - 659450 ) with @xmath101 is formed by repeating the basic string @xmath102 with the length @xmath103 , where @xmath104 .
although the transference of non - periodic nucleotide strings might also contribute to the periodic correlation structures , they are mainly formed by repeating the basic periodic nucleotide strings .
for example , the non - periodic nucleotide string @xmath105 ( 59514 - 59537 ) with @xmath106 is transferred to the places ( 59640 - 59663 ) and ( 59724 - 59747 ) with the correlation distances @xmath107 and @xmath108 , respectively .
so both the repetition of basic periodic nucleotide strings and the transference of non - periodic nucleotide strings would form the periodic correlation structures with approximately the same increasing periods .
\(3 ) the correlation distance has an increasing quasi - period .
the bacillus subtilis subsp .
subtilis str .
168 ( _ bsub _ ) genome has a basic quasi - periodic length of 5000 .
4 shows the correlation intensity at different correlation distances with @xmath58 for the _ bsub _ genome .
it is evident that there exist some approximately equidistant parallel lines at the positions @xmath109 , 10605 , 15427 and 20468 , to form a quasi - periodic correlation structure with a basic quasi - periodic length @xmath110 .
although a stronger correlation intensity appears at the position @xmath111 , it is far away from the quasi - periodic correlation structure and is not accounted . in table ii
, there are no periodic nucleotide strings with the length @xmath112 to make a contribution to the quasi - periodic correlation structure . for example , the non - periodic nucleotide string @xmath113 ( 167978 - 169382 ) with @xmath114 is transferred to the place ( 172974 - 174378 ) with the correlation distance @xmath115 .
the non - periodic nucleotide string @xmath116 ( 161449 - 161666 ) with @xmath117 is transferred to the places ( 167057 - 167274 ) , ( 172056 - 172273 ) and ( 946761 - 946798 ) with the correlation distances
@xmath118 , @xmath119 and @xmath120 , respectively .
so the transference of non - periodic nucleotide strings would form the quasi - periodic correlation structure .
\(4 ) the correlation distance contains a combination of several increasing periods and an increasing quasi - period .
firstly , the _ hpyl _ genome has two basic periodic lengths 7 , 8 and a basic quasi - periodic length of 114 . fig .
5(a ) shows the correlation intensity at different correlation distances with @xmath58 for the _ hpyl _ genome , with a local region magnified .
it is evident that there exist some equidistant parallel lines with basic periodic lengths , to form periodic correlation structures in a short range of the correlation distance .
the two basic periodic lengths are determined as @xmath121 and @xmath122 .
moreover , in fig .
5(a ) , there also exist some approximately equidistant parallel lines at the positions @xmath123 , 207 , 324 , 438 , 552 , 666 and 780 , to form a quasi - periodic correlation structure in a long range of the correlation distance .
the quasi - periodic correlation distance is described as @xmath124 , where the basic quasi - periodic length @xmath125 is 114 and @xmath126 . in table ii , there exist some periodic nucleotide strings with basic string lengths @xmath87 , @xmath88 and their integer multiples , but no periodic nucleotide strings with the basic string length @xmath89 .
for example , the nucleotide string @xmath127 ( 1 - 181 ) with @xmath128 is formed by repeating the basic string @xmath129 with the length @xmath93 , where @xmath130 .
the nucleotide string @xmath131 ( 444403 - 444490 ) with @xmath132 is formed by repeating the basic string @xmath133 with the length @xmath98 , where @xmath134 .
although the transference of non - periodic nucleotide strings might also contribute to the periodic correlation structures , they are mainly formed by repeating the basic periodic nucleotide strings .
for example , the non - periodic nucleotide string @xmath135 ( 84905 - 84926 ) with @xmath136 is transferred to the place ( 84929 - 94850 ) with the correlation distance @xmath137 .
moreover , for the quasi - periodic correlation structure , the non - periodic nucleotide string @xmath138 ( 556196 - 556224 ) with @xmath139 is transferred to the places ( 556634 - 556662 ) , ( 556748 - 556776 ) , ( 556862 - 556906 ) , ( 557300 - 557328 ) , ( 557414 - 557442 ) and ( 557852 - 557880 ) with the correlation distances
@xmath140 , @xmath141 , @xmath142 , @xmath143 , @xmath144 and @xmath145 , respectively .
so both the repetition of basic periodic nucleotide strings and the transference of non - periodic nucleotide strings would form the periodic correlation structures with approximately the same increasing periods in a short correlation distance , but only the transference of non - periodic nucleotide strings would form the quasi - periodic correlation structure in a long correlation distance .
secondly , the _ synecho _ genome has two basic periodic lengths 6 , 888 and a basic quasi - periodic length of 296 .
5(b ) shows the correlation intensity at different correlation distance with @xmath58 for the _ synecho _ genome , with a local region magnified .
it is evident that there exist some equidistant parallel lines with basic periodic lengths , to form periodic correlation structures in short and long ranges of correlation distances , respectively .
two basic periodic lengths are determined as @xmath146 and @xmath147 . moreover , in fig .
5(b ) , there also exist some approximately equidistant parallel lines at the positions @xmath148 and @xmath149 , where the basic quasi - periodic length @xmath125 is 296 and @xmath150 .
they form quasi - periodic correlation structures in a long range of the correlation distance . in table ii , there exist some periodic nucleotide strings with integer multiples of @xmath151 and the basic string length @xmath88 , but no periodic nucleotide strings with the basic string length @xmath89 .
for example , the nucleotide string @xmath152 ( 527703 - 527770 ) with @xmath153 is formed by repeating the basic string @xmath154 with the length @xmath155 , where @xmath156 .
the nucleotide string @xmath157 ( 2354010 - 2355833 ) with @xmath158 is formed by repeating the basic string @xmath159 with the length @xmath98 , where @xmath160 .
moreover , the nucleotide string @xmath161 ( 527395 - 527434 ) with @xmath162 is transferred to the places ( 527473 - 527512 ) , ( 527491 - 527530 ) , ( 527509 - 527548 ) , ( 527527 - 527566 ) and ( 527545 - 527584 ) with the correlation distances
@xmath163 , @xmath164 , @xmath165 , @xmath166 , and @xmath167 , respectively .
the non - periodic nucleotide string @xmath168 ( 2354010 - 2354300 ) with @xmath169 is transferred to the places ( 2356674 - 2356964 ) , ( 2357562 - 2357852 ) and ( 2358450 - 2358740 ) with the correlation distances
@xmath137 , @xmath170 and @xmath171 , respectively . both the repetition of basic periodic nucleotide strings and the transference of non - periodic nucleotide strings would form the periodic correlation structures with approximately the same increasing periods in short and long correlation distances , but only the transference of non - periodic nucleotide strings would form the quasi - periodic correlation structures in a long correlation distance .
\(5 ) the correlation distance contains almost no increasing periods .
the genomes of aquifex aeolicus vf5 ( _ aquae _ ) , rhizobium sp . ngr234 plasmid pngr234a ( _ pngr234 _ ) , chlamydophila pneumoniae cwl029 ( _ cpneu _ ) and chlamydia trachomatis d / uw-3/cx ( _ ctra _ ) are among cases with such characteristics .
consider the _
genome as an example .
6 shows the correlation intensity at different correlation distances with @xmath58 for the _ aquae _ genome .
it is evident that there exist some equidistant parallel lines with a basic periodic length , which is determined as @xmath42 .
however , for the basic periodic length @xmath42 , the maximal correlation intensity @xmath43 is only 179 and the correlation structure has only three peaks of the correlation intensity at the positions @xmath44 , @xmath45 and @xmath172 . the weak correlation intensity with a few peaks in the correlation distance may not make any periodic correlation structures . in table ii , there are also no periodic nucleotide strings for the almost non - periodic correlation structure .
so the _ aquae _ genome almost has no periodic correlation structures .
in summary , using the metric representation and the recurrence plot method , we have observed periodic correlation structures in bacterial and archaeal complete genomes .
all basic periodic lengths in the periodic correlation structures are determined .
on the basis of the periodic correlation structures , the bacterial and archaeal complete genomes , as classified into five groups , display four kinds of fundamental transferring characteristics : a single increasing period , several increasing periods , an increasing quasi - period and almost noincreasing period .
the mechanism of the periodic correlation structures is further analyzed by determining all long periodic nucleotide strings in the bacterial and archaeal complete genomes and is explained as follows : both the repetition of basic periodic nucleotide strings and the transference of non - periodic nucleotide strings would form the periodic correlation structures with approximately the same increasing periods . in comparison with the complete genome of the saccharomyces cevevisiae yeast@xcite , it is found that the bacterial , archaeal and yeast complete genomes have the same four kinds of fundamental transferring characteristics of nucleotide strings .
they choose preferably the basic periodic length @xmath173 or its double @xmath174 in the periodic correlation structures , even they do not have basic string lengths of long periodic nucleotide strings , which are equal to the basic periodic lengths .
the basic periodic length @xmath174 was also found in the correlation analysis of the human genomes@xcite .
although more and more biological functions of the junk dna in cells are found , the mystery of transposable elements in the whole genomes remains unraveled .
the purpose of this work is to depict the genome structure in the bacterial and archaeal complete genomes and explain the genome dynamics in terms of nucleotide string transfer .
the proposed periodic correlation structures with approximately the same increasing periods may have fundamental importance for the biological functions of the junk dna . *
acknowledgments * we would like to thank the national science foundation for partial support through the grant no .
11172310 and the imech / sccas shenteng 1800/7000 research computing facility for assisting in the computation .
p. j. deschavanne , a. giron , j. vilain , g. fagot , and b. fertil , genomic signature : characterization and classification of species assessed by chaos game representation of sequences .
* 16 * ( 1999 ) 1391 .
p. katsaloulis , t. theoharis , w .- m .
zheng , b .-
hao , a. bountis , y. almirantis and a. provata , long - range correlations of rna polymerase ii promoter sequences across organisms , _ physica a _ * 366 * ( 2006 ) 308 .
eckmann , s. o. kamphorst and d. ruelle , recurrence plots of dynamical systems , _ europhys .
_ , * 5 * , ( 1987 ) 973 ; n. marwan , m. c. romano , m. thiel and j. kurths , recurrence plots for the analysis of complex systems , _ phys .
* 438 * ( 2007 ) 237 . | the periodic transference of nucleotide strings in bacterial and archaeal complete genomes is investigated by using the metric representation and the recurrence plot method .
the generated periodic correlation structures exhibit four kinds of fundamental transferring characteristics : a single increasing period , several increasing periods , an increasing quasi - period and almost noincreasing period .
the mechanism of the periodic transference is further analyzed by determining all long periodic nucleotide strings in the bacterial and archaeal complete genomes and is explained as follows : both the repetition of basic periodic nucleotide strings and the transference of non - periodic nucleotide strings would form the periodic correlation structures with approximately the same increasing periods . *
keywords * bacterial and archaeal complete genomes , periodic correlation structures , metric representation , recurrence plots + |
a linear collider with both beams polarized will be an excellent tool not only to discover supersymmetric particles but also to determine the underlying susy model .
associated neutralino production @xmath2 could be the first kinematically accessible process that may allow a discrimination between the minimal supersymmetric standard model ( mssm ) and its nonminimal extensions .
polarization of both beams will clearly facilitate this challenging task .
an extensive analysis of polarization and spin effects in neutralino production and subsequent decay including the complete spin correlations in the mssm has been performed in @xcite .
this study also includes extended susy models : the next - to - minimal supersymmetric standard model ( nmssm ) with an additional higgs singlet superfield and an @xmath1 inspired model with a higgs singlet and a new @xmath3 gauge boson .
significant differences between the mssm and susy models with gauge singlets arise by the neutralino singlet components which do not couple to ( scalar ) fermions and standard gauge bosons . within a special scenario , where the masses of the light neutralinos were fixed in the three susy models and
the lightest neutralino had a large singlino component in the extended models , the cross sections for neutralino production at an @xmath4 linear collider with polarized beams have been discussed in @xcite , and complete different polarization asymmetries have been found .
now the study of the polarization effects is extended to a large region of the parameter space where the lightest neutralino is mainly a singlino . comparing the longitudinally polarized cross sections we identify the parameter regions where polarization effects may help to specify nonminimal supersymmetric models .
of course for an exact determination of the susy parameters and discrimination between susy models a precise analysis also of the decay characteristics is indispensable . for a special scenario
this was done in @xcite , the extension to the whole parameter space is planned . in the following sections 2 and 3 we introduce the considered susy models and show the unpolarized cross section for the neutralino production .
crucial for the understanding of the polarization effects are the couplings of the left and right selectrons to the produced neutralinos that will be discussed in section 4 .
a comprehensive comparison of the longitudinally polarized cross sections in the mssm and the extended models in section 5 concludes our contribution .
we shortly present the parameters and the neutralino sectors of the supersymmetric models . in the mssm the neutralino sector depends on the gaugino mass parameters @xmath6 and @xmath7 , the higgsino mass parameter @xmath8 and the ratio of the higgs vacuum expectation values @xmath9 . in this paper
we assume the gut - relation @xmath10 and scan over the @xmath11-@xmath8 plane with fixed @xmath12 .
the longitudinally polarized cross sections for neutralino production are compared with two extended susy models .
the nmssm is the simplest extension of the mssm by a higgs singlet field with vacuum expectation value @xmath13 and hypercharge @xmath14 @xcite .
the masses and couplings of the five neutralinos depend on @xmath6 , @xmath7 , @xmath15 , @xmath13 and the trilinear couplings @xmath16 and @xmath17 in the superpotential . within the nmssm ,
a light singlino - like neutralino can not be experimentally excluded by lep @xcite . in order to obtain a light singlino - like neutralino @xmath18 we choose in the following @xmath19 gev and @xmath20 . then nearly independently of the other parameters the @xmath18 has singlino character and a mass between about 9 and 28 gev while the masses and mixings of the heavier neutralinos correspond to the mssm with @xmath21 .
models with additional u(1 ) factors in the gauge group are a further extension of the mssm beyond the nmssm
. they can be motivated by superstring theory @xcite and imply an @xmath1 group as effective gut group , which is broken to an effective low energy gauge group of rank 5 with one additional @xmath22 factor .
this @xmath1 model contains one new gauge boson @xmath23 and an extended higgs sector with one singlet superfield with vacuum expectation value @xmath13 as in the nmssm @xcite . to respect the experimental mass bounds of new gauge bosons @xcite the vacuum expectation value of the singlet
must be larger than 1 tev .
the extended neutralino sector in the @xmath1 model contains six neutralinos , one @xmath24 gaugino and one singlino @xmath25 in addition to the mssm @xcite .
the @xmath26 neutralino mass matrix depends on six parameters : the @xmath27 , u(1)@xmath28 and @xmath22 gaugino mass parameters @xmath6 , @xmath7 and @xmath29 , @xmath15 , @xmath13 and the trilinear coupling @xmath16 in the superpotential . with the gut relation @xmath30 between the u(1 ) gaugino mass parameters the four lighter neutralinos always have mssm - like character if the mssm parameter @xmath8 is identified with @xmath31 @xcite , and the two heaviest neutralinos have large @xmath24 and @xmath25 components . assuming @xmath32 , however
, the lightest neutralino can be a nearly pure singlino @xcite .
then @xmath33 have mssm - like character and the @xmath34 has a mass @xmath35 .
such scenarios where the spectrum of the lighter neutralinos is similar to the nmssm will be discussed in the following .
table [ scentab ] gives an overview over the described fixed parameters , while the remaining parameters @xmath6 and @xmath8 ( or @xmath31 ) were scanned in the region @xmath36 gev , @xmath37 gev observing the experimental bounds from the unsuccessful chargino and neutralino search at lep2 @xcite .
also shown are the masses and singlino components of the lightest neutralino in the extended models which do not extensively vary over the scanned parameter region .
the selectron masses @xmath38 gev and @xmath39 gev are motivated by the desy / ecfa reference scenario @xcite and are fixed in all three models , for comparison .
.[scentab]parameters of the supersymmetric models . in the extended models
the @xmath40 has dominant singlino component @xmath41 in the whole @xmath6-@xmath31 parameter space .
the selectron masses @xmath38 gev and @xmath39 gev are fixed in all models . [ cols="^,^,^,^,^,^,^,^,^,^",options="header " , ]
in this section we address the question if a neutralino with a dominant singlino component can be directly produced with a sufficient cross section at a linear collider .
the neutralino production cross section in the mssm is well known @xcite .
[ mssm](a ) shows the contour lines for the unpolarized cross section @xmath42 of @xmath43 at a c.m.s .
energy @xmath44 gev in the @xmath6-@xmath8 parameter space .
depending on the masses and mixing characters of the light neutralinos it takes values up to 200 fb over most of the parameter space .
note the discontinuities ( bold lines ) for both positive and negative @xmath8 where @xmath45 and therefore the mixing character of @xmath46 changes @xcite . in the nmssm with a light singlino - dominated neutralino @xmath40 the unpolarized cross section ( fig .
[ nmssm](a ) ) is significantly smaller since the singlino component does not couple to the @xmath47 boson and the selectrons .
it is only the small contribution from the other components of @xmath18 that accounts for the cross section .
nevertheless one gets even for a neutralino with a singlino component larger than 90% production cross sections up to 30 fb which are expected to be clearly visible at a high luminosity linear collider ( @xmath48fb ) @xcite .
the use of polarized beams can still enhance these cross sections as we discuss in the following sections . in the @xmath1 model we explicitly study the dependence of the unpolarized cross sections on the parameter @xmath29 in figs .
[ e6minus](a ) [ e6plus35](a ) . for the large negative value @xmath49 tev ( fig .
[ e6minus](a ) ) the masses and the mixings of the light neutralinos are similar as in the nmssm .
therefore one also obtains similar cross sections up to 9 fb . for positive @xmath29 ( figs .
[ e6plus](a ) , [ e6plus35](a ) ) , however , the singlino component is larger than @xmath50 in the whole parameter space .
thus the cross sections are reduced to maximum values of about 0.2 fb for @xmath51 and 0.5 fb for @xmath52 leading to at least 100 respective 250 events at the expected high luminosity linear collider .
( 16,8.5 ) ( 0,.5 ) ( 3.5,.1)@xmath53gev ( 0,8.1)@xmath54gev ( 3.85,8.1)(a ) ( .7,2.75)120 fb ( 5.55,3)120 fb ( .7,3.9)80 fb ( 6.8,3.4)80 fb ( .7,4.8)40 fb ( 6.8,4.9)40 fb ( 2.8,7.5)160 ( 4.9,7.5)160 ( 2.9,7.2)fb ( 5.0,7.2)fb ( 8.3,.5 ) ( 11.8,.1)@xmath53gev ( 8.3,8.1)@xmath54gev ( 12.1,8.1)(b ) ( 10.7,6 ) ( 14.2,6 ) ( 9.7,3.5 ) ( 15,3.5 ) ( 16,8.5 ) ( 0,.5 ) ( 3.5,.1)@xmath55gev ( 0,8.1)@xmath54gev ( 3.85,8.1)(a ) ( .7,3.05)5 fb ( 7.0,3.95)5 fb ( .7,5.5)1 fb ( 6.95,7.2)1 fb ( 0.7,7.15)0.5 fb ( 8.3,.5 ) ( 11.8,.1)@xmath55gev ( 8.3,8.1)@xmath54gev ( 12.1,8.1)(b ) ( 9.5,4.5)@xmath56 ( 13.9,3.7)@xmath56 ( 16,8.5 ) ( 0,.5 ) ( 3.5,.1)@xmath55gev ( 0,8.1)@xmath54gev ( 3.85,8.1)(a ) ( .7,2.55)5 fb ( 7.0,2.85)5 fb ( .7,3.65)1 fb ( 7.0,4.1)1 fb ( .7,4.75)0.5 fb ( 6.75,5.55)0.5 fb ( 8.3,.5 ) ( 11.8,.1)@xmath55gev ( 8.3,8.1)@xmath54gev ( 12.1,8.1)(b ) ( 9.5,4.5)@xmath57 ( 13.9,3.7)@xmath57 ( 16,8.5 ) ( 0,.5 ) ( 3.5,.1)@xmath55gev ( 0,8.1)@xmath54gev ( 3.85,8.1)(a ) ( .7,3.15)0.1 fb ( 5.15,7.5)0.5 fb ( 6.2,7.5)0.1 fb ( 8.3,.5 ) ( 11.8,.1)@xmath55gev ( 8.3,8.1)@xmath54gev ( 12.1,8.1)(b ) ( 14.5,4.5 ) ( 11.2,7 ) ( 9.8,5.5 ) ( 16,8.5 ) ( 0,.5 ) ( 3.5,.1)@xmath55gev ( 0,8.1)@xmath54gev ( 3.85,8.1)(a ) ( .7,2.55)0.1 fb ( 5.18,7.5)0.5 fb ( 6.05,7.5)0.1 fb ( 8.3,.5 ) ( 11.8,.1)@xmath55gev ( 8.3,8.1)@xmath54gev ( 12.1,8.1)(b ) ( 11.3,6 ) ( 14.3,4 ) ( 15.3,7.3 ) ( 10,4 ) ( 15.4,5.6 )
if apart from the singlino component which does not contribute to the production process both produced neutralinos @xmath18 and @xmath46 have dominant gaugino components , the contribution of @xmath47 exchange can be neglected for energies above the @xmath47 peak .
then for longitudinally polarized beams the cross section consists of two terms describing the exchange of left and right selectrons @xmath58 .
the structure of @xmath59 is @xcite @xmath60 \tilde{\sigma } ( s , m_{\tilde{e}_{l / r}},m_{\tilde{\chi}^0_{1/2}})\ ] ] where @xmath61 and @xmath62 are the longitudinal polarizations of the electron and positron beam , respectively , and @xmath63 is a function of the beam energy and the particle masses . for selectron masses of the same order of magnitude @xmath64 the polarization behavior is significantly determined by the ratio @xmath65 of the couplings of the produced neutralinos to the left handed and right handed electrons and selectrons @xmath66 @xcite .
the effects of different masses , especially the extreme cases @xmath67 @xmath68 , are shortly mentioned in the following section . in this contribution
we consider longitudinal beam polarizations of @xmath69 for electrons and @xmath70 for positrons @xcite .
the solid lines in figs .
[ mssm](b ) , [ e6plus](b ) and [ e6plus35](b ) show the contour for @xmath71 . in the nmssm and the @xmath1 model with negative @xmath29 ( figs .
[ nmssm](b ) and [ e6minus](b ) ) it is always @xmath72 .
this contour line clearly separates the gray shaded regions with maximum cross sections for polarization configurations @xmath73 ( right handed polarized electrons and left handed polarized positrons ) or @xmath74 ( vice versa ) in the mssm ( fig .
[ mssm](b ) ) in the gaugino region ( @xmath75 ) . in the @xmath1 model ( figs .
[ e6plus](b ) and [ e6plus35](b ) ) the situation is more complicated because the higgsino components of the @xmath40 are larger than the gaugino components in the whole parameter space .
nevertheless the contour @xmath71 approximately describes the polarization behavior .
the polarization effects in detail are discussed in the following section .
we compare the longitudinally polarized cross sections @xmath76 and @xmath77 where @xmath73 and @xmath74 denotes the polarization configuration @xmath78 and @xmath79 , respectively . these polarization asymmetries may help to distinguish between the susy models .
[ mssm](b ) [ e6plus35](b ) present the regions where the different polarizations @xmath73 or @xmath74 lead to maximum cross sections .
polarization configurations with both beams polarized in the same direction or only one beam polarized are always smaller .
the following remarks apply to all models : if the gaugino - character of @xmath40 and @xmath80 is not too small ( which is the case in the mssm for @xmath81 ) @xmath77 dominates if the gaugino - coupling of the left selectron is larger than that of the right one , @xmath82 , and the mass difference between the selectrons is rather small .
analogously for @xmath83 , @xmath76 turns out to be the dominating cross section .
this situation may change for other selectron masses . if @xmath84 the exchange of the left ( right ) selectron in the neutralino production is strongly suppressed and @xmath76 ( @xmath77 ) dominates in the whole gaugino region , cf .
@xcite .
if the gaugino components of the produced neutralinos can be neglected , the ratio of the neutralino - selectron - electron couplings plays no role and the asymmetry in the mssm and nmssm @xmath85 depends only on the electron-@xmath47 couplings @xmath86 and @xmath87 .
so @xmath77 is generally largest in the mssm higgsino region @xmath88 , whereas in the nmssm the gaugino components of @xmath80 always cause @xmath89 for @xmath90 gev . in the @xmath1 model formula ( [ asyhiggs ] ) becomes more complicated due to @xmath23 exchange . because of @xmath91 in the nmssm and in the @xmath1 model with @xmath49 tev @xmath76 dominates in the whole parameter space . in the @xmath1 model with positive @xmath29 both cases @xmath72 and @xmath92 occur .
thus @xmath76 as well as @xmath77 can be largest .
the form of the respective parameter regions crucially depends on the value of @xmath29 ( compare figs .
[ e6plus](b ) and [ e6plus35](b ) ) and is only approximately given by the contours @xmath93 because here the higgsino character of @xmath40 is always rather large . in all models the cross sections for both beams polarized are more than two times larger than the unpolarized cross sections in a large fraction of the parameter space .
to describe this fact we define the ratio @xmath94 in the mssm for @xmath75 ( gaugino - region ) one nearly always obtains @xmath95 or @xmath96 in the respective regions . in the nmssm and the @xmath1 model with negative @xmath29 it is always @xmath97 and even @xmath98 ( nmssm ) or @xmath99 ( @xmath1 model ) in a large fraction of parameter space . and
also for positive @xmath29 @xmath100 or @xmath101 holds in large parameter regions .
thus polarization of both beams is an important tool to increase the event rates .
finally we describe the parameter regions ( @xmath36 gev , @xmath102 gev ) where significant differences between the susy models arise in neutralino production at a linear collider if both beams are polarized . in the extended models with a light singlino - like neutralino ( @xmath103 gev ) the polarization configuration @xmath76 dominates for @xmath52 , while in the mssm @xmath77 is largest apart from a narrow band between @xmath104 and @xmath105 . if the mass of the singlino - like neutralino , however , increases also the @xmath1 model with @xmath106 tev allows @xmath107 in some parameter regions . in the mssm , nmssm and @xmath1 model with negative @xmath29
the situation for @xmath8 or @xmath51 is similar to @xmath108 . in the @xmath1 model with positive @xmath29
the polarization asymmetries strongly depend on @xmath29 . for @xmath109 tev with a lighter singlino @xmath76 dominates for @xmath110 and @xmath111 contrary to the mssm , whereas for @xmath112 tev @xmath76 dominates only for large @xmath6 .
we summarize our results in two points : * a high luminosity linear collider is sufficient for the direct production of neutralinos with dominant singlino components in extended supersymmetric models .
polarization of both beams is an important tool to increase the event rates by a factor of 2 or 3 . *
the minimal supersymmetric standard model and extended supersymmetric models show a significant different polarization behavior for neutralino production in wide regions of the parameter space .
we thank g. moortgat - pick for many valuable discussions . this work was supported by the deutsche forschungsgemeinschaft ( dfg ) under contract no.fr 1064/4 - 1 , by the bundesministerium fr bildung und forschung ( bmbf ) under contract no .
05 ht9wwa 9 and by the fonds zur frderung der wissenschaftlichen forschung of austria , project no .
p13139-phy .
g. moortgat - pick , s. hesselbach , f. franke , and h. fraas , wue - itp-99 - 023 , hep - ph/9909549 , contribution to the proceedings of the _ 4th international workshop on linear colliders ( lcws99 ) _ , sitges , barcelona , spain , april 28 may 5 , 1999 .
gunion , l. roszkowski , and h.e .
haber , phys .
d 38 * ( 1988 ) 105 ; + m.m .
boyce , m.a .
doncheski , and h. knig , phys . rev . *
d 55 * ( 1997 ) 68 ; + m. cveti , d.a .
demir , j.r .
espinosa , l. everett , and p. langacker , phys .
d 56 * ( 1997 ) 2861 ; + t. gherghetta , t.a .
kaeding , and g.l .
kane , phys .
d 57 * ( 1998 ) 3178 .
j. ellis , k. enqvist , d.v .
nanopoulos and f. zwirner , nucl .
* b 276 * ( 1986 ) 14 ; + s. nandi , phys
b 197 * ( 1987 ) 144 ; + e. keith and e. ma , phys . rev . *
d 56 * ( 1997 ) 7155 ; + d. suematsu , phys . rev . *
d 57 * ( 1998 ) 1738 .
s. hesselbach , f. franke , and h. fraas , in _
@xmath4 linear colliders : physics and detector studies , part e _ , contributions to the workshops , frascati , london , munich , hamburg , ed .
r. settles ( , hamburg , 1997 ) p. 479 .
p. abreu et al .
( delphi collaboration ) , phys .
* b 446 * ( 1999 ) 75 ; + r. barate et al .
( aleph collaboration ) , cern - ep/99 - 014 ; + g. abbiendi et al .
( opal collaboration ) , cern - ep/99 - 123 , hep - ex/9909051 ; + m. acciarri et al .
( l3 collaboration ) , cern - ep/99 - 127 , hep - ex/9910007 .
s. ambrosanio , g.a .
blair , and p.m. zerwas , efca - desy linear collider workshop , available via www at + ` http://www.hep.ph.rhbnc.ac.uk/~blair/susy ` .
a. bartl , w. majerotto , and b. msslacher , in _
@xmath4 collisions at 500 gev : the physics potential , part b _ , proceedings of the workshop , munich , annecy , hamburg , ed .
p.m. zerwas , ( desy 92 - 123b , hamburg , 1992 ) p. 641 .
martyn , talk presented at the workshop _
2nd ecfa / desy study on physics and detectors for a linear electron - positron collider _ , oxford , march 20 23 , 1999 , transparencies available via www at + ` http://hepnts1.rl.ac.uk/ecfa_desy_oxford/scans/0120_martyn.pdf ` .
g. moortgat - pick , a. bartl , h. fraas , and w. majerotto , desy 00 - 002 , hep - ph/0002253 , contribution to the proceedings of the _
2nd joint ecfa / desy study on physics and detectors for a linear electron - positron collider_. | we discuss associated neutralino production @xmath0 with both beams polarized in the mssm , nmssm and an @xmath1 model .
it is shown that neutralinos with a large singlino component can be produced at a high luminosity linear collider in a broad parameter range .
polarization effects in the extended models are compared over the whole parameter space where the lightest neutralino is mainly a singlino .
we explain the complete different behavior of these models in some parameter regions , which may help to identify the respective supersymmetric model .
wue - itp-2000 - 008 lc - th-2000 - 025 hep - ph/0003272 * neutralino production with polarized beams in extended supersymmetric models * s. hesselbach , f. franke , h. fraas institut fr theoretische physik , universitt wrzburg , d-97074 wrzburg , germany |
charmonium has played a crucial role in the recent history of particle physics .
it has been nearly 30 years since its discovery in both @xmath20 interactions @xcite and in collisions of protons on a beryllium target @xcite . over the last two decades ,
the charmonium and bottomonium systems have served as a laboratory for testing qcd . in the weak sector
, charmonium also serves as a critical tool in extracting ckm @xcite phases in @xmath21-meson decays .
however , even after 30 years of studying @xmath22 systems , we still lack a complete understanding of their production mechanisms in glue - rich environments . about a decade ago , the cdf experiment reported production rates of charmonium in proton - antiproton collisions which exceeded the existing theoretical calculations by a factor of about 10 for @xmath8 and about a factor of 50 for @xmath15 @xcite .
an explanation of this excess was given by the so - called _ color - octet _
mechanism @xcite , whereby a gluon fragments into a color - octet @xmath23 @xmath24 pair , which then evolves non - perturbatively into a color - singlet by emission of a soft gluon .
the size of this non - perturbative matrix element is not predicted and was determined by a fit to the cdf data . because of the glue - rich environment and the @xmath25 suppression of the color - singlet process relative to the color - octet process , it was argued @xcite that the latter contribution is likely to be important . while this model can explain the rate and momentum spectrum of @xmath8 and @xmath15 production at the tevatron it appears that it does not properly describe recent @xmath8 polarization data from cdf @xcite .
furthermore , when the same matrix elements determined at cdf are applied to photoproduction of @xmath8 at hera , the color - octet contribution is about a factor of ten too large @xcite . over the last several years
, the role of the color - octet mechanism in @xmath8 production in @xmath26 collisions has been under theoretical study @xcite .
the dynamics of the color - octet processes are expected to give rise to significant differences in the @xmath8 momentum spectrum and production angle as compared to color - singlet production .
recently , both babar @xcite and belle @xcite have reported measurements of the cross section and the momentum spectra of @xmath8 s in @xmath20 collisions on the @xmath27 ( @xmath21 decays excluded ) or just below the @xmath28 ( _ i.e. , _ in the continuum ) .
@xmath29 ) pb , whereas belle measures a number which is 40% lower , ( [email protected]@xmath60.13 ) pb ( about 3 standard deviations below than the babar result ) .
the two experiments both observe similar shapes for the @xmath8 momentum spectrum , which are softer than the predictions of the color - octet model @xcite which predict a peaking of the @xmath8 momentum spectrum near the kinematic endpoint .
however , recent theoretical studies of the color - octet subprocesses , @xmath30 @xcite and @xmath31 @xcite , show that the perturbative expansion breaks down near the kinematic endpoint , and the authors appeal to soft - collinear effective theory ( scet ) to systematically include the non - perturbative effects . in ref .
@xcite , it is shown that by using scet the color - octet model predictions can be sufficiently softened and reasonably good agreement with @xmath32 data can be achieved , although the calculation is not completely predictive because it uses a shape function which is fit to the @xmath32 data @xcite .
belle also reports on production of @xmath15 in @xmath20 collisions , with a measured ratio @xmath33 .
that is , the production rates for @xmath8 and @xmath15 in @xmath20 collisions are approximately equal .
the color - singlet mechanism can yield the @xmath15 final state , but the expected ratio is @xmath3410% @xcite .
belle has extended their inclusive @xmath8 analysis to search for associated charmed particles , and they find : @xmath35 @xcite .
the color - octet contribution is expected to be at the level of 1% @xcite of the inclusive rate .
the disagreement here indicates that the production mechanisms of charmonium are not well understood , and more theoretical and experimental input is required .
several theoretical papers @xcite have suggested that the study of @xmath8 production in @xmath0 decays could provide an alternate probe of the charmonium system in that the @xmath0 decay provides a glue - rich environment in which @xmath8 mesons can be produced abundantly through the color - octet mechanism .
the kinematics of such @xmath8 s are expected to exhibit signatures distinct from other production mechanisms , such as a peak in the @xmath8 momentum spectrum near the kinematic endpoint .
the predicted branching fraction from color - octet processes is @xmath36 @xcite , with approximately 10% feed - down expected from @xmath15 and another 10% from @xmath37 @xcite ( summed over all @xmath38 ) .
color - singlet processes , such as @xmath39 start at @xmath40 , and are therefore suppressed relative to color - octet processes which enter at @xmath41 . however , computations of the color - singlet process @xmath42 @xcite indicate a sizeable branching fraction of @xmath43 , with about 10% coming from @xmath15 feed - down . the enhancement here arises because the non - perturbative color - singlet matrix element for @xmath44 may be 210 - 360 times larger than the corresponding color - octet matrix element , which is enough to compensate for the perturbative suppression .
moreover , unlike the color - octet processes , this process inherently results in a soft @xmath8 momentum spectrum because of the two additional charm quarks in the final state . as a result
, the @xmath8 momentum can not exceed about 3.3 gev/@xmath45 in this process .
therefore , while the color - octet and color - singlet processes give similar predictions for total rate , their momentum distributions are significantly different . figure [ jpsi_feyn2 ] shows feynman diagrams for ( a ) two of the more important color - octet processes and
( b ) the @xmath42 color - singlet diagram .
it should be noted that color - singlet production would also be signalled by the presence of additional charmed particles ( open charm ) in association with the @xmath8 . to capitalize on the small yield of @xmath8 s in @xmath0 decay , many @xmath46 decay channels , both inclusive and exclusive
will need to be explored .
we therefore relegate the search for open charm in association with @xmath8 in @xmath0 decay to a future report .
feynman diagrams for production of charmonium in @xmath0 decays from ( a ) color - octet processes and ( b ) color - singlet @xmath42 . for the color - octet processes , the @xmath8
is produced in a color - octet and becomes a color - singlet through emission of a soft gluon.,width=480 ] the process @xmath47 has been previously observed by cleo @xcite , where the branching fraction was measured to be ( [email protected]@xmath60.2)@xmath48 based on @xmath4920 observed events .
cleo also reported a soft momentum spectrum for the @xmath8 , albeit with limited statistical precision .
the argus collaboration reported an upper limit of 0.68@xmath48 @xcite at 90% confidence level .
the cleo collaboration has collected large data samples on the @xmath50(n@xmath51 ) resonances and currently has the world s largest samples of @xmath0 , @xmath50(2@xmath51 ) and @xmath50(3@xmath51 ) decays .
consequently , cleo is in a unique position to help clarify the roles of color - singlet and color - octet models in @xmath8 production . in this paper
, we present vastly improved measurements of the rate , momentum spectrum , and angular distributions in @xmath47 decays .
we also present first observations of the decays @xmath9 and evidence for @xmath52 .
the paper is organized as follows . in section [ sec : evsel ] we discuss the data samples used , the @xmath8 backgrounds , event selection and @xmath8 reconstruction .
section [ sec : upstojpsix ] details the measurement of the @xmath47 branching fraction and momentum spectrum in @xmath0 decays .
this section also includes a measurement of the cross section @xmath53 using data on and below the @xmath28 resonance , which is used to estimate and subtract the continuum contribution at the @xmath0 .
the section concludes with an examination of some event - level distributions .
section [ sec : psitwos ] presents the measurement of the @xmath9 branching fraction .
the report then discusses in section [ sec : upstochicx ] the measurement @xmath54 . for each of these analyses ,
we present a cross - check by measuring the corresponding branching fraction in @xmath21-meson decay .
lastly , we discuss in section [ sec : sys ] the systematic uncertainties in each of these analyses , and the paper is concluded in section [ sec : summary ] .
the analysis presented here uses data collected using the cleo iii detector @xcite .
the primary data sample includes 1.2 @xmath55 of data collected on the @xmath0 , and amounts to ( [email protected])@xmath56 @xmath0 decays . for background determinations and systematic checks ,
we also utilize 5.0 @xmath55 of data on the @xmath28 resonance ( 10.4 million @xmath21-meson decays ) and 2.3 @xmath55 just below ( @xmath4910.56 gev ) the @xmath28 resonance .
we also use the on-@xmath28 data for cross - checks on charmonium yields in @xmath21-meson decays .
the backgrounds to @xmath57 on the @xmath0 are : ( a ) radiative bhabha events , ( b ) @xmath58 fusion producing @xmath37 which subsequently produces @xmath59 , ( c ) radiative return processes such as @xmath60 or @xmath61 , and ( d ) continuum production ( @xmath32 ) .
various event selection requirements are targeted at reducing or eliminating these backgrounds .
radiative bhabha events produce background in the @xmath8 mass region when one of the hard leptons is combined with a soft lepton from the converted photon .
such events are suppressed by requiring that the invariant mass of either electron from the @xmath4 candidate with any other electron in the event have @xmath62 mev/@xmath63 .
events produced through @xmath64 fusion typically only leave two charged tracks in the cleo iii detector , and these events are therefore easily rejected by a requirement of at least 3 charged tracks .
the radiative return backgrounds are suppressed through event selection criteria which take advantage of the special kinematics of these processes , namely a low particle multiplicity coupled with either the detection of a high energy photon ( @xmath494 gev ) or large missing event momentum .
events are required to have their missing event momentum magnitude , @xmath65 gev/@xmath45 , or , if the number of charged tracks , @xmath66 , we require @xmath67 gev/@xmath45 . when the high energy photon is detected ( or an @xmath20 pair with invariant mass less than 100 mev/@xmath63 ) , the event is vetoed if @xmath66 and the ( converted ) photon has energy greater than 3.75 gev .
the remaining background from these three sources to the @xmath47 signal is negligible .
however , because of the small signal in @xmath9 , the remaining background can not be neglected . this background is determined using the evtgen mc followed by a geant - based detector simulation , and the resulting contribution is subtracted from the observed yields .
continuum background is reduced by requiring that the second fox - wolfram moment @xcite , @xmath68 .
the remainder of this background is estimated using @xmath28 data , and is statistically subtracted from the observed @xmath0 yields .
the estimate of this background is discussed in sections [ sec : bkest ] and [ sec : bkexp ] .
candidate @xmath8 s are formed by pairing oppositely charged electron or muon candidates .
these charged track candidates are required to have momentum in the range from 0.1 to 5.3 gev/@xmath45 and have at least 50% of the maximum number of expected hits in the tracking system .
we also require these tracks to be consistent with coming from the interaction point in three dimensions .
electron candidates are additionally required to have a shower profile which is consistent with expectations for an electron and an energy deposition in the calorimeter , @xmath69 , which is compatible with its measured momentum , @xmath70 , by requiring @xmath71 . for these electron candidates ,
we correct for radiated photons by adding back the momentum of the highest energy photon which lies within a 5@xmath72 cone of the initial particle direction .
muon candidates are formed using charged tracks which penetrate at least 3 hadronic interaction lengths of iron absorber in the muon chambers @xcite .
figure [ mass_1s_all ] shows the invariant mass distribution of @xmath8 candidates for ( a ) @xmath3 and ( b ) @xmath4 in the @xmath0 on - resonance sample satisfying all selection criteria .
the shaded histograms show the corresponding distributions for @xmath28 continuum data , scaled by a factor of 0.65 , which accounts for the differences in luminosities and center of mass energies .
the mass distributions from the @xmath0 data set are fit to the sum of a linear background and a gaussian signal whose means and widths are allowed to float .
the fitted peaks have a resolution of 13.4 mev/@xmath63 and 14.2 mev/@xmath63 for the @xmath3 and @xmath4 channels , respectively .
the fitted yields are 399@xmath625 @xmath3 and 449@xmath627 @xmath4 signals events .
dilepton invariant mass distributions for ( a ) @xmath73 and ( b ) @xmath74 candidates for data taken on the @xmath0 resonance ( points ) and data taken just below the @xmath28 resonance ( shaded ) .
the @xmath28 distributions are scaled to account for the different integrated luminosities and center of mass energies for the two data samples.,width=336 ] to study the momentum distribution , we divide the data into bins of scaled momentum , @xmath75 , where @xmath76 . here
, @xmath77 is the maximum @xmath8 momentum assuming the @xmath8 is recoiling against a massless particle , @xmath78 is the square of the center of mass energy , @xmath79 is the momentum of the @xmath8 candidate and @xmath80 is the @xmath8 mass @xcite .
the data are binned in intervals of @xmath81 .
this scaled momentum variable removes the beam - energy dependence which is useful in comparing spectra on the @xmath0 and the @xmath28 .
the invariant mass distributions for @xmath3 and @xmath4 for @xmath0 data in bins of @xmath75 are shown in fig .
[ mass_x_mm ] and fig .
[ mass_x_ee ] , respectively .
if the @xmath75 distribution has a sharp peak near the kinematic endpoint , there may be smearing into the @xmath82 region .
the absence of any signal in the @xmath83 bin shows that all events are containe within the physically - allowed region .
a simulation of the @xmath8 signal ( see section [ sec : receff ] ) indicates that the widths of the invariant mass distributions are independent of @xmath8 momentum , and therefore these distributions are fit using a width fixed to the values obtained from the full sample .
invariant mass distributions for @xmath3 candidates in @xmath75 bins of size 0.2 for data taken on the @xmath0 resonance.,width=336 ] invariant mass distributions for @xmath4 candidates in @xmath75 bins of size 0.2 for data taken on the @xmath0 resonance.,width=336 ] the continuum contribution to the @xmath47 signal is estimated using data taken on and below the @xmath28 .
this measurement is interesting in itself in light of the disagreement in the rates for @xmath32 measured by babar @xcite and belle @xcite .
we employ the same event selection criteria as for the data taken on the @xmath0 , except that for the on-@xmath28 data , we require the @xmath8 to have momentum larger than 2 gev/@xmath45 , which eliminates contributions from @xmath21-meson decay . the measured signal for @xmath32 below the @xmath28 is shown in fig .
[ y4scont_jpsix_tot ] for ( a ) @xmath3 and ( b ) @xmath4 .
the fitted numbers of events are 112@xmath617 ( @xmath3 ) and 116@xmath619 ( @xmath4 ) .
the corresponding distributions for data taken on the @xmath28 resonance are shown in fig .
[ y4s_p2_jpsix_tot ] .
the fitted yields are 130@xmath617 @xmath3 and 193@xmath624 @xmath4 events .
the yields per unit luminosity are statistically compatible , after correcting the on-@xmath28 yield for the 2 gev/@xmath45 momentum requirement .
the correction is determined from the @xmath8 momentum spectrum from the below-@xmath28 continuum and is estimated to be @xmath84 .
invariant mass distribution for ( a ) @xmath3 and ( b ) @xmath4 in data taken below the @xmath28 .
the data are integrated over all momenta.,width=336 ] invariant mass distributions for ( a ) @xmath3 and ( b ) @xmath4 in data taken on the @xmath28 resonance . to reject @xmath8 s from @xmath21 decay
, we require the momentum of the @xmath8 to be larger than 2.0 gev/@xmath45.,width=336 ] the data are corrected for geometric acceptance and analysis requirements using the pythia monte carlo @xcite and a geant - based detector simulation @xcite . the reconstruction efficiency as a function of @xmath75 and @xmath85 , where @xmath86 is the polar angle of the @xmath8 in the lab frame ,
is shown in fig .
[ upstojpsi_receff ] .
the circular points are for @xmath3 and the triangles are for @xmath4 .
the efficiencies decrease slightly with increasing momentum and @xmath87 and have average values of ( 40@xmath62)% for @xmath3 and
( 50@xmath62)% for @xmath4 .
the small drop in efficiency with momentum is a result of not reconstructing the softer lepton which is emitted backward in the @xmath8 rest frame .
the lower @xmath3 reconstruction efficiency is due to the requirement that both muons penetrate at least three layers of iron absorber , which limits the muon momentum to be larger than about 1 gev/@xmath45 .
the momentum distributions of both the @xmath47 signal as well as the on-@xmath28 and below-@xmath28 yields are corrected using these @xmath75-dependent efficiencies .
this is justified since the reconstruction efficiency is not sensitive to small differences in the event environment ( between @xmath47 and @xmath32 ) .
continuum - produced @xmath8 s have a similar charged - track multiplicity to @xmath47 ( @xmath497 - 8 ) and the @xmath0 data peak at low @xmath88 ( see section [ sec : evlev_dists ] , and figures therein ) as do @xmath88 measurements in the continuum ( see fig.1 ( c),(d ) in ref .
@xcite ) .
efficiency for reconstructing @xmath8 s in @xmath0 decays as a function of ( a ) scaled @xmath8 momentum , and ( b ) @xmath85 .
the circles are for @xmath3 and the triangles are for @xmath4.,width=336 ] the resulting differential cross sections , @xmath89 , versus @xmath75 , are shown in fig .
[ y4s_x_dist ] using the combined on-@xmath28 and below-@xmath28 data .
the circles represent @xmath3 and the triangles are @xmath4 .
the distributions clearly peak at large @xmath75 values with a mean of about 0.7 . integrating these distributions , and using @xmath90=5.9% @xcite
, we find @xmath91(stat ) pb for @xmath3 and @xmath92(stat ) pb for @xmath4 .
combining these results we obtain @xmath93(stat ) pb . the results using the different lepton species are consistent with one another and lie between the babar and belle measurements of ( @xmath94 ) pb and ( [email protected]@xmath60.13 ) pb , respectively . the rates found in the continuum are about a factor of 6 - 7 lower than on the @xmath0 .
distributions in @xmath75 for @xmath32 using the combined data taken on and just below the @xmath28 .
the circles show the results obtained using @xmath3 and the triangles show the corresponding distribution obtained using @xmath4.,width=336 ] the extrapolation of the differential cross section for @xmath32 on and below the @xmath28 ( see fig .
[ y4s_x_dist ] ) to the @xmath0 requires that we take into account the differences between these two samples , and includes two factors ( other than the luminosity scaling ) : the ratio of partonic cross sections for @xmath32 and a phase space correction for producing the @xmath95 final state .
for the former , we assume @xmath96 scaling , since the process proceeds through a virtual photon , and therefore the parton - level cross section at 9.46 gev is 1.25 times larger than at 10.58 gev . for the phase space extrapolation
, we bound this factor at unity by assuming the phase space at 9.46 gev is equal to that at 10.58 gev . to obtain a lower bound
, we assume that the @xmath8 s are always produced in association with a pair of @xmath46 mesons , which has a significantly reduced phase space at 9.46 gev as compared to 10.58 gev . using pythia ,
we determine that the probability , of producing @xmath97 at 9.46 gev is 55% of the corresponding value at 10.58 gev . using these values as extremes , and assuming that the `` true '' value has a flat probability of lying somewhere in that interval
, we estimate the phase space ratio is @xmath98 . combining the two factors , we determine the continuum extrapolation factor , @xmath99
. for the @xmath8 momentum spectrum in @xmath0 decays , we are primarily interested in the shape for the gluonic intermediate states .
the @xmath100 intermediate state , which proceeds through the coupling of the @xmath0 to a virtual photon , is assumed to have the same shape in @xmath75 as in @xmath32 , and therefore is more closely related to the predictions for @xmath8 production in the continuum .
therefore for the purposes of the momentum spectrum , we subtract the expected @xmath101 contribution .
this contribution is included for the branching fraction measurement .
any potential interference between the continuum and the @xmath102 contributions is neglected .
we express the @xmath102 rate relative to the corresponding rate for @xmath32 using , @xmath103 here , @xmath104 is shorthand for @xmath105 .
the measured value for @xmath106 is [email protected] nb @xcite . in that same reference ,
the theoretical value for @xmath107 at 9.46 gev is estimated to be 1.12 nb @xcite . a more recent estimate based on the fpair monte carlo ( mc )
simulation @xcite gives a larger cross section of about 1.38 nb .
taking the average of these two cross sections as our central value and half their difference as the uncertainty , we obtain @[email protected] nb .
we therefore estimate that the @xmath101 contribution is @xmath108 of the @xmath32 at @xmath109=9.46 gev . adding this contribution to the continuum extrapolation factor , @xmath110
, we obtain an overall extrapolation factor for the @xmath75 spectrum of @xmath111 .
figure [ y1s_y4s_x_dist ] shows the differential cross sections in @xmath75 for ( a ) @xmath3 and ( b ) @xmath4 using data taken on the @xmath0 ( solid circles ) and averaged results from the data taken on and below the @xmath28 ( triangles ) .
the latter have been scaled as discussed above to include both the continuum and @xmath112 contributions .
the differential cross section ( versus @xmath75 ) for @xmath47 is given by the difference of these two distributions , and reflects only the contributions from gluonic intermediate states .
the results are shown in fig .
[ y1s_sub_x_dist ] using @xmath3 ( circles ) and @xmath4 ( triangles ) .
the figure also shows the theoretical predictions of the color - octet @xcite ( solid line ) and the color - singlet @xmath113 @xcite ( dashed line ) model .
differential cross sections in @xmath75 for data taken on the @xmath0 ( points ) and data taken both on and just below the @xmath28 ( triangles ) .
the latter have been scaled to account for the @xmath112 contribution , as discussed in the text .
the upper figure shows the results obtained using @xmath3 and the lower figure shows the corresponding distributions obtained using @xmath4.,width=336 ] differential cross sections in @xmath75 for @xmath47 obtained using @xmath3 ( circles ) and @xmath4 ( triangles ) .
we also show the theoretical expectations based on the color - octet ( solid line ) @xcite and color - singlet ( dashed line ) @xcite models.,width=336 ] the branching fraction for @xmath47 is computed by integrating the differential cross section distribution .
we only subtract the expected continuum contribution ( we use @xmath110 as our extrapolation factor as opposed to @xmath114 ) , so that the branching fraction includes the three intermediate hadronic states : @xmath115 , @xmath116 and @xmath100 .
the resulting branching fractions for the @xmath3 and @xmath4 final states are : @xmath117 systematic uncertainties are presented in section [ sec : sys ] . using @xmath118 )
kev @xcite , our measurement corresponds to partial widths , @xmath119 of ( @xmath120 ) ev and ( @xmath121 ) ev for the @xmath3 and @xmath4 channels , respectively .
subtracting the expected @xmath112 contribution , we obtain @xmath122 of ( @xmath123 ) ev and ( @xmath124 ) ev .
in other words , about 90% of the @xmath8 rate comes from the @xmath115 and @xmath116 intermediate states .
the @xmath116 contribution is only expected to be at the level of about 5% @xcite of the @xmath115 rate .
theoretical estimates of this rate based only on color - octet contributions , which neglect the @xmath100 intermediate state , give a total branching fraction of @xmath125 @xcite .
those predictions are in good agreement with the measurements reported here . on the other hand ,
our measured momentum spectrum is significantly softer than predicted by the color - octet model , which is expected to peak near the kinematic limit ( see fig .
[ y1s_sub_x_dist ] ) .
however , it has been recently pointed out @xcite that in a similar process , @xmath32 , the non - relativistic calculations break down near the kinematic endpoint where there are large perturbative and non - perturbative corrections .
these effects may be systematically treated using so - called soft - collinear effecive theory ( scet ) and are expected to soften the @xmath8 momentum spectrum .
using scet , the shape of the measured @xmath8 momentum spectrum in @xmath32 , which peaks near @xmath126 , was shown to be reproducible @xcite .
it will be interesting to see if these corrections , when applied to @xmath47 , can soften the color - octet predictions sufficiently to bring them into agreement with our data .
our measured rate is also consistent with the predictions of the color - singlet process @xmath42 , which predicts a branching fraction of @xmath43 and a soft momentum spectrum which peaks at @xmath127 and has a kinematic limit of @xmath128 .
while the data is somewhat softer than the color - singlet predictions , it should be noted that this is a parton - level calculation and neglects the hadronization process .
inclusion of the hadronization of the charm quarks to charm hadrons softens the @xmath8 momentum spectrum , with more softenining occuring as the mass of the recoiling system increases .
further softening of the @xmath8 momentum spectrum occurs when including the feed - down of @xmath15 and @xmath37 to @xmath8 . using a pythia simulation of the color - singlet process
, we are able to obtain reasonably good agreement in the @xmath129 region using our measured values for the feed - down from @xmath15 , @xmath37 to @xmath8 along with a reasonable , but arbitrary admixture of recoiling @xmath46 , @xmath130 , and @xmath131 states .
this is not necessarily evidence for color - singlet production , but it is suggestive .
angular distributions have the potential to differentiate the mechanisms for @xmath8 production in @xmath20 collisions .
theoretical predictions for the production and helicity angle distributions for continuum production are available @xcite , but the calculations are yet to be done for @xmath0 decay . in the same spirit
, we present distributions of the ( polar ) production angle , @xmath85 , of the @xmath8 and the helicity angle , @xmath132 , where @xmath133 is the angle between the positive lepton momentum in the @xmath8 rest frame and the @xmath8 momentum in the lab frame .
the efficiency - corrected @xmath3 and @xmath4 channels are combined and shown in fig .
[ y1s_ang_dists ] . here , we subtract the expected @xmath112 and continuum contributions to extract the @xmath115 and @xmath116 shapes .
the normalizations are arbitrary .
the top figure shows the distribution of @xmath132 for @xmath0 data ( points ) , pythia simulation ( dashed ) , and a fit ( dotted ) to the form @xmath134 , from which we find @xmath135 ( @xmath136/dof=0.50 ) .
the bottom figure shows the distribution in @xmath85 for @xmath0 data ( points ) , pythia simulation ( dashed ) and a fit ( dotted ) to the form @xmath137 , from which we find @xmath138 ( @xmath136/dof=1.44 ) .
the functional forms are the same as those used to describe the angular distributions for continuum production of @xmath8 mesons @xcite .
the negative value of @xmath139 indicates that the @xmath8 has a significant longitudinal polarization component ( a positive value would indicate transverse polarization ) . for continuum production of @xmath8 ,
the color - octet and color - singlet models differ greatly on their expectations for @xmath21 at large values of scaled momentum , with @xmath140 for the color - singlet model @xcite and @xmath141 for the color - octet model @xcite .
if a large difference persists for @xmath0 decay , the production angle distribution could be useful in differentiating these two mechanisms .
we note that the pythia simulation ( using default parameters ) , which produces @xmath8 via @xmath142 , appears to be in reasonable agreement with data . the helicity angular distributions , @xmath132 ( top ) and production angle , @xmath85 ( bottom ) of the @xmath8 in @xmath47 .
in each case , the points are the @xmath0 data , the dashed histogram is a pythia simulation of @xmath47 and the dotted line is a fit to the @xmath0 data as described in the text.,width=336 ] additional information on the @xmath47 process can be obtained by studying various event - level distributions .
we present distributions of : * number of reconstructed charged tracks , @xmath143 ( fig . [ y1s_dists](a ) ) * reconstructed neutral energy in the crystal calorimeter , @xmath144 ( fig . [ y1s_dists](b ) )
* the second fox - wolfram moment , @xmath88 ( fig . [ y1s_dists](c ) ) * invariant mass recoiling against the @xmath8 , @xmath145 ( fig .
[ y1s_dists](d ) ) in each case , we have performed a sideband subtraction , where the signal region is defined to be from @xmath146 gev/@xmath63 and the sideband region includes the @xmath147 regions from @xmath148 and @xmath149 gev/@xmath63 .
the relatively small continuum contribution has not been subtracted . for each distribution
, we also show the corresponding distribution from the pythia mc simulation , which primarily produces a final state which consists of @xmath97 .
the data are shown as points ( @xmath3 are circles and @xmath4 are triangles ) and the simulation is the histogram .
we find that the charged particle multiplicity , which includes all charged particles , has a mean of about nine . the neutral energy , which comprises all energy in the calorimeter which is not associated with charged tracks , has an average of about ( @xmath150 ) gev , with most of the events having less than about 3.5 gev . the fox - wolfram moment , @xmath88 , peaks at low @xmath88 in @xmath47 data which indicates that these events tend to be more spherical than collimated ( jetlike ) .
the recoil mass can be used to discern whether there is another particle recoiling against the @xmath8 .
it is defined by : @xmath151 where @xmath78 is the square of the center - of - mass energy , and @xmath79 and @xmath152 are the momentum and energy of the @xmath8 candidate .
we do not observe any significant peaks in the recoil mass spectrum , indicating that the @xmath8 is usually not accompanied only by a second ( bound ) @xmath22 meson .
the color - octet model predicts @xmath1531% contribution to the inclusive rate whereas the color - singlet model does not predict the fraction of recoiling charm which is in the form of charmonium .
sideband - subtracted distributions of ( a ) number of charged tracks , @xmath143 , ( b ) neutral energy , @xmath144 , ( c ) the second fox - wolfram moment , @xmath88 , and ( d ) the recoil mass , @xmath145 , for @xmath0 data and pythia simulation .
the points ( triangles ) correspond to @xmath3 ( @xmath4 ) , and the histograms are the corresponding distributions from the simulation.,width=336 ] as a cross - check of our detector simulation and analysis procedure , we use the same tools to measure @xmath155 ) in @xmath28 data .
the efficiencies for reconstructing @xmath8 in @xmath154 events are about 5% lower than in @xmath47 .
in addition to the selection requirements described in section [ sec : evsel ] , we require the @xmath8 momentum to be less than 2.0 gev/@xmath45 .
the yields are corrected for the expected continuum contribution , which is typically at the level of 1 - 2% of the @xmath154 yield .
the resulting branching fractions are found to be @xmath156(stat))% and @xmath157(stat))% for @xmath3 and @xmath4 , respectively .
these results are slightly higher than the world average value of @xmath158% @xcite .
this difference is included as a systematic uncertainty in the @xmath8 reconstruction efficiency .
we search for @xmath9 using the decay mode @xmath159 .
pion candidates must pass the previously mentioned track selection criteria and must have a measured energy loss in the tracking chambers within four standard deviations of the expected value . using all pairs of oppositely charged pion candidates , we compute the invariant mass difference , @xmath160 , a quantity which has better resolution than @xmath161 .
we also require @xmath162 to be in the range from 3.00 - 3.14 gev/@xmath63 . the resulting distribution for @xmath160 is shown in fig . [ ups1s_psi2s_em ] , where we have summed over both lepton species .
the distribution is fit to the sum of a gaussian signal shape and a second - order polynomial background .
the width of the gaussian is fixed to 2.5 mev/@xmath63 , the value determined from @xmath163 data .
the 0.3 mev intrinsic width @xcite of the @xmath15 is negligible compared to the detector resolution , and is therefore ignored .
the fitted yield is 56@xmath611 events .
the significance of the signal , @xmath164 , where @xmath51 is the fitted signal and @xmath21 is the estimated background , varies from @xmath165 , depending on whether @xmath21 is estimated from the sidebands or the background function .
the radiative return background , @xmath61 , is estimated using the evtgen @xcite simulation package and published cross sections in ref .
the events are processed using geant and analyzed using the same analysis tools as the @xmath0 data .
the efficiency for these events to pass a loose hadronic event selection is ( [email protected])% for @xmath3 decays and
( [email protected])% for @xmath4 decays . for these subsamples
, a fraction , @xmath166 of @xmath3 decays also pass the analysis - specific selection criteria discussed in section [ sec : evsel ] .
the corresponding fraction for @xmath4 decays is @xmath167 . the larger efficiency for the electron channel to pass the loose hadronic event selection results from the use of the calorimeter in defining this subsample of events . with the assumption that all data events which fail the analysis requirements are radiative return ( discussed below )
, the expected background contribution in data from radiative return is computed using where the quantity in parentheses is the ratio of simulated radiative return events which pass the analysis - specific selection to those that are rejected .
the quantity @xmath169 is the number of rejected events in @xmath0 data for each lepton species .
we find @xmath170 and @xmath171 in the @xmath15 signal region , obtained through sideband subtraction .
we therefore estimate radiative return contributions of [email protected] events and [email protected] events in the @xmath3 and @xmath4 channels , respectively , and therefore a total of [email protected] background events from this source .
the assumption that the rejected events in data are from radiative return is supported by comparing event - level distributions of these rejected events between data and simulated @xmath61 radiative return events .
figure [ y1s_psitwos_rej ] shows comparisons of ( a ) number of reconstructed charged tracks , @xmath143 , ( b ) neutral energy in the calorimeter , @xmath144 , ( c ) missing event momentum , @xmath172 , and ( d ) the cosine of the angle between the @xmath15 direction and the beam axis , @xmath173 . in all cases
, the radiative return simulation reproduces the rejected events in @xmath0 data , indicating that the rejected data events are mostly from radiative return .
event - level distributions for events rejected in the @xmath15 analysis .
distributions shown are ( a ) number of charged tracks , ( b ) neutral energy , ( c ) missing event momentum , and ( d ) cosine of the @xmath15 production angle .
solid points are rejected @xmath9 candidate events and the histogram is the @xmath174 radiative return simulation.,width=336 ] in fig .
[ y1s_psitwos_acc ] we show the analogous distributions for @xmath0 data ( points ) passing all analysis selection requirements .
the corresponding distributions from a mc simulation of @xmath175 are overlaid ( histogram ) .
these distributions are clearly quite different than the distributions for rejected events ( see fig . [ y1s_psitwos_rej ] ) . event - level distributions for events accepted in the @xmath15 analysis .
distributions shown are ( a ) number of charged tracks , ( b ) neutral energy , ( c ) missing event momentum , and ( d ) cosine of the @xmath15 production angle .
solid points are accepted @xmath9 candidate events and the histogram is a @xmath176 mc simulation.,width=336 ] we use the integrated luminosity @xmath179 , and branching ratios of @xmath180 @xcite and @xmath181 .
the reconstruction efficiencies are determined using a pythia simulation of @xmath9 which is used to model the continuum as well as the signal , for the same reasons as mentioned previously .
the efficiencies for both @xmath3 and @xmath4 final states are nearly independent of momentum with average values , @xmath182=(17@xmath61)% for @xmath183 and ( @xmath184% for @xmath185 . for the background extrapolation
, we have assumed the same phase space suppression for @xmath15 as @xmath8 , and assign a 50% uncertainty to its value .
we therefore expect a continuum background contribution of [email protected] @xmath3 and [email protected] @xmath4 events , which sum to [email protected] events .
the error is dominated by the uncertainties in @xmath186 and @xmath110 . as a consistency check ,
we have searched our 2.3 @xmath55 continuum data sample for @xmath15 , and we find @xmath187 and 12@xmath64 events in the @xmath188 and @xmath189 channels , respectively . using the belle cross section measurement , we would have expected @xmath190 @xmath188 and @xmath191 @xmath189 events , which is consistent with our observations . combining the radiative return and continuum backgrounds
, we estimate a total background of [email protected] events . the uncertainty in the central value
is included as a systematic uncertainty ( see section [ sec : sys ] ) .
where @xmath194 ( @xmath195 ) is the total number of @xmath15 ( @xmath8 ) signal candidates for lepton species @xmath185 and @xmath183 , @xmath196 ( @xmath197 ) is the expected @xmath15 ( @xmath8 ) background , and @xmath182 ( @xmath198 ) is the average reconstruction efficiency for @xmath15 ( @xmath8 ) .
a summary of the inputs used for the @xmath9 branching fraction computation are presented in column 4 of table [ tab : psi2s ] .
the table also shows in columns 2 and 3 the values for the @xmath3 and @xmath4 channels separately .
the event yields are consistent with one another .
the @xmath47 branching fraction is consistent with predictions of both the color - octet mechanism for @xmath8 production in @xmath0 decays @xcite and color - singlet production via @xmath42 @xcite , each which predict a branching fraction at the level of 6@xmath7 .
the observed scaled momentum spectrum is relatively soft , peaking around @xmath199 , in contrast to @xmath8 s produced in the continuum , which peak at about 0.7 .
the peaking at low momentum is in sharp contrast to the prediction of the color - octet model which predicts a peaking of @xmath75 near 1 .
it is possible that incorporation of final state interactions could improve this agreement as was shown for @xmath32 @xcite . the observed spectrum is closer to , although softer than , the expectation of the color - singlet process @xcite , @xmath42 , which peaks near @xmath2000.5 .
when this parton - level calculation is simulated using pythia we are able to achieve satisfactory agreement in the region @xmath129 when hadronization of the recoiling charm quarks into charm hadrons is included , and our measured feed - downs of @xmath15 and @xmath37 to @xmath8 are incorporated .
the observation of @xmath9 is the first to a @xmath22 final state other than @xmath8 .
the feed - down to @xmath8 constitutes ( 24@xmath66@xmath65)% of the inclusive rate for @xmath47 , which is significantly larger than expected in either the color - octet @xcite or color - singlet model @xcite , each which predict a feed - down to @xmath8 at the level of 10% . our measured rates for @xmath10 yield feed - down contributions of ( 11@xmath63@xmath62)% for the @xmath38=1 state and ( 10@xmath62@xmath62)% for the @xmath38=2 state , which is also larger than the expected contribution of about 10% , summed over @xmath38=0 , 1 , and 2 @xcite .
these measurements can shed additional light on the role of the color - octet and color - singlet mechanisms in producing charmonium , not only in @xmath0 decays but also in @xmath20 and @xmath201 collisions . in this regard , it would be of great interest to determine whether the same softening mechanism applied to the color - octet prediction for @xmath32 @xcite can account for the @xmath8 momentum spectrum in @xmath47 . moreover ,
computation of the angular distributions for the color - octet and color - singlet mechanisms may provide additional discrimination between these two processes . from an experimental perspective , the additional information on the roles of color - singlet versus color - octet mechanisms may be obtained by measuring the ratio @xmath202 at the tevatron .
the unexpectedly large value for @xmath203 reported by belle @xcite may point to a large rate in @xmath201 collisions as well .
we gratefully acknowledge the effort of the cesr staff in providing us with excellent luminosity and running conditions .
this work was supported by the national science foundation , and the u.s .
department of energy .
we also thank kingman cheung and wai - yee keung for their assistance with the color - octet predictions and shi - yuan li for providing color - singlet predictions .
[ cleo collaboration ] , nucl .
instrum . meth .
a * 384 * , 61 ( 1996 ) ; a. wolf _
[ cleo collaboration ] , nucl .
instrum . meth .
a * 408 * , 58 ( 1998 ) ; g. viehhauser _
[ cleo collaboration ] , nucl .
instrum . meth .
a * 462 * , 146 ( 2001 ) ; y. kubota _ et al . _
[ cleo collaboration ] , nucl .
instrum . meth .
a * 320 * , 66 ( 1992 ) .
b. aubert _ et al . _ [ babar collaboration ] , phys .
d * 67 * , 032002 ( 2003 ) [ hep - ex/0207097 ] ; k. abe _ et al . _ [ belle collaboration ] , phys .
* 89 * , 011803 ( 2002 ) [ hep - ex/0202028 ] ; s. chen _ et al . _
[ cleo collaboration ] , phys .
d * 63 * , 031102 ( 2001 ) [ hep - ex/0010036 ] . | using data collected by the cleo iii detector at cesr , we report on measurements of @xmath0 decays to charmonium final states .
the data sample used for this analysis consists of @xmath1 @xmath0 decays , representing about 35 times more data than previous cleo @xmath0 data samples .
we present substantially improved measurements of the branching fraction @xmath2 using @xmath3 and @xmath4 decays .
the branching fractions for these two modes are averaged , thereby obtaining : @xmath5(stat)@xmath60.6(syst))@xmath7 .
a greatly improved measurement of the @xmath8 momentum distribution is presented and indicates a spectrum which is much softer than predicted by the color - octet model and somewhat softer than the color - singlet model .
first measurements of the @xmath8 polarization and production angle are also presented .
in addition , we report on the first observation of @xmath9 and evidence for @xmath10 .
their branching fractions are measured relative to @xmath2 and are found to be : @xmath11(stat)@xmath60.08(syst ) , @xmath12 = [email protected](stat)@xmath60.06(syst ) , @xmath13 = [email protected](stat)@xmath60.09(syst ) , and @xmath14 at 90% confidence level .
the resulting feed - down contributions to @xmath8 are ( 24@xmath66(stat)@xmath65(syst))% for @xmath15 , ( 11@xmath63(stat)@xmath62(syst))% for @xmath16 , ( 10@xmath62(stat)@xmath62(syst))% for @xmath17 , and @xmath18 at 90% confidence level for @xmath19 . these measurements ( apart from @xmath19 ) are about a factor of two larger than expected based on the color - octet model . 0 ( 1s)_c0+x 1 ( 1s)_c1+x 2 ( 1s)_c2+x f 4s ( 4s ) 4 ( 4s ) |
reliable lattice calculations of hadronic matrix elements and quark masses require both high precision numerical simulations and non - perturbative determinations of renormalization constants ( @xmath6-factors ) .
the cp - pacs collaboration recently carried out a sophisticated spectrum calculation in @xmath7 full qcd @xcite using a renormalization group ( rg ) improved gauge action and a tadpole improved clover quark action .
however , non - perturbative @xmath6-factors were not available for this action combination .
hence analyses had to rely on one - loop perturbative values .
as a first step toward a systematic study of non - perturbative renormalization for this action , we apply the schrdinger functional method @xcite to calculations of @xmath6 factors for vector ( @xmath0 ) and axial - vector ( @xmath1 ) currents in quenched qcd with the same improved action .
we examine in particular whether a large scaling violation of meson decay constants observed for this action @xcite is improved with non - perturbative @xmath6-factors .
we report preliminary results in these proceedings .
we follow the method developed by the alpha collaboration @xcite . namely , we use a lattice geometry of @xmath8 with @xmath9 for @xmath0 with a vector operator at @xmath10 , and @xmath11 for @xmath1 with two axial vector operators at @xmath10 and @xmath12 , except at @xmath13 and 2.4 for @xmath1 ( see sec .
[ sec : exceptional ] for details of this exception ) .
tree - level values are used for coefficients of boundary counter terms of the action . for improving the axial current
, we use the one - loop perturbative value for the coefficient @xmath14 .
values of @xmath0 and @xmath1 are determined for @xmath13 8.0 which almost covers the range of the cp - pacs quenched spectrum calculation @xcite , @xmath15 2.575 .
physical size is normalized at @xmath16 2.6 on an @xmath17 lattice . for other @xmath18 values
, two lattice sizes are analyzed to match the physical size using the string tension .
our action has @xmath19 errors since we employ a tadpole improved value of @xmath20 .
therefore we extrapolate / interpolate results linearly in @xmath21 .
we have analyzed 3004000 configurations depending on @xmath18 value and lattice size .
it is straight - forward to calculate @xmath6-factors for @xmath22 for @xmath0 and @xmath23 for @xmath1 .
we find reasonable plateaux in the ratio of green functions for the @xmath6-factors in spite of the @xmath19 error of the action , which implies viability of the schrdinger functional method for our action .
however , for lower @xmath18 values on a large lattice , anomalously large values appear in the ensemble of @xmath24 , @xmath25 and @xmath26 where @xmath27 and @xmath28 .
this is illustrated with a time history of @xmath24 and @xmath25 at @xmath29 in fig .
[ fig : f1fv ] .
= 6.0 cm in order to estimate @xmath6-factors at low @xmath18 values , we have investigated the properties of these `` exceptional configurations '' .
we find : i ) large values of @xmath24 and @xmath25 for @xmath0 and @xmath24 and @xmath26 for @xmath1 are strongly correlated ( see fig .
[ fig : f1fv ] ) .
ii ) histograms of @xmath30 s have a long tail toward very large values as shown in fig .
[ fig : hist ] .
we then impose a cutoff in taking the average of the @xmath30 s , and find that @xmath6-factors are stable against change of the cutoff as long as anomalously large values are discarded , as the numerator and denominator for @xmath6-factors are correlated and effects mostly cancel out .
see fig .
[ fig : cutoff ] .
we then estimate @xmath6-factors for low @xmath18 values taking a certain value of the cutoff . for @xmath1 at @xmath13 and 2.4 , the lattice geometry is also changed from @xmath11 to @xmath9 because `` exceptional configurations '' appear very frequently for the original geometry to the extent that the cutoff analysis above does not work .
we have checked at @xmath31 that the change of geometry does not lead to any significant difference in @xmath1 .
= 6.4 cm = 6.4 cm
= 6.4 cm = 6.4 cm in fig .
[ fig : result ] we show results of @xmath6-factors together with pad fits ( solid curves in the figure ) to them .
non - perturbative estimates give values smaller than the one - loop perturbative ones ( dashed lines ) by about 10 % ( 6% ) for @xmath0 ( @xmath1 ) at the largest coupling of the cp - pacs simulation , @xmath32 .
= 6.4 cm = 6.4 cm we compare in fig .
[ fig : decay ] @xmath4 and @xmath5 determined with non - perturbative ( filled circles ) and perturbative ( open circles ) @xmath6-factors . also shown
are the results from the standard plaquette and wilson action ( squares)@xcite using the perturbative @xmath6-factors .
we observe that , even with the non - perturbative @xmath6-factors , large scaling violation of meson decay constants remains for the range we have investigated .
a possible reason is the necessity of non - perturbatively fixing the @xmath19 and perhaps higher terms in the currents themselves . for the axial vector current
, it will be worth investigating if non - perturbative estimates of the @xmath19 coefficient @xmath14 yield a large value .
we have successfully applied the schrdinger functional method to calculations of @xmath0 and @xmath1 for the combination of a rg - improved gauge action and a tadpole improved clover quark action down to the lattice spacings @xmath33 gev where the quenched cp - pacs data for decay constants were taken .
while @xmath6-factors estimated non - perturbatively are smaller by @xmath2 than perturbative ones for this range , there still remain large scaling violations of @xmath19 and higher in meson decay constants with non - perturbative @xmath6-factors .
further work is needed to examine if hadronic matrix elements could be reliably extracted at lattice spacings much coarser than @xmath34 gev with operators improved non - perturbatively at @xmath19 and beyond .
this work is supported in part by grants - in - aid of the ministry of education ( nos . 10640246 , 10640248 , 11640250 , 11640294 , 12014202 , 12304011 , 12640253 , 12740133 , 13640260 ) .
vl is supported by the research for future program of jsps ( no . jsps - rftf 97p01102 ) . | renormalization constants of vector ( @xmath0 ) and axial - vector ( @xmath1 ) currents are determined non - perturbatively in quenched qcd for a renormalization group improved gauge action and a tadpole improved clover quark action using the schrdinger functional method .
non - perturbative values of @xmath0 and @xmath1 turn out to be smaller than the one - loop perturbative values by @xmath2 at @xmath3 gev . a sizable scaling violation of meson decay constants @xmath4 and @xmath5 observed with the one - loop renormalization factors remains even with non - perturbative renormalization . |
with the advances in acoustic communication technology , the interest in study and experimental deployment of underwater networks has been growing @xcite .
however , underwater acoustic channels impose many constraints that affect the design of wireless networks .
they are characterized by a path loss that depends on both the transmission distance and the signal frequency , a feature that distinguishes an underwater acoustic system from a terrestrial radio system .
thus , not only the power consumption , but also the useful bandwidth depend on the transmission distance @xcite . from an information theoretic perspective , both the distance between two nodes and the required capacity determine the power consumption for that link and the optimal transmission band .
it is thus of interest to have a simple , closed - form expression that relates the transmission power to the desired capacity .
this would enable an efficient design of both point to point links and underwater networks , eventually leading to a minimum cost overall network optimization .
thus , these expressions may be useful from both a theoretic and an engineering standpoint . in this paper ,
simple closed - form approximations for the power consumption and operating frequency band as functions of distance and capacity are presented .
this approximate model stems from an information theoretic analysis that takes into account a physical model of acoustic propagation loss , and colored gaussian ambient noise .
it was shown in @xcite that the transmission power as a function of the distance could be well approximated by @xmath0 .
a similar relationship was shown to exist for the operating bandwidth .
the coefficients in this model were determined as functions of the required signal to noise ratio .
the present work extends this idea of modeling the power and bandwidth as functions of distance , but the problem is cast into a slightly different framework .
namely , instead of using the snr as a constraint , i.e. a fixed design parameter , the desired link capacity is used as a figure of merit . in few words
, this work proposes approximate models for the parameters as functions of the capacity .
this resulting model is useful for a broad range of capacities and distances .
the paper is organized as follows . in section 2 , a model of an underwater channel
is outlined . in section 3 , a brief description of the numerical evaluation procedure is described . in section 4 ,
closed - form expressions for the parameters of interest are presented .
section 5 gives numerical results for different ranges of distance and capacity .
conclusions are summarized in the last section .
, @xmath1 and @xmath2 , by incrementing @xmath3 at each step by @xmath4 until a stopping condition is fulfilled , width=336,height=288 ] and @xmath5 for @xmath6 and approximate model.@xmath7 $ ] , @xmath8 $ ] , @xmath9,@xmath10 and @xmath11 , width=336,height=336 ] and @xmath5 for @xmath1 and approximate model.@xmath7 $ ] , @xmath8$],@xmath9,@xmath10 and @xmath11 , width=336,height=336 ] and @xmath5 for @xmath2 and approximate model.@xmath7 $ ] , @xmath8 $ ] , @xmath9,@xmath10 and @xmath11 , width=336,height=336 ] and @xmath5 for @xmath6 and approximate model.@xmath12 $ ] , @xmath13 $ ] , @xmath9,@xmath10 and @xmath11 , width=336,height=336 ]
an underwater acoustic channel is characterized by a path loss that depends on both distance _
l _ and signal frequency _
f _ as @xmath14 where _ k _ is the spreading factor and @xmath15 is the absorption coefficient @xcite .
the spreading factor describes the geometry of propagation , e.g. @xmath16 corresponds to spherical spreading , @xmath17 to cylindrical spreading , and @xmath9 to practical spreading .
the absorption coefficient can be expressed in db / km using thorp s empirical formula for @xmath18 in khz : @xmath19 for frequencies above a few hundred hz . for lower frequencies ,
the model is : @xmath20 the noise in an acoustic channel can be modeled through four basic sources : turbulence , shipping , waves , and thermal noise@xcite .
the following formulas give the power spectral density ( psd ) of these noise components in db re @xmath21 pa per hz as a function of frequency in khz : @xmath22 where the shipping activity @xmath23 ranges from 0 to 1 , for low and high activity , respectively , and @xmath24 corresponds to the wind speed measured in m / s .
the overall psd of the ambient noise is given by @xmath25 let us assume that this is a gaussian channel .
then , the capacity of this channel can be obtained using the waterfilling principle @xcite .
also , assume that the power and band of operation can be adjusted to reach a certain capacity level .
thus , the capacity of a point - to - point link is @xmath26 where @xmath27 is the optimum band of operation .
this band could be thought of as a union of non - overlapping intervals , @xmath28 $ ] , where each non - overlapping band @xmath29 has the lower - end frequency @xmath30 and the higher - end frequency @xmath31 associated to it . in its simplest form
@xmath32 $ ] .
the power consumption associated with a particular choice of @xmath33 is given by @xmath34 where @xmath35 .
evidently , these expressions are quite complicated to be used in a computational network analysis .
also , they provide little insight into the relationship between power consumption , @xmath36 and @xmath37 , in terms of the pair @xmath33 .
this motivates the need for an approximate model that will represent these relations for ranges of @xmath38 and @xmath39 that are of interest to acoustic communication systems .
the model should also provide flexibility to changing other parameters , such as the spreading factor @xmath40 , wind speed @xmath24 and shipping activity @xmath23 .
the dependence on the spreading factor @xmath40 is quite simple .
let us assume that a model for @xmath6 has been developed for a particular value of @xmath41 , i.e. @xmath42 . to determine @xmath43 for @xmath44 ,
let us note that for a change in @xmath40 , the product @xmath45 constitutes a constant scaling factor with respect to @xmath18 .
therefore , for a link of distance @xmath39 the term @xmath27 will remain unchanged .
thus , if the same capacity @xmath38 is required for @xmath46 and @xmath47 , equation shows that the only other term that can vary is @xmath3 , i.e. @xmath48 .
then , @xmath49 .
finally , let us use the equation to determine the relationship between @xmath50 and @xmath51 .
the dependence on the spreading factor @xmath40 is quite simple .
let us assume that a model for @xmath6 has been developed for a particular value of @xmath41 , i.e. @xmath42 . to determine @xmath43 for @xmath44 .
note that for a change in @xmath40 , the product @xmath45 constitutes a constant scaling factor with respect to @xmath18 .
therefore , for a link of distance @xmath39 the term @xmath27 will remain unchanged .
thus , if the same capacity @xmath38 is required for @xmath46 and @xmath47 , equation , shows that the only other term that can vary is @xmath3 , i.e. @xmath48 .
then , @xmath49 .
finally , let us use equation to determine the relation between @xmath50 and @xmath51 .
@xmath52 thus , any model for the transmission generated for some parameter @xmath40 has a simple extension . also , note that the transmission bandwidth remains the same for any value of @xmath40 .
and @xmath5 for @xmath1 and approximate model.@xmath12 $ ] , @xmath13 $ ] , @xmath9,@xmath10 and @xmath11 , width=336,height=336 ] and @xmath5 for @xmath2 and approximate model.@xmath12 $ ] , @xmath13$],@xmath9,@xmath10 and @xmath11 , width=336,height=336 ] a numerical evaluation procedure similar to that in @xcite is used to compute the value of @xmath6 , @xmath53 and @xmath1 , for a region of values of @xmath33 .
the procedure starts by fixing a target value of the capacity @xmath38 .
then , for each distance @xmath39 , the initial value of @xmath3 is set to the minimum value of the product @xmath54 , i.e. @xmath55 . the frequency at which this occurs , i.e. @xmath56 , is called the optimal frequency . after this ,
@xmath3 is increased iteratively by a small amount ( figure [ anplot.tag ] ) , until the target capacity value @xmath38 is met .
finally , this procedure is repeated for each value of @xmath38 in a range of interest . at the @xmath57-th step of the procedure ,
when @xmath58 is increased by a small amount , the band @xmath59 is determined for that iteration .
this band is defined as the range of frequencies for which the condition @xmath60 .
then , the capacity @xmath61 is numerically determined for the current @xmath58 and @xmath59 , using the equation ( 9 ) . if @xmath62 , a new iteration is performed . otherwise , the procedure stops .
by applying the above procedure for varying @xmath39 and @xmath38 , one arrived at the complete model for the power consumption @xmath63 where @xmath64 below , two ranges of operation were studied .
the first one is for @xmath65 km$ ] , @xmath66 kbps$ ] , and propagation factor of @xmath9 , which will be called case 1 from here on .
the second one is for @xmath67 kms$ ] , @xmath68 kbps$ ] , and propagation factor of @xmath9 , which will be called case 2 . for both regions and different ranges of @xmath23 and @xmath24 ,
the power consumption @xmath6 can be approximated by equations , and .
similar model are found to provide a good fit for the high / end frequency @xmath1 and for the bandwidth @xmath69 .
these models are given by @xmath70 where @xmath71 @xmath72
the transmission power , highest frequency and bandwidth of transmission band were computed for a variety of values of @xmath23 , @xmath24 and two ranges of interest of the pair @xmath33 , i.e. @xmath65 kms$ ] , @xmath66 kbps$ ] , and @xmath67 kms$ ] , @xmath68 kbps$ ] .
the models proposed fitted these cases quite well .
results are presented for the case of @xmath74 , @xmath75 and @xmath10 , for both cases . also for case 1
, it will be seen that the @xmath76 and @xmath77 parameters show almost no dependence on the shipping activity factor @xmath23 , especially if the wind speed is @xmath78 .
thus , the approximate model for this case could be simplified to only consider @xmath24 as part of the model , instead of the pair @xmath79 .
as function of @xmath24 . @xmath7 $ ] , @xmath8$],@xmath9,@xmath10 , width=336,height=403 ] as function of @xmath24 .
@xmath7 $ ] , @xmath8$],@xmath9,@xmath10 , width=336,height=403 ] figures [ figplc_k_15_lowc.tag ] , [ fhighlc_k_15_lowc.tag ] and [ blc_k_15_lowc.tag ] show parameters @xmath80 and @xmath5 for @xmath6,@xmath1 , and @xmath2 , respectively .
this approximation was carried out for the first case with a propagation factor of @xmath9 , a shipping activity of @xmath10 and a wind speed of @xmath11 .
the values of @xmath76 s and @xmath77 s are shown in table [ table_a1_lowc ] and [ table_a2_lowc ] , for parameters @xmath80 and @xmath5 , respectively .
these tables also show the mean square error ( mse ) of the approximation with respect to the actual parameters . in figure
[ blc_k_15_lowc.tag ] , there is a considerable variation in the values of parameter @xmath81 .
however , note that the y - axis of the plot shows very little variation .
[ table_a1_lowc ] .@xmath80 approximation parameter values for @xmath6 , @xmath1 and @xmath2 , with @xmath7 $ ] , @xmath8$],@xmath9,@xmath10 and @xmath11 [ cols="^,^,^,^,^,^",options="header " , ]
this paper offers an insight into the dependence of the transmission power , bandwidth , and the band - edge frequency of an underwater acoustic link on the capacity and distance .
it provides closed - form approximate models for the time - invariant acoustic channel , taking into account a physical model of acoustic path loss and the ambient noise , assuming that the channel is gaussian .
these approximate models where shown to provide a good fit to the actual empirical values by numerical evaluation for different ranges of distance @xmath39 and capacity @xmath38 , as well as noise profiles corresponding to different shipping activity factor and wind speed .
the band - edge frequency @xmath1 and the bandwidth @xmath2 were also shown to be invariant to the spreading factor @xmath40 , while the power scales as @xmath82 . for a certain range of values ( l , c ) , the approximate model of @xmath6
was shown to be almost independent of the shipping activity factor @xmath23 while having a marked dependency on the wind speed @xmath24 .
this dependence , however , is quite smooth and could be approximated by a simple model , thus resulting in a complete model for the @xmath6 for a range of values ( l , c ) that is of interest to a typical underwater communciation system .
hence , these models can be used in network optimization problems to determine the optimal power consumption for some required data rate .
future work will focus on studying convexity properties of the @xmath6 model and using it in network optimization problems .
this work was supported in part by the nsf grants # 0520075 and onr muri grant # n00014 - 07 - 1 - 0738 , and darpa bae systems national security solution , inc .
subcontract # 060786 . 1 partan , j. , kurose , j. , levine , b. n. , `` a survey of practical issues in underwater networks '' , in proc .
wuwnet 06 , pp .
17 - 24 , los angeles , sept .
2006 stojanovic , m. , `` on the relationship between capacity and distance in an underwater acoustic communication channel '' , in proc .
wuwnet 06 , pp .
41 - 47 , los angeles , sept . 2006 | the underwater acoustic channel is characterized by a path loss that depends not only on the transmission distance , but also on the signal frequency . as a consequence ,
transmission bandwidth depends on the transmission distance , a feature that distinguishes an underwater acoustic system from a terrestrial radio system .
the exact relationship between power , transmission band , distance and capacity for the gaussian noise scenario is a complicated one .
this work provides a closed - form approximate model for 1 ) power consumption , 2 ) band - edge frequency and 3 ) bandwidth as functions of distance and capacity required for a data link .
this approximate model is obtained by numerical evaluation of analytical results which takes into account physical models of acoustic propagation loss and ambient noise .
the closed - form approximations may become useful tools in the design and analysis of underwater acoustic networks . |
10 y. yu , chinaxiv:201607.00065 , in submission .
y. yu , ann .
phys . * 323 * , 2367 ( 2008 ) , and references therein ; y. yu , mod .
* b * * 29 * , 1550068 - 1 ( 2015 ) .
f. bloch , phys .
rev . * a * * 7 * , 2187 ( 1973 ) .
a. g. leggett , rev .
phys . * 73 * , 307 ( 2001 ) .
see , e. g. , j. g. daunt and k. mendelssohn , nature , * 143 * , 719 ( 1939 ) .
p. kaptiza , nature * 141 * , 74 ( 1938 ) ; j. f. allen and a. d. misener , nature * 141 * , 75 ( 1938 ) .
y. yu , in preparation . here
a medium could also refer to vacuum , and to a specific geometric arrangement of a superflow . for experimental observations of persistent @xmath1he superflow ,
see , e. g. , j. d. reppy and d. depatie , phys .
* 12 * , 187 ( 1964 ) . | we review on a recently proposed quantum exception to the second law of thermodynamics .
we emphasize that @xmath0he superflows , like any other forms of flows , shall carry entropy or heat in a thermal environment .
following that , one can use a heterogeneous @xmath0he superflow loop to realize entropy - decreasing processes .
we also mention that the heat content of a superflow has an unusual dependence on flow velocity , which is an important factor contributing to the entropy - decreasing processes .
_ keywords _ : superflow , entropy , many - body spectra + the second law of thermodynamics ( slt ) prohibits the decrease of the entropy in an isolated system .
it rules out the desired possibility of utilizing the endless colossal thermal energy in the environment as the major energy source for the civilization . but is this slt truely universal ?
we know that first law of thermodynamics is exact universal , because it is a manifestation of conservation of energy and quantum mechanics obeys the conservation of energy as a built - in law . unlike the first law
, slt does nt corresponds to a built - in rule in quantum mechanics .
imagine a situation in which that numberless physical processes obey slt , but there is one quantum process which violates slt , then in this situation we shall accept that slt is not an exact law of nature and that there is an exception to slt .
it is quantum mechanics , rather than slt , which is the true governing law of nature . in @xcite , we show that some @xmath0he - superflow - involving quantum processes contradict slt , and that it is possible to convert the thermal energy in the environment into useful energy .
we shall review this quantum exception to slt .
the entropy - decreasing quantum processes rely on a fundamental property of @xmath0he superflows .
whether does a @xmath0he superflow carry entropy or thermal energy ?
the phenomenological two - fluid model of superfluid @xmath1he gives a no answer .
but this is at odds with quantum mechanics .
a @xmath0he superflow corresponds to a number ( @xmath2 ) of @xmath0he atoms flowing together . if this superflow carries zero entropy and heat , that is equivalently to say that the quantum state of the @xmath2 @xmath0he atoms in the superflow is the lowest energy state at given momentum ( or at given current value ) .
we know that the time evolution of the whole system ( the superflow plus its environment ) is governed by the many - body schrodinger equation . in this time evolution ,
the quantum state of the environment is a mixed quantum state , corresponding to a certain temperature .
the subsystem of the @xmath2 @xmath0he atoms ( superflow ) can be expected to be a mixed quantum state too .
it is hard to imagine how this subsystem could reach a zero entropy and zero thermal energy quantum state following the time evolution .
even if at a moment this superflow reaches this zero entropy quantum state dramatically , it shall evolve into a mixed state at the following moments .
in other words , the superflow shall be a thermal flow in a thermal environment , similar to any conventional flows . plane .
dots ( in gray and in black ) represent levels .
the horizontal and vertical coordinates of a dot correspond to the momentum and the energy of the level , respectively .
@xmath3 denotes the height of an energy barrier between two `` valleys '' .
dots in black are the occupied levels of the system ( at low temperature ) , corresponding to a superflow state with a momentum of @xmath4 . ]
can we understand that a @xmath0he superflow is both thermal and superfluidic ( i.e. , keeping its current from decay ) ?
a recently developed quantum theory of superfluidity @xcite provides an affirmative answer .
this microscopic theory shows that the many - body spectrum of a superfluid has a characteristic feature @xcite .
consider a superfluid composed of @xmath2 particles in a limited geometry , the hamiltonian of this many - body system can be written as @xmath5 where @xmath6 is the mass of a particle and @xmath7 represents the inter - particle interaction .
consider the eigenspectrum of the hamiltonian operator , labelled by @xmath8 , @xmath9 where @xmath10 is eigenlevel and @xmath11 is the eigenwavefunction .
we shall consider also the momentum carried by eigenwavefunction @xmath12 , @xmath13 where @xmath14 is the total momentum operator along a superflow direction .
@xmath15 is not required to be an eigenwavefunction to the operator @xmath14 .
if one schematically plots the energy levels of this system at the energy - momentum ( @xmath16 ) plane ( see fig .
[ fig : sflevels ] ) , one then sees that the low - lying levels form valley - like structures at the boundary region @xcite .
the relevant quantum states at low temperature corresponds to those levels within the bottoms of the valleys and one can ignore the levels at higher energies .
a superflow state ( at low temperature ) corresponds to some occupied levels at a valley at a non - zero momentum .
the system is not allowed to jump from the occupied levels at this valley to left - side valleys with less momenta , due to the energy barriers which separate the valleys .
the prohibition of inter - valley jumps ensures that a superflow keeps its momentum from decay . for a given superflow
, the occupation probability distribution at a relevant valley follows a boltzmann distribution , namely , the occupation probability of a level at the valley is proportional to @xmath17 , where @xmath18 is the energy of the level , @xmath19 is the boltzmann constant and @xmath20 is the temperature of the superflow .
microscopically this thermal distribution is caused by the quanta exchanges between the superflow and its surroundings , which in turn is caused by microscopic interactions between the atoms in the superflows and the particles ( atoms or molecules ) in the surroundings . once realizing that @xmath0he superflows are still thermal flows
, one can expect a phenomenon of @xmath0he superflows similar to the peltier effect of electric currents .
when a @xmath0he superflow passes from a medium to a different medium , it shall experience a temperature change .
the heat content ( specific enthalpy ) of a superflow in the first medium as a function of temperature differs from the heat content function in the second medium . for the conservation of energy , the temperature of the superflow
shall be adjusted so that heat content remains the same ( roughly ) when the superflow exits the first medium ( see fig .
[ fig : sfpeltier ] ) . in literature , the temperature change of @xmath0he superflow exiting a medium to a rather large empty space was reported @xcite shortly after the discovery of superfluidity @xcite
. one could note that the flow velocity of the superflow generally changes when it passes from the first medium to the second medium .
if the first medium has a porosity smaller than the second medium , then the flow velocity in the first medium is larger than the flow velocity at the second medium .
it can be shown that the heat content of a superflow in a medium depends on the flow velocity @xcite , which is quite unusual from a classical viewpoint .
qualitatively , the larger the flow velocity , the smaller the heat content .
such a dependence of a superflow s heat content on the flow velocity could make a large contribution to this `` peltier effect '' of the superflow .
he superflow changes when exiting a medium to a different medium . ] in @xcite , we consider a system referred as to a heterogeneous superflow loop ( hsl ) .
it is a circling @xmath21he superflow filling a torus - shaped vessel , where half of the vessel is packed with one kind of medium and the other half is packed with a different medium @xcite ( see fig .
[ fig : hsloop ] ) .
one can find an interesting stable temperature configuration along an isolated hsl .
the superflow in one medium has a temperature @xmath22 higher than the temperature of the superflow at the other medium ( denoted by @xmath23 ) .
with such a temperature difference , the heat content of superflow in one medium is ( approximately ) the same as the heat content of superflow in the other medium , so that net heat transfer from one medium to the other , caused by the superflow , becomes zero ( approximately ) . to show a dramatical counterexample to slt , consider to prepare two infinitely large heat baths ( besides a hsl ) , with one bath at a temperature @xmath24 and the other at a temperature @xmath25 . @xmath24 and @xmath25
is set to satisfy @xmath26 .
consider making a good thermal contact between the high temperature part of the hsl and the bath at @xmath24 , and making a good thermal contact between the low temperature part of hsl and the bath at @xmath25 .
then , the superflow absorbs some heat from the heat reservoir at @xmath27 , and this ( moving ) superflow transfers the heat to the high temperature part of the hsl , and eventually the heat is passed to the heat reservoir at @xmath28 ( see fig .
[ fig : sltce2 ] ) .
this is a process in which the entropy of the whole system decreases .
the entropy - decreasing process can run for a cosmologically long time , since that the decay of the superflow is prevented by some energy barriers @xcite . in @xcite
, we sketch out a prototype of a system for extracting thermal energy from the environment .
this potential energy approach is environmentally friendly to a great extent
. the energy - generating process can be purely physical , without consuming any materials or producing any unwanted materials .
it can be roughly estimated that a system with one kilogram of liquid helium generates ideally a power of several kilowatts or more , but some relevant cryogenic technology needs to be developed for achieving such an efficiency . to another heat reservoir at a higher temperature .
] to summarize , a quantum exception to slt can be constructed with @xmath0he superflows .
financial supports from chinese nsf ( grant no .
11474313 ) , from cas ( grant no .
xdb21030300 ) , from nkrdp of china ( grant no .
2016yfa0301503 ) , are acknowledged . |
in recent years , analogies have taken on a major role in physics and mathematics @xcite .
they provide the new ways of looking at problems that permit cross - fertilization of ideas among different branches of science .
analogy methods have practical worth from two point of views : first , they provide new light on perplexing theoretical questions @xcite .
second , some of them are interesting for experimental set - up @xcite .
+ in this contribution we use the method developed in @xcite , in a multilayer stack to show how we can set up an experimental model for special relativity and rindler space as a special case of general relativity .
+ in the present section we review some of the relevant results achieved in @xcite . as it is well known , @xmath0 group is locally isomorphic ( or a 2 to 1 homomorphic ) to lorentz group , @xmath1 .
in fact , the transfer matrix , used in the classical optics for description of the interaction of light and matter , can be considered as an element of @xmath2 ( see @xcite and references therein ) .
@xcite , authors gave a geometrical interpretation for the transfer matrix of an absorbing system .
in fact , they considered an isolated absorbing film embedded between two identical ambient @xmath3 and substrate @xmath4 media .
a monochromatic , linear polarized plane wave falls from the ambient , as well as a plane wave of the same frequency and polarization falls from the substrate .
if the amplitudes of the field are characterized by a column vector , @xmath5 where @xmath6 ( @xmath7 ) is denoted backward ( forward)-travelling plane wave , and the amplitudes of the ambient and substrate are related by @xmath8 where @xmath9 is the transfer matrix and is defined as , @xmath10 and @xmath11 and @xmath12 are the transmission and reflection coefficients , respectively . + by defining the coherency matrix @xcite as @xmath13 for both @xmath3 and @xmath4 , it is easy to show that , @xmath14 it is useful to identify the variables @xmath15 , @xmath16 as the coordinates of a @xmath17-dimensional minkowski space , in which @xmath18 ( the identity ) and @xmath19 s are the standard pauli matrices .
the conjugation ( [ 0.4 ] ) induces a transformation on the variables @xmath20 of the form @xmath21 where , @xmath22 it is easy to see that the transformation @xmath23 is a lorentz transformation @xcite which is generated by @xmath24 and @xmath25 .
thus it is a @xmath26 homomorphism map between @xmath27 and @xmath28 .
+ we use the model which is developed in @xcite , and consider a monochromatic , linearly polarized plane electromagnetic wave incident on a non - absorption multilayer thin film . in this model basis
coordinates are defined by the components of the electromagnetic wave , i.e. , @xmath29 and @xmath30 .
+ the organization of this paper is as follows . in section [ s2 ]
we review the transfer matrix method for description of the multilayer films . in section [ s3 ]
we introduce spinor map between @xmath31 and @xmath28 , and we draw analogy between an em field which passes through a multilayer film and transformation of coordinates under the lorentzian group . in addition
, two special cases , a multilayer , that is , quarter - wave films and a multilayer with very low thickness films , are studied in detail . in section [ s4 ] , by using the multilayer films , we suggest an experimental set - up for investigation of two kinematical properties , that is , time delay and length contraction as some of predictions of the special relativity . in section [ s5 ]
we introduce an analogy between the rindler space , as a curved space , and a suitable multilayer film .
finally , a summary and conclusion remarks are given in section [ s6 ] .
we consider an electromagnetic radiation falling onto a structure consisting of a thin film with uniform thickness @xmath32 and index of refraction @xmath33 .
this isolated ( homogeneous and isotropic ) film , is embedded between two media , @xmath3 and @xmath4 , with indexes of refraction , @xmath34 and @xmath35 , respectively .
we consider a monochromatic , linearly polarized plane falling wave which is injected into the thin film from the above with amplitude @xmath36 and @xmath37 .
the field after passing the end layer is characterized by @xmath38 and @xmath39 , as it is shown in figure ( [ fig1 ] ) .
we use transfer matrix method which is a method used in optics to analyse the propagation of electromagnetic through a stratified ( layered ) medium @xcite . + the transfer - matrix method is based on the fact that , according to maxwell s equations , there are simple continuity conditions for tangential component of electric field and normal component of magnetic field across boundaries . in addition , this method can be developed to the multilayer films . @xcite . for a stack with @xmath40 layers ,
the transfer - matrix can be defined as a product of matrices given by @xmath41 where the transfer - matrix of the @xmath42 layer , @xmath43 , is given by @xmath44 where @xmath45 and @xmath46 for @xmath47-polarization and @xmath48-polarization is characterized by @xmath49 and @xmath50 respectively @xcite .
it is obvious that the angle @xmath51 is related to the angle of incidence @xmath52 by snell s law , that is , @xmath53
in this section , we use the spinor map , from @xmath31 , onto the lorentz group , @xmath1 .
we define a coherence matrix @xcite as follows : @xmath54 for both @xmath3 and @xmath4 , so that @xmath55 and @xmath56 are related acording to the eq .
( [ 0.4 ] ) .
we explicitly define the coordinates @xmath15 , by using equation ( [ 0.5 ] ) , @xmath57 it is worth to note that these coordinates behave as a light - like interval : @xmath58 therefore , the conjugation relation ( [ 0.4 ] ) induces a transformation on the variables @xmath20 in the form of equation ( [ 2.5 ] ) .
+ however , we can decompose the transfer matrix of a thin film , according to equations ( [ 1.3 ] ) and ( [ 12.1 ] ) with @xmath59 , as @xmath60 thus , under the homomorphism ( [ 2.6 ] ) , @xmath61 generates a boost of velocity @xmath62 ( in natural units , where @xmath63 ) , while @xmath64 induces a pure spatial rotation . + to gain physical insights ,
we write explicitly the equation ( [ 2.6 ] ) for @xmath65 and @xmath66 . under the homomorphism ( [ 2.6 ] ) , @xmath67 generates the following matrix @xmath68 which is a rotation in the plane @xmath69 of angle twice the phase of the transmission coefficient .
+ furthermore , diagonal matrix @xmath61 , under the homomorphism ( [ 2.6 ] ) , generates a boost in the plane @xmath70 @xcite , @xmath71 where rapidity @xmath72 is obtained from the relations @xmath73 and velocity @xmath74 is defined by @xmath75 in figure ( [ fig3 ] ) , we have plotted the variation of @xmath74 with respect to the @xmath76 for different values of the index of refraction , @xmath33 , when the electromagnetic field have @xmath47- , and @xmath48-polarization .
it is seen that for every values of the index of refraction , @xmath77 has maximum or minimum in @xmath78 .
in addition , for a fixed @xmath79 , the @xmath74 is increased by increasing the refraction index , for both @xmath47- , and @xmath80-polarization .
however , for the large values of the refraction index , @xmath81 , for both polarization , the result is the same , as shown in figure ( [ fig3 ] ) .
finally , by using equations ( [ 2.7 ] ) , ( [ 2.8 ] ) and ( [ 2.9 ] ) we obtain @xmath82 as @xmath83 where @xmath84 . also , with some calculations , we can rewrite equation ( [ 2.10.1 ] ) , in the following form @xcite .
@xmath85 in addition , by using einstein rule for the composition of velocities @xcite , @xmath86 we can write @xmath82 as a rotation and a boost , @xmath87 where @xmath88,{\nonumber}\\ \beta_{3}^{\prime}&=&-\frac{\beta } { 1-\beta^{2}\cos(2\omega t ) } \big ( \cos(2\omega t)+\frac{\beta^{2}\sin^{2}(2\omega t ) } { 1+\sqrt{1-\beta^{2}}}\big).\end{aligned}\ ] ] in figures ( [ fig5 ] ) and ( [ fig6 ] ) , we have plotted the variations of @xmath89 , @xmath90 and @xmath91 with respect to @xmath32 for @xmath92 and @xmath93 , for @xmath94-polarization and @xmath48-polarization , respectively .
as it is seen , @xmath95 , @xmath96 and latexmath:[$ we now turn to the multi - layer stack with @xmath40 films . by employing the method which is used to obtain equation ( [ 2.100.2 ] ) , and by using equation ( [ 2.200 ] )
, we can write @xmath98 as @xmath99 where @xmath100 and @xmath101 with @xmath102 and @xmath103,\\ \beta_{3}^{(i)}&=&-\frac{\beta^{(i ) } } { 1-(\beta^{(i)})^{2}\cos(2\omega^{(i ) } t^{(i ) } ) } \big ( \cos(2\omega^{(i ) } t^{(i)})+\frac{(\beta^{(i)})^{2}\sin^{2}(2\omega^{(i ) } t^{(i)})}{1+\sqrt{1-(\beta^{(i)})^{2}}}\big).\end{aligned}\ ] ] also , by using einstein addition rule ( [ 3.2 ] ) , we can rewrite equation ( [ 3.3 ] ) as @xmath104 where @xmath105 and @xmath106 are given by applying the rule ( [ 3.2 ] ) consecutively .
+ in this section , we consider a multilayer with quarter - wave films @xcite , that is @xmath107 and assume that @xmath108 , so that @xmath109 and @xmath110 . as an example , for a stack consists of two layers , matrix @xmath111 , is given by , @xmath112 which under the homomorphism ( [ 2.6 ] ) , generates a boost as @xcite , @xmath113 where velocity @xmath114 is given by @xmath115 and rapidity @xmath116 is obtained from the relations @xmath117 in general , for a stack consists of @xmath118 quarter - wave layers , in which the thickness of @xmath119 layer is @xmath120 , after some simple calculation , we can achieve @xmath121 as , @xmath122 where velocity @xmath123 is calculated by the rule ( [ 3.2 ] ) . as an example for @xmath124
, we have @xmath125 in this section we study a multilayer , with @xmath40 layers , in such a way that @xmath126 s are near to the unit , @xmath127 .
thus , by using equation ( [ 2.11 ] ) and keeping only first - order term of @xmath126 , we obtain @xmath128 and therefore the lorentzian transformations are replaced by the galilean transformations .
thus , by these assumption , we can rewrite the equation ( [ 3.7 ] ) as the following relation : @xmath129 where @xmath130,{\nonumber}\\ \mathcal{b}_{3}^{(\mathcal{m})}&=&\sum_{i=1}^{m}\beta^{(i)}\cos[2(\omega^ { ( i)}+\omega^{(i ) } ) t^{(i)}].\end{aligned}\ ] ] in addition , when we assume @xmath32 is very thin , from equation ( [ 2.10.1 ] ) , by keeping only first - order terms in @xmath32 , the rotation @xmath131 appears as a pure boost , @xmath132 , which is followed by another pure boost in another direction , @xmath133 , so that the result is like a thomas precession .
that is , we can write @xcite , @xmath134 where @xmath135 and @xmath136 is the acceleration along the second axis , @xmath137 .
we can now propose some specific experiments bases on the properties discussed in the preceding sections .
the basic experimental set - up consists essentially of a monochromatic plane wave , linear polarized ( @xmath94 or @xmath48 ) with the specific values of @xmath138 and @xmath139 .
+ we use a stack consists of @xmath140 quarter - wave layers . according to the equation ( [ 20.9 ] ) , the coordinates @xmath141 and @xmath142 under the action of transformation matrix remain invariant .
therefore , it is possible to consider @xmath143 as a boost which belongs to @xmath144 : @xmath145 to show how the experimental set - up works , let us suppose that the electromagnetic fields , @xmath146 and @xmath147 , fall onto the multi - layer . from equation ( [ 2.4 ] ) we have @xmath148 and @xmath149 these equations clearly show that there are many possible experimental methods to choose the input fields in such a way that two different points in a reference frame may be appear in the other frame as either with the same temporal or same spatial coordinate .
+ we consider two experimental set - ups with different inputs @xmath150 and @xmath151 and corresponding to outputs @xmath152 and @xmath153 , where @xmath154 .
thus , it is easy to show that @xmath155 when @xmath156 and @xmath157 when @xmath158 by using equation ( [ 5.4 ] ) , the first set - up can be used to demonstrate the length contraction , @xmath159 in order to put forward the phenomenon of the time dilation , let us consider another experimental set - up in which @xmath160 , @xmath161 , @xmath162 and @xmath163 satisfy the equation ( [ 5.5 ] ) .
therefore , measurement of @xmath164 may be regarded as a time interval measurement .
therefore , we obtain @xmath165 in summary , we have proposed a series of specific experiments by using a multilayer , which contains two films and the thickness of their films are given by @xmath166 , that clearly demonstrate the most important predictions of the lorentz kinematics .
maxwell s equations in general coordinates are written as @xcite @xmath167 and @xmath168 e_{k , j}=\dfrac{\partial ( \sqrt{g}{\hspace{0.04 cm}}g^{ij}b_{j})}{\partial t},\hspace{.5 cm } \left [ ijk \right ] h_{k , j}=\frac{1}{c^{2}}\dfrac{\partial ( \sqrt{g}{\hspace{0.04 cm}}g^{ij}e_{j})}{\partial t}+\mu_{0}\sqrt{g}{\hspace{0.04 cm}}j^{i},\end{aligned}\ ] ] where permutation symbol , @xmath169 $ ] , is defined by @xmath170=\bigg\lbrace \begin{array}{ccc } + 1 , & \text{if}\ ijk \hspace{.1 cm } $ is an even permutation of $ 123,\\ -1 , & \text{if}\ ijk \hspace{.1cm}$ is an odd permutation of $ 123,\\ 0 , & $ otherwise$. \end{array}\end{aligned}\ ] ] and the einstein summation convention is used .
as it is well known in optics , the maxwell s equations in empty - curved space can be replaced by maxwell s equations in the flat space which is filled by dielectric media with the following parameters @xcite , @xmath171 therefore , we can see a correspondence between geometry and matter , as in the general relativity .
thus , in this case , the index of refraction is obtained by @xcite @xmath172 now , if we consider the metric , @xmath173 by using equations ( [ 2.24 ] ) and ( [ 2.267 ] ) , the index of refraction is given by @xmath174^{-1 } $ ] .
+ on the other hand , the principle of equivalence make the gravitational field connect to the acceleration .
therefore , to study gravitational effects , we can generalize concept of lorentz transformations so as to include observers who are moving with a constant acceleration with respect to an inertial frame .
this metric is called rindler metric and described as @xcite , @xmath175 it is easy to see that the metrics ( [ 2.22 ] ) and ( [ 2.25 ] ) are conformal . accordingly , by using the equation ( [ 2.267 ] ) , we obtain the refractive index of the analogue media with the rindler space as the following relation : @xmath176 thus , for analogue rindler space we assume a multilayer , for which the refractive index of the @xmath119 layer is characterized by , @xmath177 however , up to the first order approximation in @xmath178 and by using equation ( [ 4.5 ] ) , for @xmath94- and @xmath80-polarization , when @xmath179 , we can rewrite equation ( [ 4.1 ] ) as , @xmath180 where @xmath181 and @xmath182 also , we can consider a stack consists of the quarter - wavelength layers , i.e. , @xmath183 . therefore , by using the equation ( [ 2.26 ] ) , the thickness of @xmath119 layer is @xmath184 after some calculation , we can obtain thickness of the @xmath185 layer as , @xmath186 and the index of refraction of this layer is obtained by @xmath187 we can now construct an analogy between the rindler space and a multilayer with @xmath118 quarter - wavelength layers , in which the thickness of @xmath42 layer is given by equation ( [ 4.101 ] ) and its refraction index is given by equation ( [ 4.102 ] ) .
accordingly we obtain , @xmath188 in refs .
@xcite , electrodynamics in the rindler space had been studied by using a phenomenological approach .
accordingly , the component of the electric field , near the horizon , @xmath189 , reads as follow : @xmath190 where @xmath191 is the modified bessel function of order @xmath192 .
+ figure ( [ fig18 ] ) displays a comparison between the output values of the electric fields are given by equation ( [ multilyer ] ) and equation ( [ rin ] ) , i.e. , @xmath193 in terms of @xmath194 and for different values of @xmath195 . as it is seen , by increasing @xmath196 ( increasing the number of layers ) , @xmath197 approaches to zero .
in addition , @xmath197 is decreased more rapidly by decreasing the @xmath195 .
this paper is an example of analogue models of einstein relativity .
firstly , we have used a geometrical tool to analyze a multilayer film in a concise way that is closely related to the special relativity .
then , this method was made an analogy between multilayer stack and curved space in the general relativity .
in fact , in the first step , we have seen how the transfer matrix can be interpreted as multiplication of a boost , a rotation and another boost . + in addition , in this case , we have suggested an experimental set - up which tests the most important predictions of the lorentz kinematics , i.e. , time delay and length contraction .
however , by using this isomorphism , apart from a relativistic presentation of the topic , which has interest in its own , we suggested an experimental set - up which is , in theoretical view , so simple and understandable , and , in practical level , was so easy to accomplish , in contrast to the similar experimental set - up @xcite .
+ finally , we have employed analogue method in replacing maxwell s equations in empty - curved space to maxwell s equations in the flat space where the later is filled by a media .
then , we have applied this method to the rindler frame , which by using principle of equivalence , is locally equivalent to gravitational fields .
this means that this method makes an analogy between accelerated frame in general relativity and a suitable multilayer films , with specific values of the refractive indexes and thickness of the films .
the authors wish to thank the office of graduate studies of the university of isfahan and shahrekord university for their support .
100 snchez - soto , l. l. ; monzn , j. j. ; barriuso , a. g. ; cariena , j. f. phys .
rep . * 2012 * , 513 , 191 .
monzn , j. j. ; barriuso , a. g. ; snchez - soto , l. l. ; montesinos - amilibia , j. m. phys .
a * 2011 * , 84 , 023830 .
novello , m. ; visser , m. ; volovik , g. _ artificial blackholes _ ; world scientific , singapore , 2002 .
leonhardt , u. nature ( london ) * 2002 * , 415 , 406 .
leonhardt , u. ; philbin , t. g. _ geometry and light : the science of invisibility _ ; dover , mineola , 2010 .
yonte , t. ; monzn , j. j. ; snchez - soto , l. l. ; cariena , j. f. ; lpez - lacasta , c. j. opt .
a * 2002 * , 19 , 603 . born , m. ; wolf , e. _ principles of optics : electromagnetic theory of propagation , interference and diffraction of light _ ; oxford , pergamon press , 1964 .
naber , g. l. _ the geometry of minkowski spacetime _ ; springer , new york , 1992 .
bass , m. ; li , g. _ handbook of optics , volume iv _ ; mcgraw - hill , new york , 2010 .
misner , ch .
w. ; thorne , k. s. ; wheeler , j. a. _ gravitation _ ; w. h. freeman , new york , 1973 .
sattinger , d. h. ; weaver , o. l. _ lie groups and algebras with applications to physics , geometry , and mechanics _ ; springer - verlag , 1993 .
cornwell , j. f. _ group theory in physics , volume i _ ; academic press , 1989 .
ungar , a. a. comput .
appl . * 2007 * , 53 , 1228 .
fowles , g. r. _ introduction to modern optics _ ; dover publications , usa , 1986 .
odonnell , k. ; visser , m. eur . j. phys . * 2011 * , 32 , 1033 .
padmanabhan , t. _ gravitation : foundations and frontiers _ ; cambridge university press , cambridge , 2010 .
brevik , i. j. math . phys .
* 1987 * , 28 , 2241 .
hacyan , sh . j. opt . *
2011 * , 13 , 105710 .
monzn , j. j. ; snchez - soto , l. l. phys .
lett a * 1999 * , 262 , 18 . ,
for @xmath94 and @xmath80-polarization , with respect to @xmath76 for different values of the index of refraction , @xmath198 are plotted by red and blue line , @xmath199 by red and blue dashing line and @xmath200 by red and blue dot line , for @xmath47 and @xmath80-polarization , respectively . ] | in this paper , to obtain an analogy between the curved spaces and the linear optics , we expand the idea of ref.@xcite to the multilayer films . we investigate effects of thickness and index of refraction of the films on the lorentzian transformations .
in addition , by using the multilayer films , we suggest very simple experimental set - up which can serve as an analogue computer for testing special relativity . finally , we draw an analogy between the rindler space , as an example of the curved spaces , and a suitable multilayer film .
* keywords * : multi - layers optics , rindler space , spinor map , transfer matrix . |
recently , the besiii collaboration claimed that four charged charmonium - like resonances @xcite were observed in the invariance mass spectra of @xmath5 , @xmath6 , @xmath7 and @xmath8 , which were respectively referred as @xmath9 , @xmath10 , @xmath11 and @xmath12 . among them , the @xmath10 has been confirmed by the belle collaboration @xcite and the cleo - c experiment @xcite .
so far it is still early to tell whether @xmath9 and @xmath10 , as well @xmath11 and @xmath12 , are the same state or not .
moreover , in very recently , the @xmath13 , which was first observed by belle collaboration @xcite , is confirmed in lhcb experiment with about 13.9 @xmath14 of significance @xcite .
the charmonium - like resonances are of particular interest as they facilitate not only investigating the dynamics of interaction between light quarks and heavy quarks , but also the testing of the standard model itself . until now
, their inner structures have not been well determined by theory . in the literature possible interpretations
include molecular states @xcite , tetraquark states @xcite , and enhancement structures @xcite , etc . in refs.@xcite ,
@xmath13 was treated as a pure tetraquark state , since there is no open charm meson thresholds close to it .
it is tempting to think this state to be a radial excitation of the @xmath10 , which was analyzed in the framework of qcd sum rules @xcite . of the hidden strange charmonium - like states ,
the cdf collaboration observed a narrow structure @xmath15 near the @xmath16 threshold in the exclusive @xmath17 decay mode produced in the @xmath18 collision at @xmath19 @xcite .
a scalar molecular state constructed by @xmath1 with @xmath20 was proposed to interpret the @xmath15 @xcite soon after its observation , and the extracted mass coincided with the experimental data . utilizing the qcd sum rules ,
people have successfully estimated the mass of @xmath10 and @xmath12 with @xmath21 and @xmath22 currents respectively @xcite .
these successes intrigue enormous interests of theorists .
it is of great interest to note that these states indicate that a new class of hadrons has been observed , but without the strange flavor until now .
namely , two resonances are expected to exist near the thresholds of @xmath0 and @xmath1 with the quantum numbers @xmath2 . in this work ,
we calculate their mass spectra with qcd sum rules , discuss the implications of our results and suggest their possible decay channels in the summary .
we hope this work may be helpful to experiment in exploring the hidden strange charmonium - like states .
primary formulae are presented after the introduction . in sec.iii ,
the numerical results and related figures are shown .
the last section is a short summary .
this work considers @xmath23 and @xmath1 molecular states with @xmath2 via qcd sum rules @xcite .
the calculations of the qcd sum rules are based on the correlator constructed by two hadronic currents . for an axial vector state ,
the two - point correlation function is given by : @xmath24 where @xmath25 is a current with the quantum numbers @xmath2 : @xmath26 , \label{eq2 } \\
j_\mu^{d_s^ * \bar{d}_s^ * } \!\!\!\!\!&=&\!\!\!\!\ ! \frac{i}{\sqrt{2}}\big[(\bar{s}_a \gamma^\alpha c_a)(\bar{c}_b \sigma_{\alpha \mu } \gamma_5 s_b ) -(\bar{s}_a \sigma_{\alpha \mu } \gamma_5 c_a)(\bar{c}_b \gamma^\alpha s_b)\big ] , \end{aligned}\ ] ] for respectively @xmath0 and @xmath1 .
here @xmath27 , @xmath28 are color indices , and these currents keep closer relations respectively with @xmath10 and @xmath12 @xcite . as @xmath29 is not a conserved current
, the two - point correlation function has two independent lorentz structures @xcite : @xmath30 where the invariant functions @xmath31 and @xmath32 are respectively related to the spin-0 and spin-1 mesons . in order to study the @xmath33 molecular state ,
@xmath32 was utilized .
the fundamental assumption of the qcd sum rules is the principle of quark - hadron duality .
accordingly , on the one hand , the correlation function @xmath32 is obtained at the hadron level where the mass and coupling constant of the hadron are introduced .
it may be calculated at the quark - gluon level , in which the operator product expansion ( ope ) is employed . on the hadron side , after separating out the ground state contributions from the pole terms , the correlation function @xmath32 is obtained as a dispersion integral over a physical regime , @xmath34 in which @xmath35 is the mass of @xmath0 or @xmath1 molecular state with @xmath2 , and @xmath36 is the spectral density which contains contributions from the higher excited states and the continuum states , @xmath37 is the threshold of the higher excited states and continuum states , and the coupling constant @xmath38 is defined by @xmath39 , where @xmath40 is the lowset lying @xmath33 @xmath0 or @xmath1 molecular state . on the quark - gluon side ,
the correlation function @xmath32 can be expressed as a dispersion relation : @xmath41 in which @xmath42 / \pi$ ] and @xmath43 here @xmath44 , @xmath45 and @xmath46 denote those contributions of the correlation function which have no imaginary parts but have nontrivial values under the borel transform .
it should be mentioned that in principle the four - gluon operator , the @xmath47 , also belongs to the dimension - eight condensate , however , in practice we find it is only 1% of the mixed condensate @xmath48 in magnitude , and hence the four gluon condensate is neglected in the evaluation of this work . after making a borel transform of the quark - gluon side : @xmath49 to consider the effects induced by the mass of the strange quark , terms which are linear in the strange quark mass @xmath50 are utilized in the following calculations .
for both the @xmath0 and @xmath1 molecular states , we put the concrete forms of spectral densities in eq.([spectral - density ] ) into the appendix . performing the borel
transform on the hadron side ( eq.([hadron ] ) ) and matching it to eq.([spectral - density ] ) , the resultant sum rule for the mass of the hidden strange molecular state h with @xmath33 is : @xmath51 where h represents the @xmath0 or @xmath1 molecular state and @xmath52
in the numerical calculation , the values of the condensates and the quark masses are used as @xcite : @xmath53 , @xmath54 , @xmath55 , @xmath56 , @xmath57 , @xmath58 , @xmath59 , @xmath60 , and @xmath61 . here
, the strange quark mass is the current quark mass in a mass - independent substraction scheme such as @xmath62 at the scale @xmath63 , whereas the charm and bottom quark masses are the running masses in the @xmath62 scheme . to provide greater clarity of choosing these parameters , the primary sources for determining these parameters can be found in refs .
@xcite . to determinate the borel parameter @xmath64 and the threshold parameter @xmath37 , we used the following limit constraints , which is the standard procedure for qcd sum rules analysis .
the sum rule parameter @xmath65 is utilized in our analysis . in the qcd sum rules , for choosing the threshold @xmath37 and the parameter @xmath66
, there are two criteria @xcite .
first , the convergence of the ope is retained .
therefore , in order to determine their convergence , one needs to compare the relative contributions of each term to the total contributions of the ope side .
the second criterion to constrain the @xmath66 is that the pole contribution ( pc ) , defined as the pole contribution divided by the total contribution ( pole plus continuum ) , is larger than the continuum contribution . in order to safely eliminate the contributions of the higher excited and continuum states ,
the pc is generally greater than @xmath67 @xcite . to find a proper value for @xmath68
, we carry out a similar analysis as in ref . @xcite .
since the continuum threshold is connected to the mass of the studied state by the relation @xmath69 , where @xmath70 is about 0.6 gev , various @xmath68 satisfying this constraint are taken into account . among these values
, one needs then to find out the proper one which has an optimal window for borel parameter @xmath64 .
that is , within this window , the physical quantity , here the molecular mass @xmath71 , is independent of the borel parameter @xmath64 as much as possible . through the above procedure one
obtains the central value of @xmath68 .
however , in practice , in the qcd sum rules calculation , it is normally acceptable to vary the @xmath68 by @xmath72 @xcite , which gives the lower and upper bounds and hence the uncertainties of @xmath68 .
the ope convergences are illustrated in fig.[convergence ] respectively for @xmath0 and @xmath1 .
complying with the first criterion , the upper limit constraints of @xmath66 are @xmath73 and @xmath74 with @xmath75 and @xmath76 . for the @xmath0 molecular state with @xmath77 ,
where @xmath65 .
( * b * ) ope convergence in the region @xmath78 for the @xmath1 molecular state with @xmath79 .
the black line denotes the fraction of perturbative contribution , and each subsequent line denotes the addition of one extra condensate , _
i.e. _ , @xmath80 ( red line ) , @xmath81 ( blue line ) , @xmath82 + @xmath83 ( red dotted line ) , @xmath84 ( blue dotted line ) .
the vertical lines respectively denote the upper limit constraints of the @xmath66 for @xmath0 and @xmath1 .
the curve including the dimension - eight contribution is defined as : ( @xmath85)/total .
hence the curve including dimension eight operator is just a straight line , and we do not show it here.,title="fig:",width=321 ] for the @xmath0 molecular state with @xmath77 , where @xmath65 .
( * b * ) ope convergence in the region @xmath78 for the @xmath1 molecular state with @xmath79 .
the black line denotes the fraction of perturbative contribution , and each subsequent line denotes the addition of one extra condensate , _
i.e. _ , @xmath80 ( red line ) , @xmath81 ( blue line ) , @xmath82 + @xmath83 ( red dotted line ) , @xmath84 ( blue dotted line ) .
the vertical lines respectively denote the upper limit constraints of the @xmath66 for @xmath0 and @xmath1 .
the curve including the dimension - eight contribution is defined as : ( @xmath85)/total .
hence the curve including dimension eight operator is just a straight line , and we do not show it here.,title="fig:",width=321 ] the result of the pc is shown in fig.[pole ] , which indicates the lower limit constraint of @xmath66 .
noting that the lower limit constraint of @xmath66 depends on the threshold value @xmath37 , for different @xmath37 , there are different lower limits of @xmath66 . to determine the value of @xmath37 ,
an analysis similar to ref.@xcite was carried out .
state with @xmath77 .
( * b * ) the relative pole contribution of @xmath1 state with @xmath79 .
the black line represents the relative contribution , whereas the red dashed line corresponds to the continuum contribution .
the vertical lines respectively denote the lower limit constraints of the @xmath66 for @xmath0 and @xmath1.,title="fig:",width=321 ] state with @xmath77 .
( * b * ) the relative pole contribution of @xmath1 state with @xmath79 .
the black line represents the relative contribution , whereas the red dashed line corresponds to the continuum contribution .
the vertical lines respectively denote the lower limit constraints of the @xmath66 for @xmath0 and @xmath1.,title="fig:",width=321 ] the dependence of @xmath71 on the parameter @xmath66 in fig.[mass ] with various @xmath68 is drawn where the two vertical lines indicate the upper and lower limits of a valid borel window for the central value of @xmath68 . for complete analysis ,
the borel windows with various @xmath68 are listed in table.[window ] .
state as a function of the sum rule parameter @xmath66 , for different values of @xmath68 .
( * b * ) the mass of the @xmath1 state as a function of the sum rule paramter @xmath66 , for different values of @xmath68 .
the two vertical lines indicate the upper and lower limits of a valid borel window.,title="fig:",width=321 ] state as a function of the sum rule parameter @xmath66 , for different values of @xmath68 .
( * b * ) the mass of the @xmath1 state as a function of the sum rule paramter @xmath66 , for different values of @xmath68 .
the two vertical lines indicate the upper and lower limits of a valid borel window.,title="fig:",width=321 ] [ cols="^,^,^,^,^,^ " , ] [ comparison ] ref .
@xcite systematically evaluated the masses of the tetraquark systems with various quantum numbers and hidden - charm .
the results for the hidden - charm and strange tetra- quark state with @xmath86 are in the region of 3.92 - 4.34 gev . comparing to our results , the masses of the molecular states in our calculation lie in 3.83 - 4.54 gev , which covers the tetraquark mass region .
since considerable errors exist , so far we think it is hard to distinguish the underlying quark configurations of the tetraquark states and molecular states .
this work used the qcd sum rules to study the masses of the @xmath0 and @xmath1 molecular states with quantum numbers @xmath2 . in our calculation , interpolating currents with definite charge parity were employed , and the contributions up to dimension eight in the operator product expansion ( ope ) were taken into account .
the numerical results were respectively @xmath87 and @xmath88 for the @xmath0 and @xmath1 molecular states .
the central value of @xmath0 molecular state was below the corresponding meson - meson threshold @xmath89 \simeq 4.08 \ , \text{gev}$ ] about 100 mev , which means that such molecular state would be tightly bound .
the reason for the relative large binding energy is that the current in eq.([eq2 ] ) is local .
hence they do not represent a object with two mesons separated in space , but rather compact one with two singlet quark - antiquark pairs .
the central value of @xmath1 molecular state was above that of its corresponding meson - meson threshold @xmath90 \simeq 4.22 \ , \text{gev}$ ] .
therefore , the @xmath1 molecular state with @xmath33 may form a resonance . in conclusion ,
considering the uncertainties , our results indicated that two hidden strange charmonium - like states , which may be probed in future experiments , may exist in the energy ranges of @xmath91 and @xmath92 . to ascertain the hidden strange charmonium - like states through their decays
, the following decay modes may be measured : @xmath93 , @xmath94 and @xmath95 for @xmath0 molecular state , and @xmath96 , @xmath97 , @xmath98 , @xmath99 , @xmath100 and @xmath101 for @xmath1 molecular state .
we suggest future experiments to search for these hidden strange charmonium - like resonances . the besiii , belle and babar and forthcoming belleii and superb will facilitate searching for such hidden strange charmonium - like resonances . for the @xmath0 molecular state , the concrete forms of spectral densities in eq.([spectral - density ] ) are obtained as : @xmath102 \ ; , \\
\rho^{\langle \bar{s } s \rangle}_{d_s \bar{d}_s^*}(s ) \!\!\ ! & = & \!\!\ ! \frac{3 \langle \bar{s } s \rangle}{2 ^ 8 \pi^4 } \int_{\alpha_{min}}^{\alpha_{max } } d \alpha \bigg{\ { } \frac{2 m_s { \cal h}^2}{(1 - \alpha ) \alpha } - \int_{\beta_{min}}^{1 - \alpha } \frac{d \beta}{\alpha \beta } { \cal f } \bigg [ [ m_c ( 1 + \alpha + \beta ) \frac{(\alpha + \beta ) } { \alpha \beta } \nonumber \\ \!\!\ ! & + & \!\!\ ! 2 m_s \alpha \beta ] { \cal f } + 8 m_s m_c^2 \bigg ] \bigg{\}}\;,\end{aligned}\ ] ] @xmath103 { \cal f } \nonumber \\ \!\!\ ! & + & \!\!\ ! m_s m_c^3 ( \alpha + \beta)^2 ( 1 - \alpha - \beta)(3 + \alpha + \beta ) ( \alpha^2 - \alpha \beta + \beta^2)\bigg ] \ ; , \\ \rho^{\langle \bar{s } g s \rangle}_{d_s \bar{d}_s^*}(s ) \!\!\ ! & = & \!\!\ ! \frac{\langle g_s \bar{s } \sigma \cdot g s \rangle}{2 ^ 7 \pi^4 } \int_{\alpha_{min}}^{\alpha_{max } } d \alpha \bigg{\ { } 2 m_s ( { \cal h } + m_c^2 ) - \frac{1}{\alpha}(3 m_c + m_s ) { \cal h } \nonumber \\ \!\!\ ! & + & \!\!\ !
\int_{\beta_{min}}^{1 - \alpha } d \beta \bigg[\frac{3 m_c}{2 } \frac{{\cal f}}{\alpha } + \frac{1}{\alpha^2}\big[(2 m_s \alpha + 3 m_c ( \alpha + \beta)){\cal f } - 3 m_s m_c^2 \alpha \big]\bigg]\bigg{\ } } \ ; , \\
g^3 \rangle}_{d_s \bar{d}_s^*}(s ) \!\!\ ! & = & \!\!\ ! \frac{\langle g_s^3 g^3 \rangle}{2^{14 } \pi^6 } \int_{\alpha_{min}}^{\alpha_{max } } \frac{d \alpha}{\alpha^3 } \int_{\beta_{\min}}^{1 - \alpha } d \beta \bigg[(1 - ( \alpha + \beta)^2 ) ( 2 { \cal f } + 4m_c^2 \beta ) \nonumber \\ \!\!\ ! & - & \!\!\ !
m_s m_c ( 1 - \alpha -\beta)(3 + \alpha + \beta)(\alpha + 6\beta)\bigg]\ ; , \\ \rho^{\langle \bar{s } s \rangle^2}_{d_s \bar{d}_s^*}(s ) \!\!\ ! & = & \!\!\ !
\frac{\langle \bar{s } s \rangle^2}{2 ^ 8 \pi^2 } \bigg[4 m_c^2 - 3 m_s m_c \bigg ] \sqrt{1 - \frac{4 m_c^2}{s } } \ ; , \\ \pi^{\langle g_s^3 g^3\rangle}_{d_s \bar{d}_s^ * } ( m_b^2 ) \!\!\ ! & = & \!\!\ !
\frac{m_s m_c^3 \langle g_s^3 g^3 \rangle}{2^{13 } \pi^6 } \int_{0}^{1 } \frac{d \alpha}{\alpha^4 } \int_{0}^{1 - \alpha } d \beta ( \alpha + \beta)(1 - \alpha - \beta ) \nonumber \\ \!\!\ ! & \times & \!\!\ ! ( 3 + \alpha + \beta ) \text{exp } \bigg [ - \frac{m_c^2 ( \alpha + \beta)}{m_b^2 \alpha \beta}\bigg ] \ ; , \\
\pi^{\langle \bar{s } s \rangle^2}_{d_s \bar{d}_s^ * } ( m_b^2 ) \!\!\ ! & = & \!\!\ ! - \frac{m_s m_c^3 \langle \bar{s } s \rangle^2}{2 ^ 4 \pi^2 } \int_{0}^{1 } \frac{d \alpha}{\alpha } \text{exp } \bigg [ - \frac{m_c^2}{m_b^2 ( 1 - \alpha ) \alpha}\bigg ] \ ; , \end{aligned}\ ] ] and the contribution from the dimension - eight condensate is obtained as : @xmath104 \bigg ] \text{exp}\bigg [ - \frac{m_c^2}{m_b^2 ( 1 - \alpha ) \alpha}\bigg ] \ ; .\end{aligned}\ ] ] here , @xmath105 is the borel parameter introduced by the borel transformation , @xmath106 , @xmath107 and the integration limits are given by @xmath108 , @xmath109 and @xmath110 . for the @xmath1 molecular state , the concrete forms of spectral densities in eq.([spectral - density ] )
are obtained as : @xmath111 { \cal f}^3 \ ; , \\
\rho^{\langle \bar{s } s \rangle}_{d_s^ * \bar{d}_s^*}(s ) \!\!\ ! & = & \!\!\ ! -\frac{3 \langle \bar{s } s \rangle}{2 ^ 8 \pi^4 } \int_{\alpha_{min}}^{\alpha_{max } } d \alpha \bigg{\ { } \frac{2m_s { \cal h}^2 } { ( 1 - \alpha)\alpha } + \int_{\beta_{min}}^{1 - \alpha } d \beta \bigg [ \big[\frac{3m_c ( \alpha + \beta)(1 + \alpha + \beta)}{\alpha^2 \beta^2 } + \frac{2 m_s}{\alpha \beta}\big ] { \cal f } \nonumber \\ \!\!\ ! & - & \!\!\
! \frac{4m_s m_c^2 ( 5 - \alpha - \beta ) } { \alpha \beta}\bigg ] { \cal f } \bigg{\ } } \ ; , \\ \rho^{\langle g_s^2
g^2 \rangle}_{d_s^ * \bar{d}_s^*}(s ) \!\!\ ! & = & \!\!\ ! \frac{\langle g_s^2 g^2 \rangle}{2^{13 } \pi^6 } \int_{\alpha_{min}}^{\alpha_{max } } \frac{d \alpha}{\alpha^2 } \int_{\beta_{min}}^{1 - \alpha } \frac{d \beta}{\beta^2 } \bigg [ 2(\alpha + \beta)(1 - 2\alpha - 2\beta ) { \cal f}^2 \nonumber \\ \!\!\ !
& + & \!\!\ !
\frac{2m_c^2 ( 1 - \alpha - \beta)}{\alpha \beta}[4 ( \alpha^4 + \beta^4 ) + 6\alpha \beta ( \alpha^2 + \beta^2 ) + 4\alpha^2 \beta^2 + 7(\alpha^3 + \beta^3 ) \nonumber \\ \!\!\ ! & + & \!\!\ ! 7\alpha \beta(\alpha + \beta ) - 5(\alpha^2 + \beta^2)]{\cal
f } + \frac{m_s m_c}{\alpha \beta}[9(\alpha^5 + \beta^5 ) + 22\alpha \beta(\alpha^3 + \beta^3 ) \nonumber \\ \!\!\ ! & + & \!\!\ ! 18 ( \alpha^4 + \beta^4 ) + 25\alpha^2 \beta^2 ( \alpha + \beta ) + 22\alpha \beta(\alpha^2 + \beta^2 ) - 27 ( \alpha^3 + \beta^3 ) \nonumber \\ \!\!\ ! & + & \!\!\ ! 12\alpha^2 \beta^2]{\cal f } - \frac{2m_c^4 } { 3\alpha \beta}(\alpha^3 + \beta^3)(1 - \alpha - \beta)^2 ( 5 + \alpha + \beta ) \nonumber \\ \!\!\ ! & - & \!\!\ ! \frac{3m_s m_c^3}{\alpha \beta}(1 - \alpha - \beta)(\alpha + \beta)(\alpha^3 + \beta^3)(3 + \alpha + \beta ) \bigg ] \ ; , \\
\rho^{\langle \bar{s } g s \rangle}_{d_s^ * \bar{d}_s^ * } ( s ) \!\!\ !
& = & \!\!\ ! \frac{\langle g_s \bar{s } \sigma \cdot g s \rangle}{2 ^ 8 \pi^4 } \int_{\alpha_{min}}^{\alpha_{max } } d
\alpha \bigg{\ { } 20 m_s m_c^2 - \frac{m_s { \cal h}}{\alpha(1 - \alpha ) } [ 1 + 18 \alpha ( 1 - \alpha ) ] - \frac{9 m_c { \cal h}}{(1 - \alpha)\alpha } \nonumber \\ \!\!\ ! & + & \!\!\ !
\int_{\beta_{min}}^{1 - \alpha } d \beta \bigg[\frac{m_c}{2\alpha^2 \beta^2 } [ 2(\alpha^3 + \beta^3 ) + 7\alpha \beta ( \alpha
+ \beta ) - 4 ( \alpha^2 + \beta^2 ) - 4\alpha \beta ] { \cal f } \nonumber \\ \!\!\ !
& - & \!\!\ !
\frac{m_s ( \alpha + \beta)}{\alpha \beta } { \cal f } - \frac{m_s m_c^2}{\alpha \beta}[(\alpha^2 + \beta^2)-2(\alpha + \beta ) ] \bigg]\bigg{\ } } \ ; , \\ \rho^{\langle g_s^3
g^3 \rangle}_{d_s^ * \bar{d}_s^*}(s ) \!\!\ ! & = & \!\!\ ! \frac{\langle g_s^3 g^3 \rangle } { 2^{13 } \pi^6 } \int_{\alpha_{min}}^{\alpha_{max } } \frac{d \alpha}{\alpha^3 } \int_{\beta_{min}}^{1 - \alpha } \frac{d \beta}{\beta^3 } \bigg[\frac{3{\cal f}}{2}[1 - ( \alpha + \beta)^2](\alpha^3 + \beta^3 ) + m_c^2(1 - \alpha - \beta ) \nonumber \\ \!\!\ ! & \times & \!\!\ ! [ 4(\alpha^5 + \beta^5 ) + 5\alpha \beta ( \alpha^3 + \beta^3 ) + 7(\alpha^4 + \beta^4 ) + \alpha^2 \beta^2 ( \alpha + \beta ) + 4\alpha \beta ( \alpha^2 + \beta^2 ) \nonumber \\ \!\!\ ! & - & \!\!\ !
5(\alpha^3 + \beta^3 ) ] - \frac{3m_s m_c}{4}(1 - \alpha - \beta)[6(\alpha^5 + \beta^5 ) + 7\alpha \beta ( \alpha^3 + \beta^3 ) \nonumber \\ \!\!\ ! & + & \!\!\ ! 18(\alpha^4 + \beta^4 ) + \alpha^2 \beta^2 ( \alpha + \beta ) + 3\alpha \beta ( \alpha^2 + \beta^2 ) ] \bigg ] \ ; , \\ \rho^{\langle \bar{s } s \rangle^2}_{d_s^ * \bar{d}_s^*}(s ) \!\!\ ! &
= & \!\!\ ! -\frac{\langle \bar{s } s\rangle^2}{2 ^ 6 \pi^2 } \int_{\alpha_{min}}^{\alpha_{max } } d \alpha \big[20({\cal h } - m_c^2 ) + 9m_s m_c \big ] \ ; , \end{aligned}\ ] ] @xmath112 + \frac{9m_s m_c^3}{2 } ( 1 - \alpha - \beta ) ( \alpha + \beta ) ( 3 + \alpha + \beta ) ( \alpha^4 + \beta^4 ) \bigg ] \nonumber \\ \!\!\ ! & \times & \!\!\ ! \text{exp}\bigg [ - \frac{m_c^2(\alpha + \beta)}{m_b^2 \alpha \beta } \bigg ] \ ; , \\ \pi^{\langle \bar{s } s \rangle^2}_{d_s^ * \bar{d}_s^*}(m_b^2 ) \!\!\ ! &
= & \!\!\ ! -\frac{3m_s m_c^3 \langle \bar{s } s \rangle^2 } { 2 ^ 5 \pi^2}\int_0 ^ 1 d \alpha \frac{1}{(1 - \alpha ) \alpha } \text{exp } \bigg[- \frac{m_c^2}{(1 - \alpha ) \alpha m_b^2 } \bigg ] \ ; , \\ \pi^{\langle \bar{s } s \rangle \langle \bar{s } g s \rangle}_{d_s^ * \bar{d}_s^*}(m_b^2 ) \!\!\ ! & = & \!\!\ !
\frac{\langle \bar{s } s \rangle \langle g_s \bar{s } \sigma \cdot g s \rangle}{3 \times 2 ^ 8 \pi^2 } \int_0 ^ 1 d \alpha \bigg[- \frac{20m_c^2}{(1 - \alpha ) \alpha}[1 + 12 \alpha ( 1 - \alpha ) ] \nonumber \\ \!\!\ ! & + & \!\!\ ! \frac{m_s m_c}{(1 - \alpha ) \alpha}[8 + 45\alpha(1 - \alpha ) ] - \frac{120m_c^4}{(1 - \alpha)\alpha m_b^2 } + \frac{m_s m_c^3}{(1 - \alpha)^2 \alpha^2 m_b^2 } [ 10 \nonumber \\ \!\!\ !
& + & \!\!\ 57 \alpha ( 1 - \alpha ) ] + \frac{30m_s m_c^5}{(1 - \alpha)^2 \alpha^2 m_b^4 } \bigg ] \text{exp } \bigg [ - \frac{m_c^2}{(1 - \alpha ) \alpha m_b^2 } \bigg ] .\end{aligned}\ ] ] m. ablikim _ et al . _
[ besiii collaboration ] , phys .
lett . * 112 * , 022001 ( 2014 ) [ arxiv:1310.1163 [ hep - ex ] ] ; m. ablikim _ et al . _
[ besiii collaboration ] , phys .
lett . * 110 * , 252001 ( 2013 ) [ arxiv:1303.5949 [ hep - ex ] ] ; m. ablikim _ et al . _
[ besiii collaboration ] , phys .
lett . * 111 * , 242001 ( 2013 ) [ arxiv:1309.1896 [ hep - ex ] ] ; m. ablikim _ et al . _
[ besiii collaboration ] , phys .
lett . *
112 * , 132001 ( 2014 ) [ arxiv:1308.2760 [ hep - ex ] ] . | this work uses the qcd sum rules to study the masses of the @xmath0 and @xmath1 molecular states with quantum numbers @xmath2 .
interpolating currents with definite c - parity are employed , and the contributions up to dimension eight in the operator product expansion ( ope ) are taken into account .
the results indicate that two hidden strange charmonium - like states may exist in the energy ranges of @xmath3 gev and @xmath4 gev , respectively .
the hidden strange charmonium - like states predicted in this work may be accessible in future experiments , e.g. besiii , belleii and superb .
possible decay modes , which may be useful in further research , are predicted .
+ + pacs numbers : 11.55.hx , 12.38.lg , 12.39.mk + keywords : qcd sum rules , molecular state , hidden strange |
half a century ago johnson and teller @xcite suggested the need for a scalar - isoscalar meson with a mass @xmath3 to provide saturation and binding in nuclei . in a way this was the starting point for one - boson - exchange ( obe ) potentials
where , in addition to the pion , all possible resonances would be included @xcite . despite their undeniable success describing nn scattering data ,
there has always been some arbitrariness on the scalar meson mass and coupling constant to the nucleon , partly stimulated by a lack of other sources of information , definitely helping the fits .
the relation of the ubiquitous scalar meson in nuclear physics and nn forces in terms of correlated two pion exchange has been pointed out many times @xcite ( see e.g. @xcite for a discussion in a chiral context ) . the quest for the existence of the @xmath4 resonance ( commonly denoted by @xmath5 ) has finally culminated with its inclusion in the pdg @xcite as the @xmath6 seen as a @xmath2 resonance , where a spread of values ranging from @xmath7 for the mass and a @xmath8 for the width are displayed @xcite .
the uncertainties have recently been sharpened by a determination based on roy equations and chiral symmetry @xcite yielding the value @xmath9 ; the lowest resonance in the hadronic spectrum .
it is mandatory and perhaps possible to scrutinize its role in hadronic phenomenology all over . here ,
we approach the problem from nn scattering in the @xmath1 channel from a renormalization viewpoint as applied to the obe potential ( without explicit inclusion of @xmath10 exchange ) and try to see the connection to @xmath2 scattering .
the field theoretical obe model of the nn interaction @xcite includes all mesons with masses below the nucleon mass , i.e. , @xmath11 , @xmath12 , @xmath13 and @xmath14 , in addition with a scalar - isoscalar boson . dropping @xmath12 and @xmath15 because of their small couplings , the @xmath16 nn potential is @xmath17 as is well known @xcite any perturbative determination of a potential suffers from off - shell ambiguities ( even in the born approximation ) , particularly because of relativistic finite mass corrections which may be smoothly shifted between entirely energy dependent and local potentials or energy independent and nonlocal potentials .
this trading between retardation and nonlocality may become sizable at short distances scales , where the interaction is unknown anyhow , and the particular choice is completely arbitrary
. our renormalization scheme will be such that , as suggested by partovi and lomon we ignore both retardation and nonlocality in the long distance limit @xcite , as well as the exponentially @xmath18 suppressed @xmath19 cut .
relativistic effects are only kept by renormalization of the couplings , but the effect is small and @xmath20 . in any case , one should bear in mind that nn scattering in the elastic region below pion production threshold involves cm momenta @xmath21 mev .
given the fact that @xmath22 we expect heavier mesons to be irrelevant , and @xmath23 itself to be marginally important , even in s - waves , which are most sensitive to short distances . in order to illustrate this , we take @xmath24mev , @xmath25mev , @xmath26mev and @xmath27 , which seem firmly established , and treat @xmath28 , @xmath29 and @xmath30 , as fitting parameters . as we show now
, this vector meson irrelevance has not been fulfilled in the conventional approach to nn scattering , forcing too large @xmath31 couplings .
actually , in the standard approach the scattering phase - shift @xmath32 is computed by solving the ( s - wave ) schrdinger equation * r*-space @xmath33 with a regular boundary condition at the origin @xmath34 .
moreover , for a short range potential such as the one in eq .
one also has the effective range expansion ( ere ) @xmath35 where the _ scattering length _ @xmath36 and the _ effective range _
@xmath37 are defined by the asymptotic behavior of the zero energy wave function . in the usual approach
@xcite everything is obtained from the potential assumed to be valid for @xmath38 .
in addition , due to the _ unnaturally large _
nn @xmath1 scattering length ( @xmath39 ) , any change in the potential @xmath40 has a dramatic effect on @xmath36 , since one obtains @xmath41 and thus the potential parameters _ must be fine tuned _ , and in particular the short distance physics .
a fit to the np data of ref .
@xcite yields two possible but incompatible scenarios : @xmath42mev , @xmath43 , @xmath44 with @xmath45 and @xmath46mev , @xmath47 , @xmath48 with @xmath49 .
the small errors should be noted .
the ambiguity in this solution is a typical inverse scattering one ; note that despite the @xmath23 being repulsive , the total potential is not repulsive at short distances , and the corresponding couplings and scalar mass are determined to high accuracy but incompatible .
this is just opposite to our expectations and we may regard these fits , despite their success in describing the data , as unnatural .
to overcome the unphysical short distance sensitivity we implement the renormalization viewpoint ( see e.g. ref .
@xcite ) . in the simplest version one proceeds as follows * for a given @xmath36 integrate in the zero energy wave function @xmath50 down to the cut - off radius @xmath51 .
this is the renormalization condition .
@xmath52 * impose self - adjointness to get the finite energy wave function @xmath53 , @xmath54 * integrate out the finite energy wave function @xmath55 , eq .
, to determine the phase shift @xmath32 from eq . .
* remove the cut - off @xmath56 to ensure _ model independence_.
this procedure allows us to compute @xmath32 ( and hence the next order s parameters @xmath37 , @xmath57 ) from @xmath58 and @xmath36 as independent information .
note that this is equivalent to consider , in addition to the regular solution , the irregular one .
a fit of the potential eq . to the np data of ref .
@xcite using the renormalization method gets @xmath59mev , @xmath60 , @xmath61 with @xmath62 .
note that @xmath31 is , not only small but mostly irrelevant , so we consider this fit natural .
a consequence of this is that we could take the su(3 ) value @xmath63 which on the basis of the ozi rule , @xmath64 , sakurai s universality @xmath65 and the ksfr relation @xmath66 yields @xmath67 for which we get @xmath68 , @xmath69 and @xmath70 with a strong linear correlation ( see fig . [
fig : fits ] ) .
confidence level ellipse in the @xmath71 plane for @xmath72 .
right : renormalized obe @xmath1 pn phase shifts ( in degrees ) as a function of cm momentum .
data from @xcite.,title="fig : " ] confidence level ellipse in the @xmath71 plane for @xmath72 .
right : renormalized obe @xmath1 pn phase shifts ( in degrees ) as a function of cm momentum .
data from @xcite.,title="fig : " ]
besides the numerical coincidence it is not obvious whether or not are we entitled to identify the @xmath0-scalar with the @xmath2-scalar because the @xmath2-scalar has a large width , which suggests that this state decouples provides an inadmissible mid - range repulsion . ] .
we suggest a large @xmath73 motivated scenario where this identification might actually become compelling .
the authors of @xcite suggest that the potential due to iterated @xmath74 scattering can be written for non - vanishing distances @xmath75 ^ 2 t_{00 } ( -q^2)\end{aligned}\ ] ] where @xmath76 is the @xmath77 sigma term and @xmath78 the @xmath2 scattering amplitude in the @xmath79 channel as a function of the cm energy @xmath80 . in the large @xmath73 limit , @xmath81 while @xmath82 yielding @xmath83 as expected @xcite . at the sigma pole @xmath84
^ 2 t_{\pi \pi}^{ii } ( s ) \to \frac{g_{\sigma
nn}^2}{s - ( m_\sigma- i \gamma_\sigma ) ^2 } \to \frac{g_{\sigma nn}^2}{s - m_\sigma^2}\end{aligned}\ ] ] where in the second step we have taken the large @xmath73 limit .
this yields @xmath85 , provided @xmath86 and @xmath87 , a point disputed in ref .
@xcite where the iam method is applied to @xmath2 scattering .
if we use instead the bethe - salpeter method to lowest order @xcite , we get a once subtracted dispersion relation , with an arbitrary constant @xmath88 \label{eq : bs}\end{aligned}\ ] ] where @xmath89 is the tree level amplitude and @xmath90 the two - pion phase space .
the difference between @xmath91 and @xmath92 is higher order in the chiral expansion but both scale as @xmath93 .
we fix the accurately determined scattering length @xmath94 @xcite . for @xmath95 , and @xmath96 we get the pole at @xmath97 although @xmath98 at @xmath99 overshoots the roy analysis value @xmath100 @xcite mainly because higher order chiral corrections @xcite and possibly subthreshold @xmath101 effects @xcite , have been omitted . scaling according to large @xmath73 counting @xmath102 and
@xmath103 the unitarity integral in eq .
( [ eq : bs ] ) can be neglected and the pole satisfies @xmath104 the limit is smooth , and while @xmath105 we get @xmath106 , closer to the @xmath0-scalar . on view of this agreement
it is tempting to think that perhaps the @xmath5 proposed by johnson and teller in 1955 might correspond to the @xmath5 determined by caprini , colangelo and leutwyler in 2006 _ in the large _
@xmath73 limit .
it remains to be seen if higher order chiral and @xmath93 corrections both for @xmath0 as well as for @xmath2 support this view . | we determine the mass and strength of the scalar meson from @xmath0 scattering data by renormalizing the one boson exchange potential .
this procedure provides a great insensitivity to the unknown short distance interaction making the vector mesons marginally important allowing for su(3 ) couplings in the @xmath1 channel .
the scalar meson parameters are tightly constrained by low energy np .
we discuss whether this scalar should be compared to the recent findings based on the roy equations analysis of @xmath2 scattering .
address = departamento de fsica atmica , molecular y nuclear , + universidad de granada , + e-18071 granada , spain .
altaddress = departamento de fsica atmica , molecular y nuclear , + universidad de granada , + e-18071 granada , spain . |
broad absorption line ( bal ) quasars show deep and wide blueshifted absorption troughs in their rest - frame uv spectra ( e.g. , weymann et al .
the intrinsic fraction of quasars with bals is @xmath5 20% ( e.g. , hewett & foltz 2003 ) ; in the most commonly accepted scenario this represents the covering factor of an outflowing bal wind .
while all bal quasars show absorption from high - ionization transitions , such as si iv and c iv , some also display absorption from lower ionization transitions , such as mg ii , and such objects tend to be more reddened ( e.g. , sprayberry & foltz 1992 ; reichard et al .
2003 ) .
although bals were once believed to be confined to radio - quiet sources , optical spectroscopy of the quasars from the _ vla _ 1.4 ghz first survey @xcite , conducted by the first bright quasar survey ( white et al .
2000 ) , and the sloan digital sky survey ( sdss ; york et al .
2000 ) has revealed many radio - loud bal quasars ( e.g. , becker et al .
2000 ; menou et al .
radio - loudness is commonly parameterized by the ratio of rest - frame radio - to - optical flux densities , @xmath7 @xcite , where those quasars with @xmath8 1 are considered radio - loud . as this definition
may somewhat exaggerate the radio - loudness of bal quasars with substantial optical reddening , it is also useful to consider the rest - frame radio luminosity density , where @xmath9 10@xmath10 erg s@xmath11 hz@xmath12 distinguishes radio - loud quasars ( e.g. , miller et al . 1993 ) .
the majority of radio - loud bal quasars discovered to date meet both criteria but tend to be of intermediate radio - loudness ( fewer than 20% in the combined samples of becker et al .
2000 and menou et al .
2001 have @xmath8 2 ) .
@xcite found that about one - third of their bal quasars were flat - spectrum radio sources ( some radio loud ) , suggesting a viewing angle for this subset well above the equatorial region where disk - associated bal outflows would be visible , and further discovered that most of their bal quasars were compact ( 80% unresolved to 0.2@xmath6 at 8 ghz ) .
these results support alternate proposals that bals may be associated with an evolutionary phase with a large bal wind covering fraction , rather than orientation ( e.g. , gregg et al .
2000 ) , although the consistency of the sub - millimeter emission of bal quasars with that of non - bal quasars is difficult to reconcile with evolutionary scenarios in which bal quasars are emerging from a dusty `` shroud '' ( lewis et al .
2003 ; willott et al .
2003 ) .
radio - quiet bal quasars , particularly those with low - ionization absorption features , are usually weaker in x - rays than would be expected from their optical luminosities ( e.g. , green et al . 2001 ; gallagher et al .
the x - ray spectra of radio - quiet bal quasars show clear evidence of x - ray absorption , often complex , with intrinsic column densities ( e.g. , gallagher et al .
although uv and x - ray absorption are clearly linked ( e.g. , brandt , laor , & wills 2000 ) , the higher column density of the x - ray absorber ( e.g. , gallagher et al .
2006 ) suggests the x - ray absorption arises interior to the uv bals , perhaps in the `` shielding gas '' postulated by murray et al .
( 1995 ) and generated naturally in the simulations of @xcite .
a _ chandra _ snapshot survey of five radio - loud bal quasars confirmed they are also x - ray weak relative to similar non - bal quasars , but with fairly soft spectra incompatible with the large column densities necessary to explain this weakness as simple neutral absorption ( brotherton et al .
2005 ) .
many radio - loud quasars display striking extended emission in the form of jets and lobes .
the classification scheme of fanaroff & riley ( 1974 ) distinguishes the edge - darkened , core - dominated fr i sources from the more luminous edge - brightened , lobe - dominated fr ii objects .
the multiwavelength properties of fr i and fr ii jets are quite distinct : fr i jets have linear or convex spectral energy distributions ( seds ) that can typically be modeled by a synchrotron spectrum extrapolated ( with a break if necessary ) from the radio through the optical to the x - ray ( e.g. , worrall et al .
2004 ) , while the seds of the prominent knots in fr ii jets are concave , frequently possessing only an upper limit at optical wavelengths ( e.g. , sambruna et al .
2004 ; marshall et al .
2005 ) . the x - ray emission in fr ii jets is often interpreted as arising from inverse - compton scattering of cosmic microwave background photons ( ic / cmb models ; e.g. , tavecchio et al .
2000 ; celotti et al . 2001 ) , a process that is increasingly efficient for highly relativistic bulk motions , for jet angles close to the line of sight , and for objects at large redshifts ( schwartz 2002 ) .
some observational and theoretical complications with ic / cmb models have been noted ( e.g. , tavecchio et al .
2003 ; hardcastle 2006 ) , and high - redshift radio - loud quasars in general do not appear to have extreme jet - to - core x - ray brightness ratios ( e.g. , siemiginowska et al .
2003 ; bassett et al .
2004 ; lopez et al .
alternative models for the origin of the x - ray emission from fr ii jets have been proposed , including synchrotron emission from a population of highly relativistic electrons ( e.g. , atoyan & dermer 2004 ) .
x - ray hotspots have been detected by _
chandra _ in a number of fr ii sources ; the most luminous hotspots are consistent with synchrotron self - compton ( ssc ) emission but other cases require a second highly energetic synchrotron component to produce the x - ray emission ( hardcastle et al .
2004 ) .
pg 1004 + 130 ( pks 1004 + 13 , 4c 13.41 ) is an optically bright ( @xmath13 = 14.98 ; e.g. , garcia et al .
1999 ) , radio - loud ( @xmath14 2.32 [ wills et al . 1999 ] and @xmath15 6.5@xmath110@xmath10 erg s@xmath16 hz@xmath12 ) , lobe - dominated quasar at a redshift of @xmath17 = 0.240 @xcite .
pg 1004 + 130 is notably x - ray weak ; it was undetected by _ einstein _ with a 0.54.5 kev flux limit of 1.4 @xmath1 10@xmath18 erg @xmath3 s@xmath12 @xcite . @xcite
analyzed _ iue _ and _ hst _ ( goddard high resolution spectrograph ; ghrs ) spectra of pg 1004 + 130 and found evidence for broad , blueshifted absorption in several uv lines , notably o vi , n v , and c iv .
they concluded that pg 1004 + 130 was likely a bal quasar with a balnicity index(weymann et al .
1991 ) of @xmath5 850 km s@xmath12 .
a recent observation with _ hst _ performed with the space telescope imaging spectrograph ( stis ) to search for bal variability has confirmed this designation ( b. j. wills , in preparation ) , making pg 1004 + 130 the only currently known low - redshift ( @xmath19 0.5 ) radio - loud bal quasar .
these results suggest that the x - ray weakness of pg 1004 + 130 is likely related to absorption .
the extended radio structure of pg 1004 + 130 is also notable : it is one of the prototypical examples discussed by gopal - krishna & wiita ( 2000 ) of a hybrid morphology source possessing both an fr i and an fr ii lobe .
the majority of the radio emission in the fr i south - east ( se ) structure is concentrated in a knotted jet that is most prominent close to the nucleus and then fades into a gradually broadening plume , whereas the north - west ( nw ) structure shows classical fr ii edge - brightened morphology .
the structure of such hybrid sources is more intuitively explained as arising from the propagation of twin jets into dissimilar large - scale environments rather than by invoking a `` lopsided '' central engine ( gopal - krishna & wiita 2000 ) .
both the extended projected size of pg 1004 + 130 ( @xmath20500 kpc ) and the ratio of lobe - to - core flux ( we measure this to be @xmath2040 at rest - frame 1.7 ghz using first data ) suggest the jet axis is inclined at a relatively large angle to our line of sight , consistent with physical models in which bal winds flow equatorially from the accretion disk . a quantitative lower limit on the inclination to the line - of - sight , @xmath21 ,
may be estimated from the core radio - to - optical luminosity ratio , @xmath22 , which wills & brotherton ( 1995 ) find improves upon the commonly used lobe - to - core ratio .
pg 1004 + 130 has @xmath23 1.1 , suggesting @xmath24 45@xmath25 ( consistent with wills et al .
this large inclination angle implies little line - of - sight beaming in the jet and constrains the interpretation of any associated x - ray emission .
we have obtained the first x - ray detections and spectra of pg 1004 + 130 in order to check for nuclear absorption and determine the quasar s general x - ray properties .
we make use of both _ xmm - newton _ and _ chandra _ , as they provide complementary information : _ xmm - newton _ has high throughput and covers a broader energy range , whereas _
chandra _ provides the angular resolution necessary to detect and characterize any extended x - ray emission .
the two x - ray observations were conducted @xmath201.7 years apart , allowing investigation of both short and long - term x - ray variability . in this paper
we adopt a standard cosmology with @xmath26 = 70 km s@xmath12 mpc@xmath12 , @xmath27 = 0.3 , and @xmath28 = 0.7 @xcite .
this choice results in a luminosity distance of 1200 mpc and an angular distance scale of 3.8 kpc arcsec@xmath12 for pg 1004 + 130 . the galactic column density toward pg 1004 + 130 ( @xmath29 = 10 07 26.10
, @xmath30 = + 12 48 56.20 ) is 3.93@xmath110@xmath31 @xmath3 @xcite .
unless otherwise noted , errors are given as 90% confidence intervals for one parameter of interest ( @xmath32 = 2.71 ) .
pg 1004 + 130 was observed by _ xmm - newton _ on 2003 may 4 ( obsid 0140550601 ) for an effective exposure time of 22.2 ks , and data were collected by all instruments .
there were insufficient counts in the rgs grating spectra for meaningful analysis ; we therefore restrict our study of the x - ray spectrum to the epic ( pn and mos ) imaging spectroscopy data .
there was no flaring during the observation in the detector - wide count rate for events with energies greater than 10 kev .
analysis was performed with the _ xmm - newton _ science analysis software ( sas v6.5.0 ) .
the nuclear spectrum was extracted from a circular region with a radius of 24.6@xmath6 , while background regions were determined separately for the pn ( a 533@xmath6 by 82@xmath6 rectangle ) and mos ( a 381@xmath6 by 192@xmath6 rectangle ) detectors to avoid detector gaps .
we generated custom redistribution matrix files ( rmfs ) and ancillary response files ( arfs ) , which together describe the instrument response to incident photons and allow forward - fitting of the spectrum .
finally , the extracted nuclear spectrum was binned to contain at least 20 counts per bin using the ftools task grppha .
there are @xmath201550 counts from 0.58 kev in the pn spectrum , and @xmath201100 total counts in the mos1 and mos2 spectra ; the expected number of background counts is @xmath2050 in the pn spectrum and @xmath2030 total in the mos spectra .
the observation included sets of images from the _ xmm - newton _ optical monitor ( om ) for each of the uvw1 , uvm2 , and uvw2 filters , with central wavelengths of 2910 , 2310 , and 2120 , respectively .
we converted background - subtracted om count rates ( automatically calculated during pipeline processing ) to flux densities using scaling factors appropriate for active galactic nuclei ( agns ) , as determined by the om calibration team .
the resulting uvw1 , uvm2 , and uvw2 observed fluxes for pg 1004 + 130 are 6.3 , 8.2 , and 8.1 @xmath1 10@xmath33 erg @xmath3 s@xmath12 @xmath12 , respectively .
the standard deviation of the individual flux density measurements in a given filter is @xmath34 0.2 @xmath1 10@xmath33 erg @xmath3 s@xmath12 @xmath12 , while the systematic uncertainty in the flux calibration is less than 10% .
we obtained a subsequent _ chandra _ observation to confirm the x - ray spectral properties , to investigate variability , and to examine ( at higher angular resolution ) any extended emission .
pg 1004 + 130 was observed by _
chandra _ on 2005 jan 5 ( obsid 5606 ) for 41.6 ks with the advanced ccd imaging spectrometer ( acis ; garmire et al .
2003 ) s - array in faint mode .
these data have been reprocessed ( repro - iii ) to incorporate the latest calibration ( caldb v3.2.0 ) including automatic application of both the acis charge - transfer inefficiency correction and the time - dependent gain adjustment .
the acis - s3 image shows a bright point source at the location of the nucleus of pg 1004 + 130 , jet - associated x - ray emission collinear with but mostly upstream of the se radio jet , and several additional sources , two of which are coincident with the outer regions of the se fr i lobe .
these two sources have point - source _ hst _ counterparts ; utilizing the methodology and results of maccacaro et al .
( 1988 ) , we find that their x - ray - to - optical flux ratios are consistent with those expected for agns .
we extracted the nuclear and the extended x - ray emission for spectral analysis , using regions as shown in figure 1a ( a circle with a 4.8@xmath6 radius for the nucleus and a 4.1@xmath6 by 6.3@xmath6 rectangle for the jet ) and measuring the background from a large source - free elliptical region ( with a 91@xmath6 major axis and a 38@xmath6 minor axis ) east of the nucleus .
there was no significant background flaring during the observation .
source - specific rmfs and arfs were generated from the caldb v3.2.0 database ; the arfs include the effects of low - energy quantum - efficiency degradation due to contamination build - up on the acis optical - blocking filter .
the extracted nuclear spectrum was binned to contain at least 20 counts per bin using the ftools task grppha .
the 0.58 kev nuclear spectrum contains @xmath201800 counts ( of which only @xmath5 5 are expected to be from background ) , and there are @xmath2030 counts in the extended x - ray emission .
we supplement the _ xmm - newton _ and _ chandra _ data with previously published and archival radio , ir , optical , and uv coverage of pg 1004 + 130 .
_ vla _ observations of pg 1004 + 130 have been performed in several configurations and bands , and we make use of 1.4 ghz , 5.4@xmath6 fwhm catalog images from the first survey @xcite as well as c band ( b configuration ) data from 1979 taken by e. fomalont ( fomalont 1982 ) and l band ( a configuration ) data from 1982 taken by j. wardle . images for these latter two observations were constructed using @xmath35 components between 20 and 160 @xmath36 , allowing investigation of the small - scale structure along the jet at a resolution of 1.5@xmath6 while excluding the large - scale structure already apparent in the first image .
ir fluxes were obtained from _ iras _ ( 10012 @xmath37 m ) ,
ground - based ( @xmath38 , @xmath39 , and @xmath40 bands ) , and _ 2mass _ ( @xmath41 , @xmath42 , and @xmath43 bands ) data . an optical spectrum and photometric @xmath44 magnitudes
were taken from the sdss database ( pg 1004 + 130 is sdss j100726.10 + 124856.2 ) and an _ hst _
wfpc2 image of pg 1004 + 130 , taken with the f606w filter and analyzed and discussed by bahcall et al .
( 1997 ) , was used to search for extended optical emission .
pg 1004 + 130 has been observed spectroscopically in the uv by _ hst _ ( initially with ghrs and later with stis ; wills et al .
1999 ; b. j. wills , in preparation ) , _ iue _ ( with both the long and short - wavelength spectrographs ) , and _ fuse _ , and portions of these data have been incorporated as well .
x - ray , optical , and radio images of pg 1004 + 130 are presented in figure 1 .
the x - ray image of figure 1a has been constructed with acis pixel randomization removed to increase angular resolution and is shown with overlaid 4.9 ghz _ vla _ contours ; the jet - associated x - ray emission commences upstream of the radio jet and extends to the first radio knot .
examination of the optical _ hst _ image with accompaning radio and smoothed x - ray contours reveals no apparent optical counterpart to the se fr i jet .
a spectral index map of pg 1004 + 130 generated from 4.9 and 1.5 ghz _
vla _ images of @xmath201.5@xmath6 resolution is presented in figure 1c , along with a high - resolution ( 0.5@xmath6 fwhm ) 5 ghz _ vla _ image of the nuclear region .
there does not appear to be any extended kpc - scale inner - jet radio emission from pg 1004 + 130 .
the spectral index @xmath45 is @xmath460.41 for the nucleus ( @xmath47 ) and varies along the jet .
the error in the spectral index increases greatly toward the edges of the emission , but the signal - to - noise ratio in both bands is quite high down the spine of the jet ; the repeated spectral steepening is genuine .
we rebinned the x - ray image to 0.05@xmath6 pixels and extracted the radial profile along the se radio jet ( using concentric annular sectors restricted to a 30@xmath25 arc ) as well as excluding the jet ( using annular sectors covering the remaining 330@xmath25 ) , subtracting background as determined from a nearby rectangular region .
the background - subtracted surface - brightness profiles are compared with normalized acis - s point spread functions ( created with mkpsf ) in figure 1d .
as the nucleus contains @xmath201080 counts from 0.52 kev and @xmath20730 counts from 28 kev , the pre - calculated 1.5 kev response data are sufficient for this analysis .
the nucleus appears unresolved , but the jet - associated emission is inconsistent with a point - source profile .
we are confident that the jet - associated x - ray emission arises in a genuine counterpart to the se fr i radio jet due to the low probability of finding a background source so close to the pg 1004 + 130 nucleus , the high x - ray - to - optical flux ratio of this feature , and the extended nature of the x - ray emission as well as its position upstream of the peak radio brightness in the jet . from bauer
et al . ( 2004 ) , the expected sky density of sources ( primarily agns ) with 0.52 kev x - ray fluxes comparable to or greater than the of the jet is @xmath48 500 deg@xmath49 .
consequently , the probability of finding a background source of appropriate brightness within 10@xmath6 and aligned ( within 30@xmath25 ) with either the jet or counterjet of pg 1004 + 130 is only @xmath5 0.002 .
further , the lack of any _ hst _ point sources coincident with the x - ray emission restricts the nature of any such background source ; the optical upper limit yields an x - ray - to - optical flux ratio , as defined by maccacaro et al .
( 1988 ) , that is more than @xmath2020 ( @xmath208 ) times greater than the maximum value for agns ( bl lacs ) observed by maccacaro et al .
the x - ray emission is extended along the path of the radio jet , and the peak x - ray brightness in the jet occurs upstream of the peak radio brightness , consistent with the observed trend for the ratio of x - ray - to - radio emission to decrease along jets ( e.g. , marshall et al .
2002 ; sambruna et al .
2004 ) .
large - scale diffuse x - ray emission is common within groups and clusters , is sometimes seen in halos around rlqs , and has been observed along radio jets .
the intra - group , intra - cluster , and halo x - ray emission is well - established to be thermal radiation from hot gas , but the origin of diffuse x - ray emission along jets is much less certain ; ic / cmb scattering has been suggested as a reasonable explanation ( e.g. , siemiginowska et al .
2002 ; schwartz et al .
we adaptively smoothed the 0.54 kev _ chandra _ image to search for diffuse extended emission in the vicinity of pg 1004 + 130 ; the resulting image is presented in figure 2a with overlaid 1.4 ghz first contours .
the local environment of pg 1004 + 130 is sparsely populated ; mclure & dunlop ( 2001 ) find that pg 1004 + 130 has the lowest clustering amplitude among the 13 rlqs they surveyed , and the nearest known companion galaxy lies 33.4@xmath6 ( 127 projected kpc ) to the south - west , with another associated galaxy 45@xmath6 ( 170 kpc ) to the west ( stockton 1978 ; bahcall et al .
there does not appear to be any detectable x - ray halo of quasi - spherical , hot , thermally emitting gas around pg 1004 + 130 .
the hybrid radio morphology of pg 1004 + 130 may suggest that the density of the surrounding medium is higher to the se , where the radio jet is quickly decollimated , than to the nw , where the unseen counterjet powers a distant , edge - brightened lobe , although the lack of any apparent interaction with the few neighboring field galaxies leaves the mechanism for generating such an asymmetrical density distribution unclear .
we do not observe any strong gradient in diffuse x - ray emission from the se to the nw , but naturally this does not prohibit large - scale , asymetrically distributed , colder gas .
however , pg 1004 + 130 does show intriguing diffuse emission along the path of the se jet and nw counterjet .
we verified that this emission was not an artifact of the smoothing process by extracting radial profiles from the rebinned image as with the psf analysis described previously ; here we use the regions indicated in figure 2a to obtain surface - brightness profiles along the jet , counterjet , and non - jet background ( excluding all point sources ) .
figure 2b shows that beyond @xmath2016@xmath6 ( past the se x - ray jet emission ) , the diffuse emission in the jet and counterjet is similar and clearly surpasses the emission from the non - jet background region ; we calculate @xmath2029.3 counts above background in the 0.54 kev diffuse emission from 16@xmath6 to 63@xmath6 to the se , and @xmath2027.4 to the nw . note that the acis readout streak can not contribute significantly to the diffuse emission , as it should contain less than 6 total counts over the jet / counterjet region , is offset by @xmath2015@xmath25 from the jet position angle , and would extend the entire length of the chip . extended x - ray emission around rlqs has occasionally been noted to follow the direction of the radio lobes ( e.g. , croston et al . 2004 ; belsole et al .
2006 ) but detection of an x - ray counterjet component is somewhat unusual ( cf .
schwartz et al .
we briefly consider the nature of this large - scale emission in @xmath504.2
. the x - ray light curves of pg 1004 + 130 do not reveal significant rapid variability within either the _ xmm - newton _ or the _ chandra _ observations .
the cumulative photon arrival times are consistent with a constant count rate with kolmogorov - smirnov probabilities of 0.64 and 0.61 for _ xmm - newton _ ( pn detector ) and _ chandra _ , respectively .
the 1@xmath51 upper limit to the kev count - rate variability on 1 ks timescales is @xmath5212% over the _ xmm - newton _ observation , and @xmath5215% over the _ chandra _ observation . after accounting for the relative number of soft ( 0.52 kev ) and hard ( 28 kev ) band counts ,
the 1@xmath51 upper limit to hard band variability is somewhat greater than that for the soft band .
we analyzed the x - ray spectra using the most recent ( v12.2.1 ) xspec ( arnaud 1996 ) software package .
data associated with energies below 0.4 kev or above 8.0 kev were discarded for the purposes of fitting ; these cutoffs were imposed due to increasing calibration uncertainties at low energies and declining source counts ( due to decreasing instrumental effective area and the spectral shape ) as well as increasing background at high energies .
all fits include galactic absorption fixed at 3.93@xmath110@xmath31 @xmath3 @xcite .
joint fitting of the spectra from the _ xmm - newton _ pn , mos1 , and mos2 detectors increases the signal - to - noise and is consequently preferred in the absence of significant cross - calibration uncertainties . fitting a simple power - law model gives consistent results ( the 90% confidence intervals for both photon index and normalization overlap for the pn , mos1 , and mos2 detectors ) , therefore all model parameters with the exception of normalization have been fit jointly throughout the remainder of the analysis . unless otherwise noted we quote the outcome of fitting various models to the binned spectra with @xmath53 minimization , but the following results agree with those obtained by fitting the ungrouped spectra with the xspec @xmath54-statistic ( after cash 1979 )
. given the bals observed in the uv by @xcite and the correlation between bals and x - ray absorption ( e.g. , gallagher et al .
2002 ) , we anticipated potential x - ray absorption at low energies . in an effort to disentangle the underlying power - law from any intrinsic absorption
, we initially fit only the data in the 28 kev range with a power - law model .
the photon index obtained by jointly fitting the pn , mos1 , and mos2 data from _ xmm - newton _ is @xmath55 = 1.57@xmath56 , consistent with the _ chandra _ result of @xmath55 = 1.52@xmath57 .
this photon index is typical for the high - energy spectra of radio - loud quasars ( rlqs ) ; e.g. , reeves & turner ( 2000 ) found @xmath58 = 1.66 with @xmath51 = 0.22 for an _ asca _ sample of 35 rlqs , while page et al .
( 2005 ) determined @xmath58 = 1.55 with @xmath51 = 0.29 for an _ xmm - newton _ sample of 16 rlqs at @xmath592 .
the 28 kev _ xmm - newton _ pn model flux is 2.57@xmath6010@xmath18 erg @xmath3 s@xmath12 , 23% lower than the _
result of 3.32@xmath6110@xmath18 erg @xmath3 s@xmath12 ; this flux difference exceeds cross - calibration uncertainties , which are on the order of 12% ( s. snowden 2005 , private communication )
. many agns with significant x - ray absorption also have strong fe k@xmath62 emission lines ( e.g. , see the _ asca _ observations of seyfert 2 galaxies by turner et al .
this was predicted by krolik & kallman ( 1987 ) , who argued that the prominence of iron features produced in the scattering and reflection regions increases as the observed x - ray continuum becomes dominated by reprocessed radiation .
examination of the residuals from the power - law fit described above did not show any obvious emission features in the x - ray spectrum of pg 1004 + 130 .
we tested for fe k@xmath62 emission by adding an unresolved ( @xmath51 = 0.01 kev ) redshifted gaussian emission line to the 28 kev power - law model . here
we fit the ungrouped data ( with the xspec @xmath54-statistic ) , as the large energy bins at high energy would tend to smooth out any narrow line in the binned spectra .
the _ xmm - newton _ pn and mos spectra show no evidence for neutral or ionized fe k@xmath62 emission and an upper limit of 105 ev ( 90% confidence ) can be placed on the rest - frame equivalent width of a narrow line at a fixed rest - frame energy of 6.4 kev from the pn spectrum .
modeling the _ chandra _ spectrum by adding a narrow line at rest - frame 6.4 kev imposes an upper limit of 121 ev on the intrinsic equivalent width .
if the line energy is allowed to vary , there is marginal evidence for emission at 6.57@xmath63 kev ; the rest - frame equivalent width of the line is only 105@xmath64 ev .
since this line energy does not match the transition energy for fe k@xmath62 emission from either neutral / low ionization ( 6.4 kev ) , he - like ( 6.7 kev ) , or h - like ( 6.9 kev ) iron , if the marginal detection is assumed to be indicative of physical conditions then there must be a range of ionization states in the scattering material .
overall , the lack of strong fe k@xmath62 emission argues against a scenario in which the nucleus of pg 1004 + 130 is heavily absorbed and the weak continuum arises chiefly from scattering or reflection .
we extrapolated the 28 kev power - law fits to lower energies to search for x - ray absorption , as illustrated in figures 3a and 3b .
many quasars , including radio - loud objects , show enhanced flux above a power - law model below @xmath202 kev ( porquet et al .
2004 ; brocksopp et al .
2006 ) ; no such soft excess is observed from pg 1004 + 130 .
while there is only minimal evidence for intrinsic absorption in the _ xmm - newton _ spectra , the large systematic negative residuals in the _ chandra _ spectrum indicate substantial absorption .
this significant change in the absorption properties of pg 1004 + 130 occurred over only 494 rest - frame days .
we characterized this absorption by adding a redshifted neutral absorber to our model and expanding the range of the fit to 0.48 kev .
the _ xmm - newton _ spectra do not require any intrinsic absorption ( the 90% confidence upper limit is @xmath65 1.6@xmath110@xmath31 @xmath3 ) and the modest negative residuals in figure 3a disappear in figure 3d ( top ) when the photon index is adjusted to @xmath55 = 1.37@xmath66 .
applying this procedure to the _ chandra _ data does indeed indicate the presence of intrinsic absorption , but the negative residuals in the _ chandra _ spectrum of figure 3b are accomodated primarily through an extreme flattening of the photon index to @xmath55 = 1.13@xmath67 .
this model is physically implausible in light of the inconsistency with the best - fit high - energy photon index , and it is also not a particularly good fit ( @xmath53/@xmath68 = 85/76 )
. a lower limit to the intrinsic absorption indicated by the _
spectrum may be obtained by constraining @xmath55 to lie within the 90% errors from the high - energy fit ; the resulting column density is @xmath69 9.29@xmath110@xmath31 @xmath3 , with @xmath55 fixed at 1.26 .
the residuals for this fit ( shown in figure 3d , middle ) suggest that further refinements to the absorption model are required .
complex absorption is common in bal quasars , of both the radio - quiet ( e.g. , gallagher et al .
2002 ) and the radio - loud ( brotherton et al .
2005 ) types , and thus it is perhaps not surprising that the simple intrinsic - absorption model is insufficient to fit the _
spectrum for pg 1004 + 130 .
a partial - covering absorber model gives a better representation ( @xmath53/@xmath68 decreases from 85/76 to 79/75 , an improvement with an @xmath70-test probability of only 0.02 of occuring by chance ; see figure 3d , bottom ) , with parameters @xmath71 = 1.20@xmath7210@xmath2 @xmath3 , @xmath73 = 0.49@xmath74 , and @xmath55 = 1.37@xmath75 . here
@xmath73 is the fraction of the source emission that passes through an intrinsic redshifted absorber with column density given by the fitted @xmath0 , while the remaining 1@xmath76 of the source emission experiences only galactic absorption .
given our photon statistics , we can not constrain the nature of the absorption complexity in detail , but the physical significance of this result is discussed further in @xmath505 . while variable absorption appears to be required in pg 1004 + 130 , a brief consideration of possible alternative explanations for the discrepancy between the _ xmm - newton _ and _
chandra _ soft - band spectra is warranted .
the small - scale jet seen in the _ chandra _ image and unresolved by _ xmm - newton _ is much too weak to explain the differences in the soft x - ray spectra : when the larger _ xmm - newton _ extraction region , which includes the jet , is used to extract a _
spectrum , the resulting parameters for the partial - covering absorber model are similar to those found above , with @xmath0 = 0.95@xmath7710@xmath2 @xmath3 , @xmath73 = 0.42@xmath78 , and @xmath55 = 1.30@xmath79 .
another possibility is that the change in the low - energy x - ray spectrum may be due to variable soft - band emission rather than variable absorption .
it has been suggested that the `` soft - excess '' in fr ii objects may be related to jet emission ( e.g. , evans et al .
2006 ) , and so perhaps a variable unresolved jet might explain the discrepancy between the _ xmm - newton _ and _ chandra _ spectra .
the _ xmm - newton _ spectra can also be adequately fitted with a partial - covering absorber with parameters fixed ( except for normalization ) to the best - fit _
chandra _ values , and an additional `` unresolved jet '' power - law component with a likely photon index fixed at @xmath55 = 1.8 .
the `` unresolved jet '' would then have had to decrease in brightness by a factor of @xmath2015 by the _
observation ; flaring on that order has been observed in the inner jet knot of m87 ( harris et al .
2006 ) , so it is possible ( albeit somewhat contrived ) that the _ xmm - newton _ observation occurred during a flaring episode .
however , the `` unresolved jet '' must contribute approximately one - fourth of the total _ xmm - newton _ 28 kev flux to avoid worsening the fit , and as the _ xmm - newton _ hard - band flux is already observed to be lower than that measured by _
chandra _ , additional variability of the nuclear x - ray emission would be required .
the coincidental brightening of the `` unresolved jet '' during the presumed flare episode to the level necessary to mimic the unbroken , continuous power - law seen in the _ xmm - newton _ spectra further suggests that variable absorption is a more logical explanation than such conspiratorial combinations of variable emission components . note that disfavoring an additional variable `` unresolved jet '' component does not impact interpretation of the underlying origin of the entire 0.58 kev nuclear x - ray emission , a topic considered further in @xmath505 .
the 0.52 kev observed fluxes for the best - fit models as described above are 1.07@xmath110@xmath18 erg @xmath3 s@xmath12 for _ xmm - newton _ and 9.35@xmath110@xmath80 erg @xmath3 s@xmath12 for _ chandra _ , with 28 kev fluxes of 2.60@xmath110@xmath18 erg @xmath3 s@xmath12 and 3.56@xmath110@xmath18 erg @xmath3 s@xmath12 .
the observed 1980 _ einstein _
kev soft - band limit would predict ( for a @xmath55 = 1.5 power - law model with galactic absorption ) 0.52 kev and 28 kev fluxes of 6.7@xmath110@xmath80 erg @xmath3 s@xmath12 and 1.5@xmath110@xmath18 erg @xmath3 s@xmath12 , respectively , with a greater uncertainly applying to the high - energy extrapolation ; both the _ xmm - newton _ and the _ chandra _ soft and hard - band fluxes are higher than suggested by the _ einstein _ non - detection .
the _ einstein _
non - detection can not be explained by simply increasing the covering factor while holding the column density fixed to the value measured in the _ chandra _ spectrum , as even with @xmath73=1 the predicted 0.54 kev flux is higher than the limit measured by _
einstein_. if the _
non - detection results from variable absorption ( rather than emission ) then @xmath0 must have been a factor of @xmath203 higher ( for @xmath73=1 ) to account for the lack of soft - band x - ray flux relative to the _ chandra _ observation .
the unabsorbed 0.52 kev ( rest - frame 0.62.5 kev ) luminosities are 2.08@xmath110@xmath81 erg s@xmath12 for _ xmm - newton _ and 2.98@xmath110@xmath81 erg s@xmath12 for _ chandra _ , with 28 kev ( rest - frame 2.59.9 kev ) luminosities of 4.52@xmath110@xmath81 erg s@xmath12 and 6.44@xmath110@xmath81 erg s@xmath12 .
there are sufficient counts to fit the _ chandra _ x - ray jet spectrum with a simple power - law model ( assuming fixed galactic absorption ) , although the small number of counts necessitates use of the xspec @xmath54-statistic rather than @xmath53 .
the best - fit photon index is @xmath55 = 1.71@xmath82 .
a power - law model is almost universally appropriate for the spectra of jets ( e.g. , worrall & birkinshaw 2006 ) , but with the limited counts available here other possibilities can not be excluded . for example , a themal bremsstrahlung model also yields an acceptable fit ( the @xmath54-statistic value is essentially unchanged ) with @xmath835 kev , although the temperature is poorly constrained . the 0.52 kev and 28 kev fluxes associated with the power - law model are 1.8 and 3.1@xmath110@xmath33 erg @xmath3 s@xmath12 , respectively , corresponding to unabsorbed luminosities of 3.6 and 5.3@xmath110@xmath84 erg s@xmath12 ( quoted for isotropic , unbeamed emission ;
see @xmath504.2 ) .
the optical / uv - to - x - ray spectral slope , @xmath85 , describes the ratio of rest - frame luminosity density at 2500 to that at 2 kev as @xmath86 ; we take @xmath87 and @xmath88 to have units of erg s@xmath12 hz@xmath12 throughout .
the optical / uv and x - ray luminosities of radio - quiet quasars ( rqqs ) are correlated ( @xmath89 ) , although not directly proportional ( e.g. , strateva et al . 2005 ; steffen et al . 2006 ) .
this relationship likely reflects a connection between optical / uv emission from the inner accretion disk and x - ray emission arising from compton upscattering of disk photons in a hot corona . in rlqs
this picture is complicated by additional jet emission , and in both rqqs and rlqs absorption can significantly depress x - ray flux .
the ratio of optical / uv to x - ray luminosity increases with increasing optical / uv luminosity , which may be expressed as an anti - correlation between @xmath85 and @xmath87 ; steffen et al . ( 2006 ) find @xmath85 = @xmath460.139@xmath90 + 2.680 for their sample of unabsorbed rqqs .
we compare the @xmath85 value for pg 1004 + 130 with those of other quasars with @xmath9123 and @xmath92 0.5 in the palomar - green ( pg ; schmidt & green 1983 ) survey .
we make use of data from steffen et al .
( 2006 ) for the pg @xmath85 values .
they determined monochromatic uv luminosities by extrapolating known 3000 values to 2500 assuming @xmath93 = @xmath460.5 , and used _ rosat _ pointed and all - sky survey pspc count rates with an assumed @xmath55 = 2 power - law to calculate monochromatic x - ray luminosities .
the parameterization of the luminosity dependence of @xmath85 for rlqs is not as accurately known as it is for rqqs .
@xcite find that rlqs with flat radio spectra also have somewhat flatter optical / uv - to - x - ray spectra than steep - spectrum rlqs ( which they attribute to stronger jet x - ray emission in the flat - spectrum sources ) , but both flat - spectrum and steep - spectrum rlqs have @xmath94 correlations with slopes consistent ( within the 90% error range ) with those of rqqs .
we therefore remove the luminosity dependence of @xmath85 using the steffen et al .
( 2006 ) relation for rqqs discussed previously , which we find also orders rlqs adequately for purposes of comparison .
the resulting @xmath95 histograms are shown in figure 4a . while there are some quasars in the negative @xmath85 tail of the rqq distribution that lack evidence of intrinsic uv or x - ray absorption , those quasars with strong uv absorption all have anomalously steep optical / uv - to - x - ray spectral slopes .
the two rlqs with confirmed intrinsic absorption ( pg 1309 + 355 and 2251 + 113 both show uv and x - ray absorption ) similarly have more negative values of luminosity - corrected @xmath85 than most of the unabsorbed rlqs . the monochromatic uv luminosity for pg 1004 + 130 was determined from the _ xmm - newton _ om fluxes and sdss spectroscopy , as the dates of these measurements are most nearly coincident with those of the x - ray observations .
the sdss spectrum was scaled by sdss photometric measurements to correct for fiber inefficiencies .
extrapolation to rest - frame 2500 was performed by renormalizing a standard @xcite rlq sed to fit the dereddened @xcite optical / uv data ; the scaled rlq sed was matched to the om uvw1 and uvm1 fluxes and is then only slightly below the continuum of the sdss spectrum , giving a luminosity density of @xmath96= 30.51 .
the monochromatic x - ray luminosity was calculated from the _ chandra _ data for two different best - fit models ( see @xmath503 for details ) , the first a simple power - law fit over the observed kev band ( yielding a `` hard '' @xmath97 = @xmath461.83 determined by the high - energy x - ray spectrum ) and the second a partial - covering neutral absorption model fit over the observed 0.58 kev band ( yielding a `` soft '' @xmath97 = @xmath461.88 primarily influenced by the low - energy x - ray spectrum ) .
the model flux densities were measured at rest - frame 2 kev ( observed 1.6 kev ) and converted to bandpass - corrected luminosity densities .
the results of this analysis are plotted on figure 3 , along with the _ einstein _
@xmath982.01 limit from @xcite .
we note that our @xmath87 is @xmath5 2.5 times less than that used by @xcite to determine @xmath85 , and the _ einstein _ x - ray flux limit with our @xmath87 measurement would give @xmath991.97 .
the 19731990 photographic monitoring of smith et al .
( 1993 ) indicates that pg 1004 + 130 fluctuates in optical brightness by @xmath5 0.5 mag on timescales of 610 yr .
more recently , garcia et al .
( 1999 ; 2006 , private communication ) found that pg 1004 + 130 brightened and then dimmed over a magnitude range of @xmath13 = 14.715.2 from 19931999 ( the sdss 2003 photometry corresponds to @xmath13 = 15.3 ) , and stalin et al .
( 2004 ) found that pg 1004 + 130 dimmed in @xmath100 by 0.09 magnitudes from 1999 march to 2000 april .
the _ einstein _ vs. _ chandra _ discrepancy in @xmath85 values arises from a combination of measurement uncertainties and genuine variability at both optical / uv and x - ray frequencies .
the optical / uv - to - x - ray spectral slope is indeed steeper for pg 1004 + 130 than for other pg rlqs , even if calculated from the hard - band emission , and it is steeper than almost all of the non - absorbed rqqs as well . to differentiate conclusively between excess `` big blue bump '' emission ( see elvis & fabbiano 1984 ) and x - ray weakness as the cause of the low value of @xmath85 , it is helpful to compare the broad - band sed of pg 1004 + 130 with those of other quasars .
the sed of pg 1004 + 130 presented in figure 4b was constructed with radio and ir fluxes , optical and uv spectra and photometry , and the x - ray best - fit models .
there are two sets of radio measurements shown : the parkes data include the extended radio emission , while the _ vla _ data are for the nucleus alone
. the optical ( sdss ) and uv ( _ hst _ ghrs and _ iue _ ) spectra were smoothed to reduce noise , and the geocoronal ly@xmath62 region has been excluded .
the sdss spectrum was scaled to match sdss photometric measurements , as described previously .
data were corrected for galactic extinction ( with @xmath101 = 0.038 mag ) following @xcite .
we have included the best - fit _ xmm - newton _ power - law model and the _ chandra _ partial - covering absorber model with parameters as given in @xmath503 ( in both cases correcting for galactic absorption ) , as well as the 2 kev flux density corresponding to the 0.54.5 kev _ einstein _ limit .
bandpass - corrected luminosity densities at radio - to - x - ray rest - frame frequencies for pg 1004 + 130 are given in table 1 .
comparing the standard quasar seds compiled by @xcite to that of pg 1004 + 130 in figure 4b , there is excellent agreement in the shapes of the seds at radio - to - optical / uv frequencies .
there does not appear to be evidence for enhanced uv emission ( relative to the radio , ir , and optical data ) , but pg 1004 + 130 is distinctly x - ray weak relative to rlqs with comparable optical / uv luminosities . removing the intrinsic absorption apparent in the _ chandra _ spectrum partially accounts for the weakness of the x - ray emission , but the unabsorbed _ chandra _ and _ xmm - newton _ power - law spectra remain below the standard rlq x - ray emission .
some of this apparent x - ray weakness may be a consequence of the chosen method of comparison , as the elvis et al .
( 1994 ) composite seds are biased toward x - ray bright objects due to selection criteria requiring an _ einstein _ detection .
indeed , the average @xmath85 of the rlqs used to contruct the standard sed plotted in figure 4b is @xmath102 = @xmath461.31 , corresponding to relatively greater x - ray luminosities than the @xmath102 = @xmath461.54 average of the pg rlqs . however , the `` hard '' @xmath85 = @xmath461.83 for pg 1004 + 130 is still 1.2@xmath51 below the average @xmath85 of the pg rlqs ; the corresponding ratio by which pg 1004 + 130 is x - ray weak relative to other pg rlqs is @xmath205.4 ( for similar optical / uv luminosities ) . the black - hole mass for pg 1004 + 130 is found by vestergaard & peterson ( 2006 ) to be 1.87@xmath103 10@xmath104 m@xmath105 ( from the fwhm of the h@xmath106 line and the monochromatic optical luminosity at 5100 ) , while falomo et al .
( 2003 ) estimate a mass of 1.35 @xmath1 10@xmath104 m@xmath105 ( from the @xmath107@xmath108 relation ) .
the black - hole masses calculated for pg 1004 + 130 using these different methods are in general agreement and correspond to an eddington luminosity of @xmath202 @xmath1 10@xmath109 erg s@xmath12 .
we measure the bolometric luminosity for pg 1004 + 130 by integrating the scaled standard sed up to x - ray frequencies , then integrating the best - fit power - law model from rest - frame 0.510 kev , and obtain @xmath110 = 2.0 @xmath1 10@xmath111 erg s@xmath12 .
the observed x - ray emission may not be representative of the true x - ray power of the source ( see @xmath505 ) ; if instead the scaled standard sed is integrated to 10 kev , the bolometric luminosity is slightly higher , @xmath110 = 2.3 @xmath110@xmath111 erg s@xmath12 .
both calculations suggest that pg 1004 + 130 is radiating at @xmath112 10@xmath12 l@xmath113 .
the se radio jet is made up of a string of emission peaks ( knots ) , presumably indicating distinct shock sites where particle acceleration takes place .
the spectral index @xmath45 steepens downstream of each knot ( see figure 1c ) , perhaps reflecting an evolution in the underlying electron distribution .
the observed - frame lifetimes for synchrotron cooling for plausible magnetic - field strengths are quite long , corresponding to scales orders of magnitude longer than these projected distances , so if spectral aging is the dominant factor behind this effect the electrons must remain trapped within the shock region for long periods before diffusing downstream .
the x - ray emission begins upstream of the first bright radio knot , and there does not appear to be any jet - associated optical emission .
radio fluxes were extracted from the region encompassing the first knot , overlapping with the end of the x - ray jet extraction region indicated in figure 1a and extending a short distance beyond it . a 5@xmath51 optical upper limit was determined from the noise in the _ hst _ image within the x - ray jet extraction region . the 2 kev @xmath114 taken from the power - law fit to the jet x - ray spectrum and the error bars on @xmath115 are plotted in figure 5 along with the _ hst _ limit and the _ vla _ fluxes .
the power - law fit for the jet is consistent with the @xmath116 1.5 photon indices found by sambruna et al .
( 2004 ) for the brightest x - ray knots in their _ chandra _ and _ hst _ survey of core - dominated fr ii quasars with known radio jets ; however , the x - ray spectra of prominent knots in fr i jets are generally steeper , with @xmath116 2.3 ( e.g. , 3c
66b : hardcastle et al .
2001 ; 3c 31 : hardcastle et al . 2002 ; m 87 : marshall et al . 2002 ;
cen a : hardcastle et al .
2003 ; b2 0755 + 37 : parma et al .
the @xmath117 = @xmath460.87 value for pg 1004 + 130 is similar to those seen for both fr i ( fr i jet references as above ) and fr ii ( sambruna et al . 2004 ; marshall et al .
2005 ) jets .
the pg 1004 + 130 jet optical limit falls well below the power law connecting the radio and x - ray data , ruling out simple single - component synchrotron models .
the broad - band spectral indices are constrained to be @xmath1181.1 , @xmath1190.46 , and consequently @xmath120 2.3 .
these values are similar to those found for fr ii jets ( sambruna et al .
2004 ) but are inconsistent with those of fr i jets , which tend to have @xmath1210.7 , @xmath1221.2 , and thus @xmath1230.6 ( fr i jet references as above ) . despite its standard fr i radio structure , the se pg 1004 + 130 jet shares many of the characteristics of well - known fr ii jets .
while the agreement between the multiwavelength characteristics of the se fr i jet and powerful fr ii jets initially appears somewhat surprising , this result might have been anticipated based on the radio luminosity and hybrid morphology of pg 1004 + 130 .
the radio luminosity density at rest - frame 178 mhz of pg 1004 + 130 is @xmath20 7.1 @xmath1 10@xmath124 erg s@xmath12 hz@xmath12 ( from the flux measurement of wright & otrupcek 1990 ) , more than an order of magnitude above the 2 @xmath1 10@xmath10 erg s@xmath12 hz@xmath12 luminosity density found by fanaroff & riley ( 1974 ) to divide empirically the lower power fr i population from the higher power fr ii sources .
more recently , the radio power separating fr i from fr ii sources has been observed to be an increasing function of the host galaxy optical luminosity ( e.g. , ledlow & owen 1996 ) , and the optical magnitude ( @xmath12524.26 ; falomo et al .
2003 ) of its elliptical host @xcite places pg 1004 + 130 somewhat closer to the observed transition line .
further , the nw lobe of pg 1004 + 130 has standard edge - brightened fr ii structure . if hybrid morphology sources are reflective of dissimilar environments rather than dissimilar jets , as suggested by gopal - krishna & wiita ( 2000 ) , then the se fr i jet should be as intrinsically powerful as the nw fr ii jet .
we have applied various models to the multiwavelength jet emission with the goal of determining the most plausible origin for the x - ray emission .
the results of this analysis are described below and representative models are shown in figure 5 .
the radio - to - optical emission in jets is well - established as synchrotron radiation , with the principle observational support coming from polarization measurements .
the radio data and optical limit allow determination of the magnetic field , assuming equipartition .
based on the _ vla _ images we estimate the emission region to be roughly circular with a radius of around 1.75@xmath6 , corresponding to a spherical volume of 3.6@xmath110@xmath126 @xmath127 . as is standard practice , we assume a power - law electron energy distribution , with index @xmath128=2 to match the spectral slope of the _ vla _ radio data ( see figure 1c ) .
the low energy cutoff for the electron spectrum is observationally unconstrained , and we choose @xmath129=50 , similar to values typically assumed for fr ii jets .
the high energy cutoff is limited by the _ hst _
non - detection , and we use @xmath130=10@xmath131 . in the case of pg 1004 + 130 ,
the large angular size , the high lobe - to - core ratio , and the optical - to - radio core luminosity ratio suggest the jet is inclined to the line - of - sight by @xmath24 45@xmath25 , which would limit the allowed beaming to @xmath132 1.4 .
we consider here the @xmath133=1 case .
we use standard synchrotron formulae ( e.g. , worrall & birkinshaw 2006 ) for a single - injection model with pitch - angle isotropization ( jaffe & perola 1973 ) . a continuous - injection model , in which the spectral slope steepens by 0.5 above a critical frequency ( e.g. , carilli et al . 1991 ) , would still require an exponential cutoff at frequencies below the optical limit , but a kardashev - pacholczyk ( kardashev 1962 ; pacholczyk 1970 ) model with no pitch - angle scattering would allow a synchrotron cutoff at higher frequencies
. the precise slope of the synchrotron spectrum above the turnover frequency does not greatly affect the x - ray emission , so a kardashev - pacholczyk or continuous - injection model would lead to similar qualitative conclusions .
we derive a magnetic field strength @xmath134 = 14 @xmath37 g , and note that in general @xmath135 = @xmath134/@xmath133 ( e.g. , harris & krawczynski 2002 ) . a natural explanation for the origin of the x - ray emission would be compton upscattering by the synchrotron electrons .
we consider two sources of seed photons : the synchrotron radiation itself ( ssc ; e.g. , hardcastle et al .
1997 ) or the cosmic microwave background ( ic / cmb ; e.g. tavecchio et al .
2000 ; celotti et al .
2001 ) . at the distance of the x - ray jet ,
the photon flux from the agn or from the host galaxy is comparatively insignificant .
the x - ray emission expected from the ssc process with the above parameters is more than three orders of magnitude less than the observed x - ray flux .
the x - ray ssc emission increases relative to the synchrotron emission if the magnetic - field strength decreases , but radically sub - equipartition fields are required to attribute the observed x - ray flux to ssc emission . a representative model of this type
is included in figure 5 . both the high and low electron - energy cutoffs
have been adjusted to accommodate the optical limit ( 10@xmath13610@xmath137 ) but given the @xmath128 = 2 power - law distribution and the consistent span of 4.3 decades in energy the equipartion magnetic field remains 14 @xmath37 g .
the actual magnetic field required to fit the x - ray flux is then 0.021 @xmath37 g , several hundred times less than the equipartion value .
ic / cmb is often put forward as an explanation for the concave seds of powerful fr ii jets , and this process is particularly efficient for high - redshift , relativistic jets inclined close to the line of sight . unlike ssc , which is actually depressed by beaming
, x - ray ic / cmb emission is boosted by an additional factor of 1+@xmath62 relative to synchrotron emission ( dermer 1995 ) . in the case of pg 1004 + 130 ,
the expected ic / cmb emission is still several hundred times less than the observed x - ray emission ; the enhanced beaming required to match the x - ray flux , @xmath133=3.0 , would place an upper limit ( where @xmath133 = @xmath55 ) on the jet angle of @xmath13819@xmath25 , smaller than the @xmath24 45@xmath25 suggested from the optical and radio luminosities as well as the radio morphology ( see @xmath501 ) .
further , the lifetime of the low - energy electrons ( @xmath139 10@xmath140 ) responsible for the x - ray emission in ic / cmb models greatly surpasses the length of the jet , and so x - ray emission would be expected to persist along the jet ( absent deceleration ; e.g. , georganopoulos & kazanas 2004 ) . despite the similarities in pg 1004 + 130 to the seds and
x - ray spectra of the core - dominated , highly beamed fr ii jets that dominate the surveys of both sambruna et al .
( 2004 ) and marshall et al .
( 2005 ) , the ic / cmb model commonly applied to such fr ii jets does not appear to be appropriate for pg 1004 + 130 .
for both ssc and ic / cmb models , the location of the x - ray emission largely upstream of the parent synchrotron electrons is difficult to explain and the predicted x - ray spectral slope is flatter than observed .
we consider it unlikely that the x - ray emission arises from either ssc or ic / cmb emission .
if the x - ray jet emission does not arise as a consequence of the low - energy synchrotron radiation ( as in the ssc and ic / cmb cases ) , then various other emission mechanisms may be considered , such as thermal bremsstahlung or synchrotron emission from a secondary population of high - energy electrons ( e.g. , atoyan & dermer 2004 ) .
the offset of the x - ray and radio emission is perhaps more easily accomodated by such models .
the x - ray emission occupies a rectangular region of approximately 5@xmath6 by 2.4@xmath6 ; for an edge - on cylinder this again corresponds to a volume of 3.6@xmath110@xmath126 @xmath127 .
the best - fit temperature for a bremsstrahlung model is @xmath1415 kev , but this is only poorly constrained . neglecting line emission ( which contributes significantly to the soft x - ray emission at lower temperatures ) , a gas cloud with an average ion charge of @xmath1421 would be required to have a density of @xmath1430.05 @xmath144 to account for the observed x - ray flux .
this corresponds to a total mass of 1.5@xmath110@xmath104 m@xmath105 and an ideal - gas pressure of 4.6@xmath110@xmath145 dynes @xmath3 .
such a large quantity of concentrated hot gas at so great a distance from the host galaxy seems unlikely , and as this gas cloud would be overpressured with respect to the surrounding igm we would have to be observing it at a favorable time before it dispersed . if instead the x - ray emission from pg 1004 + 130 arises from a second synchrotron component generated by a population of highly relativistic electrons , the low - energy cutoff must be high enough to avoid over - predicting the optical flux . taking the magnetic field to be 14 @xmath37 g and setting the electron energy index to be @xmath128=2.4
as indicated by the x - ray photon index , this model provides an acceptable explanation of the x - ray emission with @xmath129=1.6@xmath110@xmath146 , as shown in figure 5 . as the lifetime of the x - ray synchrotron electrons is quite short ( electrons initially associated with 1 kev emission would have half lives of @xmath20600 years ) , multiple acceleration or injection sites are required along the extent of the x - ray jet .
electrons with these injection parameters would lose sufficient energy within @xmath2030,000 years to produce fluxes at optical frequencies in excess of the observed _ hst _ limit , suggesting that the energetic electrons escape the shock region before cooling to energies below @xmath1473@xmath110@xmath131 .
the diffuse emission extends 4050@xmath6 ( 150190 projected kpc ) from the nucleus to both the se and the nw , tracing the path of the fr i radio jet and providing supporting evidence for the presence of the hidden fr ii counterjet .
there are insufficient counts for spectral analysis , but with an assumed @xmath55=1.8 power law the 0.58 kev x - ray flux of the se component is 4.5@xmath110@xmath33 erg @xmath3 s@xmath12 , while that of the nw component is 4.2@xmath110@xmath33 erg @xmath3 s@xmath12 .
most radio jets are one - sided , with the absence of a detectable counterjet generally attributed to doppler boosting and hence yielding a constraint upon the line - of - sight angle .
as the line - of - sight angle for pg 1004 + 130 is likely @xmath24 45@xmath25 , the jet / counterjet flux ratio for twin jets is expected to be less than @xmath2015 ( using @xmath148^{\alpha-2}$ ] and @xmath55 = @xmath149 = 1/@xmath150 = 1.4 ; e.g. , worrall & birkinshaw 2006 ) . however , the observed ratio of radio emission in the se jet to that in the undetected nw counterjet exceeds 100 , indicating that doppler boosting alone can not explain the lack of a radio counterjet in pg 1004 + 130 .
the diffuse x - ray emission precludes the possibility of an intrinsically one - sided jet , leading us to hypothesize that the entraining environment is indeed less dense to the nw , as suggested by the hybrid radio morphology .
if the diffuse x - ray emission is thermal radiation , then the required gas density is @xmath204.5@xmath110@xmath151 @xmath144 with a total mass of 2.3@xmath110@xmath152 m@xmath105 and a pressure of 1.5@xmath110@xmath153 dynes @xmath3 ; while these parameters are not as restrictive as those for the thermal jet models , similar concerns apply .
the mechanism for heating the gas along the jet is unclear , and a non - thermal origin seems somewhat more plausible .
ssc emission alone is not a viable explanation for the diffuse x - ray emission , as the absence of a detectable radio counterjet to the nw imposes a stringent limit to the available synchrotron photon density above 1 ghz , and the concurrent lack of energetic electrons makes boosting to x - ray frequencies difficult .
the diffuse x - ray emission could result from unbeamed ic / cmb emission , and indeed its apparently smooth extent along the entire jet suggests an association with long - lived electrons .
the paucity of radio emission in the nw counterjet region does not greatly affect the ic / cmb x - ray yield , which is driven by low - energy electrons associated with sub - ghz synchrotron radiation .
diffuse x - ray emission unaccompanied by detectable radio emission has been observed in a handful of additional sources and can be successfully attributed to ic / cmb processes in those cases as well ( e.g. , siemiginowska et al . 2002 ; schwartz et al .
the _ vla _
c band limit on radio emission in the nw counterjet allows for considerable leeway in determing the high - energy cutoff and magnetic field required to generate sufficient diffuse x - ray emission via the ic / cmb process . for @xmath154 = 50 and @xmath155
= 10@xmath156 , the magnetic field is required to be lower ( @xmath202 @xmath37 g ) than in the jet region , with equipartition electron densities @xmath2030 times less than in the knots .
the primary results from our analysis of the first x - ray detections and spectra of pg 1004 + 130 are the following : \1 .
variable complex absorption : although the _ xmm - newton _ observation of pg 1004 + 130 shows only minimal intrinsic absorption , the _ chandra _ spectrum reveals significant soft x - ray absorption that can not be modeled by a simple redshifted neutral absorber .
the best - fit parameters for the preferred partial - covering model are @xmath157 = 1.20@xmath7210@xmath2 @xmath3 , @xmath73 = 0.49@xmath74 , and @xmath55 = 1.37@xmath75 .
the 28 kev _ xmm - newton _ flux is 23% lower than the 28 kev _ chandra _ flux .
x - ray weakness : pg 1004 + 130 has the lowest luminosity - corrected value of @xmath85 among the pg rlqs , and after correcting for intrinsic absorption it is @xmath205.4 times weaker in x - rays than the other pg rlqs , when normalizing to similar optical / uv luminosities .
examination of the sed of pg 1004 + 130 confirms that the anomalous value of @xmath85 is due to x - ray weakness rather than optical / uv brightness .
x - ray jet : the _ chandra _ image reveals an x - ray counterpart to the se radio fr i jet .
the jet is undetected by _ hst _ , ruling out simple single - component synchrotron emission models , and both the x - ray spectrum and the multiwavelength sed shape are similar to those observed for prominent knots in fr ii jets . diffuse x - ray emission
is observed along the path of the jet and counterjet .
complex x - ray absorption is common in radio - quiet bal quasars ( e.g. , gallagher et al .
2002 , 2006 ) and may be inferred by x - ray weakness coupled with relatively soft x - ray spectra for radio - loud bal quasars as well ( brotherton et al .
there is precedent for variable x - ray absorption such as that seen in pg 1004 + 130 ; gallagher et al .
( 2004 ) found that pg 2112 + 059 showed a factor of @xmath207 increase in intrinsic @xmath0 over three years ( 483 rest - frame days ) .
the more absorbed _ chandra _
spectrum for pg 2112 + 059 required either an ionized or partially covering absorber to fit the flat soft x - ray region and also revealed fe k@xmath62 emission undetectable against the higher continuum of the earlier _ asca _ observation , with a rest - frame equivalent width of 1050@xmath158 ev .
the best - fit @xmath0 for pg 1004 + 130 is relatively low compared with that of most radio - quiet bal quasars , as is the covering fraction ( cf .
green et al . 2001 ) .
together these results suggest that the nature of the x - ray absorber is broadly similar in radio - quiet and radio - loud bal quasars , but that the column density associated with the observed x - ray absorption is lower in radio - loud bal quasars , somewhat analogous to the trend for uv absorption to be stronger in radio - quiet bal quasars than in radio - loud bal quasars @xcite .
although notably x - ray weak relative to non - bal rlqs , pg 1004 + 130 is not as x - ray weak as the bal rlqs examined by brotherton et al .
( 2005 ) , and it is possible that pg 1004 + 130 is simply an intrinsically x - ray weak rlq that also shows variable absorption .
alternative hypotheses as to the cause of the x - ray weakness of pg 1004 + 130 are constrained by the characteristics of the _ xmm - newton _ and _ chandra _ spectra .
the x - ray weakness can not be attributed solely to attenuation by simple intrinsic neutral absorption , as the required high column densities would result in x - ray spectra significantly harder than observed .
if much of the intrinsic x - ray emission is hidden by heavy absorption of significantly higher column density ( @xmath1595@xmath110@xmath160 @xmath3 ) than that indicated by the flat spectral shape , the observed x - ray spectrum could result from photons leaking through `` holes '' in the absorber or scattering off a `` torus '' or electron - cloud mirror into the line of sight .
however , the absence of strong fe k@xmath62 emission is somewhat surprising if the latter scenario is correct , and pg 1004 + 130 is actually brighter in x - rays than expected for a reflection - dominated continuum .
if the nucleus is obscured for photon energies up to @xmath5 8 kev ( or extremely weak in x - rays ) , it is also possible that we are viewing x - ray emission from a subparsec - scale jet ; this may provide a natural explanation for the absence of prominent fe k@xmath62 emission , as well as the observed lack of excess low - energy emission that distinguishes the x - ray spectra of pg 1004 + 130 from those of other quasars .
motivated by the observed correlation between radio and optical luminosity for narrow - line radio galaxies , chiaberge et al .
( 2000 ) suggested a synchrotron origin for the nuclear optical emission from these sources .
this correlation has been extended to the x - ray band by evans et al .
( 2006 ) , who argue that fr i rlqs derive a significant fraction of their ( generally unabsorbed ) nuclear x - ray emission from an unresolved jet , while fr ii rlqs are dominated by ( absorbed ) accretion - powered x - ray emission but also contain a jet spectral component . in the particular case of pg 1004 + 130 ,
the complexity and intermediate column density of the absorption in the observed _ chandra _ x - ray spectrum could plausibly be attributed to viewing the jet through progressively diminishing bal - wind column densities with increasing distance from the nucleus .
x - ray spectra with improved photon statistics and resolution are required to constrain better the physical nature of the nuclear x - ray emission , or it may also be possible to detect direct x - rays in the @xmath5 8200 kev band if the putative absorption does not exceed @xmath1612@xmath110@xmath162 @xmath3 ( e.g. , matt 2002 ) .
detection of rapid x - ray variability would restrict the size of the emission region and limit the degree to which reprocessed radiation could contribute to the observed continuum , but neither the _ xmm - newton _ nor the _ chandra _ observation shows such variability .
identification of the x - ray emission mechanism in the pg 1004 + 130 jet would be aided by an optical detection of the jet and determination of the frequency and nature of the break in the sed between radio and x - ray wavelengths .
however , we consider it unlikely that the x - ray jet is dominated by ic / cmb emission , and instead favor the x - ray emission arising from a second synchrotron component .
as with other x - ray jets ( e.g. , marshall et al .
2002 ; sambruna et al .
2004 ) , the x - ray jet of pg 1004 + 130 peaks in brightness upstream of the brightest radio knot .
this suggests that x - ray synchrotron emission may be more prominent in the inner jet , either because of stronger magnetic fields or a supply of highly energetic electrons ( e.g. , sambruna et al .
the agreement in the multiwavelength properties of the pg 1004 + 130 jet with those of other fr ii jets indicates that the fr i radio morphology is likely due to propagation into a dense environment rather than intrinsically lower power , as suggested by .
the orientation of the jet axis demonstrates that concave jet seds can arise without substantial beaming .
the diffuse x - ray emission traces the path of the jet and hidden counterjet and suggests that low - energy electrons inhabit the entire length from nucleus to lobes , more consistent with continuous jet emission from the central engine than with sporadic activity ( e.g. , stawarz et al .
2004 ) .
we gratefully acknowledge the financial support of nasa grant sao sv4 - 74018 ( gpg , principal investigator ) , nasa ltsa grant nag5 - 13035 ( bpm , wnb , dps ) , _ xmm - newton _ grant nag5 - 13541 ( bpm , wnb ) , the _ spitzer _ fellowship program , under award 1256317
( scg ) , and nasa grant go-09432 from the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas5 - 26555 ( bjw ) .
ed fomalont kindly reduced and analyzed the _ vla _ data used in this paper , and we appreciate his generous assistance with this project . we thank george chartas for helpful discussions , and we thank an anonomous referee for many constructive suggestions .
this work includes observations obtained with _ xmm - newton _ , an esa science mission with instruments and contributions directly funded by esa member states and nasa .
funding for the sdss and sdss - ii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , the u.s .
department of energy , the national aeronautics and space administration , the japanese monbukagakusho , the max planck society , and the higher education funding council for england .
the sdss web site is http://www.sdss.org/. cccll band & frequency & luminosity & date & source + & 8.00 & 34.26 & & pks 80 mhz + & 8.34 & 33.85 & & pks 178 mhz + radio & 8.70 & 33.58 & & pks 408 mhz + & 9.24 & 33.26 ( 31.92 ) & & pks ( vla ) 1.4 ghz + & 9.53 & 33.05 & & pks 2.7 ghz + & 9.79 & 32.77 ( 31.70 ) & & pks ( vla ) 5 ghz + & 12.57 & 32.85 & 1983 & iras 100 @xmath37 m + & 12.79 & 32.40 & 1983 & iras 60 @xmath37 m + & 13.17 & 32.24 & 1983 & iras 25 @xmath37 m + & 13.49 & 31.98 & 1983 & iras 12 @xmath37 m + & 13.55 & 31.56 & 1986/01 & @xmath38 ( bjw ) + ir & 13.89 & 31.40 & 1986/01 & @xmath39 ( bjw ) + & 13.98 & 31.23 & 1986/01 & @xmath163 ( bjw ) + & 14.03 & 31.17 & 1986/01 & @xmath40 ( bjw ) + & 14.23 & 30.83 & 2000/04/06 & 2mass @xmath41 + & 14.35 & 30.69 & 2000/04/06 & 2mass @xmath42 + & 14.47 & 30.67 & 2000/04/06 & 2mass @xmath43 + & 14.62 & 30.64 & 2003/01/28 & sdss @xmath17 + & 14.70 & 30.66 & 2003/01/28 & sdss @xmath164 + optical & 14.78 & 30.61 & 2003/01/28 & sdss @xmath165 + & 14.90 & 30.61 & 2003/01/28 & sdss @xmath166 + & 15.02 & 30.61 & 2003/01/28 & sdss @xmath167 + & 15.11 & 30.48 & 2003/05/04 & om uvw1 + uv & 15.21 & 30.44 & 2003/05/04 & om uvm1 + & 15.24 & 30.38 & 2003/05/04 & om uvw2 + & 15.59 & 29.48 & 2000/12/17 & fuse + & 17.68 & 25.36 & 1980/05/09 & _ einstein _ + x - ray & 17.68 & 25.59 & 2005/01/05 & _ chandra _ + & 17.68 & 25.56 & 2003/05/04 & _ xmm - newton _ + | we investigate the x - ray properties of pg 1004 + 130 , a radio - loud broad absorption line ( bal ) quasar with a hybrid fr i / fr ii radio morphology .
this optically bright , low - redshift quasar was undetected by _ einstein _ , marking it as anomalously x - ray weak relative to other radio - loud quasars . the 22.2 ks _ xmm - newton _ and 41.6 ks _
chandra _
observations presented here are the first detections of pg 1004 + 130 and constitute the highest spectral quality x - ray observations of a radio - loud bal quasar available to date .
the _
chandra _
spectrum shows evidence for complex soft x - ray absorption not detected in the data obtained 1.7 yr previously with _
xmm - newton _
, with a best - fit intrinsic column density of @xmath0=1.2 @xmath1 10@xmath2 @xmath3 for the preferred partial - covering model .
there is no significant difference in the hard - band power - law photon index of @xmath41.5 between the two observations .
the _
chandra _ image also reveals extended x - ray emission @xmath58@xmath6 ( 30 kpc ) south - east of the nucleus , aligned with the fr i jet but upstream of the 1.4 ghz radio - brightness peak .
the jet is not detected by _
hst _ , and the optical upper limit rules out a simple single - component synchrotron interpretation of the radio - to - x - ray emission . the multiwavelength characteristics of the pg 1004 + 130 jet , including its relatively flat x - ray power law and concave spectral energy distribution , are similar to those of powerful fr ii .
the lack of strong beaming in pg 1004 + 130 limits the efficiency of inverse compton upscattering , and we consider the x - ray emission to most likely arise from a second synchrotron component generated by highly energetic electrons . |
in searching for new physics , low - energy high - precision measurements are complementary to the lhc high - energy frontier .
the long - standing ( 34)@xmath5 discrepancy between the experimental value of the muon anomalous magnetic moment @xmath6 and the standard model ( sm ) prediction , @xcite , has been considered during these years as one of the most intriguing indications of physics beyond the sm . however , the accuracy of the sm prediction , @xmath7 , is limited by strong interaction effects , which can not be computed perturbatively at low energies . long time ago , by using analyticity and unitarity , it was shown @xcite that the leading - order ( lo ) hadronic contribution to the muon @xmath4 - 2 , @xmath8 , could be computed via a dispersion integral of the hadron production cross section in @xmath9 annihilation at low - energy . the present error on @xmath8 , @xmath10 , with a fractional accuracy of 0.6% ,
constitutes the main uncertainty of the sm prediction .
an alternative evaluation of @xmath8 can be obtained by lattice qcd calculations @xcite .
even if the current results are not yet competitive with those obtained with the dispersive approach via time - like data , their errors are expected to decrease significantly in the next few years @xcite .
the @xmath11 hadronic light - by - light contribution , @xmath12 , which has the second largest error in the theoretical evaluation , contributing with an uncertainty of ( 2.54)@xmath13 , can not at present be determined from data and its calculation relies on the use of specific models @xcite . from the experimental side ,
the error achieved by the bnl e821 experiment , @xmath14 ( corresponding to 0.54 ppm ) @xcite , is dominated by the available statistics .
new experiments at fermilab and j - parc , aiming at measuring the muon @xmath4 - 2 to a precision of @xmath15 ( 0.14 ppm ) , are underway @xcite . and @xmath16 .
dhmz is ref .
@xcite , hlmnt is ref .
@xcite ; smxx is the average of the two previous values with a reduced error as expected by the improvement on the hadronic cross section measurement ; bnl - e821 04 ave . is the current experimental value of @xmath17 ; new ( @xmath4 - 2 ) exp . is the same central value with a fourfold
improved precision as planned by the future @xmath4 - 2 experiments at fermilab and j - parc @xcite.,scaledwidth=50.0% ] fig .
[ fg : g-2new - v3 ] , from ref .
@xcite , shows the status of the @xmath4 - 2 discrepancy compared with what could be expected after the new @xmath4 - 2 measurements at fermilab and j - parc , assuming that the central value would remain the same .
together with a fourfold improved precision on the experimental side , an improvement on the lo hadronic contribution is highly desirable .
differently from the dispersive approach , which uses time - like data from annihilation cross sections , our proposal is to determine @xmath8 from a measurement of the effective electromagnetic coupling in the space - like region , where the vacuum polarization is a smooth function of the squared momentum transfer .
this approach to evaluate @xmath8 has been recently proposed @xcite by using bhabha scattering data . a method to determine the running of the electromagnetic coupling by using small - angle bhabha scattering was proposed in @xcite and applied to lep data in @xcite .
the hadronic contribution to the running of @xmath18 could also be determined unambiguously by using the @xmath19-channel @xmath20 scattering process , from which @xmath8 could be obtained , as detailed in the following .
the paper is organized as follows . after a short review of the theoretical framework in sect .
[ sec : theory ] , we present our experimental proposal in sect . [ sec : exp ] .
preliminary considerations on the detector and systematic uncertainties are given in sect .
[ sec : det ] and sect .
[ sec : unc ] , respectively , while our conclusions are drawn in sect .
[ sec : concl ] .
with the help of dispersion relations and the optical theorem , the lo hadronic contribution to the muon @xmath4 - 2 is given by the well - known formula @xcite @xmath21 where @xmath22 is the ratio of the total @xmath23 and the born @xmath24 cross sections , @xmath25 is a smooth function and @xmath26 ( @xmath27 ) is the muon ( pion ) mass .
we remark that @xmath22 in the integrand function of eq.([amu ] ) is highly fluctuating at low energy due to resonances and threshold effects .
the dispersive integral in eq .
( [ amu ] ) is usually calculated by using the experimental value of @xmath22 up to a certain value of @xmath28 @xcite and by using perturbative qcd ( pqcd ) @xcite in the high - energy tail . for the calculation of @xmath29 , an alternative formula
can also be exploited @xcite , namely @xmath30,\ ] ] where @xmath31 is a space - like ( negative ) squared four - momentum and @xmath32 is the hadronic contribution to the running of the fine - structure constant .
in contrast with the integrand function of eq .
( [ amu ] ) , the integrand in eq .
( [ amu_xalpha ] ) is smooth and free of resonances . by measuring the running of @xmath33 , @xmath34 where @xmath35 and @xmath36 is the fine - structure constant in the thomson limit
, the hadronic contribution @xmath32 can be extracted by subtracting from @xmath37 the purely leptonic part @xmath38 , which can be calculated order - by - order in perturbation theory ( it is known up to three loops in qed @xcite and up to four loops in specific @xmath39 limits @xcite ) .
involves @xmath40 evaluated at negative space - like momenta @xmath41 .
[ xfunction ] ( left ) shows @xmath42 and @xmath43 as functions of the scale @xmath44 defined in eq .
( [ t ] ) .
the integrand function of eq .
( [ amu_xalpha ] ) , @xmath45 $ ] , is plotted in fig .
[ xfunction ] ( right ) , using the output of the routine ` hadr5n12 ` @xcite ( which uses time - like hadroproduction data and perturbative qcd ) .
the range @xmath46 corresponds to @xmath47 , with @xmath48 for @xmath49 .
the peak of the integrand occurs at @xmath50 ( @xmath51 gev@xmath52 ) and @xmath53 ( see fig .
[ xfunction ] ( right ) ) .
we propose to use eq .
( [ amu_xalpha ] ) to determine @xmath8 by exploiting the measurement of @xmath32 in the space - like region . measuring the running of @xmath33 in the space - like region , using a muon beam with @xmath54 gev on a fixed electron target , with a technique similar to the one described in @xcite for a measurement of the pion form factor , is very appealing for the following reasons : * it is a @xmath19-channel process , making the dependence on @xmath19 of the differential cross section proportional to @xmath55 : @xmath56 where the first factor on the right - hand side @xmath57 refers to the effective born cross section including virtual and soft photons , analogously to ref .
@xcite , where small - angle bhabha scattering at high energy was considered . the vacuum polarization effect , in the leading photon @xmath19-channel exchange ,
is incorporated in the running of @xmath18 and gives rise to the second factor @xmath58 .
it is understood that for a high precision measurement also higher - order radiative corrections must be included . for a detailed discussion see refs .
@xcite ; * given the incoming muon energy @xmath59 , in a fixed - target experiment the @xmath19 variable is related to the energy of the scattered electron @xmath60 or its angle @xmath61 : @xmath62 @xmath63 @xmath64 where @xmath65 * the boosted kinematics of the collision allows all the scattering angles to be accessed by a single detector element in the laboratory system .
the angle @xmath61 spans the range ( 031.85 ) mrad for the electron energy @xmath60 in the range ( 1139.8 ) gev ( the low - energy cut at 1 gev is arbitrary ) . by using the same detector to cover the whole acceptance , many systematic errors , _
e.g. _ on the efficiency , will cancel out ( at least at first order ) in the relative ratios of event counts in the high and low @xmath39 regions ( signal and normalization regions ) ; * for @xmath66 gev , @xmath67 gev@xmath52 and @xmath68 gev@xmath69 gev@xmath52 ( @xmath70 , where @xmath71 is the klln function ) , which implies that @xmath19 spans the peak region of the integrand function of eq .
( [ amu_xalpha ] ) , as visible in fig .
[ xfunction ] ( right ) , corresponding to an electron scattering angle @xmath72 1.5 mrad ; * the angle of the scattered electron and muon are correlated as shown in fig .
[ giovanni ] ( for muons of @xmath73 gev ) .
this constraint is extremely important to select elastic scattering events , rejecting background events from radiative or inelastic processes and to minimize systematic effects in the determination of @xmath19 .
note that for scattering angles of ( 23 ) mrad there can be an ambiguity between the outgoing electron and muon , as their angles and momenta are similar . to associate them
correctly it is necessary to identify the two particles by means of downstream dedicated detectors ( calorimeter and muon detectors ) .
+ we estimate the statistical sensitivity of this experiment ( supplemented with large @xmath74 contributions derived from pqcd ) to be @xmath3 on the value of @xmath75 after two years of data taking , using 30 experimental points in @xmath76 , assuming a running time of @xmath77 s / yr , and a muon beam of 150 gev with an average intensity of @xmath2 muon / s .
such a beam is currently available at the cern north area @xcite , and a preliminary detector layout is described in the following section .
in order to perform the planned measurement to the required precision , a dedicated detector is necessary . we describe here a possible setup to measure the following observables : * direction and momentum of the incident muon ; * directions of the outgoing electron and muon .
the cern muon beam m2 , used at 150 gev , has the characteristics needed for such a measurement .
the beam intensity appears to be adequate to provide the required event yield .
the beam time structure allows to tag the incident muon while keeping low the background related to incoming particles ( _ e.g. _ electrons ) .
the electrons contamination is very small .
the beam provides both positive and negative muons , which we plan to use .
the target consists of atomic electrons . to reach the required statistics ,
the target must consist of an adequate amount of material to give a sufficient number of electron scattering centres .
the target has to be made of a low-_z _ material to minimize the impact of multiple scattering and the background due to bremsstrahlung and pair production processes . a promising idea for
the detector is to use 20 layers of be ( or c ) coupled to si planes , spaced by intermediate air gaps , located at a relative distance of one meter from each other .
[ be - detector ] shows the basic layout , presently under study .
target ( be or c ) and two silicon tracking stations located at a distance of one meter .
( b ) to perform the @xmath78 discrimination in the case of small scattering angles ( both @xmath79 and @xmath80 below 5 mrad ) the detector is equipped with an electromagnetic calorimeter and a muon detector . ]
the arrangement provides both a distributed target with low-_z _ and the tracking system . as downstream particle identifiers we plan to use a calorimeter for the electrons and a muon system for the muons ( a filter plus active planes )
this particle identifier system is required to solve the muon - electron ambiguity for electron scattering angles around ( 23 ) mrad ( @xmath81 . fig .
[ giovanni ] ) .
the preliminary studies of such an apparatus , performed by using geant4 , indicate an angular resolution for the outgoing particles of @xmath82 mrad .
the detector acceptance must cover the region of the signal , with the electron emitted at extremely forward angles and high energies , as well as the normalization region , where the electron has much lower energy ( around 1 gev ) and an emission angle of some tens of mrad .
the boosted kinematics of the process allows the detector to cover almost 100% of the acceptance .
the incoming muons have to be tagged and their direction and momentum precisely measured . to this purpose ,
a detector similar to those used by compass @xcite or na62 @xcite can be employed .
significant contributions of the hadronic vacuum polarization to the @xmath0 differential cross section are essentially restricted to electron scattering angles below 10 mrad , corresponding to electron energies above 10 gev .
the net effect of these contributions is to increase the cross section by a few per mille : a precise determination of @xmath8 requires not only high statistics , but also a high systematic accuracy , as the final goal of the experiment is equivalent to a determination of the differential cross section with @xmath8310 ppm systematic uncertainty at the peak of the integrand function ( @xmath81 . fig .
[ xfunction ] ) . such an accuracy can be achieved if the efficiency is kept highly uniform over the entire @xmath39 range , including the normalization region , and over all the detector components .
this motivates the choice of a purely angular measurement : an acceptance of tens of mrad can be covered with a single sensor of modern silicon detectors , positioned at a distance of about one meter from the target .
it has to be stressed that particle identification ( electromagnetic calorimeter and muon filter ) is necessary to solve the electron - muon ambiguity in the region below 5 mrad .
the wrong assignment probability can be measured with the data by using the rate of muon - muon and electron - electron events .
another requirement for reaching very high accuracy is to measure all the relevant contributions to systematic uncertainties from the data themselves . an important effect , which distinguishes the normalization from the signal region , is multiple scattering , as the electron energy in this region is as low as 1 gev .
multiple scattering breaks the muon - electron two - body angular correlation , moving events out of the kinematic line in the 2d plot of fig .
[ giovanni ] .
in addition , multiple scattering in general causes acoplanarity , while two - body events are planar , within resolution .
these facts allow multiple scattering effects to be modelled and measured by using data .
an additional handle on multiple scattering could be the inclusion of a thin layer in the apparatus , made of the same material as the main target modules .
this possibility will be studied in detail with simulation . in experiments dedicated to high - precision measurements
, several systematic effects can be explored within the experiment itself . in this respect
the proposed modularity of the apparatus will help .
a test with a single module could provide a proof - of - concept of the proposed methods . from the theoretical point of view
, the control of the systematic uncertainties requires the development of high - precision monte carlo tools , including the relevant radiative corrections to reach the needed theoretical precision . to this aim ,
qed radiative corrections at leading - logarithmic level resummed at all orders of perturbation theory and matched to the exact @xmath84 correction ( as for instance implemented in the ` babayaga ` event generator @xcite for bhabha scattering ) are mandatory , ensuring a theoretical precision at the level of @xmath85 on the differential cross section .
moreover , by using the ratio of the cross sections in the signal and normalization regions , we expect that the theoretical uncertainties will be further reduced to the level of @xmath86 , due to partial cancellation of common radiative corrections .
work is in progress to extend ` babayaga ` to @xmath87 scattering and to quantify the actual accuracy on the computation of the ratio of cross sections by means of dedicated monte carlo simulations .
any further improvement in the theoretical accuracy would require the matching of qed resummation with exact two - loop corrections , which are not available at present for the @xmath87 process but are within reach .
we presented a novel approach to determine the running of @xmath33 in the space - like region and @xmath1 , the leading hadronic contribution to the muon @xmath4 - 2 , by scattering high - energy muons on atomic electrons of a low-@xmath88 target through the process @xmath0 .
the experiment is primarily based on a precise measurement of the scattering angles of the two outgoing particles as the @xmath39 of the muon - electron interaction can be directly determined by the electron ( or muon ) scattering angle .
an advantage of the muon beam is the possibility of employing a modular apparatus , with the target subdivided in subsequent layers .
a low-_z _ solid target is preferred in order to provide the required event rate , limiting at the same time the effect of multiple scattering as well as of other types of muon interactions ( pair production , bremsstrahlung and nuclear interactions ) .
the normalization of the cross section is provided by the very same @xmath0 process in the low-@xmath39 region , where the effect of the hadronic corrections on the fine - structure constant is negligible .
such a simple and robust technique has the potential to keep systematic effects under control , aiming to reach a systematic uncertainty of the same order as the statistical one . for this purpose
a preliminary detector layout has been described . by considering a beam of 150 gev muons with an average intensity of @xmath89 muons
/ s , currently available at the cern north area , a statistical uncertainty of @xmath3 can be achieved on @xmath1 in two years of data taking .
a test of a single module could provide a validation of the proposed method .
we would like to thank carlo broggini , lau gatignon and fred jegerlehner for fruitful discussions , and fedor ignatov for fruitful discussions and help in the simulation .
this work was supported in part by the italian ministry of university and research under the prin project 2010yj2nyw .
m.p . acknowledges partial support by fp10 itn elusives ( h2020-msca - itn-2015 - 674896 ) and invisibles - plus ( h2020-msca - rise-2015 - 690575 ) . c. bouchiat , l. michel , j. phys
. radium 22 ( 1961 ) 121 ; + l. durand , phys .
* 128 * ( 1962 ) 441 [ erratum - ibid . * 129 * ( 1963 ) 2835 ] ; + m. gourdin , e. de rafael , nucl .
b * 10 * ( 1969 ) 667 . c. aubin , t. blum , phys . rev .
d * 75 * ( 2007 ) 114502 ; + p. boyle , l. del debbio , e. kerrane , j. zanotti , phys .
d * 85 * ( 2012 ) 074504 ; + x. feng , k. jansen , m. petschlies , d.b .
renner , phys .
lett . * 107 * ( 2011 ) 081802 ; + m. della morte , b. jager , a. juttner , h. wittig , jhep * 1203 * ( 2012 ) 055 ; + t. blum _ et al .
_ , phys .
* 116 * ( 2016 ) no.23 , 232002 ; + b. chakraborty , c. t. h. davies , p. g. de oliviera , j. koponen and g. p.
lepage , arxiv:1601.03071 [ hep - lat ] .
t. blum , m. hayakawa , t. izubuchi , pos lattice * 2012 * ( 2012 ) 022 ; + b. chakraborty , c. lehner , r. van de water , talks at ichep 2016 , available at ` https://indico.cern.ch/event/432527/contributions/speakers ` .
m. knecht , a. nyffeler , phys .
d * 65 * ( 2002 ) 073034 ; + k. melnikov , a. vainshtein , phys .
d * 70 * ( 2004 ) 113006 ; + j. prades , e. de rafael , a. vainshtein , arxiv:0901.0306 [ hep - ph ] .
+ a. nyffeler , phys .
d * 79 * ( 2009 ) 073012 . c. m. carloni calame , m. passera , l. trentadue and g. venanzoni , phys .
b * 746 * ( 2015 ) 325 a.b .
arbuzov , d. haidt , c. matteuzzi , m. paganoni and l. trentadue , eur .
j. c * 34 * ( 2004 ) 267 .
g. abbiendi _ et al . _
[ opal collaboration ] , eur .
j. c * 45 * ( 2006 ) 1 .
f. jegerlehner , `` the anomalous magnetic moment of the muon '' , springer tracts mod .
* 226 * , 2008 .
m. davier , a. hoecker , b. malaescu , z. zhang , eur .
j. c * 71 * ( 2011 ) 1515 [ erratum - ibid .
c * 72 * ( 2012 ) 1874 ] .
k. hagiwara , r. liao , a.d .
martin , d. nomura , t. teubner , j. phys .
g * 38 * ( 2011 ) 085003 .
harlander , m. steinhauser , comput .
commun .
* 153 * ( 2003 ) 244 .
lautrup , a. peterman , e. de rafael , phys .
rept . * 3 * ( 1972 ) 193 .
m. steinhauser , phys .
b * 429 * ( 1998 ) 158 .
baikov , k.g .
chetyrkin , j.h .
kuhn and c. sturm , nucl .
b * 867 * ( 2013 ) 182 ; + c. sturm , nucl .
b * 874 * ( 2013 ) 698 ; + p.a .
baikov , a. maier , p. marquard , nucl .
b * 877 * ( 2013 ) 647 .
f. jegerlehner , in proceedings of `` fifty years of electroweak physics : a symposium in honour of professor alberto sirlin s 70th birthday '' , new york university , 27 - 28 october 2000 , j. phys .
g * 29 * ( 2003 ) 101 .
s. eidelman and f. jegerlehner , z. phys .
c * 67 * ( 1995 ) 585 ; + f. jegerlehner , nucl .
* 181 - 182 * ( 2008 ) 135 g. balossini , c. m. carloni calame , g. montagna , o. nicrosini and f. piccinini , nucl .
b * 758 * ( 2006 ) 227 ; + c. m. carloni calame , phys .
b * 520 * ( 2001 ) 16 ; + c. m. carloni calame , c. lunardini , g. montagna , o. nicrosini and f. piccinini , nucl .
b * 584 * ( 2000 ) 459 . | we propose a new experiment to measure the running of the fine - structure constant in the space - like region by scattering high - energy muons on atomic electrons of a low-_z _ target through the process @xmath0 .
the differential cross section of this process , measured as a function of the squared momentum transfer , provides direct sensitivity to the leading - order hadronic contribution to the muon anomaly @xmath1 . by using a muon beam of 150 gev , with an average rate of @xmath2 muon / s , currently available at the cern north area ,
a statistical uncertainty of @xmath3 can be achieved on @xmath1 after two years of data taking .
this direct measurement of @xmath1 will provide an independent determination , competitive with the time - like dispersive approach , and consolidate the theoretical prediction for the muon @xmath4 - 2 in the standard model .
it will allow therefore a firmer interpretation of the measurements of the future muon @xmath4 - 2 experiments at fermilab and j - parc . |
the experimental realization of quasi-2d frustrated antiferromagnets,@xcite such as @xmath3 , @xmath4 , and @xmath5 has recently generated a renewed interest in the physics of the so - called _ collinear _ antiferromagnets . in these compounds
the relevant superexchange interactions involve @xmath6 ions on weakly coupled stacked planes , and the magnetic behavior is likely to be described by the @xmath0 heisenberg model on the square lattice , @xmath7 where @xmath8 and @xmath9 are the ( positive ) nearest - neighbor ( _ n.n . _ ) and next - nearest - neighbor ( _ n.n.n .
_ ) couplings , respectively . in the experimentally relevant case @xmath10
, frustration is known to originate in the classical ( @xmath11 ) limit a collinear low - temperature phase.@xcite in fact , for @xmath12 , the ground state for @xmath11 is a state where the spins are ferromagnetically aligned along one direction and antiferromagnetically in the other , corresponding to magnetic wave vectors @xmath13 or @xmath14 .
these two families of states break not only the su(2 ) symmetry and the translational invariance of the hamiltonian , as the conventional nel state , but also its invariance under @xmath15 real - space rotations .
interestingly , this additional two - fold degeneracy is expected to generate non - trivial finite - temperature properties .
in fact , an ising - like order parameter can be defined in order to discriminate between the two classes of states which are connected by real space @xmath15 rotations , and a finite - temperature second - order phase transition is expected.@xcite how these low - temperature properties are affected by quantum fluctuations , especially for the experimentally interesting @xmath16 case , is an open problem currently preventing a faithful comparison with the experiments.@xcite in fact , due to numerical instabilities induced by frustration ( _ sign - problem_),@xcite this issue can not be investigated with the powerful stochastic numerical methods successfully employed for the _
n.n _ heisenberg model @xcite , so that approximate approaches have to be followed.@xcite in this paper , we address the reliability of the spin - wave ( @xmath17 ) theory as an analytical tool to give _ quantitative _ predictions on the low - energy properties of the @xmath0 heisenberg model in the collinear phase .
to this purpose , a previously introduced finite - size @xmath17 theory @xcite is generalized to the collinear phase and a direct comparison with lanczos exact diagonalization results @xcite is performed for @xmath16 .
this allows us to demonstrate the effectiveness of @xmath17 theory in describing the low - energy properties of the @xmath0 heisenberg antiferromagnet in the experimentally relevant regime @xmath10 .
the finite - size @xmath17 theory of zhong and sorella,@xcite is a rigorous generalization of the standard @xmath17 approach to finite clusters with @xmath18 sites , which allows one to avoid the spurious goldstone - mode divergences related to the su(2 ) symmetry breaking assumption in a straightforward way , without imposing any _ ad hoc _ holonomic constraint on the sublattice magnetization.@xcite in the collinear case , a systematic finite - size @xmath17 expansion can be derived for @xmath12 considering as the approximated ground state the symmetric linear combination @xmath19 and @xmath20 are the ( normalized ) @xmath17 ground states obtained assuming the classical collinear order with @xmath13 , and @xmath14 , respectively . since on any finite size the ground state is expected to have all the spatial symmetries of the hamiltonian it is important to consider the symmetric linear combination ( [ gs ] ) in order to obtain an accurate description of its correlations . in addition , this is also crucial to generalize rigorously the finite - size @xmath17 theory to the collinear phase . both @xmath21 and @xmath20 in eq . ( [ gs ] )
can be obtained according to the standard @xmath17 procedure .
the first step is to apply the unitary transformation which defines a spatially varying reference frame pointing along the local direction of the spins in the classical state , assumed in the @xmath22 plane : @xmath23 $ ] .
then , using the holstein - primakoff spin - boson transformation to first - order in @xmath24 , @xmath25 , @xmath26 , @xmath27 , and translational invariance , the leading term of the hamiltonian in the @xmath17 expansion reads @xmath28\ ] ] where @xmath29 is the classical ground - state energy , and @xmath30 a_r^\dagger$ ] , with @xmath31 varying in the first brillouin zone of the lattice . here
@xmath32 and @xmath33 for @xmath14 and with @xmath34 for @xmath13 .
such leading part of the hamiltonian , being bilinear in the bose operators , can be diagonalized for @xmath35 using the well - known bogoliubov transformation , @xmath36 , with @xmath37 , @xmath38 , @xmath39 being the @xmath17 dispersion relation .
in contrast , the goldstone modes can not be diagonalized by this transformation since @xmath40 and @xmath41 are not defined for the @xmath42-vectors for which @xmath43 . for infinite systems such divergences
are integrable in two dimensions and do not lead to any contribution . on any finite - size system , instead , they are important and they must be treated independently . to this end , it is possible to define a set of hermitian operators commuting with each other and with the hamiltonian : @xmath44 for @xmath45 and @xmath46 and @xmath47 for @xmath48 and @xmath49 . with these definitions
the goldstone - mode contribution to ( [ hsw ] ) reads @xmath50 $ ] , with the sum extended only to the four singular modes .
this allows us to diagonalize the hamiltonian for any finite size in a basis where the @xmath51 s have definite quantum numbers . in particular ,
being @xmath52 for @xmath12 , @xmath53 favors a ground state with @xmath54 and it is easy to show that this implies to order @xmath24 @xmath55 and @xmath56 where @xmath57 is the @xmath58 component of the magnetic structure factor .
these relations together with the sum rules @xmath59 , following from the translational invariance of the @xmath17 hamiltonian , are consistent with a singlet ground - state with a zero value of the total spin on each sublattice .
this is rigorously true according to the lieb - mattis theorem @xcite for @xmath60 .
in fact , in this limit the system decouples into two independent @xmath61 heisenberg models on the two sublattices . in general , a singlet ground state is also expected for the @xmath0 model on any finite size with an even number of sites and non - frustrating boundary conditions .
the sum rules @xmath62 , instead , are not necessarily true on a finite size and for finite values of @xmath63 but they turn out to be fulfilled with good approximation in the collinear phase ( see below ) . according to this procedure ,
the @xmath17 estimate of the ground - state energy reads , @xmath64 $ ] , and it turns out to be in remarkable agreement with the exact results for @xmath16 and @xmath65 in the whole range @xmath66 ( fig . [ erg ] ) , i.e. , far enough from the classical critical point @xmath67 separating the collinear from the nel zero - temperature phases .
in fact , in this strongly frustrated regime , a non - magnetic ground state of purely quantum mechanical nature is likely to be stabilized for @xmath16 ( ref . ) .
it is therefore natural to expect @xmath17 theory to fail in describing its correlations . indeed , approaching the fully frustrated point @xmath67 , the @xmath17 expansion eventually breaks down leading to a zero value of the antiferromagnetic order parameter @xcite and of the uniform susceptibility ( see below ) .
a little more involved calculation allows us also to access the spin - rotation invariant spin - spin correlation functions : @xmath68 .
the different contributions coming from the two degenerate states in eq .
( [ gs ] ) , @xmath69 , can be found by adding to the hamiltonian a term of the form @xmath70 where the vectors @xmath71 are the @xmath72 equivalent lattice vectors that can be reached from a site @xmath73 with @xmath74 .
the resulting @xmath17 hamiltonian has the same form of eq .
( [ hsw ] ) with field - dependent coefficients and it can be diagonalized similarly to the zero - field case . then , according to the hellmann - feynman theorem,@xcite the spin - spin correlations can be obtained by differentiating the @xmath17 energy in presence of the perturbation with respect to @xmath75 , after a careful treatment of the singular modes .
in particular , in the collinear case , such rigorous treatment is possible only assuming for the ground state the symmetric form ( [ gs ] ) .
the resulting correlation functions have different expressions depending on the form of the vector @xmath76 .
in particular , for @xmath76 connecting the same sublattice , they read : @xmath77 if @xmath78 , and @xmath79 if @xmath80 , with the sums extended to all the non singular ( @xmath81 ) @xmath42 s . for
@xmath76 connecting different sublattices , instead , the spin - spin correlations are : @xmath82 where @xmath83 with @xmath84 , and @xmath85 with @xmath86 . the @xmath17 spin - spin correlation functions calculated as @xmath87 are invariant under the symmetry operations of the lattice , as expected for the ground - state expectation values on any finite size . using these expressions the spin - isotropic magnetic structure factor , @xmath88 ,
can be also calculated .
in addition , having obtained an ordered expansion in @xmath24 for the spin - spin correlation functions , @xmath89 , the antiferromagnetic order parameter , @xmath90,@xcite can be expanded as @xmath91 with @xmath92 , and it reads : @xmath93~.\ ] ] note that for @xmath94 the known @xmath17 result @xcite for the thermodynamic order parameter , @xmath95 , is recovered . as shown in fig .
[ esseq ] , the @xmath17 magnetic structure factor , @xmath96 , is in remarkable quantitative agreement with the exact diagonalization result for @xmath16 and @xmath97 for all the wave vectors @xmath98 , and the @xmath17 sum rule @xmath99 is fulfilled almost exactly . in particular
, the @xmath17 approach provides a very accurate estimate ( within a few percent ) of the antiferromagnetic order parameter in the whole range of @xmath100 .
the sharp drop of the order parameter at @xmath101 is suggestive of a first order transition between the gapped and the collinear phase , as also indicated by a recent variational study.@xcite this leads in particular to a very fast saturation of the order parameter to the value of the @xmath61 heisenberg model,@xcite expected for @xmath102 .
remarkably , the @xmath17 prediction for @xmath103 is in excellent agreement with the result of recent neutron scattering measurements @xcite on @xmath3 , giving @xmath104 in the zero temperature limit .
this compound is known to have @xmath105 , even if the precise value of the frustration ratio is presently much debated .
@xcite unfortunately , the weak dependence of the order parameter on @xmath63 for @xmath103 does not allow to use the @xmath17 prediction to determine more precisely this ratio . within the finite - size @xmath17 theory , it is also possible to study the structure of the low - energy spin excitation spectrum which is connected to the ground - state antiferromagnetic correlations and to the uniform susceptibility .
following lavalle _
et al._,@xcite in order to stabilize an excitation of total spin @xmath106 , a zeeman term , @xmath107 , is added to the spin hamiltonian .
classically , for magnetic fields not large enough to induce a spin - flop transition , the classical solution is simply obtained by canting the spins of an angle @xmath108 along the direction of the field @xmath109 [ @xmath110 . after a new rotation of the reference frame in order to align it to the new directions of the spins , the finite - size @xmath17 expansion is straightforward and again it leads to a linearized hamiltonian of the same form of eq .
( [ hsw ] ) with field - dependent coefficients @xmath111 , @xmath112 , and classical energy . in this case , due to the reduced spin - rotation symmetry of the hamiltonian only the @xmath113 and the @xmath46 modes are singular .
the former favors a value of @xmath114 consistent with the applied field , at the classical level , while the latter generates the usual approximated sum rule @xmath115 , which is not affected by the uniform field along the @xmath116 direction .
the total spin @xmath117 of the excitation can be related to the magnetic field @xmath109 by means of the hellmann - feynman theorem @xmath118~,\end{aligned}\ ] ] where @xmath119~,\ ] ] and @xmath120 . using these expressions
the energy spectrum @xmath121 can be calculated by means of a legendre transform , @xmath122 . besides
, the uniform susceptibility in the thermodynamic limit , @xmath123 , can be obtained from eq .
( [ eh ] ) by direct derivation and it reads : @xmath124 where @xmath125 is the susceptibility in the classical limit . due to the discreteness of the energy spectrum ,
a direct comparison of the @xmath17 susceptibility with the exact diagonalization results is not possible on a finite size .
however , in a quantum antiferromagnet with a long - range ordered ground state , the susceptibility is directly related to the properties of the low - energy spectrum and this can be exploited to establish the accuracy of the @xmath17 predictions .
in fact , whenever long - range order is present in the thermodynamic limit , the low - lying spin excitations ( @xmath126 ) order themselves as in the spectrum of a free quantum rotator , @xmath127 , where @xmath128 is the so - called _ momentum of inertia _ per site .
this definition is very close to the definition of the uniform spin susceptibility , which can be calculated by taking first the infinite - volume limit of the energy per site at fixed magnetization @xmath129 , @xmath130 , and then letting @xmath131 : @xmath132 .
as it is known from the low - energy theory of a quantum antiferromagnet,@xcite this allows for an identification between @xmath128 and @xmath133 so that the quantity @xmath134^{-1 } = n\left(e(s)-e_0\right)\left[s(s+1)\right]^{-1}$ ] approaches for infinite size the physical inverse susceptibility @xmath135 , for any spin - excitation @xmath136 .
this features are clearly verified by the @xmath17 scheme as it is illustrated in the lower panel of fig .
[ chisw ] for @xmath97 , where @xmath137 is plotted for @xmath138 and approaches the thermodynamic value calculated according to eq .
( [ eq.chisw ] ) ( @xmath139 ) .
the comparison with the exact results of the @xmath17 spectrum @xmath140 vs @xmath141 ( inset of fig . [ chisw ] ) , and , equivalently , of the quantity @xmath142 ( upper panel of fig . [ chisw ] ) demonstrates the accuracy of @xmath17 theory in estimating the spin susceptibility of the @xmath0 model in the collinear phase for @xmath1 .
as it is shown in the same figure ( upper panel ) , the classical uniform susceptibility is strongly renormalized by quantum fluctuations for @xmath16 at the @xmath17 level . as expected , approaching @xmath67 such reduction is enhanced due to the increased frustration and it leads eventually to the vanishing of the susceptibility for @xmath143 and to the expected transition to a spin - gapped non - magnetic phase.@xcite however , from the comparison with the exact diagonalization results for both the uniform susceptibility and the order parameter , it appears clear that the effects of quantum fluctuations are underestimated by the @xmath17 expansion in the regime of strong frustration so that the transition to the non - magnetic region is likely to occur at slightly higher values of the @xmath63 ratio ( @xmath144 ) .
in conclusion , we have presented a finite - size spin - wave study of the @xmath0 heisenberg antiferromagnet in the collinear phase , focusing in particular on the ground - state spin - spin correlations and the low - energy excitations . for @xmath16 ,
the comparison with exact diagonalization results reveals a remarkable agreement for @xmath1 .
in particular , both the antiferromagnetic structure factor and the spin susceptibility are very close to the spin - wave predictions .
the accuracy of the results on finite sizes indicates that the spin - wave expansion provides a quantitatively reliable description of the ground state of the @xmath0 model in the experimentally relevant regime @xmath105 .
this is confirmed by recent neutron scattering measurements on @xmath3 , providing a value of the order parameter in excellent agreement with the spin - wave prediction in the collinear phase , @xmath2 .
the accuracy of spin - wave theory in describing the low - energy excitations also suggests that theoretical approaches to the thermodynamics based on a gaussian treatment of quantum fluctuations , such as effective hamiltonian methods , @xcite are likely to successfully describe the finite - temperature properties of the @xmath0 model in the collinear phase . work on this line of research would allow to establish a closer contact between the frustrated heisenberg model and the physical properties of the recently synthesized collinear antiferromagnets , helping in their precise characterization and in the determination of the actual frustration ratios .
i wish to thank alberto parola , sandro sorella , and valerio tognetti for their teachings , support , and continuous encouragement .
special thanks to pietro carretta for a very useful correspondence and for kindly communicating the experimental results on @xmath3 before publication .
thanks to federico becca for help in performing the lanczos calculations , and to frederic mila , and doug scalapino for many fruitful discussions .
special thanks also to liz osterberg for a careful reading of the manuscript .
cpu time kindly provided by sissa and infm - democritos is gratefully acknowledged .
this work was supported by nsf under grant dmr-9817242 .
r. melzi , p. carretta , a. lascialfari , m. mambrini , m. troyer , p. millet , and f. mila , , 1318 ( 2000 ) ; r. melzi , s. aldovrandi , f. tedoldi , p. carretta , p. millet , and f. mila , , 024409 ( 2001 ) .
p. carretta , r. melzi , n. papinutto , and p. millet , , 47601 ( 2002 ) ; p. carretta , n. papinutto , c. b. azzoni , m.c .
mozzati , e. pavarini , s. gonthier , and p. millet , cond - mat/0205092 .
the nature of the non - magnetic phase is presently a debated issue .
see for instance : l. capriotti , f. becca , a. parola , and s. sorella , , 097201 ( 2001 ) ; , 212402 ( 2003 ) ; v.n .
kotov , j. oitmaa , o.p .
sushkov , and z. weihong , _ phil . mag .
b _ , * 80 * 1483 ( 2000 ) . | the ground - state and low - energy properties of the two - dimensional @xmath0 heisenberg model in the collinear phase are investigated using finite - size spin - wave theory [ q. f. zhong and s. sorella , _ europhys .
lett . _
* 21 * , 629 ( 1993 ) ] , and lanczos exact diagonalizations . for spin one - half
where the effects of quantization are the strongest the spin - wave expansion turns out to be quantitatively accurate for @xmath1 . in this regime , both the magnetic structure factor and the spin susceptibility are very close to the spin - wave predictions . the spin - wave estimate of the order parameter in the collinear phase , @xmath2 , is in remarkable agreement with recent neutron scattering measurements on @xmath3 . |
perhaps the most perplexing and long - standing problem associated with galactic and cluster cooling flows is the uncertain physical nature and spatial distribution of the gas that cools .
the apparent absence of large masses of cooled gas in elliptical galaxies has led some to argue that little or no cooling actually occurs and to postulate some source of heating that offsets the radiative losses in x - ray emission .
but the energy required to balance radiative losses is prohibitively large and appropriate heating sources may not be universally available .
if the expected radiative cooling actually occurs , two questions arise : ( 1 ) what is the nature of the objects that condense from the hot gas ? and
( 2 ) where is most of the cooled mass located in the galaxy ? regarding the first question ,
a variety of physical arguments support the hypothesis , or even the inevitability , of low mass star formation ( fabian , nulsen & canizares 1982 ; thomas 1986 ; cowie & binney 1988 ; vedder , trester , & canizares 1988 ; sarazin & ashe 1989 ; ferland , fabian & johnstone 1994 ; mathews & brighenti 1999 ) . here
we shall address the second question in the context of cooling flows in elliptical galaxies where the known stellar mass and light profiles strongly constrain the spatial distribution of cooled gas .
we adopt the generally accepted hypothesis that only stars of very low mass ( e.g. @xmath5 @xmath6 ) form in cooling flows ( e.g. ferland , fabian & johnstone 1994 ) , so that the mass to light ratio of the young stellar population formed from the cooling gas is essentially infinite . in view of the difficulties we encounter with this hypothesis , described below
, it seems more likely that the stellar population formed from cooled gas extends to somewhat more massive stars that are optically luminous .
our gas dynamical models for the evolution of hot interstellar gas in giant ellipticals indicate that the origin of the gas varies with galactic radius .
most of the gas in the inner , optically luminous regions originates from the ejected envelopes of evolving stars ; gas in the outer halo is supplied by cosmological secondary infall or tidal acquisitions from neighboring galaxies ( mathews & brighenti 1998b ) .
circumgalactic gas around massive ellipticals is enriched by type ii supernovae that accompanied early star formation .
the variability of circumgalactic gas among luminous ellipticals is responsible for some of the enormous dispersion in x - ray luminosity @xmath7 among ellipticals of similar optical luminosity @xmath8 ( mathews & brighenti 1998a ) .
since the hot interstellar gas in a bright elliptical emits observable x - rays , it is clearly losing energy .
however , as the gas loses energy it is compressed toward the galactic center by gravitational forces and @xmath9 work maintains the high temperatures observed , @xmath10 k , producing a galactic cooling flow .
the positive interstellar temperature gradients typically observed within a few effective radii are often cited as evidence of radiative cooling in a cooling flow , but this cooling is due instead to the mixing of somewhat cooler , locally virialized gas ejected from stars with hotter gas arriving from larger galactic radii ( mathews & brighenti 1998b ; brighenti & mathews 1998 , 1999a ) .
if large entropy fluctuations are present in the hot gas , catastrophic cooling can occur at any radius in the flow .
regions of low entropy ( low temperature , high density ) radiate more and cool sooner .
the amplitude distribution of entropy fluctuations in the interstellar gas determines the radius where cooling mass dropout occurs in the cooling flow . for example , if the entropy in some region in the flow is only slightly less than in the ambient flow , the differential radiative cooling will be small and the region will cool out of the flow at small galactic radii ; conversely , localized regions with entropy much lower than the ambient flow cool rapidly and deposit their mass at large radii . some possible sources of interstellar entropy variations are stellar winds , explosions of type ia supernovae , non - uniform snii heating at early times , and mergers with small , gas - rich galaxies . the total rate that mass cools and drops out of the flow is closely related to the x - ray luminosity @xmath7 .
the x - ray luminosity can be approximately expressed as the product of the total cooling rate @xmath11 and the enthalpy per gram in the hot gas , or @xmath12 here we have used data from the giant virgo elliptical ngc 4472 : @xmath13 k ; @xmath14 ; @xmath15 .
if @xmath7 and @xmath16 are reasonably constant over the hubble time , a mass @xmath17 @xmath6 of cold gas is expected to condense from the hot ism somewhere within ngc 4472 .
although this mass is very large , it is only about 4 percent of the total stellar mass in ngc 4472 today .
the mass that cools can therefore be ignored if it is widely distributed throughout the galactic volume . however , the central concentration of h@xmath18 emission in ellipticals ( e.g. macchetto et al .
1996 ) suggests that the cooling is concentrated toward the galactic center where the interstellar density is highest and the bulk of the x - ray energy is emitted .
the motivation of this paper is to explore a variety of options for the mass dropout profile of cooled gas in bright ellipticals appropriately constrained by the known radial distributions of total stellar and non - baryonic mass .
the radial mass dropout profile of cooled gas can not be determined from first principles because the distribution and amplitude of the entropy and magnetic fluctuations in the hot gas are unknown and difficult to evaluate from simple physical arguments .
nevertheless , the total mass of cooled gas inside an effective ( half - light ) radius @xmath3 must be consistent with the mass to light ratio determined from stellar velocities and with the total mass inferred from x - ray observations within @xmath3 .
assuming that the stellar mass to light ratio is uniform with radius , the stellar mass @xmath19 and the x - ray mass @xmath20 appear to be in nearly perfect agreement for two bright virgo ellipticals in the range @xmath21 ( brighenti & mathews 1997a ) .
because of constraints on the mass distribution of cooled gas provided by x - ray and stellar dynamical observations , galactic cooling flows provide a critical venue for testing the physics of mass deposition in cooling flows .
the mass of cold ( @xmath22 k ) gas @xmath23 actually observed in ellipticals is many orders of magnitude less than the total cooled mass estimated above .
for example , neither hi nor h@xmath24 gas has been observed in ngc 4472 , only upper limits , @xmath25 @xmath6 ( bregman , roberts & giovanelli 1988 ; braine , henkel & wiklind 1988 ) .
if stars form they must either be indistinguishable from the old stellar population or non - luminous .
we explore here the possibility that most of the cooled gas forms dense baryonic clouds or stars that are dark at optical and radio frequencies .
the very low mass stars advanced by ferland , fabian & johnstone ( 1994 ) satisfy this invisibility criterion , while the star formation models of mathews & brighenti ( 1999 ) indicate that ( luminous ) stars of mass @xmath26 @xmath6 can form in galactic cooling flows .
x - ray studies indicate significant masses of cold , absorbing gas in cluster and galactic cooling flows ( e.g. white et al .
1991 ; allen et al . 1993 ; fabian et al .
1994 ; allen & fabian 1997 ; buote 1999 ) , but these results are inconsistent with the absence of radio frequency emission from the cold gas ( braine & dupraz 1994 ; donahue & vogt 1997 ) and should be regarded as controversial until this inconsistency is resolved .
even taken at face value , the total mass of cooled gas implied by the x - ray absorption in cluster cooling flows is typically only a small fraction of the total mass that should have cooled in a hubble time ( allen & fabian 1997 ; wise & sarazin 1999 ) , implying that most of the cooled gas may have formed stars . if cooled gas forms into small stars , these stars will have apogalatica near their point of origin where they will spend most of their orbital time .
we shall assume that the gas mass that cools and drops out of the flow contributes optically dark ( stellar ) mass at the radius where the cooling occurred . in the following we describe a series of gas dynamical calculations for the evolution of x - ray emitting interstellar gas over the hubble time and
investigate a variety of assumptions about the radial distribution of optically dark cooled gas . to be specific ,
we compare our models with the well - observed massive elliptical ngc 4472 .
we find that the mass of cooled gas contributes significantly to dynamical mass to light determinations within @xmath3 based on stellar velocities .
the e2 elliptical ngc 4472 is a luminous , slowly rotating [ @xmath27 galaxy in the virgo cluster . with an adopted distance @xmath28 mpc ,
its optical luminosity is @xmath29 @xmath30 and its half light or effective radius @xmath31 is @xmath32 kpc ( faber et al . 1989 ) .
the total stellar mass @xmath33 @xmath6 is found from the mass to light ratio @xmath34 determined by van der marel ( 1991 ) with a two - integral stellar distribution function .
this mass to light ratio is appropriate to the galactic region within about 0.4@xmath3 where stellar velocities are well determined , although the mass determined from x - ray observations suggests that @xmath35 remains constant to at least @xmath3 ( brighenti & mathews 1997a ) .
if @xmath35 is spatially constant , the stellar mass also has a de vaucouleurs profile . within a central core or break radius @xmath36 ( @xmath37 pc ) the stellar density profile flattens ( faber et al .
1997 ) , but we shall not consider this small feature here .
ngc 4472 contains a central black hole of mass @xmath38 @xmath6 ( magorrian et al .
1998 ) . the total mass distribution in luminous ellipticals can most easily be determined from the radial variation of density and temperature in the hot interstellar gas , assuming hydrostatic equilibrium .
figure 1 illustrates the interstellar density and temperature profiles in ngc 4472 .
the filled circles in figure 1 are _ einstein _
hri observations ( trinchieri , fabbiano & canizares 1986 ) and open circles are rosat hri and pspc data from irwin & sarazin ( 1996 ) ( also see forman et al . 1993 ) . the @xmath39 and @xmath40 profiles have been fit with analytic curves as described by brighenti & mathews ( 1997a ) .
hydrostatic equilibrium in the hot interstellar gas is an excellent approximation since the cooling flow velocity is very subsonic .
the total mass interior to radius @xmath41 determined from x - ray observations is @xmath42 where @xmath43 is the proton mass and @xmath44 is the mean molecular weight .
the last term , representing the possibility of magnetic pressure @xmath45 , is negative if @xmath46 as seems likely .
if the magnetic term is important but not included , the total mass will be underestimated .
assuming @xmath47 , the total mass @xmath20 in ngc 4472 is shown with a solid line in figure 1b . in the outer halo , @xmath48 ,
the total mass is dominated by the dark halo .
the dark halo mass distribution in ngc 4472 can be approximated with an nfw halo profile ( navarro , frenk & white 1996 ) of virial mass @xmath49 @xmath6 although the observed halo is somewhat less centrally peaked than nfw ( brighenti & mathews 1999a ) .
within @xmath3 the contribution of the dark nfw halo mass in the model is small ; for example at @xmath50 the total mass to light ratio is @xmath51 , only 10 percent greater than the dynamic value 9.2 determined in @xmath52 .
it is remarkable that the total mass found from the x - ray data @xmath20 is nearly identical to the expected dynamical mass @xmath19 ( based on stellar velocities and @xmath53 ) in the range @xmath21 ( figure 1 ) .
an almost identical agreement in this radius range is indicated by x - ray observations of another bright virgo elliptical , ngc 4649 ( brighenti & mathews 1997a ) . in this important region
the hot gas is confined by the _ stellar _ potential .
the excellent agreement of the stellar and x - ray masses supports the consistency of two radically different mass determinations : from stellar velocities and from the radial equilibrium of hot interstellar gas .
the apparent agreement of the x - ray and stellar masses in this range of galactic radii also indicates that the hydrostatic support of the hot gas is not strongly influenced by local magnetic fields and rotation .
however , it is not obvious why @xmath35 would be constant with galactic radius , particularly when the cooling dropout mass is considered , and why the agreement between stellar dynamic and x - ray masses no longer obtains in the central regions @xmath54 . in this central region
the total mass indicated by the x - ray observations in figure 1 is considerably less than the expected mass based on an assumed de vaucouleurs profile and constant mass to light ratio .
this type of deviation could be due to magnetic or other non - thermal pressure in this region , to rotation , or to local cooling dropout in the hot interstellar gas .
the lower mean temperature in cooling regions lowers the total apparent gas temperature and results in an underestimate of the total internal mass ( equation 1 ) .
we assume that currently available _ einstein _ and rosat observations are accurate in the central region of ngc 4472 , @xmath54 .
these observations have been reduced assuming no ( non - galactic ) photoelectric absorption by low temperature gas in the central regions .
if x - ray absorption is present , the true hot gas density would be more centrally peaked than shown in figure 1 and the total mass indicated by the x - ray observations would increase .
buote ( 1999 ) finds that two - temperature models fit the x - ray spectrum for ngc 4472 quite well .
the two temperatures do not necessarily need to be spatially mixed ; they could also approximately represent the range of the radial temperature variation observed in ngc 4472 . in buote s two temperature model ,
only the cooler component ( @xmath55 kev located in @xmath56 ) requires an absorption column @xmath57 @xmath58 in excess of the galactic value .
however , the influence of cold gas having this column density on the hot gas density plotted in figure 1 is small . for a worst case example
, suppose that absorbing material with column density @xmath57 @xmath58 is in a disk oriented perpendicular to the line of sight and that this disk absorbs _ all _ x - rays from the back side of the galaxy .
the radius of this disk would be @xmath59 pc if it contained the maximum mass @xmath60 @xmath6 allowed by co and hi observations ( bregman , roberts & giovanelli 1988 ; braine , henkel & wiklind 1997 ) or 20 kpc if it contained all of the gas that has cooled , @xmath61 @xmath6 .
the x - ray surface brightness within the opaque disk would be reduced by 2 , but the corresponding gas density would be lowered by only @xmath62 since the volume emissivity @xmath63 .
such a small correction in the density ( gradient ) could not account for the large mass discrepancy between the x - ray mass and the stellar dynamical mass shown in figure 1 within 0.1@xmath3 .
the densities and temperatures in figure 1 were determined from x - ray data assuming the abundance of the hot gas is uniformly solar . since the gas is likely to be more metal rich at smaller galactic radii ( matsushita 1997 ; brighenti & mathews 1999b ) , an allowance for this gradient would tend to lower the derived density gradient and the internal mass , _ increasing _ the discrepancy in @xmath64 in figure 1 by a small amount . in the following discussion we shall ignore the relatively small possible influence of absorption or metallicity gradients on the results shown in figure 1
the hydrodynamical models we use in this paper are similar to those in our recent papers ( e.g. brighenti & mathews 1999a ) so we provide only a brief review here .
hot interstellar gas in ellipticals has a dual origin : ( i ) mass loss from an evolving old stellar population and ( ii ) secondary infall into the overdensity perturbation that formed the galaxy group within which the elliptical formed by early merging events . for a given set of cosmological parameters , dark and baryonic matter flow toward an overdensity region .
the dark matter forms an nfw halo , growing in size with time .
spherical geometry is assumed . within the accretion shock at time @xmath65 gyr ,
when enough baryons have accumulated , we form the current de vaucouleurs stellar profile and release the energy of all type ii supernovae according to a salpeter imf ( slope : @xmath66 , mass limits : @xmath67 and @xmath68 @xmath6 ) .
all stars greater than 8 @xmath6 produce type ii supernovae each of energy @xmath69 ergs .
we assume that a fraction @xmath70 of this energy is delivered to the internal energy of the gas .
we have shown ( brighenti & mathews 1999b ) that such a galaxy formation scheme can work in a variety of cosmologies : flat or low density , with or without a lambda term .
the evolution of gas within the optical effective radius @xmath3 , of most interest here , is insensitive to these cosmological parameters . for simplicity
therefore we assume a simple flat universe with @xmath71 , @xmath72 km s@xmath73 mpc@xmath73 and @xmath74 .
we characterize the dark halo with an nfw profile ( navarro , frenk & white 1996 ) having a virial mass @xmath49 @xmath6 at the current time @xmath75 gyrs .
the models we discuss here are identical to the standard model of ( brighenti & mathews 1999a ) except we now use a finer central spatial zoning ( 65 pc for the innermost zone ) , a snii efficiency @xmath70 , and a mass `` dropout '' function @xmath76 with more adjustable parameters ( see below ) .
our objective is to seek solutions of the gasdynamical equations including mass dropout that jointly satisfy several observational constraints at time @xmath77 : ( i ) the observed hot gas density , temperature and x - ray surface brightness profiles , ( ii ) the known dynamical mass in the galactic center usually attributed to a massive black hole , ( iii ) the apparent dynamical mass to light ratio @xmath34 determined in @xmath52 , and ( iv ) an apparent internal mass @xmath20 in @xmath78 based on equation ( 1 ) that agrees with the constant-@xmath79 de vaucouleurs profile as shown in figure 1 .
the baryonic component in our models has a complex evolution .
much of the initial baryonic mass is consumed in creating the stellar system .
when the type ii supernova energy is released , a significant mass of gas is expelled as a galactic wind .
after these early events , the interstellar gas is re - established and sustained by stellar mass loss and by inflow of circumgalactic gas ( secondary infall ) , most of which was previously enriched and expelled by snii .
we assume that the stars form during a short epoch that can be described by a single burst salpeter imf as discussed above .
the stellar mass loss rate for this imf varies as @xmath80 where @xmath81^{-1.3}$ ] sec@xmath73 .
although galactic stars form at @xmath82 gyr , their mass loss contribution to the ambient interstellar gas is assumed to begin at a later time , @xmath65 gyrs . since
galactic stars have been enriched by supernova ejecta , the single burst model can not be strictly correct , but our approximation @xmath83 is consistent with several early starbursts closely spaced in time and allows for metal enrichment of old galactic stars that are not in the first single - burst population . if the de vaucouleurs profile is a result of largely dissipationless merging , some or most of the star formation must have occurred at a time @xmath84 before the important merging events at @xmath65 gyrs . by taking @xmath85 we reduce by @xmath86 @xmath6 the total amount of gas ejected by stars within the galactic potential .
we recognize the inconsistencies in these approximations of complex stellar formation and dynamical processes that are poorly understood .
however , once the galaxy is formed , we follow the interstellar gas dynamical evolution in full detail , conserving mass and energy . continued heating by type ia supernova
is assumed to vary inversely with time , snu@xmath87 snu@xmath88@xmath89 , where the current rate , snu@xmath90 snia per 100 yrs per @xmath91 @xmath30 , is near the lower limit of observed values , snu@xmath92 ( cappellaro et al . 1997 ) , as required to maintain the low interstellar iron abundance . for the models discussed here the equation of continuity includes a `` mass dropout '' term : @xmath93 where @xmath94 is the time for gas to cool locally by radiative losses at constant pressure ( see , e. g. sarazin & ashe 1989 ) .
the cooling is assumed to be instantaneous without advection in the cooling flow , i.e. @xmath95 , although in practice this inequality may not always be satisfied . while this type of cooling dropout has been widely used in past models ,
there is no adequate physical model for mass dropout .
clearly , the gas must cool somewhere the emission of x - rays indicates a large net energy ( and mass ) loss from the interstellar medium .
when small regions of low entropy cool , the pressure remains constant since the sound crossing time is much less than the flow time .
following fabian , nulsen & canizares ( 1982 ) and ferland , fabian , & johnstone ( 1994 ) , we assume that cooled gas converts to a second population of optically dark , low mass stars . regarding h@xmath18 emission as a tracer for the cooling gas , mathews & brighenti ( 1999 ) have shown that the cooling occurs at a multitude ( @xmath96 ) of cooling sites distributed throughout the inner galaxy and that only stars with mass @xmath97 can form at each cooling site .
this is supported by the observed absence of young massive stars and snii in elliptical galaxies .
nevertheless , in the following discussion we assume that the maximum stellar mass in the dropout population is sufficiently low that the optical light from these stars is unobservable .
we expect cooling and low mass star formation to be concentrated within at least 2 kpc , which is the observed extent of h@xmath18 emission in ngc 4472 ( macchetto et al .
while the region containing optical emission lines provides a natural guideline for selecting the dropout profile , we consider a wider range of constant or variable dropout coefficients @xmath76 parameterized by @xmath98 which concentrates the cooling within radius @xmath99 .
note that even when @xmath100 is constant , the mass dropout term is spatially concentrated , @xmath101 .
in a series of recent papers in this journal we have presented evolutionary cooling flow models for ngc 4472 that agree quite well with the observed distributions of interstellar gas density , temperature and metallicity ( brighenti & mathews 1999a ; 1999b ) , particularly at intermediate and large radii . since the total mass of cooled gas in these calculations was only a few percent of the stellar mass , the gravitational contribution of cooled gas was ignored , even in models including mass dropout . for a better understanding of the inner galaxy
, we now include the gravity of all gas , both hot and cold , except when specifically noted . in this section
we begin with models having @xmath102 everywhere , so that cooling to low temperatures occurs in a small central region , then we investigate models in which @xmath100 is constant throughout the cooling flow .
we begin by considering a perfectly homogeneous interstellar flow in which all the gas reaches the central computational zone ( @xmath104 pc ) or its neighboring zones where it then cools to @xmath105 k. for comparison we discuss two cases : in model 1 we ignore the gravity of this cooled gas ; in model 2 ( and subsequent models ) we include its gravitational influence on the flow .
the mass of gas that cools into the central gridzone is dynamically equivalent to a massive black hole , but we do not necessarily regard our calculation as a realistic model for black hole formation . in figure 2a
we illustrate the radial variation of gas density and temperature for models 1 and 2 after the interstellar gas has evolved to @xmath106 gyrs .
several global parameters for these and subsequent models are listed in table 1 .
when the gravity of the cooled gas is considered ( model 2 ) , the interstellar gas within a few kpc of the galactic center is compressed and sharply heated in the local potential .
such hot central thermal cores are not generally observed ( but see colbert & mushotzky 1998 ) .
the central mass in both models , @xmath107 @xmath6 is about 16 times larger than the black hole mass found in ngc 4472 , @xmath38 @xmath6 ( magorrian et al .
for these reasons neither model 1 nor 2 provides a realistic interstellar mass distribution for this galaxy .
rosat band x - ray surface brightness profiles @xmath108 for models 1 and 2 are shown in figure 2b and the total rosat band luminosities are listed in table 1 . for model 1
the x - ray brightness peaks strongly in the galactic center , diverging from the observations within about 3 kpc ; it was just this sort of disagreement that initially led to the assumption of distributed mass dropout in cooling flows ( thomas 1986 ; vedder , trester , & canizares 1988 ; sarazin & ashe 1989 ) .
if the efficiency of heating by early snii were lowered , models 1 and 2 could be made to agree better with @xmath109 observations in @xmath110 kpc , but the computed @xmath109 would rise even further above the observations within a few kpc .
the x - ray luminosities @xmath7 for models 1 and 2 listed in table 1 are unreliable in part because of numerical inaccuracies caused by the extremely steep variation of gas temperature and density in the central 2 or 3 computational zones .
this numerical difficulty is common to all cooling flows that proceed to the very center of the galaxy before cooling , for any reasonable central grid spacing . since more @xmath9 work is done on the flow in model 2 ( where the gravity of cooled gas is included )
, we expect that @xmath7 should also be larger than for model 1
. the opposite sense of the change in @xmath7 shown in table 1 evidently results from numerical inaccuracies near the central singularity where zone to zone variations are no longer linear .
if the gas has not cooled before flowing into the core ( @xmath111 pc ) , as in models 1 and 2 , a significant fraction of the total x - ray emission should come from this very central region , again in disagreement with observations . in models 1 and 2 a sphere of cold ( @xmath112 k ) , dense gas accumulates in the central gridzones and grows in mass and size over the hubble time .
this sphere is an unrealistic artifact of our computational assumptions .
therefore , to explore further the central numerical difficulty with @xmath7 encountered in models 1 and 2 , we considered two additional models using higher spatial resolution ( radius of central zone is only 15 pc ) , both of which include the gravitational influence of cooled gas .
the first model ( model 2.1 ) is similar to model 2 with @xmath102 in all zones . in the second model ( model 2.2 )
we set @xmath102 in all zones except the central zone where @xmath113 .
model 2.2 is the appropriate limit of the series of models discussed below in which the cooling dropout is more and more centrally concentrated .
while the flow in @xmath114 kpc is very similar in models 2.1 and 2.2 , the behaviour at smaller radii is quite different .
for example the flow parameters for model 2.1 at 100 pc and @xmath75 gyrs are : @xmath115 k , @xmath116 @xmath117 , and @xmath118 km s@xmath73 . at the same radius and time for model 2.2 the flow is quite different : @xmath119 k , @xmath120 @xmath117 , and @xmath121 km s@xmath73 . at projected radius @xmath122 pc , the rosat x - ray surface brightness in model 2.2 is 20 times larger than model 2.1 and the total rosat x - ray luminosity integrated over the entire cooling flow in model 2.2 is larger by a factor 32 .
the differences between these two models result from upstream propagation of information about the central boundary conditions made possible by subsonic flow near the origin .
fortunately , these numerical and physical difficulties near the origin do not arise in more realistic cooling flows discussed below in which @xmath123 at larger galactic radii . while solutions of the gasdynamical equations provide the temperature as a function of physical radius @xmath39
, the observed temperature @xmath124 is an emission - weighted mean temperature along the line of sight at projected radius @xmath125 .
for symmetric galaxies and at small galactic radii , these two temperatures are nearly identical because of the steep radial variation in x - ray emissivity .
the variation of temperature with physical radius @xmath126 in the background flow is shown with light solid lines in figure 2a . in the presence of spatially distributed cooling dropout
, the local temperature is an emission - weighted mean of the background ( uncooled ) gas and the cooling regions .
cooling is assumed to occur at a large number of cooling sites where the gas remains in pressure equilibrium as it cools , apparently unrestricted by magnetic stresses ( see mathews & brighenti 1999 for details ) .
the heavy solid lines in figure 2a show the mean apparent temperature @xmath127 including contributions from locally cooling regions , @xmath128 where @xmath16 is the background flow temperature and the slowly - varying functions @xmath129 are plotted in brighenti & mathews ( 1998 ) .
note that the effective temperature is independent of the local gas density .
the temperature that is actually observed is an average of @xmath130 along the line of sight ; this temperature @xmath131 is shown with dotted lines in figure 2a .
lacking an acceptable physical model for spatially distributed mass dropout in galactic cooling flows , we are at liberty to choose any variation for the dropout coefficient @xmath76 , appropriately constrained _ a posteriori _ by the known dynamical mass from stellar velocities , the x - ray mass , the central black hole mass and the observed radial variation of hot gas density , temperature and x - ray surface brightness .
it is natural to begin with constant @xmath100 solutions similar to those considered by sarazin & ashe ( 1989 ) in their steady state cooling flow solutions . for simplicity
we assume that the mass of cooled gas remains at the cooling site where it contributes to the gravitational potential . in the central panels of figure 2a we illustrate the hot interstellar gas density and temperature that results after @xmath75 gyrs assuming uniform @xmath137 and 4 ;
these are listed in table 1 as models 3 and 4 respectively .
when @xmath100 is a constant independent of radius , the ratios of true to apparent values
@xmath138 , @xmath139 and @xmath140 are also approximately uniform with galactic radius . here
@xmath141 is the true total mass within @xmath41 and @xmath20 is the value that would be determined from x - ray observations of the models by assuming hydrostatic equilibrium ( as in fig .
the influence of constant-@xmath100 mass dropout is similar at all galactic radii since the factors that convert background temperature and density to effective values in equations ( 3 ) and ( 4 ) depend only on @xmath39 which is slowly varying , not on @xmath40 . because of the enormous volume and time available to gas dropping out at large @xmath41 , the total mass of cooled gas ( listed in table 1 ) becomes very large in the outer galaxy . in figure 3
we compare the radial distribution of stellar and dark halo mass with the distributed dropout mass that results after 13 gyrs with @xmath137 and 4 . the de vaucouleurs `` stellar '' mass profile in figure 3
is constructed assuming uniform @xmath34 . as @xmath100 increases from 1 to 4 , the background hot gas density ( light solid lines in fig .
2a ) decreases and its radial gradient flattens , causing a rise in temperature to provide enough pressure to support the cooling flow atmosphere .
note that the total dropped - out mass within @xmath142 decreases with larger @xmath100 ( figure 3 and table 1 ) .
although more overall cooling dropout occurs as @xmath100 increases , most of this cooling occurs at very large @xmath41 and , somewhat paradoxically , less gas remains for dropout closer to the galactic center , @xmath56 .
however , the effective ( i.e. apparent ) density ( heavy solid lines in fig .
2a ) is less sensitive to @xmath100 .
the @xmath137 solution ( model 3 ) is preferred for its fit to observed temperatures while the @xmath113 solution ( model 4 ) agrees better with the observed density in @xmath143 kpc and almost exactly with the x - ray surface brightness ( fig .
in addition , the centrally concentrated dropped out mass in the @xmath137 solution ( model 3 ) contributes a larger fraction of the dynamical mass in @xmath144 and the total mass within the central gridzone ( @xmath104 pc ) is almost equal to the known mass of the black hole in ngc 4472 ( table 1 ) .
evidently , cooled masses in models with @xmath145 would exceed the central dark mass observed . for both values of @xmath100 considered ,
the dropped out mass listed in table 1 contributes appreciably to the total mass within @xmath146 . when both old stellar and dropout mass are included , the @xmath35 values at @xmath142 shown in table 1 12.19 and 11.21 exceed both the dynamical value @xmath34 found by van der marel ( 1991 ) within @xmath147 and the value @xmath148 in our models which includes a small additional contribution of non - baryonic dark matter .
to quantify the influence of the dark baryonic dropout mass on the mass to light ratio we include in table 1 an entry for @xmath149 @xmath150 where @xmath151 @xmath6 is the combined mass of luminous stars and dark halo matter at @xmath142 and @xmath152 also includes the dropout mass within this radius .
since our computed total mass @xmath153 exceeds the observed dynamical value 9.2 ( and also 10.18 ) , to be fully consistent we should have chosen @xmath154 for the old stellar population ( see below ) , provided @xmath35 for stars produced from the cooled gas is infinite as we have assumed .
if these constant @xmath100 models are physically appropriate , the agreement between @xmath20 and @xmath19 in @xmath78 , as shown in figure 1 ( and for ngc 4649 ) , is surprising since @xmath155 and @xmath156 differ by 10 - 40 percent ( table 1 ) for the constant @xmath100 models .
this suggests that the mass to light ratio of the dropout stellar population is not infinite as we have assumed , but comparable with that of the old stellar population .
we now seek evolutionary gasdynamical solutions with variable dropout coefficients @xmath76 that strongly concentrate the mass dropout in the inner galaxy , @xmath56 .
the limited spatial extent of optical emission lines in the cores of bright ellipticals suggests that cooling dropout occurs well within @xmath3 .
for example , the h@xmath18 + [ nii ] image of ngc 4472 is observed out to approximately @xmath157 or 2 kpc ( macchetto et al .
as hot interstellar gas cools in this region , its temperature pauses at @xmath158 k where the gas is heated and ionized by stellar uv radiation .
therefore , we consider parameters @xmath99 and @xmath159 in equation ( 2 ) that concentrate the mass dropout in the central region , but not at the very center as in model 2 . the dropout parameters @xmath160 , @xmath99 and @xmath159 , are listed in table 1 for models 5 , 6 and 7 .
the mass dropout in these three models is progressively more concentrated toward the galactic center . the results for model 5 , for which the dropout scale length is very large , @xmath161 kpc , are similar in most respects to those of model 4 except the massive dropout in the outer galaxy is no longer present . in particular
, the mass to light ratio at @xmath142 in table 1 is very similar for models 4 and 5 . the current interstellar density and temperature variations for model 6 with @xmath162 kpc
are shown in figure 2a and @xmath108 is plotted in figure 2b .
the apparent density and @xmath109 ( heavy solid lines ) are considerably greater than observed values in @xmath163 kpc , although the projected apparent temperature is a reasonable fit to the ngc 4472 data .
the radial distribution of dropout mass for model 6 is shown in figure 3 . from table 1
the total ( old stars , halo and dropout ) mass to light ratio at @xmath142 is @xmath164 , 29 percent higher than van der marel s value and 22 percent greater than the corresponding value for our background model galaxy . also shown in figures 2a and 2b are similar results for model 7 in which the mass dropout , now approximated with a gaussian , is concentrated within @xmath165 pc .
model 7 is constructed so that most of the mass dropout occurs in @xmath166 , corresponding to the region of apparent disagreement between @xmath167 and @xmath168 in figure 1 .
it is of interest that the gas density and @xmath109 within 1 kpc for model 7 , where the dropout is greatest , exceeds those of model 2 in which there is no dropout at all .
this can be understood from the flow velocity distribution .
for model 2 without distributed dropout the inward moving gas velocity increases through the entire region illustrated and shocks at @xmath169 kpc . in the presence of dropout ,
the flow velocity at corresponding radii in model 7 is much slower and reaches a maximum ( negative ) value near
@xmath170 kpc then approaches zero subsonically at the origin .
the density and @xmath109 enhancements in the background flow for model 7 at @xmath163 kpc are due to a local compression as the gas flows into slowly moving gas in the core . within about 3 kpc , both
the background and apparent densities exceed the observations by a larger factor than those of model 6 and @xmath109 also peaks unrealistically in this same region ( fig .
the dropout mass for model 7 shown in figure 3 equals that of the old population stellar mass at @xmath171 kpc ( @xmath172 ) . for these reasons model 7 seems less satisfactory than model 6 , but neither is as generally successful as model 3 ( @xmath137 ) .
although increased mass dropout can decrease the x - ray surface brightness @xmath109 , as when uniform @xmath100 increased from 1 to 4 ( models 3 and 4 in fig .
2b ) , this is not always the case .
when the dropout is concentrated more toward the galactic center , @xmath109 actually increases , as in the transition from model 6 to 7 .
it is interesting to determine the influence of distributed dropout on the total apparent mass @xmath20 found from the model by assuming hydrostatic equilibrium ( equation 1 ) . due to the contribution of low temperature cooling regions to the total x - ray emission , @xmath39 and
therefore @xmath20 is always lower than the true mass @xmath173 in cooling dropout regions .
a difference in this sense is apparent in table 1 at @xmath146 for models 3 - 7 ; this is similar to the mass discrepancy in figure 1 at @xmath54 .
the high central apparent gas density and surface brightness for model 6 shown in figures 2a and 2b are obvious problems for this model .
however , the gas density can be reduced if a strong magnetic field or other non - thermal energy density is present in @xmath174 kpc .
dynamically important magnetic fields may also be required to fit the x - ray data of ngc 4636 ( brighenti & mathews 1997a ) and may be generally expected in luminous ellipticals ( mathews & brighenti 1997 ; godon , soker & white 1998 ) .
additional non - thermal pressure support is also implied for model 6 since the effect of central cooling dropout fails to lower the apparent mass @xmath20 below the actual mass @xmath175 as much as the observed deviation shown in figure 1 .
like many bright ellipticals , ngc 4472 has a weak non - thermal radio source within the central @xmath176 kpc , indicating @xmath177 @xmath178 g ( ekers & kolanyi 1978 ) . for all calculated models in table 1 with distributed dropout ,
the total mass @xmath179 significantly exceeds the mass of the old stellar population plus dark halo @xmath180 throughout the inner galaxy , indicating that dropout material makes an important additional contribution to the total mass .
however , the apparent mass @xmath167 determined from equation ( 1 ) is less than @xmath175 in the inner galaxy and , for model 6 , can also be less than @xmath181 .
of particular interest is the region @xmath182
( i.e. , @xmath183 ) in figure 1 .
although the agreement in figure 1 is excellent in this region , for models 6 and 7 the total mass is larger than @xmath173 and values of @xmath184 in table 1 suggest that the dropped out mass contributes 25 - 35 percent of the total mass in this region .
therefore , if the dropout mass is optically dark , the true stellar mass to light ratio of the old stellar population must be @xmath185 rather than the value @xmath34 found by van der marel which includes the dropout mass . to investigate such a possibly more self - consistent old stellar component
, we consider model 8 based on the same @xmath76 used in model 6 , but with @xmath186 for the old stellar population .
for additional consistency in model 8 , @xmath187 is increased by the ratio of assumed stellar mass to light ratios 9.2/6 = 1.53 as described below ; the total stellar mass ejected is identical to that in model 6 .
figure 3 illustrates the dropout mass profile for model 8 .
the apparent density , temperature and @xmath109 profiles for model 8 shown in figures 4a and 4b are very similar to those of model 6 , so this adjustment of the stellar @xmath35 has had little effect .
the radial mass profiles of models 6 and 8 are compared in figure 5 . in this plot
the open circles show the observed x - ray mass @xmath20 for ngc 4472 using equation ( 1 ) and the solid lines shows @xmath20 based on equation ( 1 ) using @xmath40 and @xmath39 from the models .
the dashed lines show @xmath141 and the dotted lines are the stellar mass @xmath19 based on a de vaucouleurs profile with @xmath34 in the upper panel ( model 6 ) and @xmath186 in the lower panel ( model 8) .
the superiority of model 8 is evident from the closer agreement between the x - ray mass data points for ngc 4472 and the solid line for that model .
this agreement for model 8 would be even closer in the range @xmath188 if we had used a dark halo mass profile less centrally peaked than nfw .
model 8 may provide the most self - consistent overall fit to the mass constraints for ngc 4472 ; if so , the old stars in ngc 4472 have a mass to light ratio @xmath185 , about @xmath189 percent lower than the mass to light ratio determined from stellar dynamics . in summary ,
none of our models is fully satisfactory in every respect . the total mass to light ratio ( including dropout mass ) in @xmath144 is 10 - 35 percent higher than the value for the underlying galaxy .
however , in model 8 in which @xmath186 for the old stars , the difference between the x - ray mass determined from the models and ngc 4472 is appreciably reduced .
nevertheless , the central apparent gas density and x - ray surface brightness in model 8 are still larger than observed , requiring additional non - thermal support .
there is no independent theoretical justification for the dropout profile @xmath76 assumed in models 6 and 8 ; as explained earlier , the dropout distribution depends on unknown interstellar entropy fluctuations .
nevertheless , for all models considered here the mass of cooled interstellar gas contributes significantly to the total mass and the dynamically determined mass to light ratio within the inner galaxy .
in this series of calculations we have taken a census of all baryons involved in the evolution of a large elliptical galaxy : the original stellar component , the interstellar medium , and of most interest the small but troublesome mass of hot gas that cools over cosmic time .
we have shown that the radial distribution of cooled interstellar gas influences dynamical and x - ray determinations of the total interior mass and the radial profiles of apparent density , temperature and x - ray brightness of the hot gas . in our models ,
cooled gas is slowly deposited in the central galaxy @xmath56 ( see figure 3 ) as indicated by the extent of observed h@xmath18 emission . since there is little or no direct observational evidence for the mass that has dropped out in ellipticals like ngc 4472 , the cooled mass must either be dark at optical and radio frequencies or indistinguishable from the old stars .
low mass stars are an obvious and physically reasonable endstate for the cooled gas ( ferland , fabian & johnstone 1994 ; mathews & brighenti 1999 ) . we also suppose that the cooled gas remains at the dropout site where it contributes to the galactic potential . in an attempt to reduce the influence of cooling and cooled gas on the models , we have altered many of the model parameters .
the chosen cosmology ( @xmath71 ; @xmath190 and @xmath191 , etc . ) or the baryon fraction @xmath192 have little influence on the total dropout mass . changing the times when the stars and the galactic potential form ( @xmath84 and @xmath193 ) or the spatial scale of the release of snii energy within reasonable limits have only a modest influence on the dropout mass that accumulates by time @xmath75 gyrs . increasing the interval @xmath194 between star and galaxy formation
does reduce the total dropout mass , but this interval can not be too large since luminous ellipticals are observed at large redshifts . the cooling dropout would not be dramatically reduced if massive ellipticals were all only a few gyrs old since most of the stellar mass loss occurs just after star formation ; however , many or most luminous ellipticals are thought to be very old . perhaps the most effective way to preserve the excellent agreement between @xmath181 and @xmath167 in figure 1 in @xmath195 , without constraining the mass dropout profile @xmath76 , would be to reduce the total mass that has cooled over cosmic time .
the most sensitive parameter influencing the cooled mass is the specific rate of stellar mass loss , @xmath187 .
although the total mass of hot gas increases with time due to the continued influx of secondary infalling gas , the mass of hot gas within the optical galaxy originates mostly from stellar mass loss and the x - ray luminosity there scales as @xmath196 ( appendix b of tsai & mathews 1995 ) . computed models similar to those described here but with arbitrarily reduced @xmath187 fit the @xmath40 and @xmath108 data rather well at time @xmath77 and produce much less cooling dropout
however , @xmath187 can not be lowered without also increasing the stellar mass .
for all reasonable power law initial mass functions , @xmath187 varies inversely with the stellar mass to light ratio : @xmath197 { 1 \over ( m_*/l_b ) } \propto { 1 \over ( m_*/l_b)}.\ ] ] the physical explanation for the constancy of @xmath198 is that both @xmath8 ( dominated by post - main sequence stars ) and @xmath199 depend on the instantaneous rate that stars leave the main sequence so this imf - dependent factor cancels out ( e.g. renzini & buzzoni 1986 ) . to illustrate this result
, we use the renzini - buzzoni procedure and plot in figure 6 the relationship between @xmath200 and @xmath35 at time @xmath75 gyrs for 84 power law imfs with slopes @xmath2010.6 , 0.8 , 1.0 , 1.2 , 1.4 , 1.6 , and 1.8 , each with lower and upper mass limits of @xmath2020.01 , 0.032 , 0.1 , 0.32 and @xmath20310 , 32 , 100 .
the remarkable linearity in figure 6 shows that the quantity in square brackets in the equation above is almost invariant to large changes in the slope or mass limits for power law imfs .
therefore , if @xmath187 is reduced by 2 or 3 in an attempt to reduce the total dropout mass , the stellar mass ( and @xmath35 ) must be increased by the same factor ; as a result the total mass ejected , @xmath204 , and the total dropout mass do not change .
although massive ellipticals are not rotationally flattened , they do rotate significantly , e.g. @xmath205 for ngc 4472 ( faber et al .
if the hot interstellar gas rotates in the same sense as the bulk of mass - losing stars , stars formed from the cooled gas should form into a disk of scale @xmath206 ( brighenti & mathews 1997b ) , although the development of such disks is likely to be suppressed by the mass dropout process . nevertheless , to the extent that the cooled gas has a disk - like distribution , its global influence on the stellar dynamics in @xmath52 would be less than if the same dark mass were distributed spherically .
remarkably , there is no observational evidence at present for rotational flattening in the x - ray images of giant ellipticals like ngc 4472 .
dynamical determinations of the mass to light ratio from stellar velocities reflect the entire mass within the stellar orbits , including non - baryonic mass .
the nfw halo we use in our models for ngc 4472 agrees with x - ray observations in the extended halo but is slightly too massive near @xmath207 relative to the x - ray mass @xmath20 observationally determined for ngc 4472 .
this may indicate that dark halos are less centrally peaked than nfw ( see kravtsov et al .
the mass of our nfw halo model contributes about 10 percent to the dynamical mass to light ratio measured within @xmath142 , but the nfw profile is probably disturbed in this region .
when the dominant baryonic mass in @xmath56 compressed to form the de vaucouleurs profile , we expect that the nfw halo was dragged inward and distorted .
however , the dark halo cores of the earlier galactic condensations that merged to form the elliptical may have expanded due to starburst driven galactic winds . because of these various counteracting effects , the small non - baryonic contribution to dynamical mass to light determinations is uncertain . for all models studied here based on a wide variety of mass dropout profiles @xmath76
the mass of cooled interstellar gas contributes substantially to the total mass within @xmath147 where the stellar mass to light ratio is determined from stellar velocities . if optically dark low mass stars form from the cooled gas , the `` stellar '' mass to light ratios in the literature refer to two distinct stellar populations having radically different initial mass functions and spatial distributions . the stellar mass to light ratio of the original ,
optically luminous stellar population is lower than published values indicate .
a complete solution of this problem requires a better understanding of the physics of star formation and the processes that control the stellar imf .
we have assumed here that a salpeter imf provides a satisfactory approximation to the original single - burst star formation at early times .
yet we argue that the younger dropout imf is strongly skewed toward low mass stars .
the imf has evolved over time .
it is possible therefore that the early , more intense mass dropout resulted in a more nearly salpeter - like imf , producing a fraction of currently observed luminous stars in ellipticals .
if so , this early dropout would not contribute to the excess dropout mass that we find in our models , but would have already been included in the original de vaucouleurs population .
high density , metal enriched stellar cores in elliptical galaxies may have derived from normal - imf star formation from early , more intense interstellar cooling .
this is similar to assumptions made for dissipative galactic core formation from the convergence of gas following major mergers ( mihos & hernquist 1996 ) .
while such notions can not be entirely dismissed , the approximate universality of the de vaucouleurs light profile among ellipticals may argue against a dual formation process for the radial distribution of luminous stars : violent relaxation and cooling flow dropout . throughout this discussion
we have assumed that the stellar population formed from cooling flow dropout is optically dark .
although there is little or no evidence that normal massive ob stars ( or snii ) are present in elliptical galaxies , it is possible that younger stars having masses up to @xmath26 @xmath6 are present and that such intermediate mass stars could form from the cooled gas ( mathews & brighenti 1999 ) .
this type of dropout stellar population could contribute to the total optical light .
if the mass to light ratio of dropout and old stellar populations are similar , the dropout component could be difficult to detect by the means we have discussed here and its perturbation on the observed @xmath35 would be greatly lessened . in this case
the dropout population would introduce an additional radial light profile that would differ slightly from that of the old stellar population .
the dropout mass profiles in figure 3 indicate that models 4 and 6 would be rather difficult to detect against the background stellar light .
intermediate mass dropout stars could therefore provide a satisfactory resolution to the problems we have discussed here .
the ensemble of elliptical galaxies is known to have global parameters that deviate slightly from the assumptions of virial equilibrium and homologous structure .
the deviation of this fundamental plane relationship is in the sense that the dynamical mass to light ratio increases with galactic mass , @xmath208 ( dressler et al .
1987 ; djorgovski & davis 1987 ) . such a non - homologous deviation could in principle be produced by the small amount of cooled interstellar gas @xmath23 provided it increases appropriately with @xmath168 .
to test this idea , we performed an identical hydrodynamic calculation for an elliptical galaxy having a mass one fourth that of ngc 4472 .
the dark halo mass was also reduced by the same factor but the cosmological environment and mass dropout distribution were identical to those used for ngc 4472 , scaled to a smaller @xmath3 .
we found that the amount of mass dropout @xmath23 is _ higher _ in smaller ellipticals relative to the total baryonic mass @xmath168 .
this is opposite to the trend observed in the fundamental plane .
however , if hot gas and dark matter in the outer halos of smaller ellipticals is tidally stripped in group environments , as suggested by mathews & brighenti ( 1998b ) , then the mass of cooled gas would be reduced and its effect on the fundamental plane would be reduced .
nevertheless , explanations of the deviations of the fundamental plane from virial scaling must recognize the possible additional influence of dropout mass , regardless of the trend of @xmath209 with @xmath168 .
\(1 ) if the hot interstellar gas cools only in the very center of ngc 4472 for @xmath210 gyrs , the total accumulated mass there would be @xmath211 times larger than the mass of the central black hole observed .
if such large concentrated masses were generally present in bright ellipticals , interstellar gas in @xmath166 would be compressed and heated to temperatures @xmath212 kev .
such hot thermal cores are not generally observed .
we conclude that the cooling dropout in massive ellipticals must occur before the gas reaches the galactic center .
the hypotheses of distributed mass dropout and low mass star formation were proposed many years ago ( fabian , nulsen & canizares 1982 ; thomas 1986 ; cowie & binney 1988 ; vedder , trester , & canizares 1988 ; sarazin & ashe 1989 ; ferland , fabian , & johnstone 1994 ) . however , these historical arguments were generally based on the notion that mass dropout would help reduce computed x - ray surface brightness profiles at small projected radii , as required by the observations , but we have shown here that in some cases enhanced dropout at small galactic radii can cause @xmath109 to increase , not decrease .
( the total bolometric x - ray luminosity @xmath7 should always be lower in distributed cooling models since the hot gas experiences only a fraction of the galactic potential . )
the best arguments for distributed cooling dropout are ( i ) limits on the central black hole mass and ( ii ) the absence of rotational flattening in x - ray images .
\(2 ) we have considered a wide variety of possible mass profiles for the radial deposition of cooled interstellar gas in ngc 4472 .
the dropout mass is assumed to be optically dark , consistent with the formation of very low mass stars . in every case
the ( stellar plus dropout ) mass to light ratio at @xmath213 significantly exceeds the mass to light ratio determined from stellar velocities .
if the dropout mass is optically dark , dynamical mass to light ratios in luminous ellipticals should be substantially enhanced by dark baryonic matter . in this case the true @xmath35 for luminous stars may be @xmath189 percent smaller than published values .
\(3 ) the excellent agreement between the x - ray and `` stellar '' mass in ngc 4472 shown in figure 1 ( and also ngc 4649 ) in the range @xmath195 may be a coincidence if our estimates of the cooling flow dropout mass are correct and if this mass is non - luminous .
\(4 ) dynamical mass to light determinations within @xmath3 refer to a superposition of two stellar populations : an old luminous population with a de vaucouleurs galactic profile and a younger population having a bottom - heavy imf and a different galactic mass profile .
if the younger population is optically dark , the mass to light ratio is not likely to be constant with galactic radius within @xmath3 .
\(5 ) it is not possible to reduce the total amount of mass deposited from the cooling flow simply by lowering the specific rate of stellar mass loss @xmath214 . for all reasonable power law
initial mass functions we show that @xmath215 . for given @xmath8 , larger stellar masses @xmath19 must accompany lower values of @xmath200 so the total amount of mass ejected from stars @xmath216 is nearly independent of the imf .
\(6 ) among the models we consider , those with centrally concentrated mass dropout perform best in minimizing the overall disagreement with the central @xmath35 and the x - ray determined mass in @xmath195 .
constant @xmath100 models in which the dropout is proportional to the local gas emissivity at every radius deposit less mass in @xmath195 , but may have gas temperatures that are too low ( @xmath217 ) or central masses that are too large ( @xmath218 ) .
\(7 ) even in the presence of mass dropout , the computed central interstellar gas densities and x - ray surface brightnesses @xmath108 are generally too large .
such deviations would be reduced if the hot gas is partially supported in @xmath166 by magnetic or other non - thermal pressure associated with the extended radio source .
\(8 ) if the stellar population formed from cooled interstellar gas extends to intermediate masses , @xmath26 @xmath6 , its mass to light ratio may blend with that of the older population .
in this case the dropout mass would already be represented in the de vaucouleurs profile representing the stellar mass distribution in our models .
dynamical determinations of @xmath35 would be a weighted mean of the two populations . if the dropout stellar population is luminous , some of the difficulties we have discussed here would be alleviated , but not those regarding @xmath108 .
\(9 ) if the influence of rotation and non - thermal pressure can be understood , high resolution images of the central regions of elliptical galaxies using the _ chandra _ ( axaf ) satellite may detect the presence of dark baryonic dropout material and an accurate determination of the mass to light ratio of the old stellar population .
thanks to karl gebhardt for providing useful information .
studies of the evolution of hot gas in elliptical galaxies at uc santa cruz are supported by nasa grant nag 5 - 3060 and nsf grant ast-9802994 for which we are very grateful .
fb is supported in part by grant murst - cofin 98 .
allen , s. w. , fabian , a. c. , johnstone , r. m. , white , d. a. , daines , s. j. , edge , a. c. & stewart , g. c. 1993 , mnras 262 , 901 allen , s. w. & fabian , a. c. 1997 , mnras 286 , 583 braine , j. & dupraz , c. 1994 , a&a 283 , 407 braine , j. , henkel , c. & wiklind , t. 1997 , a&a 321 , 765 bregman , j. n. , roberts , m. s. & giovanelli , r. 1988 apj 330 , l93 brighenti , f. & mathews , w. g. 1997a , apj 486 , l83 brighenti , f. & mathews , w. g. 1997b , apj 490 , 592 brighenti , f. & mathews , w. g. 1998 apj 495 , 239 brighenti , f. & mathews , w. g. 1999a , apj 512 , 65 brighenti , f. & mathews , w. g. 1999b , apj ( in press ) buote , d. a. 1999 , mnras ( submitted ) cappellaro , e. , turatto , m. , tsvetkov , d. yu . , bartunov , o. s. , pollas , c. , evans , r. , & hamuy , m. 1997 , astron . & ap .
322 , 431 colbert , e. & mushotzky , r. f. 1998 , ( astro - ph/9809150 ) cowie , l. l. & binney , j. 1988 , apj 215 , 723 djorgovski , s. g. & davis , m. 1987 , apj 313 , 59 donahue , m. & vogt , m. 1997 , in _ galactic and cluster cooling flows _ , ed .
n. soker , asp conf .
series 115 , 48 dressler , a. , lynden - bell , d. , burstein , d. , davies , r. l. , faber , s. m. , terlevich , r. j. , & wegner , g. 1987 , apj 313 , 42 ekers , r. & kolanyi , c. 1978 , a&a 67 , 47 faber , s. m. , wegner , g. , burstein , d. , davies , r. l. , dressler , a. , lynden - bell , d. & terlevich , r. j. 1989 , apjs 69 , 763 faber , s. m. , tremaine , s. , ajhar , e. a. , byun , y - i , dressler , a. , gebhardt , k. , grillmair , c. , kormendy , j. , lauer , t. r. , & richstone , d. 1997 , aj 114 , 1771 fabian , a. c. , nulsen , p. e. j. & canizares , c. r. 1982 , mnras 201 , 933 fabian , a. c. , arnaud , k. a. , bautz , m. w. & tawara , y. 1994 , apj 436 , l63 ferland , g. j. , ferland , a. c. & johnstone , r. m. 1994 , mnras 266 , 399 forman , w. , jones , c. , david , l. , franx , m. , makishima , k. , & ohashi , t. 1993 , apj 418 , l55 godon , p. , soker , n. & white iii , r. e. 1998 , aj 116 , 37 irwin , j. a. & sarazin , c. l. 1996 , apj , 471 , 683 kravtsov , a. v. , klypin a. a. , bullock , j. s. & primack , j. r. 1998 , apj 502 , 48 macchetto , f. , pastoriza , m. , caon , n. , sparks , w. b. , giavalisco , m. , bender , r. & capaccioli , m. 1996 , a&as 120 , 463 magorrian , j. , tremaine , s. , richstone , d. , bender , r. , bower , g. , dressler , a. , faber , s. m. , gebhardt , k. , green , r. , grillmair , c. , kormendy , j. , & lauer , t. 1998 , aj 115 , 2285 mathews , w. g. & brighenti , f. 1997 , apj 488 , 595 mathews , w. g. & brighenti , f. 1998a , apj 493 , l9 mathews , w. g. & brighenti , f. 1998b , apj 503 , l15 mathews , w. g. & brighenti , f. 1999 , apj ( submitted ) matsushita , k. 1997 , ph.d .
thesis , university of tokyo mihos , j. c. & hernquist , l. 1996 , apj 464 , 641 navarro , j. f. , frenk , c. s. , & white , s. d. m. 1996 , apj , 462 , 563 renzini , a. & buzzoni , a. 1986 , in _ the spectral evolution of galaxies _ eds . c. chiosi & a. renzini ( reidel : dordrecht ) , p. 195 sarazin , c. l. & ashe , g. a. 1989 , apj 345 , 22 thomas , p. a. 1986 , mnras 220 , 949 trinchieri , g. , fabbiano , g. , & canizares , c. r. 1986 , apj , 310 , 637 trinchieri , g. , fabbiano , g. , & kim , d .- w .
1997 , a&a 318 , 361 tsai , j. c. & mathews , w. g. 1995 , apj 448 , 84 van der marel , r. p. 1991 , mnras , 253 , 710 vedder , p. w. , trester , j. j. , & canizares , c. r. 1988 , apj 332 , 725 white , d. a. , fabian , a. c. , johnstone , r. m. , mushotzky , r. f. , & arnaud , k. a. 1991 , mnras 252 , 72 wise , m. w. & sarazin , c. l. 1999 , apj ( submitted ) crccccccccccc 1 & 0 & ... & ... & @xmath219 & 4.64 & @xmath220 & @xmath221 & @xmath222 & 0.00 & @xmath223 & 28.6 & 28.6 2 & 0 & ... & ... & @xmath224 & 4.66 & @xmath225 & @xmath226 & @xmath227 & 0.34 & @xmath228 & 21.7 & 21.7 3 & 1 & ... & ... & @xmath229 & @xmath230 & @xmath231 & @xmath232 & @xmath233 & 0.19 & @xmath234 & 18.7 & 36.4 4 & 4 & ... & ... & @xmath235 & @xmath236 & @xmath237 & @xmath238 & @xmath239 & 0.10 & @xmath240 & 8.8 & 42.6 5 & 4 & 15 & 1 & @xmath241 & @xmath242 & @xmath243 & @xmath244 & @xmath245 & 0.11 & @xmath240 & 13.9 & 27.7 6 & 4 & 2 & 1 & @xmath241 & @xmath246 & @xmath247 & @xmath248 & @xmath249 & 0.27 & @xmath250 & 22.1 & 33.5 7 & 4 & 0.8 & 2 & @xmath241 & @xmath251 & @xmath252 & @xmath253 & @xmath254 & 0.34 & @xmath255 & 29.1 & 43.3 8 & 4 & 2 & 1 & @xmath256 & @xmath257 & @xmath258 & @xmath259 & @xmath260 & 0.40 & @xmath261 & 20.0 & 28.1 | typically @xmath0 percent of the total baryonic mass in luminous elliptical galaxies is in the form of cooled interstellar gas . although the mass contributed by cooled gas is small relative to the mass of the old stellar system in these galaxies , it is almost certainly concentrated within the optical effective radius where it can influence the local dynamical mass .
however , the mass of cooled gas can not be confined to very small galactic radii ( @xmath1 ) since its mass would greatly exceed that of known central mass concentrations in giant ellipticals , normally attributed to massive black holes .
we explore the proposition that a population of very low mass , optically dark stars is created from the cooled gas . for a wide variety of assumed radial distributions for the interstellar cooling
, we find that the mass of cooled gas contributes significantly ( @xmath230 percent ) to stellar dynamical mass to light ratios which , as a result , are expected to vary with galactic radius . however ,
if the stars formed from cooled interstellar gas are optically luminous , their pertubation on the the mass to light ratio of the old stellar population may be reduced .
cooling mass dropout also perturbs the local apparent x - ray surface brightness distribution , often in a positive sense for centrally concentrated cooling . in general
the computed x - ray surface brightness exceeds observed values within @xmath3 , suggesting the presence of additional support by magnetic stresses or non - thermal pressure .
the mass of cooled gas inside @xmath3 is sensitive to rate that old stars lose mass @xmath4 , but this rate is nearly independent of the initial mass function of the old stellar population .
.2 in |
if the next - generation reactor- and accelerator - based experiments fail to observe any positive signal for non - zero @xmath1 , more powerful experiments will be needed .
in addition to this , even if the forthcoming experiments could measure @xmath1 , it is quite unlikely that they will obtain neither a cp violating signal nor a measurement of @xmath12 .
therefore , a new generation of experiments will be needed also in this case .
two main options have been envisaged for this purpose : the neutrino factory @xcite , and the so - called `` @xmath0-beam '' @xcite . while the former produces the ( anti)neutrino beams through muon decay , the latter entails producing @xmath0-unstable radioactive ions and letting them decay in a storage ring .
we propose a @xmath0-beam set - up where we produce , accelerate and store ions of the four kinds ( @xmath3he , @xmath2ne , @xmath6li and @xmath6b @xmath13 ) at cern , each of them running for a period of 2.5 years .
we aim the @xmath3he - and @xmath2ne -generated low - energy neutrino beams to a megaton water detector located at @xmath5 km from the source , and the @xmath6li - and @xmath6b -generated high - energy neutrino beams to a 50 kton iron detector at a distance close to the magic baseline .
we have also reconsidered the storage ring design and feasibility : as we consider different ions for the two baselines , we have reduced the size of the ring that is aiming at the far detector , increasing at the same time the @xmath14 factor for the @xmath6b and @xmath6li ions in order not to lose sensitivity to the main observables we are interested in .
it is well - known that for a baseline of _ o _ @xmath15 , all the @xmath16-dependent terms in the golden channel probability vanish @xcite .
this provides a clean bedrock to determine @xmath1 , because the intrinsic degeneracy disappears .
in addition to this , the density encountered by the ( anti)neutrinos at this baseline allows for a resonant enhancement in the probability when @xmath17 gev if the mass hierarchy is normal ( inverted ) .
the advantage of this resonance is two - fold : first , it compensates the loss of events due to the very long baseline ; second , it only occurs for ( anti)neutrinos if the mass hierarchy is normal ( inverted ) , therefore providing an extremely good probe of the mass ordering . in order to reach this range of energies ,
@xmath6b and @xmath6li ions are the best candidates because of their high end point energies
. however , @xmath16 can not be measured at the magic baseline , and matter effects can fake true cp violation .
therefore , short baselines are better if we want to probe @xmath16 . in the small matter effect regime
, maximizing the cp - violating terms translates into fixing @xmath18 km / gev .
the mean neutrino energy of neutrinos coming from @xmath3he and @xmath2ne decays at @xmath19 is @xmath20 gev , which translates to an on - peak baseline of @xmath21 km , while for @xmath6b and @xmath6li , due to their much higher end point energies , this baseline would be @xmath22 km .
the main problem in this case is that the flux is proportional to @xmath23 .
therefore , @xmath3he and @xmath2ne turn out to be the better candidates to probe cp violation .
the original design of the storage ring proposed in @xcite must be modified when the boost factor @xmath14 is increased .
if we use lhc dipolar magnets ( with a maximum magnetic field of 8.3 t ) to bend the ions , and keeping the straight sections untouched , the useful fraction of ion decays ( also called livetime " ) for this ring would be ] _
= 0.28 , the total length of the decay ring being @xmath24 m. the tilt angle needed to aim at a detector located at 650 km from the source is @xmath25 : this means that the maximum depth of the far end of the ring is @xmath26 m. however , if a ring of the same type is used to aim at a detector located at @xmath27 km from the source , the tilt angle to be considered is @xmath28 . in this case
, the maximum depth of the far end of the ring is @xmath29 m , something well beyond any realistic possibility . in our proposal ,
two different storage rings will be used to aim to the detectors .
therefore , it is possible to design two rings of different characteristics and reduce the size of the ring that is aiming at the far detector . note that with the refurbished sps ( sps+ ) the @xmath6b and @xmath6li ions could be accelerated up to @xmath30 and @xmath31 , respectively . due to the resonance ,
a @xmath32 increase in the @xmath14 factor for @xmath6li produces an increase of a @xmath33 in the number of antineutrino events at the detector .
therefore , we can use a ring with a much shorter straight section , @xmath34 m and the physics reach of the setup will remain practically unaffected , in spite of the fact that it will have a lower livetime _
= @xmath35 . the maximum depth
this ring would reach is @xmath36 m , which is still much larger than what is needed for the neutrino factory but is almost 1 km less deep than in the original design .
we make the following choices for our detectors : ( 1 ) since cp measurements are better at lower energies with @xmath2ne and @xmath3he as source ions @xcite , it is preferable to have a detector with lower threshold and good energy resolution .
we opt for a water detector with 500 kton fiducial mass ( as in refs .
@xcite ) . this detector could be housed at canfranc , for example , at a distance of 650 km from the @xmath0-beam at cern ; ( 2 ) mass hierarchy measurement is the main motivation for the experiment at the magic baseline , for which higher energy neutrinos from highly boosted @xmath6b and @xmath6li ions will be used .
we prefer thus to use a 50 kton magnetized iron detector at this baseline
. this far detector could be the ical@ino detector in india @xcite which is at a distance of 7152 km , and which will soon go under construction .
the efficiencies and beam - induced backgrounds expected in a water detector for the @xmath19 @xmath0-beam fluxes from @xmath2ne and @xmath3he decays are given in @xcite as migration matrices that we use to simulate the detector .
unfortunately , a similarly detailed analysis of the performance of the iron detector exposed to the @xmath0-beam fluxes is lacking .
we therefore follow the efficiencies and backgrounds derived in @xcite for the neutrino factory fluxes instead .
notice that this is a very conservative assumption , since for a @xmath0-beam charge identification is not mandatory , unlike in the nf .
therefore , we could take advantage of this and use the charge i d capability of the detector to further reduce the background . finally , note that the largest uncertainties in the performance of the iron detector are on the efficiencies and backgrounds for the events of lowest energy , around @xmath37 gev
. however , the main role of the iron detector is to observe the resonance in the probability , which takes place around @xmath38 gev , so these uncertainties will practically have no effect at all in the performance of the setup .
we quantify the sensitivity reach of the experiments in terms of three different performance indicators : 1 .
the @xmath39 discovery reach : this is the minimum true value of @xmath39 for which the experiment can rule out at the 1 d.o.f .
@xmath40 the value @xmath41 in the fit , after marginalizing over all the other parameters . 2 .
the cp - violation reach : this is the range of @xmath16 as a function of @xmath39 which can rule out no cp - violation ( @xmath42 and @xmath43 ) at the 1 d.o.f . @xmath40 , after marginalizing over all the other parameters .
the @xmath44 reach in @xmath39 : this is defined as the limiting value of @xmath39 for which the wrong hierarchy can be eliminated at @xmath40 . in fig .
[ fig : sens ] we present the comparison of our set - up with respect to other high @xmath14 @xmath0-beam proposals , defined as follows : 1 .
solid , black lines : these correspond to our proposal @xcite .
blue , dotted lines : the two - baseline @xmath0-beam set - up proposed in @xcite . here
neutrino beams from decay of @xmath6b and @xmath6li with boost factor @xmath19 , are detected in two 50 kton magnetized iron detector located at 2000 km and 7000 km respectively .
3 . orange , dashed lines : the two - baseline @xmath0-beam set - up proposed in @xcite . here
all four ions are used .
beams from decays of @xmath2ne and @xmath3he accelerated to @xmath45 are detected in a 50 kton totally active scintillator detector ( tasd ) at 730 km .
beams from decays of @xmath6b and @xmath6li accelerated to @xmath46 are detected in a 50 kton magnetized iron detector at 7000 km .
purple , dot - dashed lines : the one - baseline @xmath0-beam set - up proposed in @xcite .
neutrino beams produced by @xmath2ne and @xmath3he decays , each accelerated to @xmath47 are detected in a 500 kton water detector located at 650 km .
the original @xmath0-beam proposals assumed `` standard '' useful fluxes of @xmath48 and @xmath49 decays per year for @xmath2ne and @xmath3he respectively .
similar standard numbers regarding @xmath6b and @xmath6li are lacking . at present , @xmath6li and @xmath3he can be easily produced with the standard isolde techniques .
@xmath6b is easily produced , but it tends to interact with the medium nearby .
finally , @xmath2ne production is still difficult , and we are far from the original goal of @xmath48 ions per year .
preliminary studies on the production rates of @xmath6b and @xmath6li show that it could be enhanced in the near future .
@xmath0-beams are facilities under study for construction in the next two decades .
therefore , we will assume that @xmath50 ions per year can be stored into the ring , for all ion species@xcite .
note that for the `` standard '' storage ring considered in set - ups 2 , 3 and 4 , the livetime is _ l _ = 0.28 , what translates into _ @xmath51 useful decays per year _ per polarity .
however , for our proposal , the storage ring for the @xmath6b and @xmath6li ions is shorter , giving a 40% smaller livetime .
we have considered 2.5% and 5% systematic errors on the signal and on the beam - induced background , respectively .
they have been included as `` pulls '' in the statistical @xmath52 analysis .
the following @xmath53 errors for the oscillation parameters were also considered : @xmath54 , @xmath55 , @xmath56 and @xmath57 .
eventually , an error @xmath58 has been considered for the earth density given by the prem model @xcite .
marginalization over these parameters has been performed for all observables .
the globes 3.0 @xcite software was used to perform the numerical analysis . [ cols="^,^ " , ] we present in fig .
[ fig : senscpf ] the comparison of the performance of our setup with the ids neutrino factory design @xcite and the high-@xmath14 @xmath0-beam set - up 4 , described above . in this case
we present the sensitivities to the observables as a function of the fraction of the values of @xmath16 for which they can be discovered instead of the true values of @xmath16 . from fig .
[ fig : senscpf ] it is clear that the facility with sensitivity to the different observables down to smallest values of @xmath39 is the neutrino factory .
this can be understood from the very large fluxes assumed for the ids baseline as compared to the ones assumed here for the @xmath0-beam set - ups : @xmath59 useful muon decays per year and per baseline to be compared to the @xmath51 assumed for the @xmath0-beams . on the other hand , the high energy of the neutrino factory beams implies a very small value of @xmath60 .
this translates into a stronger suppression of the cp violating term of the oscillation probability with respect to the one suppressed by two powers of @xmath1 for large values of this parameter .
therefore , the cp discovery potential of @xmath0-beams outperforms that of the neutrino factory in fig .
[ fig : senscpf ] when @xmath61 .
since this large value of @xmath62 also guarantees a discovery of the mass hierarchy and @xmath39 regardless of the value of @xmath16 , this makes @xmath0-beams the better option when @xmath61 .
furthermore , even if the statistics in the near @xmath0-beam detector is reduced by half in our proposal compared to the one in ref .
@xcite , the cp - discovery potential for @xmath61 is better in the two - baseline set - up due to the lifting of the degeneracies that can mimic cp - conservation when combining the information from the two detectors .
we have presented a new @xmath0-beam set - up that combines the strengths of the best set - ups in the literature , trying to probe with the same facility @xmath1 , the existence of leptonic cp - violation and the neutrino mass ordering in the challenging regime of small @xmath1 . in order to achieve good sensitivity to the cp phase , we have chosen the highest @xmath14 accesible at the sps+ but exploiting the decay of the ions with smallest end - point energy .
this guarantees good statistics at the detector , since the flux and cross sections grow with @xmath14 , while maintaining the mean energy around @xmath63 gev , which allows to consider a water detector .
furthermore , the oscillation baseline can be kept short , so as to further increase the statistics and to avoid strong matter effects that could spoil the cp discovery potential . on the other hand , to achieve sensitivity to the mass hierarchy ,
a far detector located at the magic baseline is mandatory . for this detector
, ions with higher end - point energies are preferred , and higher @xmath14 values help to increase the statistics . as for this case
the neutrino fluxes are peaked around @xmath64 gev , we opted for an iron detector .
while two - baseline @xmath0-beam set - ups have been proposed and studied before , our proposal is unique .
we propose two different racetrack geometry decay rings , as we use different ions for the two baselines .
we have kept the initial design for the ring that is aiming at the near detector .
however , the @xmath6b and @xmath6li beam has to be sent over a baseline @xmath7 km , and hence its storage ring requires an inclination of @xmath8 . this would require a maximum depth @xmath9 m at the far end of the storage ring if we use the same design . in order to alleviate this problem ,
one necessarily has to reduce the size of the straight sections of the ring that is aiming to the far detector . to compensate for the consequent loss in the number of events due to this reduction
, we propose to increase the @xmath14 for the @xmath6b and @xmath6li ions .
we have also made a full comparison with the rest of proposed facilities . while the presently assumed @xmath0-beam fluxes can not compete with the expectations from a neutrino factory and can not probe values of @xmath1 much smaller than @xmath65 , we find that @xmath0-beam set - ups are better optimized for regions with @xmath66 , providing sensitivity to the different observables in larger fractions of the parameter space .
we believe that the combination of ions and baselines proposed here represents an optimal @xmath0-beam set - up , that takes advantage of the properties of the different achievable beams , with very good sensitivity to all of the three observables considered .
a. de rujula , m. b. gavela and p. hernandez , nucl .
b * 547 * , 21 ( 1999 ) [ arxiv : hep - ph/9811390 ] .
p. zucchelli , phys .
b * 532 * , 166 ( 2002 ) .
p. huber and w. winter , phys .
d * 68 * , 037301 ( 2003 ) .
[ arxiv : hep - ph/0301257 ] .
p. coloma , a. donini , e. fernandez - martinez and j. lopez - pavon , jhep * 0805 * , 050 ( 2008 ) [ arxiv:0712.0796 [ hep - ph ] ]
. s. k. agarwalla , s. choubey and a. raychaudhuri , nucl .
b * 805 * , 305 ( 2008 ) [ arxiv:0804.3007 [ hep - ph ] ] .
j. burguet - castell , d. casper , e. couce , j. j. gmez - cadenas and p. hernandez , nucl .
b * 725 * , 306 ( 2005 ) [ arxiv : hep - ph/0503021 ] .
j. burguet - castell , d. casper , j. j. gmez - cadenas , p. hernandez and f. sanchez , nucl .
b * 695 * , 217 ( 2004 ) [ arxiv : hep - ph/0312068 ] .
m. s. athar _ et al . _
[ ino collaboration ] , a report of the ino feasibility study , + http://www.imsc.res.in/ ino / openreports / inoreport.pdf t. abe _ et al . _ [ iss detector working group ] , arxiv:0712.4129 [ physics.ins-det ] .
s. choubey , p. coloma , a. donini and e. fernandez - martinez , arxiv:0907.2379 [ hep - ph ] .
a. donini and m. lindroos , _ talk at nufact08 , valencia , spain , 30 jun - 5 jul 2008 _
huber , m. lindner and w. winter , comput .
phys . commun .
* 167 * ( 2005 ) 195 [ arxiv : hep - ph/0407333 ] .
p. huber , j. kopp , m. lindner , m. rolinec and w. winter , comput .
commun .
* 177 * ( 2007 ) 432 [ arxiv : hep - ph/0701187 ] .
a. bandyopadhyay _ et al . _
[ iss physics working group ] , arxiv:0710.4947 [ hep - ph ] . | we propose a @xmath0-beam experiment with four source ions and two baselines for the best possible sensitivity to @xmath1 , cp violation and mass hierarchy .
neutrinos from @xmath2ne and @xmath3he with lorentz boost @xmath4 are detected in a 500 kton water detector at a distance @xmath5 km from the source .
neutrinos from @xmath6b and @xmath6li are detected in a 50 kton magnetized iron detector at a distance @xmath7 km from the source . since a tilt angle @xmath8
is required to send the beam to the magic baseline , the far end of the ring has a maximum depth of @xmath9 m. we alleviate this problem by proposing to trade reduction of the decay ring with the increase in the boost factor of the @xmath6li and @xmath6b ions up to @xmath10 and @xmath11 , such that the number of events at the detector remains almost the same .
we study the sensitivity reach of this two - baseline two - storage ring @xmath0-beam experiment , and compare it with the corresponding reach of the other proposed facilities .
address=
instituto de fsica terica , universidad autnoma de madrid , + 28049 madrid , espaa |
the two states @xmath0 and @xmath1 observed in the invariant mass spectrum of @xmath5 in @xmath2 by the lhcb collaboration @xcite has immediately attracted a lot of attention from the whole community since they could be the long - searching - for pentaquark states in the heavy flavor sector .
recent theoretical studies can be found in refs .
their masses are @xmath6 mev and @xmath7 mev , respectively , and their widths are @xmath8 mev and @xmath9 mev , respectively .
their preferred @xmath10 are either @xmath11 and @xmath12 , or @xmath13 and @xmath14 , respectively , based on the detailed experimental analysis .
the possible existence of multiquark states has always been regarded as a natural consequence of qcd .
although the conventional quark model has made tremendous successes in the description of the hadron spectroscopy , it also raised questions on why and how those multiquark states kept out of our sight for such a long time .
the lhcb results and the recent results from belle @xcite and besiii @xcite certainly make a big step forward to our understanding the multiquark system .
but at the same time , they also give us chances to ask more questions . in this letter
we will examine the role of the non - perturbative anomalous triangle singularity ( ats ) in the decay of @xmath2 .
as recently pointed out in ref .
@xcite , the pronounced narrow threshold states in the elastic channels should indicate non - perturbative rescatterings which will eventually generate pole structures after sum over all the loops to infinity .
it was also stressed that if the ats is present the threshold peak will be enhanced and mix with the dynamic pole structure in the inelastic channel .
therefore , in order to understand the nature of the threshold enhancements a careful analysis of the triangle process should be necessary .
there are several interesting features arising from the decay channel as the data have shown .
for instance , there are clear structures for @xmath15 resonances in the lower end of the @xmath16 invariant mass spectrum while the higher mass region appear to smooth out . in contrast , the observed @xmath0 and @xmath1 , in particular , the @xmath1 , are clear structures above the phase space .
note that some interferences occur over a rather broad mass range above the @xmath1 in the @xmath5 invariant mass spectrum ( see e.g. fig . 2 of ref .
this feature suggests that there are intermediate processes in @xmath2 which will give rise to final state interactions .
our motivation is to investigate whether the ats will accumulate the events at the two peak positions .
as follows , in the next section we will first analyze the transition mechanisms for @xmath2 and then investigate the loop diagrams where the ats can be present .
a summary will be given in the last section .
the transition of @xmath2 is dominated by the flavor changing weak decay of @xmath17 via the @xmath18 current .
this leads to an intuitive expectation that the @xmath19 and @xmath20 quark in @xmath4 should be a spectator , hence their isospin quantum number @xmath21 should be conserved . as a consequence ,
if the @xmath22 pair is to be combined with a @xmath23 quark to form a charmed baryon , it will favor @xmath24 baryons instead of @xmath25 ones . in refs .
@xcite the authors propose that @xmath0 and @xmath1 are molecular states of @xmath26 and @xmath27 .
but it was not discussed why the production of @xmath3 in the @xmath4 decay should be significant . also , in ref .
@xcite the authors refer one of the pentaquark candidates to a molecular state of @xmath26 with @xmath28 . in ref .
@xcite the authors proposed to produce the @xmath1 via the @xmath5 rescattering to @xmath29 which , however , should be higher order terms . in a followed - up paper @xcite
the authors investigate other isospin 0 channel recoiled by @xmath30 .
actually , the production of the @xmath3 is not necessarily to be suppressed at leading order .
since the @xmath31 quark decay is a short - distance process , the final @xmath23 , @xmath32 and @xmath33 quark will be created locally .
they can easily split the @xmath22 quark and rearrange them with the created @xmath34 to form , e.g. @xmath35 .
then the rescattering of @xmath29 into @xmath5 can occur .
this process is illustrated in fig .
[ fig-1 ] ( c ) with ( a ) and ( b ) the same mechanism analyzed in ref .
@xcite as a contrast . for fig .
[ fig-1 ] ( b ) , a decomposition of the isospin gives the relative decay amplitudes @xmath36 where @xmath37 denotes charmed baryon @xmath38 or @xmath3 .
we denote the possible @xmath38 and @xmath39 by the ground state symbols and simply write the amplitudes by the final state particles .
the flavor wavefunction @xmath40 for @xmath4 is implied .
similarly , the decomposition for fig . [ fig-1 ] ( c ) can be written as @xmath41 \ , \end{aligned}\ ] ] where @xmath42 and @xmath43 would allow the formation of pentaquark states via the @xmath44 rescatterings as studied in refs .
@xcite . from eqs .
( [ trans-1 ] ) and ( [ trans-2 ] ) we can see that fig .
[ fig-1 ] ( b ) dominantly contribute to the production of @xmath45 pentaquark system recoiled by @xmath46 . and
@xmath1 in @xmath2 .
, scaledwidth=60.0% ] if the production of the pentaquark candidates @xmath0 and @xmath1 is indeed via the @xmath47 ( table [ tab-3 ] ) or @xmath48 ( table [ tab-1 ] ) interactions , then the relevant thresholds will play non - trivial roles in the coupled - channel decays into e.g. @xmath5 . considering the major rescattering processes arising from fig .
[ fig-1 ] , we can reexpress the transitions as fig . [ fig-2 ] where three type of rescatterings will contribute .
figure [ fig-2 ] ( a ) and ( b ) are triangle diagrams while ( c ) is two - point loop interactions .
we note that the creation of @xmath3 via fig .
[ fig-1 ] ( c ) should be driven by short - distance interactions .
this process will favor the formation of @xmath0 and @xmath1 if they are states generated by @xmath47 interactions . the interesting property of fig .
[ fig-2 ] ( a ) and ( b ) is that given the masses of the involved states located within certain ranges it will allow the internal states to be on - shell simultaneously .
this is different from the kinematic cusp effects which only recognizes the on - shell condition for two internal particles and contributions from such a branch point is subleading compared to the ats @xcite .
therefore , we do not expect that such cusp effects produce narrow and strongly enhanced structures in the invariant mass spectrum .
in contrast , when the ats condition is satisfied , the singularity behavior of the integral will produce strong enhancements at the singular points of which the effects can be measured in the experiment .
in particular , the singular points will mostly locate in the vicinity of the two - body thresholds but not necessarily to be exactly at the thresholds .
it should be realized that the positions of the singularity will not change even when higher partial waves contribute at the interaction vertices .
the reason is because the singular term will always be kept in the decomposition of the integrand in the feynman parametrization .
in other words , even though the contribution from the singular term relative to other contributions might be small , its enhancement at the singular point may not be negligible .
nevertheless , in the case of @xmath2 there are several thresholds close to each other .
even a small singularity enhancement can build up and produce measurable effects . since quite a lot of thresholds can appear in the decays of fig .
[ fig-2 ] and we are still lack of information about the vertex couplings , we only consider low partial waves and thresholds which are close to the masses of interest and we discuss separately the properties of those three types of loops in fig . [ fig-2 ] . where the ats and kinematic cusp can be recognized . ,
scaledwidth=50.0% ] figure [ fig-2 ] ( a ) is a consequence of fig .
[ fig-1 ] ( a ) where the rescattering between @xmath49 and an exchanged proton from the decay @xmath50 is considered .
note that the mass thresholds for @xmath51 ( @xmath52 ) are close to the peak masses for @xmath0 and @xmath1 as listed in table [ tab - chic ] .
also , the @xmath53-wave scatterings of @xmath54 and @xmath55 is suppressed in @xmath56 decay or any hadron with only one heavy @xmath31 quark without charm or anti - charm quark as discussed in ref .
@xcite , we can still estimate how large the next leading order contributions are or what is the behavior after considering it .
] can access the quantum numbers of @xmath13 and @xmath12 for the threshold enhancement .
the @xmath57 and @xmath58 scattering can access the quantum numbers of @xmath59 and @xmath13 .
the @xmath60 can reach @xmath61 and @xmath11 via a @xmath62 wave interaction .
it is interesting to notice that the significant enhancement to the @xmath63 ( with @xmath64 via the ats would prefer that the mass of @xmath15 within the mass regions @xmath65 , @xmath66 , @xmath67 , respectively . from fig . 2 ( a ) of ref .
@xcite , it shows that the cross section for @xmath68 is smooth but non - zero .
note that as long as the kinematics approaching the ats condition , all the cross sections will contribute to the threshold singularity . in fig .
[ fig - chic ] we show the structures in the invariant mass of @xmath5 via the triangle diagram of fig .
[ fig-2 ] ( a ) .
as discussed before , since @xmath69 and @xmath70 can access the possible quantum numbers via the @xmath53 wave , we only consider loops of @xmath57 and @xmath55 at this moment . on the other hand , since the branching ratio of @xmath71 is one order of magnitude smaller than that of @xmath72 , we expect that it is also the case in the @xmath4 decay .
then , the upper limit of the contributions from @xmath55 can be estimated by requiring @xmath73 . as shown in fig .
[ fig - chic ] , the ats can produce significant threshold enhancement of @xmath74 while the effect from the @xmath70 loop is strongly suppressed .
note that the singularity from the @xmath74 loop exactly locates at the mass position of the observed @xmath75 .
the pole trajectories of the @xmath74 loop and the properties of the poles on different riemann sheet were first considered in ref . @xcite .
.the @xmath76 thresholds which can be enhanced by the ats via fig .
[ fig-2 ] ( a ) . [ cols="^,^,^,^",options="header " , ] for fig .
[ fig-2 ] ( c ) , the kinematic effects will be just cusp structures in the @xmath5 invariant mass spectrum .
we do not discuss the dynamic consequences if the intermediate @xmath47 ( table [ tab-3 ] ) may have strong couplings , then they may generate dynamic poles near threshold after proper summation over the bubble loops .
instead , we only show the kinematic cusp for which has turned out not to lead to pronounced structures as recently studied in ref .
@xcite . in another word , the observed pronounced peaks may either be produced by possible pole structures or the ats mechanisms . in order to try to distinguish the behavior from a pole structure and the ats , we generate the invariant mass spectra of @xmath5 in different @xmath16 invariant mass regions the same as fig .
8 of ref .
@xcite ( see fig .
[ fig - cutoff ] ) .
since the ats contributions will vary in terms of different kinematics , such a quantity will be able to distinguish the behavior of the pole and the ats mechanism .
the results for those three loop processes of fig .
[ fig-2 ] ( a ) , ( b ) and ( c ) are illustrated by the solid , dashed and dotted lines which are different from the symmetric breit - wigner lineshape .
in particular , the structures created by the cusp effects appear to be negligible .
note that the kinematics of higher invariant mass of @xmath16 is favored by fig .
[ fig-2 ] ( a ) .
it means even small couplings for @xmath63 scattering are introduced , the ats enhancement can still be observable . since we introduce several mechanisms to generate the kinematic singularities , it is also necessary to discuss their similar and different characteristics .
both @xmath74 and @xmath77 thresholds are very close to the mass of @xmath1 .
note that @xmath74 can scatter into @xmath5 via multi - gluon exchange process , while @xmath78 can scatter into @xmath5 via quark interchange process which is a rearrangement of the quark flavors .
the @xmath62-wave scattering between @xmath57 and proton will imply that the quantum numbers of the @xmath5 system can be @xmath79 , but the @xmath62-wave scattering between @xmath80 and @xmath81 will imply that the quantum numbers can be @xmath82 .
some of these quantum numbers are compatible with the experimental fitting results .
if one hopes the rescattering mechanism can favor the @xmath83 assignment of @xmath1 , the @xmath84-wave rescattering of @xmath74 or @xmath77 will be required .
usually the higher partial wave rescattering amplitudes will be suppressed to some extent .
however , the analytic properties of the kinematic singularities will mainly depend on the kinematics of the loop integrals , which will not be affected too much by the coupling forms of the vertices . to clarify whether the @xmath76 rescatterings or the @xmath85 rescatterings would be dominant in producing the resonance - like peaks
, we suggest that @xmath86 would be a promising channel .
namely , the transition @xmath87 would require the heavy quark spin flip , thus , will be suppressed .
in contrast , in @xmath85@xmath88@xmath89 such a spin flip does not necessarily occur . furthermore , since the @xmath89 threshold is much larger than the lower thresholds , some higher resonance - like peaks induced by the rescatterings , similar to @xmath1 , can be expected in the @xmath89 invariant mass distributions . with different @xmath16 momentum cuts which are same as fig .
8 of ref .
@xcite , i.e. ( a ) @xmath90,(b ) @xmath91 , ( c ) @xmath92 , ( d ) @xmath93 .
the vertical dashed lines indicate the masses of @xmath0 and @xmath1 , respectively.,scaledwidth=60.0% ]
in this work we analyze the role played by anomalous triangle singularity in @xmath2 as a possible contribution to the observed hidden - charm pentaquark candidates @xmath0 and @xmath1 .
we first show that the @xmath44 can be produced at leading order which seems to be overlooked in the literature .
we then demonstrate that the ats can generate threshold enhancements which can mimic the experimental observations .
we have to admit that more detailed model construction is needed in order to determine better the ats behavior and strengths .
but we emphasize that the kinematics of @xmath2 are in favor of the ats mechanism when intermediate states are involved .
since all the triangle amplitudes can build up the contributions from the ats near the mass region that we are interested in , they seem to be able to produce structures similar to what observed in experiment . meanwhile , we find that the @xmath94 threshold can give rise to a third peak around 4.5 gev .
we also discuss the kinematic feature of @xmath47 which can produce kinematic cusp effects . in this sense ,
the observed pronounced structures can be either produced by pole structures or the ats mechanism .
the kinematic dependence of the ats is investigated by looking at the ats cross sections in different @xmath16 invariant mass regions .
we also show that the ats mechanism should have rather strong dependence of the kinematics by looking at the @xmath5 invariant mass distributions at different energy cuts for @xmath68 . in brief , since the ats could play a role in the production of threshold states , it may mix with the pole if the singular threshold is close to the pole position .
such a mixing and interference might generate complicated structures such as to distort the lineshape or produce narrower enhancements that behaves very differently from expected pole contributions .
in order to better understand the nature of these newly observed pentaquark candidates , a combined study of the ats and dynamically generated pole structures should be necessary .
useful discussions with ulf - g . meiner and m. oka are acknowledged .
this work is supported , in part , by the japan society for the promotion of science under contract no .
p14324 , the jsps kakenhi ( grant no .
25247036 ) , the sino - german crc 110 symmetries and the emergence of structure in qcd " ( nsfc grant no .
11261130311 ) , the national natural science foundation of china ( grant nos .
11425525 ) , and the national key basic research program of china under contract no . | we find that several thresholds can contribute to the enhancements of the newly observed heavy pentaquark candidates @xmath0 and @xmath1 via the anomalous triangle singularity ( ats ) transitions in the specific kinematics of @xmath2 .
apart from the observed two peaks we find that another peaks around 4.5 gev can also be produced by the ats .
we also show that the @xmath3 can be produced at leading order in @xmath4 decay .
this process is different from the triangle diagram and its threshold enhancement only appears as cusp effects if there is no pole structure or the ats involved .
the threshold interaction associated with the presence of the ats turns out to be a general phenomenon and plays a crucial role in the understanding of candidates for exotic states . |
we consider the fowler equation @xcite : @xmath0 ( x ) - \partial_{x}^2 u(t , x ) = 0,\quad x \in \r , t>0 , \\ &
u(0,x ) = u_0(x),\quad x \in \r , \end{aligned } \right .
\label{fowlereqn5}\ ] ] where @xmath1 represents the dune height and @xmath2 is a nonlocal operator defined as follows : for any schwartz function @xmath3 and any @xmath4 , @xmath5 ( x ) : = \int_{0}^{+\infty } |\xi|^{-\frac{1}{3 } } \varphi''(x-\xi ) \ ,
d\xi . \label{nonlocalterm4}\ ] ] we refer to @xcite for theoretical results on this equation .
the nonlocal term @xmath2 is anti - diffusive .
indeed , it has been proved in @xcite that @xmath6\right)(\xi ) = - 4 \pi^2 \gamma\left(\frac{2}{3}\right ) \left(\frac{1}{2}-i \ , \mbox{sgn}(\xi ) \frac{\sqrt{3}}{2}\right)|\xi|^{4/3 } , \ ] ] where @xmath7 denotes the fourier transform normalized in .
thus , @xmath2 can been seen as a fractional power of order @xmath8 of the laplacian , with the `` bad '' sign .
it will be clear from the analysis below that our results can easily be extended to the case where @xmath2 is replaced with a fourier multiplier homogeneous of degree @xmath90,2[$ ] , as in @xcite , and not only @xmath10 .
we assume that the initial data @xmath11 belongs to @xmath12 , and thus has a unique solution belonging to @xmath13 , h^3(\r))$ ] for all @xmath14 , from @xcite
. we will denote @xmath15 by @xmath16 ; @xmath17 maps @xmath18 to itself .
duhamel s formula for the continuous problem reads @xmath19 where @xmath20 is the kernel of the operator @xmath21 , and @xmath22 is defined by @xmath23 where @xmath24 are positive constants .
recently , to solve the fowler equation some numerical experiments have been performed using mainly finite difference approximation schemes @xcite .
however , these schemes are not effective because if we opt for an explicit scheme , numerical stability requires that the time step @xmath25 is limited by @xmath26 and , if we choose an implicit scheme , we have to solve a large system which is a computationally expensive operation .
thus , the splitting method becomes an interesting alternative to solve the fowler model . to our knowledge
, there is no convergence result in the literature for the splitting method associated to the fowler equation .
this method is more commonly used to split different physical terms , such as reaction and diffusion terms , see for instance @xcite .
splitting methods have also been employed for solving a wide range of nonlinear wave equations .
the basic idea of this method is to decompose the original problem into sub - problems and then to approximate the solution of the original problem by solving successively the sub - problems .
various versions of this method have been developed for the nonlinear schrdinger , korteweg - de - vries and modified korteweg - de - vries equations , see for instance @xcite . + for the fowler model , we consider , separately , the linear cauchy problem @xmath27 - \eta \ , \partial_{x}^2 v = 0 ; \quad v(0,x ) = v_0(x ) , \label{nlocal}\ ] ] and the nonlinear cauchy problem @xmath28 where @xmath29 are fixed _ positive _ parameters such that @xmath30 . equation is simply the viscous burgers equation .
we denote by @xmath31 and @xmath32 , respectively , the evolution operator associated with and : @xmath33 where @xmath34 with @xmath35 where @xmath36 is the heat kernel defined by @xmath37 furthermore , the following @xmath38-estimate holds @xmath39 let us explain the choice of this decomposition .
first , we can remark that if we do not consider the nonlinear term in , the analytical solutions are available using the fourier transform .
thus , the linear part may be computed efficiently using a fast fourier transform ( hereafter fft ) algorithms .
note also that the laplacian and the fractional term @xmath2 can not be treated separately .
indeed , the equation @xmath40 = 0 $ ] is ill - posed .
we next decide to handle the nonlinear term by adding a bit of viscosity in order to avoid shock problems in the standard burgers equation .
therefore , the splitting approach presented in this article differs from e.g. the one analyzed in @xcite , which corresponds to assuming @xmath41 in the above definitions .
the splitting operators associated to this approach when @xmath42 ( which amount to considering alternatively the heat equation and the burgers equation ) have been studied in @xcite ( as well as other equations involving the burgers nonlinearity , such as the kdv equation , see also @xcite ) .
the main difference with the result presented in @xcite is that the operator @xmath2 is not a differential operator , so its action on nonlinear terms is rather involved , while this point would be needed to compute the lie commutator between the two operators @xmath43,\quad b(v ) = -v\d_x v + \eps^2\d_x^2 v,\ ] ] as in e.g. @xcite .
note that since we consider the lie splitting operator , we could in principle use the exact formula established in @xcite for the local error .
but then again , we face the problem to compute @xmath44 $ ] , which is the only term that we can not estimate in ( * ? ? ?
* theorem 1 ) .
we finally point out that we use smoothing effects associated to the viscous burgers equation ; see corollaries [ cor : condest ] and [ cor : split1 ] .
we motivate this choice by the presence of artificial diffusion in classical numerical schemes used to solve the convection equations .
an alternative to reduce this effect is to consider numerical schemes of high order which are usually computationally expensive and do not seem to be very useful for the fowler model because of the diffusion term .
we consider the lie formula defined by @xmath45 the alternative definition @xmath46 could be studied as well , leading to a similar result . also , the following evolution operators @xmath47 corresponding to the strang method @xcite could be considered .
following the computations detailed in the present paper for the case , it would be possible to show that the other lie formula generates a scheme of order one , and to prove that the strang method is of order two ( for smooth initial data ) , in the same fashion as in , e.g. , @xcite .
this fact is simply illustrated numerically in section [ numerique ] , to avoid a lengthy presentation . with @xmath48
given by , our main result is : [ theoreme1 ] for all @xmath49 and for all @xmath50 , there exist positive constants @xmath51 and @xmath52 such that for all @xmath530,\delta t_0]$ ] and for all @xmath54 such that @xmath55 , @xmath56 here , @xmath51 and @xmath52 depend only on @xmath57 , @xmath58 } \|s^t u_0\|_{h^2(\r)}$ ] , and @xmath59
. it will follow from lemma [ lemme8 ] that @xmath60 } \|s^t u_0\|_{h^2(\r)}\le c_t(\|u_0\|_{h^1(\r)})\|u_0\|_{h^2(\r)},\ ] ] for some nonlinear ( increasing ) function @xmath61 depending on @xmath57 . in this paper , we begin by estimating the @xmath38-stability for error propagation .
we next prove that the local error of the lie formula is an approximation of order two in time .
finally we prove that this evolution operator represents a good approximation , of order one in time , of the evolution operator @xmath17 in the sense of theorem [ theoreme1 ] .
+ furthermore , we apply lie and strang approximations in order to make some numerical simulations using the split - step fourier experiments . +
this paper is organised as follows . in the next section
we give some properties related to the kernels @xmath36 and @xmath62 , and we prove two fractional gronwall lemmas . in section [ sectionlemme ] , we establish some estimates on @xmath31 , @xmath32 , @xmath48 and @xmath17 . in section [ errorlocal ] , we prove a local @xmath38 error estimate .
theorem [ theoreme1 ] is proved in section [ sec : proof ] .
we finally perform some numerical experiments which show that the lie and strang methods have a convergence rate in @xmath63 and @xmath64 , respectively
. * notations . * + - we denote by @xmath7 the fourier transform of @xmath65 which is defined by : for all @xmath66 , @xmath67 we denote by @xmath68 its inverse .
+ - we denote by @xmath69 a generic constant , strictly positive , which depends on parameters @xmath70 , and @xmath57 .
@xmath71 is assumed to be a monotone increasing function of its arguments .
we begin by recalling the properties of kernels involved in the present analysis .
[ kernel2 ] the kernel @xmath62 satisfies : 1 .
@xmath72 , @xmath73 and @xmath740,\infty[\times \r\right)$ ] .
2 . @xmath75 .
@xmath76 such that for all @xmath770,t]$ ] , @xmath78 .
@xmath76 such that for all @xmath790,t]$ ] , @xmath80 . 5 . for any @xmath81 and @xmath82 , @xmath83 where @xmath84 [ heat ] the kernel @xmath36 satisfies : 1 .
@xmath850,\infty[\times \r\right)$ ] .
2 . @xmath86 , @xmath87 .
3 . @xmath88 , @xmath89 .
@xmath90 such that for all @xmath91 , @xmath92 .
@xmath93 such that for all @xmath94 , @xmath95 .
[ remarkconvolution ] the kernel @xmath96 of @xmath97 has similar properties to the kernel @xmath62 .
moreover , for all @xmath14 , we have @xmath98 [ lemme1 ] let @xmath99 \to \r_+$ ] be a bounded measurable function , and suppose that there are positive constants @xmath100 and @xmath1010,1[$ ] such that for all @xmath102 $ ] , @xmath103 where @xmath104 is the riemann liouville operator defined by @xmath105 then there exists @xmath106 such that @xmath107.\ ] ] the proof of this lemma is well - known and is based on an iteration argument ; see for instance ( * ? ? ?
* lemma 4.3 ) or ( * ? ? ?
* corollary 2 ) .
we sketch the argument for the sake of completeness .
iterating inequality once , we have @xmath108 from fubini s theorem , we get @xmath109 where @xmath110 is the beta function .
therefore , we have @xmath111 iterating the estimate @xmath112 times , with @xmath113 , we get the following estimate : @xmath114 with @xmath115 , and where @xmath116 is a positive constant which depends on @xmath57 and @xmath117 .
the lemma then follows from the classical gronwall lemma .
[ lemme2 ] let @xmath99 \to \r_+$ ] be a bounded measurable function and @xmath118 be a polynomial with positive coefficients and no constant term .
we assume there exists two positive constants @xmath71 and @xmath1190,1[$ ] such that for all @xmath102 $ ] , @xmath120 then there exists @xmath106 such that for all @xmath121 $ ] , @xmath122 arguing as in the proof of lemma [ lemme1 ] , we iterate the previous inequality .
after one iteration , we get @xmath123 where we have used the assumptions on @xmath118 , and fubini s theorem again for the last term . iterating sufficiently many times
, we infer like in the proof of lemma [ lemme1 ] : @xmath124 with @xmath125 .
set @xmath126 then @xmath127 using to control the second term , and choosing @xmath128 sufficiently large , we infer : @xmath129 since @xmath130 @xmath131 , for all @xmath102 $ ] , @xmath132 this completes the proof .
in this paragraph , we collect several estimates concerning the convolutions with @xmath96 , @xmath62 and @xmath36 , which will be useful in the estimates of the local error of the scheme .
[ estimnonlocal ] let @xmath133 and @xmath134 . then @xmath135 \in h^{s-4/3}(\r)$ ] and we have @xmath136\|_{h^{s-4/3}(\r ) } \le 4 \pi^2 \gamma\(\frac{2}{3}\ ) \|\varphi\|_{h^s(\r)}.\ ] ] for all @xmath133 and all @xmath134 , we have , using @xmath137&\|_{h^{s-4/3}(\r ) } = \left ( \int_\r ( 1 + |\xi|^2)^{s-4/3 } |\f(\i[\varphi])(\xi)|^2 \ , d\xi \right)^{1/2 } \\ & = 4 \pi^2 \gamma\(\frac{2}{3}\ ) \left ( \int_\r ( 1 + |\xi|^2)^{s-4/3 } \left|\frac{1}{2}-i \mbox { sgn}(\xi ) \frac{\sqrt{3}}{2}\right|^2 |\xi |^{8/3 } |\f(\varphi)(\xi)|^2 \ , d\xi \right)^{1/2 } \\ & = 4 \pi^2 \gamma\(\frac{2}{3}\ ) \left ( \int_\r \left ( \frac{|\xi|^2 } { 1 + |\xi|^2}\right ) ^{4/3 } ( 1 + |\xi|^2)^s | \f(\varphi)(\xi)|^2 \ , d\xi \right)^{1/2 } \\ &
\le 4 \pi^2 \gamma\(\frac{2}{3}\ ) \left[\int_\r ( 1 + |\xi|^2)^s | \f(\varphi)(\xi)|^2 \ , d\xi \right]^{1/2 } = 4 \pi^2 \gamma\(\frac{2}{3}\ ) \|\varphi\|_{h^s(\r)},\end{aligned}\ ] ] hence the result .
[ lemme3 ] @xmath138 let @xmath139 then , for all @xmath140 and all @xmath14 , @xmath141 where @xmath142 .
+ @xmath143 let @xmath139 there exists @xmath71 such that for all @xmath144 and all @xmath14 , @xmath145 using plancherel formula , we have @xmath146 moreover , since @xmath147 then , from again plancherel formula , we have @xmath148 hence the first point of the lemma .
let @xmath149 , @xmath150 .
we have @xmath151 but from the definition of @xmath31 , @xmath152 is given by @xmath153 = \eta x^s \d_x^2 v - \i[x^s v],\ ] ] since @xmath154 .
thus , using proposition [ estimnonlocal ] and the first point of this lemma , we get @xmath155 \|_{h^n(\r ) } \ , ds \\ & \le \eta \ , t \ , e^{\beta_0 t}\|v\|_{h^{n+2}(\r ) } + \int_0^t \|\i[x^s v ] \|_{h^n(\r)}ds \\ & \le \eta \ , t \ , e^{\beta_0 t } \|v\|_{h^{n+2}(\r ) } + 4 \pi^2 \gamma\(\frac{2}{3}\ ) \int_0^t \|x^s v \|_{h^{n+4/3}(\r)}ds \\ & \le \eta \ , t \ , e^{\beta_0 t } \|v\|_{h^{n+2}(\r ) } + 4 \pi^2 \gamma\(\frac{2}{3}\ ) \int_0^t \|x^s v \|_{h^{n+2}(\r)}ds \\ & \le \left ( \eta + 4 \pi^2 \gamma\(\frac{2}{3}\ ) \right ) t \ , e^{\beta_0 t } \|v\|_{h^{n+2}(\r ) } , \end{aligned}\ ] ] hence the result .
recalling that @xmath62 corresponds to @xmath96 in the case @xmath156 , we readily infer : [ corokernel ] for all @xmath157 and all @xmath14 , @xmath158 where @xmath71 is a positive constant independent of @xmath159 and @xmath160 .
we conclude this paragraph with an analogous result on the heat kernel @xmath36 : [ lemmechaleur ] for all @xmath161 and all @xmath82 , @xmath162 proceeding as above , we have : @xmath163 taking the norm @xmath38 and using proposition [ heat ] , young s inequality yields @xmath164 hence the result .
we now turn to the viscous burgers equation : @xmath165 it is well - known that the change of unknown function @xmath166 turns the viscous burgers equation into the heat equation : @xmath167 we infer the explicit formula : @xmath168 however , this formula does not seem very helpful in order to establish proposition [ prop : vb ] .
[ prop : vb ] let @xmath169 . then has a unique solution @xmath170 .
in addition , there exists @xmath171 such that for all @xmath172 , @xmath173 if in addition @xmath174 , for some @xmath175 , then @xmath176 and for all @xmath50 , there exists @xmath177 such that for all @xmath121 $ ] , @xmath178 the existence and uniqueness part being standard , we focus on the estimates . the @xmath38 estimate yields ( formally , multiply by @xmath160 and integrate ) @xmath179 and the @xmath180 estimate ( differentiate with respect to @xmath181 , multiply by @xmath182 and integrate ) , @xmath183 the @xmath38 estimate shows that the map @xmath184 is non - increasing : @xmath185 an integration by parts and cauchy schwarz inequality then yield @xmath186 in order to take advantage of the smoothing effect provided by the viscous part , integrate in time and write @xmath187 gagliardo nirenberg inequality yields @xmath188 so using , we infer : @xmath189 in view of hlder inequality in the last integral in time , @xmath190 where we have used young inequality @xmath191 .
we infer @xmath192 gagliardo nirenberg inequality @xmath193 now yields @xmath194 where we have used , hlder inequality and , successively .
integrate the @xmath180 estimate with respect to time , and now discard the viscous part whose contribution is non - negative : @xmath195 the first part of the proposition then follows from the gronwall lemma . to complete the proof of the proposition , we use the general @xmath196 estimate , for @xmath149 : set @xmath197 . applying @xmath198 to yields @xmath199 integrating by parts , the last term is non - positive , since @xmath200 write @xmath201 integrating by parts the first term yields @xmath202 in view of kato - ponce estimate @xcite @xmath203 we have ( with @xmath204 and @xmath205 ) @xmath206 leaving out the viscous term , gronwall lemma yields the _ a priori _
estimate @xmath207 where @xmath71 depends only on @xmath149 .
in particular , for @xmath208 , gronwall lemma implies @xmath209 , where @xmath210 .
we bootstrap , thanks to gagliardo nirenberg inequality again : @xmath211 therefore , for @xmath121 $ ] , @xmath212 the last estimates of the proposition then follow from .
[ estimnl1 ] let @xmath50 . for all @xmath213 ,
there exists @xmath214 such that @xmath215 .
\label{estmnle1}\ ] ] differentiating the duhamel formula in space , we have @xmath216 using young inequality and inequality , we infer : @xmath217 in view of proposition [ heat ] , this implies : @xmath218 writing @xmath219 and invoking the fractional gronwall lemma [ lemme1 ] with @xmath220 , the lemma follows .
[ cor : condest ] let @xmath221 and @xmath222 .
let @xmath223 . if @xmath224 then there exists @xmath225 depending only on @xmath57 and @xmath226 such that @xmath227.\ ] ] denote @xmath228 .
from lemma [ estimnl1 ] , @xmath229 where @xmath71 depends only on @xmath226 and @xmath57 .
from the proof of the first part of proposition [ prop : vb ] ( see the estimate after ) , we infer @xmath230.\ ] ] the corollary then stems from . combining the estimates on @xmath31 and @xmath32 established in the previous two sections ,
we infer : [ cor : split1 ] 1 . for all @xmath231 and all @xmath14 , @xmath232 where @xmath233 .
2 . let @xmath234 and @xmath235 .
there exists @xmath236 such that for all @xmath102 $ ] , @xmath237 the first point is a direct consequence of the relation and lemma [ lemme3 ] .
+ the second point is readily established with lemma [ lemme3 ] and corollary [ cor : condest ] . [ estimapriori2 ]
let @xmath238 and @xmath50 . then , the unique mild solution @xmath239 ; l^{2}(\r))\cap c(]0,t ] ; h^{2}(\r))$ ] of satisfies , for all @xmath102 $ ] @xmath240 where @xmath241 . multiplying by @xmath242 and integrating with respect to the space variable
, we get : @xmath243-u_{xx}\right)u \ : dx = 0\ ] ] because the nonlinear term is zero .
using and the fact that @xmath242 and @xmath244-\partial_{xx}^2 u ) u \ , dx$ ] are real , we get @xmath245-\partial_{xx}^2 u ) u \ , dx = \int_\r \f^{-1}(\psi_\i \f u ) u \ , dx = \int_\r \psi_\i |\f u|^2 \ , d\xi = \int_\r
\mbox{re}(\psi_\i ) |\f u|^2 \ , d\xi.\ ] ] we infer @xmath246 where @xmath247 .
the result then follows from the gronwall lemma .
[ lemme8 ] let @xmath248 @xmath249 and @xmath50 .
there exists @xmath250 such that the unique mild solution @xmath251 ; h^{n}(\r))$ ] satisfies @xmath252 the proof is similar to the one given in lemma [ estimnl1 ] .
differentiating the duhamel formula in space , we have @xmath253 using young inequality and proposition [ kernel2 ] , we infer , for any integer @xmath221 : @xmath254 in view of proposition [ kernel2 ] , this implies : @xmath255 for @xmath256 , we use lemma [ estimapriori2 ] to have @xmath257 the fractional gronwall lemma [ lemme1 ] with @xmath258 then yields @xmath259 where @xmath71 depends only on @xmath57 and @xmath260 . from
, this implies the lemma in the case @xmath256 . for @xmath221 , leibniz rule and cauchy schwarz inequality yield @xmath261 the lemma
then easily follows by induction on @xmath112 .
we will also need the fact that the flow map @xmath17 is uniformly lipschitzean on balls of @xmath262 .
[ prop : lipschitz ] let @xmath263 .
there exists @xmath264 such that if @xmath265 then @xmath266.\ ] ] set @xmath267 , @xmath268 and @xmath269 .
it solves @xmath270-\partial_{xx}^2 w = v\d_x v - u\d_x u = -u\d_x w - w\d_x v.\ ] ] the @xmath38 energy estimate yields : @xmath271 where the term @xmath272-\partial_{xx}^2 w\)$ ] has been estimated as in the proof of lemma [ estimapriori2 ] . we infer @xmath273 where we have used sobolev embedding and lemma [ lemme8 ] .
gronwall lemma yields the result .
as pointed out in the introduction , the formalism of lie derivatives does not seem to be easy to use in the framework of this paper , since the action of the nonlocal operator @xmath2 on nonlinearities is rather involved . as a consequence , we prove an @xmath38 error estimate by a rather pedestrian way . [ localerror ] let @xmath49 .
there exists @xmath274 such that for all @xmath275 $ ] , @xmath276 from the definition of @xmath277 and remark [ remarkconvolution ] , we have @xmath278 thus , from duhamel formula for the fowler equation and the lie formula , we have : @xmath279 where the remainder @xmath280 is written as @xmath281 then , from proposition [ kernel2 ] , corollary [ cor : split1 ] and lemma [ estimapriori2 ] , we have , for @xmath282 $ ] : @xmath283 where @xmath71 is a positive constant .
to estimate the remainder , we decompose it as follows @xmath284 where @xmath285 let us first study the term @xmath286 . from corollaries [ corokernel ] and [ cor : split1 ] , we have ( recall that @xmath282 $ ] ) @xmath287 in the same way , from lemma [ lemmechaleur ] and corollary [ cor : condest ] , we control the term @xmath288 as @xmath289 from lemma [ lemme3 ] and corollary [ cor : condest ] , @xmath290 for the term @xmath291 , write @xmath292 by linearity of the evolution operator @xmath31 , we have @xmath293 hence @xmath294 now from sobolev embedding , lemma [ lemme3 ] and corollary [ cor : condest ] , we get : @xmath295 finally , since @xmath296 then for @xmath297 , @xmath298 and by integration for @xmath299 $ ] , @xmath300 we conclude by applying the modified fractional gronwall lemma [ lemme2 ] .
the proof follows the same lines as in ( * ? ? ?
* section 5 ) . denote by @xmath301 the numerical solution , and @xmath302 which corresponds to the exact evolution of the numerical value @xmath303 at time @xmath304 up to time @xmath305 . from lemma [ estimapriori2 ] ,
there exists @xmath306 such that @xmath307.\ ] ] we prove by induction that there exists @xmath308 such that if @xmath309 , for all @xmath149 with @xmath310 , @xmath311 where @xmath312 .
the above properties are satisfied for @xmath313 .
let @xmath221 , and suppose that the induction assumption is true for @xmath314 .
since @xmath315 and @xmath316 , we estimate @xmath317 for @xmath318 , @xmath319 and proposition [ prop : lipschitz ] yields , along with the induction assumption , @xmath320 which is bounded by @xmath321 if @xmath322 .
up to replacing @xmath62 with @xmath323 , we obtain , for @xmath324 and @xmath325 , @xmath326 using proposition [ localerror ] , we infer @xmath327 for some uniform constant @xmath71 . therefore , @xmath328
which yields the first two estimates of the induction , provided one takes @xmath329 , which is uniform in @xmath112 and @xmath25 .
finally , the last estimate of the induction stems from corollary [ cor : split1 ] .
the aim of this section is to numerically verify the lie method convergence rate in @xmath330 for the fowler equation .
+ to solve the linear sub - equation , discrete fourier transform is used and for the nonlinear sub - equation , different numerical approximations can be used . here
, we use the finite difference method .
+ since the discrete fourier transform plays a key role in these schemes , we briefly review its definition , which can be found in most books . in some situation , when the mesh nodes number @xmath331 is chosen to be @xmath332 for some integer @xmath333 , a fast fourier transform ( fft ) algorithm
is used to further decrease the computation time . in this work
we will use a subroutine implemented in matlab . in this program ,
the interval @xmath334 $ ] is discretized by @xmath331 equidistant points , with spacing @xmath335 .
the spatial grid points are then given by @xmath336 , @xmath337 .
if @xmath338 denotes the approximate solution to @xmath339 , the discrete fourier transform of the sequence @xmath340 is defined by @xmath341 for @xmath342 , and the inverse discrete fourier transform is given by @xmath343 for @xmath344 . here @xmath345 denotes the discrete fourier transform and @xmath346 its inverse . in what follows , the linear equation is solved using the discrete fourier transform and time marching is performed exactly according to @xmath347 to approximate the viscous burgers equation , we use the following explicit centered scheme : @xmath348 + \varepsilon \ , \delta t\frac{u_{j+1}^n-2 u^n_j + u_{j-1}^n } { \delta x^2},\ ] ] which is stable under the cfl - peclet condition @xmath349 where @xmath350 is an average value of @xmath242 in the neighbourhood of @xmath351 ) . in the case where the linear sub - equation is solved using a finite difference scheme instead of a fft computation ,
an additional stability condition is required , see @xcite . moreover
, the computation time becomes very long because of the discretization of the nonlocal term which is approximated using a quadrature rule .
indeed , in @xcite , the fowler equation has been discretized using finite difference method and the numerical analysis showed that this operation is computationally expensive .
this observation has also motivated the use of splitting methods , in particular the implementation of split - step fourier methods . in order to avoid numerical reflections due to boundaries conditions and to justify the use of the fft method , we consider initial data with compact support displayed in figure [ cinitiale ] to perform numerical simulations . since we do not know the exact solution of the fowler equation , a classical numerical way to determine the convergence numerical order of schemes is to plot the logarithm of the error @xmath352 in function of the logarithm of the step time @xmath25 , where @xmath353 and @xmath354 are computed for time steps @xmath355 and @xmath356 , respectively , up to the final time @xmath57 .
hence , the numerical order corresponds to the slope of the curve , see figures [ ordcvglie ] , [ ordcvgstrang ] . for reference , a small line of slope one ( resp .
two ) is added in figure [ ordcvglie ] ( resp . [ ordcvgstrang ] ) .
we see that the slopes for the three initial data match well and so we can conclude that numerical simulations are consistent with the theoretical results established above .
+ we also study numerical convergence of strang splittings using initial data displayed in figure [ cinitiale ] .
results are plotted in figure [ ordcvgstrang ] .
we can see that the strang formulation is of order two in time for smooth initial data .
, _ the lie
trotter splitting for nonlinear evolutionary problems with critical parameters . a compact local error representation and application to nonlinear schrdinger equations in the semi - classical regime _ , i m a j. numer .
2012 ) . to appear | we consider a nonlocal scalar conservation law proposed by andrew c. fowler to describe the dynamics of dunes , and we develop a numerical procedure based on splitting methods to approximate its solutions .
we begin by proving the convergence of the well - known lie formula , which is an approximation of the exact solution of order one in time .
we next use the split - step fourier method to approximate the continuous problem using the fast fourier transform and the finite difference method .
our numerical experiments confirm the theoretical results . |
the general scattering tensor of a system of triclinic symmetry and an arbitrary magnetization is given by @xmath10 which defines a non - symmetric , non - hermitian , complex tensor .
naturally this tensor can be diagonalized to give a complex basis defining the principal axes and a diagonal tensor . in general the principal axes
will be @xmath6 dependent .
even in high symmetries there will already be a strong @xmath6 dependence on the principal axes when a magnetization in an arbitrary direction is introduced .
it is therefore often more intuitive to stay in a cartesian basis for the scattering tensor , with @xmath11 , @xmath12 , and @xmath13 aligned along high symmetry crystal axes .
based on general symmetry arguments the scattering tensor simplifies .
let @xmath14 be the set of symmetry operations present in the paramagnetic phase of a material . in the magnetically ordered phase the local moments order in the @xmath15 $ ] direction .
we will assume that the crystal structure does not change between the paramagnetic and ordered phase .
in the case that the local moment direction does not coincide with one of the high symmetry axes of the paramagnetic phase , only the identity operation is left of the original symmetry operations as all other symmetry operations will rotate the local moment .
one can however define a new set of symmetry operations @xmath16 that rotates first the entire system and then the local moment , i.e. the spin and orbital momentum , back .
this would be the same as rotating the system , but not the local moment . in order to define symmetry operations that rotate the system but not the local moment
, the scattering tensor needs to be written as a product of a function that is independent of the direction of the local moment , but depends on @xmath6 and the polarization and a function that depends purely on the direction of the local moment .
this can be done by expanding each element of the scattering tensor on spherical harmonics in the coordinates of the local moment .
@xmath17 @xmath18 and @xmath19 define the direction of the local moment ( @xmath20 , @xmath21 , and @xmath22 ) , @xmath23 is a spherical harmonic function and @xmath24 is the @xmath25 component of the scattering tensor on a basis of linear polarized light in the coordinate system of the crystal ( @xmath26 ) .
note that both the local magnetization direction as well as the polarization are expressed in the same cartesian coordinate system .
finding the symmetry allowed components of the scattering tensor implies going through all symmetry operations possible and solving the equation @xmath27 .
this will lead to a set of allowed values for the expansion coefficients .
for example a @xmath28 operation acting on @xmath29 i.e. rotating the system , but not rotating the magnetization direction , would give @xmath30 which leads to sets of equations of the form @xmath31 , which have to be solved .
let us first discuss spherical symmetry .
in spherical symmetry there should be no @xmath19 dependence and only terms with @xmath32 remain .
furthermore the infinite sum truncates at @xmath33 due to the triangular equations .
this leaves only the expansion coefficients proportional to @xmath34 , @xmath35 and @xmath36 . for a local moment in the @xmath13 direction ( @xmath37 $ ] ) one
gets the known result : @xmath38}=\left ( \begin{array}{ccc } f^{(0)}-\frac{1}{3}f^{(2 ) } & f^{(1 ) } & 0 \\ -f^{(1 ) } & f^{(0)}-\frac{1}{3}f^{(2 ) } & 0 \\ 0 & 0 & f^{(0)}+\frac{2}{3}f^{(2 ) } \\
\end{array } \right)\ ] ] and for a local moment in the @xmath39 direction : @xmath40}=\left ( \begin{array}{ccc } f^{(0)}+(x^2-\frac{1}{3})f^{(2 ) } & ( z)f^{(1)}+(x y)f^{(2 ) } & -(y)f^{(1)}+(x z)f^{(2 ) } \\
-(z)f^{(1)}+(x y)f^{(2 ) } & f^{(0)}+(y^2-\frac{1}{3})f^{(2 ) } & ( x)f^{(1)}+(y z)f^{(2 ) } \\ ( y)f^{(1)}+(x z)f^{(2 ) } & -(x)f^{(1)}+(y z)f^{(2 ) } & f^{(0)}+(z^2-\frac{1}{3})f^{(2 ) } \\ \end{array } \right)\ ] ] the result for a local moment in an arbitrary direction could also be obtained from the tensor of the local moment in the @xmath13 direction by rotating the scattering tensor .
@xmath41 then @xmath40}=r f_{[001 ] } r^{t}\ ] ] @xmath42 is related to the gyromagnetic vector by @xmath43 and describes the circular dichroism or the faraday effect .
the @xmath44 components describe the linear dichroism . the magneto optical kerr effect is given by both the @xmath42 and @xmath44 spectra depending on the experimental geometry . in symmetries lower than spherical
the expansion of the spin direction in spherical harmonics does not truncate at finite @xmath45 .
this has often been neglected previously , but is quite obvious as angular momentum is not a conserved quantum number in real crystals .
elements like @xmath46 , @xmath47 , etc . are allowed by symmetry .
there is thus in principle an infinite number of fundamental spectra .
not all of them are important and most of them will be very small . below we will discus the higher order expansions in more detail and give several numerical examples for realistic parameters . furthermore , as previously discussed by carra and thole @xcite the fundamental spectra of order @xmath45 will branch according to their symmetry representations in the corresponding point - group . in cubic symmetry ( @xmath48 ) the scattering tensor becomes ( @xmath49 ) : @xmath50}=\\ \nonumber & & \left ( \begin{array}{ccc } f^{(0)}_{a_{1g}}+(x^2\!-\!\frac{1}{3})f^{(2)}_{e_g } & z f^{(1)}_{t_{1u}}+z(z^2\!-\!\frac{3}{5 } ) f^{(3)}_{t_{1u}}+x y f^{(2)}_{t_{2 g } } & -y f^{(1)}_{t_{1u}}-y(y^2\!-\!\frac{3}{5 } ) f^{(3)}_{t_{1u}}+x z f^{(2)}_{t_{2 g } } \\ -z f^{(1)}_{t_{1u}}-z(z^2\!-\!\frac{3}{5 } ) f^{(3)}_{t_{1u}}+x y f^{(2)}_{t_{2 g } } & f^{(0)}_{a_{1g}}+(y^2\!-\!\frac{1}{3})f^{(2)}_{e_g } & x f^{(1)}_{t_{1u}}+x(x^2\!-\!\frac{3}{5})f^{(3)}_{t_{1u}}+y z f^{(2 ) } _ { t_{2 g } } \\ y f^{(1)}_{t_{1u}}+ y(y^2\!-\!\frac{3}{5})f^{(3)}_{t_{1u}}+ x z f^{(2)}_{t_{2 g } } & - x f^{(1)}_{t_{1u}}- x(x^2\!-\!\frac{3}{5 } ) f^{(3)}_{t_{1u}}+ y z f^{(2)}_{t_{2 g } } & f^{(0)}_{a_{1g}}+(z^2-\frac{1}{3})f^{(2)}_{e_g } \\
\end{array } \right)\end{aligned}\ ] ] whereby the expansion series can be continued by summing for the diagonals the @xmath51 and @xmath52 cubic harmonics of order @xmath45 multiplied by a fundamental spectrum and for the off diagonal components the @xmath53 and @xmath54 cubic harmonics .
the important change between cubic and spherical symmetry is that @xmath44 becomes different ( @xmath55 or @xmath56 ) for diagonal and off diagonal elements in @xmath4 .
@xmath56 defines the magnetic linear dichroic spectrum one measures if the sample is magnetized along a @xmath57 direction , whereas @xmath55 defines the magnetic linear dichroic spectrum one measures if the sample is magnetized along a @xmath58 direction .
for very small deviations from spherical symmetry @xmath55 must be roughly equal to @xmath56 , however as we will show below by several numerical examples one finds for real systems that @xmath56 and @xmath55 are very different . on top of
that one finds that for realistic parameters and large local moments the contribution of @xmath59 can not be neglected . in tetragonal symmetry ( @xmath60 )
the scattering tensor becomes ( @xmath61 , c_2 \parallel < 100>$ ] ) .
@xmath50}=\\ \nonumber&&\left ( \begin{array}{ccc } f^{(0)}_{a_{1g}^b}+\frac{1}{2}(x^2\!-\!y^2)f^{(2)}_{b_{1g}}-\frac{1}{2}(z^2\!-\!\frac{1}{3})f^{(2)}_{a_{1g}^b } & z f^{(1)}_{a_{2u}}+z(z^2\!-\!\frac{3}{5 } ) f^{(3)}_{a_{2u}}+x y f^{(2)}_{b_{2 g } } & -yf^{(1)}_{e_u}-y(y^2\!-\!\frac{3}{5 } ) f^{(3)}_{e_u}+x z f^{(2)}_{e_g } \\ -z f^{(1)}_{a_{2u}}-z(z^2\!-\!\frac{3}{5 } ) f^{(3)}_{a_{2u}}+ x y f^{(2)}_{b_{2 g } } & f^{(0)}_{a_{1g}^b}-\frac{1}{2}(x^2\!-\!y^2)f^{(2)}_{b_{1g}}-\frac{1}{2}(z^2\!-\!\frac{1}{3})f^{(2)}_{a_{1g}^b } & x f^{(1)}_{e_u}+x(x^2\!-\!\frac{3}{5 } ) f^{(3)}_{e_u}+ y z f^{(2)}_{e_g } \\
y f^{(1)}_{e_u}+y(y^2\!-\!\frac{3}{5 } ) f^{(3)}_{e_u}+ x z f^{(2)}_{e_g } & - x f^{(1)}_{e_u}-x(x^2\!-\!\frac{3}{5 } ) f^{(3)}_{e_u}+ y z f^{(2)}_{e_g } & f^{(0)}_{a_{1g}^a}+(z^2\!-\!\frac{1}{3})f^{(2)}_{a_{1g}^a } \\
\end{array } \right)\end{aligned}\ ] ] the difference between @xmath62 and @xmath63 defines the natural linear dichroic spectra , also present in a paramagnetic sample .
there are five fundamental spectra up to order @xmath33 that describe the magnetic linear dichroism .
let us give examples how to obtain each one of these spectra . placing the polarization into the @xmath13 direction , the dichroism resulting from changing the magnetization direction from @xmath11 to @xmath13
is described by @xmath64 .
when the polarization is along @xmath65 $ ] and the magnetisation changes from @xmath11 to @xmath13 , @xmath66 describes the magnetic linear dichroism . placing the polarization along @xmath11
, changing the magnetization from @xmath11 to @xmath12 will result in a magnetic linear dichroism determined by @xmath67 .
there are two off - diagonal elements ( @xmath68 and @xmath69 ) that define the dichroism when the magnetization is in the @xmath70 $ ] direction : @xmath68 for the dichroism between polarizations along @xmath65 $ ] and @xmath71 $ ] , @xmath72 for polarizations along @xmath73 $ ] and @xmath74 $ ] .
the isotropic spectrum , i.e. the spectrum measured on a powdered sample , is given by the trace of the conductivity tensor .
the isotropic spectrum on a single crystal is measured by averaging three orthogonal polarization directions . for magnetic systems in cubic symmetry
the trace is independent of the magnetization direction .
the isotropic spectrum of a para - magnet is thus equal to the isotropic spectrum of the magnetically ordered system .
it is interesting to note that the trace of the conductivity tensor in @xmath60 symmetry becomes dependent on the magnetization direction .
this is due to the @xmath56 spectrum in cubic symmetry that branches to an @xmath75 spectrum for @xmath13 polarization and to an @xmath76 and @xmath77 spectrum for @xmath11 or @xmath12 polarization .
this finding does not contradict that the spectrum of a paramagnet only depends on @xmath78 components .
it is also valid that the spectrum of a multidomain sample averaged over all possible spin directions such that the final magnetic symmetry is at least cubic equals the spectrum of a para - magnet .
the dependence of the trace of the conductivity tensor on the spin direction might therefore come as a surprise .
however it is allowed by symmetry , which can be understood with an example . for a @xmath79 system one
has three low energy eigen - states , @xmath80 , @xmath81 or @xmath82 . with these three eigen - states
any spin direction can be created . within cubic symmetry
these three states still belong to the same irreducible representation ( @xmath53 ) and rotating the spin therefore does not change the symmetry of the ground - state . in tetragonal symmetry
the @xmath83 states however have a different symmetry ( @xmath84 ) from the @xmath85 state ( @xmath86 ) .
naively one might expect that this is unimportant as these states will be degenerate as long as spin - orbit coupling for the @xmath87 shell is neglected in the initial state .
when spin - orbit coupling is included one might still expect that the splitting will be smaller than the magnetic field applied or the internal exchange fields present in the sample .
but formally the ground - state symmetry changes depending on the magnetization direction and therefore the isotropic spectra ( the trace of the conductivity tensor ) might have a different line - shape depending on having an @xmath52 ( @xmath83 ) or @xmath86 ( @xmath85 ) ground - state .
the size of these effects will be shown below by numerical calculations . in tetragonal symmetry the antisymmetric part of order @xmath88
can not be represented by a single spectrum .
the circular dichroic spectrum will be different for a magnetization direction parallel or perpendicular to the @xmath89 axes . in scattering experiments
this fundamental spectrum determines the line - shape of the first order bragg reflection of an antiferromagnetically ordered crystal .
it will therefore have a large impact on the interpretation of scattering data .
in orthorhombic symmetry ( @xmath90 ) the scattering tensor becomes ( @xmath91 ) : @xmath50}=\\ \nonumber&&\left ( \begin{array}{ccc } f^{(0)}_{a_g^{xx}}+\!(x^2\!-\!\frac{1}{3})f^{(2)}_{a_g^{xxa}}+\!(y^2\!-\!z^2)f^{(2)}_{a_g^{xxb } } & z f^{(1)}_{b_{1u}}+z(z^2\!-\!\frac{3}{5 } ) f^{(3)}_{b_{1u}}+x y f^{(2)}_{b_{1 g } } & -y f^{(1)}_{b_{2u}}-y(y^2\!-\!\frac{3}{5 } ) f^{(3)}_{b_{2u}}+x z f^{(2)}_{b_{2 g } } \\
-z f^{(1)}_{b_{1u}}-z(z^2\!-\!\frac{3}{5 } ) f^{(3)}_{b_{1u}}+x y f^{(2)}_{b_{1 g } } & f^{(0)}_{a_g^{yy}}+\!(y^2\!-\!\frac{1}{3})f^{(2)}_{a_g^{yya}}+\!(z^2\!-\!x^2)f^{(2)}_{a_g^{yyb } } & x f^{(1)}_{b_{3u}}+x(x^2\!-\!\frac{3}{5 } ) f^{(3)}_{b_{3u}}+y z f^{(2)}_{b_{3 g } } \\ y f^{(1)}_{b_{2u}}+y(y^2\!-\!\frac{3}{5 } ) f^{(3)}_{b_{2u}}+x z f^{(2)}_{b_{2 g } } & -x f^{(1)}_{b_{3u}}-x(x^2\!-\!\frac{3}{5 } ) f^{(3)}_{b_{3u}}+y z f^{(2)}_{b_{3 g } } & f^{(0)}_{a_g^{zz}}+\!(z^2\!-\!\frac{1}{3})f^{(2)}_{a_g^{zza}}+\!(x^2\!-\!y^2)f^{(2)}_{a_g^{zzb } } \\
\end{array } \right)\end{aligned}\ ] ]
the scattering tensor can not be measured directly .
absorption , reflection , scattering or diffraction experiments measure different parts or combinations of the tensor .
an absorption measurement probes the imaginary part of @xmath5 .
@xmath92\ ] ] with @xmath93 the polarization of the light , @xmath94 the conductivity tensor of atom @xmath95 .
the sum is over all atoms in the sample ( neglecting self absorption effects ) and @xmath96 . the real part can then be obtained from a kramers - kronig transformation . in a reflection , scattering or diffraction experiment one measures @xmath97 with @xmath98 the polarization of the incoming ( outgoing ) light , @xmath99 the wave vector of the incoming ( outgoing ) light , @xmath100 the position of atom @xmath95 , @xmath101 the scattering tensor of atom @xmath95 and the sum over all atoms in the sample ( neglecting self absorption effects ) .
the polarization can not be chosen arbitrarily in an diffraction experiment , as the @xmath102-vectors of the light must fulfill the bragg condition .
the polarization has to be perpendicular to @xmath102 , leaving two options for the polarization : @xmath93 can be in the scattering plane ( @xmath103 polarization ) or @xmath93 can be perpendicular to the scattering plane ( @xmath104 polarization ) .
theoretically it is easier to work in cartesian coordinates and describe the material properties independent of the measurement geometry .
we express the magnetization direction , the pointing vectors of the light as well as the polarization in the same coordinate system .
it is thus needed to find an easy expression of @xmath104 and @xmath103 polarization in this coordinate frame .
we will take that @xmath105 , @xmath106 , and @xmath107 , ( the lattice vectors ) and the scattering vector ( @xmath108 $ ] ) are known in real space cartesian coordinates.@xcite in order to define the scattering plane we will need a second vector that might arbitrarily be chosen , but must be perpendicular to @xmath109 , which we will call @xmath110 . then @xmath19 will be the azimuthal angle , defined as the angle between @xmath110 and the scattering plane .
the angle between @xmath109 and @xmath111 will be written as @xmath112 .
( the angle between @xmath109 and @xmath113 then is @xmath114 ) .
this leads to the following definitions for the @xmath102 vectors of the light and @xmath104 and @xmath103 polarization in cartesian coordinates .
@xmath115 the polarization of the incoming ( outgoing ) light can be written in cartesian coordinates as @xmath116 , with @xmath117 and @xmath118 complex numbers such that @xmath119 .
the intensity can be obtained in terms of dot and cross products of the polarization and a unit vector in the local moment direction ( @xmath3 ) .
hannon and blume _ et al . _
@xcite came to the following formula in spherical symmetry . @xmath120 which can be derived by dotting the scattering tensor in spherical symmetry for arbitrary spin direction with @xmath121 and @xmath122 and some algebra . in cubic ( @xmath48 ) symmetry @xmath123
becomes @xmath124 with `` @xmath125 '' the standard vector cross product , `` @xmath126 '' the vector dot product and `` @xmath127 '' stands for vector multiplication per index ( @xmath128 ) .
in tetragonal symmetry ( @xmath60 ) one obtains @xmath129 with @xmath130 a vector with only the @xmath13 component of @xmath3 non - zero , i.e. @xmath131 $ ] .
the orthorhombic ( @xmath90 ) relations are @xmath132
the question that still has to be answered is whether the differences in intensity between spherical and lower symmetries are large enough to be observed .
we therefore calculated a few examples at the transition metal @xmath0 edge using the multiplet crystal field approach .
the parameters for such calculations are well discussed in the literature , @xcite and we adopted those values . in order to show the evolution from spherical to cubic and from cubic to tetragonal symmetry we calculated the spectra for several crystal - field parameters .
we did not include the @xmath134 spin - orbit coupling nor covalency ( crystal - field theory instead of ligand - field theory ) in order to make the interpretation easier . while we did not include spin orbit coupling on the @xmath87 shell the magnetization direction directly coincides with the spin direction .
the spectra are calculated at 0 k and assume a full magnetization .
the calculations have been done with the full multiplet ligand field theory program xtls 8.3 .
@xcite the fundamental spectra can be obtained by doing calculations for specific spin orientations . on top of those calculations
a basic check has been made where the optical conductivity tensor has been calculated for many different magnetization directions .
the result is expanded on spherical harmonics in coordinates of the magnetization direction .
both methods gave the same results within the numerical accuracy of the calculations .
in the top panel of fig . [ ohmnni ] we show the @xmath0 edge x - ray absorption spectra of mn@xmath133 as a function of the cubic splitting parametrized by @xmath135 .
the top spectrum has @xmath136 and is therefore in spherical symmetry .
the ground - state has always 5 electrons with parallel spin occupied and the ground - state wave function is independent of the crystal - field splitting .
all changes in the spectra are due to final - state effects , whereby the @xmath137 core hole is excited into a @xmath54 or @xmath52 electron and the energy difference between the two changes as a function of @xmath135 .
the left panel shows @xmath138 , which is the isotropic spectrum and the spectrum of the para - magnetic phase .
these spectra are the same as published earlier by de groot _
the middle column shows @xmath56 and @xmath55 .
these two spectra describe the magnetic linear dichroic effect . in spherical symmetry ( top curve , @xmath136 )
these two spectra are the same .
however when a cubic distortion is present these two spectra become different . for
a commensurate spin spiral one would expect a different resonant profile at the second order bragg peak depending on the phase of the spiral with respect to the cubic crystal structure .
i.e. a spiral with period @xmath139 and the spins in the @xmath11 , @xmath12 , @xmath140 , @xmath141 direction would show an @xmath142 like second order resonance profile , whereas a spiral at the same @xmath143 vector , but shifted in phase such that the spin directions are @xmath144 , @xmath145 , @xmath146 , @xmath147 would show an @xmath148 like second order resonance profile . for x - ray absorption
this has implications for the measured magnetic linear dichroism as explained previous by arenholz and van der laan _ _
et al.__@xcite .
the right column shows the @xmath149 and @xmath59 spectra . in @xmath48 symmetry the @xmath88 spectra branch to a single representation ( a @xmath53 ) .
additionally angular momentum ( @xmath45 ) is not a good quantum number in cubic symmetry and therefore the @xmath59 spectra becomes non - zero . for spectroscopy
this means a difference in the magnetic circular dichroism spectra for a system magnetized in the [ 001 ] direction where one measures @xmath150 $ ] and the [ 111 ] direction where the measured spectrum is @xmath151 $ ] . for resonant diffraction at an anti - ferromagnetic bragg reflection
this will lead to a different resonance profile depending if the spins are ordered in the [ 001 ] or [ 111 ] direction .
it will also lead to a different azimuthal dependence of the scattered intensity at an anti - ferromagnetic bragg reflection . for spins oriented in the [ @xmath152 direction ferromagnetically aligned in ( 111 ) planes ,
antiferromagnetically stacked in the [ @xmath153 direction and a scattering geometry such that @xmath154 one finds a scattered intensity proportional to @xmath155 in the @xmath156-@xmath5 ( @xmath5-@xmath156 ) scattering channel . for different spin alignments the @xmath19 independent scattering due to the @xmath42 scattering tensor
would be related to the projection of the spin on the @xmath109 vector .
measurements at a single resonant energy could thus lead to incorrect determination of the spin direction .
however @xmath46 has a different @xmath6 dependence than @xmath42 and this can be used to obtain the correct spin direction from such experiments . the bottom panel of fig .
[ ohmnni ] shows the same spectra , for ni@xmath133 .
the major difference between a @xmath157 system ( ni@xmath133 ) and a @xmath158 system ( mn@xmath133 ) is that for a @xmath157 system the ground - state is modified when a cubic crystal - field is included . in spherical symmetry both the @xmath52 and @xmath54 orbitals
are equally occupied , in cubic symmetry this is not the case .
when @xmath136 and a spherical charge distribution is assumed for the initial state one finds that the @xmath56 and @xmath55 spectra are equivalent ( middle panel top curve labeled as @xmath4 ) . if one however assumes a @xmath159 orbital occupation as one would find as a ground - state in cubic symmetry one finds a large difference between the @xmath56 and @xmath55 spectra ( middle panel second curve from top labeled as @xmath160 ) .
it is important to note the difference between mn@xmath133 and ni@xmath133 . for a perturbation that does not break the ground - state symmetry the changes in spectra scale with the size of the distortion , whereas for a perturbation that breaks the ground - state symmetry there will be an immediate large effect on the spectral line - shape as long as the distortion is bigger than temperature .
the symmetry breaking is important for orbitally ordered systems . above the ordering temperature
, the system can be described by the scattering tensor in the higher symmetry .
although the distortions may be small in the orbitally ordered phase , breaking the ground state symmetry will have an immediate effect on the spectra .
thus even small distortions lead to relevant changes in the scattering tensor .
in tetragonal symmetry the @xmath162 spectrum branches to @xmath62 and @xmath63 , the @xmath42 spectrum branches to @xmath163 and @xmath164 , and the @xmath44 spectrum branches to @xmath165 , @xmath166 , @xmath77 , @xmath75 and @xmath76 . to estimate the numerical differences between these spectra at the @xmath0 edge we again did crystal field calculations . the top panel of fig .
[ d4hnimncu ] shows the spectra for ni@xmath133 as a function of tetragonal distortion . in the middle
we show cubic symmetry and towards the top ( bottom ) a tetragonal contraction ( elongation )
. the ni@xmath133 ground - state does not change as a function of tetragonal distortion and therefore the difference between spectra that branch from the same irreducible representation in cubic symmetry scale with the size of the tetragonal distortion .
the left panel shows the natural linear dichroism in tetragonal distorted ni@xmath133 which has been discussed before .
@xcite the second panel from the left shows the linear dichroism that branches from the cubic @xmath52 representation .
interestingly , _ three _ fundamental spectra represent the linear dichroism that branches from the @xmath52 representation in cubic symmetry and for feasible distortions they are all different .
the trace of the conductivity tensor , or the isotropic spectrum , is thus dependent on the magnetization direction and the difference is large enough to be detectable .
the isotropic spectrum in @xmath60 symmetry is given by @xmath167 and @xmath168 is , as clearly can be seen in fig .
[ d4hnimncu ] , non zero .
the linear dichroism that branches from the cubic @xmath54 representation is shown in the middle panel .
the second panel from the right shows the circular dichroic spectra .
differences in the antisymmetric part of the scattering tensor are almost always neglected , but for reasonable distortions , present in e.g. layered perovskite structures large differences can be seen between the @xmath169 and @xmath170 spectra . going beyond @xmath171 symmetry thus also affects the antisymmetric part of @xmath4 . the middle panel of fig .
[ d4hnimncu ] shows the spectra of mn@xmath161 as a function of tetragonal distortion .
mn@xmath161 has a @xmath172 configuration and depending on the distortion it will have an occupied @xmath173 orbital ( top ) or an occupied @xmath174 orbital . as in the case of ni@xmath133 , for the spherical to cubic distortion , one has an immediate effect for all fundamental spectra and in particular also in the antisymmetric part of the scattering tensor .
this is important for magnetic circular dichroism and rxd in all kinds of manganates where orbital order and magnetic order co - exists . the bottom panel of fig .
[ d4hnimncu ] shows the fundamental spectra of the scattering tensor for cu@xmath133 a @xmath175 configuration .
this is a particularly simple and instructive example as all calculations can be done by hand and everything can be done in a one particle picture .
the ground - state has one hole in the @xmath87 shell and the final state one hole in the @xmath137 shell .
the ground - state hole in the @xmath87-shell will either be in the @xmath176 orbital ( bottom of fig .
[ d4hnimncu ] ) or in the @xmath177 orbital ( top of fig .
[ d4hnimncu ] ) . for the natural dichroism ( left panel ,
@xmath178 and @xmath179 ) one finds that for an @xmath173 hole there is no intensity in the @xmath179 as one can not excite into the @xmath173 electron with @xmath13 polarized light . for an @xmath174 hole
there is a ratio of 1 to 3 for the different fundamental spectra .
note that the spectra are independent of the size of the distortion and only depend on the symmetry of the ground - state .
the symmetric part of the magnetic scattering tensor is zero for a @xmath175 configuration , which is a general property of any kramers doublet .
the spin - up and spin - down state are related by a mirror symmetry and therefore should have the same symmetric part of the scattering tensor .
at the same time the sum of the scattering tensor for the spin - up and spin - down state should be equal to the paramagnetic tensor which is given by the @xmath162 components .
therefore the symmetric magnetic part of the scattering tensor should be zero for a ground - state kramers doublet .
the antisymmetric part shows a very strong crystal orientation dependence . for a @xmath173 hole in the ground - state ,
the @xmath169 spectra is finite , but the @xmath170 spectrum is zero .
this is again related to the fact that one can not excite into the @xmath173 hole with @xmath13 polarized light .
for the cuprates , the rxd spectra are thus only sensitive to the out - of - plane moment .
we have shown that the fundamental spectra of the scattering or conductivity tensor are strongly influenced by the crystal field symmetry at the transition metal @xmath0 edge .
both the symmetric part as noted previously , @xcite but also the antisymmetric part show strong deviations from the spectra calculated in spherical symmetry . when the symmetry is lowered the fundamental spectra known from spherical symmetry branch into several different spectra . at the same time
higher order spectra become non - zero and are important . in general more fundamental spectra are needed in order to describe the conductivity tensor as one might have expected .
we have shown to what extent the crystal symmetry effects the magneto - optical effects and how they influence rxd and xas experiments .
the reason for the deviations from the simple rule that the scattered intensity at an antiferromagnetic bragg reflection is proportional to @xmath1 is given by the fact that in spherical symmetry there is only one tensor element that describes the antisymmetric part of the conductivity tensor . in cubic symmetry
this already becomes more complicated as @xmath46 can not be neglected if @xmath180 is larger then 1 .
in tetragonal or lower symmetry also @xmath42 branches to different spectra .
when the spin aligns in a high symmetry direction of the crystal there still will be only one tensor element describing the antisymmetric part of @xmath4 and therefore the simple scattering rules as derived in spherical symmetry will still hold . even when this tensor element is made from a linear combination of @xmath42 and @xmath46
; as long as their ratio is the same for all polarizations the simple rules hold . in most cases
the , single ion anisotropy will dictate the preferred direction of the magnetic moments , aligning it along high symmetry directions . in this case
the simple rules are valid .
local magnetic moments are not always aligned in a high - symmetry crystal direction .
well known examples are the simple rock - salts like nio , mno and coo .
an other class would be magnetic spiral structures as found in the cuprates and manganates . in these cases
a more involved spectral line shape analysis is needed in order to determine the magnetization direction .
one then can not measure the scattered intensity at a single photon energy , but needs to measure the energy dependence at all azimuthal directions in order to obtain more information . by comparing the line - shape to theory at different azimuthal directions or by the use of sum - rules one then can still obtain the full information as present in the data .
the present calculations could be proved by magnetic circular and linear dichroism experiments .
temperature dependence of the magneto - optical effects as well as non - full magnetization of the sample will be discussed in a future work .
the research in kln is supported by the deutsche forschungsgemeinschaft through grant no .
sfb 608 and the cologne bonn graduate school .
j. p. hannon , g. t. trammell , m. blume , and doon gibbs , phys .
lett . * 61 * , 1245 ( 1988 ) .
m. blume , j. appl .
* 57 * , 3615 ( 1985 ) .
p. m. platzman and n. tzoar , phys .
b * 2 * , 3556 ( 1970 ) .
f. debergevin and m. brunel , acta crystallogr .
a * 37 * , 314 ( 1981 ) .
m. blume and doon gibbs , phys .
b * 37 * , 1779 ( 1988 ) .
s. w. lovesey , j. phys .
c * 20 * , 5625 ( 1987 ) .
p. rennert , phys .
b * 48 * , 13559 ( 1993 ) . j. p. hill and d. f. mcmorrow , acta cryst . *
a52 * , 236 ( 1996 ) .
see hannon _ et al .
_ @xcite and articles citing this paper .
p. carra and b. t. thole , rev .
66 * , 1509 ( 1994 ) . s. di matteo , y. joly , a. bombardi , l. paolasini , f. de bergevin , and c. r. natoli , phys .
lett . * 25 * , 257402 ( 2003 ) .
r. g. newton , am .
* 44 * , 639 ( 1976 ) .
e. arenholz , g. van der laan , r. v. chopdekar , and y. suzuki , phys .
b * 74 * , 094407 ( 2006 )
. e. arenholz , g. van der laan , r. v. chopdekar , and y. suzuki , phys .
lett . * 98 * , 197201 ( 2007 ) .
g. van der laan , e. arenholz , r. v. chopdekar , and y. suzuki , phys .
b * 77 * , 064407 ( 2008 ) .
j. kunes and p. m. oppeneer , phys .
b * 67 * , 024431 ( 2003 ) .
p. m. oppeneer , t. kraft , and h. eschrig , phys .
b * 52 * , 3577 ( 1995 ) .
a. uspenskii , e. t. kulatov , and s. v. halilov , phys .
b * 54 * , 474 ( 1996 ) .
s. uba , a. n. yaresko , l. uba , a. ya .
perlov , v. n. antonov , r. gontarz , and h. ebert , phys . rev .
b * 57 * , 1534 ( 1998 ) . within this paper
the notation @xmath181 $ ] is used to define a vector in real space cartesian coordinates .
the vector @xmath181 $ ] is thus not necessarily equal to the vector @xmath182+@xmath183+@xmath184 with @xmath105 , @xmath106 , and @xmath107 the primitive vectors spanning the bravis lattice . | a general symmetry analysis of the optical conductivity or scattering tensor is used to rewrite the conductivity tensor as a sum of fundamental spectra multiplied by simple functions depending on the local magnetization direction . using this formalism
, we present several numerical examples at the transition metal @xmath0 edge . from these numerical calculations
we can conclude that large deviations from the magneto - optical effects in spherical symmetry are found .
these findings are in particular important for resonant x - ray diffraction experiments where the polarization dependence and azimuthal dependence of the scattered bragg intensity is used to determine the local ordered magnetization direction .
resonant x - ray diffraction or reflectivity ( rxd ) has developed into a powerful method to study charge , orbital and magnetic ordering in transition metal compounds and artificially created superlattices . in principle
the resonant energy dependence of the scattering profile and its polarization and azimuthal dependence contain the information about the local magnetic , charge , and orbital order .
obtaining this information from such spectra is often still a theoretical and experimental challenge .
one of the more straightforward methods to analyze magnetic bragg reflections has been presented by hannon _
_ et al.__@xcite and several other authors .
@xcite they showed that the azimuthal intensity dependence can be related to the local orientation of the magnetic moment by a simple relation .
the scattered intensity is proportional to @xmath1 , whereby @xmath2 is the polarization of the incoming ( outgoing ) light and @xmath3 is a unit vector in the direction of the magnetization .
these relations are used relatively often in the analyses of rxd nowadays .
@xcite the crux is that these relations are derived in spherical symmetry and strictly speaking hold only in spherical symmetry .
an extended formalism has been presented , @xcite but not seen a large audience so far .
the question the present paper tries to answer is : how large are the changes of the scattered intensity when the effects of the real crystal symmetry is included and when is it important to include the real crystal symmetry .
in other words we answer the question when one can use the formulas derived in spherical symmetry and when the true crystal symmetry has to be included .
we will focus on purely magnetic bragg reflections and neglect spin - orbit coupling for the initial or ground - state when doing the numerical calculations .
a hint for the importance of including the correct crystal symmetry might be obtained from absorption spectroscopy . by the optical theorem @xcite it
is well known that the scattering tensor @xmath4 and the conductivity tensor @xmath5 are related by a factor of @xmath6 ( @xmath7 ) .
therefore all magneto - optical effects known from absorption spectroscopy should return in diffraction or reflectivity experiments .
recently it has been realized that the symmetric part of the conductivity tensor due to magnetism i.e. the magnetic linear dichroism effect is not given by a single spectrum as in spherical symmetry , but behaves more complicated in cubic or lower symmetries@xcite .
for an antiferromagnetic bragg reflection the symmetric part of the scattering tensor does not contribute to the diffracted intensity .
how the antisymmetric part of the conductivity tensor behaves for less than spherical symmetry is not well known in the x - ray regime .
for infra - red and visible wave - lengths there are several experiments and numerical calculations showing that for systems with less than cubic symmetry , the magnetic part of the optical conductivity tensor @xmath5 becomes considerably different from what is expected in spherical symmetry .
@xcite one might thus expect that also for x - ray wavelengths the spectral line - shape of circular dichroism and magnetic bragg reflections should be crucially influenced by the crystal symmetry .
in the present paper , we show several calculations at the transition metal @xmath0 edge in order to exemplify when deviations from spherical symmetry become important .
to present the results in a consistent picture , we first introduce a symmetry analysis of the magnetization direction dependence of the conductivity tensor or scattering tensor .
this allows us to define a few fundamental spectra that describe the entire optical response of a system independent of the magnetization direction .
we will assume a certain symmetry in the paramagnetic phase and in the ordered phase an additional , symmetry breaking sub - lattice magnetization in an arbitrary direction , given by @xmath8 $ ] .
@xcite the tensor formalism used is in principle valid for any type of magneto - optical experiment and closely related , though not entirely equivalent , to previous publications .
@xcite the formalism can be used for the description of the crystal orientation dependence of the kerr angle in the infra - red regime on cro@xmath9 @xcite or the magnetic bragg diffraction of the cuprates at 930 - 950 ev .
the first section presents a general derivation of the scattering tensor or conductivity tensor as a function of magnetization direction for several different point - group symmetries . in the second section
we will show how these tensors can be rewritten to simple dot products relating the scattered intensity to the polarization of the incoming and outgoing light in the same way as previously presented@xcite in spherical symmetry .
the third section is reserved for calculations of the fundamental spectra of several example materials in cubic point - group symmetry ( the symmetry relates to the paramagnetic phase ) .
the fourth section presents results in tetragonal symmetry . within the conclusions we will discuss when deviation from spherical symmetry becomes visible in experiment and how the magnetization direction can be deduced from resonant x - ray diffraction experiments |
in a wide variety of galactic environments , polycyclic aromatic hydrocarbons ( pahs ) li 2007b ) . ] are the dominant emitters in the mid - infrared ( mir ) from about @xmath25 - 15 @xmath1m@xmath7 ( vogler et al .
2005 , smith et al .
2007 , galliano et al .
2008b ) . at longer wavelengths approaching the far - infrared ( fir ) ,
thermal emission dominates the spectral energy distribution ( sed ) ( galliano et al .
pahs have been observed in many galaxies in locations often associated with star forming regions .
however , new discoveries of pah emission extending to significant distances ( 6 - 10 kpc ) from the disks of several galaxies ( irwin & madden 2006 ; engelbracht et al . 2006 ; irwin et al . 2007 )
raise important questions as to how these large molecules can be present at such high altitudes above the plane and also their source of excitation .
presumably , far - uv ( fuv ) emission from in - disk star forming regions provides the excitation .
however , a fit to a sed is really required to understand both the contributions from various emission components ( cool stars , pahs , classical grains , etc . ) as well as the stellar component that is required for the excitation .
better yet , a sed fit at different locations in a galaxy , especially at different locations in the disk and/or in extraplanar gas , can further help us to understand the complex interplay between various ir - emitting interstellar medium ( ism ) components and the origin of the excitation .
we have therefore chosen to study mir - submm emission in a very well known edge - on galaxy , ngc 891 ( figure [ fig : n891_wboxes ] ) , that has known extraplanar emission in a variety of wavebands ( e.g. h i , h@xmath8 , dust and co - see sect . [
sec : n891 ] ) .
since there is so much spatially resolved ir data available for this galaxy , it is an excellent target for modeling its sed at different positions in both the disk and halo regions . for this paper , we have examined and compared no less than 20 independent data sets @xmath1 m image was also examined , but found to be of too low spatial resolution to be useful . ]
ranging in wavelength from @xmath21.2 @xmath1 m to 850 @xmath1 m using both space - based and ground - based instruments ( sect .
[ sec : obs_datared ] ) .
to the limits of the variety of spatial resolution , coverage , and sensitivity achieved , we compare the extraplanar extent of the emission and , most importantly , we derive the sed at different disk and extraplanar locations . because of the proximity of ngc 891 , its almost exactly edge - on orientation , its current active star formation ( sf ) and its prominent extraplanar emission ( see sect . [
sec : n891 ] ) , this galaxy has been a prime target for studies of the disk - halo interaction . although the specific vertical height , @xmath4 , at which gas might be considered in the halo is debatable , a conservative estimate is 1 kpc ( 21.5 arcsec ) . since all of the 20 data sets examined have higher resolution than 21.5 arcsec ( see sect .
[ sec : obs_datared ] ) , it is possible to distinguish the disk from extraplanar ( halo ) region in the observations presented in this paper . in the next section
we describe ngc 891 . in sect .
[ sec : obs_datared ] , we discuss the observations and data reduction .
the morphology of ngc 891 is discussed in sect .
[ sec : results ] , and the sed results are given in sect . [
sec : sed ] . sects .
[ sec : discussion ] and [ sec : conclusion ] present the discussion and conclusion , respectively .
ngc 891 ( figure [ fig : n891_wboxes ] , table [ tab : galaxy_params ] ) is believed to be similar to the milky way ( mw ) in b - band luminosity , rotational velocity , and hubble type .
it is 2.5 times brighter in the ir , and may have twice as much gas and dust than the mw ( see scoville et al .
1993 ) , although the difference is not strong , given uncertainties .
there has been much discussion in the literature as to whether or not ngc 891 is a ` starburst ' galaxy , since its star formation rate ( sfr ) appears to be intermediate between that of a normal spiral and a starburst galaxy ; it may be in a phase that saw more powerful starbursting in the past ( see temple et al .
ngc 891 s current sf activity is clearly indicated by a variety of measures , including the detection , in h@xmath8 , of many bubbles , shells , and supershells in the mid - plane of this galaxy ( rossa et al .
2004 ) .
.basic galaxy parameters of ngc 891 [ cols="<,<,<",options="header " , ] an important result of this work is our first attempt at a sed model in the extraplanar ( halo ) region of an external galaxy ( table [ tab : model_params ] ) .
the extra - planar dust mass results are consistent with those determined from extra - planar absorption ( howk @xmath9 savage 1997 ) , after reasonable corrections for different apertures are taken into account .
the @xmath5 ratios in the halo are similar to those in the disk .
this result was unexpected given the observations of the vertical profiles in figure [ fig:77onk450_24 ] , which show that the cool dust ( at @xmath2450 @xmath1 m , for example ) extends further into the halo vertically than the pahs ( at @xmath27.7 @xmath1 m , for example ) .
however , the error bars are very large for these halo results , and the sed model does not take into account the variation of the stellar populations from the disk to the halo
. a decrease of the hardness of the isrf outward , would decrease the pah strength , for a constant pah - to - dust mass ratio , which is what we see here .
the quantity , @xmath6 , which is close to 1 in the halo , is consistent with an interpretation in which few hot photons are available for excitation . while ngc 891 is known for large amounts of neutral gas , dig , and dust in its halo ( sect .
[ sec : intro ] ) , its pah emission is not as extended in @xmath4 as has been seen in several other galaxies which are _ not _ known for abundant halo gas and dust ( see irwin @xmath9 madden 2006 , irwin et al .
2007 ) . since the excitation source for the pahs appears to be within the disk ( see above ) , we speculate that fuv photons that could potentially excite high latitude pah emission in ngc 891 may instead be absorbed by other gaseous and dust components .
that is , other components in the halo of ngc 891 have some optical depth to fuv photons that originate in the disk .
we have examined 20 spatially resolved ir data sets for ngc 891 , 14 of which are newly reduced and/or newly presented data . the emission spans wavelengths from @xmath21.2 to 850 @xmath1 m .
although there are a variety of resolutions and s / n for the different data sets , which include both ground - based and space - based observations , we have compared the spatial distributions , both along and perpendicular to the major axis for a subset of the data .
pah emission is found to correlate well in the disk with molecular gas , likely because of higher densities of gas and dust in these regions as well as the abundant supply of uv photons from sf regions . of particular interest
is the extraplanar , or ` halo ' gas and dust in ngc 891 .
the h i , dig , and dusty halos have been established by others .
our work has compared the @xmath4 extents of various components and also shown that pahs also exist in the halo of this galaxy to heights of 2.3 @xmath10 0.2 kpc at a 3@xmath11 level .
ngc 891 is only the 4th galaxy in which pahs have been detected in the halo .
we find that cool halo gas , as measured by @xmath2450 @xmath1 m emission , is broader in @xmath4 than both warm halo gas @xmath224 @xmath1 m and pah emission , suggesting that the dust heating sources and pah excitation sources are within the disk .
we have used our multi - frequency data set to construct complete ir spectra for four different locations in the disk of ngc 891 and two in its halo .
this has allowed us to model seds in different environments both within the disk of ngc 891 and , for the first time , in the halo .
the sed models have allowed us to determine the mass of dust , fraction of pahs , and fraction of ionized pahs for these different environments ( table [ tab : model_params ] ) .
we find that the pah fraction , @xmath5 , is similar to galactic values ( within a factor of two ) and this value is lowest in the galaxy s center , consistent with previous studies of pahs in galaxies .
the fraction of dust exposed to a colder ( i.e. equivalent to an equilibrium dust temperature less than 25 k ) radiation field , @xmath6 , is also lowest in the center , suggesting the presence of a hotter radiation field at the center of the galaxy . in the extraplanar dust
, @xmath5 is similar to @xmath5 in the disk .
therefore we can not conclude a lower pah abundance , especially since the sed model used does not take into account a changing radiation field that is more typical of a halo .
the model used could take into account the changing radiation field , but we would have needed additional constraints .
the fraction of pahs exposed to a cold radiation field is very high in the halo and is almost unity .
together with the vertical halo extents indicated above , these results suggest that the halo pahs are excited from escaping disk photons .
we thank dr . r. swaters for supplying us with the ngc 891 h
i fits file , dr .
p. alton for the scuba fits files .
this publication makes use of the iso and spitzer data archives from the european space agency and the national aeronautics and space administration ( nasa ) , respectively .
it also 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 / california institute of technology , funded by nasa and the national science foundation .
99 alton p. , bianchi s. , rand r. , xilouris e. , davies j. , trewhella m. 1998 , apj 507 , l125 alton p. , rand r. , xilouris e. , bevan s. , ferguson a. , davies j. , bianchi s. 2000 , a@xmath9as 145 , 83 baldwin j. , pooley g. 1973 , mnras , 161 , 127 bendo g. et al .
2003 , apj , 125 , 2361 bendo g. et al .
2008 , mnras , 389 , 629 biviano a. 1998 , the isocam calibration error budget rep . , ver 3.1 blommaert j. et al .
2003 , the iso handbook : cam - the iso camera ( v 2.0 ) , esa sp-1262 , volume ii burgdorf m. , ashby m. l. n. , & williams r. 2007 , apj , 668 , 918 chanial p. 2003 phd thesis , university of paris cohen m. , wheaton w. , megeath s. 2003 , aj 126 , 1090 coia d. et al .
2005 , a@xmath9a , 430 , 59 dahlem m. , dettmar r - j . , hummel e. 1994 , a@xmath9a , 290 , 384 dale d. , helou g. , contursi a. , silbermann n. , kolhatkar s. 2001 , apj , 549 , 215 dale d. et al .
2005 , apj , 633 , 857 de vaucouleurs g. , de vaucouleurs a. , corwin h. , buta r. , paturel g. , fouque p. 1991
, third reference catalogue of bright galaxies , vols 1 - 3 .
springer - verlag , berlin , heidelberg , new york draine b. @xmath9 li a. 2007a , apj , 657 , 810 draine b. et al .
2007b , apj , 663 , 866 dupac x. et al .
2003 , mnras , 344 , 105 dunne l. @xmath9 eales s. 2001 , mnras , 327 , 697 dwek e. et al .
1997 , apj , 475 , 565 engelbracht c. w. et al .
2006 , 642 , l127 fioc m. @xmath9 rocca - volmerange b. 1997 , a@xmath9a , 326 , 950 fouks b. , schubert j. 1995 , proc .
spie , 2475 , 487 fraternali r. , oosterloo t. , sancisi r. , swaters r. 2005 , asp conference series , 331 , 239 449 galliano f. 2004 , phd thesis , university of paris xi galliano f. , dwek e. , chanial p. 2008a
, apj , 672 , 214 galliano f. , madden s. , tielens a. , peeters e. , jones a. 2008b , apj , 679 , 310 garcia - burillo s. , guelin m. , cernicharo j. , dahlem m. 1992 , a@xmath9a , 266 , 21 gil de paz a. , et al .
2007 , apjs , 173 , 185 gordon k. d. et al .
2005 , pasp , 117 , 50 guelin m. , zylka r. , mezger p. , haslam c. , kreysa e. , lemke r. , sievers a. 1993 , a@xmath9a , 279 , l37 haas m. , klaas u. , bianchi s. 2002 , a@xmath9a , 385 , l23 hoopes c. g. et al .
2005 , apj , 619 , 99 howk c. @xmath9 savage b. 1997 , aj , 114 , 2463 howk c. @xmath9 savage b. 2000 , aj , 119 , 644 irwin j. , madden s. 2006 , a@xmath9a , 445 , 123 irwin j. , kennedy h. , parkin t. , madden s. 2007 , a@xmath9a , 474 , 461 israel f. , van der werf p. , tilanus r. 1999 , a@xmath9a , 344 , l83 jarrett t. , chester t. , cutri r. , schneider s. , huchra j. 2003 , aj , 125 , 525 kamphuis p. , holwerda b. , allen r. , peletier r. , van der kruit p. 2007 , a@xmath9a , 471 , l1 le coupanec p.
the universe as seen by iso .
p. cox @xmath9 m. f. kessler .
esa - sp 427 , p. 917
mathis j. , mezger p. , panagia n. 1983 , a@xmath9a , 128 , 212 mattila k. , lehtinen k. , lemke d. 1999 , a@xmath9a , 342 , 643 oosterloo t. , fraternali f. , sancisi r. 2007 , aj , 134 , 1019 otte b. , reynolds r. , gallagher j. , ferguson a. 2001 , apj , 560 , 207 pagani l. et al .
2003 , a@xmath9a , 406 , l59 popescu c. , & tuffs r. 2003 , a@xmath9a , 410 , l21 popescu c. , tuffs r. j. , kylafis n. d. , & madore , b. f. 2004 , a@xmath9a , 414 , 45 rand r. j. , wood k. , & benjamin r. a. 2008 , apj , 680 , 263 reach w. , et al . 2005 , pasp , 117 , 978 regan m. et al .
2006 , apj , 652 , 1112 rossa j. @xmath9 dettmar r .- j . 2003 , a@xmath9a , 406 , 493 rossa j. , dettmar r. , walterbos r. , norman c. 2004 , aj , 128 , 674 scoville n. , thakkar d. , carlstrom j. , sargent a. 1993 , apj , 404 , l59 smith j. et al . 2007 , apj , 656 , 770 sofue y. , @xmath9 nakai n. 1993 , pasj , 45 , 139 sofue y. 1997 , pasj , 45 , 139 strickland d. , hekcman t. , colbert e. , hoppes c. , weaver k. 2004 , apj , 606 , 829 swaters r. , sancisi r. , van der hulst j. 1997 , apj , 491 , 140 temple r. , raychaudhury s. , stevens i. 2005 , mnras , 362 , 581 tilanus r. 2005 http://www.jach.hawaii.edu/jcmt/continuum/calibration/sens/gains.html vogler a. , madden s. , beck r. , lundgren a. , sauvage m. , vigroux l. , ehle m. 2005 , a@xmath9a , 441 , 491 zubko v. , dwek e. , arendt r. 2004 , apjs , 152 , 211 | we present a multi - wavlength infrared study of the nearby , edge - on , spiral galaxy ngc 891 .
we have examined 20 independent , spatially resolved ir images of this galaxy , 14 of which are newly reduced and/or previously unpublished images .
these images span a wavelength regime from @xmath0 @xmath1 m in which the emission is dominated by cool stars , through the mir , in which emission is dominated by pahs , to @xmath2850 @xmath1 m , in which emission is dominated by cold dust in thermal equilibrium with the radiation field .
the changing morphology of the galaxy with wavelength illustrates the changing dominant components .
we detect extra - planar dust emission in this galaxy , consistent with previously published results , but now show that pah emission is also in the halo , to a vertical distance of @xmath32.5 kpc .
we compare the vertical extents of various components and find that the pahs ( from @xmath27.7 and 8 @xmath1 m data ) and warm dust ( @xmath224 @xmath1 m ) extend to smaller @xmath4 heights than the cool dust ( @xmath2450 @xmath1 m ) . for six locations in the galaxy for which the s / n was sufficient , we present seds of the ir emission , including two in the halo the first time a halo sed in an external galaxy has been presented .
we have modeled these seds and find that the pah fraction , @xmath5 , is similar to galactic values ( within a factor of two ) , with the lowest value at the galaxy s center , consistent with independent results of other galaxies . in the halo environment ,
the fraction of dust exposed to a colder radiation field , @xmath6 , is of order unity , consistent with an environment in which there is no star formation .
the source of excitation is likely from photons escaping from the disk .
[ firstpage ] ism : molecules - galaxies : haloes - galaxies : ism - infrared : ism |
bilayer fluid membranes@xcite spontaneously self assemble when aliphatic molecules are dissolved in water at a sufficiently high concentration . at high temperature ,
these membranes have no internal order and can be modeled as fluctuating structureless surfaces characterized by a bare bending rigidity @xmath0 .
the bending rigidity is length - scale dependent and becomes zero at the persistence length @xmath1 where @xmath2 is a molecular length and @xmath3 is the temperature . at length scales
less than of order @xmath4 , the membrane is flat ; at longer length scales , it is crumpled .
a flat rigid membrane can have quasi - long - range ( qlr ) hexatic order @xcite at low temperature and undergo a kosterlitz - thouless ( kt ) disclination unbinding transition @xcite to a disordered high - temperature phase .
a fluctuating membrane can also have qlr hexatic order @xcite .
hexatic order stiffens the bending rigidity so that , rather than scaling to zero at long length scales , it approaches a constant times the hexatic rigidity @xmath5 @xcite .
the hexatic membrane is thus more rigid than a fluid membrane , and it is said to be crinkled " rather than crumpled .
a fluctuating hexatic membrane can undergo a kt transition from the crinkled to the crumpled state .
reference discussed two possible mechanisms for the crinkled - to - crumpled transition : disclination melting and crumpling .
the latter mechanism is analogous to that producing the flat - to - crumpled transition in tethered membranes@xcite and is argued to be associated with the buckling instability@xcite of a membrane with a single disclination .
figure [ gtfig ] shows a schematic phase - flow diagram in the @xmath6-@xmath7 plane ( @xmath8 ) for a @xmath9 membrane embedded in three dimensions adapted from ref . .
the vertical line at @xmath10 is the kosterlitz - thouless disclination unbinding line of a flat membrane .
the curved line joining the vertical line at @xmath11 is an estimate of the crumpling transition obtained by equating the energy of a single positive disclination in a buckled membrane to its entropy .
thus , the crumpling transition in this estimate is a kosterlitz - thouless transition in a buckled membrane .
this schematic phase diagram describes qualitatively features that are in agreement with simple physical reasoning : for large @xmath0 , there should be a disclination melting to the crumpled phase as temperature is increased , and at fixed @xmath5 , there should be a transition to the crumpled phase as @xmath0 is decreased .
it , however , has features that are either unexpected or unexplained .
the discontinuous change in curvature where the melting and crumpling lines join is surely an artifact .
it suggests that the physics of the crinkled - to - crumpled transition produced by decreasing @xmath5 at fixed @xmath0 and by decreasing @xmath0 at fixed @xmath5 are totally different , the latter being associated with the buckling instability of a membrane with a single disclination .
it leaves unanswered whether the buckling instability line for the zero temperature membrane has any significance for a membrane in thermal equilibrium , which is allowed to choose disclination configurations to minimize its free energy and thereby to reject highly energetic configurations with an excess of positive or negative disclinations . in this paper , we present a detailed analysis of the low - temperature crinkled - to - crumpled transition in hexatic membranes .
our approach treats both disclinations and gaussian curvature in the same real - space renormalization procedure and allows us to obtain recursion relations for @xmath0 , @xmath5 and the disclination fugacity @xmath12 .
previous treatments used momentum space renormalization procedures to calculate the recursion relation for @xmath5 and did not actually provide a complete set of recursion relations for @xmath0 , @xmath5 and @xmath12 nor a prescription for doing so .
our procedure shows that thermally induced shape fluctuations cause a @xmath0-dependent reduction in @xmath5 , missed in previous calculation , that leads to the phase - flow diagram shown in fig .
[ hexphased ] . the vertical line in fig .
[ gtfig ] is now curved , and the mechanism for the crinkled - to - crumpled phase transition in the vicinity of @xmath13 , the termination of the crinkled line , is the same for decreasing both @xmath0 and @xmath5 .
our renormalization equations allow us to study the fluid phase in the vicinity of the transition point @xmath13 and to show that near @xmath13 the persistance length is @xmath14 rather than @xmath15 , where @xmath16 is the kt correlation length ( with @xmath17 a constant ) , and @xmath18 is the bending rigidity at length scale @xmath19 .
our calculations are strictly speaking restricted to @xmath20 , @xmath21 and @xmath22 , i.e. , to the region below the curve @xmath23 in fig .
[ hexphased ] .
we can not , therefore , make any definitive statement about the interesting @xmath24 limit , which should be related to the crumpling of tethered membranes .
however , we believe that the phase diagram shown in fig . [ hexphased ] makes physical sense beyond the region of validity of our calculations . in this scenario , the transition from the crinkled to the crumpled phase would be controlled by the fixed point @xmath13 in fig .
[ hexphased ] for all @xmath25 and @xmath26 .
the transition at @xmath27 would , however , be controlled by another fixed point .
we begin in sec.ii with a review of how to describe tangent plane order on fluctuating surfaces .
we pay particular attention to correlation functions of orientational order . in order to compare tangent plane vectors at two different points on the surface ,
it is necessary to parallel transport one of the vectors along some path @xmath28 to the position of the other vector .
when gaussian curvature is nonzero , the direction of a parallel transported vector depends on @xmath28 even when there are no disclinations present .
physical correlation functions are invariant with respect to local coordinate transformations and for a particular membrane shape and distribution of disclinations depend on @xmath28 . when correlation functions are averaged over shape and position of disclinations , the dependence on @xmath28 vanishes . in sec.iii
, we discuss various models for hexatic membranes .
we begin with the hamiltonian for hexatic membranes expressed in terms of the orientational angle @xmath29 and a height variable .
we then transform this model into a coulomb gas model in which
charge " density arises both from disclinations and from gaussian curvature .
there are two coulomb - like terms in this hamiltonian : one which is zero when the local disclinations density equals the local gaussian curvature and one , arising from the imposition of a covariant cutoff@xcite , that couples gaussian curvature to gaussian curvature .
finally , we transform the coulomb - gas hamiltonian to a sine - gordon hamiltonian with a term coupling the sine - gordon field @xmath30 to the gaussian curvature with an imaginary coefficient .
the latter term is analogous to the dilaton coupling of string theory @xcite . in sec.iv
, we relate the dielectric constant and hexatic rigidity to correlation functions of the disclination - gaussian curvature density .
we show in particular that the renormalized hexatic rigidity appearing in orientational correlation function is the same as that calculated from the free energy .
we also calculate the charge - density correlation functions . finally in sec.v , we derive renormalization group recursion relations for @xmath5 , @xmath0 , and the disclination fugacity @xmath12 .
these equations show that height fluctuations renormalize the hexatic rigidity in the absence of disclinations .
this renormalized rigidity then flows under renormalization in the presence of disclination in exactly the same way as the rigidity of a flat membrane .
if the initial height renormalized rigidity is less than the critical rigidity for a flat membrane , there is no rigid phase .
thus , height fluctuations can destroy the crinkled phase . in sec .
vi , we review results and discuss some unanswered questions .
in particular , we address the effect of asymmetry in the energies of positive and negative disclinations@xcite on our results .
the focus of this paper is on the nature of tangent - plane order and the kosterlitz - thouless transition on fluctuating surfaces .
it will not present a complete derivation of the rg equations used in sec.v because they involve some two loop graphs , which should be treated in a sophisticated regularization procedure . in a companion paper
, we will derive the complete recursion relations from the sine - gordon model using a real space regularization procedure generalized from that used by amit et al . @xcite to treat the flat space problem .
we are concerned with the nature of tangent - plane order on fluctuating membranes . in this section
, we will review relevant concepts in differential geometry , mostly to establish notation .
we will also discuss how to describe long - range order ( or lack thereof ) on a metric with fluctuating curvature .
points @xcite on a two - dimensional surface embedded in three - dimensional euclidean space are specified by a three - dimensional vector @xmath31 with components @xmath32 as a function of a two - dimensional parameter @xmath33 .
covariant tangent - plane vectors are then defined as @xmath34 where @xmath35 .
we will use greek letters @xmath36 to denote components of covariant and contravariant tangent - plane tensors and roman letters @xmath37 to denote components of vectors and tensors in euclidean space .
the metric tensor is @xmath38 its inverse @xmath39 satisfying @xmath40 allows us to define contravariant tangent - plane vectors @xmath41 satisfying @xmath42 .
any vector @xmath43 in the tangent plane can be expressed as @xmath44 where @xmath45 and @xmath46 are , respectively , the covariant and contravariant components of @xmath43 . a unit normal @xmath47 to the surface can be constructed from @xmath48 and @xmath49 : @xmath50 the curvature tensor is then @xmath51 from the curvature tensor , one can construct the mean curvature , @xmath52 and the gaussian curvature , @xmath53 where @xmath54 and @xmath55 are the principal radii of curvature at the point of the surface in question .
the integral of the gaussian curvature is a topological invariant , @xmath56 where @xmath57 is the number of handles and @xmath58 is the euler characteristic . in the monge gauge , @xmath59 and @xmath60 , and the metric tensor @xmath61
is written as @xmath62 and the curvature tensor @xmath63 is @xmath64 where @xmath65 .
the anti - symmetric tensor @xmath66 will be particularly useful in what follows .
it is defined via @xmath67 where @xmath68 and @xmath69 is the anti - symmetric tensor with @xmath70 .
the contravariant tensor @xmath71 equals @xmath72 and satisfies @xmath73 .
finally the mixed tensor @xmath74 rotates vectors by @xmath75 since @xmath76 and @xmath77 .
we are interested primarily in order in the tangent plane of unit ( or fixed ) magnitude . for this
, it is useful to introduce orthonormal tangent - plane basis vectors @xmath78 and @xmath79 satisfying @xmath80 a tangent vector @xmath43 can be expressed in the basis @xmath81 as well as that defined by the covariant or contravariant vectors : @xmath82 where @xmath83 .
we will use roman subscripts @xmath84 to denote vector and tensor components with respect to the local orthonormal basis .
covariant derivatives are derivatives projected into the tangent plane
. components of the covariant derivative of a vector @xmath43 relative to the orthonormal basis are @xmath85 where @xmath86 is the spin - connection whose curl is the gaussian curvature : @xmath87 we will also find it useful to use a circular basis defined by the vectors @xmath88 satisfying @xmath89 with @xmath90 . in this basis , @xmath91 , and the covariant derivative , @xmath92 has a particularly simple form .
a vector order parameter @xmath93 that is restricted to lie in the tangent plane of a surface can be written as @xmath94 or @xmath95 .
if @xmath93 is a unit length vector , it is conveniently expressed in terms of its angle in the local orthonormal basis as @xmath96 where @xmath97 and @xmath98 .
the unit vector @xmath93 can thus be written in the circular basis as @xmath99 where @xmath100 is a complex order parameter with unit length @xmath101 .
we now turn to tensor tangent - plane order .
the simplest nontrivial tensor order parameter is a symmetric - traceless tensor @xmath102 with cartesian components @xmath103 the traceless constraint implies @xmath104 .
@xmath102 is a uniaxial tensor with a principal axis lying along a unit vector @xmath105 in the tangent plane .
if we require that @xmath102 have a fixed magnitude defined by @xmath106 , then we can write @xmath107 or @xmath108.\ ] ] this tensor , like the vector @xmath93 , can be expressed in terms of the real part of a complex tensor .
introduce the direct product tensor @xmath109 with components @xmath110 .
then @xmath111 generalization of this construction to higher order tensors is straightforward
. let @xmath112 with @xmath113 be a @xmath114th rank complex tangent - plane tensor . if we define the inner product of two tensors by @xmath115 the tensor @xmath116 describes vector order such as is present in the smectic-@xmath117 phase in which the long axes of molecules comprising the membrane tilt relative to the normal @xmath47 .
@xmath118 describes tangent - plane nematic order with inversion symmetry .
@xmath119 describes hexatic order .
@xmath120 would describe `` 4-atic '' order , etc .
thus @xmath121 for general @xmath114 describes what we call `` @xmath114-atic '' order . in flat space ,
the basis vectors @xmath122 or @xmath123 are independent of spatial coordinate .
information about the existence of long - range order and about @xmath114-atic order parameter correlations is contained in the correlation function @xmath124 .
basis vectors at different points on curved surfaces are not identical ( or even in the same plane ) and the simple dot product of vectors at different points does not carry information about tangent - plane order . in order to compare order parameters @xmath125 and @xmath126 at two different points
@xmath127 and @xmath128 , we need to parallel transport the order parameter at @xmath127 along some path @xmath28 to @xmath128 .
the parallel transported tensor @xmath129 is now in the tangent plane at @xmath128 like @xmath130 and we can take its dot product with @xmath126 . thus the correlation function , @xmath131 where the `` @xmath132 '' means parallel transported inner product , is what is appropriate for describing correlations in tangent - plane order . note that this correlation function depends on the path @xmath28 when the membrane has nonzero gaussian curvature .
we will find , however , that it becomes independent of @xmath28 on nearly flat membranes after averaging over height fluctuations .
the parallel transported order parameter can be expressed in terms of the spin - connection .
let @xmath133 be a tangent vector at @xmath127 and let @xmath134 be the vector @xmath135 parallel transported to a nearby point @xmath136 . then by definition @xmath137 with @xmath138
alternatively in terms of the circular basis , @xmath139 or @xmath140 applying this result to the vector @xmath141 , we obtain @xmath142 or @xmath143 thus we have @xmath144 this function is invariant under changes in the local coordinate system ( _ i.e. _ under nonlocal rotations of the vectors @xmath78 and @xmath79 ) . the generalization to @xmath114-atic
order is straight forward since by construction @xmath145 so that @xmath146 again , this is invariant under coordinate system changes .
in this section , we will derive various equivalent representations for the hamiltonian and associated partition function describing @xmath114-atic order on a fluctuating surface .
it is now well established that the long wavelength properties of a fluid membrane are well - described by the helfrich - canham hamiltonian @xmath147 @xcite , which can be expressed as a sum of three terms , @xmath148 the first term is the mean curvature energy , @xmath149 where @xmath150 , and the second form is valid for the monge gauge . the second term is the gaussian curvature energy , @xmath151 where @xmath152 .
this term is a topological invariant depending only on the genus of the surfaces .
since we will consider surfaces of fixed genus , we will drop this term .
finally @xmath153 is the surface tension energy .
we are mostly interested in free membranes for which the renormalized surface tension obtained by differentiating the total free energy @xmath154 with respect to the total surface area @xmath155 @xmath156 , is zero .
since there are entropic contributions to @xmath157 as well as contributions vfrom internal order , the value of the bare surface tension @xmath158 will have to be adjusted to keep @xmath157 zero . in
what follows , we will ignore @xmath159 with the understanding that it is really present if we want to keep track of how @xmath157 actually becomes zero .
the partition function for a fluid membrane , @xmath160 is obtained by integrating over all physically distinct realizations of the surface , which is specified by the vector @xmath161 .
this means we have to specify a gauge or parametrization for @xmath161 .
thus the measure @xmath162 for distinct surface configurations contains a fadeev - popov determinant @xcite .
in addition , it in general contains a factor to correct for the fact that different surface configurations arising from a fixed parametrization surface can have different areas @xcite . as discussed in the preceding section
, the complex tangent - plane tensor @xmath121 distinguishes a fluid membrane from one with @xmath114-atic tangent - plane order . since @xmath121 has a fixed magnitude and there are no external fields aligning @xmath121 along a particular direction , the lowest nontrivial contribution to the energy associated with @xmath121 arises from its gradients , @xmath163 where @xmath164 is a covariant derivative . using eq .
( [ eq : cov - circ ] ) and eq .
( [ eq : ppsi ] ) , it is straight forward to show @xmath165 and @xmath166 where we set @xmath167 .
( [ eq : patic - ener ] ) is the simplest energy arising from @xmath114-atic order .
it is rotationally isotropic , and it describes correctly the lowest order gradient contribution to the energy for all @xmath168 . for @xmath169 or @xmath170 , however , this energy should have anisotropic contributions @xcite . in this paper
, we will ignore these anisotropies .
the @xmath114-atic order parameter @xmath121 can have disclinations of strength @xmath171 where @xmath172 is an integer @xcite .
a disclination at @xmath173 with strength @xmath174 gives rise to a singular contribution @xmath175 to @xmath29 satisfying @xmath176 where @xmath28 is a contour enclosing @xmath177 .
thus , in general @xmath178 where @xmath179 is nonsingular , @xmath180 and @xmath181 where @xmath182 is the disclination density .
the vector @xmath183 can always be chosen so that it is purely transverse , so that @xmath184 . in the @xmath114-atic hamiltonian
, @xmath185 always occurs in the combination @xmath186 . the spin - connection @xmath187 can and will in general have both a longitudinal and a transverse component . however , one can always redefine @xmath179 to include the longitudinal part of @xmath187
. this amounts to choosing locally rotated orthonormal vectors @xmath188 and @xmath189 so that @xmath190 .
thus we may take both @xmath183 and @xmath187 to be transverse and the @xmath114-atic hamiltonian , @xmath191 can be decomposed into a regular longitudinal part , @xmath192 and a transverse part , @xmath193 where it is understood that @xmath194 .
it costs an energy @xmath195 to create the core of a disclination of strength @xmath172 .
( we assume for the moment that the core energies of the positive and negative disclinations are the same .
see , however , ref . and the summary section . )
thus , partition sums should be weighted by a factor @xmath196 for each disclination of strength @xmath172 . since @xmath197 , we may at low temperature
restrict our attention to configurations in which only configurations of strength @xmath198 appear .
let @xmath199 be the number of disclinations of strength @xmath198 and let @xmath200 be the coordinate of the core of the disclination with strength @xmath198 labeled by @xmath201 .
the @xmath114-atic membrane partition function can then be written as @xmath202 where @xmath203 , and @xmath204 .
@xmath205 depends on all of the disclination coordinates @xmath200 where @xmath206 and @xmath207 where @xmath2 is a molecular length .
the kronecker factor @xmath208 in @xmath209 imposes the topological constraint @xcite that the total disclination strength on a surface with euler characteristic @xmath210 be equal to @xmath210 .
thus , on a sphere with @xmath211 , @xmath212 .
if @xmath213 for example , there must be 12 more positive than negative disclinations .
on a nearly flat surface , the coulomb interaction effectively restricts @xmath214 to equal @xmath215 , and we use eq .
( [ eq : trace ] ) with @xmath216 even though @xmath210 for an open surface is strictly speaking equal to one @xcite . the @xmath114-atic model of eq .
( [ eq : partition ] ) can easily be converted to a coulomb gas model using @xmath217 which follows from eq .
( [ eq : gaussian ] ) and eq .
( [ eq : vortex ] ) .
the quantity @xmath218 is a `` charge '' density with contributions arising both from disclinations and gaussian curvature .
equation ( [ eq : curl - vel ] ) implies @xmath219 where @xmath220 is the laplacian on a surface with metric tensor @xmath61 acting on a scalar . recall [ eq .
( [ eq : gam - rot ] ) ] that @xmath221 rotates a vector by @xmath75 so that @xmath222 is perpendicular to @xmath223 and is thus manifestly transverse .
using eq .
( [ eq : va - aa ] ) in eq .
( [ eq : h - perp ] ) , we obtain @xmath224 where @xmath225 is the coulomb hamiltonian associated with the charge @xmath226 .
the longitudinal variable @xmath179 appears only quadratically in @xmath227 and the trace over @xmath179 can be done directly : @xmath228 where @xmath229 is the liouville action @xcite arising from the conformal anomaly .
the first term in this expression is proportional to the surface area and can be incorporated into the surface tension energy @xmath159 by shifting the surface tension @xmath158 .
we assume , as discussed earlier , that the total surface tension is zero .
we will , therefore , ignore this term in what follows .
the second term can be viewed as a coulomb interaction for gaussian curvature but with a negative coupling constant , _
i.e. _ , an imaginary charge .
the coulomb gas partition function can thus be written @xmath230 where @xmath231 is the `` total '' coulomb hamiltonian in which there are two distinct charge densities , @xmath226 and @xmath232 , or , equivalently , @xmath233 and @xmath232 .
the coupling constant for gaussian curvature interactions in the absence of disclinations is @xmath234 this is the shifted coupling arising from the liouville term discussed in ref .
@xcite and @xcite .
the coupling constant for disclinations and between disclinations and gaussian curvature are not shifted by the liouville term and remain equal to @xmath5 . the coulomb gas model can be converted following standard procedures into a sine - gordon model .
the first step is to carry out a hubbard - stratonovich transformation on @xmath235 : @xmath236 where the liouville factor @xmath237 is needed to ensure that @xmath238 be one when @xmath239 . inserting this in eq .
( [ eq : tr - hs ] ) , we obtain @xmath240 where @xmath241 the only dependence on disclinations is now in the term linear in @xmath233 .
thus to carry out @xmath209 , we need only to evaluate @xmath242 -\omega \ } } \right)^{n_{+ } } \left ( y \int\frac{d^{2}u\sqrt{g}}{a^{2 } } e^{-i \ { 2\pi [ \phi({\bf u})/p ] -\omega \ } } \right)^{n_{- } } \nonumber \\ & = & \int \frac{d\omega}{2\pi } e^{i\omega p\chi } e^{(2y / a^{2})\int d^{2}u\sqrt{g } \cos[2\pi(\phi / p)-\omega]}.\end{aligned}\ ] ] thus @xmath243 } e^{-i\int d^{2}u\sqrt{g } s\phi}. \label{eq : tr - sg}\ ] ] we can now change variables , letting @xmath244 .
the term linear in the gaussian curvature then becomes @xmath245 where we used eq .
( [ eq : eulerc ] ) relating @xmath210 to the integral of the gaussian curvature .
thus the @xmath246-dependent part of this term exactly cancels the @xmath247 term arising from the integral representation of the kronecker delta .
the integral over @xmath246 in eq .
( [ eq : tr - sg ] ) is now trivial , and we obtain @xmath248 where @xmath249 is the sine - gordon action on a fluctuating surface of arbitrary genus .
the first two terms of this action are the gradient and cosine energies present in a flat space .
the final term provides the principal coupling between @xmath30 and fluctuations in the metric .
it is analogous to the dilaton coupling @xcite of string theory though here the coupling constant is imaginary rather than real .
note that the liouville action is not explicitly present in eq .
( [ eq : par - sg ] ) .
the requirement that the cutoff @xmath30 be applied in a covariant fashion is , however , still present .
thus , if @xmath250 , the integral over @xmath30 will yield the liouville factor .
when @xmath12 is not zero , care must be taken to implement any cutoffs in integrals in a covariant fashion .
in the preceding section , we derived several equivalent expressions for the partition function of a fluctuating membrane with internal @xmath114-atic order . in this section
, we will derive an expression for the inverse dielectric constant associated with the charge density @xmath226 and show that it is equivalent to the renormalized @xmath114-atic rigidity controlling the gauge - invariant @xmath114-atic correlation function of eq .
( [ eq : gp - corr1 ] ) .
we will begin by reviewing the derivation of the longitudinal dielectric constant in flat space .
we will then derive the dielectric constant and renormalized rigidities on fluctuating surfaces . consider a system with an internal charge density @xmath226 and fixed or controllable external charge density @xmath251 . to be concrete
, @xmath251 could be the free charge at the metal electrodes of a capacitor .
the total charge density is @xmath252 .
the electric field @xmath253 is determined by the total charge : @xmath254 here @xmath255 sets the units of charge ( for example @xmath256 in _ mks _ units ) .
the displacement vector @xmath257 , where @xmath258 is the dielectric constant , is controlled by the external charge : @xmath259 where @xmath260 is the longitudinal part of @xmath261 .
the electric potential can be introduced in the usual way : @xmath262 , @xmath263 and @xmath264 where @xmath265 and @xmath266 . the inverse longitudinal dielectric constant is obtained by expanding the free energy to second order in @xmath267 : @xmath268 a wave - number dependent dielectric constant can be obtained if @xmath269 is translationally invariant or by carrying the usual maxwell averaging procedure .
the result is @xmath270 where @xmath271 is the volume of the system .
we can use this formula to relate @xmath272 to correlations of the charge density .
the coulomb hamiltonian can be written as @xmath273 expanding @xmath274 , where @xmath275 is the non - coulombic contribution to the hamiltonian to second order in @xmath276 , we obtain @xmath277 and @xmath278 where @xmath279 is the density correlation function .
( [ eq : del - f ] ) is the well - known expression @xcite relating @xmath280 to the charge susceptibility @xmath281 . in what follows
, we will use the notation introduced in eq .
( [ eq : crr ] ) for arbitrary variables , @xmath282 where we have set @xmath283 and @xmath284 .
we can now determine the dielectric constant associated with the charge @xmath285 on a fluctuating membrane following exactly the procedures outlined in the preceding subsection .
we impose external charges @xmath251 to create a constant slowly varying external potential @xmath286 .
these charges lead to a total charge density @xmath252 .
they should not , however , change the gaussian curvature because otherwise they would change the liouville energy .
thus @xmath251 must arise from disclinations which we can , for concreteness , place at the boundary of the membrane like the charge on capacitor electrodes . in analogy with flat space , we can introduce external and induced electric potentials @xmath286 and @xmath30 satisfy @xmath287 and @xmath288 with @xmath289 the total potential .
the coulomb hamiltonian can then be written as @xmath290 we are interested in the long wavelength dielectric constant for our presumed macroscopically flat ( or nearly flat ) membrane .
for this purpose it is most convenient to use the monge gauge in which @xmath291 measures position in the average plane of the membrane . in this case
@xmath292 is a slowly varying function of @xmath293 , which , in the limit of fixed charge density at membrane boundaries , grows linearly with @xmath293 .
the long wavelength dielectric constant for our nearly flat membrane can be obtained using eq .
( [ eq : diele - inv ] ) , where @xmath294 is the fourier variable dual to position @xmath293 , and eq .
( [ eq : cab ] ) for @xmath295 .
using eq .
( [ eq : cab ] ) and expanding @xmath154 to second order in @xmath286 , we obtain @xmath296 and @xmath297 is the charge density correlation function for @xmath298 rather than @xmath226 .
@xmath299 is in fact the effective charge density for the coarse grained flat surface ( with metric tensor equal to unity ) because the total charge in an area patch @xmath300 should not change under coarse graining .
the quantity @xmath301 is equal to @xmath302 as can be verified easily to lowest order in @xmath3 in the monge gauge or more generally using the conformal gauge @xcite .
thus we have @xmath303 @xmath304 is the charge density correlation function evaluated in the presence of all interactions , including the coulomb like @xmath305 .
we will return at the end of this section to its evaluation .
in this section we will show that the renormalized hexatic rigidity @xmath306 is equal to @xmath307 .
to calculate @xmath306 , we introduce a shift @xmath308 in @xmath309 via @xmath310 in eq .
( [ eq : h - theta ] ) and evaluate the free energy to second order in @xmath308 . using eq .
( [ eq : del - free ] ) , we obtain @xmath311 as discussed previously , @xmath179 decouples from other variables . for any given realization of the surface @xmath312 .
we can therefore combine the first two terms in eq .
( [ eq : shift - free ] ) to obtain @xmath313 where @xmath314 is the transverse projection operator and @xmath315 is manifestly transverse from eq .
( [ eq : va - aa ] ) .
thus , as expected , @xmath316 depends only on the transverse part of @xmath308 .
a longitudinal function such as @xmath317 maintains its form under coarse graining .
a transverse function does not because the direction of a transverse function depends on the local value of the rotation matrix @xmath318 .
thus to coarse - grain our transverse function @xmath319 , we set it equal to @xmath320 .
in the monge gauge @xmath321 has the same value regardless of the particular configuration of the membrane . then using @xmath322 ,
we obtain @xmath323 setting @xmath324 where @xmath325 is the coarse grained @xmath308 , we find @xmath326 where @xmath327 is the base area and @xmath328 . because @xmath329 , @xmath304 tends to zero as @xmath330 in the ordered phase .
we can , therefore , define @xmath331 where in the final step we used the fact that @xmath332 depends only on @xmath333 in translationally invariant systems . thus @xmath334 @xmath304 is derived in eq .
( [ eq : crr ] ) .
we have argued that the renormalized rigidity is determined by the dielectric constant associated with @xmath226 rather than that associated with some other combination of the independent charge densities @xmath233 and @xmath232 .
we will now show that it is indeed this rigidity that controls the long distance properties of the @xmath114-atic correlation function @xmath335 , [ eqs .
( [ eq : gp - corr1 ] ) and ( [ eq : gp - corr2 ] ) ] .
our calculations follow closely those in ref .
@xcite for correlation functions in flat space . to keep things simple ,
we calculate only to lowest order in height fluctuations in the monge gauge . as discussed in sec.ii , the angle @xmath29 in eq .
( [ eq : gp - corr2 ] ) can be broken up into a regular part and a singular part : @xmath336 .
in addition , we can always choose a gauge so that @xmath187 is purely transverse .
in this case @xmath179 decouples from other variables , and we have @xmath337 where @xmath338 is evaluated for a fixed metric tensor @xmath61 determined by @xmath339 , and @xmath340 to the order we are working , we can factorize @xmath341 into a `` spin - wave '' part arising from @xmath179 and a transverse part : @xmath342 where @xmath343 and @xmath344 where @xmath345 and the double brackets in eq .
( [ eq : doub - bra ] ) refers to an average over @xmath29 followed by an average over @xmath339 . in the monge gauge , to the order we are working , @xmath346 where @xmath347 thus @xmath348 to evaluate @xmath349 , we use eq .
( [ eq : va - aa ] ) to lowest order in the monge gauge to obtain @xmath350 where @xmath351 then using the cauchy - riemman relation , @xmath352 where @xmath353 , we find @xmath354 and @xmath355 setting @xmath356 and @xmath357 , and expanding in @xmath358 , we obtain @xmath359 then using eqs .
( [ x2rr ] ) , ( [ eq : uy ] ) , ( [ eq : gv - xx ] ) and @xmath360 , we find @xmath361 \nonumber \\ & \sim & \frac{2b}{2\pi}\ln\frac{|{\bf x}-{\bf x}'|}{a } , \label{eq : gv - sim}\end{aligned}\ ] ] where we have treated @xmath362 as a non - divergent constant because @xmath363 is restricted to be greater than @xmath2 . combining eqs .
( [ eq : gp - gen ] ) , ( [ eq : gv - gen ] ) , ( [ eq : gsw - res ] ) and ( [ eq : gv - sim ] ) , we obtain @xmath364 with @xmath365 this equation agrees with eq . (
[ eq : k - ren ] ) to lowest order in @xmath366 .
as we discussed at the beginning of this section , the correlation function @xmath304 must be evaluated in the presence of all interactions including that arising from the liouville term . in this subsection
, we will derive a general expression for @xmath304 , which we will evaluate in the low - temperature , low - fugacity limit .
we will also discuss effective coulomb interactions .
both @xmath0 and @xmath5 have units of energy .
thus , both @xmath367 and @xmath6 are unitless measures of temperature . in addition
either of these quantities multiplied by any function of the ratio @xmath368 is also a measure of temperature . in our low - temperature expansion
, we will require @xmath369 , where @xmath370 smaller than @xmath371 , as well as @xmath372 and @xmath6 to be small .
thus , we will not consider the limit @xmath24 .
we will . nonetheless , be able to explore a large region of the phase diagram .
the evaluation of @xmath304 is obtained most directly in terms of diagrams .
figure [ diagrams ] shows the basic ingredients of a diagrammatic perturbation theory : the height propagator @xmath373 , the three coulomb interactions [ eq . ( [ eq : tot - c - h ] ) ] @xmath374 , @xmath375 , and @xmath375 coupling , respectively , @xmath376 to @xmath376 , @xmath377 to @xmath377 and @xmath376 to @xmath377 , the @xmath378 propagator when @xmath379 is zero , and non - linear vertices arising from the curvature .
the total coulomb hamiltonian of eq .
[ eq : tot - c - h ] contains @xmath380 and @xmath381 terms and a part @xmath382 coupling @xmath233 to @xmath232
. we will first calculate correlation functions when @xmath379 is zero .
then we will calculate the effects of turning on @xmath379 .
if @xmath379 is zero , then @xmath383 where @xmath384 is zero , and the gaussian curvature correlation function @xmath385 is simply that of a coulomb gas with coupling @xmath386 .
it can be represented as a sum of polarization bubbles shown in fig .
[ coulomb ] linked by a coulomb interaction as shown in fig .
the result is @xmath387 to lowest order in @xmath388 , the gaussian curvature is @xmath389 and , to lowest order in @xmath390 , @xmath391 is determined by a diagram ( a ) in fig . [ coulomb ] : @xmath392 note that @xmath391 is proportional to @xmath393 even though the numerator in the integrand is proportional to @xmath394 .
there is an infrared singularity in the integral over @xmath395 that converts the @xmath394 to the @xmath393 .
higher order contributions to @xmath396 are shown in figs .
[ coulomb ] ( b)-(f ) .
the contribution to @xmath397 from fig .
[ coulomb]b is of order @xmath398 , i.e. , of order @xmath399 smaller than eq.([eq : p - polar ] ) .
the diagrams of fig .
[ coulomb]c and [ coulomb]d are smaller by another factor of @xmath400 .
the diagram in fig .
[ coulomb]e , which arises from the @xmath401 factor and nonlinear terms in @xmath402 , is of order @xmath403 times eq .
( [ eq : p - polar ] ) .
it , however , has an infrared divergence , which would have to be handled with care if we wished to calculate to the next order in @xmath3 .
the diagram in fig .
[ coulomb]f is of order @xmath404 . when @xmath405 , the disclination correlation function @xmath406 can be calculated as a power series in @xmath12 .
to lowest order in @xmath3 , we can replace @xmath401 by 1 in @xmath377 .
then to lowest order in @xmath12 , @xmath407 where @xmath408 .
when @xmath379 is turned on , there will be additional contributions to the polarization bubble such as shown in fig .
[ newp ] . the correlation functions @xmath410 , @xmath385 and @xmath383 can be obtained from the diagrams in fig .
[ diagram2 ] : @xmath411 where @xmath412 is given by eq .
( [ eq : p - polar ] ) with @xmath391 corrected by contribution such as that in fig . [ newp ] . combining eqs .
( [ eq : c - nnss ] ) and ( [ eq : c - ns ] ) , we obtain the charge - density correlation function determining @xmath306 : @xmath413 } { 1 + ( \beta k / q^{2 } ) { \cal p}({\bf q})[(k'/k ) - ( \beta\kappa / q^{2 } ) c^{0}_{\tilde{n}\tilde{n}}({\bf q})]}.\end{aligned}\ ] ] in the limit of low - temperature and low - fugacity , this reduces to @xmath414 where @xmath391 is given by eq .
( [ eq : p - polar ] ) and @xmath415 by eq .
( [ eq : cnn - form ] ) .
the second term in the coefficient of @xmath416 is of order @xmath399 and will be neglected .
we can now calculate the dielectric constant or renormalized rigidity to lowest order in temperature and fugacity using eqs .
( [ x2rr ] ) , ( [ eq : k - ren ] ) , ( [ eq : p - polar ] ) , ( [ eq : cnn - form ] ) and ( [ crr - ap ] ) .
@xmath417 where @xmath418 we assume @xmath419 so that @xmath420 thus , the equation for @xmath421 is identical to that for a rigid flat membrane but with @xmath422 replacing @xmath5 .
to determine the properties of the kosterlitz - thouless transition on a fluctuating membrane , we need to determine the renormalization flow equations for @xmath0 , @xmath5 , and @xmath12 .
equation ( [ eq : rge - k ] ) for @xmath423 is identical to the equation for @xmath423 on a flat membrane@xcite with @xmath5 replace by @xmath424 .
flow equations for @xmath425 and @xmath426 can be obtained in the usual way@xcite from eq .
( [ eq : rge - k ] ) . by breaking up the integral @xmath427 into @xmath428 , putting the first part into a renormalized @xmath425 and rescaling the second part by @xmath429
, we obtain @xmath430 with @xmath431 observe that the bending rigidity does not appear explicitly in these equations ; it only appears via the dependence of @xmath432 on @xmath0 .
this suggests that there should be an effective theory for a flat membrane with a bare hexatic rigidity equal to @xmath433 rather than @xmath5 .
indeed , if we integrate our height fluctuations to produce an effective flat space hamiltonian via @xmath434 we obtain @xmath435 thus , the fluctuating membrane behaves like a flat membrane with a hexatic rigidity @xmath422 .
note that the renormalized rigidity @xmath436 appears in eq .
( [ heff ] ) because fluctuations at _ all _ wavenumbers @xmath294 have been integrated out .
if we had calculated an effective hamiltonian by removing @xmath437 for @xmath294 within a shell @xmath438 , there would have been no shift in @xmath5 .
the renormalization of @xmath0 can be calculated using eqs .
( [ eq : tr - hs ] ) and ( [ eq : tot - c - h ] ) .
diagrams contributing to the renormalized bending rigidity @xmath439 to lowest order in @xmath440 are shown in fig .
[ diagram1 ] .
they yield @xmath441 where @xmath442 is the linear dimension of the membrane , which we take to be infinite , and @xmath443 .
we can now follow exactly the same procedure to obtain the recursion relation for @xmath0 we used to obtain those for @xmath422 and @xmath12 .
to lowest order in @xmath390 , we can replace @xmath386 by @xmath5 .
we break the integral @xmath444 into two parts , and rescale to obtain @xmath445 this equation was derived from the low - order diagrams shown in fig .
[ diagram1 ] in which the coulomb energy @xmath446 is treated as a perturbation .
one can also structure a perturbation theory in which the coulomb propagator @xmath374 coupling @xmath376 to @xmath376 [ eq . ( [ eq : tot - c - h ] ) ] is replaced by @xmath447 where @xmath448 as shown in fig .
[ sceenkappa ] .
similar expressions can be derived for the effective coulomb potentials coupling @xmath377 to @xmath377 and @xmath377 to @xmath376 . in this scheme ,
[ diagram1]b and c are replaced by fig .
[ sceenkappa]a .
in addition , there are higher order diagrams such as those shown in fig .
[ sceenkappa]b .
these diagrams have two or more loops and are difficult to treat using our naive renormalization scheme . in the companion paper
, we will employ a controlled regularization scheme to treat these graphs using the sine - gordon hamiltonian .
it shows that @xmath5 should be replaced by @xmath422 in eq .
( [ kappa1 ] ) .
thus we have @xmath449 the renormalization flow equations now depend on @xmath422 and not on @xmath5 .
equations ( [ rg1 ] ) , ( [ rg2 ] ) , and ( [ rg3 ] ) define renormalization flows for fluctuating hexatic membranes .
they have a fixed point at @xmath450 we can obtain equations in the vicinity of this fixed point by defining @xmath451 then @xmath452 @xmath453 the flow lines for these equations are shown in figs .
[ fig1 ] and fig .
[ fig2 ] . figure [ fig1]a shows flows in the @xmath454-plane , which are identical to those in flat space @xcite .
figure [ fig1]b shows flows in the @xmath455-plane , which are similar to those in the @xmath454-plane .
figure [ fig2]a shows flows in the @xmath456-@xmath457 plane . as in previous treatments@xcite , there is a fixed line in this plane at @xmath458 ( @xmath459 ) , and the crinkled - to - crumpled transition occurs at @xmath460 independent of @xmath461 , and the kosterlitz - thouless transition occurs at @xmath462 independent of @xmath463 . the separatrix @xmath464 in fig . [ fig1]a is the transition line dividing the crinkled hexatic phase from the melted crumpled phase .
systems whose initial values of @xmath465 and @xmath12 lie below and to the left of the separatrix flow to the low - temperature crinkled hexatic phase ; points above and to the right of the separatrix flow away to the crumpled fluid .
the initial value of @xmath465 is @xmath466 , which increases with decreasing @xmath467 .
decreasing the bending rigidity @xmath0 decreases @xmath422 and eventually drives @xmath465 to the right of the separatrix .
thus , there will be a transition from the crinkled to the crumpled phase as @xmath0 is decreased . an alternative way to see
this is to use @xmath5 rather than @xmath424 as the independent variable . to do this
, we define @xmath468 to satisfy eq .
( [ eq : k - kappa ] ) with @xmath422 and @xmath0 replaced , respectively , by @xmath469 and @xmath470 .
flows in the @xmath471 plane are shown in fig .
the vertical transition line in the @xmath472 plane is now curved and crosses the @xmath473 axis at a finite value of @xmath372 .
points to the left and below the separatrix flow to the crinkled fixed line .
points outside it flow to the crumpled phase . in deriving our recursions relations
, we restricted ourselves to @xmath457 , @xmath6 and @xmath474 .
thus , our calculations strictly speaking only apply below the curves @xmath23 in figs .
[ hexphased ] and [ fig2]b .
the flows in fig .
[ fig2 ] we calculated for @xmath169 .
figure [ hexphased ] shows flows for the physically more interesting case of @xmath213 .
our approximation applies to a considerable region around the fixed point @xmath13 .
higher order terms in @xmath403 and @xmath475 would have to be included to get an accurate picture of what happens above the curve @xmath23 .
is seems likely to us , however , that these higher order terms will not lead to any qualitative modifications to fig .
[ hexphased ] . for any finite value of @xmath476 and @xmath477 , the crinkled - to - crumpled transition
is governed by the fixed point @xmath13 .
for infinite hexatic rigidity , @xmath478 , the transition is controlled by different physics , probably analogous to that of tethered membranes@xcite .
there could , of course , be some phase boundary at large @xmath479 separating @xmath480-like behavior from finite-@xmath5 behavior , but we do not see any particular reason why this should be so .
the persistence length in the fluid phase is @xmath481 where @xmath482 is determined by @xmath483 i.e. , by @xmath484 .
integration of eqs .
( [ nrecur1 ] ) and ( [ nrecur3 ] ) yield @xmath485 and thus @xmath4 .
equation ( [ nrecur3 ] ) for @xmath463 is a nonlinear and can not be determined analytically .
we can , however , obtain a very good estimate of @xmath4 when @xmath486 , where @xmath487 is the kosterlitz - thouless correlation length .
the latter length is @xmath488 , where @xmath489 or @xmath490 .
the recursion relation , eq .
[ rg3 ] for @xmath485 is only valid for @xmath491 .
when @xmath492 , we approach a fluid phase with recursion relations for @xmath0 determine by @xmath0 alone .
we , therefore , assume that @xmath493 for @xmath494 .
then @xmath495 this equation can be integrated subject to the boundary condition that @xmath496 be the value of @xmath485 at @xmath497 determined by eqs .
[ rg1 ] and [ lineariz ] .
thus , @xmath498 @xmath499 , and @xmath500 where we used @xmath501 .
this expression is to be compared with the result @xmath502 for a pure fluid membrane .
it is what one would naively have expected .
the microscopic length @xmath2 is replaced by the kt coherence length @xmath19 , and @xmath0 at length scale @xmath2 is replace by @xmath0 at length scale @xmath19 .
equation [ xip ] for @xmath4 is valid provided @xmath486 ( or @xmath503 ) . near the kt critical point
, this inequality may not be satisfied .
we have investigated the kosterlitz - thouless transition on fluctuating hexatic membranes .
we developed three equivalent hamiltonians for describing these membranes : the hexatic elastic hamiltonian expressed in terms of gradients of the hexatic angle variable , a coulomb - gas hamiltonian , and a sine - gordon hamiltonian .
the coulomb gas is characterized by gaussian - curvature and disclination charge densities , which interact via potentials partially determined by the liouville action arising from a covariant cutoff .
the sine - gordon hamiltonian has a linear coupling between the sine - gordon field and gaussian curvature .
we showed that height fluctuations , when integrated over all wave number , soften the hexatic stiffness @xmath5 .
as a result , the disclination melting transition from the hexatic crinkled phase to the fluid crumpled phase is brought about both by increasing temperature and by decreasing the bending rigidity @xmath0 .
we derived renormalization group recursion relations for @xmath5 , @xmath0 , and the disclination fugacity @xmath12 that explicitly verify this . in a companion paper , we provide an alternative derivation of these recursion relations using the sine - gordon hamiltonian .
though the picture we present of the kosterlitz - thouless transition from the hexatic crinkled phase to the crumpled fluid phase makes a great deal of sense , it does leave some incompletely answered or unanswered questions .
first , we believe that the nature of shape fluctuation in the crinkled is not completely resolved .
it is generally believed@xcite that the crinkled phase is characterized by power - law correlations in layer normals : @xmath504 , where @xmath505 .
this result is obtained via exponentiation of the expansion @xmath506 ^ 2 \rangle \nonumber \\ & & = 1 - { t \over 2 \pi \kappa } \ln ( | { \bf x } | /a ) \label{nn}\end{aligned}\ ] ] using @xmath507 .
the exponentiation of this series has not been explicitly justified with , for example , a calculation of second order terms in @xmath508 .
it is interesting to note that the de gennes - taupin persistence length@xcite @xmath509 beyond which a fluid membrane is crumpled was calculated using eq.([nn ] ) and setting @xmath510 .
this observation would suggest that the possibility that the crinkled phase is in fact crumpled , but with a longer persistence length than the fluid phase , can not be ruled out .
second , we have in our calculations imposed a constraint of charge neutrality . in flat membranes ( i.e. , films ) , this constraint is imposed by the prohibitive energy cost of having excess charge of either sign .
we believe that there is a similar energy cost for breaking charge neutrality in free membranes , which are not constrained to be flat , though the situation here is more complicated .
nelson@xcite has pointed out that the plus - minus symmetry present in flat membranes is broken in free membranes .
membranes with a single disclination undergo a mechanical buckling transition from a flat configuration to a cone configuration for a positive disclination or a saddle configuration for a negative disclination . the energy of the cone is lower than that of the saddle , though both energies are proportional to @xmath511 where @xmath512 is the linear dimension of the membrane .
does this asymmetry modify the picture presented in this paper ?
we believe not .
a free membrane will choose configurations that will minimize its free energy .
any configuration , whether flat or buckled , with an excess of one sign of charge will have a contribution to its energy proportional to @xmath513
. charge neutral configurations , on the other hand , have energies that are finite in the @xmath514 limit .
thus , there is a prohibitive energy cost in both flat and free membranes to the violation of charge neutrality .
the shape of a membrane near the core of positive and negative disclination may nonetheless differ and lead to different fugacities @xmath515 and @xmath516 .
we generalize our treatment to include this possibility in a companion paper .
the results are that the ratio of the two fugacities is a marginal variable . in the ordered phase ,
both @xmath515 and @xmath516 scale to zero , and the kt transition to the disordered phase is not affected . in the disordered phase ,
both fugacities grow with a fixed ratio .
a more complete treatment of gaussian curvature will be needed to interpret this result .
we are grateful to mark bowich , david nelson , phil nelson , burt ovrut , and leo radzihovski for helpful conversations .
this work was supported in part by the penn laboratory for research in the structure of matter under nsf grant no .
91 - 20668 .
further support was provided by the institute for theoretical physics ( under nsf grant no .
89 - 04035 ) , where a portion of this work was carried out .
y. kantor , m. kardar , and d.r .
nelson , phys .
* 57 * , 791 ( 1986 ) : phys
a * 35 * , 3056 ( 1987 ) ; m. paczuski , m. kardar , and d.r .
nelson , phys .
* 60 * , 2638 ( 1988 ) ; j. a. aronovitz , l. golubovi , and t.c .
lubensky , j. phys .
( paris ) * 50 * , 609 ( 1988 ) . | we investigate the kosterlitz - thouless transition for hexatic order on a free fluctuating membrane and derive both a coulomb gas and a sine - gordon hamiltonian to describe it .
the coulomb - gas hamiltonian includes charge densities arising from disclinations and from gaussian curvature .
there is an interaction coupling the difference between these two densities , whose strength is determined by the hexatic rigidity , and an interaction coupling gaussian curvature densities arising from the liouville hamiltonian resulting from the imposition of a covariant cutoff . in the sine - gordon hamiltonian , there is a linear coupling between a scalar field and the gaussian curvature .
we discuss gauge - invariant correlation function for hexatic order and the dielectric constant of the coulomb gas .
we also derive renormalization group recursion relations that predict a transition with decreasing bending rigidity @xmath0 . |
the study of light - matter interactions is interesting and important both from the point of view of fundamental fundamental physics and device applications . in this context
, the structure which attracted a particular attention is a semiconductor microcavity ( mc ) with a quantum well ( qw ) embedded into it .
tuning the energy of the excitonic transition in resonance with the energy of the photonic cavity mode one can reach a regime of strong coupling accompanied by formation of hybrid quasiparticles called exciton - polaritons .
these half - light , half - matter particles exhibit a number of extraordinary properties . due to their extremely small effective mass and bosonic nature
, polaritons provide an opportunity to study various quantum collective phenomena , ranging from polariton bec @xcite and josephson effect @xcite to polariton - mediated superconductivity @xcite .
strong polariton - polariton interactions make it possible to observe several remarkable non - linear effects such as superfluidity @xcite , bi- and multi - stability @xcite , and soliton - like propagation @xcite .
although free polaritons are 2d particles , it was recently noted @xcite that it can be interesting to consider 1d polariton wires ( see setup in fig .
[ channel ] ) provided by the lateral confinement of the polaritons in one of the directions @xcite .
such wires have the potential to become basic building blocks in future spinoptronic devices , including polariton berry phase interferometer @xcite and polariton datta and das spin transistor @xcite . in the nonlinear regime ,
the bistability effects in 1d polariton channels allow for the realization of logical circuits based on polariton neurons @xcite , in which a local switching between states propagates throughout the wire .
the above mentioned phenomena require a proper theoretical basis for description of all relevant processes which is a non - trivial task .
a successful theoretical consideration should include the effects of interaction with a reservoir of acoustic phonons which leads to polariton thermalization , and the effects of polariton - polariton scattering leading to the blueshifts and nonlinearities .
additionally , it should be taken into account that polaritons have finite lifetime , and the correct description of their dynamics should necessarily take into account the effects of pump and decay .
currently , two main ways have been pursued to describe the dynamics of interacting polaritons .
firstly , with the assumption of full coherence , mean - field approximation gives the gross - pitaevskii equation ( gpe ) commonly used for the description of spatially inhomogeneous polariton condensates @xcite .
however , while gpe includes polariton - polariton scattering it does not describe interaction with a phonon reservoir . assuming oppositely that the polaritons are completely incoherent , the dynamics in reciprocal space can be described by the semiclassical boltzmann equations @xcite .
unfortunately , this technique fails to describe the real space dynamics of inhomogeneous systems .
recently , we proposed a novel formalism based on full density matrix approach to describe the dynamics of the interacting polariton system with dissipation in real space and time @xcite .
the method was applied for the study of the propagation of 1d polariton droplets . however , we neglected the pumping terms and thus consideration of the cw regime most relevant from the experimental point of view was not done .
this paper is devoted to the bridging of this evident gap .
here we investigate the phenomena of the bistability and domain wall propagation in polaritonic systems with dissipation provided by the interaction with an incoherent phonon bath and present the results on the temperature dependence of the hysteresis in the polariton system and polariton neuron switching .
the general formalism for the time evolution of a spatially inhomogeneous bosonic system without account of the coherent pumping terms was developed in our previous paper @xcite .
we refer to this paper for details and derivations and give here only a brief overview , adding the detailed consideration of the pumping terms only .
the state of the whole system is described by the combined polariton and phonon density matrix @xmath0 ( factorization corresponds to born approximation ) .
the phonon part of the system is assumed to be time - independent and thermalized , @xmath1 , while we need to determine the time - dependence of the single - particle polariton density matrix in real space .
the convenient way of its representation is the use of the polariton field operators @xmath2 : @xmath3 where the trace is performed by all the degrees of freedom of the system .
a fourier transform can be performed to work in reciprocal space , which makes the calculations easier .
@xmath4 where @xmath5 is the dimensionality of the system ( @xmath6 for non - confined polaritons , @xmath7 for the polariton channel ) , @xmath8 is its linear size , @xmath9 , @xmath10 are creation and annihilation operators of the polaritons with momentum
* k*. if the time dependence of the density matrix in reciprocal space is determined , an inverse fourier transform allows to obtain the dynamics in real space straightforwardly .
the hamiltonian of the system can be represented as a sum of the terms corresponding to various physically relevant processes in the system : @xmath11 where @xmath12 corresponds to free polariton propagation , @xmath13 to polariton - polariton scattering , @xmath14 @xmath15 to coherent laser pumping , @xmath16 to incoherent pumping and finite polariton lifetime ( see expressions for these two terms below ) . in the above formulae @xmath17
defines the dispersion of the free polaritons and the quantities @xmath18 and @xmath19 correspond to the polariton- polariton and polariton- phonon scattering .
the hamiltonian can be separated into the sum of the coherent part and the part introducing decoherence , @xmath20 the effect of the coherent and incoherent parts should be treated in different ways . as for the coherent processes in the system , their effect can be accounted for using the liouville - von neumann equation @xmath21 \label{liouville}\ ] ] on the contrary , the incoherent part of the evolution is described by the lindblad equation @xcite which reads @xmath22\right]= \label{liouville_int}\\ \nonumber & = \delta_{\delta e}\left[2\left(h^+\rho h^-+h^-\rho\right . h^+\right)- \nonumber \\ \nonumber & \left.-\left(h^+h^-+h^-h^+\right)\rho-\rho\left(h^+h^-+h^-h^+\right)\right]\label{lindblad}\end{aligned}\ ] ] where the coefficient @xmath23 denotes energy conservation and has dimensionality of inverse energy and is in the calculation taken to be equal to the broadening of the polariton state @xcite . the terms @xmath24 and @xmath25 correspond to the processes when the thermal reservoir particle ( phonon or other reservoir boson , see below ) is created or destroyed .
the effects of the terms corresponding to the polariton - phonon and polariton - polariton interactions was in detail considered in ref . .
the corresponding equations for the elements of the density matrix and their derivation can be found there , and we do not reproduce them in current manuscript . for polariton- polariton scattering ,
these equations reproduce an analog of gross- pitaevskii equation written for the density matrix , and for polariton- phonon interaction they are generalizations of semiclassical boltzmann equations . here
we only consider the terms provided by coherent and incoherent pumps , which should be added to the dynamic equations of ref . to make our consideration complete .
we start from the case of the coherent pump .
its physical meaning is coupling to an electric field with a well defined phase , provided e.g. by an external laser beam .
mathematically , the corresponding hamiltonian can be introduced as @xmath26 the coefficients @xmath27 are fourier transforms of the pumping amplitude in a real space by @xmath28 , which can be cast as @xmath29 where @xmath30 is the pumping spot profile in real space , @xmath31 is an in - plane pumping vector resulting from the inclination of the laser beam as respect to the vertical and @xmath32 is the pumping frequency of the single - mode laser .
let us now check the effect of @xmath15 on the evolution of the polariton density matrix .
insertion of the hamiltonian into liouville- von neumann equation yields the following result : @xmath33 one sees , that the equations for matrix elements contain a new quantity- an average value of the annihilation operator of the polariton field , which is nothing but the order parameter ( also called a macroscopic wavefunction ) of the system .
one thus needs to obtain the dexpression for this quantity to close the system of the equations .
straightforward derivation gives : @xmath34 where @xmath18 is a matrix element of polariton- polairton interaction and @xmath35 are transition rates of phonon- assisted processes ( see @xcite for the definition of the parameters ) . by incoherent pump
we mean the exchange of particles between the polariton system and some incoherent bosonic reservoir , whose nature depends on the pumping scheme .
usually , this will be an ensemble of incoherent excitons created either by an electrical pump or by incoherent optical excitation or a reservoir of the external photonic modes providing the leakage of the photons from the cavity .
the corresponding hamiltonian written in dirac representation reads @xmath36 where @xmath37 is a secondary quantization operator corresponding to the bosonic reservoir in question , @xmath38 are constants characterizing the coupling between the polariton system and the reservoir the introduction of this hamiltonian into lindblad equation leads to the standard terms , whose derivation can be found elsewhere@xcite : @xmath39 where the terms @xmath40 and @xmath41 denote the intensity of the incoherent pump and broadening of the polaritonic levels connected with lifetimes of the polariton states , @xmath42 .
they are usually taken as phenomenological parameters , but can be connected with the quantities entering in hamiltonian : @xmath43 where @xmath44 are the occupancies of the bosonic reservoir . in our further consideration
we will consider only the case of the coherent pump , thus putting all @xmath45 and retaining in the resulting equations only the terms corresponding to the lifetime , which corresponds to the case of the empty bosonic reservoir , @xmath46 for all the * k*.
to get a full system of equations for the dynamics of the polariton system with pump and decay , one should combine the equations derived in ref . with expressions , , .
the resulting formalism is suitable both for describing 2d polaritons and polaritons confined in 1d channels .
the consideration of the former case , however , needs powerful computing facilities and in the present paper we focus only on consideration of the latter one .
we considered a @xmath47 @xmath48 wide polariton channel in a microcavity with an active region based on inalgaas alloys with rabi splitting 15 mev .
the matrix elements of the polariton- polariton and polariton- phonon interactions were estimated using the standard formulae @xcite .
the first phenomenon we modeled was the effect of the thermalization in the polariton system provided by polariton- phonon interaction on its bistable behavior .
it is well known that if the polariton ensemble is fully coherent and its dynamics is described by the gross- pitaevskii equation containing coherent pumping and lifetime terms , and if the energy of the pumping laser lies slightly above the bottom of the lower polariton branch , the dependence of the concentration of the polaritons on pump intensity is described by an s - shaped curve , characteristic for systems revealing the effects of bistability and hysteresis @xcite .
such a behavior is due to the polariton- polariton interactions which introduce nonlinearity into the system . on the other hand , in the approach based on the semiclassical boltzmann equations corresponding to the limit of strong decoherence
the bistability is absent and the dependence of the occupancy of the ground state on the pump intensity is described by a single defined threshold function @xcite .
one can expect that a transition between these two regimes should occur if one raises the temperature in the system which leads to the intensification of the polariton- phonon interactions and decoherence in the system .
this was indeed observed in our calculations .
the computational results are presented in fig .
[ fig : bistability ] .
the system is pumped with a spatially homogenious laser beam oriented perpendicular to the qw ( i.e. at @xmath49 in @xmath50-space ) at a slightly higher frequency than the @xmath49 polariton frequency ( 0.24 mev detuning ) . at low pump intensity , the pump frequency is not in resonance with the condensate .
consequently , the condensate occupation remains fairly low .
as the pumping is increased , the polariton energy is blueshifted into resonance with the pump and there is a sudden jump in the occupation of the @xmath49 polariton state at some characteristic pump intensity @xmath51 .
if one then decreases the pump , the polariton occupancy jumps down at different value @xmath52 which corresponds to the hysteresis behavior .
however , the increase of the temperature leads to the intensification of the phonon scattering , and bistable behavior becomes less and less pronounced : the hysteresis area narrows and is quenched completely above some critical temperature @xmath53 .
this correspond to the transition between the gross - pitaevskii and boltzmann regimes in the polariton system .
a similar phenomenon was earlier predicted for microcavity- based terahertz emitting device @xcite .
the effect of bistability can form a basis for creation of a variety of devices based on nonlinear polariton transport @xcite . among them are polariton neurons @xcite- the building blocks of polariton based optical integrated circuits @xcite utilizing the phenomenon of domain wall propagation in bistable systems .
the underlying idea is following .
imagine that the polariton system is driven by a spatially homogeneous background cw pump with intensity corresponding to the bistable regime and the steady state occupancy of the system corresponds to the lower branch of the s- shaped curve .
then a short localized pulse is applied in the middle of the wire .
its intensity should be enough to send the condensate locally to the upper branch of the s - shaped curve .
due to diffusion , the polariton wavepacket spreads to the neighboring regions and switches them to the upper branch .
this way , the area of high occupancy steadily expands .
this phenomenon is analogous to the propagation of the domain wall in ferromagnetic materials .
it should be noted that although the polaritons have finite lifetime , this does not limit the length of signal propagation in a polariton neuron , and the signal keeps propagating as long as the background cw pumping persists .
as bistability in the polaritonic system strongly depends on temperature as was discussed above , the same should be true for the domain wall propagation in the polariton neurons .
this can indeed be seen from fig .
[ fig : neuron ] representing the results of our calculations .
at 15k , the switching is clear and one can easily detect the propagation of a sharp well defined domain wall . at 20k and using the same pump intensity as before , the switching is less pronounced , partly because the higher intensity is lowered and partly because the spreading of the high intensity is slower , which leads to the washing out of the domain wall . at 25k
there is no bistability for this pump intensity anymore , and the system goes back to its low population state after the application of the pulse . finally , the effect of phonon scattering on pure dephasing in the system was considered .
we pumped the system with a coherent pulse having gaussian profile in the real space .
for various temperatures of the system state we investigated in the steady state the spatial profiles of both total polariton density and its coherent part determined as @xmath54 , where @xmath55 the results are shown in fig .
[ fig : decoherence ] .
one can see that at low temperatures the coherent fraction is quite large at the center of the coherent pumping spot , but it dramatically decays outside it .
however , as temperature is increased , the density profile gets more spread out over the wire , forming an almost constant background density made up of the decoherent population . , no detuning .
the solid lines show the total population , the dashed lines the coherent population and the dotted ones the incoherent population . ]
in conclusion , we have considered the effects of coherent pumping and finite lifetime in a polaritonic system accounting for all physically relevant processes in the system .
we applied our theory for the consideration of nonlinear polariton propagation in a 1d polariton wire .
we have shown that the increase of temperature dramatically affects such processes as bistability switching and domain wall propagation in polariton neurons .
we thank dr . t.c.h .
liew , dr .
g. malpuech , dr . d.d .
solnyshkov and mr .
o. kyriienko for useful discussions .
the work was supported by rannis `` center of excellence in polaritonics '' and fp7 irses project `` polaphen '' .
i.a.s acknowledges the support from cost polatom project and thanks mediterranean institute of fundamental physics for hospitality .
j. kasprzak , m. richard , s. kundermann , a. baas , p. jeambrun , j. m. j. keeling , f. m. marchetti , m. h. szy- manska , r. andre , j. l. staehli , v. savona , p. b. lit- tlewood , b. deveaud and le si dang , nature * 443 * , 409 ( 2006 ) .
a. t. hammack , m. griswold , l. v. butov , l. e. smallwood , a. l. ivanov , and a. c. gossard , phys .
lett . * 96 * , 227402 ( 2006 ) ; r. b. balili , d. w. snoke , l. pfeiffer , and k. west , appl .
. lett . * 88 * , 031110 .
( 2006 ) ; o. el daif , a. baas , t. guillet , j .-
brantut , r. idrissi kaitouni , j. l. staehli1 , f. morier - genoud , and b. deveaud , appl .
. lett . * 88 * , 061105 ( 2006 ) ; r. i. kaitouni , o. el daif , a. baas , m. richard , t. paraiso , p. lugan , t. guillet , f. morier - genoud , j. d. ganiere , j. l. staehli , v. savona , and b. deveaud , phys . rev .
b * 74 * , 155311 ( 2006 ) ; m. m. kaliteevskii , s. brand , r. abram , i. iorsh , a. kavokin , and i. shelykh , appl . phys . lett .
* 95 * , 251108 ( 2009 ) | we investigate the phenomena of the bistability and domain wall propagation in polaritonic systems with dissipation provided by the interaction with incoherent phonon bath .
the results on the temperature dependence of the polariton bistability behavior and polariton neuron switching are presented . |
by definition , the process - nonclassicality quasiprobability distribution ( pnqd ) is given by @xmath58.\end{aligned}\ ] ] by using the kraus decomposition of the quantum process , @xmath59 , where @xmath60 , @xmath61 and @xmath62 are some kraus operators on @xmath63 , the above equation becomes @xmath64 consequently , we can write the pnqd as an expectation value @xmath65 with respect to the input coherent state , where we have defined @xmath66 and @xmath67 it has been shown in @xcite that @xmath68 is bounded , and @xmath69 . therefore , also the operator @xmath70 is bounded @xcite . according to @xcite
, this implies that this function can be expressed as an everywhere convergent power series in terms of @xmath54 and @xmath71 .
hence , for any unknown quantum process @xmath72 and any complex number @xmath51 , @xmath30 is uniquely determined by any small compact set of input coherent states @xmath31 .
we reconstruct the nonclassicality quasiprobability distribution ( nqd ) @xmath4 from the quadrature distributions with the help of the relation @xmath73 the pattern function @xmath74 is given by @xmath75 where @xmath76 is the nonclassicality filter . in practice , the nonclassicality quasiprobability can be estimated from quadrature - phase value pairs @xmath77 as @xmath78 the data consists of @xmath79 points taken at ten different phase values . to avoid certain computational artifacts ,
the evaluation is performed as described in the supplementary material of @xcite . as the quantum efficiency slightly differs for each state
, we decided to remove its effect and show the results for the states with quantum efficiency @xmath80 .
this can be achieved by some simple rescaling : @xmath81 on the left side of the equation , we have the nonclassicality quasiprobability of the ideal state .
therefore , it is sufficient to sample the nonclassicality quasiprobability of the lossy state with rescaled width and @xmath51 . for details , see also @xcite .
if we want to predict the outcome of a phase - insensitive quantum process for a phase - insensitive input quantum state ( like a thermal state ) , it is sufficient to examine only the phase - randomized output states of the process . in general , however , even a classical process can lead to a phase - sensitive output state from a phase - insensitive input state , for example by coherent displacement . for the phase - randomized nonclassicality quasiprobability , @xmath82 we can use the phase - randomized pattern function @xmath83 as it can be seen , the phase argument of the pattern function disappears .
there is one critical point in the consideration of probabilistic quantum processes , such as the single - photon addition process , see @xcite : the output of the quantum process formalism is given by @xmath84 however , for probabilistic quantum processes the term on the right side is not a valid quantum state , since the density matrix is not correctly normalized .
therefore , we do not observe the right side directly , but its quantum state @xmath85 therefore , if we insert the @xmath3 representation of the quantum state @xmath86 and the output state for input coherent states @xmath87 we find that @xmath88 where @xmath89 is the mean photon number of the input state for which we want to predict the output .
so far , we have sampled the nonclassicality quasiprobabilities of the output states @xmath90 for different coherent input states , @xmath30 . from these
, we can predict the nqd of the output state for an input state as @xmath91 where @xmath92 is the @xmath3 function of the input state .
if the latter is independent of the phase , the integral simplifies to @xmath93 where @xmath94 is the phase - randomized pnqd described above . in practice
, we evaluate this integral from the final number of measured input coherent state with the trapezoidal rule .
a systematic error is estimated by the comparison of the result with the integration over a cubic spline interpolation function . | we propose a definition of nonclassicality for a single - mode quantum - optical process based on its action on coherent states .
if a quantum process transforms a coherent state to a nonclassical state , it is verified to be nonclassical . to identify nonclassical processes ,
we introduce a representation for quantum processes , called the process - nonclassicality quasiprobability distribution , whose negativities indicate nonclassicality of the process . using this distribution
, we derive a relation for predicting nonclassicality of the output states for a given input state .
we experimentally demonstrate our method by considering the single - photon addition as a nonclassical process , and predicting nonclassicality of the output state for an input thermal state .
[ [ introduction . ] ] introduction .
+ + + + + + + + + + + + + the ability of detecting any nonclassicality generated by any quantum device enables us to manipulate and control the evolution of a quantum state .
in particular , this plays a central role in implementation of quantum information processing and communication @xcite . in general ,
the problem of the characterization of an unknown quantum device is addressed by means of quantum process tomography @xcite . a general method for quantum process tomography
was recently proposed that is based on probing a quantum process ( described by a completely positive and linear map @xmath0 ) using coherent states to characterize the process tensor in the fock basis , with a fixed maximum number of photons @xcite .
however , any photon - number cutoff will transform a classical state into a nonclassical one , as a finite sum of nonclassical states is always nonclassical .
therefore , the previously known methods are not able to distinguish quantum processes whose outputs are classical for any classical input state from those that may convert classical input states into nonclassical output states
. for this purpose a universal nonclassicality test of the output states is indispensable .
nonclassicality of quantum states is characterized by the glauber - sudarshan representation @xcite of the density operator @xmath1 , @xmath2 if the @xmath3 function has the form of a classical probability density , the corresponding quantum state is said to have classical analogue @xcite , otherwise the state is referred to as nonclassical @xcite .
however , in practice the @xmath3 function is highly singular for many quantum states , so that it can not be used to experimentally check the nonclassicality in general .
a recently proposed method for verifying nonclassicality of quantum states is to use a regularized version of the @xmath3 function , referred to as the nonclassicality quasiprobability distribution ( nqd ) , @xmath4 .
its negativities indicate the nonclassicality of any quantum state @xcite , and it can be directly sampled by homodyne detection @xcite .
the relation between the nqd and the @xmath3 function is easily formulated by their fourier transforms , i.e. , the characteristic functions @xmath5 and @xmath6 , respectively . the function @xmath7 is obtained by multiplying @xmath8 with a proper filter function , for a detailed discussion of the requirements we refer to @xcite .
an example of such a filter function is @xmath9 where @xmath10 ensures the normalization @xmath11 . in this letter , we propose a definition of nonclassicality for single - mode quantum - optical processes . we introduce a method for detecting nonclassical processes by testing the nonclassicality of the output states for coherent states at the input .
if there exists an input coherent state leading to a nonclassical output state , the quantum process is nonclassical .
this method enables us to identify nonclassical quantum processes that may transfer classical input states into nonclassical output states .
we derive a relation for predicting the nqd of the output state for given input states .
moreover , we experimentally demonstrate our method by verifying the single - photon addition to be a nonclassical process and predicting nonclassicality of the output state for an input thermal state .
[ [ nonclassical - processes . ] ] nonclassical processes .
+ + + + + + + + + + + + + + + + + + + + + + + for a general input quantum state @xmath1 , the output of a quantum process described by the map @xmath12 is obtained by using eq .
( [ glauber - sudarshan ] ) and the linearity of the map , @xmath13 where @xmath14 is the @xmath3 function of the input state .
as the map may not be trace - preserving in general , the output state , @xmath15 , is obtained from this expression simply by normalization .
the @xmath3 function of the output state is given by @xmath16 where @xmath17 is the @xmath3 function of the output state of the process conditioned on the input state being the coherent state @xmath18 . from it
follows that if the output states of a quantum process for input coherent states are classical states , i.e. , having @xmath3 functions that are positive semidefinite , then the output of the process for any classical input quantum states will be always a classical state .
this motivates us to define nonclassicality of quantum processes as follows .
a quantum process is nonclassical if it transforms an input coherent state to a nonclassical state .
therefore , based on this definition , a _
classical process _ transforms all coherent states to classical states , and the output state is classical for any classical input state .
also , nonclassicality of the output state for only one coherent state is sufficient evidence that the process is nonclassical . as the regularized version of the @xmath3 function , @xmath19 , is an appropriate representation of quantum states for experimentally verifying nonclassicality
, we introduce a related characterization of quantum processes .
the regularized version of @xmath20 , denoted as @xmath21 , a regular function of two complex variables , is a representation of the process that we use to verify its nonclassicality .
this conditioned quasiprobability distribution is denoted as the process - nonclassicality quasiprobability distribution ( pnqd ) , which unambiguously identifies the nonclassicality of a given quantum process .
for a nonclassical process there exists an input coherent state @xmath22 and @xmath23 such that @xmath24 .
we note that the nonclassicality of a quantum process does not imply that the output state is nonclassical for any classical input states , as we shall see in the following example of the cat - generation process .
having knowledge of the pnqd , @xmath21 , and using , one can find the nqd of the output state for an input state described by @xmath14 via @xmath25 in case the @xmath3 function of the input state is highly singular , by using the parseval identity the nqd of the output state can be obtained as @xmath26 where @xmath27 is the characteristic function of the @xmath3 function and @xmath28 is the fourier transform of @xmath29 . for unknown quantum processes
the pnqd can be estimated by sampling the nqd of the output states @xcite for a sufficiently large number of input coherent states . in principle , for any unknown quantum process , @xmath30 can even be uniquely determined by knowing the action of the process on only an arbitrary small compact set of input coherent states @xmath31 @xcite .
[ [ examples - of - classical - and - nonclassical - processes . ] ] examples of classical and nonclassical processes .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + examples of classical processes include the photon subtraction and the interaction of a state with a thermal bath ; for further details on classical maps , cf . ref .
@xcite . for the photon - subtraction process
, the map @xmath32 yields a classical state .
consequently , the output will be classical for any classical input state @xcite .
however , a nonclassical input state may be transformed to an output state with modified nonclassical properties @xcite , see fig . [
fig : pomega1 ] .
hence the output state of a classical process for a nonclassical input state may not be classical . and @xmath33 .
we used the filter function with the filter width @xmath34 .
the negativities indicate the nonclassicality of the state . ]
let us consider a model for the process of decoherence caused by a thermal bath of mean occupation number @xmath35 @xcite , the characteristic function of the @xmath3 function of the state at time @xmath36 is given by @xcite @xmath37 \phi_q(\xi e^{-\gamma t},0 ) \ , \ ] ] where @xmath38 ( @xmath39\phi(\xi,0)$ ] ) is the characteristic function of the @xmath40 function of the state at time @xmath41 , @xmath36 is the interaction time , and @xmath42 is the damping rate . the @xmath40 function of any quantum state is a positive semidefinite function @xcite .
hence , for @xmath43 the @xmath3 function is positive semidefinite , as it is given by the convolution of two positive semidefinite functions , and the output state for any initial nonclassical state is always classical . in this case
, this process is classical .
an example of a nonclassical process is the cat - generation process .
the unitary evolution associated with the hamiltonian @xmath44 , generates the schrdinger cat state at time @xmath45 ( @xmath46 ) @xcite @xmath47 this is a nonclassical process , as the pnqd takes on negative values ; see fig .
[ fig : pomega1b ] . as would be expected ,
this nonclassical process converts a classical state to a nonclassical one .
however , for certain classical input states the output state can be classical .
as the corresponding unitary operator is a function of the photon number operator @xmath48 , it leaves the photon number states @xmath49 unchanged . as a consequence ,
any statistical mixture of photon number states remains unchanged .
therefore , the output state of this process for an input thermal state is the same thermal state , which yields a classical output from a nonclassical process with classical input . .
the filter function is the same as in fig .
[ fig : pomega1 ] . ]
last but not least , even a nonclassical process with nonclassical input can have a classical output state .
a simple example is the application of the squeezing operation on a squeezed input state .
for example , when a squeezed vacuum input state is squeezed again with the same amount of squeezing but in the quadrature orthogonal to the original squeezing , the squeezed state transforms into the vacuum state .
[ [ experimental - demonstration - of - a - nonclassical - process . ] ] experimental demonstration of a nonclassical process .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + based on our definition , the single - photon addition process is a nonclassical process , as it transforms the vacuum state to the single - photon state , i.e. , a coherent state with zero amplitude to a nonclassical state . in the following ,
we experimentally demonstrate our method by applying it to the single - photon addition process .
stimulated parametric down - conversion is used to generate the single - photon - added coherent states @xcite .
the core of the experimental setup is a @xmath50-nonlinear crystal [ @xmath51-barium borate ( bbo ) , type i ] pumped by a 90-mw uv beam obtained by frequency doubling 1.5-ps pulses at 785 nm from a mode - locked ti : sapphire laser .
the spontaneous parametric down - conversion from the crystal consists in pairs of entangled photons emitted in two well - defined directions called signal and idler .
when a seed coherent state is injected in the crystal along the signal direction , stimulated emission also takes place .
a single - photon avalanche silicon detector is placed along the idler beam after spatial and spectral filtering .
a click from this detector heralds the generation of a single - photon - added coherent state in a well - defined spatiotemporal signal mode , which is then characterized by time - domain homodyne detection @xcite . in the experiment
, we analyzed the photon - added states with 13 different input coherent - state amplitudes .
for each acquisition the homodyne phase was varied between zero and @xmath52 and actively locked to 10 different values by monitoring the dc level from the homodyne receiver .
now we prove experimentally that this process is a nonclassical one . to estimate the pnqd of this process from experimentally recorded quadrature distributions , we use the sampling approach which has already been applied to determine the nqd in @xcite .
the pnqd is reconstructed by using the filter function with the filter width @xmath53 .
the effect of the quantum efficiency is removed as described in @xcite .
the obtained results are shown for three different input coherent states in fig .
[ fig : nqp : spacs ] .
we observe negativities for different amplitudes of the coherent input state @xmath31 , with decreasing negativity for increasing @xmath54 .
obviously , the negativity appears close to the origin of phase space , i.e. , at @xmath55 .
therefore , we examine the dependence of the pnqd on the input amplitude @xmath54 at @xmath55 more closely ; see fig .
[ fig : nqp : spacs : cross ]
. it is clearly seen that the negativity is statistically significant for low input amplitudes @xmath54 , which eventually yields the sought experimental proof of the nonclassicality of the process . for larger amplitudes
, the negativity vanishes at @xmath55 .
however , this does not mean that the output state for an input coherent state with large amplitude is classical .
as the single - photon - added coherent states are nonclassical for any input amplitude @xcite , one will find negativities of the pnqd at values of @xmath51 different from zero . for different input amplitudes @xmath54 for the single - photon - addition process , evaluated at @xmath56 .
the error bars correspond to one standard deviation .
the green solid line represents the theoretical expectation . ] by using the experimentally estimated pnqd for the single - photon - addition process , we are able to estimate the nqd of the output state via eq .
( 5 ) for a thermal input state with low mean photon number .
the fact that photon addition is probabilistic is properly taken into account ; for details see @xcite . in fig .
[ fig : pomega : spats ] we show the predicted nqd of the output state for a thermal input state , displaying strongly significant negativities , which prove nonclassicality of the output .
this estimate of the nqd of the output state is in good agreement with the directly measured nqd of single - photon - added thermal states @xcite . .
the standard deviation , shown by the blue shaded area , is mostly hidden by the linewidth , the systematic error is displayed by the red shaded area , see @xcite . ] [ [ conclusions . ] ] conclusions .
+ + + + + + + + + + + + we have proposed a definition of nonclassicality of a quantum process through its action on coherent states .
based on this definition , any quantum process that transforms a coherent state to a nonclassical one is identified as a nonclassical process , which may transfer a classical state to a nonclassical one . for classical processes
the output state is guaranteed to be classical for any input classical states .
a classical process can also be useful to transform the nonclassical properties of the input state into another form , which is desired for some applications .
nonclassical processes are necessary for the generation of nonclassical states , and subsequently they can be used to create entanglement by overlapping them on a beam splitter @xcite .
conversely , interference of classical states will not generate entanglement .
the presented method enables us to check whether an unknown quantum device can generate nonclassical states and to predict nonclassicality of the output state .
one useful application for the method is to verify nonclassicality ( classicality ) in multimode quantum devices and channels , by considering each input to output connection as a single - mode process , when output states are required to be entangled ( unentangled ) states .
[ [ acknowledgments . ] ] acknowledgments .
+ + + + + + + + + + + + + + + + we thank jan sperling and tim ralph for helpful discussions .
this work was supported by the deutsche forschungsgemeinschaft through sfb 652 .
s.r . also acknowledges support from the australian research council centre of excellence for quantum computation and communication technology project no .
ce110001027 ) .
a.z . and m.b .
acknowledge support of ente cassa di risparmio di firenze , and the eu under era - net chist - era project qscale .
m. a. nielsen and i. l. chuang , _ quantum computation and quantum information _
( cambridge university press , cambridge , england , 2000 ) .
j. f. poyatos , j. i. cirac , and p. zoller , phys .
rev .
lett . * 78 * , 390 ( 1997 ) .
g. m. dariano and p. lo presti , phys .
rev .
lett . * 86 * , 4195 ( 2001 ) .
m. mohseni , a. t. rezakhani , and d. a. lidar , phys .
rev .
a * 77 * , 032322 ( 2008 ) .
m. lobino , d. korystov , c. kupchak , e. figueroa , b. c. sanders , and a. i. lvovsky , science * 322 * , 563 ( 2008 ) .
s. rahimi - keshari , a. scherer , a. mann , a. t. rezakhani , a. i. lvovsky , and b. c. sanders , new j. phys .
* 13 * , 013006 ( 2011 ) .
m. lobino , c. kupchak , e. figueroa , and a. i. lvovsky , phys .
rev .
lett . * 102 * , 203601 ( 2009 ) .
r. kumar , e. barrios , c. kupchak , and a. i. lvovsky , phys .
rev .
lett . * 110 * , 130403 ( 2013 ) .
r. j. glauber , phys .
rev . * 131 * , 2766 ( 1963 ) .
e. c. g. sudarshan , phys .
rev .
lett .
* 10 * , 277 ( 1963 ) .
u. m. titulaer and r. j. glauber , phys . rev .
* 140 * , b676 ( 1965 ) .
l. mandel , phys .
scr .
* t12 * , 34 ( 1986 ) . t. kiesel and w. vogel , phys . rev .
a * 82 * , 032107 ( 2010 ) .
t. kiesel , w. vogel , m. bellini , and a. zavatta , phys .
rev .
a * 83 * , 032116 ( 2011 ) .
t. kiesel , w. vogel , b. hage , and r. schnabel , phys .
rev .
lett . * 107 * , 113604 ( 2011 ) .
see supplementary material . c. gehrke , j. sperling , and w.vogel , phys .
rev .
a * 86 * , 052118 ( 2012 ) .
a. zavatta , v. parigi , m. s. kim , and m. bellini , new j. phys . * 10 * , 123006 ( 2008 ) .
a. ourjoumtsev , r. tualle - brouri , j. laurat , and p. grangier , science * 312 * , 83 ( 2006 ) . t. richter and w. vogel , phys . rev .
a * 76 * , 053835 ( 2007 ) .
p. marian and t. a. marian , j. phys .
a * 33 * , 3595 ( 2000 ) .
a. m. perelomov , _ generalized coherent states and their applications _
( springer - verlag , berlin , 1986 ) . g. j. milburn , phys .
rev . a * 33 * , 674 ( 1986 ) .
b. yurke and d. stoler , phys .
rev .
lett . * 57 * , 13 ( 1986 ) .
natexlab#1#1bibnamefont # 1#1bibfnamefont # 1#1citenamefont # 1#1url # 1`#1`urlprefix[2]#2 [ 2][]#2 , , , * * , ( ) . , , ,
* * , ( ) .
, , , , , * * , ( ) .
m. s. kim , w. son , v. buek , and p. l. knight , phys .
rev .
a * 65 * , 032323 ( 2002 ) .
wang xiang - bin , phys .
rev .
a * 66 * , 024303 ( 2002 ) .
saleh rahimi - keshari@xmath57 , thomas kiesel , werner vogel , samuele grandi , alessandro zavatta , and marco bellini + |
let @xmath1 be a finite partially ordered set ( poset , for short ) .
one can associate a simplicial complex with @xmath1 in a canonical way .
let @xmath2 denote a simplicial complex with a vertex set @xmath1 consisting of such simplices @xmath3 that for some permutation @xmath4 we have @xmath5 in @xmath1 .
let @xmath1 and @xmath6 be posets .
a map @xmath7 : @xmath8 is called a morphism ( or map ) of posets if it preserves the non - strict order , i.e. for all @xmath9 , @xmath10 such that @xmath11 we have @xmath12 in @xmath6 . the map @xmath7 induces a simplicial map @xmath13 from @xmath2 into @xmath14 by an obvious rule .
thus any poset inclusion @xmath15 induces an inclusion map of the associated simplicial complexes : @xmath16 .
following d.quillen we use the construction @xmath17 to assign topological concepts to posets .
for example , we call @xmath1 contractible provided @xmath2 is contractible , and we define the homology groups of @xmath1 to be those of @xmath2 . consider a finite group @xmath18 .
the set of all subgroups of @xmath18 ordered by inclusion forms a lattice with a proper part @xmath19 .
the set of cosets of all subgroups ( including @xmath20 and @xmath18 ) ordered by inclusion also forms a lattice and @xmath21 is its proper part ( see @xcite ) .
thus the natural question arises : what homotopy type can spaces @xmath22 and @xmath23 have for arbitrary finite group @xmath18 ?
k.brown , c.kratzer and j.thevenaz proved that if @xmath18 is solvable , then both @xmath24 and@xmath25 are homotopy equivalent to wedges of equidimensional spheres ( see @xcite ) .
this fact is really intriguing , because a poset not associated to a finite group can have almost any homotopy type .
namely , for any finite simplicial complex @xmath26 there exists a finite poset @xmath1 such that @xmath26 and @xmath2 are homotopy equivalent ( see @xcite ) .
the problem of determining the homotopy type of @xmath24 and @xmath25 for any finite group @xmath18 is still open .
the case of simple groups seem to offer the main difficulty . the homotopy complementation formula by bjrner and walker ( see @xcite ) and
similar results which allowed to compute the exact homotopy type of lattices of solvable groups depend on the existence of a normal subgroup and therefore can not be used in this case .
studying the shellability property of subgroup lattices of finite groups shareshian proved that for certain series of finite simple groups ( @xmath27 for prime @xmath28 , @xmath29 , @xmath30 and @xmath31 for prime @xmath32 ) the homotopy type of lattice @xmath24 is that of a wedge of @xmath33 circles ( see @xcite ) .
the proof was based on the fact that in each case a subgroup lattice can be reduced to a 1-dimensional connected using quillen s fiber lemma .
any such complex is obviously homotopy equivalent to a wedge of circles .
therefore , another question may be of interest : does there exist a minimal simple group whose subgroup lattice has homotopy type different from a wedge of equidimensional spheres ?
we give an example of such group by determining the exact type of @xmath34 which is a wedge of 48 circles and 48 spheres . in this work we also use spectral sequence method to obtain estimates for betti numbers of complexes @xmath24 and @xmath25 for any finite group @xmath18 as well as more precise results for certain groups ( @xmath35 , @xmath36 and @xmath37 , for prime @xmath32 and @xmath38 ) .
for an element @xmath39 we will use the following notation : @xmath40 ( posets @xmath41 , @xmath42 and @xmath43 are defined similarly ) , @xmath44 and @xmath45 for any subset @xmath46 .
the main tools for dealing with topological properties of posets are the quillen s fiber lemma and the homotopy complementation formula by bjrner and walker : let @xmath7 : @xmath47 be a map of finite posets such that upper fibers @xmath48 are contractible for all @xmath49 ( respectively , lower fibers @xmath50 are contractible for all @xmath49 ) .
then @xmath7 induces a homotopy equivalence between @xmath1 and @xmath6 . the join ( or the least upper bound ) of elements @xmath9 and @xmath51 of a poset @xmath1 is defined as an element @xmath52 ( it it exists ) .
the meet ( or the greatest lower bound ) @xmath53 is defined similarly .
a poset @xmath1 is called a lattice if for any elements @xmath9 and @xmath51 of @xmath1 there exist @xmath54 and @xmath53 in @xmath1 .
a lattice is bounded provided it contains the greatest element @xmath55 and the least element @xmath56 .
obviously , every finite lattice is bounded .
the proper part of a bounded lattice @xmath57 is a subposet @xmath58 .
it is easy to check that if a poset @xmath1 contains an element @xmath59 which is comparable to all element of @xmath1 ( e.g. @xmath60 or @xmath61 ) , then @xmath2 is contractible as @xmath2 is a cone over @xmath62 in this case : @xmath63 .
particularly , every finite lattice is contractible . therefore by topological properties of a lattice @xmath57 we mean those of its proper part .
let @xmath57 be a finite lattice and @xmath64 .
denote @xmath65 .
then : 1 .
@xmath66 is contractible , 2 .
if @xmath67 is an antichain ( i.e. any two elements of @xmath67 are incomparable ) , then @xmath68 where @xmath69 denotes a join of topological spaces @xmath26 and @xmath70 , and @xmath71 denotes a suspension over topological space @xmath26 .
we also mention the following well - known corollary of quillen s fiber lemma ( see @xcite ) : [ coatoms ] let @xmath57 be a proper part of some finite lattice @xmath72 , @xmath73 be a set of all elements @xmath74 such that @xmath75 , where @xmath76 is some subset of maximal elements of @xmath57 .
then @xmath57 and @xmath73 are homotopy equivalent .
the last lemma allows us to reduce the complexity of a lattice being examined greatly .
for example , it follows that the poset @xmath77 is homotopy equivalent to a 1-dimensional complex which is connected as well as the original poset .
once can easily check that its reduced euler characteristics is -60 , hence @xmath77 is homotopy equivalent to a wedge of 60 circles . for many minimal simple groups ( @xmath27 for prime @xmath28 , @xmath29 , @xmath30 and @xmath31 for prime @xmath32 )
the situation is similar , however one still needs to use quillen s fiber lemma to remove a number of elements to get a 1-dimensional complex .
keep the notation of the previous lemma .
let @xmath78 be a proper part of some sublattice of @xmath72 such that @xmath79 .
then @xmath57 and @xmath78 are homotopy equivalent .
the sets of all non - empty intersections of maximal elements of @xmath57 and @xmath78 coincide ( with @xmath73 ) .
unfortunately , lemma [ coatoms ] can not be used iterative as we can not delete any new element .
thus , it is naturally to ask a question : is it possible to `` get rid '' of some maximal elements ?
we were able to show that in a more general case the homotopy type of a poset @xmath1 can be determined using the topology of its subposets .
let @xmath1 be a finite poset , @xmath80 . then @xmath81 any element of @xmath82 is greater than any element of @xmath83 , thus any chain in @xmath84 is a union of some chain in @xmath82 and some chain in @xmath83 ( note , that either chain may be empty ) .
but such chains correspond to the simplices of a join of spaces @xmath85 and @xmath86 .
r5.5 cm let @xmath1 be a finite poset , @xmath80 .
let the simplicial complex @xmath87 be contractible by @xmath88 , then @xmath89 a simplicial complex @xmath90 represents a cone with a point @xmath91 over the base @xmath92 .
the complex @xmath2 is a gluing of @xmath88 and @xmath93 by @xmath94 ( see figure [ p_neq_m ] ) .
the homotopy map , contracting @xmath92 by @xmath88 to some point @xmath9 , maps the cone @xmath93 to a suspension @xmath95 .
this suspension is glued to @xmath88 by exactly a point @xmath9 .
if the complex @xmath96 is not connected , then the basepoint @xmath9 of the wedge @xmath9 must belong to the same component as @xmath92 ( because @xmath92 is contractible by @xmath96 , it must be contained in a single connected component ) . [ vee_sigma ] let @xmath73 be an antichain of elements of @xmath1 .
assume that the complex @xmath97 is contractible by @xmath98 .
then @xmath99 the proof of the last theorem is similar to the proof of the preceeding lemma , so we omit it .
nevertheless , it is worth mentioning that @xmath73 is an antichain and thus for any @xmath100 and @xmath101 the cones @xmath102 and @xmath103 may intersect only by their bases .
if @xmath104 for all @xmath105 and the complex @xmath98 is @xmath106-connected ( i.e. @xmath107 ) , then ( [ vee_m_djoin ] ) holds .
[ rubanok ] let @xmath73 be a set of some ( possibly not all ) maximal elements of @xmath1 .
assume that the complex @xmath108 is contractible by @xmath98 , then @xmath109 any two maximal elements in a poset @xmath1 are incomparable .
hence , any subset @xmath46 consisting of maximal elements is an antichain .
this implies that we can use theorem [ vee_sigma ] .
it is only left to mention that @xmath110 for any topological space @xmath26 .
note that if @xmath1 is a finite lattice and @xmath73 is a set of some maximal elements of the proper part of @xmath1 , then @xmath111 is also a lattice .
thus , lemma [ rubanok ] is likely to give good results together with lemma [ coatoms ] : we leave only the intersections of maximal elements , then we delete some maximal elements , then we apply lemma [ coatoms ] again etc .
note , that at any moment we can do the same for minimal elements . combining the homotopy methods described above we will be able to determine the exact homotopy type of @xmath34 .
shareshian made a conjecture ( see @xcite ) that for any finite group @xmath18 the simplicial complexes @xmath22 and @xmath23 are homotopy equivalent to wedges of spheres of possibly different dimensions .
however , it is not even known if homologies in some dimension are torsion - free for arbitrary finite group . up to now the attention was focused mainly to minimal simple groups with @xmath24 homotopy equivalent to a wedge of circles .
we shall consider a minimal simple group @xmath0 and prove that the proper part of its subgroup lattice is a wedge of spheres of two different dimensions .
.,width=604 ] the subgroup lattice of @xmath0 is depicted in figure [ psl27 ] ( see @xcite ) .
each vertex corresponds with a conjugacy class , the cardinality of the class is represented by a number next to a vertex . if two conjugacy classes are connected by an arrow that means that a group from the upper class contains subgroups from the lower class .
the number of such subgroups is equal to a small number next to an arrow .
figure [ psl27 ] does not contain all possible arrows : if @xmath112 , then we omit the arrow @xmath113 and draw @xmath114 and @xmath115 . by lemma [ coatoms ]
one can consider a smaller poset @xmath6 of all nontrivial intersections of maximals subgroups ( i.e. conjugacy classes of @xmath116 , @xmath117 and @xmath118 are omitted ) .
suppose that @xmath119 is a set of all subgroups of type @xmath120 .
each elements of @xmath121 is maximal in @xmath6 , hence the complex @xmath122 is connected .
note that for any @xmath123 the complex @xmath124 ( i.e. it consists of @xmath125 points ) and by lemma [ rubanok ] and its corollaries we conclude : @xmath126 thus , we have isolated a wedge of @xmath127 circles .
denote @xmath128 by @xmath129 .
from the fact that @xmath130 ( see @xcite ) , we conclude that @xmath131 ( as we removed a wedge of @xmath127 @xmath132 ) . we will show that @xmath133 is homotopy equivalent to a wedge of @xmath127 spheres .
for each subgroup in @xmath134 we consider @xmath135 .
this is a connected 1-dimensional complex with reduced euler characteristics equal to @xmath136 as @xmath137 consists of exactly @xmath138 vertices and @xmath139 edges .
consequently , @xmath140 .
we delete @xmath141 vertices from the right conjugacy class of @xmath134 and show that the resulting poset @xmath142 is contractible . by theorem [ vee_sigma ] this means that @xmath143 again we use lemma [ rubanok ] to consider the poset of all nontrivial intersections of maximal subgroups instead of @xmath142 ( note that @xmath142 is still a proper part of some lattice , it is depicted in figure [ end ] ) .
it is well - known that @xmath144 and thus admits a natural group action on a 3-dimensional vector space over @xmath145 .
r5.5 cm without loss of generality it can be assumed that the left conjugacy class of @xmath134 consists of the stabilizers of non - zero vectors . any pair of non - zero vectors @xmath146 can be extended to some basis @xmath147 .
the group acts transitively on the set of all bases , hence @xmath148 is exactly the subgroup of operators mapping @xmath147 to any basis @xmath149 isomorphic to @xmath150 .
the stabilizer of any 3 vectors is obviously trivial .
denote the only element of the right class of @xmath134 by @xmath151 .
suppose that some subgroup @xmath152 in @xmath142 is not contained in @xmath151 , then it is either @xmath134 ( vector stabilizer ) or @xmath150 ( stabilizer of two vectors ) . again without loss of generality we assume that @xmath151 ( as a line stabilizer ) consists of all invertible matrices of the form @xmath153 one can easily check that the set of such matrices contains a nontrivial stabilizer of any two non - zero vectors .
in fact , we proved that any subgroup in @xmath142 intersects with @xmath151 nontrivially .
consequently , @xmath154 is empty and by homotopy complementation formula we conclude that @xmath142 is contractible .
thus , we proved the following theorem on the homotopy type of the subgroup lattice of @xmath0 : simplicial complex of the subgroup lattice of @xmath0 is homotopy equivalent to a wedge of @xmath127 circles and @xmath127 spheres : @xmath155
if we consider groups @xmath156 or @xmath157 , it becomes evident that the subgroup lattice of a finite group can be a wedge of spheres of any given number of dimensions and thus its euler characteristics can not give the complete information on the homotopy type of this lattice .
we shall use the standard tool of studying the homologies of a topological space a spectral sequence .
let @xmath1 be a finite poset , consider a natural filtration on @xmath1 : @xmath158 that is , @xmath159 exactly when the maximal length of a chain @xmath160 , @xmath161 , is @xmath106 .
obviously , @xmath162 thus the maximal length of the chain in @xmath1 is @xmath163 : @xmath164 and @xmath165 .
we define @xmath166 we call each set @xmath167 a level ( namely , the @xmath106-th level ) .
note , that if @xmath168 and @xmath169 , then @xmath170 and @xmath171 .
now we are ready to proof the main lemma of this section : [ nadstrojka1 ] assume that @xmath172 , @xmath173 and @xmath174 .
then the quotient space @xmath175 coincides with a suspension over @xmath176 : @xmath177 r5.5 cm consider a topological space @xmath178 .
it is obviously a cone over @xmath179 .
furthermore , as @xmath180 , the base of the cone @xmath176 lies in @xmath181 .
space @xmath182 represents a union of spaces @xmath181 and @xmath178 intersecting by the base of the cone @xmath176 ( see figure [ x_h ] ) : @xmath183 consequently , @xmath184 the last lemma shows a strong connection between @xmath2 and all of its subspaces @xmath176 .
now we need the following theorem generalizing lemma [ nadstrojka1 ] : [ buket_nadstrojka ] quotient space @xmath185 is a wedge of suspensions over @xmath176 indexed by all @xmath172 : @xmath186 r5.5 cm we have @xmath187 , futhermore , assume that @xmath9 and @xmath51 are in @xmath188 and @xmath168 , then @xmath189 and either @xmath190 or @xmath191 .
consequently , the space @xmath192 is a union of @xmath181 and the cones @xmath178 index by all @xmath172 ( see figure [ x_hg ] ) .
assume @xmath193 and @xmath194 then the intersection of posets @xmath195 coincides with @xmath196 .
hence , the cones @xmath197 and @xmath198 may intersect only by their bases , but their bases lie in a space @xmath181
. thus , @xmath199 by lemma [ nadstrojka1 ] the last space is a wedge of suspensions over @xmath176 at the vary of all @xmath200 .
note that for any topological spaces @xmath26 and @xmath70 the reduced homologies have the following properties : @xmath201 and @xmath202 ( see @xcite ) .
the filtration of @xmath1 defined above induces a natural filtration on @xmath2 : @xmath203 the structure of @xmath185 is determined , so we are ready to write down the spectral sequence for that filtration .
first , we describe @xmath204 ( we assume that @xmath205 for all @xmath206 and @xmath207 ) : @xmath208 now we should say some words about @xmath209 ( the arrows ) : they act from @xmath210 into @xmath211 . thus @xmath212 is a quotient of the kernel of @xmath209 from @xmath210 by the image of @xmath213 in @xmath210 ( in a spectral sequence the image always lies in a kernel ) : @xmath214 next , from the inductive construction of @xmath215 we see that 1 .
if @xmath216 for some @xmath106 and @xmath217 , then for all @xmath218 we have @xmath219 .
2 . for all @xmath106 , @xmath217 and @xmath220
we have @xmath221 ( by @xmath222 we will mean the torsion - free rank of an abelian group ) .
all the arrows @xmath209 starting from the diagonal @xmath223 point to cells on a diagonal @xmath224 regardless of @xmath225 , and vice versa , the arrows ending on a diagonal @xmath223 start from the cells on a diagonal @xmath226 regardless of @xmath225 .
thus we have @xmath227 from the dimension argument we conclude that spectral sequence stabilizes on @xmath228th step : @xmath229 .
@xmath230 can be used to determine the homologies of @xmath231 from the diagonal @xmath223 ( for fixed @xmath91 ) : denote all nontrivial groups on this diagonal starting from the top as @xmath232 , @xmath233 , @xmath234 , @xmath235 .
then @xmath232 is a subgroup in @xmath236 ; @xmath233 is a subgroup in @xmath237 , @xmath238 is a subgroup in @xmath239 etc .
the last group @xmath235 coincides with the considered quotient group ( see @xcite ) . using the fact that every diagonal @xmath223 in @xmath204 contains sums of groups of type @xmath240 ( except for the 0th column @xmath241 , where the only non - zero cell is @xmath242 ) , it is possible to formulate the following theorems ( we assume that @xmath243 and @xmath244 for any @xmath245 ): [ h_m_leq ] for any @xmath246 we have the following estimate for the betti numbers of the poset @xmath1 ( i.e. the ranks of its homology groups ) : @xmath247 fix @xmath248 . by the properties of the spectral sequence given above we conclude : 1 .
@xmath249 is a sum of dimensions @xmath250 for all @xmath251 on the diagonal @xmath223 .
2 . the dimension of @xmath210 can not grow with the increase of @xmath225 .
it easily follows that @xmath252 if @xmath253 , then the diagonal @xmath254 can contain more than one non - zero element : @xmath255 , and @xmath256
. the spectral sequence constructed above may be just as well applied to prove the absence of torsion : [ homology_0 ] suppose that for some fixed @xmath246 and for any @xmath39 we have @xmath257 , then @xmath258 . furthermore ,
if for all @xmath39 the homologies @xmath259 are torsion - free , then @xmath236 are also torsion - free . if for some @xmath246 the homologies @xmath257 for all @xmath39 , then all the elements on a diagonal @xmath226 in the spectral sequence are zeroes for all @xmath218 and thus @xmath258 .
moreover , for all @xmath260 the diagonal @xmath226 in the spectral sequence contains only zeroes , hence for all @xmath218 the image of the map @xmath209 for @xmath226 is equal to @xmath261 .
thus , in the inductive construction of @xmath210 for @xmath223 we will take quotients by @xmath261 .
this means that for @xmath262 each group @xmath251 on a diagonal @xmath223 is a subgroup of @xmath263 . if @xmath264 is torsion - free , then @xmath265 is torsion - free , hence @xmath236 is also torsion - free .
theorem [ homology_0 ] is very useful for dealing with torsion in higher non - vanishing homologies of a subgroup lattice or a coset lattice of a finite group @xmath18 : it is sufficient to prove that all the subgroups of a given group @xmath18 have torsion - free higher homologies and that higher non - vanishing homologies of @xmath18 have dimension at least one more than that of any of its subgroups .
[ sled_hdim_leq ] suppose that there exist such @xmath266 that for all @xmath267 and for all @xmath39 the homologies @xmath268 vanish , then for all @xmath269 the homologies @xmath270 also vanish : @xmath271 [ h_m_geq ] the following lower estimate holds for the betti numbers of a poset @xmath1 : if @xmath248 , then @xmath272 if @xmath253 , then @xmath273 and obviously @xmath274 the estimates can be easily deduced from the fact the sum of all dimensions @xmath251 on a diagonal @xmath223 with respect to the total dimension of @xmath263 on the same diagonal can not decrease by more than the sum of all dimensions of @xmath263 on the diagonals @xmath226 and @xmath224 . unfortunately , the right part of these expressions is often negative .
but if we know some homology groups of @xmath2 ( for example if @xmath2 is connected , then @xmath275 ) , then the method demonstrated above allows us to obtain sharper estimates for the betti numbers .
it is natural that this technic is useful in the case when euler characteristics does not contain the full information on the homotopy type of @xmath2 .
that is the case when @xmath2 is not homotopy equivalent to a wedge of equidimensional spheres .
we also note that to use theorems [ h_m_leq ] , [ homology_0 ] and [ h_m_geq ] one needs to know only the homologies of complexes @xmath276 , not the way they are linkes with each other .
thus for the subgroup lattice and the coset lattice of some finite group @xmath18 it is sufficient to know only the types of its subgroups and their number , but not the exact structure of the whole lattice .
for a topological space @xmath26 define its homology dimension as the maximal dimension of its non - vanishing reduced homologies : @xmath277 if all the reduced homologies @xmath26 are vanishing or @xmath278 , we assume that @xmath279 . for example , if @xmath26 is a wedge of spheres of possibly different dimensions , then @xmath280 is the maximal dimension of spheres in the wedge .
the cell homology theorem ( see @xcite ) states that homology dimension of any simplicial complex does not exceed its ordinary dimension : @xmath281 corollary [ sled_hdim_leq ] can be reformulated in the new notation as @xmath282 now we define the concept of decreasing poset and decreasing level inductively .
we start from the bottom : the empty set @xmath20 is not decreasing .
let @xmath1 be some poset and the property of being decreasing is defined for subposets @xmath276 for all @xmath39 .
we will define this property for @xmath1 . as in the previous section
we divide the poset into levels : @xmath283 , @xmath284 , @xmath234 , @xmath285 .
we say that the level @xmath167 is decreasing provided the poset @xmath276 is decreasing for all @xmath286 .
let @xmath287 be the number of decreasing levels in @xmath1 .
we say that the poset @xmath1 is decreasing if @xmath288 thus , we have defined the concept of being `` decreasing '' for 1-dimensional posets , then for 2-dimensional etc .
for example , any contractible poset ( including a single point ) is decreasing , but any 0-dimensional poset except for a point is not .
[ ponij_basa ] suppose that @xmath289 and the level @xmath290 is decreasing
. then @xmath291 . if @xmath292 , then @xmath293 .
the poset @xmath276 is decreasing , so its homology dimension can not be maximal . hence , @xmath294 .
now suppose that @xmath295 , then @xmath296 . from ( [ hdim_leq ] )
we conclude that @xmath291 .
[ ponij_basa_basa ] suppose that @xmath289 and the level @xmath167 is decreasing for some @xmath173
. then @xmath291 . if @xmath297 , then the statement is equivalent to lemma [ ponij_basa ] .
suppose @xmath298 .
again by lemma [ ponij_basa ] the top level of @xmath276 is decreasing for all @xmath172 .
hence , @xmath299 .
we use the induction : if for some @xmath300 and for all @xmath301 the homology dimension @xmath302 does not exceed @xmath303 , then for all @xmath304 from the level @xmath305 we have : @xmath306 thus , when we `` move '' to the next level , the homology dimension @xmath302 can not increase by more than 1 . inductively considering the levels from the bottom we reach @xmath1 itself . [ thm_ponij_s_levels ] if the poset @xmath1 containt exactly @xmath287 decreasing levels , than @xmath307 .
moreover , @xmath1 is decreasing exactly when @xmath308 .
moving from @xmath309 to the next level we see that homology dimension of posets @xmath276 does not increase , if @xmath310 is decreasing , and possibly increases by 1 otherwise .
as @xmath1 contains exactly @xmath287 decreasing levels , we conclude @xmath311 as we shall see later , this theorem has a useful application to the group lattices .
thus , the existence of decreasing subposets @xmath276 allows one to bound the homology dimension of @xmath2 .
it is natural to wish for some tools able to determine , whether a given poset is decreasing or not .
so we consider a fixed level @xmath167 in a poset @xmath1 .
let @xmath312 be the number of decreasing posets in @xmath1 below @xmath167 .
suppose there exists some non - decreasing level @xmath167 such that for any @xmath286 one of the following is true : 1 .
the poset @xmath276 is decreasing ; 2 .
the poset @xmath276 contains at least @xmath313 decreasing levels
. then the poset @xmath1 is decreasing .
as for each @xmath286 a complex @xmath276 contains at least @xmath312 decreasing levels and @xmath314 , theorem [ thm_ponij_s_levels ] states that @xmath315 does not exceed @xmath316 .
if @xmath276 is decreasing , then @xmath317 . in the opposite case by the conditions of the theorem it contains at least @xmath313 decreasing levels and
again we conclude that @xmath317 .
thus , @xmath317 for all @xmath286 , while the level @xmath167 is non - decreasing .
it follows that there was no dimension increase between levels @xmath318 and @xmath167 and therefore @xmath319 .
let @xmath18 be a finite group .
we use the following notation : @xmath320 is a proper part of the subgroup lattice ( ordered by inclusion ) , @xmath321 is a proper part of the coset lattice ( by inclusion ) , @xmath322 .
note that for @xmath323 we have @xmath324 .
this fact is really handy for performing calculations in @xmath25 as this poset contains lots of isomorphic fibers @xmath325 : if @xmath326 and @xmath327 are cosets of isomorphic subgroups @xmath328 and @xmath329 respectively , then @xmath330 . the same holds for @xmath24 and @xmath331 .
now we reformulate the results proved by the spectral sequence method using the language of group lattices .
[ group_h_m_leq ] betti numbers of @xmath24 , @xmath25 or @xmath331 for all @xmath246 can be majorized by betti numbers of all nontrivial proper subgroups in the following way : @xmath332 for @xmath24 the statement is just a reformulation of theorem [ h_m_leq ] .
for @xmath25 and @xmath331 we need to mention that subposets @xmath333 and @xmath325 are isomorphic for every subgroup @xmath152 and every coset @xmath334 of @xmath152 and the total number of cosets of @xmath152 is its index @xmath335 .
[ group_homology_0 ] let @xmath336 be one of the posets @xmath24 , @xmath25 or @xmath331 . if for some @xmath246 and for all subgroups @xmath337 we have @xmath338 , then @xmath339 .
furthermore , assume that for all @xmath337 homologies @xmath340 are torsion - free , then @xmath341 are torsion - free .
the statement is a reformulation of [ homology_0 ] .
let @xmath342 be one of the posets @xmath24 , @xmath25 or @xmath331 . if the maximal dimension of higher non - vanishing homologies of complexes @xmath343 at the vary of all proper subgroups @xmath337 is @xmath266 ( if there are no such @xmath152 , i.e. @xmath344 , we assume @xmath345 ) , then the dimension of higher non - vanishing homologies of @xmath346 does not exceed @xmath347 : @xmath348 [ group_no_tors ] if for any proper subgroup @xmath337 higher non - vanishing homologies @xmath349 are torsion - free and the poset @xmath336 is decreasing ( i.e. the higher non - zero homologies of @xmath350 are exactly of dimension @xmath347 ) , then higher homologies of @xmath350 are torsion - free . consider a coset lattice of @xmath0 .
the simplicial complex @xmath351 has dimension @xmath352 ( the longest chain is @xmath353 ) .
howerver , coset posets of any subgroup : @xmath145 , @xmath354 , @xmath117 , @xmath118 , @xmath150 , @xmath355 , @xmath116 , @xmath120 , @xmath356 and @xmath134 have non - zero homologies either in dimension @xmath261 ( @xmath145 , @xmath354 , @xmath117 , @xmath118 ) , or in dimension @xmath357 ( @xmath150 , @xmath116 , @xmath355 , @xmath120 and @xmath356 ) , or in dimension @xmath358 ( @xmath134 ) . as all proper subgroups of @xmath0 are solvable , their homologies are torsion - free ( see @xcite ) . applying thorem [ group_homology_0 ] and corollary [ group_no_tors ]
we obtain @xmath359 obviously , @xmath351 is connected .
moreover , @xmath35 proved to be simply connected ( see @xcite ) .
therefore by theorems [ group_h_m_leq ] and [ h_m_geq ] we immediately obtain some estimates for homology ranks of @xmath35 : @xmath360 now we shall find a connection between homologies of posets @xmath24 , @xmath25 and @xmath331 for any finite @xmath18 . consider the opposite filtration @xmath25 ( starting from the maximal cosets and ending by single elements ) : assume that @xmath361 , then @xmath362 , as for any @xmath363 the subposet of cosets containing @xmath364 is isomorphic to @xmath24 , hence , @xmath365 .
[ lcs_g ] consider a poset @xmath366 .
assume that for some @xmath91 we have @xmath367 , then the homology groups @xmath368 can be embedded into homology groups @xmath369 and there exists a surjection @xmath370 .
the easiest way to prove this theorem is to use an exact sequence of a pair ( see @xcite ) : consider a topological pair @xmath371 . by theorem [ buket_nadstrojka ]
we have @xmath372 the exact sequence for the pair @xmath371 is the following : @xmath373 if @xmath374 , then we get an injection @xmath375 and a surjection @xmath370 .
[ cg_lg ] assume that @xmath376
. then @xmath361 and @xmath377 are embedded into @xmath378 .
particularly , if homologies @xmath379 ( or are torsion - free ) , then @xmath380 ( or are torsion - free , respectively ) .
as @xmath381 , it is sufficient to use theorem [ lcs_g ] for the case @xmath382 .
[ cg_lg_dim ] the following estimate for the betti numbers holds : @xmath383 now we apply the concept of decreasing posets to the case of subgroup lattices .
it is important to mention that if the hypothesis that @xmath24 , @xmath25 and @xmath331 are all homotopy equivalent to wedges of spheres of possibly different dimensions ( see @xcite ) , then their homology dimensions coincide with maximal dimensions of spheres in the wedges .
we should say some facts about the structure of levels in group lattices : for any @xmath173 levels @xmath384 , @xmath385 and @xmath386 in the posets @xmath24 , @xmath25 and @xmath331 respectively contain the same subgroups . a level in each poset @xmath336 is decreasing provided the corresponding posets @xmath387 are decreasing for all subgroups on this level .
let @xmath336 be one of the posets @xmath24 , @xmath25 or @xmath331 and @xmath388 .
assume that @xmath336 contains exactly @xmath389 decreasing levels
. then @xmath390 .
moreover , @xmath336 is decreasing if and only if @xmath391 . a direct reformulation of theorem [ thm_ponij_s_levels ] .
corollary [ group_no_tors ] can be reformulated in the following way : suppose that @xmath18 is a finite group , the higher homologies of @xmath387 are torsion - free for all the proper subgroups @xmath152 of @xmath18 and the poset @xmath336 is not decreasing . then the higher homologies of @xmath336 are torsion - free .
posets @xmath24 and @xmath25 proved to be connected in the sense of being `` decreasing '' : suppose neither @xmath24 nor @xmath25 contains a decreasing level
. if @xmath24 is decreasing , then @xmath25 is decreasing .
an obvious corollary of [ cg_lg ] .
consider a group @xmath392 , where @xmath32 is prime and @xmath393 ( for @xmath394 the homotopy type of @xmath24 was completely determined in @xcite ) .
then all subgroups of @xmath18 are either solvable or isomorphic to @xmath395 for some @xmath396 .
obviously , the number of subgroups conjugated to a given subgroup @xmath397 in @xmath18 coincides its index @xmath335 , as @xmath152 is self - normalizing .
furthermore , all subgroups isomorphic to @xmath395 are contained in a single conjugacy class in @xmath18 for all @xmath396 ( see @xcite ) .
thus we have @xmath398 in particular , every subgroup @xmath395 is contained in a single subgroup @xmath399 , hence , using quillen s fiber lemma we can drop all subgroups isomorphic to @xmath395 for @xmath400 without changing the homotopy type of @xmath24 . let @xmath78 be a poset of all solvable subgroups of @xmath18 and @xmath151 be a set of subgroups of the type @xmath401 . then @xmath402 in fact , shareshian proved ( see @xcite ) that @xmath405 is homotopy equivalent to a wedge of circles : @xmath406 .
thus , we are ready to use the spectral sequence method : consider a filtration of a complex @xmath407 : @xmath408 . in the resulting spectral sequence @xmath204 will contain only two non - zero cells ( except for @xmath409 ) : @xmath410 1 .
@xmath411 for @xmath393 .
if @xmath394 , then obviously @xmath412 .
the reduced homologies of @xmath37 for @xmath393 have the following structure : @xmath413 where @xmath414 and @xmath415 is a finite abelian group ( torsion part ) .
consider a group @xmath416 , where @xmath32 and @xmath38 are different primes .
let @xmath417 be again a set of solvable subgroups of @xmath18 , and @xmath151 be a set of all simple subgroups of @xmath18 , i.e. a union of two conjugacy classes of @xmath418 and @xmath419 .
then @xmath420 and lemma [ nadstrojka1 ] yields @xmath421 thus , a spectral sequence constructed using the same filtration of @xmath422 will contain only two non - zero cells ( except for uninteresting @xmath409 ) : @xmath423 99 brown k.s . the coset poset and probabilistic zeta function of a finite group // j. algebra .
p. 989 - 1012 .
bjrner a. , walker j.w . a homotopy complementation formula for partially ordered sets //
european j. combin . 1983 .
4 . p.11 - 19 . dickson l.e .
linear groups with an exposition of the galois theory . new york : dover , 1984 .
hall p. the eulerian functions of a group // quart .
. 7 . p.134 - 151 .
kratzer c. , thevenaz j. type dhomotopie des treillis et treillis des sous - groupes dun groupe fini // comment .
p.85 - 106 .
shareshian j. on the shellability of the order complex of the subgroup lattice of a finite group
p.26892703 .
shareshian j. topology of order complexes of intervals in subgroup lattices // j. algebra .
p.677686 .
shareshian j. combinatorial properties of subgroup lattices of finite groups .
ph.d.thesis .
rutgers university , 1996 .
quillen d. homotopy properties of the poset of nontrivial @xmath32-subgroups of a group . // adv . in math .
28 , 2 . p.101 - 128 .
ramras d.a .
connectivity of the coset poset and the subgroup poset of a group // j. group theory .
p.719746 .
suzuki m. on a class of doubly transitive groups //
annals of math . 1962 .
75 . p.105 - 145 .
fomenko a.t .
, fucks d.b . a course in homotopy topology .
moscow : nauka , 1989 . | in this article we study group lattices using the ideas by k.s.brown and d.quillen of associating a certain topological space to a partially ordered set . we determine the exact homotopy type for the subgroup lattice of @xmath0 , find a connection between different group lattices and obtain some estimates for the betty numbers of these lattices using the spectral sequence method .
[ section ] [ section ] [ section ] [ section ] [ section ] |
high mass x - ray binaries ( hmxbs ) are among the brightest x - ray sources in the sky .
they consist of an accreting compact object ( a neutron star or black hole ) in orbit with an early type star ( o or b star ) .
they represent an important stage in the evolution of binary stars with early - type components , and they dominate the x - ray output of galaxies with young stellar populations .
hmxbs emit x - rays via accretion onto the compact object by gas from the strong wind from the companion star , or from roche lobe overflow .
the x - rays interact with the accretion flow and stellar wind , producing observable signatures in timing and spectra .
this provides a means to study the hot star wind and the accretion flow and the interaction of x - rays with these structures .
one consequence of this interaction is the production of polarized x - rays via scattering .
polarization provides unique information into the geometry of the stars , wind and scattering region . in this paper
, we explore the polarization signatures associated with this interaction .
there are @xmath010 hmxb sources with properties of their stellar winds and binary orbit which are well enough determined for quantitative modeling of the interaction of the accretion flow with x - rays @xcite .
the number of known hmxbs is many times greater when sources with less well constrained properties are included @xcite ; these include sources discovered by integral which are seen primarily at hard x - ray energies owing to obscuration , and transient sources which are detected primarily via long time - baseline monitoring .
the most thoroughly studied systems include the objects cyg x-1 , vela x-1 , and cen x-3 .
the majority of hmxbs contain pulsating neutron stars .
if so , orbital eclipses , x - ray pulse timing and primary radial velocities @xcite provides strong constraints on the orbital separations and masses .
simple estimates suggest that the accretion rate onto the compact object can be provided by the wind from the companion star in fewer than half the sources , while roche lobe overflow is required in the remainder @xcite . though they are outnumbered by low mass x - ray binaries in our galaxy ,
these sources dominate the x - ray output from galaxies with larger star formation rates @xcite
. their mass transfer can be very rapid ; in a system in which the primary has a radiative envelope and fills its roche lobe , the primary radius will shrink as a consequence of mass transfer , but if the mass transfer is conservative the roche lobe will shrink more rapidly , potentially leading to a common envelope phase .
if the primary is an evolved star with a convective envelope this outcome is more probable @xcite .
common envelope evolution when the primary is an evolved star is likely to lead to ejection of much of the envelope of the primary , leaving a binary consisting of a neutron star and the core of the primary . whether due to nuclear evolution or to unstable mass transfer , their evolutionary lifetimes must be short ( @xmath1 yrs ) .
the wind from the companion absorbs and scatters the x - rays from the accreting compact object in hmxbs , and the column densities traversed by the x - rays range from @xmath2@xmath3 .
the x - ray source luminosities range from @xmath4 erg s@xmath5 . in sources where the x - ray source is luminous and the wind is weak ,
the wind is ionized throughout most of the region between the two stars . in systems with more massive winds and weaker x - ray sources
x - ray - ionization effects are a perturbation on the wind properties .
key questions about hmxbs include the nature of the compact object : its mass , variability and intrinsic radiation pattern . also of interest are the properties of the companion star wind and its uniformity in density and temperature .
the x - ray source can provide a probe for the study of wind regions which are only weakly affected by ionization or the gravity of the compact object .
many of the x - ray properties of hmxbs are affected by the stellar wind and accretion flow .
this includes the variability around the orbit , which shows gradual eclipse transitions due to photoelectric absorption in the wind @xcite , enhanced absorption at late orbital phases due to a wake or stream trailing the compact object @xcite .
the wind and accretion flow also provide torque to the compact object and regulate its spin or angular momentum .
the structure of the wind and accretion flow is uncertain ; winds from comparable single o or b stars have radiatively driven winds @xcite which are affected by instabilities @xcite , shocks @xcite and clumping @xcite ; observations provide the most reliable means for understanding these processes , via fitting of parameterized models .
in addition , the compact object affects the gas flow , via its gravity and also x - ray heating and ionization
. the next new astrophysical window will be the advent of measurements of x - ray polarization @xcite . among other things
, polarization allows for potentially sensitive tests of the geometry of astrophysical sources , on scales which are far too small to be imaged directly .
there is only one source in the sky whose x - ray polarization is known , the crab nebula @xcite .
owing to visibility constraints , it is likely that the first astronomical x - ray polarimetry observations will be of objects which have never before been observed with this technique .
this motivates thorough and accurate modeling of the polarization properties of the brightest and ( otherwise ) best understood x - ray sources , for use as calibrators and test sources for x - ray polarimetry .
hmxbs are among the sources best suited for this .
their orbital elements are relatively well understood and their orbital variability provides a predictably changing view with respect to an important source of polarization : the strong stellar wind from the companion star .
x - ray polarimetry provides a means to probe the structure and the physical processes occuring in and near compact objects , on length scales too small to be imaged directly . in spite of scanty astrophysical detections so far @xcite it is of interest to consider the possible signals and diagnostic use of x - ray polarization in known classes of cosmic sources , as it is likely that instruments with improved sensitivity will become available eventually .
polarization properties of hmxbs have been discussed previously in the context of scattering of optical and uv from the primary star @xcite .
these authors predicted that the linear polarization has a characteristic variability around the binary orbit , associated with scattering in the stellar envelope which is distorted by the gravity of the compact object .
this variability can be used as a diagnostic of the shape of the star , and also of the inclination of the binary orbit .
these effects are observed at the predicted level in some systems ; their absence in other cases suggests that the optical light may be affected by contributions from structures whose shape is not symmetric around the line of centers .
related work has been carried out by @xcite .
useful formalism for the calculation of polarization has been provided by @xcite .
the polarization properties of the accretion columns in x - ray pulsars have been calculated by @xcite .
predictions of x - ray polarization by the large scale structures in binaries has been carried out for distorted stars @xcite ; for compton scattering within the accretion column of magnetic cataclysmic variables @xcite , following up on pioneering work by @xcite . x - rays have an important advantage over optical light which is that the source is almost certainly compact compared with the size of the binary system .
that is , we expect x - rays to be emitted from a region smaller than the alfven shell , defined as the distance from the compact object where the pressure due to a dipole magnetic field balances the ram pressure of accretion . for fields
@xmath6 gauss this region has a size @xmath7 cm @xcite .
far outside this region the influence of the field on the dynamics is negligible .
the alfven shell size can be compared with the primary star radius and orbital separation which are @xmath6 cm .
this disparity in sizes greatly simplifies the geometrical considerations associated with the use of x - ray polarization since the x - ray source is effectively a point source when considering the effect of the scattering by the stellar wind . calculating the polarization from these structures is straightforward , given certain simplifying assumptions , though it is important to take into account the effects of atomic absorption and resonance line scattering , which in turn depends on the gas dynamics , along with electron scattering .
the polarization properties of hmxbs will be affected by the intrinsic polarization of the compact x - ray source , and by the polarization imprinted by the stellar wind and accretion flow .
these can be distinguished by their differing variability behavior : the compact source generally varies on a pulsation timescale for sources containing a pulsar , which is @xmath8 seconds minutes .
the very few sources which contain black holes also have their strongest intrinsic variability on comparably short timescales .
wind variability , on the other hand , is associated with the orbital timescale ( @xmath8 days ) or possibly on the wind flow timescale which is @xmath9hours . in some hmxbs
the x - ray source is luminous enough to ionize the wind almost completely , so the light observed during and near eclipse , and its polarization , is dominated by electron scattering . in this paper
we present a general discussion of the behavior of x - ray polarization from hmxbs as a function of the parameters of the system and the viewing position .
we present calculations of the polarization signature for various simple analytic models for the wind density and ionization structure , and discuss the dependence of these on parameters : wind optical depth , inclination , and viewing direction or orbital phase .
we also explore the effects of wind hydrodynamics , i.e. departures from spherical symmetry due to the effects of the x - ray source gravity and heating .
we utilize sample three - dimensional dynamical models .
we take the intrinsic polarization of the compact object as unpolarized and explore the combined effects of geometry and scattering physics on the predicted linear polarization in the 0.1 - 10 kev x - ray band . in section [ sec1 ] we present numerical calculations for spherical winds using only electron scattering .
the effects of resonance line scattering are discussed in section [ sec3 ] . in section [ sec3b ]
we present models which include photoelectric absorption and an ensemble of resonance scatterers . in section [ sec4 ] we present models utilizing three - dimensional hydrodynamic models for the wind density and velocity field .
we use these to derive approximate predictions for the polarization levels expected for several well known systems in section [ sec5 ] .
the appendix presents simple analytic estimates of the polarization for idealized hmxb conditions ; many of these mirror earlier results .
the computational techniques and level of detail we employ are similar to those of our previous work @xcite on the polarization from warm absorbers in syefert galaxies .
the basic geometry of an hmxb is illustrated in figure [ fig1 ] .
this shows a schematic of an hmxb with a typical orbital separation and a sample x - ray ionized zone .
it also shows the geometry we use when calculating the polarization : the x - ray source is at the origin and the primary orbits in a plane inclined by an angle @xmath10 to the line of sight .
the orbital phase is described by the angle @xmath11 relative to the line of centers .
the system is viewed along the y axis . in this section
we will consider the simple case in which the wind density is spherically symmetric about the primary star , and the only interaction between the x - rays and the wind is electron scattering .
if so , the scattering phase function follows the rayleigh form @xcite .
the scattered emissivity at each point depends only on the local gas density and on the flux from the x - ray source .
we assume the x - ray source radiates isotropically with a constant total luminosity @xmath12 and is unpolarized .
we calculate the polarization signatures in the form of the three stokes parameters for linearly polarized light using the formal solution to the equation of transfer @xcite .
@xmath13 where @xmath14 is the photon energy , @xmath15 , is the source function , @xmath16 is the opacity , @xmath17 is the optical depth from a point @xmath18 to a distant viewer , @xmath19 is the scattering angle and @xmath20 is the angle between the scattering plane and a reference direction on the sky . here and in what follows
we describe the stokes parameters in terms of the luminosity seen by a distant observer , i.e. the total energy ( in ergs s@xmath5 sr@xmath5 ) radiated by the system in that direction . equation [ eq1 ] defines the scattered luminosity , @xmath21 ( polarized plus unpolarized ) ; observations are also affected by an unscattered component , @xmath22 where @xmath23 is the position of the x - ray source . for the purpose of evaluating these quantities
we adopt a cylindrical coordinate system with the x - ray source at the origin and the viewing direction along the @xmath24 axis .
the angle @xmath20 corresponds to the azimuthal angle on the plane of the sky and is measured relative to a line which is the intersection between the orbital plane and the plane of the sky . in the case of pure electron scattering the source function is @xmath25 where @xmath26 is the distance from the x - ray source and the opacity is @xmath27 .
some useful results are presented in the appendix .
the fractional polarization relevant to observation is @xmath28 .
this quantity is zero when the system is viewed at inclination @xmath29 and at either conjunction , i.e. orbital phase 0.5 or 1 , @xmath30 or @xmath31 .
the polarization at @xmath29 is a maximum at quadrature , phase 0.25 or 0.75 .
we can also define the polarization of the scattered radiation only , i.e. @xmath32 , and the value of this quantity at quadrature and @xmath29 depends only on the orbital separation and on the distribution of gas density in the wind .
the value of @xmath33 is less than the value of @xmath34 by a factor proportional to the thomson depth whenever the x - ray source is not eclipsed . at @xmath35 ,
i.e. when the system is viewed face - on , the polarization rotates with orbital phase at a rate twice the orbital rotation rate .
as an illustration of these results , we calculate the polarization produced by a single - scattering calculation by numerically evaluating equation [ eq1 ] .
the wind velocity is assumed to be radial relative to the star with a speed given by @xmath36 where the terminal velocity is @xmath37=1000 km s@xmath5 and @xmath38=100 km s@xmath5 , and @xmath39 cm .
the wind mass loss rate is @xmath40 .
the orbital separation is @xmath41 . in the remainder of this paper
we discuss quantitative calculations of polarization produced by wind scattering . these include calculations based on both the very simple spherical wind with parameters given in the previous paragraph and also on hydrodynamic calculations of the wind density structure which make no assumptions about spherical symmetry . in the former case ,
it is illuminating to estimate the accretion rate onto the compact object if the accretion is supplied solely by the wind .
this can be done using simple estimates based on @xcite and @xcite , i.e. @xmath42 where the accretion radius is @xmath43 , @xmath44 is the mass of the compact object and @xmath45 is the speed of the wind at the x - ray source .
this corresponds to @xmath46 where @xmath47 , @xmath48 or an x - ray luminosity @xmath49 using the parameters given in the previous paragraph and assuming an efficiency of converting accreted mass into energy of 0.1 .
this luminosity is less than the time averaged luminosities of most hmxbs by factors @xmath85 100 , which likely reflects the fact that wind mass loss rates may be greater , additional mass can be supplied by roche lobe overflow , and the wind and accretion flow dynamics can be enhanced by the influence of the compact object gravity and ionization .
the results in what follows , those which are based on the simple spherical wind approximation , must interpreted subject to this caveat .
an additional caveat which applies to essentially all of the results presented in this paper is the assumption of single scattering .
the validity of this assumption depends on the wind optical depth from the x - ray source ; multiple scattering is important when this quantity approaches or exceeds unity . for parameters similar to those discussed so far
, this quantity for a distant observer corresponds to @xmath50 , where @xmath51 . comparing this expression with the accretion luminosity estimate given above implies that multiple scattering can be important when the accretion luminosity exceeds @xmath52 for a 1 @xmath53 compact object .
the likely effect of multiple scattering is to produce smaller net polarization than for single scattering , since the second and subsequent scatterings have a wider range of scattering angles and planes than for single scattering .
if the depth is moderate , i.e. @xmath54 10 , there will still be a significant fraction of photons which reach the observer after a single scattering .
if so the net polarization will likely be less than predicted here , by factors of order unity , and our results should be modified to include multiple scattering effects for such sources .
maps of polarization projected against the sky are shown in figure [ fig2 ] .
this shows contours of constant intensity as solid colors separated by solid black curves , along with lines corresponding to polarization vectors , which appear as dashed curves .
these are plotted vs position in units of @xmath55 cm .
these illustrate many of the results presented in the appendix : at high inclination , @xmath29 , and at orbital phase angles 0 and @xmath31 ( orbital phases 0 and 0.5 ) polarization vectors are perfectly circumferential , contours of constant intensity are also circular and the net polarization is zero . at orbital phase
angles @xmath56 and @xmath57 ( orbital phases 0.25 and 0.75 ) the net polarization is maximum . at inclination
@xmath35 the star always influences the polarization , so that the net polarization fraction is constant , but the position rotates with the orbit .
more insight comes from the stokes quantities as a function of the orbital phase .
these are shown in figure [ fig3 ] for inclinations @xmath35 and @xmath29 .
u is green while q is shown in red ; solid is @xmath29 and dashed is @xmath35 .
this further illustrates the results from the previous figure : at @xmath35 the polarization oscillates between q and u along orbital phase , and the vector sum is constant . at @xmath29 u
is always small ( in this convention ) and q is a maximum at phase angles @xmath56 and @xmath57 .
another way of displaying the same thing is shown in figure [ fig4 ] , which shows u plotted vs q , for @xmath35 in green , and @xmath29 in red .
this shows that the trajectory of u vs. q is circular at @xmath35 and becomes linear at @xmath29 .
as shown by @xcite , this trajectory is an ellipse for intermediate inclinations , and the cases in figure [ fig3 ] are the extremes of eccentricity .
this has been suggested as a means for measuring inclination @xcite .
figure [ fig5 ] shows the scattered polarization fraction vs. orbital phase , i.e. @xmath32 .
this shows that the maximum polarization fraction is @xmath010@xmath58 for the parameters chosen here , when the polarized component is compared with the total scattered component .
this is comparable to the result in the appendix , calculated for a thin shell at the star rather than for an extended wind .
we emphasize that this value depends primarily on geometric quantities : the extent of the wind , and orbital separation relative to the stellar radius .
it does not depend on the wind density or column density since it is a comparison of scattered quantities .
the difference between @xmath35 and @xmath29 is clearly apparent : the former produces approximately constant polarization fraction and the latter oscillates between 0 and a value comparable to the @xmath29 value .
figure [ fig6 ] shows the polarization fraction measured relative to the total radiation , scattered plus direct , i.e. i.e. @xmath59 .
this illustrates the diluting effect of the direct radiation for this model , which has @xmath60
. maximum linear polarization at @xmath29occurs just following the eclipse transition , where the departures from circular symmetry on the sky are not negligible , and where the direct radiation is blocked by the primary star . at mid - eclipse the polarization at @xmath29 is zero due to symmetry .
figure [ fig7 ] shows the polarization angles from the same set of models .
the angle sweeps through 180 degrees twice per orbital period for @xmath35 , while the angle is constant ( though undefined near conjunctions ) for @xmath29 .
the models discussed in section [ sec1 ] include solely electron scattering ; they do not include resonant scattering in bound - bound transitions . this process ( when associated with uv and optical transitions )
is the dominant driving mechanism for the winds from early type stars . here
we consider scattering in the x - ray band . in the absence of x - ray ionization ,
the ions which are most abundant in early - type star winds are in charge states from @xmath81 5 times ionized , and do not have many strong x - ray resonance line transitions .
ionization by a compact x - ray source can produce ions of arbitrary charge state , depending on the x - ray flux and gas density , and so resonance scattering can affect the scattered x - ray intensity and polarization .
the ionization structure in hmxb winds has been explored by @xcite who showed that the surfaces of constant ionization in a spherical wind are either nested spheres surrounding the x - ray source , or else open surfaces enclosing the primary star .
the ionization distribution depends on a parameter @xmath61 where @xmath62 where @xmath63 is the ionizing x - ray luminosity , @xmath64 is the gas density at the x - ray source and @xmath65 is the orbital separation .
this is a particular example of the scaling of ionization in optically thin photoionized gas , which depends on the ionization parameter @xmath66 @xcite .
the properties of resonance scattering as applied to x - ray polarization in hmxbs are dominated by the fact that , owing to the restricted regions where ionization is suitable , most lines can only have strong opacity over a fraction of the wind .
the regions where this occurs most often resemble spherical shells surrounding the x - ray source .
resonance scattering differs from electron scattering in its sensitivity to the velocity structure of the wind . at a given observed photon energy ,
scattering can occur only over a resonant surface with shape determined by the wind velocity law ; in a spherical wind with monotonic velocity law these surfaces are open surfaces of revolution symmetric about the line of sight to the x - ray source .
resonance scattering affects polarization by redistributing the polarization among the components of the stokes vector according to the phase matrix which is a linear combination of the rayleigh and isotropic phase matrices ; coefficients of the matrix depend on the angular momentum quantum numbers of the initial and final states of the transition , @xmath67 @xcite .
the matrices describing this process have been calculated by @xcite . in practice
most of the strongest x - ray resonance lines have initial and final values @xmath68 and @xmath69 , which correspond to the rayleigh phase matrix , and we adopt this in our calculations .
similar assumptions were employed in @xcite .
polarization also depends on the scattering angle , and the resonance condition constrains the scattering geometry , thereby imposing a relation between the energy of the scattered photon and its polarization .
resonance scattering can have a cross section which is much greater than electron scattering , but only over the relatively narrow energy band spanned by the resonance line , including the effects of doppler broadening associated with the wind motion . within this band ,
the polarization can be greater than that produced by electron scattering , owing to the greater cross section , given suitable geometry .
in the remainder of this section we illustrate these effects by generalizing the single scattering spherical wind calculations to include resonance scattering .
we consider a single resonance line , chosen to crudely resemble a line such as o viii l@xmath70 . the parent ion is assumed to exist over a range of ionization parameter @xmath71 .
we calculate the optical depth using the sobolev expression @xcite : @xmath72 where @xmath73 is the oscillator strength , @xmath74 is the gas number density , @xmath75 is the ion fraction , @xmath76 is the element abundance , @xmath77 is the position along the line of sight , and @xmath78 is the line wavelength . for the source function
we adopt the same expression used for electron scattering @xmath25 since thermalization is unimportant and the size of the x - ray source is small compared with the other length scales of interest for hmxbs . results of our simple single line calculation are shown in figure [ fig8 ] .
this shows the line profile as a function of orbital phase for a system viewed edge on ( @xmath29 ) . for each phase we display the luminosity , with the transmitted luminosity shown in black and the scattered luminosity shown in green ( bottom panel ) and polarization fraction ( top panel ) .
the polarization angle is not shown because there is no significant dependence on energy or orbital phase when the system is viewed at high inclination ; the polarization is always perpendicular to the orbital plane .
the horizontal axis is energy in units of the wind terminal velocity . at phase 0 ( superior conjunction of the x - ray source )
the shape of the profile in luminosity is similar to p - cygni profiles familiar from uv resonance lines in hot stars : the outflow produces blue - shifted absorption ( shown as negative energy in these figures ) of the continuum .
this absorption is offset from the zero of energy owing to the fact that the continuum source in this case is the compact x - ray source , and the wind speed at the x - ray source is approximately half the terminal speed .
in addition , the x - ray source creates an ionized region within which line scattering can not occur . as a result ,
the absorption occurs in a relatively narrow region of energy near the energy corresponding to the wind terminal velocity .
the scattered emission , on the other hand , is essentially symmetric in energy , since scattered emission comes from a region which is not necessarily along the line of sight to the x - ray source .
projection effects make the scattered emission appear at all energies @xmath79 where @xmath80 is the line rest energy .
the scattered emission is unpolarized at phase 0 for the same reason as in the pure electron scattering case . at phase 0.25 ( quadrature )
the wind is viewed perpendicular to its velocity vector at the x - ray source , and therefore the absorption covers a large fraction of the energy spanned by the wind .
the scattered emission is polarized up to @xmath850@xmath58 ; the degree of polarization is symmetric around the center of the line , owing to the fact that the wind velocity structure is symmetric around the line of centers . at phase 0.5 ( inferior conjunction of the x - ray source )
there is no transmitted flux ( due to occultation by the star ) and the emission is predominantly red - shifted owing to the location of the x - ray ionized zone in the receding part of the wind . in this case , even though the flux is all scattered , the polarization fraction is negligible owing to the circular symmetry of the scattering region of the wind as viewed in the plane of the sky .
the ionization balance in an hmxb wind is determined at each point by the x - ray flux and by the gas density .
the x - ray flux in turn depends on the effects of geometrical dilution and on attenuation . in order to do so
, we utilize results from xstar @xcite at each point in the wind in order to determine the opacity .
this is done along radial rays originating from the x - ray source , so that the opacity at each point is calculated using an ionization parameter which corresponds to the flux transmitted along the corresponding ray .
this single stream transfer treatment is similar to the transfer treatment used by xstar with one difference : xstar uses the local spectral energy distribution ( sed ) to calculation the ionization along a ray ; here we use the same unattenuated sed everywhere for the ionization calculation ( so we can use a stored table ) , but we calculate the ionization parameter self - consistently , i.e. by calculating the transmitted flux at each position and then using that value to calculate the ionization parameter .
this simplification allows for more efficient computation without unduly sacrificing physical realism . this approximation will be least accurate when the column density is highest , i.e. @xmath81 @xmath3 . in such regions ,
the ionization is expected to be low , owing to attenuation , and x - ray scattering will be relatively unimportant . our approximation will tend to produce less attenuation at a given column density than a self - consistent calculation would .
x - ray ionization creates an ensemble of resonance lines in an hmxb wind from the many trace elements such as c , n , o , ne , mg , si , s , and fe . at any point in the wind the ionization depends most sensitively on the ionization parameter , defined in the previous section .
figure [ fig9 ] shows the distribution of ionization parameter produced in a wind with a density distribution which is spherically symmetric around the primary star , as discussed so far .
we adopt a spherical wind with mass loss rate @xmath82 , corresponding to a total wind column density of @xmath83 @xmath3 , and an x - ray source luminosity of @xmath84 erg s@xmath5 .
each spatial region with a distinct ionization parameter will have a different distribution of ion abundances , and hence a unique distribution of resonance line opacity .
these can be calculated under the assumption of ionization equilibrium .
figure [ fig10 ] show examples of the spectrum and the polarization fraction produced by such a distribution from the spherical wind shown in figure [ fig9 ] .
ionization and line opacities were calculated as a function of ionization parameter using the xstar code @xcite .
these were binned into 1000 energy bins over energy from 0.1 ev to 10@xmath85 ev .
since the doppler shifts associated with the wind speed is less than the energy bin size , we adopt an approximate form for the source function each resonance line : @xmath86 where the quantity @xmath87 takes into account the fact that the line does not conver the entire energy bin .
this treatment assumes that the line optical depths are not large , which is an adequate approximation for our situation .
depolarization effects at large sobolev optical depths associated with multiple scatterings are not taken into account .
these calculations serve to illustrate the magnitude of the polarization effects expected from resonance scattering in hmxbs .
simulations suitable for quantitatively diagnosing the wind or the x - ray source properties will require reexamination of these assumptions . results of spectra and polarization fractions are shown in figure [ fig10 ] .
these demonstrate that the resonance lines cover a significant fraction of the x - ray energy band and that they can scatter a significant flux of x - rays , and create polarization fractions as high as 0.5 , much greater than would be produced by electron scattering alone .
a difference between resonance scattering and electron scattering is that resonance scattering in a given line occurs within a relatively narrow spatial region where the parent ion is most abundant . for x - ray lines ,
these are most likely to be approximately spherical surface surrounding the x - ray source .
this region tends to produce a small polarization which is almost independent of the viewing angle .
thus resonance scattering produces weaker modulation of the polarization with orbital phase than does electron scattering .
it is also worth noting that the effects of scattering can be either polarizing or not polarizing , depending on the scattering angle .
also , lines which appear in absorption in the spectrum can have enhanced polarization in their troughs owing to the presence of scattered light in the residual intensity . in the spectrum shown in figure [ fig10 ]
the polarization in the absorption lines is generally lower than in the emission features , or in the adjacent continuum .
this is due to the fact that the scattered light in the residual intensity is forward scattered , and so is not polarized because of the scattering angle .
the stellar wind in hmxbs is not spherically symmetric .
physical processes which affect the wind and accretion flow in hmxbs include : the three - dimensional geometry of the flow ( i.e. the flow both in and out of the orbital plane ) , rotational forces , the influence of gravity and radiation pressure from both stars in the binary , transport of the radiation from the stars into the flow and the reprocessed radiation out of the flow , x - ray heating and ionization , and departures from thermal equilibrium due to advection and adiabatic heating and cooling .
the dynamics of x - ray heated winds have been discussed by @xcite , and @xcite , and multi - dimensional models have been calculated by @xcite in this section we illustrate the effects of the wind hydrodynamics on the polarization .
we use two sample wind hydrodynamic models calculated using a numerical model similar to that described in @xcite but extended to three dimensions .
the hydrodynamic models are computed on one hemisphere of a spherical grid , assuming reflection symmetry about the orbital plane . a non - uniform grid of 448 ( r ) by 128 ( @xmath88 ) by 512(@xmath89 )
zones is used with highest resolution near the surface of the primary star and in the vicinity of the accreting neutron star .
this model calculates the wind dynamics in three dimensions taking into account the radiative driving by the uv / optical radiation from the primary , and also the gravity of the star and the compact x - ray source . a fixed x - ray luminosity of 10@xmath90 erg / s
is used to calculate x - ray heating using the approximate formulae given in blondin ( 1994 ) and the effects of x - ray ionization on the dynamics via changes in the uv radiation force multiplier .
the two models have identical dimensions : primary radius 2.4 @xmath91 cm and orbital separation 3.6 @xmath91 cm .
they differ in their mass loss rates , which are 4 @xmath92 and 1.7 @xmath93 .
the maximum wind speed in both models is approximately 1600 km s@xmath5 .
a plot showing the density contours and velocity vectors is shown in figure [ fig11 ] .
this clearly shows the influence of the gravity of the compact object in creating a region of higher density and non - radial wind flow in the vicinity of the x - ray source .
the distorted stellar wind of the hydrodynamic models is illustrated with velocity vectors in the orbital plane and a series of transparent density isosurfaces .
right panel is the model with a higher mass loss rate ( mdot = 1.7e-6 ) ; left panel is a lower mass loss rate ( mdot = 4.0e-7 ) .
the lowest density isosurface corresponds to a density of 4e9 @xmath94 in both panels .
velocity vectors are not shown for values less than 800 km / s .
the high mass loss rate model has a denser spherical component of the wind , but a less pronounced accretion wake .
the low mass loss rate model has a larger volume of wind moving at low velocity due to photoionization of the wind in the vicinity of the x - ray source .
the result is a denser wind coming off the primary along the line of centers of the binary system and a larger , denser accretion wake that wraps more tightly around the primary star .
we apply the same calculation of x - ray scattering to the hydrodynamic wind as was done in section [ sec3 ] .
we assume an x - ray source luminosity of @xmath95 erg s@xmath5 and a @xmath96=2 power law ionizing spectrum .
the distribution of ionization parameter in the orbital plane is shown in figure [ fig12 ] .
this shows a spiral structure , owing to a corresponding structure in the gas density .
figures [ fig13 ] shows the spectra and polarization produced by the hydrodynamic wind model with @xmath97 at orbital phases 0 , 0.25 and 0.5 .
comparison with figure [ fig10 ] shows that the hydrodynamic model produces a spectrum which is similar to that from a smooth spherical wind .
however , the degree of polarization is significantly greater , due to the departure from spherical symmetry created by the x - ray source . in particular , the spectrum during eclipse is greater , owing to the fact that the wake structure extends beyond the disk of the primary star during eclipse .
the spherical wind produces fractional polarization averaged over the 0.1 10 kev energy band , of at most 10@xmath58 during eclipse .
the spherical wind produces polarization values which are much less than this at mid - eclipse .
the hydrodynamic wind produces fractional polarization of approximately 21@xmath58 at mid - eclipse .
polarization calculations for hydrodynamic models have been carried out for the two mass loss rates shown in figure [ fig11 ] , and for several choices of ionizing x - ray luminosity .
note that the hydrodynamic models assume a fixed x - ray luminosity ( used to calculate x - ray heating and ionization within the simulation ) of 10@xmath90 erg / s , independent of the x - ray luminosity used to calculate the spectrum and polarization .
moreover , the prescribed x - ray luminosity is not generally consistent with the mass accretion rate derived from the hydrodynamic simulation , and thus these models are not fully self - consistent .
nonetheless , they serve to illustrate the behavior of the polarization and its dependence on wind density and x - ray luminosity
. values for polarization fractions during eclipse are shown in table [ table1 ] .
these are averages of the linear polarization over energy in the range from 0.1 10 kev .
polarization values reflect competing effects . for highly ionized gas ,
the wind is essentially transparent and the polarization is due to thomson scattering .
the polarization fraction is proportional to the thomson depth ( when small ) and also to departures from circular symmetry of the scattering material in the plane of the sky .
partially ionized gas can produce larger polarization , owing to the greater cross section associated with resonance scattering , although each line has very limited spectral range and the ensemble of lines for any given model seldom has a width which exceeds @xmath98 . on the other hand , partially ionized and near - neutral gas
also is affected by photoelectric absorption .
this limits the size of the x - ray scattering region , and tends to reduce the net polarization .
table [ table1 ] shows that , for the limited range of parameters spanned by our models , more highly ionized models tend to produce greater polarization .
that is , the effects of photoelectric absorption offset the effects of resonance scattering in partially ionized gas , leading to very weak dependence of the polarization on x - ray luminosity when the luminosity is not large .
we expect that at very low x - ray luminosities this effect will be even stronger , since the partially ionized zone containing the resonance scattering gas will shrink and become more round , and photoelectric absorption will remove a larger fraction of photons .
our results so far on the polarization produced by wind scattering in hmxbs can be summarized as follows : polarization depends on inclination : high inclination produces variable polarization fraction plus constant angle ; low inclination produces constant polarization fraction plus variable polarization position angle .
the maximum attainable continuum polarization fraction scales approximately proportional to electron scattering optical depth , for depths less than unity ; this is partly a consequence of our single scattering assumption and must be suitably modified if multiple scattering is important . at high inclination , polarization fraction of the scattered radiation ( @xmath99 ) at phase 0 and 0.5 is less than at phases 0.25 or 0.75 ; for a spherical wind the phase 0 and 0.5 polarization is zero .
polarization fraction of the total radiation , including unscattered , is small for the parameters considered here for all orbital phases out of eclipse .
a spherical wind thus produces narrow intervals of high polarization during eclipse but away from phase 0 . resonance line optical depths are greater than for electron scattering , and so can produce greater linear polarization . on the other hand ,
the optical depths are often large , and the x - ray scattering regions tend to be small , i.e. nearly circular around the compact object , producing smaller orbital phase modulation .
the hydrodynamics of the interaction between the wind and the compact object breaks the spherical symmetry and increases the net polarization .
predictions of the polarization signatures for particular known hmxbs require hydrodynamic simulations for each system incoporating known physical parameters : the sizes and masses of the components , the primary wind mass loss rate and the x - ray source luminosity and spectrum .
these would then need to be analyzed using a transfer calculation such as the single scattering calculations we have presented . in this paper
we have presented generic models for the wind densities and x - ray source properties , both for spherical winds and incoporating the hydrodynamic effects of the two gravitating centers .
these generic models can be used to infer , in a very simple way , the polarization signatures expected from various sources .
table [ table2 ] shows the known parameters of the 5 brightest and best - studied hmxbs .
the relevant quantities are the source x - ray luminosity , primary type , mass loss rate , and average x - ray luminosity .
these quantities are taken from @xcite and from @xcite .
we do not include several systems which generally have smaller observed x - ray fluxes or less well constrained properties .
with these quantities we can use our spherical wind models to crudely predict the polarization fractions and orbital phase modulation of the polarization for these sources .
we do this in the following way : we use the results in table [ table1 ] to construct an approximate scaling of mid - eclipse net polarization with the wind mass loss rate and x - ray luminosity .
we then apply this to the known hmxbs in table 1 .
the results are shown graphically in figure [ fig14 ] .
the most relevant quantities for each of the table 1 hmxbs , x - ray luminosity and wind mass - loss rate , are plotted on the axes and the values for each object are plotted as points .
contours of constant predicted mid - eclpise polarization are shown as solid curves , and labeled .
this shows that high polarizations are expected for some objects , those with the strongest winds and weakest x - rays generally .
this demonstrates that polarization fractions in the range from 5 30 @xmath58 are expected at mid - eclipse for these systems .
more detailed information is contained in the spectra and the time variation of the net polarization during eclipse ; interpretation of these signals requires modeling which is tailored to each particular system .
crrr & & @xmath100&erg s@xmath5 + vela x-1&b0.5 iab&7 @xmath105&0.05 + cen x-3&o6.5 ii - iii&2 @xmath101&0.3 + cyg x-1&o9.7 iab&2.5 @xmath105&0.1 + 4u1700 - 37&o6.5 iaf+&1 @xmath106&0.01 + smc x-1&b0.6
iab&5 @xmath101&3 + illustrative analytic results can be derived for an idealized hmxb system : a spherical primary star with radius @xmath107 cm in a circular orbit with an x - ray source with separation @xmath108 .
we view the system from some phase angle @xmath11 relative to the line of centers where @xmath30 corresponds to superior conjunction of the x - ray source .
in addition , the system can have inclination @xmath10 defined such that @xmath29 corresponds to viewing in the orbital plane .
the x - ray source has a luminosity @xmath109 erg s@xmath5 and illuminates the primary star and wind isotropically .
the wind has a mass loss rate @xmath110 yr@xmath5 , a terminal velocity @xmath111 cm s@xmath5 , and a velocity law @xmath112 which depends only on the distance from the center of the star .
if so , the effects of scattering in the wind depend on the wind column density , which can be characterized by the column density @xmath113 and an approximate value for this is @xmath114 where @xmath115 is the density at the x - ray source , @xmath116 is the mean particle weight , and @xmath117 is the hydrogen mass . for the fiducial parameters given above , @xmath118
@xmath94 @xmath119 and @xmath120 @xmath3 @xmath121 .
the ratio of the total intensity of scattered radiation to the total emitted radiation can be characterized by the quantity @xmath122 where @xmath123 is a characteristic scattering cross section and @xmath124 is the thomson cross section .
where @xmath127 are the radial , angular and axial cylindrical coordinates with the axis along y. these are indicated on figure [ fig1 ] . in the case a star located at a distance @xmath65 from the x - ray source , orbiting in a plane with a normal which is inclined at an agle @xmath10 with respect to line of sight , and with an orbital phase described by an angle @xmath128 , this expression can be rewritten in terms of coordinates centered on the star : from equation [ eq4 ] it is apparent that , in the case where the density is independent of @xmath20 , the net polarization is zero and the surfaces of constant polarization in the integrand are circular .
this is expected to be the case for a binary system where the gas density is circularly symmetric around the line of centers , and when viewed at conjunction from the orbital plane ( i.e. @xmath128=0 or @xmath31/2 and @xmath10=@xmath56 ) . in the case of a binary system viewed along the angular momentum axis , i.e. @xmath10=0
, the intensities can be shown to be : in the case of a binary viewed from within the orbital plane , it is straightforward to show that the polarization is zero at phases 0 and 0.5 , i.e. superior and inferior conjunction of the x - ray source . at phase 0.25 or 0.75 ,
i.e. quadrature , the stokes quantities can be written : this can be evaluated analytically in the limit that the density distribution is a thin shell at @xmath143 in which case the net polarization @xmath144 varies from unity for @xmath145 to 0.886 for @xmath146 lamers , h. j. g. l. m. , & cassinelli , j. p. 1999
, introduction to stellar winds , by henny j. g. l. m. lamers and joseph p. cassinelli , pp .
isbn 0521593980 .
cambridge , uk : cambridge university press , june 1999 . , | x - ray astronomy allows study of objects which may be associated with compact objects , i.e. neutron stars or black holes , and also may contain strong magnetic fields .
such objects are categorically non - spherical , and likely non - circular when projected on the sky .
polarization allows study of such geometric effects , and x - ray polarimetry is likely to become feasible for a significant number of sources in the future .
potential targets for future x - ray polarization observations are the high - mass x - ray binaries ( hmxbs ) , which consist of a compact object in orbit with an early type star . in this paper
we show that x - ray polarization from hmxbs has a distinct signature which depends on the source inclination and orbital phase .
the presence of the x - ray source displaced from the star creates linear polarization even if the primary wind is spherically symmetric whenever the system is viewed away from conjunction .
direct x - rays dilute this polarization whenever the x - ray source is not eclipsed ; at mid - eclipse the net polarization is expected to be small or zero if the wind is circularly symmetric around the line of centers .
resonance line scattering increases the scattering fraction , often by large factors , over the energy band spanned by resonance lines .
real winds are not expected to be spherically symmetric , or circularly symmetric around the line of centers , owing to the combined effects of the compact object gravity and ionization on the wind hydrodynamics .
a sample calculation shows that this creates polarization fractions ranging up to tens of percent at mid - eclipse . |
braid groups of the plane were defined by artin in 1925 @xcite , and further studied in @xcite .
braid groups of surfaces were studied by zariski @xcite , and were later generalised using the following definition due to fox @xcite .
let @xmath0 be a compact , connected surface , and let @xmath11 .
we denote the set of all ordered @xmath12-tuples of distinct points of @xmath0 , known as the _ @xmath13 configuration space of @xmath0 _ , by :
@xmath14 configuration spaces play an important rle in several branches of mathematics and have been extensively studied , see @xcite for example .
the symmetric group @xmath15 on @xmath12 letters acts freely on @xmath16 by permuting coordinates .
the corresponding quotient space will be denoted by @xmath17 .
notice that @xmath16 is a regular covering of @xmath17 .
the _ @xmath13 pure braid group @xmath2 _ ( respectively the _ @xmath13 braid group @xmath18 _ ) is defined to be the fundamental group of @xmath16 ( respectively of @xmath17 ) .
if @xmath19 , then we may define a homomorphism @xmath20 $ ] induced by the projection @xmath21 $ ] defined by @xmath22 .
representing @xmath23 geometrically as a collection of @xmath24 strings , @xmath5 corresponds to forgetting the last @xmath6 strings .
* we adopt the convention , that unless explicitly stated , all homomorphisms @xmath25 in the text will be this one*. if @xmath0 is the @xmath26-disc ( or the plane @xmath27 ) , @xmath28 and @xmath29 are respectively the classical artin braid group @xmath30 and pure braid group @xmath31 @xcite .
if @xmath0 is without boundary , fadell and neuwirth study the map @xmath9 , and show ( ( * ? ? ? * theorem 3 ) ) that it is a locally - trivial fibration .
the fibre over a point @xmath32 of the base space is @xmath33 which we interpret as a subspace of the total space via the map @xmath34 $ ] defined by @xmath35 applying the associated long exact sequence in homotopy , we obtain the _ pure braid group short exact sequence of fadell and neuwirth _ : @xmath36 where @xmath37 if @xmath0 is the sphere @xmath38 @xcite , @xmath39 if @xmath0 is the real projective plane @xmath40 @xcite , and @xmath41 otherwise @xcite , and where @xmath42 and @xmath5 are the homomorphisms induced by the maps @xmath43 and @xmath9 respectively
. the short exact sequence ( [ eq : split ] ) has been widely studied , and may be employed for example to determine presentations of @xmath2 ( see section [ sec : pres ] ) , its centre , and possible torsion .
it was also used in recent work on the structure of the mapping class groups @xcite and on vassiliev invariants for surface braids @xcite . in the case of @xmath31 , and taking @xmath44
, @xmath45 is a free group of rank @xmath12 . the short exact sequence ( [ eq : split ] ) splits for all @xmath41 , and so @xmath31 may be described as a repeated semi - direct product of free groups .
this decomposition , known as the ` combing ' operation , is the principal result of artin s classical theory of braid groups @xcite , and yields normal forms and a solution to the word problem in @xmath30 .
more recently , it was used by falk and randell to study the lower central series and the residual nilpotence of @xmath46 @xcite , and by rolfsen and zhu to prove that @xmath46 is bi - orderable @xcite .
the problem of deciding whether such a decomposition exists for braid groups of surfaces is thus fundamental .
this was indeed a recurrent and central question during the foundation of the theory and its subsequent development during the 1960 s @xcite .
if the fibre of the fibration is an eilenberg - maclane space then the existence of a section for @xmath5 is equivalent to that of a cross - section for @xmath9 @xcite ( cf .
@xcite ) . but with the exception of the construction of sections in certain cases ( for @xmath38 @xcite and
the @xmath26-torus @xmath47 @xcite ) , no progress on the possible splitting of ( [ eq : split ] ) was recorded for nearly forty years . in the case of orientable surfaces without boundary of genus at least two ,
the question of the splitting of ( [ eq : split ] ) which was posed explicitly by birman in 1969 @xcite , was finally resolved by the authors , the answer being positive if and only if @xmath10 @xcite . as for the non - orientable case ,
the braid groups of @xmath40 were first studied by van buskirk @xcite , and more recently by wang @xcite and the authors @xcite . for @xmath10 , we have @xmath48 .
van buskirk showed that for all @xmath39 , neither the fibration @xmath49 $ ] nor the homomorphism @xmath50 $ ] admit a cross - section ( for @xmath9 , this is a manifestation of the fixed point property of @xmath40 ) , but that the fibration @xmath51 $ ] admits a cross - section , and hence so does the corresponding homomorphism @xmath5 .
using coincidence theory , we showed that for @xmath52 and @xmath53 , neither the fibration nor the short exact sequence ( [ eq : split ] ) admit a section @xcite . in @xcite
, we gave a complete answer to the splitting problem for @xmath40 : if @xmath54 , the homomorphism @xmath55 $ ] and the fibration @xmath56 $ ] admit a section if and only if @xmath57 and @xmath44 .
in other words , van buskirk s values ( @xmath57 and @xmath44 ) are the only ones for which a section exists ( both on the geometric and the algebraic level ) . in this paper
, we study the splitting problem for compact , connected non - orientable surfaces without boundary and of genus @xmath1 ( every non - orientable compact surface @xmath0 without boundary is homeomorphic to the connected sum of @xmath58 copies of @xmath40 , @xmath59 being the _ genus _ of @xmath0 ) . in the case of the klein bottle @xmath60 ( @xmath61 )
, the existence of a non - vanishing vector field implies that there always exists a section , both geometric and algebraic ( cf .
our main result is : [ th : nosplitg ] let @xmath0 be a compact , connected , non - orientable surface without boundary of genus @xmath1 , and let @xmath54 . then the homomorphism @xmath62 $ ] and the fibration @xmath7 $ ] admit a section if and only if @xmath10 . applying theorem [ th : nosplitg ] and the results of @xcite , we may solve completely the splitting problem for surface pure braid groups : [ th : complete ] let @xmath54 and @xmath63 .
let @xmath64 be a compact , connected surface possibly with boundary , let @xmath65 be a finite subset in the interior of @xmath64 , let @xmath66 , and let @xmath67 $ ] be the standard projection .
a. if @xmath68 or if @xmath0 has non - empty boundary then @xmath5 admits a section for all @xmath6 and @xmath12 .
b. suppose that @xmath69 and that @xmath0 is without boundary .
then @xmath5 admits a section if and only if one of the following conditions holds : a. @xmath0 is @xmath38 , the @xmath26-torus @xmath47 or the klein bottle @xmath60 ( for all @xmath6 and @xmath12 ) .
b. @xmath70 , @xmath57 and @xmath44 . c. @xmath71 and @xmath10 .
the rest of the paper is organised as follows . in section [ sec : pres ] , we determine a presentation of @xmath2 ( theorem [ th : basicpres ] ) . in section [ sec :
setup ] , we study the consequences of the existence of a section in the case @xmath44 and @xmath39 , _ i.e. _ @xmath72 $ ]
. the general strategy of the proof of theorem [ th : nosplitg ] is based on the following remark .
suppose that ( [ eq : split ] ) splits .
if @xmath73 is any normal subgroup of @xmath74 contained in @xmath45 , the quotiented short exact sequence @xmath75 must also split . in order to obtain a contradiction , we seek such a subgroup @xmath73 for which this short exact sequence does _ not _ split .
however the choice of @xmath73 needed to achieve this may be somewhat delicate : if @xmath73 is too ` small ' , the structure of the quotient @xmath76 remains complicated ; on the other hand , if @xmath73 is too ` large ' , we lose too much information and can not reach a conclusion . in section [ sec : proofthm ] , we first show that we may reduce to the case @xmath44 , and then go on to prove theorem [ th : nosplitg ] using the analysis of section [ sec : setup ] . as we shall see in section [ sec : proofthm ] , it suffices to take @xmath73 to be abelianisation of @xmath45 , in which case the quotient @xmath77 is a free abelian group
. we will then deduce theorem [ th : complete ] .
this work took place during the visit of the second author to the departmento de matemtica do ime - universidade de so paulo during the periods 14 29 april 2008 , 18 july
8 august 2008 and 31 october
10 november 2008 , and of the visit of the first author to the laboratoire de mathmatiques nicolas oresme , universit de caen during the period 21 november 21 december 2008 .
this work was supported by the international cooperation usp / cofecub project n 105/06 , and by the cnrs / cnpq project n 21119 .
let @xmath78 be a compact , connected , non - orientable surface without boundary of genus @xmath79 .
if @xmath11 and @xmath80 is a topological disc , the inclusion induces a homomorphism @xmath81 $ ] .
if @xmath82 then we shall denote its image @xmath83 simply by @xmath84 . for @xmath85 , we consider the following elements of @xmath2 : @xmath86 where @xmath87 are the standard generators of @xmath88 .
the geometric braid corresponding to @xmath89 takes the @xmath90 string once around the @xmath91 string in the positive sense , with all other strings remaining vertical . for each @xmath92 and @xmath93
, we define a generator @xmath94 which is represented geometrically by a loop based at the @xmath95 point and which goes round the @xmath96 twisted handle .
these elements are illustrated in figure [ fig : gens ] that represents @xmath0 minus a disc .
a presentation of the braid groups of non - orientable surfaces was originally given by scott @xcite .
other presentations were later obtained in @xcite . in the following theorem
, we derive another presentation of @xmath2 .
[ th : basicpres ] let @xmath0 be a compact , connected , non - orientable surface without boundary of genus @xmath79 , and let @xmath11 .
the following constitutes a presentation of the pure braid group @xmath2 : 1 .
@xmath89 , @xmath97 , and @xmath94 , where @xmath98 and @xmath93 .
2 . a. [ it : rel1 ] the artin relations between the @xmath89 emanating from those of @xmath99 : @xmath100 b. [ it : rel2 ] for all @xmath97 and @xmath101 , @xmath102 c. [ it : rel3 ] for all @xmath103 , the ` surface relations ' @xmath104 .
d. [ it : rel4 ] for all @xmath97 , @xmath98 , @xmath105 , and @xmath93 , @xmath106 we apply induction and standard results concerning the presentation of an extension ( see ( * ? ? ?
* theorem 1 , chapter 13 ) ) .
the proof generalises that of @xcite for @xmath40 , and is similar in spirit to that of @xcite . first note that the given presentation is correct for @xmath10 ( @xmath107 has a presentation @xmath108 ) .
so let @xmath41 , and suppose that @xmath2 has the given presentation .
taking @xmath44 in ( [ eq : split ] ) , we have a short exact sequence : @xmath109 in order to retain the symmetry of the presentation , we take the free group @xmath45 to have the following one - relator presentation : @xmath110 together with these generators of @xmath45 , the elements @xmath89 , @xmath97 , and @xmath94 , @xmath98 and @xmath111 , of @xmath74 ( which are coset representatives of the generators of @xmath2 ) form the given generating set of @xmath74
. there are three classes of relations of @xmath74 which are obtained as follows .
the first consists of the single relation @xmath112 of @xmath45 .
the second class is obtained by rewriting the relators of the quotient in terms of the coset representatives , and expressing the corresponding element as a word in the generators of @xmath45 . in this way , all of the relations of @xmath2 lift directly to relations of @xmath74 , with the exception of the surface relations which become @xmath113 along with the relation of @xmath45
, we obtain the complete set of surface relations ( relations ( [ it : rel3 ] ) ) for @xmath74 .
the third class of relations is obtained by rewriting the conjugates of the generators of @xmath45 by the coset representatives in terms of the generators of @xmath45 : a. [ it : relex1 ] for all @xmath97 and @xmath114 , @xmath115 b. [ it : relex2 ] @xmath116 for all @xmath97 and @xmath111 . c. [ it : relex3 ] for all @xmath103 and @xmath117 , @xmath118 d. [ it : relex4 ] for all @xmath119 and @xmath111 , @xmath120 then relations ( [ it : rel1 ] ) for @xmath74 are obtained from relations ( [ it : rel1 ] ) for @xmath29 and relations ( [ it : relex1 ] ) , relations ( [ it : rel2 ] ) for @xmath74 are obtained from relations ( [ it : rel2 ] ) for @xmath29 and relations ( [ it : relex3 ] ) , and relations ( [ it : rel4 ] ) for @xmath74 are obtained from relations ( [ it : rel4 ] ) for @xmath29 , and relations ( [ it : relex2 ] ) and ( [ it : relex4 ] ) .
for the whole of this section , we suppose that @xmath1 and @xmath39 . by theorem [ th : basicpres ] , @xmath121 is generated by the union of the @xmath89 , @xmath97 , and of the @xmath94 , @xmath98 , @xmath93 .
let us consider the homomorphism @xmath122 $ ] . in this section ,
we suppose that @xmath5 admits a section , denoted by @xmath123 .
applying ( [ eq : split ] ) , we thus have a split short exact sequence @xmath124 & k \ar[r ] & p_{n+1}(m_g ) \ar@<4pt>[r]^-{p_{\ast } } & p_n(m_g ) \ar@<4pt>@{-->}[l]^-{s_{\ast } } \ar[r ] & 1,}\ ] ] where @xmath125 is a free group of rank @xmath126 , generated by @xmath127 , and subject to the relation @xmath128 let @xmath129 $ ] be the commutator subgroup of @xmath130 . then @xmath131 is a free abelian group of rank @xmath126 . in what follows , we shall not distinguish notationally between the elements of @xmath130 and those of @xmath131 .
the quotient group @xmath131 thus has a basis @xmath132 and the relation @xmath133 holds in the abelian group @xmath131 .
since @xmath73 is normal in @xmath134 and @xmath5 admits a section , it follows from equation ( [ eq : fnmgnp2 ] ) that we have a split short exact sequence @xmath135 & k / h \ar[r ] & p_{n+1}(m_g)/h \ar@<4pt>[r]^-{\overline{p } } & p_n(m_g ) \ar@<3pt>@{-->}[l]^-{\overline{s } } \ar[r ] & 1,}\ ] ] where @xmath136 is the homomorphism induced by @xmath5 , and @xmath137 is the induced section .
consider the subset @xmath138 of @xmath139 .
if @xmath140 then @xmath141 , and so @xmath142 . then the integer coefficients @xmath143 , where @xmath144 and @xmath145 , are ( uniquely ) defined by the equations : @xmath146 there is an equation for each element of @xmath147 .
most of the elements of @xmath147 commute with the elements of the basis @xmath148 of @xmath131 given in equation ( [ eq : basisb ] ) .
we record the list of conjugates of such elements for later use .
in what follows , @xmath149 , @xmath150 and @xmath151 . in @xmath131 , we have @xmath152 ( this follows from equation ( [ eq : bijrel ] ) and the fact that the elements @xmath153 , @xmath154 , belong to @xmath131 and so commute pairwise ) , and @xmath155 by equation ( [ eq : relspn ] ) .
thus @xmath89 belongs to the centraliser of @xmath131 in @xmath139 . also by equation ( [ eq : relspn ] ) , we have
@xmath156 so @xmath157 where @xmath158 is the kronecker delta . by equation ( [ eq : relsrhoik ] )
, we have @xmath159 so @xmath160 combining equations ( [ eq : commbijgen ] ) and ( [ eq : commrr ] ) , we obtain @xmath161 so @xmath162 furthermore , by equation ( [ eq : commbijgen ] ) , @xmath163 commutes with @xmath164 , and therefore @xmath165 hence @xmath163 also belongs to the centraliser of @xmath131 in @xmath139 . from equations ( [ eq : commbijgen ] ) and ( [ eq : commrr ] ) , we obtain the following relations : @xmath166 and @xmath167 setting @xmath168 for all @xmath169 and @xmath170 yields : @xmath171 equations ( [ eq : rhojlcomm1 ] ) and ( [ eq : rhojlcomm2 ] ) will be employed repeatedly in the ensuing calculations .
we now investigate the images under @xmath137 of some of the relations ( [ it : rel2])([it : rel4 ] ) of theorem [ th : basicpres ] ( it turns out that the analysis of the other relations , including ( [ it : rel1 ] ) , will not be necessary for our purposes ) .
a. let @xmath97 and @xmath101 .
we examine the three possible cases of equation ( [ eq : defcoeffs ] ) ( relation ( [ it : rel2 ] ) of theorem [ th : basicpres ] ) .
a. @xmath172 : then @xmath173 in @xmath121 .
the respective images under @xmath137 are : @xmath174 and @xmath175 the relation @xmath173 in @xmath134 implies that @xmath176 . comparing coefficients of the elements of @xmath148 in @xmath131 ( cf .
equation ( [ eq : basisb ] ) ) , if @xmath177 , we have @xmath178 while if @xmath179 , applying equation ( [ eq : surfacerelgen ] ) yields @xmath180 and thus equation ( [ eq : relbcommi ] ) also holds for @xmath179 .
so for all @xmath97 and @xmath181 , @xmath182 b. @xmath183 : then @xmath184 in @xmath121 for all @xmath97 and @xmath185 .
the respective images under @xmath137 are : @xmath186 and @xmath187 since @xmath184 in @xmath134 , we obtain @xmath188 if @xmath177
then all of the terms in equation ( [ eq : kequall ] ) are expressed in terms of the basis @xmath148 of @xmath131 of equation ( [ eq : basisb ] ) , and so for all @xmath189 , @xmath190 if @xmath179 then substituting for @xmath191 in equation ( [ eq : kequall ] ) using equation ( [ eq : surfacerelgen ] ) and comparing coefficients in @xmath131 of the elements of @xmath148 yields @xmath192 but @xmath168 , so for all @xmath193 and @xmath185 , @xmath194 c. @xmath195 : then @xmath196 in @xmath121 .
the respective images under @xmath137 are : @xmath197 and @xmath198 since @xmath196 in @xmath134 , we see that @xmath199 if @xmath177 , it follows by comparing coefficients of the elements of @xmath148 in @xmath131 that for all @xmath200 and @xmath201 , @xmath202 if @xmath179 then @xmath203 and comparing coefficients of the elements of @xmath148 in @xmath131 , we observe that equations ( [ eq : rhocomm ] ) also hold if @xmath179 .
so for all @xmath149 and @xmath201 , @xmath204 b. let @xmath103
. then @xmath104 in @xmath121 by relation ( [ it : rel3 ] ) of theorem [ th : basicpres ] . for @xmath93 , note that @xmath205 as we saw in equations ( [ eq : commrhokr1 ] ) and ( [ eq : commrhokr2 ] ) , @xmath206 belongs to the centraliser of @xmath131 in @xmath207 , so @xmath208 further , @xmath209 now in @xmath139 , @xmath210 , hence @xmath211 thus for all @xmath103 , @xmath212
in this section , we use the calculations of section [ sec : setup ] to prove theorem [ th : nosplitg ] , from which we shall deduce theorem [ th : complete ] .
the case @xmath10 was treated in theorem 1 of @xcite , using the fact that if @xmath78 , where @xmath213 , then @xmath0 is homeomorphic to the connected sum of one or two copies of @xmath40 with a compact , orientable surface without boundary of genus at least one .
conversely , suppose that there exist @xmath19 and @xmath214 for which the homomorphism @xmath67 $ ] admits a section .
we shall argue for a contradiction . by ( * ? ? ?
* proposition 3 ) , it suffices to consider the case @xmath44 .
we first analyse the general structure of the coefficients @xmath143 defined by equation ( [ eq : defcoeffs ] ) .
a. [ it : itema ] taking @xmath179 in equation ( [ eq : etaikj ] ) implies that @xmath215 for all @xmath193 and @xmath216 .
b. [ it : itemb ] by equation ( [ eq : aage ] ) , @xmath217 for all @xmath97 and @xmath201 .
taking @xmath179 , we obtain @xmath218 since @xmath168 by definition and @xmath219 for all @xmath193 and @xmath220 by equation ( [ eq : alinr ] ) .
+ it thus follows from ( [ it : itema ] ) and ( [ it : itemb ] ) that @xmath221 c. by equation ( [ eq : alinr ] ) , @xmath222 for all @xmath193 and @xmath185 .
so @xmath223 d. [ it : itemd ] by equation ( [ eq : etaikj ] ) , for all @xmath181 and @xmath97 , we have @xmath224 using equation ( [ eq : gamikl0 ] ) .
so by taking @xmath225 we obtain @xmath226 e. [ it : iteme ] by equation ( [ eq : aage ] ) @xmath227 for all @xmath97 and @xmath201 . but
@xmath215 by equation ( [ eq : gamikl0 ] ) , and @xmath228 by equation ( [ eq : alij1 ] ) if @xmath229 and by equation ( [ eq : alinr ] ) if @xmath179 .
setting @xmath230 , it follows that @xmath231 by ( [ it : itemd ] ) and ( [ it : iteme ] ) we thus have @xmath232 f. [ it : itemf ] suppose that @xmath233 .
then @xmath234 so taking @xmath225 , we have @xmath235 for all @xmath236 .
further , for all @xmath237 , @xmath238 + hence it follows from equation ( [ eq : etaikj0a ] ) and ( [ it : itemf ] ) that @xmath239 g. from equation ( [ eq : betins ] ) , we obtain @xmath240 h. by equations ( [ eq : betijj ] ) and ( [ eq : etaikj0 ] ) , we see that @xmath241 i. [ it : iteml ] by equations ( [ eq : betiji ] ) and ( [ eq : etaikj0 ] ) , we obtain @xmath242 analysing equation ( [ eq : surface ] ) , we are now able to complete the proof of theorem [ th : nosplitg ] as follows .
let @xmath243 .
then the coefficient of @xmath244 yields : @xmath245 now @xmath246 and @xmath247 so using equation ( [ eq : gamikk ] ) , equation ( [ eq : bin1 ] ) becomes @xmath248 summing over all @xmath249 , and setting @xmath250 and @xmath251 , we obtain @xmath252 now let @xmath253 , and let @xmath254 . since @xmath255 , the coefficient of @xmath256 in equation ( [ eq : surface ] ) yields : @xmath257 hence @xmath258 .
summing over all @xmath259 , we obtain @xmath260 equating equations ( [ eq : ng2 g ] ) and ( [ eq : ng2ga ] ) , we see that @xmath261 .
since @xmath1 , it follows that @xmath262 , and therefore @xmath263 by equation ( [ eq : ng2 g ] ) .
this yields a contradiction to the fact that @xmath264 is an integer , and thus completes the proof of theorem [ th : nosplitg ] .
although some of the relations derived in ( [ it : itema])([it : iteml ] ) do not exist if @xmath57 , one may check that the above analysis from equation ( [ eq : bin1 ] ) onwards is also valid in this case ( with @xmath265 and @xmath266 ) .
a. if @xmath267 then the result follows applying the methods of the proofs of proposition 27 and theorem 6 of @xcite .
if @xmath69 and @xmath0 has non - empty boundary , let @xmath268 be a boundary component of @xmath0
. then @xmath269 is homeomorphic to a compact surface with a single point deleted ( which is the case @xmath270 ) , so ( [ eq : split ] ) splits for @xmath271 .
the inclusion of @xmath271 in @xmath0 not only induces a homotopy equivalence between @xmath0 and @xmath271 , but also a homotopy equivalence between their @xmath272 configuation spaces
. therefore their @xmath272 pure braid groups are isomorphic , and the sequence ( [ eq : split ] ) for @xmath0 splits if and only it splits for @xmath271 .
b. suppose that @xmath69 and that @xmath0 is without boundary .
if @xmath273 , @xmath44 and @xmath274 then the statement follows from @xcite .
the geometric construction of fadell may be easily generalised to all @xmath19 . if @xmath275 , the result is obvious since @xmath276 is trivial . if @xmath277 or @xmath60 , the fact that @xmath5 has a section is a consequence of @xcite and the fact that @xmath47 and @xmath60 admit a non - vanishing vector field . if @xmath70 then @xmath5 admits a section if and only if @xmath57 and @xmath44 by @xcite . finally , if @xmath71 then @xmath5 admits a section if and only if @xmath10 by theorem [ th : nosplitg ] for the non - orientable case , and by @xcite for the orientable case .
d. l. gonalves and j. guaschi , on the structure of surface pure braid groups , _ j. pure appl.algebra_ * 182 * ( 2003 ) , 3364 ( due to a printer s error , this article was republished in its entirety with the reference * 186 * ( 2004 ) , 187218 ) .
d. l. gonalves and j. guaschi , classification of the virtually cyclic subgroups of the pure braid groups of the projective plane , _ j. group theory _ , to appear .
j. gonzlez - meneses , new presentations of surface braid groups , _
j. knot theory and its ramifications _ * 10 * ( 2001 ) , 431451 . | let @xmath0 be a compact , connected non - orientable surface without boundary and of genus @xmath1 .
we investigate the pure braid groups @xmath2 of @xmath0 , and in particular the possible splitting of the fadell - neuwirth short exact sequence @xmath3 where @xmath4 , and @xmath5 is the homomorphism which corresponds geometrically to forgetting the last @xmath6 strings .
this problem is equivalent to that of the existence of a section for the associated fibration @xmath7 $ ] of configuration spaces , defined by @xmath8 .
we show that @xmath9 and @xmath5 admit a section if and only if @xmath10 .
together with previous results , this completes the resolution of the splitting problem for surfaces pure braid groups . |
in this paper we study the spaces of conformal blocks for one dimensional lattice vertex operator algebras . spaces of conformal blocks or modular functor ( see @xcite ) play an important role in conformal field theory .
they can be defined for a wide class of vertex operator algebras and they parameterizes the states of the corresponding field theory . to avoid cumbersome notion
let us describe our approach for affine algebras ( which corresponds to the wznw theory ) .
first , recall the definition of the space of conformal blocks .
let @xmath1 be a simple lie algebra , by @xmath2 denote the central extension of the lie algebra of @xmath1valued laurent polynomials .
fix a level @xmath3 , that is , identify the central element of @xmath2 with the positive integer @xmath3 . as a geometric part of the data consider a complex curve @xmath4 and a set of pairwise distinct points @xmath5 with a local coordinate in a small vicinity of each point .
then define the lie algebra @xmath6 of @xmath1-valued meromorphic functions on @xmath4 with poles only at @xmath7 .
clearly we have the inclusion @xmath8 mapping any @xmath1-valued function to its laurent expansions at the points @xmath9 .
another part of the data is a set of integrable representations @xmath10 of @xmath2 at level @xmath3 .
their external tensor product is a representation of @xmath11 and , therefore , of @xmath12 .
then the space of conformal block is defined as the space of coinvariants @xmath13 i.e. the quotient of @xmath14 by the action of @xmath15 .
it is well known that these spaces are finite - dimensional and form a flat family over the module space of curves with punctures . in @xcite
another construction is proposed .
let , as above , @xmath4 be a curve , @xmath9 and @xmath16 be pairwise distinct points on @xmath4 .
then consider the subalgebra @xmath17 of functions with zeroes at @xmath9 and the subalgebra @xmath18 of functions whose values at @xmath9 belongs to the nilpotent borel subalgebra @xmath19 .
then for any integrable representation @xmath20 at level @xmath3 we have the natural isomorphisms ( see @xcite ) @xmath21 where sum goes over all the sets of integrable representations at level @xmath3 and @xmath22 denotes the finite - dimensional representation of @xmath1 generated by the highest vector of @xmath23 .
so the space of coinvariants of a single representation @xmath20 contains the spaces of conformal blocks where @xmath20 appears .
an advantage of this construction is that these spaces of coinvariants is well defined if some of the points @xmath24 coincides . in this case
we just consider zeroes with multiplicities . for multiple points we have not isomorphisms , , but it is easy to show that these spaces form a coherent sheaf over the variety @xmath25 of sets of points on @xmath4 .
first question is whether the dimension of the space of coinvariants preserves when some points coincides .
indeed , this question is local and it is enough to check this when @xmath4 is the riemann sphere .
the answer is positive for @xmath26 ( see @xcite for @xmath27 and @xcite for @xmath28 ) and negative for @xmath29 .
also there is the positive answer for the similar question about the @xmath30-minimal models of the virasoro algebras ( see @xcite ) .
but the general case is far from been studied . here
we prove the equality of dimensions in the case of one dimensional lattice voas .
another question is about the natural filtration on these spaces .
namely , the affine lie algebra @xmath1 and its integrable representations are graded by the action of the energy element .
but the subalgebras defined above are not homogeneous ( except the case of riemann sphere and @xmath31 ) , hence the spaces of coinvariants can be considered only as filtered spaces .
nevertheless we consider the hilbert polynomials @xmath32 of the adjoint graded spaces and interpret them as a @xmath0-analog of the correspondent verlinde numbers ( which are dimensions of spaces of conformal blocks ) .
this object is interesting even in the case of the riemann sphere ( and well studied only in this case ) . in @xcite
there is a conjecture that expresses @xmath32 in terms of generalized @xmath0-supernomial coefficients . for @xmath30-minimal models of the virasoro
algebras the polynomials @xmath32 take part in the andrews - gordon identities and therefore can be decomposed into sums of @xmath0-multinomial coefficients as well . in this paper
we prove a formula ( see theorem [ th : char ] , ( ii ) ) which expresses @xmath32 for one - dimensional lattice voas as a sum of @xmath0-supernomial coefficients . note that this formula can be considered as another definition of @xmath0-supernomials .
these two question are connected .
namely , suppose that the dimension preserves when the points coincides .
then in the case of the riemann sphere it is enough to find @xmath32 for the graded subalgebra with @xmath31 .
it makes possible to degenerate the algebra together with the subalgebra into an abelian one or a rather simple voa and obtain a fermionic type ( in terms of @xcite ) formula for @xmath32 .
such a formula exists for @xmath33 ( see @xcite , @xcite , @xcite ) , for @xmath30-minimal models of the virasoro algebra ( see @xcite ) and for one - dimensional lattice voas ( see theorem [ th : char ] , ( i ) ) . in more general cases ( @xmath34-algebras , lattice voas , etc . )
it is not evident which subalgebras should be chosen instead of @xmath35 and @xmath36 .
let us explain what we are doing in the case of affine lie algebras . introduce some `` elementary '' graded subalgebras @xmath37 . as the subalgebras are graded , they consist of @xmath1-valued functions with some local conditions at the zero point
. then define the `` fused '' subalgebra @xmath38 of functions with pole only in @xmath16 and local conditions in @xmath7 prescribed by the subalgebras @xmath39 .
then we have @xmath40 in particular , the subalgebra @xmath41 is fused from @xmath42 copies of the affine nilpotent borel subalgebra @xmath43 and the subalgebra @xmath44 is fused from @xmath42 copies of the affine nilpotent parabolic subalgebra @xmath45 , so and follows from .
the paper is organized as follows . in sec .
[ sec : lvoa ] we recall some facts about lattice vertex operator algebras ( `` voa '' ) and their representations . in sec . [ sec : ass ] we introduce an associative algebra that acts in each representation of the voa .
in this way representations of the voa can be obtained as induced representations of the associative algebra . in sec . [
sec : verlinde ] we introduce spaces of coinvariants and give basic statements about their dimensions . in sec .
[ sec : element - sub ] elementary subalgebras are introduced and the fused subalgebras are studied . sec .
[ sec : algebras ] and [ sec : combinator ] are rather technical . in sec .
[ sec : algebras ] the character of some auxiliary space of coinvariants is calculated . in sec .
[ sec : combinator ] we show that characters of these auxiliary spaces can be expressed in terms of @xmath0-supernomial coefficients .
this allows us to prove the main results of the paper in sec .
[ sec : main - result ] .
we use the standard notation for the @xmath0-binomial coefficients @xmath46_{q } } = \begin{cases } \frac{(q)_n}{(q)_{n - m}(q)_m},&n\geq m\geq0,\\ 0,&\text{otherwise } , \end{cases}\qquad\text{where}\quad ( q)_n=(1-q)\dots(1-q^n).\ ] ]
recall that a lattice vertex operator algebra ( voa ) is determined by the following data : @xmath47 , @xmath48 , @xmath49 , where @xmath48 is a scalar product in @xmath47 , @xmath49 is a lattice in @xmath47 and @xmath48 takes integer values on vectors from @xmath49 .
such an algebra is generated by vertex operators @xmath50 , @xmath51 . in this paper
we concentrate at one dimensional lattice voas with positively defined scalar product .
let @xmath52 be the basis vector of @xmath49 and @xmath53 , @xmath54 . denote this voa by @xmath55 for @xmath56 .
the algebra @xmath55 is generated by @xmath57 and @xmath58 .
the generators @xmath59 are bosonic ( @xmath60 ) whenever @xmath16 is even and fermionic ( @xmath61 ) otherwise and satisfy the following relations @xmath62 whenever @xmath16 is even and @xmath63 whenever @xmath16 is odd .
introduce currents @xmath64 for @xmath65 as the currents appear in the ope @xmath66 currents @xmath67 and @xmath68 correspond to @xmath69-current and energy
momentum tensor respectively .
choose virasoro @xmath70 so that @xmath71 and @xmath72 be primary fields with conformal weights @xmath73 and @xmath74 respectively @xmath75 the mode decomposition corresponding the choice is @xmath76 this gives the following commutation relations @xmath77}=((p-2)n - i){{\cg^+}}_{i+n}\,,\qquad { \left[\cl_n,\,{{\cg^-}}_i\right]}=-i{{\cg^-}}_{i+n}\,,\qquad { \left[\cl_0,\,{{\ch^{m}}}_i\right]}=-i{{\ch^{m}}}_i\,.\ ] ] the algebra @xmath55 admits the following family of automorphisms ( spectral flow ) @xmath78 @xmath79 where @xmath80 .
the automorphisms with @xmath81 play the role of the affine part of the weyl group .
note that in the case @xmath82 the algebra @xmath55 can be considered as the infinite dimensional clifford algebra , in the case @xmath83 as @xmath84 at level @xmath85 and in the case @xmath86 as the @xmath87-superconformal algebra with the central charge @xmath88 .
recall some facts about representations of lattice voas .
the category of representations of any lattice voa is semisimple , that is , any representation decomposes into the direct sum of irreducible ones .
the irreducible representations are enumerated by the quotient @xmath89 , where @xmath90 is the dual lattice .
namely , the representation corresponding to the element @xmath91 is the space of descendants of @xmath92 , and if two elements of @xmath90 differs by an element of @xmath49 then the corresponding representations are isomorphic .
it is also known that these representations form a minimal model with the verlinde algebra isomorphic to the group ring of @xmath89 .
let us describe it in detail for the one - dimensional case .
denote the representation corresponding to @xmath93 by @xmath94 with @xmath95 .
this representation contains the highest weight vector @xmath96 satisfying the following conditions @xmath97 another description of @xmath94 is as follows @xmath98 where @xmath99 is the fock module over @xmath67 with the highest weight vector @xmath100 . in this realization
@xmath101 act as vertex operators .
we call the representation @xmath102 a _ vacuum _ representation .
representations @xmath94 with @xmath95 compose a minimal model and the verlinde algebra ( the fusion algebra ) is @xmath103 the vacuum representation @xmath102 is the unit in the verlinde algebra .
note also that the module @xmath94 is isomorphic to the contragradient to @xmath104 .
this equips the verlinde algebra with the scalar product .
automorphisms @xmath105 naturally act on the set of representations @xmath94 as @xmath106 in particular , the action of @xmath107 on the set of representations @xmath94 coincides with the fusion with @xmath108 .
each space @xmath94 is bi - graded by operators @xmath109 and @xmath110 .
characters of representations @xmath94 corresponding to this bi - grading are by definition @xmath111 where the trace is calculated over @xmath112 and implies @xmath113 for any graded subspace an analogous series can be considered .
we will denote them by the symbol @xmath114 in front of the symbol of the subspace .
representations of @xmath55 can be considered from a more algebraic viewpoint .
operators @xmath115 act on each representation @xmath94 therefore the associative algebra generated by @xmath115 acts on each @xmath94 . denote this algebra by @xmath116 and let us describe its defining relations .
note first of all that @xmath55 contains a virasoro algebra hence the algebra of vector fields on the circle acts by derivatives ( commutators with @xmath117 ) on @xmath116 . by
@xmath118 denote the representation of the algebra of vector fields in weight @xmath119 tensor fields @xmath120(dz)^\lambda$ ] .
then the space @xmath121 is isomorphic to @xmath122 and @xmath123 is isomorphic to @xmath124 .
this isomorphism is given by the identification @xmath125 , @xmath126 .
generators @xmath127 commute whenever @xmath16 is even , anticommute otherwise and in both cases satisfy quadratic relations @xmath128 for any polynomial @xmath129 of degree not bigger than @xmath130 . the relations between @xmath131 and @xmath132 are also quadratic .
introduce the space @xmath133 of all commutators of @xmath131 and @xmath132 for even @xmath16 and anticommutators for odd @xmath16 .
let us describe @xmath133 as a quotient of the tensor product @xmath134 .
we identify @xmath134 with the space of laurent polynomials @xmath135(dz_1)^{2-p}(dz_2)^0 $ ] .
then introduce the subspaces @xmath136(z_1-z_2)^{p-2-j}(dz_1)^{2-p}(dz_2)^0 , \qquad j\in{\mathbb{z}}.\ ] ] note that the quotient @xmath137 is isomorphic to @xmath138 as a module over the algebra of vector fields .
the ope implies that the space @xmath139 acts on any @xmath94 by zero and @xmath140 acts by a scalar .
the scalar can be found as follows .
we know that @xmath141 , but @xmath142 is the space of @xmath74-forms , so there exists the functional @xmath143 let @xmath144 be the kernel of the mapping @xmath145 . then @xmath146 is a subspace in @xmath134 and we have @xmath147 .
the defining relations in @xmath116 are and relations between @xmath148 and @xmath149 following from . we do nt use this theorem anywhere in the paper , so let us omit the proof .
but this consideration gives us the following description of @xmath150 in terms of laurent polynomials @xmath151 this notion is very convenient for us . in particular , automorphism acts on the space of laurent polynomials as the multiplication with @xmath152 .
one can show that the category of representations of @xmath116 with highest weight is equivalent to the category of representations of @xmath55 .
so irreducible representations of @xmath116 with highest weight are precisely @xmath94 .
define for the algebra @xmath116 some class of representations .
each of these representations contains a cyclic vector that is annihilated by some subspaces in @xmath153 , which we now describe .
let @xmath154 be a @xmath155-dimensional vector with integer components that satisfy @xmath156 we define the subspace @xmath157 as follows @xmath158z_1^{n_+}(dz_1)^{2-p}\subset { \mathbb{c}}[z_1,z_1^{-1}](dz_1)^{2-p}\,,\\ w_-({{\mathbf{n } } } ) = \,&\,{\mathbb{c}}[z_2]z_2^{n_-}(dz_2)^0\ \ \ , \subset { \mathbb{c}}[z_2,z_2^{-1}](dz_2)^0\end{aligned}\ ] ] and the space @xmath159 is the sum ( not direct ! ) of the images of the subspaces @xmath160z_1^{n_j - n_-}z_2^{n_-}(z_1-z_2)^{p - j-2}(dz_1)^{2-p } ( dz_2)^0\ ] ] for @xmath161 under the mapping @xmath162 .
note that if @xmath163 then @xmath159 is the subspace of ( anti)commutators of @xmath164 and @xmath165 .
clearly we have @xmath166 , where @xmath167 the irreducible representation @xmath94 is induced from the subspace @xmath168 .
let @xmath83 .
then we can identify @xmath169 , @xmath170 , @xmath171 , where @xmath172 is the standard basis of @xmath173 .
so our voa is isomorphic to @xmath84 at level @xmath74 .
then we have @xmath174\oplus h\cdot t^{n_0}{\mathbb{c}}[t ] \oplus f\cdot t^{n_-}{\mathbb{c}}[t]\ ] ] so for @xmath85 we cover the cases considered in @xcite .
let @xmath175 be the representation induced from the trivial representation of the subalgebra generated by the subspace @xmath176 .
we know that the representation @xmath175 is a direct sum of @xmath94 .
the multiplicity of @xmath94 in @xmath175 is the dimension of the space @xmath177 , which is isomorphic to the space of @xmath176-invariants in @xmath94 .
it is more convenient to study the dual space to the space of @xmath176-invariants in @xmath94 namely the space of @xmath176-coinvariants in @xmath178 , which is by definition @xmath179 .
note that the dual representation @xmath178 differs from @xmath94 only by exchanging positive and negatives modes .
the space @xmath176 can be deformed , i.e. can be included in a family of subspaces of @xmath34 depending on a set of parameters that are points on @xmath180 . given @xmath3 vectors
@xmath181 satisfying and given a vector @xmath182 we define @xmath183 as the direct sum of the spaces @xmath184 and @xmath185 , where @xmath186(dz_1)^{2-p } \prod_{i=1}^k ( z_1 - a_i)^{n_+^i}\subset { \mathbb{c}}[z_1,z_1^{-1}](dz_1)^{2-p}\,,\\ w_-({{\mathbf{n}}}^1,\dots , { { \mathbf{n}}}^k;\va ) = & \,\ , { \mathbb{c}}[z_2](dz_2)^0 \prod_{i=1}^k ( z_2 - a_i)^{n_-^i}\subset { \mathbb{c}}[z_2,z_2^{-1}](dz_2)^0\end{aligned}\ ] ] and the space @xmath185 is the sum ( not direct ! ) of the images of the subspaces @xmath187(z_1-z_2)^{p - j-2}(dz_1)^{2-p } ( dz_2)^0 \prod_{i=1}^k ( z_1 - a_i)^{n_j^i - n_-^i}(z_2 - a_i)^{n_-^i}\ ] ] for @xmath161 under the mapping @xmath162 .
then @xmath188 for @xmath189 , so the family of subalgebras @xmath183 is a deformation of @xmath176 .
such a deformation can be applied to the calculation of dimensions for the coinvariants . to simplify the notation
denote by @xmath190 $ ] the space of coinvariants @xmath191 .
[ thm : verlinde ] suppose that points @xmath192 are pairwise distinct .
then the dimension of @xmath193 $ ] coincides with the coefficient before @xmath94 in the following expression @xmath194 \cdot { { \mathsf{i}_{0 } } } + \dots + \dim { { \mathsf{i}_{p-1}}}[w({{\mathbf{n}}}^i ) ] \cdot { { \mathsf{i}_{p-1 } } } \right),\ ] ] where product means the verlinde multiplication .
first recall the fusion procedure in the case of one input and @xmath3 outputs .
suppose that @xmath195 and @xmath196 are two sets of vectors such that @xmath197 for any @xmath198 , @xmath199 .
then we have the natural projection @xmath200 \to { { \mathsf{i}_{r}}}[w({{\mathbf{m}}}^1,\dots , { { \mathbf{m}}}^k;\va)]$ ] .
therefore the dual spaces form an inductive system and we can consider the limit @xmath201 = \lim_\to \left({{\mathsf{i}_{r}}}[w({{\mathbf{n}}}^1,\dots , { { \mathbf{n}}}^k;\va)]\right)^*.\ ] ] the action of the algebra @xmath116 on @xmath94 can be extended to the action of the algebra @xmath202 on the limit @xmath203 $ ] , where summands correspond to points @xmath204 .
the statement that the set of representations @xmath94 form a minimal model means that the representation @xmath203 $ ] decomposes into the direct sum of external tensor products @xmath201 = \bigoplus_{{\genfrac{}{}{0pt}{}{0\le s_1 , \dots , s_k < p}{{{\mathsf{i}_{s_1 } } } \cdot { { \mathsf{i}_{s_2}}}\cdot \dots \cdot { { \mathsf{i}_{s_k } } } = { { \mathsf{i}_{r } } } } } } { { \mathsf{i}_{s_1 } } } \boxtimes \dots \boxtimes { { \mathsf{i}_{s_k}}}\ ] ] emphasize that the sum goes over the set of representations whose product in the verlinde algebra is @xmath94 .
clearly , the space @xmath193 $ ] is dual to the space of @xmath205-invariants in @xmath203 $ ] . as the subspace @xmath205 is dense with respect to the topology of direct limit in the subspace @xmath206 , we have @xmath207 = \sum_{{\genfrac{}{}{0pt}{}{0\le s_1 , \dots , s_k < p}{{{\mathsf{i}_{s_1}}}\cdot { { \mathsf{i}_{s_2 } } } \cdot \dots \cdot { { \mathsf{i}_{s_k } } } = { { \mathsf{i}_{r } } } } } } \dim { { \mathsf{i}_{s_1}}}[w({{\mathbf{n}}}^1 ) ] \cdot \dots \cdot \dim { { \mathsf{i}_{s_k}}}[w({{\mathbf{n}}}^k)],\ ] ] that is equivalent to the statement of the theorem . here
we have a family of subalgebras parametrized by complex numbers @xmath208 . the deformation argument ( see for example @xcite ) implies that dimension of the space of coinvariants with respect to a special subalgebra is not less than one with respect to a generic subalgebra .
[ prop : defarg ] the dimension of @xmath193 $ ] is minimal whenever @xmath209 are pairwise distinct and maximal whenever @xmath209 are equal each to other . introduce elementary spaces @xmath210 with @xmath211 and @xmath212 , @xmath213 .
[ ? ] we have @xmath214 = \sharp\{n | 0 \le pn + j + r \le i\}\ ] ] in other words , @xmath215 = \begin{cases } 2 & i = p,\ ( r+j ) \
{ \rm mod}\ p = 0\\ 1 & \mbox{otherwise}\\ 0 & ( r+j)\ { \rm mod } \ p
\ > \ i \end{cases}\ ] ] note that @xmath216 contains the subalgebra @xmath217 therefore the space of @xmath216-invariants in @xmath94 is spanned by several vectors @xmath218 , @xmath219 . taking the invariants of @xmath220 and @xmath221 selects the vectors @xmath218 with @xmath222 .
we need to show that these vectors are invariant .
note that if @xmath222 then @xmath223 . and we know that the vector @xmath224 is annihilated by the subspace @xmath225 .
according to , and we define the positive integers @xmath226 $ ] by the following formula @xmath227\cdot \xi^r,\ ] ] where @xmath228 is a primitive root of unity of degree @xmath16 .
then by theorem [ thm : verlinde ] and proposition [ prop : defarg ] we have @xmath229\geq \dim{{\mathsf{i}_{r}}}[w({{\mathbf{n}}}^{i^1,j^1},\dots,{{\mathbf{n}}}^{i^k , j^k};\va)]\geq d_{p , r}[i^1,j^1;\dots;i^k , j^k]\end{gathered}\ ] ] for any @xmath230 .
one aim of this paper is to show that ( [ eq : part - of - inequ ] ) is indeed an equality for arbitrary vector @xmath230 ( see theorem [ thm : verlinde-1 ] ) .
it is easy to see from that @xmath226 $ ] is a sum with the step @xmath16 of supernomial coefficients . to write this sum introduce the @xmath231 matrix @xmath232 so @xmath233 for @xmath234 . with the vector @xmath235 we associate the other vector
@xmath236 define the vector @xmath237 as follows @xmath238 let @xmath239 and @xmath240 .
then @xmath241= \sum_{a\in { \mathbb{z } } } { \left({\genfrac{}{}{0pt}{}{{\mathbf{l}}}{pa+n_--r}}\right)_{}}^{p+1}\,,\ ] ] in the case where at least one of @xmath242 sum can be easily calculated .
we have @xmath243=2^{l_1}\,3^{l_2}\ , \dots\,(p-1)^{l_{p-2 } } p^{l_{p-1}-1 } ( p+1)^{l_p}\ ] ] whenever @xmath244 or equivalently at least one @xmath245 is equal to @xmath73 . at last
let us describe the class of subalgebras @xmath176 which can be deformed in this way .
suppose that all the components of the vector @xmath246 are non - negative .
then we have @xmath247 for some @xmath248 .
[ [ section ] ] let @xmath249 be a set with @xmath250 elements . let @xmath251 with @xmath252 be the standard basis in @xmath253 .
we write elements @xmath254 from @xmath253 as rows .
let @xmath255 denote the standard scalar product in @xmath253 .
let @xmath256 be the lattice with the basis @xmath251 and the scalar product determined by a symmetric matrix @xmath257 @xmath258 consider the algebra @xmath259 generated by vertex operators @xmath260 , @xmath252 .
emphasize that we do not involve vertex operators corresponding to @xmath261 .
let us fix the decomposition of @xmath260 into formal power series @xmath262 the algebra @xmath259 is naturally @xmath256-graded .
decomposition determines another one grading by operator @xmath110 @xmath263}=-n v({{\mathbf{e}}}_a)_n\,.\ ] ] this algebra acts in the space @xmath264 where @xmath265 is the fock module with the momentum @xmath266 and @xmath267 is the lattice dual to @xmath256 .
the representation @xmath268 contains the subrepresentation @xmath269 generated from the highest weight vector @xmath270 in @xmath265 . the representation @xmath269 plays a role of verma module , namely , it is maximal among representation with the same annihilation conditions satisfied by the vector @xmath270 .
clearly @xmath269 admits the @xmath267-grading ; to fix it let us set @xmath271 .
the dual space @xmath272 can be identified with the space of correlation functions .
namely , any homogeneous linear functional @xmath273 of degree @xmath274 determines and is determined by the function @xmath275 [ prop : corr - funcs ] correlation functions @xmath276 are products of the polynomial @xmath277 and a polynomial symmetric in each group of variables @xmath278 .
the numbers @xmath279 are determined by the annihilation conditions satisfied by the vector @xmath266 . in other words
, we have the decomposition @xmath280 and @xmath281 can be identified with the space of correlation functions of degree @xmath254 . using proposition [ prop : corr - funcs ] we obtain the following formula of gordon type .
the character of the space @xmath269 is given by the formula @xmath282 where @xmath283 and @xmath284 .
let @xmath285 be an @xmath250-dimensional vector with integer components .
define the subspaces @xmath286 as follows : @xmath287 then @xmath288-coinvariants of @xmath272 can be easily calculated in terms of correlation functions .
namely the dual space consists of functions given in proposition [ prop : corr - funcs ] with a restriction on the degrees in each variable .
this gives the following [ prop : char ] the character of @xmath289={{\mathsf{j}_{{{\mathbf{y}}}}}}^*/ w^a({{\mathbf{n}}}){{\mathsf{j}_{{{\mathbf{y}}}}}}^*$ ] is given by the formula @xmath290(q,{{\mathbf{z}}})=
\sum_{{{\mathbf{n}}}\in{\mathbb{z}}^\ell } { { \mathbf{z}}}^{{{\mathbf{n}}}}\ , q^{{\frac{1}{2}}{{\mathbf{n}}}a\cdot{{\mathbf{n } } } + { { \mathbf{v}}}\cdot{{\mathbf{n } } } } \prod\limits_{a\in i } { \left[{\genfrac{}{}{0pt}{}{{{\mathbf{e}}}_a\cdot({{\mathbf{n}}}+{{\mathbf{n}}}-{{\mathbf{n}}}a+{{\mathbf{w}}})}{{{\mathbf{e}}}_a\cdot{{\mathbf{n}}}}}\right]_{q}}\ ] ] where @xmath291 .
[ [ section-1 ] ] consider the family of algebras @xmath292 obtained from the algebra @xmath116 by the following recursive procedure .
introduce a filtration on @xmath116 by assigning degree 1 to the generators @xmath71 and @xmath72 .
the associated graded algebra @xmath293 is generated by @xmath294 , @xmath295 .
define the currents @xmath296 by the following ope : @xmath297 note that the original algebra @xmath116 is the central extension of @xmath298 . at the next step
we assign degrees @xmath74 to @xmath294 , @xmath295 and @xmath299 . after taking graded object
we obtain the algebra @xmath300 whose generators we denote abusing notations again by @xmath294 , @xmath295 and @xmath299 .
the algebra @xmath300 is generated by currents @xmath294 , @xmath295 and @xmath299 that have opes with the following properties @xmath301 to make the step @xmath302 we introduce a filtration by attaching degree @xmath74 to @xmath294 , @xmath295 and @xmath299 , , @xmath303 .
as the result we obtain the algebra @xmath304 with generators @xmath294 , @xmath295 and @xmath305 , @xmath306 that have the following opes @xmath307 determined by the following @xmath308 matrix @xmath309 where the first row and column corresponds to the index @xmath310 the second ones to @xmath311 the third to @xmath312 and so on till @xmath313 .
there is the surjective homomorphism @xmath314 acting on generators as @xmath315 from we can deduce that all relations between @xmath260 hold for their images in @xmath292 .
so defines a homomorphism .
as the image contains all the generators of @xmath292 , this map is surjective .
note that the map is graded .
namely , it preserves the grading with respect to @xmath110 and specialize the @xmath49-grading of @xmath316 into @xmath67-grading of @xmath292 .
this specialization is defined by the functional @xmath317 on @xmath49 such that @xmath318 , @xmath319 , @xmath320 ( see ) .
any representation @xmath94 of @xmath116 contains the highest weight vector .
therefore the filtration in @xmath116 determines a filtration in @xmath94 and associated graded algebra @xmath298 acts in the associated graded module @xmath321 .
we can repeat this procedure with @xmath321 and the image of the highest weight vector . in this way we obtain the family of representations @xmath322 of algebras @xmath323 .
evidently characters of all @xmath322 , @xmath324 are equal to each other and equal to character of @xmath94 given by .
note that allows us to treat any @xmath292-module as a @xmath316-module . according to this
we have let @xmath325 .
then there is the surjective mapping of @xmath316-modules @xmath326 mapping the vector @xmath270 into the highest weight vector .
it follows from the observation that the highest weight vector of @xmath327 satisfy the same annihilation conditions as @xmath270
. we will show ( see corollary [ cor : iso ] ) that is indeed an isomorphism . note that this map is bi - graded with respect to @xmath110 and @xmath328 .
define the subspaces @xmath329 recursively from @xmath176 as the adjoint graded subspaces .
[ prop : map - sub]let @xmath330 .
then @xmath331 contains the image of @xmath332 under the map .
note that @xmath176 contains linear combinations of the form @xmath333 for @xmath334 .
therefore @xmath335 contains @xmath336 for @xmath334 hence ( according to and ) it contains the image of @xmath332 .
the most important case of this statement is @xmath337 , where the condition on @xmath285 is empty . under the conditions of proposition
[ prop : map - sub ] we have the surjective map @xmath338\to{{\bar{\mathsf{i}}_{r}}}^d[w({{\mathbf{n}}})]\ ] ] this map is an isomorphism only under some additional conditions ( see corollary [ cor : iso ] ) .
the deformation argument ( see for example
@xcite ) implies that for any filtered representation @xmath339 of filtered ( sub)algebra @xmath34 we have @xmath340 ) \ge \dim v[w].\ ] ] in our case it implies @xmath341\geq\dim{{\bar{\mathsf{i}}_{r}}}^{d-1}[w^{d-1}({{\mathbf{n } } } ) ] \geq\dots\geq\dim{{\bar{\mathsf{i}}_{r}}}^0[w^0({{\mathbf{n}}})]\geq\dim{{\mathsf{i}_{r}}}[w({{\mathbf{n}}})]\ ] ] taking , and proposition [ eq : part - of - inequ ] together we obtain the _ main inequality
_ for @xmath330 we have @xmath342\geq \dim{{\mathsf{i}_{r}}}[w({{\mathbf{n}}}^{i^1,j^1},{{\mathbf{n}}}^{i^2,j^2},\dots,{{\mathbf{n}}}^{i^k , j^k};0)]\geq\\ \dim{{\mathsf{i}_{r}}}[w({{\mathbf{n}}}^{i^1,j^1},{{\mathbf{n}}}^{i^2,j^2},\dots,{{\mathbf{n}}}^{i^k , j^k};{{\mathbf{a}}})]\geq d_{p , r}[i^1,j^1;i^2,j^2;\dots;i^k , j^k ] \end{gathered}\ ] ]
[ [ section-2 ] ] following @xcite introduce @xmath0-supernomial coefficients as follows @xmath343_{q}}{\left[{\genfrac{}{}{0pt}{}{l_{k-1}+n_{k}}{n_{k-1}}}\right]_{q } } \dots { \left[{\genfrac{}{}{0pt}{}{l_{1}+n_{2}}{n_{1}}}\right]_{q}},\ ] ] where @xmath344 is a @xmath3-dimensional vector with nonnegative integer entries .
the sum in the rhs is a well - defined polynomial in @xmath0 because @xmath0-binomial coefficients vanish whenever the bottom index becomes greater than the top one .
let @xmath345 be a vector in @xmath346 ; let @xmath347 and let @xmath348 .
the aim of this section is to prove the following polynomial identity .
[ ta ] let @xmath349 .
suppose that @xmath350 and additionally @xmath351 then @xmath352_{q } } = \sum_{a\in { \mathbb{z } } } z^a q^{pa^2/2 } { \left({\genfrac{}{}{0pt}{}{{\mathbf{l}}}{pa+n_-}}\right)_{q}}^{d+2}.\ ] ] where @xmath353 we call ( [ wellb ] ) well - balanced conditions . without them identity ( [ cmain ] ) can fails .
to eliminate the well - balanced conditions we use an improved version of @xmath0-binomial coefficients : @xmath354_{q}}^+ = \left\ { \begin{array}{ll } { \left[{\genfrac{}{}{0pt}{}{n}{m}}\right]_{q } } & n\ge 0\\ ( -1)^{n - m } q^{-\frac{(n - m)^2 + ( n - m)}2 } { \left[{\genfrac{}{}{0pt}{}{-m-1}{-n-1}}\right]_{q^{-1 } } } & n<0 \end{array}\right.\ ] ] in particular , for @xmath355 the integer @xmath356_{1}}^+$ ] is the coefficient before @xmath357 in the laurent expansion of @xmath358 at @xmath359 .
this definition is motivated by the following relations .
[ prop : exqbin ] the polynomials @xmath356_{q}}^+$ ] are uniquely defined by the @xmath0-pascal identities @xmath360_{q}}^+ & = \ q^m { \left[{\genfrac{}{}{0pt}{}{n-1}{m}}\right]_{q}}^+ + { \left[{\genfrac{}{}{0pt}{}{n-1}{m-1}}\right]_{q}}^+ & \mbox{for any \ $ n$ ,
$ m$}\\ { \left[{\genfrac{}{}{0pt}{}{n}{m}}\right]_{q}}^+ & = { \left[{\genfrac{}{}{0pt}{}{n-1}{m}}\right]_{q}}^+ + q^{n - m } { \left[{\genfrac{}{}{0pt}{}{n-1}{m-1}}\right]_{q}}^+ & \mbox{for any \ $ n$ ,
$ m$ } \label{qpascal}\end{aligned}\ ] ] and conditions @xmath361_{q}}^+ = 1 , \qquad { \left[{\genfrac{}{}{0pt}{}{n}{0}}\right]_{q}}^+ = 0\ \ \mbox{for $ n<0$}\ ] ] which need to be checked only for certain @xmath362 and @xmath42 .
indeed any set of polynomials @xmath356_{q}}^?$ ] satisfying and can be expressed in the following way : @xmath363_{q}}^ ?
= \left\ { \begin{array}{ll } p_1 \cdot ( -1)^{n - m } q^{-\frac{(n - m)^2 + ( n - m)}2 } { \left[{\genfrac{}{}{0pt}{}{-m-1}{n - m}}\right]_{q^{-1 } } } & n<0,\ m<0\\ p_2 \cdot ( -1)^m q^{-\frac{m^2 + m}2 } { \left[{\genfrac{}{}{0pt}{}{-(n - m)-1}{m}}\right]_{q^{-1 } } } & n<0,\ m\ge 0\\ ( p_1 + p_2 ) \cdot { \left[{\genfrac{}{}{0pt}{}{n}{m}}\right]_{q } } & n\ge 0 \end{array}\right.\quad \begin{picture}(80,30)(-10,40 ) \epsfig{file = qbin.eps , height=80pt } \end{picture}\ ] ] where @xmath364 and @xmath365 are certain polynomials in @xmath0 . in our case we need only to check that @xmath366 and @xmath367 . note that for usual @xmath0-binomial coefficients and fails for @xmath368 . [ tb ] let @xmath349 .
suppose that @xmath350 .
then @xmath369_{q}}^+ = \sum_{a\in { \mathbb{z } } } z^a q^{pa^2/2 } { \left({\genfrac{}{}{0pt}{}{{\mathbf{l}}}{pa+n_-}}\right)_{q}}^{d+2}.\ ] ] emphasize that in the left hand side the sum goes over all integer vectors .
[ [ section-3 ] ] here we show that the left hand side of ( [ cimp ] ) is a well defined polynomial and that theorem [ tb ] implies theorem [ ta ] .
let @xmath370_{q}}^+$ ] .
we need to prove that if @xmath350 for any @xmath371 then there is only finite number of vectors @xmath372 such that @xmath373 . and
if additionally we impose the well - balance conditions ( [ wellb ] ) then @xmath373 implies @xmath374 , @xmath375 , where @xmath376 . if @xmath373 then @xmath377 condition follows from the fact that @xmath356_{q}}^+ = 0 $ ] if @xmath378 .
condition follows from the fact that @xmath356_{q}}^+ = 0 $ ] if @xmath379 and @xmath380 .
let @xmath381 . if @xmath350 and @xmath382 then @xmath383 .
suppose that @xmath382 and @xmath384 . then adding for @xmath385 twice and subtracting for @xmath386 ( if @xmath387 ) and @xmath388 ( if @xmath389 ; if @xmath390 then for @xmath391 and @xmath392 ) we obtain @xmath393 . if @xmath394 and @xmath395 , @xmath396 then @xmath383 .
proof is similar . here
we add for @xmath391 and for @xmath392 then subtract for @xmath397 .
suppose that @xmath398 and @xmath399 .
then @xmath383 for @xmath400 suppose that @xmath398 and @xmath373 . then adding for @xmath391 twice and subtracting for @xmath397 we obtain @xmath401 taking into account @xmath399 and @xmath402 we deduce @xmath403 .
clearly , we have the similar statement for @xmath404 and @xmath405 .
so if @xmath350 for any @xmath371 and the well - balanced conditions are satisfied then @xmath406 only if @xmath374 for any @xmath407 , hence theorem [ tb ] implies theorem [ ta ] . under the conditions of theorem [ tb ] we have @xmath408 for @xmath409 and @xmath410 , @xmath411 otherwise @xmath383 .
adding ( [ rall ] ) for @xmath412 and @xmath392 , we obtain @xmath413 as all the coefficient in the right hand side of this inequality are positive , there exists only finite number of such vectors @xmath372 .
[ [ section-4 ] ] here we compare the left hand side @xmath414(q , z)$ ] and the right hand side @xmath415(q , z)$ ] of ( [ cimp ] ) . note that for @xmath0-supernomial coefficients there is an identity @xmath416 therefore , for the right hand side we have @xmath417=\chi^r_{p , d}[{\mathbf{l}}+ \ve_{d},n_- -1 ] + z^{-1 } \,q^{n_- -p/2}\ , \chi^r_{p , d}[{\mathbf{l } } , n_- - p].\ ] ] also as @xmath418 , we have @xmath419 = \chi^r_{p , d-1}[(l_1 , \dots , l_d ) , n_-].\ ] ] the idea of the proof is to check ( [ rec1 ] ) and ( [ rec2 ] ) for the left hand side .
let @xmath257 be a symmetric @xmath231 matrix with integer components , let @xmath420 , @xmath421 be vectors in @xmath422 . consider the polynomial @xmath423(q , z)= \sum_{{{\mathbf{n}}}\in{\mathbb{z}}^m } z^{{{\mathbf{u}}}\cdot{{\mathbf{n}}}}\ , q^{{\frac{1}{2}}{{\mathbf{n}}}a\cdot{{\mathbf{n } } } } \prod\limits_{a } { \left[{\genfrac{}{}{0pt}{}{{{\mathbf{e}}}_a\cdot({{\mathbf{n}}}+{{\mathbf{n}}}-{{\mathbf{n}}}a)}{{{\mathbf{e}}}_a\cdot{{\mathbf{n}}}}}\right]_{q}}^+.\ ] ] then for any @xmath424 we have @xmath425= \chi_{a,{{\mathbf{u}}}}[{{\mathbf{n}}}-{{\mathbf{e}}}_a ] + z^{u_a}\ , q^{n_a - a_{aa}/2 } \ , \chi_{a,{{\mathbf{u}}},{{\mathbf{v}}},{{\mathbf{w}}}}[{{\mathbf{n}}}-{{\mathbf{e}}}_aa].\ ] ] identity follows from the identity applied to the factor @xmath426_{q}}^+$ ] in each summand . applying ( [ eq : main - recurrence ] ) with @xmath392 to @xmath414 $ ] , we check ( [ rec1 ] ) for the left hand side .
note that for @xmath427 ( [ eq : main - recurrence ] ) implies recurrence relations for the @xmath0-supernomial coefficients same as in @xcite .
to check ( [ rec1 ] ) for the left hand side we need the following identity .
there is the following relation @xmath428_{q}}^+ { \left[{\genfrac{}{}{0pt}{}{m}{m}}\right]_{q}}^+= \sum_{l\in{\mathbb{z}}}q^{(n - l)(m - l ) } { \left[{\genfrac{}{}{0pt}{}{n - m}{n - l}}\right]_{q}}^+ { \left[{\genfrac{}{}{0pt}{}{m - n}{m - l}}\right]_{q}}^+ { \left[{\genfrac{}{}{0pt}{}{n + m - n - m + l}{l}}\right]_{q}}^+\ ] ] first let us check that the sum in the right hand side is indeed finite
. note that if @xmath429 then the third factor is always zero , so entire sum is zero . if @xmath430 then either @xmath431 or @xmath432 . in the first case we have non - zero
summands only if @xmath433 , in the second case only if @xmath434 .
let us denote the polynomial in the right hand side by @xmath435 .
the idea of proof is to apply proposition [ prop : exqbin ] to @xmath435 as a set of polynomials indexed by @xmath436 and @xmath42 .
it means that we prove that @xmath437 for any @xmath436,@xmath42,@xmath438,@xmath362 and @xmath439_{q}}^+ \qquad \rh^{n , m}_{0,m } = 0\ ] ] for any @xmath438 , @xmath362 and certain @xmath440 , @xmath42 . to prove fix for a moment @xmath436 , @xmath438 , @xmath42 , @xmath362 and introduce the notation @xmath441 } = q^{(n - l)(m - l ) } { \left[{\genfrac{}{}{0pt}{}{n - m + a}{n - l + d}}\right]_{q}}^+ { \left[{\genfrac{}{}{0pt}{}{m - n + b}{m - l + e}}\right]_{q}}^+ { \left[{\genfrac{}{}{0pt}{}{n + m - n - m + l + c}{l + f}}\right]_{q}}^+.\ ] ] then substitution @xmath442 into @xmath443 leads to the identity @xmath444 } = \sum_{l\in{\mathbb{z } } } q^{2l - m - n+1 } { \left [ \begin{array}{rrr } a & b & c \\ d & e & f \end{array}\right]}.\ ] ] using this identity together with we obtain @xmath445 } = \sum
q^{n - l } { \left [ \begin{array}{rrr } -1 & 0 & 0 \\ 0 & 0 & 0 \end{array}\right ] } + \sum { \left [ \begin{array}{rrr } -1 & 0 & 0 \\ -1 & 0 & 0 \end{array}\right ] } = \\ = \sum q^{n } { \left [ \begin{array}{rrr } -1 & 0 & -1 \\ 0 & 0 & 0 \end{array}\right ] } + \sum
q^{n - l } { \left [ \begin{array}{rrr } -1 & 0 & -1 \\ 0 & 0 & -1 \end{array}\right ] } + \sum { \left [ \begin{array}{rrr } -1 & 0 & 0 \\ -1 & 0 & 0 \end{array}\right ] } = \\
= \sum q^{n } { \left [ \begin{array}{rrr } -1 & 0 & -1 \\ 0 & 0 & 0 \end{array}\right ] } + \sum q^{l - m } { \left [ \begin{array}{rrr } -1 & 0 & 0 \\ -1 & -1 & 0 \end{array}\right ] } + \sum { \left [ \begin{array}{rrr } -1 & 0 & 0 \\ -1 & 0 & 0 \end{array}\right ] } = \\ = \sum
q^{n } { \left [ \begin{array}{rrr } -1 & 0 & -1 \\ 0 & 0 & 0 \end{array}\right ] } + \sum q^{l - m } { \left [ \begin{array}{rrr } -1 & 1 & 0 \\ -1 & 0 & 0 \end{array}\right ] } = q^n \rh^{n-1,m}_{n , m } + \rh^{n-1,m}_{n-1,m},\end{gathered}\ ] ] where sums go over @xmath446 .
proof of the second part of is similar . to check we take @xmath447 and any negative @xmath448 .
we have only one non - zero summand in @xmath449 , namely for @xmath450 , and it coincides with the @xmath0-binomial given in . if @xmath448 and @xmath451 then @xmath452 and , as described above , the result is zero .
so proposition implies the lemma . applying ( [ rdc ] ) to the factors in @xmath453
$ ] corresponding to @xmath454 and @xmath455 we obtain @xmath456 = \chi^l_{p , d}[(n_+,n_-;n_0 \dots , n_{d-2},n_++n_-)],\ ] ] that is , ( [ rec2 ] ) for the left hand side .
now we can prove theorem [ tb ] by induction on @xmath457 and @xmath458 .
to complete the proof we need only to check the case @xmath459 manually .
suppose that @xmath460 .
then @xmath461_{q}}^+{\left[{\genfrac{}{}{0pt}{}{s}{k}}\right]_{q}}^+= { \left[{\genfrac{}{}{0pt}{}{m+s}{s+a}}\right]_{q}}\ ] ] first let us check that the sum in the left hand side is indeed finite . as @xmath460 we have either @xmath462 or @xmath463 .
in the first case we have non - zero summands only if @xmath464 , in the second case only if @xmath465 .
let us denote the polynomial in the left hand side by @xmath466 .
clearly , we have @xmath467 and the same identity for the right hand side . therefore we can restrict ourselves to the case @xmath468 and prove that @xmath469_{q}}^+ \quad \mbox{for $ s \ge 0 $ and any $ m$.}\ ] ] applying to the second factor in @xmath466 , we obtain @xmath470 that is for the right hand side of .
therefore it remains to check for @xmath471 .
if @xmath471 then the only non - zero summand in @xmath466 is for @xmath472 and it is equal to the right hand side of the case @xmath459 of theorem [ tb ] can be obtained by substituting @xmath473 , @xmath474 , @xmath475 and @xmath476 in .
[ thm : verlinde-1][flatness ] for any @xmath230 we have @xmath477=d_{p , r}[i^1,j^1;i^2,j^2;\dots;i^k , j^k]\,.\ ] ] the main idea of the proof is to check that the main inequality ( [ eq : the - main - inequality ] ) is indeed an equality .
here we have a chain of inequalities , where we know the first element by proposition [ prop : char ] and the last element .
as soon as we check that these two numbers are equal we obtain that other numbers in ( [ eq : the - main - inequality ] ) are also equal .
suppose that the vector @xmath421 satisfies the well - balance conditions ( [ wellb ] ) .
then this equality follows from theorem [ ta ] by setting @xmath478 and @xmath479 .
but if conditions ( [ wellb ] ) are not satisfied then this equality can be violated .
so we should deduce general case from the well - balanced one . to this end
consider the family of automorphisms @xmath480 .
they preserve the representation @xmath102 and permute the subalgebras @xmath481 .
namely , @xmath480 preserves @xmath482 , increases @xmath483 by @xmath484 and decreases @xmath485 by @xmath484 .
first consider the case @xmath487 .
then we should set @xmath337 .
let @xmath488 be the minimal number such that @xmath489 .
so we have @xmath490 and as @xmath491 we obtain @xmath492 .
hence after applying the automorphism @xmath480 the well - balance conditions will be satisfied . in the case
@xmath493 we set @xmath494 .
let as above @xmath488 be the minimal number such that @xmath495 .
so we have @xmath496 and as @xmath497 we obtain @xmath498 .
hence after applying the automorphism @xmath480 the well - balance conditions will be satisfied .
[ cor : iso ] let @xmath330 .
let @xmath499 be the matrix given by .
then we have @xmath500 if @xmath501 and @xmath285 satisfies the well - balance condition @xmath502 then we have @xmath503 \cong { { \bar{\mathsf{i}}_{r}}}^d[w({{\mathbf{n}}})]\ ] ] 1 .
let the set of indices be @xmath510 .
consider the vectors @xmath507 , @xmath508 , @xmath511 .
then @xmath512 ( q , z ) = z^{-\frac{r}{p}}\,q^{\frac{r}{2p}(r - p+2)}\ , \sum_{{{\mathbf{n}}}\in{\mathbb{z}}^\ell } z^{{{\mathbf{u}}}\cdot{{\mathbf{n}}}}\ , q^{{\frac{1}{2}}{{\mathbf{n}}}{{\mathcal{a}}}\cdot{{\mathbf{n } } } + { { \mathbf{v}}}\cdot{{\mathbf{n } } } } \prod\limits_{a\in i } { \left[{\genfrac{}{}{0pt}{}{{{\mathbf{e}}}_a\cdot({{\mathbf{n}}}+{{\mathbf{n}}}-{{\mathbf{n}}}{{\mathcal{a}}}+{{\mathbf{w}}})}{{{\mathbf{e}}}_a\cdot{{\mathbf{n}}}}}\right]_{q}}^+,\ ] ] where the notion @xmath356_{q}}^+$ ] is given by .
we have @xmath512 ( q , z)= z^{-\frac{r}{p}}\,q^{\frac{r}{2p}(r - p+2)}\ ,
\sum_{m\in { \mathbb{z } } } z^m q^{\frac{p}{2}(m^2+m ) -(r+1)m } { \left({\genfrac{}{}{0pt}{}{{\mathbf{l}}}{pm - r+n_-}}\right)_{q}}^{p+1}.\ ] ] we know that general case can be reduced to the well - balanced case by the action of @xmath480 . as @xmath480 acts on the rhs of ( ii ) by the same substitution @xmath516 into @xmath514
, we have ( ii ) in general case . the statement ( i )
can be deduced from ( ii ) by theorem [ tb ] .
let @xmath517 .
suppose that @xmath518 .
let the set of indices be @xmath519 and , as above , @xmath507 , @xmath508 , @xmath511 .
then @xmath512 ( q , z ) = z^{-\frac{r}{p}}\,q^{\frac{r}{2p}(r - p+2)}\ , \sum_{{{\mathbf{n}}}\in{\mathbb{z}}^\ell } z^{{{\mathbf{u}}}\cdot{{\mathbf{n}}}}\ , q^{{\frac{1}{2}}{{\mathbf{n}}}{{\mathcal{a}}}\cdot{{\mathbf{n } } } + { { \mathbf{v}}}\cdot{{\mathbf{n } } } } \prod\limits_{a\in i } { \left[{\genfrac{}{}{0pt}{}{{{\mathbf{e}}}_a\cdot({{\mathbf{n}}}+{{\mathbf{n}}}-{{\mathbf{n}}}{{\mathcal{a}}}+{{\mathbf{w}}})}{{{\mathbf{e}}}_a\cdot{{\mathbf{n}}}}}\right]_{q}}^+,\ ] ] where the notion @xmath356_{q}}^+$ ] is given by .
feigin , d.b .
fuchs , _ cohomology of some nilpotent subalgebras of the virasoro and kac - moody algebras _ ,
geometry and physics , essays in honor of i.m.gelfand on the occasion of his 75th birthday , eds .
s.gindikin and i.m.singer , p. 209 - 235 .
north - holland , 1991 .
b. feigin and s. loktev , _ on generalized kostka polynomials and the quantum verlinde rule _
, differential topology , infinite - dimensional lie algebras , and applications , amer .
ser . 2 , 194 ( 1999 ) , p. 6179
, math.qa/9812093 . | we propose an alternative definition of @xmath0-supernomial coefficients as characters of coinvariants for one dimensional lattice vertex operator algebras .
this gives a new formula for @xmath0-supernomial coefficients . along
the way we prove that the spaces of coinvariants form a bundle over the configuration space of complex points ( the configuration space includes diagonals ) .
= 1.6pt |
single - molecule magnets ( smm s ) consist of identical molecules , each of which made up of several magnetic ions surrounded by many different species of atoms . a single molecule of the smm s mn@xmath1-acetate@xcite and fe@xmath0@xcite has an effective ground - state spin of @xmath4 and a strong crystal - field anisotropy .
a zero - field energy barrier against magnetization reversal is approximately 65 k ( 30 k ) for uniaxial mn@xmath1 ( biaxial fe@xmath0 ) .
@xcite despite their large effective spin , these single - molecule magnets have shown quantum coherence@xcite and quantum tunneling between the energy levels of the two potential wells.@xcite although dipolar interactions between different molecules are weak , in the low - temperature limit and near zero applied field , a dipolar interaction could stimulate the quantum tunneling and thus explain the non - exponential magnetization relaxation observed at early times.@xcite recently , electron paramagnetic resonance ( epr ) experiments on single crystals of fe@xmath0 and mn@xmath1 have revealed interesting effects in the widths and positions of the epr peaks as functions of energy level , resonance frequency , and temperature when the applied field is along the easy axis .
@xcite for both fe@xmath0 and mn@xmath1 , at fixed frequency , the linewidths increase with decreasing energy levels ( the largest linewidth corresponding to the transition between the ground state and the first excited state ) , and for a particular transition the linewidths increase with decreasing frequency . on the other hand , the details of the temperature dependence of the linewidths for the two materials are quite different . for mn@xmath1 ,
the linewidths increase smoothly with increasing temperature , showing a rather weak temperature dependence . for fe@xmath0 , for transitions at low resonant fields , the linewidths increase sharply with temperature at low temperatures , reach a maximum , and then decrease slowly with temperature at higher temperatures .
the exception is the transition associated with the ground state , for which the linewidth decreases with increasing temperature in the whole temperature range studied ( 2@xmath550 k ) . on the other hand , for the transitions with high resonant fields , the linewidths increase monotonically with increasing temperature . for fe@xmath0 ,
the line positions change non - monotonically with energy level and temperature . in ref . ,
which was our first attempt to understand the energy level and resonance frequency dependence of the linewidths , we showed that for both single crystals of fe@xmath0 and mn@xmath1 the distribution in the uniaxial anisotropy parameter @xmath2 of the single - spin hamiltonian , caused by defects in the samples , contributes substantially to the inhomogeneous linewidths at constant temperature .
this was also recently supported by terahertz spectroscopy for mn@xmath1.@xcite the microscopic origin of the distribution in @xmath2 has not yet been fully understood.@xcite the analysis further showed that for fe@xmath0 the dipolar interactions between molecules contribute to the linewidths as significantly as the distribution in @xmath2 , while for mn@xmath1 their contribution is less significant.@xcite in a recent millimeter - wave study on fe@xmath0,@xcite the dipolar field may have been overestimated because the distribution in @xmath2 was not included . since the approximations made in ref .
reasonably well explained the linewidth behavior at fixed temperature , in the present paper we investigate the temperature dependence of the linewidths and line positions . to explain this additional feature ,
we take into account the inter - molecular spin - spin interactions ( exchange and dipolar ) , as well as distributions in the uniaxial anisotropy parameter @xmath2 and the land @xmath3-factor .
we find that the distributions in @xmath2 and @xmath3 do not contribute to shifts in line positions with temperature , but the exchange and dipolar interactions produce nonzero local fields which are temperature dependent ( at low temperatures ) , so that they can make the line positions change with temperature .
thus , to explain the _ temperature _ dependence of the measured line shifts , both the _ exchange _ and dipolar interactions are needed . without the exchange interaction , we can not explain the observed non - monotonic line shifts with temperature .
this was not included in our earlier study , ref .. although the exchange interaction was not considered in our earlier study , its effect on the linewidths is negligible as long as the magnitude of the exchange constant is much smaller than the linewidths.@xcite this paper is organized as follows .
a brief summary of the experimental procedures is presented in sec .
the models for fe@xmath0 and mn@xmath1 are described separately in sec .
iv describes our calculated linewidths and line shifts vs temperature , and their discussed in comparison with the experimental data .
our conclusions are provided in sec .
all of the epr experiments were performed on single crystals with the magnetic field aligned along or close to the direction of the easy magnetization axis , in a temperature range from 2 k to 50 k. for mn@xmath1 , this direction coincides with that of the longest dimension of the essentially needle - shaped crystals . for fe@xmath0 ,
the direction was determined by measurement of the largest spread of the resonant field , by rotating the crystal around the approximately known orientation of the easy axis .
the single crystals were prepared as described earlier.@xcite epr measurements were made in the 100@xmath5190 ghz range with a resonant microwave cavity system described by hill et al.@xcite , which enables observation of distortion - free epr line shapes .
the linewidths were determined by computer - fitting of the observed experimental spectra to either a gaussian or a lorentzian function for the spectra obtained at various temperatures .
all the spectra were obtained by keeping the frequency fixed and sweeping the field to obtain the resonance peaks , as is usual in epr spectroscopy.@xcite
for the examined single - crystal sample of fe@xmath0,@xcite there are two sources of the epr line _ shifts _ : temperature dependence of @xmath2 and the electronic spin - spin interactions ( dipolar and exchange interactions ) between different molecules .
there are also two sources of the epr line _ broadening _ : the @xmath2-strain effect ( distribution in @xmath2 ) and the spin - spin interactions.@xcite for the examined mn@xmath1 sample,@xcite the sources of the line broadening are the @xmath2-strain , the @xmath3-strain,@xcite and the spin - spin interactions.@xcite in our model , hyperfine interactions are not considered for the following reasons : ( i ) the observed line shifts and changes in the linewidths were much larger for fe@xmath0 than for mn@xmath1 .
( ii ) on the other hand , hyperfine fields are orders of magnitude larger for mn@xmath1 ( in which all nuclei have spins of @xmath6 ) than for fe@xmath0 ( 98% of fe nuclei with @xmath7 ) . (
iii ) any residual effect of the hyperfine fields was included in background linewidths .
possible mis - alignment of the external field and a spread in the in - plane fields are also not considered since mis - alignment of the external field can not provide the temperature dependence of the linewidths and shifts .
we checked that for fe@xmath0 the effect of a distribution in the transverse anisotropy parameter @xmath8 on the linewidths was negligible .
we also found that a spread in the in - plane fields ( possibly caused by nuclear spins ) can give rise to a distribution in the easy axis of each molecule , leading to asymmetries in epr lineshapes.@xcite the model and technique used in this study are similar to those in ref .
, except that we here take into account the non - spherical sample shapes and the actual crystal structures .
therefore , we here summarize them only briefly , focusing on what causes the temperature dependence of the line shifts and linewidths for each source .
the temperature dependence of @xmath2 is discussed in sec .
iv a.1 . in our convention ,
the energy level @xmath9 is the ground state when the field is applied along the positive @xmath10 axis , while in refs .
the ground state is @xmath11 . for clarity
, we discuss fe@xmath0 and mn@xmath1 separately .
we consider an effective single - spin hamiltonian which satisfies approximate @xmath13 symmetry , @xmath14 where the uniaxial anisotropy parameter @xmath15 , the transverse anisotropy parameter @xmath16,@xcite @xmath3 is the land @xmath3-factor which is close to 2 , and @xmath17 is the bohr magneton . here @xmath18 is the @xmath19th component of the spin angular momentum operator , and @xmath20 is the longitudinal static applied magnetic field .
we assume that the longitudinal magnetic field is applied along the easy axis ( the @xmath10 axis ) , and we ignore the small transverse anisotropy terms ( @xmath8 terms ) in calculating the linewidths .
thus , the energy level @xmath21 is a good quantum number of the spin operator @xmath22 .
according to our convention , the ground state is @xmath9 . using the density - matrix equation @xcite with the hamiltonian , eq .
( [ eq : ham ] ) , and an interaction between the spin system and an oscillating transverse field , we calculate the power absorption between the levels @xmath21 and @xmath23 for a fixed value of the uniaxial anisotropy parameter @xmath2 . in the power absorption , the line - shape function includes a natural linewidth , which is a function of temperature .
next we calculate the average power absorption with a gaussian distribution in @xmath2 , where @xmath2 itself is assumed to be temperature independent . as a consequence ,
the line broadening due to the @xmath2-strain effect becomes weakly temperature dependent because of the temperature dependence of the natural linewidths . to calculate the natural linewidths , we use the strength of the coupling between the spin system and a surrounding phonon heat bath obtained in ref .. for example , for @xmath9 , the order of magnitude of the natural linewidths is several to several tens of gauss at temperatures below several tens of kelvin , and the widths increase with decreasing @xmath21 .
the spin - spin interactions are calculated separately with @xmath2 fixed , and then combined with the @xmath2-strain effect to obtain the total linewidths . at low temperatures , all energy levels are not equally populated , and the populations of the levels change with temperature according to the boltzmann factor .
thus , the local field on a particular molecule caused by surrounding molecules changes with temperature .
therefore , the spin - spin interactions are mainly responsible for the temperature dependence of both the line shifts and the linewidths .
@xcite to calculate the line shifts and the line broadening due to the spin - spin interactions , we use a multi - spin hamiltonian which commutes with @xmath24 , where the sum runs over all molecules .
details of the technique can be found in ref .. @xmath25 \ : + { \cal h}^{(1 ) } \ : , \\ { \cal h}^{(1 ) } & \equiv & { \cal h}^{\mathrm dipole } + { \cal h}^{\mathrm exch } \ : , \\
{ \cal h}^{\mathrm dipole}&=&\frac{1}{2 } \sum_{jk}\,^{\prime } a_{jk } ( \vec{s}_j \cdot \vec{s}_{k } - 3 s_{j}^z s_{k}^z ) \ : , \ : \ : a_{jk } \equiv \left ( \frac{\mu_0}{4 \pi } \right ) \frac{(g \mu_b)^2}{2 r_{jk}^3 } ( 3 \zeta_{jk}^2 - 1 ) \ : , \label{eq : dp } \\ { \cal h}^{\mathrm exch}&= & \ : \frac{1}{2 } \sum_{jk}\,^{\prime } j_{jk } \ : \vec{s}_j \cdot \vec{s}_{k } \ : , \label{eq : ex}\end{aligned}\ ] ] where @xmath26 is the single - spin hamiltonian for the @xmath27th molecule , the sum @xmath28 runs over all molecules , and @xmath29 is the interaction between the @xmath27th molecule and the oscillating transverse field . here
@xmath30 is the dipolar interaction between the molecules , and @xmath31 are the direction cosines of the vector between molecules @xmath32 and @xmath33 ( @xmath34 ) relative to the easy axis ( @xmath10 axis ) .
the sum @xmath35 runs over all molecules , so that any two indices are not the same .
@xmath36 is the isotropic exchange interaction between the spins of nearest - neighbor molecules where the exchange coupling constant @xmath37 is @xmath38 if the @xmath27th and @xmath32th spins are nearest neighbors and zero otherwise .
it is reasonable to assume that @xmath39 is much smaller than the dipolar and exchange interactions , which again are much smaller than the sum of the single - spin hamiltonians .
since the field is swept at constant frequency , the energy levels change with the sweeping field .
we neglect slight changes of the energy levels during resonances and use @xmath40 , where @xmath41 is an epr frequency , as the field @xmath42 in the spin hamiltonian to calculate the energy levels for the particular resonance .
to calculate the @xmath43th moment of the resonant field deviation , we formulate the @xmath43th moment for a frequency sweep and then convert it to a field sweep .
this is justifiable because we neglect the slight energy change caused by the change of the field during a resonance .
the probability density function of the epr frequency @xmath41 is given by @xmath44 where @xmath45 is the energy eigenvalue of @xmath46 , @xmath47 is the corresponding eigenvector , @xmath48 denotes the sum over all states @xmath49 such that @xmath50 , and @xmath51 denotes the sum over all states for which @xmath52 . using eq .
( [ eq : fhh ] ) , we calculate the @xmath43th moment , @xmath53 where the temperature dependence of the linewidths is included through the boltzmann factors , and the eigenvalues of @xmath54 can contribute to @xmath55 and/or the boltzmann factors . assuming that the dipolar and exchange interactions , @xmath54 , are much smaller than the thermal energy , we expand the boltzmann factor into @xmath56&= & \exp \left ( -\frac{\sum_j { \cal h}_{0j}}{k_b t } \right ) \left [ 1 - \frac{{\cal h}^{(1)}}{k_b t } + \cdots \right ] \label{eq : bolt}\end{aligned}\ ] ] and consider only the first term on the right - hand side . in ref .
, the boltzmann factor was not included since the temperature of interest was quite high , so that all energy levels were equally populated . in our calculations , we use a mean - field approximation , so that the sums of @xmath57 and @xmath58 [ eqs .
( [ eq : dp ] ) and ( [ eq : ex ] ) ] can be separated from the spin operators . to compare with the measured _ line shifts
_ , we calculate perturbatively the first moment , @xmath59 , where @xmath60 is the resonant field without the spin - spin interactions , and subtract from it the first moment at a reference temperature , 30 k. this reference temperature was chosen because at higher temperatures additional line shifts can be expected from the temperature dependence of @xmath2 .
the calculated line shifts to zero order in @xmath61 contain the exchange coupling constant @xmath38 and the effective dipole field @xmath62 ( @xmath63 is the number of molecules in the sample and the summation runs over all molecules ) as variable parameters .
this zero - order result in @xmath61 depends on @xmath38 and @xmath64 through the terms @xmath65 in eq .
( [ eq : lth_moment ] ) .
since @xmath64 depends on the sample shape due to the field - induced net magnetization , ewald s method@xcite is not sufficient to estimate its value in our case . to compare with the measured _ linewidths
_ , we need to calculate the second central moment , @xmath66 , which is equivalent to @xmath67 , where @xmath68 and @xmath69 are calculated perturbatively to zero order in @xmath61.@xcite the square root of the second central moment is proportional to the broadening due to the spin - spin interactions .
the quantity @xmath68 includes the following six terms , @xmath70 , @xmath71 , @xmath72 , @xmath73 , @xmath74 , and @xmath75 .
the quantity @xmath69 includes three terms , @xmath76 , @xmath77 , and @xmath78 . here
@xmath79 and @xmath80 run over all molecules with the constraint that no two indices in the summations may be the same , and @xmath37 and @xmath58 are nonzero for nearest - neighbor molecules only .
the coefficients of the last three terms ( the summations over three indices @xmath27 , @xmath32 , and @xmath33 ) in @xmath68 are the same as those of the three terms in @xmath69 .
to simplify the second central moment , we use the identities @xmath81 summations over four different indices do not appear in eq .
( [ eq:3rd ] ) because @xmath82 by translational invariance .
we thus have only three undetermined terms , @xmath70 , @xmath72 , and @xmath71 in the calculated second central moment .
the exchange coupling constant @xmath38 could be determined from the line - shift analysis .
the remaining two terms , @xmath72 and @xmath71 , can , in principle , be calculated from the exact geometry of the system . however , in our study , we take the two terms as variable parameters , and compare their optimum values with the calculated values . physical justification for this
is provided in sec .
iv.a.2 .
we fix the epr frequency at @xmath83 ghz and vary the temperature from 4 k to 35 k for the linewidth analysis ( 2 k to 30 k for the line - shift analysis ) .
we do not analyze the experimental data above 35 k because at higher temperatures excited states ( effective spin @xmath84 ) might play a role .
only good - quality ( high signal - to - noise ratio ) experimental data were selected .
for the line - shift analysis , the exchange constant @xmath38 and effective dipole field @xmath64 are varied , while for the linewidth analysis , @xmath85 , @xmath86 , and the standard deviation of @xmath2 are varied within experimentally acceptable ranges in order to fit the experimental data .
note that in ref .
the molecules of fe@xmath0 were assumed to be distributed on a simple cubic lattice in a spherical sample .
for spherical samples the value of @xmath64 is zero for all effective dipole distances .
thus only one parameter , either @xmath87 or the effective dipole distance , sufficed in the linewidth analysis .
the real samples , however , were not spherical .
the fe@xmath0 sample examined@xcite was a thin rhombic platelet with acute angle of about 60 degrees , edges of length 0.7 mm , and thickness 0.17 mm , and the molecules in fe@xmath0 are distributed on a triclinic lattice .
therefore , we here use two fitting parameters ( @xmath64 and @xmath87 ) to consider the dipole - dipole distributions in the real experimental samples described above . for mn@xmath1
we consider an effective single - spin hamiltonian which satisfies tetragonal symmetry , @xmath89 with @xmath2@xmath90@xmath91 , @xmath92@xmath90@xmath93 , and @xmath3@xmath90@xmath94 .
@xcite here we consider the case in which the applied field is along the easy axis ( the @xmath10 axis ) , and we neglect the small transverse fourth - order anisotropy term , @xmath95 .
the technique is the same as for fe@xmath0 , except for the following : ( i ) the @xmath3-strain effect provides a weak temperature dependence to the linewidths , caused by the temperature dependence of the natural linewidths .
( ii ) the resonant field without the spin - spin interactions is modified to @xmath96 ( iii ) to calculate the natural linewidths , we use the strength of the coupling between the spin system and a surrounding phonon heat bath from ref .. ( iv ) dipoles are distributed on a centered tetragonal lattice with sample dimensions @xmath97 mm@xmath98 .
( in ref .
, the dipoles were assumed to be distributed on a body - centered cubic lattice , and the sample was assumed to be spherical . )
( v ) the easy anisotropy axis is along the long side of the needle - shaped sample . ( vi ) for mn@xmath1
the measured line shifts are negligible compared to the measured linewidths , so that we do not have to consider the exchange interaction and the effective dipole field ( @xmath99 and @xmath100 , so @xmath101 ) .
thus , the second _ central _ moment @xmath66 , which is proportional to the measured linewidths , is identical to the second moment @xmath102 . to compare with the experimental data ,
the frequency is fixed at @xmath103 ghz , and the temperature is varied from 10 k to 40 k. our analysis ends at 40 k because at higher temperatures excited states ( effective spin @xmath84 ) might play a role .
@xcite only epr spectra of good quality were selected , and @xmath87 and the standard deviations of @xmath2 and @xmath3 were varied within acceptable ranges in order to fit the experimental data .
we show that the spin - spin interactions ( dipolar and/or exchange interactions ) alone determine the trend of the temperature dependence of the line shifts and the linewidths . from the line - shift analysis
, we can estimate the orders of magnitude of the exchange interaction and the effective dipole field and obtain their signs . using this information
, we can also explain quantitatively the linewidths , including the @xmath2-strain and/or the @xmath3-strain effects which give rise to a strong @xmath21 dependence but weak temperature dependence of the linewidths .
the spin - spin interactions contribute more to the linewidths for fe@xmath0 than for mn@xmath1 , mainly because the @xmath2-strain effect is dominant over the spin - spin interactions for mn@xmath1 , while it is comparable with the spin - spin interactions for fe@xmath0 .
this explains the different temperature behavior of the linewidths for fe@xmath0 and mn@xmath1 .
the set of parameter values which best explains the experimental data has some systematic theoretical uncertainties , which are difficult to calculate exactly .
it is known that the value of the uniaxial anisotropy parameter @xmath2 may vary smoothly with temperature.@xcite to gauge the importance of this effect , we first assume that @xmath2 has a temperature dependence such as fig .
[ fig : tdep_d](a ) .
then the line shift , @xmath104@xmath5@xmath105(@xmath106@xmath9030 k ) , becomes monotonically temperature dependent , as shown in fig .
[ fig : tdep_d](b ) . comparing fig .
[ fig : tdep_d](b ) with the experimental data in fig .
[ fig : exp_shifts](a ) , we see that the monotonic temperature dependence of @xmath2 can not by itself explain the complicated temperature dependence of the measured line shifts .
it could presumably contribute together with other factors described below .
however , we hereafter take @xmath2 as a temperature independent parameter , for the sake of simplicity and because the exact temperature behavior is not yet known .
we also note that the distribution in @xmath2 does not change the line positions .
next , we consider the effect of the spin - spin interactions between molecules on the measured line shifts .
if we first ignore the exchange interaction and consider the dipolar interaction only for a spherical sample with dipoles distributed on a simple cubic lattice ( which is the assumption made in ref . ) , then the effective dipole field @xmath64 vanishes , so that there is no line shift to zero order in @xmath61 .
higher - order corrections [ the second term in eq .
( [ eq : bolt ] ) ] provide a much smaller and qualitatively different temperature dependence from that seen in the measurements [ compare fig .
[ fig : shifts](a ) with fig .
[ fig : exp_shifts](a ) ] .
if we include a non - zero effective dipole field @xmath64 only in the zero - order calculation , then a negative effective dipole field moves the line shifts for all the transitions down below zero [ fig .
[ fig : shifts](b ) ] . on the other hand ,
if we include an exchange interaction only in the zero - order calculation , then a ferromagnetic exchange interaction ( negative @xmath38 ) moves the line shifts up above zero for all transitions except @xmath9 [ fig .
[ fig : shifts](c ) ] . in both cases ,
the calculated line shifts behave very differently from the measured shifts [ fig .
[ fig : exp_shifts](a ) ] .
therefore , we need to include both the effective dipole field and the exchange interaction in order to explain the measured line shifts . an inter - molecular exchange interaction was recently observed for the different types of single - molecule magnets , mn@xmath107 and mn@xmath107 dimer.@xcite since the effective dipole field @xmath64 depends on the sample shape due to the field - induced net magnetization , we do not use ewald s method@xcite to estimate @xmath64 and thus leave it as a fitting parameter . for the exchange interaction
, we assume that the coupling constant @xmath37 is isotropic along the @xmath108 , @xmath109 , and @xmath110 directions of the triclinic unit cell ( although in experimental samples the exchange interactions are highly anisotropic ) , so that the coordination number is 6 .
the optimum values of @xmath64 and @xmath38 are @xmath111 gauss and @xmath112 gauss ( @xmath113 1 mk ) respectively . with these optimum values , the calculated line shifts [ fig .
[ fig : exp_shifts](b ) ] reproduce well the trends of the temperature dependence of the experimental data [ fig .
[ fig : exp_shifts](a ) ] . figure [ fig : exp_shifts](c ) shows a direct comparison between theory and experiment for a few transitions .
the negative sign of the effective dipole field ( @xmath114 ) indicates that dipoles are antiferromagnetically coupled .
this result seems to be in conflict with the prediction that the dipolar ising spin system with the same structure as fe@xmath0 is ferromagnetically ordered.@xcite however , as pointed out in ref .
, the energy difference between the ferromagnetic and antiferromagnetic states is so small that any neglected effects may shift the ground state to an antiferromagnetic state .
the negative sign of the exchange coupling constant @xmath115 corresponds to ferromagnetic coupling between the effective spins of the molecules .
thus , if there exists any ordering for fe@xmath0 , then the ordering temperature should be estimated by considering both the exchange and dipolar interactions .
the two interactions compete with each other , thereby reducing the possible ordering temperature to a lower value than the ordering temperature with only one of the two interactions considered .
finally , we show the calculated line shifts with several other parameter values that are different from the optimum ones . if the effective dipole field and the exchange interaction both change signs , then the calculated line shift also changes its sign .
if the sign of @xmath64 is opposite to the sign of @xmath38 , then the magnitude of the calculated line shift for @xmath9 is much smaller than those for the other transitions , which does not agree with the experimental data [ compare fig .
[ fig : shifts](d ) with fig .
[ fig : exp_shifts](a ) ] . figure [ fig : shifts](e ) shows the calculated line shifts with @xmath116 and @xmath117 .
figure [ fig : shifts](f ) shows the calculated line shifts with @xmath118 and @xmath119 .
all three figures [ figs . [
fig : shifts](d)-(f ) ] are significantly different from fig .
[ fig : exp_shifts](b ) with the optimized values . for fe@xmath0 the distribution in @xmath2 and
the spin - spin interactions contribute approximately equally to the inhomogeneous line broadening .
figure [ fig : d ] shows the calculated line broadening due to the @xmath2-strain effect only as a function of temperature at @xmath83 ghz . here the standard deviation of the gaussian distribution in @xmath2 , @xmath120 , is approximately @xmath121 .
the line broadening caused by the @xmath2-strain only becomes temperature dependent , because the natural linewidths depend on temperature .
the distribution in @xmath2 makes each molecule subject to a slightly different resonant field .
a measured line shape is a sum of many lorentzian line shapes with a natural linewidth and different resonant fields .
we can calculate the variance of the resonant field , @xmath122 , due to the distribution in @xmath2 from the expression for @xmath60 .
if the natural linewidths are comparable with the variance of the resonant field , then the effect of temperature is significant .
if the natural linewidths are much smaller than the variance , then the effect of temperature is negligible .
the natural linewidth at 10 k ( 35 k ) varies from 7 g ( 29 g ) to 79 g ( 235 g ) as @xmath21 changes from + 10 to + 3 , using the parameter values in ref .. the variance of the resonant field , @xmath122 , varies from 260 g to 70 g as @xmath21 is varied from + 10 to + 3 with @xmath123 .
thus , we find that for small @xmath21 the natural linewidths are comparable with the variance , while for large @xmath21 the natural linewidths are much smaller than the variance . therefore , for small @xmath21
the calculated linewidths show a substantial temperature dependence , while for large @xmath21 there is only a very weak temperature dependence ( see fig .
[ fig : d ] ) . in fig .
[ fig : dp ] the calculated line broadening caused solely by the spin - spin interactions at fixed @xmath2 is shown vs temperature at @xmath83 ghz . here
we use the exchange constant , @xmath124 gauss , which was estimated from the measured line shifts ( sec .
iv a.1 ) , @xmath125 gauss@xmath126 , and @xmath127 gauss@xmath126 . for the ground state @xmath9 ,
the linewidths decrease with increasing temperature in the whole examined temperature range . for @xmath128 , and + 7 ,
the widths first increase sharply with temperature at low temperatures , and then decrease slowly with temperature at high temperatures . for @xmath129 , and
+ 3 , the widths increase with increasing temperature in the whole examined temperature range . as the temperature increases , the @xmath21 dependence of the line broadening due to the spin - spin interactions decreases .
this trend was also seen in the experimental linewidths ( shown as symbols in fig .
[ fig : d+dp ] ) , confirming that the spin - spin interactions are essential to understanding the temperature dependence of the linewidths .
the trends of the temperature dependence can be qualitatively understood through the relative magnitude difference of the thermal energy and the zeeman energy splitting between the states @xmath9 and @xmath11 .
if the zeeman energy splitting is much larger than the thermal energy ( this occurs at low temperatures ) , then the system is polarized .
thus , higher temperature provides larger populations in higher energy levels within the same potential well where the ground state is located .
this leads to an increase in the randomness of the spin orientation so that linewidths become larger with increasing temperature .
if the zeeman energy splitting is much smaller than the thermal energy ( this occurs at high temperatures ) , then some energy levels in _ both _ potential wells are populated . in this case ,
thermal fluctuations increase rapidly with increasing temperature , so that the duration time of the local magnetic field due to neighboring molecules becomes shorter than the spin - spin relaxation time @xmath130 ( the inverse of the natural linewidths ) .
eventually , at very high temperatures the local field is averaged out . therefore , the linewidths decrease with increasing temperature , which usually occurs in paramagnetic materials with very small or zero single - ion anisotropy.@xcite ( this effect is called motional narrowing.@xcite ) therefore , the `` crossover '' temperature where the maximum of the linewidth occurs must be proportional to the zeeman energy splitting between @xmath9 and @xmath11 , which is @xmath131 . for example ,
for the transition @xmath132 at @xmath83 ghz , the resonant field is less than 0.1 t , so the zeeman splitting is about 2@xmath133 .
consequently , its crossover temperature is below the examined temperature range . the crossover temperature increases with decreasing @xmath21 because the resonances are observed at increasing fields for decreasing @xmath21 . for the transitions @xmath134 , and @xmath135 , the crossover temperatures are within the examined temperature range ( see the inset in fig .
6 of ref . )
for @xmath129 , and + 3 , the crossover temperatures are above the studied temperature range .
figure [ fig : d+dp ] shows the experimental data ( symbols ) , and our calculated linewidths ( curves ) , including both the @xmath2-strain effect and the spin - spin interactions with @xmath136 , @xmath124 gauss , @xmath137 gauss@xmath126 , and @xmath138 gauss@xmath126 .
( the spread in @xmath2 here is different from that reported in ref . , because @xmath120 is sample dependent and the samples examined were different .
the value of @xmath87 corresponding to an effective dipole distance of 12 ( fe@xmath0 ) in ref . was about 203 gauss@xmath126 . ) as shown in fig .
[ fig : d+dp ] , our calculated linewidths agree well with the experimental data , except in the low - temperature range for large @xmath21 transitions ( @xmath139 ) .
the experimental linewidths for @xmath9 show @xmath140 dependence in the whole examined temperature range.@xcite for @xmath141 , and + 8 , the calculated linewidths are appreciably smaller than the experimental linewidths below 10 k. as a possible explanation for this discrepancy , we speculate that at low temperatures and large @xmath21 , ( i ) our assumption , @xmath142 , may break down , and/or ( ii ) there might be other linewidth contributions that we have neglected , which should be included along with the dipolar and exchange interactions . in principle , when @xmath61 is not much smaller than unity , a first - order calculation in @xmath61 produces corrections of @xmath143 .
but its implementation is quite complicated , and a first - order calculation may anyway not be sufficient to explain fully the measured linewidths . introducing the concept of the crossover temperature to explain the temperature dependence of the widths seems to be successful for @xmath83 ghz .
however , recent epr experiments ( fig . 8 in ref . ) showed that even when the frequency increased to @xmath144 ghz such that the resonant field for the ground state transition ( @xmath145 ) is approximately 1 tesla , the linewidths for this transition still increased with _ decreasing _ temperatures down to 2 k. this can not be explained using the reasoning given above , because the crossover temperature for the transition is approximately 15 k , so that the linewidths should decrease with decreasing temperature below about 15 k. at present , we do not fully understand the broadening of this ground - state transition .
( at @xmath144 ghz , for other transitions than the ground - state transition , the temperature dependence of the linewidths can be understood using the concept of the crossover temperature . )
consideration of exchange interaction in the linewidths slightly reduces @xmath120 ( from 0.0076d to 0.0064d ) and the dipolar interaction ( from @xmath146 gauss@xmath126 to @xmath137 gauss@xmath126 ) . however , the quality of the linewidth fit including exchange interaction is comparable to that without exchange interaction since the exchange coupling constant is very small compared with the linewidths.@xcite the two fitting parameters , @xmath137 gauss@xmath126 , and @xmath138 gauss@xmath126 , can be calculated using the exact geometry of the system .
the calculated values are @xmath147 gauss@xmath126 and @xmath148 gauss@xmath126 , when the easy axis is 9@xmath149 off from the @xmath108 axis toward the positive @xmath109 axis , and 7@xmath149 off from the @xmath150 plane.@xcite the optimum value of @xmath151 is quite close to @xmath152 , in contrast with @xmath87 .
possible reasons that the optimum value of @xmath87 is much smaller than @xmath153 are as follows : ( i ) in our calculation , we considered each molecule to be a point dipole .
if we consider the atomic positions of the eight fe ions in each molecule and calculate the dipolar interaction between fe ions in different molecules , then the sum of the squared dipolar interaction , @xmath87 , can be significantly reduced .
( ii ) recent nmr experiments for the single - molecule magnet mn@xmath1 showed some spin - density leakage onto the ligands.@xcite this indicates indirectly that for fe@xmath0 the spin density in a single molecule may not be confined only on the core , which would thus reduce the magnetic moments of the eight fe ions .
the above two reasons are our speculations to explain the discrepancy , but it is still unclear why considering the atomic structure within each molecule does not substantially change the value of @xmath151 .
the experimental data for mn@xmath1 are limited to resonance frequencies below 190 ghz , so that the large @xmath21 transitions where the line shifts are significant can not be observed for these low - frequency measurements .
additionally , the linewidths for mn@xmath1 are an order of magnitude larger than those for fe@xmath0 . therefore
, relatively small line shifts are probably masked for mn@xmath1 .
thus , hereafter , the small line shifts are ignored in our analysis , so the exchange interaction and the effective dipole field @xmath64 need not be considered in the linewidth analysis for mn@xmath1 .
the sources of the line broadening are then @xmath2-strain , @xmath3-strain , and the dipolar interaction .
line broadening caused by the @xmath2-strain and @xmath3-strain effects for mn@xmath1 is found to have a weak temperature dependence ( not shown ) , which is similar to the line broadening due to the @xmath2-strain effect for fe@xmath0 .
the contribution of the dipolar interaction to the linewidths is shown vs temperature in fig .
[ fig : dp - mn ] . here @xmath154 gauss@xmath126 . the dipolar broadening increases with increasing temperature for @xmath155 , and @xmath156 .
we do not see the regime where the dipolar broadening decreases with increasing temperature , because the crossover temperature for @xmath157 , about 32 k ( the resonant field is about 1.6 t ) , is close to the highest temperature analyzed ( 40 k ) . unlike fe@xmath0 , the @xmath21 dependence of the dipolar broadening does not decrease with increasing temperature ( the curves are almost parallel ) .
this is also observed in the experimental data ( shown as symbols in fig .
[ fig : d+g+dp ] ) .
we combine the three effects ( @xmath2-strain , @xmath3-strain , and dipolar interactions ) to find that the calculated linewidths agree well with the measured linewidths with @xmath158 , @xmath159 , and @xmath160 gauss@xmath126 , as shown in fig .
[ fig : d+g+dp ] .
( the value of @xmath87 corresponding to an effective dipole distance of 14 for mn@xmath1 in ref . turned out to be the same as that obtained for fe@xmath0 in ref . . ) here the standard deviation of @xmath3 is quite small , so that we can not rule out the possibility of @xmath161 . note that @xmath120 and @xmath162 vary from sample to sample .
the optimum parameter values found here are different from those estimated in ref . , because the examined mn@xmath1 sample was different . the calculated value for @xmath87 , with the exact geometry of mn@xmath1 from ref .
( with each molecule considered as a point dipole ) , is @xmath163 gauss@xmath126 ; this is , again , quite a bit higher than the optimum value for @xmath87 , probably for the same reasons as for fe@xmath0 although the origin of this discrepancy remains unclear .
overall , the temperature dependence of the linewidths for mn@xmath1 is weaker than for fe@xmath0 , because the distribution in @xmath2 for mn@xmath1 is roughly three times as wide as for fe@xmath0 , and the dipolar broadening for fe@xmath0 is comparable to that for mn@xmath1 .
thus , the distribution in @xmath2 conceals the significant temperature dependence of the dipolar broadening for mn@xmath1 . as a consistency check
, we also used the same values of the three parameters ( @xmath158 , @xmath159 , and @xmath160 gauss@xmath126 ) to analyze the measured linewidths@xcite as functions of the energy level @xmath21 for several frequencies ( @xmath164 , and 189.1 ghz ) at a fixed temperature ( t=20 k ) .
our calculated linewidths are in good agreement with the experimental data as shown in fig .
[ fig : mn12_ms ] . due to a dominant contribution of the distribution in @xmath2 to the linewidths
, the linewidths do not depend much on the resonance frequency .
we have investigated how the epr line shifts and linewidths vary with temperature for different energy levels @xmath21 with the applied field along the easy axis for the single - molecule magnets fe@xmath0 and mn@xmath1 .
our calculations consider the spin - spin interactions between molecules , as well as distributions in @xmath2 and @xmath3 .
we have found that the distributions in @xmath2 and @xmath3 provide a weak temperature dependence to the linewidths , and that the spin - spin interactions ( exchange and dipolar interactions ) dominate the temperature dependence of the line shifts and the linewidths . for fe@xmath0 , the line - shift analysis ( figs . [
fig : exp_shifts ] and [ fig : shifts ] ) provides possible evidence of an exchange interaction between molecules , and it determines the sign and order of magnitude of the exchange interaction . the competition of the suggested exchange interaction with the dipolar interaction would tend to lower a possible magnetic ordering temperature
. a small exchange interaction does not affect the linewidth analysis significantly because the exchange coupling constant is much smaller than the typical linewidths .
table [ tab : opt ] summarizes the optimized values of the parameters used in our analysis .
those parameters are , in principle , independent of the resonance frequencies , but some of them are expected to be somewhat batch and shape dependent , as indicated in table [ tab : opt ] . because of the much broader distribution in @xmath2 for mn@xmath1 , the linewidths for fe@xmath0 show a stronger temperature dependence than those for mn@xmath1 .
this conclusion also corroborates our assumption that @xmath2 is distributed for both materials,@xcite although the microscopic origin of this spread is not yet well understood.@xcite
this work was funded by nsf grant nos .
dmr-9871455 , dmr-0120310 , dmr-0103290 , and dmr-0196430 , research corporation ( s.h . ) , and by florida state university through the school of computational science and information technology and the center for materials research and technology .
s. hill , s. maccagnano , n. s. dalal , r. achey and k. park : `` d - strain , g - strain , and dipolar interactions in the fe@xmath0 and mn@xmath1 single - molecule magnets : an epr lineshape analysis '' , submitted to _ physical phenomena in high magnetic fields - iv _ , edited by g. boebinger , l. p. gorkov , a. lacerda and j. r. schrieffer ( world scientific , singapore , in press ) .
.[tab : opt]optimum values of the parameters used in the line - shift and linewidth analysis . here
@xmath120 is the standard deviation of @xmath2 , @xmath162 is the standard deviation of @xmath3 , @xmath38 is the exchange coupling constant between nearest - neighbor molecules ( negative sign means ferromagnetic interaction ) , @xmath165 which vanishes for spherical samples , @xmath85 , and @xmath86 . with experimentally determined values of @xmath2 and @xmath3,@xcite for fe@xmath0
we optimize @xmath38 and @xmath64 for the line shifts and @xmath120 , @xmath87 , and @xmath151 for the linewidths , while for mn@xmath1 we optimize @xmath120 , @xmath162 , and @xmath87 for the linewidths .
the parameters are essentially independent of the measurement frequencies , but some of them are expected to be somewhat batch and shape dependent , as marked by x in the table below . [ cols="^,<,<,^,^,^,^,^",options="header " , ] ( a ) hypothetical smooth temperature dependence of the uniaxial anisotropy parameter @xmath2 for fe@xmath0 .
the functional form used here is @xmath166\}$ ] .
( b ) the resulting calculated line shifts ( the peak position at a given temperature minus the position at @xmath167 k ) due to this temperature dependence of @xmath2 , shown vs temperature at @xmath83 ghz.,title="fig : " ] ( a ) hypothetical smooth temperature dependence of the uniaxial anisotropy parameter @xmath2 for fe@xmath0 .
the functional form used here is @xmath166\}$ ] .
( b ) the resulting calculated line shifts ( the peak position at a given temperature minus the position at @xmath167 k ) due to this temperature dependence of @xmath2 , shown vs temperature at @xmath83 ghz.,title="fig : " ] ( a ) measured line shifts vs temperature at @xmath83 ghz for fe@xmath0 .
( b ) calculated line shifts vs temperature at @xmath83 ghz for fe@xmath0 . here
the calculation is performed to zero order in @xmath61 .
we use the effective dipole field @xmath168 gauss and the exchange coupling constant @xmath124 gauss .
( c ) measured line shifts ( symbols in ( a ) ) , superimposed on calculated line shifts ( curves in ( b ) ) , for several transitions for comparison.,title="fig : " ] ( a ) measured line shifts vs temperature at @xmath83 ghz for fe@xmath0 .
( b ) calculated line shifts vs temperature at @xmath83 ghz for fe@xmath0 . here
the calculation is performed to zero order in @xmath61 .
we use the effective dipole field @xmath168 gauss and the exchange coupling constant @xmath124 gauss .
( c ) measured line shifts ( symbols in ( a ) ) , superimposed on calculated line shifts ( curves in ( b ) ) , for several transitions for comparison.,title="fig : " ] ( a ) measured line shifts vs temperature at @xmath83 ghz for fe@xmath0 .
( b ) calculated line shifts vs temperature at @xmath83 ghz for fe@xmath0 . here
the calculation is performed to zero order in @xmath61 .
we use the effective dipole field @xmath168 gauss and the exchange coupling constant @xmath124 gauss .
( c ) measured line shifts ( symbols in ( a ) ) , superimposed on calculated line shifts ( curves in ( b ) ) , for several transitions for comparison.,title="fig : " ] calculated line shifts vs temperature at @xmath83 ghz for fe@xmath0 , with ( a ) only the dipolar interaction for a spherical sample in higher - order calculations ( @xmath99 , @xmath169 , and @xmath170 ) ( b ) only the effective dipole field , @xmath171 gauss , ( c ) only the exchange coupling constant , @xmath124 gauss , ( d ) @xmath171 gauss @xmath172 0 and @xmath173 gauss @xmath174 0 , ( e ) @xmath171 gauss and @xmath175 gauss , and ( f ) @xmath176 gauss and @xmath124 gauss.,title="fig : " ] calculated line shifts vs temperature at @xmath83 ghz for fe@xmath0 , with ( a ) only the dipolar interaction for a spherical sample in higher - order calculations ( @xmath99 , @xmath169 , and @xmath170 ) ( b ) only the effective dipole field , @xmath171 gauss , ( c ) only the exchange coupling constant , @xmath124 gauss , ( d ) @xmath171 gauss @xmath172 0 and @xmath173 gauss @xmath174 0 , ( e ) @xmath171 gauss and @xmath175 gauss , and ( f ) @xmath176 gauss and @xmath124 gauss.,title="fig : " ] calculated line shifts vs temperature at @xmath83 ghz for fe@xmath0 , with ( a ) only the dipolar interaction for a spherical sample in higher - order calculations ( @xmath99 , @xmath169 , and @xmath170 ) ( b ) only the effective dipole field , @xmath171 gauss , ( c ) only the exchange coupling constant , @xmath124 gauss , ( d ) @xmath171 gauss @xmath172 0 and @xmath173 gauss @xmath174 0 , ( e ) @xmath171 gauss and @xmath175 gauss , and ( f ) @xmath176 gauss and @xmath124 gauss.,title="fig : " ] calculated line shifts vs temperature at @xmath83 ghz for fe@xmath0 , with ( a ) only the dipolar interaction for a spherical sample in higher - order calculations ( @xmath99 , @xmath169 , and @xmath170 ) ( b ) only the effective dipole field , @xmath171 gauss , ( c ) only the exchange coupling constant , @xmath124 gauss , ( d ) @xmath171 gauss @xmath172 0 and @xmath173 gauss @xmath174 0 , ( e ) @xmath171 gauss and @xmath175 gauss , and ( f ) @xmath176 gauss and @xmath124 gauss.,title="fig : " ] calculated line shifts vs temperature at @xmath83 ghz for fe@xmath0 , with ( a ) only the dipolar interaction for a spherical sample in higher - order calculations ( @xmath99 , @xmath169 , and @xmath170 ) ( b ) only the effective dipole field , @xmath171 gauss , ( c ) only the exchange coupling constant , @xmath124 gauss , ( d ) @xmath171 gauss @xmath172 0 and @xmath173 gauss @xmath174 0 , ( e ) @xmath171 gauss and @xmath175 gauss , and ( f ) @xmath176 gauss and @xmath124 gauss.,title="fig : " ] calculated line shifts vs temperature at @xmath83 ghz for fe@xmath0 , with ( a ) only the dipolar interaction for a spherical sample in higher - order calculations ( @xmath99 , @xmath169 , and @xmath170 ) ( b ) only the effective dipole field , @xmath171 gauss , ( c ) only the exchange coupling constant , @xmath124 gauss , ( d ) @xmath171 gauss @xmath172 0 and @xmath173 gauss @xmath174 0 , ( e ) @xmath171 gauss and @xmath175 gauss , and ( f ) @xmath176 gauss and @xmath124 gauss.,title="fig : " ] calculated full width at half maximum ( fwhm ) caused by the gaussian distribution in the uniaxial anisotropy parameter @xmath2 only , shown vs temperature from 4 k to 35 k at @xmath83 ghz for fe@xmath0 .
@xmath145 indicates the transition from the energy level @xmath145 to @xmath177 , etc .
the standard deviation of @xmath2 is approximately @xmath121 . ]
calculated fwhm caused by the spin - spin interactions only , shown vs temperature at @xmath83 ghz for fe@xmath0 . here
the exchange constant @xmath38 is @xmath178 gauss , @xmath179 gauss@xmath126 and @xmath180 gauss@xmath126 . ]
calculated ( curves ) and measured fwhm ( symbols ) vs temperature at @xmath83 ghz for fe@xmath0 .
here we use the standard deviation of @xmath2 , @xmath136 , the exchange coupling constant , @xmath181 gauss , @xmath137 gauss@xmath126 , and @xmath138 gauss@xmath126 .
the solid curves , from bottom to top , correspond to @xmath182 .
see the text for possible sources of the discrepancy between theory and experiment for low temperatures and large @xmath21 . ]
calculated fwhm for mn@xmath1 caused by the dipolar interactions only , shown vs temperature , at @xmath103 ghz , with the sum of the squared dipolar interactions @xmath160 gauss@xmath126 .
the examined temperature range for mn@xmath1 is from 10 k to 40 k. ] calculated ( curves ) and measured ( symbols ) fwhm vs temperature at @xmath103 ghz for mn@xmath1 . here
the @xmath2-strain ( @xmath158 ) , @xmath3-strain ( @xmath159 ) , and the dipolar interactions ( @xmath183 gauss@xmath126 ) are included in the calculated linewidths . ]
calculated ( curves ) and measured ( symbols ) fwhm vs energy level @xmath21 at @xmath184 k for @xmath185 , @xmath186 , and @xmath187 ghz for mn@xmath1 . here
the values of @xmath120 , @xmath162 , and @xmath87 are the same as those in fig .
[ fig : d+g+dp ] . because of a relatively small contribution of the dipolar interaction to the linewidths , the linewidths do not change much with the resonance frequency . ] | we examine quantitatively the temperature dependence of the linewidths and line shifts in electron paramagnetic resonance experiments on single crystals of the single - molecule magnets fe@xmath0 and mn@xmath1 , at fixed frequency , with an applied magnetic field along the easy axis .
we include inter - molecular spin - spin interactions ( dipolar and exchange ) and distributions in both the uniaxial anisotropy parameter @xmath2 and the land @xmath3-factor .
the temperature dependence of the linewidths and the line shifts are mainly caused by the spin - spin interactions .
for fe@xmath0 and mn@xmath1 , the temperature dependence of the calculated line shifts and linewidths agrees well with the trends of the experimental data .
the linewidths for fe@xmath0 reveal a stronger temperature dependence than those for mn@xmath1 , because for mn@xmath1 a much wider distribution in @xmath2 overshadows the temperature dependence of the spin - spin interactions . for fe@xmath0 ,
the line - shift analysis suggests two competing interactions : a weak ferromagnetic exchange coupling between neighboring molecules and a longer - ranged dipolar interaction .
this result could have implications for ordering in fe@xmath0 at low temperatures . |
recently , the lhcb collaboration has reported the observation of the decay @xmath2 @xcite .
the interest in this decay is related to the possible existence of the exotic glueball state decaying into the @xmath0 pair @xcite .
the spin - parity quantum numbers of the resonance states decaying into @xmath0 are reported to be @xmath11 , and @xmath12 @xcite .
the lhcb collaboration has also reported the observation of four resonance structures in the decay @xmath13 , in the mass range 4140 4700 mev , decaying into @xmath1 @xcite .
this group of resonances is widely discussed because of their possible exotic nature @xcite , side by side with the explanations based on the dynamical rescattering effects @xcite .
two of them , @xmath14 and @xmath15 , have masses in the range 4100 4350 mev attained in the @xmath1 mass spectrum in the decay @xmath2 @xcite .
the preferable spin - parity quantum numbers of these resonances are @xmath6 @xcite . the data of ref .
@xcite were plotted against the phase space distribution which was shown to be inadequate because the resonance bumps were seen in both the @xmath0 and @xmath1 mass spectra @xcite .
the aim of the present work is to write down the amplitudes and the mass spectra of the above final states in the decay @xmath2 upon taking into account possible intermediate resonance states in the @xmath0 and @xmath1 channels , irrespective of the model assumptions about their nature .
the task of obtaining the effective amplitudes is greatly simplified when one takes into account the near - threshold kinematics of the decay @xmath2 .
then one can neglect the higher partial waves in the decay amplitudes .
the expressions for the @xmath0 and @xmath1 mass spectra obtained under such an approximation upon taking into account the intermediate resonances @xmath16 and @xmath17 are used to extract from the fits the masses and coupling constants of these resonances .
it should be emphasized that the presentation of the results in terms of coupling constants is more informative than the popular representation in terms of the partial widths .
the paper is organized as follows .
section [ sec2 ] contains the expressions for the lowest momenta effective vertices of the @xmath18 , @xmath19 , @xmath16 , and @xmath17 transitions assuming @xmath3 for the @xmath20 resonances and @xmath6 for the @xmath21 ones .
the partial decay widths @xmath16 and @xmath22 are given there .
section [ amps ] is devoted to the derivation of the @xmath2 decay amplitudes upon taking into account the @xmath9 mixing due to the common @xmath0 decay channel , and @xmath10 mixing due to the @xmath1 one .
the modulus squared of the @xmath2 decay amplitude and the expressions for the @xmath0 and @xmath1 mass spectra are calculated in sec .
[ spectra ] . in sec .
[ sec5 ] , these expression are applied to the description of the lhcb data @xcite .
section [ sec6 ] contains the brief discussion of the obtained results .
section [ sec7 ] serves as a conclusion .
some details used in the derivation of expressions in the main text are given in the appendices .
we assume the existence of the resonances at @xmath23 and 2.2 gev in the @xmath0 mass spectrum which will be called as @xmath20 , and the resonances @xmath14 and @xmath15 with masses in the range 4100 4350 mev , in the @xmath1 mass spectrum @xcite which will be called @xmath21 .
one needs the effective vertices @xmath24 , @xmath25 , @xmath26 , @xmath27 .
see fig .
[ fig1 ] . since all particles in the final state of the decay @xmath28 have unit spin , the number of effective contributions to the decay amplitude is frighteningly large , especially when taking into account the fact that space parity is not conserved because the @xmath29 meson decays due to the weak interactions , so that only the angular momentum conservation , not the space parity , restricts the number of possible lorenz structures in the effective decay amplitudes .
the couplings of @xmath30 with @xmath0 and @xmath31 with @xmath1 are considered to be due to the strong interactions , hence they conserve parity , but again the unit spin admits many independent effective contributions . the situation can be greatly simplified if one takes into account the fact that the kinematics of the decay @xmath32 is such that all particles in the final state have relatively low momenta . indeed , the invariant mass of the @xmath0 pair varies in the range @xmath33 that is , @xmath34 gev .
the maximum momentum of the @xmath35 is reached when the final @xmath36 meson is at rest while two @xmath35 mesons move in opposite directions , and this gives @xmath37 .
analogously , the invariant mass of the @xmath1 state varies in the range @xmath38 that is , @xmath39 gev
. the maximum momentum of the @xmath36 meson is reached when one of the @xmath35 mesons is at rest , while other moves oppositely to @xmath36 resulting in @xmath40 .
however , these relatively small ratios are not in fact reached because the above kinematic situations taking place at the border of phase space are suppressed by the final state phase space factors .
see eqs .
( [ s12 ] ) and ( [ s23 ] ) below .
this permits one to take into account only the effective decay vertices with the lowest nonvanishing powers of momenta .
let us start with the parity - conserving effective vertices of @xmath25 and @xmath27 .
the required lowest momenta @xmath25 vertices for various quantum numbers are @xmath41 where @xmath42 is the polarization tensor of the spin two resonance , @xmath43 and @xmath44 stand for the polarization four - vector and four - momentum of the @xmath45 meson , and @xmath46 is the totally antisymmetric unit levy - civit tensor .
the quantum numbers of the @xmath14 and @xmath15 resonances are now established : @xmath47 @xcite . then the effective lowest momentum @xmath27 vertex looks like @xmath48 in the above expressions , @xmath49 , @xmath50 stand for the polarization four - vectors of the @xmath31 , @xmath36 mesons , respectively , and @xmath51 is the four - momentum of the @xmath31 resonance .
the justification of this expression is given in the appendix [ b ] .
all effective vertices are assumed to be real , and the possible dependence on the momentum squared is neglected .
let us give the effective amplitudes for the weak decays of the @xmath29 meson .
the lowest momenta amplitudes of the transitions @xmath24 for different quantum numbers of the resonance @xmath30 are @xmath52where @xmath53 .
note that , in the first expression in eq .
( [ br ] ) , the amplitude conserves ( breaks ) the space parity in the case of @xmath30 with @xmath54 ( @xmath55 ) , respectively . the expression of the decay amplitude to the tensor resonance in the second line of eq .
( [ br ] ) breaks parity .
the parity - conserving amplitude is @xmath56 , hence it has the @xmath57 wave form to be neglected side by side with the parity - breaking @xmath57 wave expression @xmath58 .
the lowest momentum @xmath26 transition amplitude , @xmath59 breaks space parity .
our goal here is to take into account the energy dependence of the partial widths of the resonances involved , as well as their mixing due to the common decay modes ( if any ) .
the particle data group ( pdg ) gives the @xmath0 , @xmath60 decay modes for the tensor @xmath61 , @xmath62 , and @xmath63 resonances and the @xmath64 one for the @xmath63 resonance @xcite .
again , because of the low statistics of the available data we will take into account only the @xmath0 decay mode relevant in the context of the data presented in ref .
the same assumption will be adopted for the scalar and pseudoscalar ones observed by the besiii collaboration @xcite .
the standard calculation gives the partial decay widths for the resonances with the given quantum numbers .
taking into account only the lowest nonvanishing momenta , one gets , using the effective vertices eq .
( [ rvert ] ) , the following expressions : @xmath65hereafter , @xmath66is the standard klln function .
correspondingly , the @xmath67 partial width calculated from eq .
( [ xvert ] ) is @xmath68 in what follows , the @xmath2 decay amplitudes to be derived below will include only the resonance contribution with specific @xmath69 . by this reason , all resonances @xmath30 with different quantum numbers
will be labeled by the single letter @xmath30 in the coupling constants , without pointing to the @xmath69 quantum numbers .
according to the diagrams shown in fig .
[ fig1 ] , let us represent the @xmath2 decay amplitude as the sum of the @xmath30 and @xmath31 resonance contributions , @xmath70 the diagrams of the decay @xmath2.,width=302 ] the lhcb data @xcite visibly demonstrate the appearance of two enhancements in the @xmath0 and @xmath1 mass spectra .
so , it will be assumed in what follows that there are two resonances @xmath4 , @xmath5 with the spin zero and ( or ) spin two in the @xmath0 system mass range from 2.0 to 2.4 gev @xcite and two resonances @xmath7 , @xmath8 with @xmath6 in the @xmath1 system mass range from 4.14 to 4.35 gev @xcite .
their masses are close , and the resonances have common decay modes like @xmath0 in the case of @xmath20 or @xmath1 in the case of @xmath21 .
they can mix inside each group . using refs .
@xcite we take into account the @xmath9 and @xmath10 mixing by introducing the nondiagonal polarization operators @xmath71 and @xmath72 , respectively .
then the simplest breit - wigner resonance contribution in the case of , say , @xmath20 , @xmath73 should be generalized to include the effects of energy dependent widths and mixing by means of introducing the amplitude @xmath74 with amputated kinematic factors ( to be included below ) .
taking into account the effect of resonance mixing in general case is outlined in appendix [ b ] by eq .
( [ ggen ] ) . in the case of two mixed resonances eq .
( [ ggen ] ) reduces to @xmath75 here , @xmath76with @xmath77 where @xmath78 for different quantum numbers of the @xmath20 resonances are given by eq .
( [ rwidth ] ) , and the nondiagonal polarization operator which includes the common @xmath0 mode is @xmath79 analogously , the mixing of @xmath31 resonances is taken into account by introducing the amplitude with amputated kinematical factors : @xmath80 the inverse propagator of the @xmath81 resonances ( @xmath77 ) that appears in eq .
( [ gx12 ] ) is @xmath82where @xmath83 is the mass of the @xmath81 resonance .
the nondiagonal polarization operator responsible for @xmath10 mixing is written analogously to eq .
( [ pir12 ] ) : @xmath84 the partial width of the @xmath81 decay to @xmath1 is given by eq .
( [ xwidth ] ) . equations ( [ gr12 ] ) and ( [ gx12 ] )
reduce to the simple sum of two breit - wigner contributions in the case of vanishing mixing @xmath85 .
the explicit alternative expression for the mixing amplitude valid in the case of the two resonances mixed via the single common decay channel is given by eq .
( [ gpart ] ) in appendix [ b ] . as for the eqs .
( [ dri ] ) and ( [ pir12 ] ) , their imaginary parts originated from the @xmath0 loop contribution are fixed by the unitarity relation .
the real parts are divergent when calculated from the dispersion relation upon neglecting the vertex form factors , @xmath86 in the case of @xmath55 and @xmath12 , when restricting to the @xmath87-wave approximation , the divergence is logarithmic and can be regularized by the subtraction at the resonance mass .
however , there is the @xmath57 wave contribution neglected in eq .
( [ rwidth ] ) which makes the divergence much stronger , and there are no model independent ways to fix the subtraction constants .
the same holds for the @xmath88 wave decay of the @xmath54 resonance .
alternatively , one may insert the vertex form factors to make the loop integrations finite , but this again requires the fixing of additional free parameters characterizing the above form factors .
the same refers to the @xmath31 resonances whose decay width contains the @xmath57-wave contribution essential in the dispersion integral at large momenta . in practice
, one can adopt the following way of action . in the present case ,
there are no sharp energy dependencies of the loop effects like those observed in the case of the difference of the @xmath89 and @xmath90 loop contributions @xcite .
the decay kinematics is such that it involves relatively narrow intervals of the invariant masses of the @xmath0 and @xmath1 states .
see eqs .
( [ m12 ] ) and ( [ m13 ] ) .
hence , one can ignore the possible dependence on energy of the really incalculable smooth real parts of the loop contributions and take them as constants . in the case of the diagonal polarization operators ,
the constants are absorbed in the masses of the resonances @xmath91 , while in the nondiagonal one , @xmath92 , it is included as the free parameter @xmath93 , all these to be determined from the fit . after these remarks , one can write the contribution of the mixed resonances in the @xmath0 state .
\(a ) @xmath94 .
@xmath95 which is odd under the parity inversion .
hereafter @xmath53 .
\(b ) @xmath96 .
@xmath97which is even under the parity inversion .
\(c ) @xmath98 .
@xmath99where @xmath100with @xmath101stands for the result of summation over the polarizations of the intermediate tensor resonance ; @xmath102 .
when calculating mass spectra , it is useful to do this in the rest reference frame of the @xmath0 pair , @xmath103 , where @xmath104 is the invariant mass squared of the @xmath0 state . in this frame
, @xmath105 reduces to the three - dimensional form expressed through the kronecker delta : @xmath106 the second necessary ingredient for obtaining the @xmath0 and @xmath1 mass spectra in the decay @xmath2 is the contribution of the @xmath31 exchange schematically depicted as the second and third feynman diagrams in fig .
it looks like @xmath107,\end{aligned}\]]where @xmath108 and @xmath109 stand for the invariant mass squared of the @xmath110 states , respectively . note that the amplitude ( [ mx ] ) is even under parity reflection and symmetric under permutation of two @xmath35 mesons .
equation ( [ mx ] ) can be simplified when taking into account the small momenta of final particles .
it is composed as the difference of two lorenz - invariant expressions each of which can be evaluated in its respective rest reference frame : @xmath111\cdot{\bm\epsilon}^{(j/\psi)}\right)_{13}-\nonumber\\ & & m_{23}g^{(x)}_{12}(m^2_{23})\left([{\bm\epsilon}_1\times{\bm\epsilon}_2]\cdot{\bm\epsilon}^{(j/\psi)}\right)_{23},\end{aligned}\ ] ] where indices 13 or 23 at the vector structures point to the rest reference frame of the state @xmath112 or @xmath113 .
now , the three - dimensional polarization vector @xmath114 of the vector meson with the four - momentum @xmath115 is expressed through its counterpart @xmath116 in the rest frame : @xmath117 hence , both three - dimensional polarization structures in eq .
( [ mxa ] ) can be represented in the form which includes only the polarization three - vectors in the rest frame , @xmath118\cdot{\bm\epsilon}^{(j/\psi)}\approx [ { \bm\xi}_1\times{\bm\xi}_2]\cdot{\bm\xi}^{(j/\psi)},\ ] ] because they differ by the terms squared in momenta which can be neglected in the considered case . under this approximation
the amplitude of @xmath31 exchange looks like @xmath119\times\nonumber\\&&[{\bm\xi}_1\times{\bm\xi}_2]\cdot{\bm\xi}^{(j/\psi)}.\end{aligned}\ ] ] the expressions for the @xmath0 and @xmath1 mass spectra in the decay @xmath2 are given in the next section .
when calculating the modulus squared of the relevant amplitude , @xmath120 , where the invariant masses squared in eqs .
( [ sm12 ] ) , ( [ m213 ] ) , and ( [ m223 ] ) are subjected to the constraint @xmath121one should take into account the approximately nonrelativistic character of the problem and keep only the lowest powers of the final particle momenta . in this case , the four - dimensional scalar product of two four - momenta is @xmath122then one obtains the required expressions for the specific quantum numbers of the resonance in the @xmath0 state .
they are the following .
\(a ) @xmath94 . in this case
@xmath123 is odd while @xmath124 is even under the space parity reflection hence they do not interfere .
the modulus squared of the decay amplitude looks like latexmath:[\ ] ] where @xmath195 and @xmath196 are the mass and the rest frame polarization three - vector of the meson @xmath193 , respectively .
one can denote the factors in front of the polarization structure as @xmath197 and use the expression @xmath198 as the effective vertex of the @xmath27 decay in the low momentum approximation adopted throughout the paper .
let us make some remarks concerning the above expressions . as for eqs .
( [ gr12 ] ) and ( [ gx12 ] ) , one can include more than two mixed resonances and more than one decay channel @xcite .
this can be done by generalizing either eq .
( [ gr12 ] ) or ( [ gx12 ] ) to the expression for the @xmath200 transition amplitude through the mixed resonances @xmath201 where @xmath202 is the number of mixed resonances , each with the inverse propagator of the specific resonance like eq .
( [ dri ] ) or ( [ dxi ] ) , and with polarization operator of mixing @xmath203 , where @xmath204 enters the partial decay width of the resonance @xmath205 in the form @xmath206 .
in fact , it is this form that is used in the present work .
one may observe that if re@xmath209 then in the vicinity of @xmath210 ( @xmath211 ) eq .
( [ gpart ] ) reduces to the simple breit - wigner form eq .
( [ bw ] ) for the resonance 1 ( 2 ) .
99 r. aaij _ et al . _
( lhcb collaboration ) , j. high energy phys .
03 , ( 2016 ) 040 .
s. j. lindenbaum and r. s. longacre , phys . lett . * 165b * , 202 ( 1985 ) .
a. etkin , k. j. foley , r. s. longacre , w .
a. love , t. w. morris , e. d. plattner , a. c. saulys , s .
j. lindenbaum , c .
s. chan , and m. a. kramer phys .
165b * , 217 ( 1985 ) .
p. s. l. booth _
nucl . phys . *
b273 * , 677 ( 1986 ) .
a. etkin _ et al .
b * 201 * , 568 ( 1988 ) .
k. a. olive _ et al . _
( particle data group ) , chin .
c * 38 * , 090001 ( 2014 ) .
m. ablikim _ et al . _
( besiii collaboration ) , phys .
d * 93 * , 112011 ( 2016 ) .
r. aaij _ et al .
_ ( lhcb collaboration ) , arxiv:1606.07895[hep - ex ] .
r. aaij _ et al .
_ ( lhcb collaboration ) , arxiv:1606.07898[hep - ex ] .
a. ali , i. ahmed , m. j. aslam , and a. rehman , phys.rev .
d * 94 * , 054001 ( 2016 ) .
r. zhu , phys .
d * 94 * , 054009 ( 2016 ) .
l. maiani , a. d. polosa , and v. riquer , phys .
d * 94 * , 054026 ( 2016 ) .
x. h. liu , arxiv:1607.01385 .
t. aaltonen _ et al . _
( cdf collaboration ) , arxiv:1101.6058[hep - ex ] .
r. aaij _ et al .
_ ( lhcb collaboration ) , phys . rev .
d * 85 * , 091103(r ) ( 2012 ) .
n. n. achasov and a. a. kozhevnikov , phys .
d * 55 * , 2663 ( 1997 ) .
n. n. achasov and a. a. kozhevnikov , phys .
d * 62 * , 117503 ( 2000 ) .
n. n. achasov , s. a. devyanin , and g. n. shestakov , phys . lett . * 88b * , 367 ( 1979 ) .
n. n. achasov , a. a. kozhevnikov , and g. n. shestakov , phys .
d * 93 * , 114027 ( 2016 ) . | the mass spectra of the @xmath0 and @xmath1 states in the decay @xmath2 recently observed by lhcb are calculated in the model which takes into account the @xmath3 intermediate resonances @xmath4 , @xmath5 in the @xmath0 channel and the @xmath6 ones , @xmath7 , @xmath8 , in the @xmath1 channel .
when obtaining the expressions for the effective amplitudes and mass spectra , the approximate threshold kinematics of the decay is used essentially .
the @xmath9 and @xmath10 mixings arising due to the common decay modes @xmath0 and @xmath1 , respectively , are also taken into account .
the obtained expressions for the mass spectra are applied for extracting the information about masses and coupling constants of the resonances in the @xmath0 and @xmath1 final states . |
let @xmath0 be a discrete or continuous - time flow defined on a metric space @xmath1 .
if @xmath2 is a compact isolated @xmath0-invariant set which is not a repeller , we introduce an spectrum , called _ metric spectrum _ , which measures the exponential rate of approximation to @xmath3 of points starting in the stable set of @xmath3 , as illustrated by figure [ figura1 ] .
the approximation of @xmath4 to @xmath3 . ]
we show that this spectrum is an invariant of the conjugation classes by bi - lipschitz maps .
in particular , when @xmath5 is a manifold , the metric spectrum is independent of riemannian metrics .
when the restriction of @xmath0 to an open neighborhood of @xmath3 is conjugated to a linear flow @xmath6 , the metric spectrum is given by the negative lyapunov exponents of @xmath6 .
in particular , if @xmath3 is normally hyperbolic , this implies that its metric spectrum is not empty .
now let @xmath7 be a linear induced flow on a flag manifold @xmath8 of a connected noncompact real semi - simple lie group @xmath9 .
assuming that @xmath7 is conformal ( its unipotent jordan component is trivial ) , we determine explicitly the metric spectrum of each morse component of the finest morse decomposition .
this is done by combining the linearization presented in @xcite with some results on the jordan decomposition of @xmath7 presented in @xcite .
this result generalizes the following simple situation .
let @xmath10 , @xmath11 and @xmath12 we have that @xmath13 $ ] and @xmath14 $ ] are , respectively , the attractor and repeller components of @xmath7 in @xmath15 . for each @xmath16 \neq [ e_2]$ ] , we have that @xmath17 \to [ e_1]$ ] , when @xmath18 .
the @xmath19-invariant distance @xmath20 in @xmath15 is such that @xmath21 , [ e_1 ] ) = \theta_t$ ] , where @xmath22 is the angle illustrated in the figure [ figura2 ] .
the approximation of @xmath17 $ ] to @xmath13 $ ] in @xmath15 . ] since @xmath23 it follows that @xmath24 , [ e_1 ] ) = -2\lambda,\ ] ] where we use that @xmath25 for the attractor component of @xmath7 in the maximal flag manifold @xmath26 , we do not need to assume that @xmath7 is conformal .
when @xmath7 is arbitrary , the metric spectrum of its attractor component in the maximal flag manifold will be called the _ flag spectrum _ of @xmath7 .
for the theory of semi - simple lie groups and their flag manifolds we refer to duistermat - kolk - varadarajan @xcite , helgason @xcite , knapp @xcite and warner @xcite . to set notation
let @xmath9 be a connected noncompact real semi - simple lie group with lie algebra @xmath27 .
let @xmath28 be the adjoint representation of @xmath9 .
an element @xmath29 acts in the lie algebra @xmath30 by the adjoint representation , so that for @xmath31 we write @xmath32 with this notation we have that @xmath33 for @xmath29 , @xmath34 .
fix a cartan involution @xmath35 of @xmath36 , a maximal abelian subspace @xmath37 and a weyl chamber @xmath38 .
let @xmath39 be the cartan decomposition and @xmath40 be the cartan inner product associated to @xmath41 .
we let @xmath42 be the set of roots of @xmath43 , @xmath44 the positive roots corresponding to @xmath45 , @xmath46 the set of simple roots in @xmath44 . for each @xmath47
, we denote by @xmath8 the associated flag manifold , which is a homogeneous space of @xmath9 .
let @xmath48 be the base point , @xmath49 be its isotropy group and @xmath50 be its isotropy algebra .
each element @xmath31 of the lie algebra induces a differentiable flow in the flag manifold @xmath51 given by @xmath52 this flow is generated by the induced vector field which is denoted by @xmath53 and its value at @xmath54 is denoted by @xmath55 .
an element @xmath29 acts in a tangent vector @xmath56 by its differential at @xmath54 .
we denote this action by @xmath57 . for induced vector fields we have that @xmath58 for a fixed @xmath59 ,
the map @xmath60 is a linear map from @xmath61 to @xmath62 whose kernel is the subalgebra of isotropy at @xmath54 .
we have that the compact group @xmath63 is transitive in @xmath8 , its isotropy subgroup at @xmath48 is denoted by @xmath64 and its elements acts in @xmath61 by @xmath40-isometries .
there exists a @xmath64-invariant subalgebra @xmath65 which complements @xmath50 in @xmath61 .
thus we can identify the tangent space of @xmath8 at @xmath48 with @xmath65 and @xmath40 determines a @xmath66-invariant riemannian metric in @xmath8 given by @xmath67 where @xmath68 with @xmath69 .
the weyl group @xmath70 acts on @xmath71 by isometries , where @xmath72 and @xmath3 are , respectively , the normalizer and the centralizer of @xmath71 in @xmath66 .
we denote by @xmath7 , @xmath73 or @xmath74 , the right invariant continuous - time flow generated by @xmath75 or the discrete - time flow generated by @xmath29 .
more precisely , when @xmath76 , we have that @xmath77 and , when @xmath78 , we have that @xmath7 is the @xmath79-iterate of @xmath80 . the flow induced by @xmath7 on @xmath8
is called a linear induced flow .
let @xmath81 be the multiplicative jordan decomposition of @xmath7 , a commutative composition of linear induced flows .
the elliptic , hyperbolic and unipotent components of @xmath7 are , respectively , the flow @xmath82 , the flow @xmath83 and the flow @xmath84 , where @xmath85 is called the hyperbolic type of the flow @xmath7 and @xmath86 is a nilpotent operator .
the hyperbolic component is gradient with respect to a given riemannian metric on @xmath51 ( see @xcite , section 3 ) .
the connected components of fixed point set of this flow are given by @xmath87 where @xmath88 is the centralizer of @xmath85 in @xmath66 .
the stable and unstable sets of @xmath89 with respect to @xmath90 are given , respectively , by @xmath91 where @xmath92 , @xmath93 for an arbitrary linear induced flow on @xmath8 , we have the following similar description for its finest morse decomposition ( see proposition 5.1 of @xcite ) . [ propmorseflag ]
let @xmath7 be a linear induced flow on @xmath8 .
the set @xmath94 is the finest morse decomposition for @xmath7 .
furthermore , the stable and unstable sets of @xmath89 with respect to @xmath7 are given , respectively , by @xmath95 now we define the subspace @xmath96 and the family of subspaces @xmath97 given by @xmath98 for @xmath99 , where @xmath100 . by proposition 3.1 of @xcite ,
the families @xmath101 are well defined , each @xmath102 is @xmath90-invariant and , for @xmath100 , we have @xmath103 furthermore we have that the map @xmath104 is a linear isomorphism , for each @xmath105 . by propositions 3.2 and 3.3 of @xcite
, we have that @xmath106 are differentiable subbundles of the tangent bundle of @xmath8 over @xmath107 such that its whitney sum @xmath108 is the normal bundle of @xmath107 and whose fiber at @xmath109 is given by @xmath110 it follows that the map @xmath111 is a linear isomorphism , for each @xmath105 .
now we define the linearization map by @xmath112 where @xmath113 and @xmath114 . by theorem 3.4 of @xcite
, we have the following result .
[ teolineariz ] the map @xmath115 takes the null section @xmath116 onto @xmath89 and satisfies : * its restriction to some neighborhood @xmath117 of @xmath116 in @xmath118 is a diffeomorphism over a neighborhood @xmath119 of @xmath89 in @xmath8 . *
its restrictions to @xmath120 are diffeomorphisms , respectively , onto @xmath121 and @xmath122 .
the above map @xmath115 is a conjugation of the flow @xmath7 on a neighborhood of the attractor component @xmath123 to the linear flow @xmath124 on a neighborhood of the null section of @xmath118 if and only if @xmath118 is invariant and @xmath125 is equivariant by the flow @xmath7 .
a sufficient condition is that the map @xmath126 be equivariant by @xmath7 , i.e , that @xmath127 . for the attractor component @xmath128 , by corollary 3.6 of @xcite
, this happens whenever either @xmath129 or @xmath130 , where @xmath131 is the annihilator of @xmath85 in the simple roots @xmath132 .
for the other morse components , by proposition 5.5 of @xcite , it is sufficient that @xmath7 be conformal , i.e. , its unipotent component be trivial .
let @xmath0 be a discrete or continuous - time flow defined on a compact metric space @xmath1 .
if @xmath2 is an isolated compact @xmath0-invariant set , its stable set is given by @xmath133 let us assume that @xmath3 is not a repeller , which is equivalent to that @xmath134 is not empty .
we say that @xmath135 is _ metric regular _ if there exists the following limit @xmath136 called the _
metric exponent of @xmath0 starting at @xmath54 _ , measuring the exponential rate of approximation of @xmath4 to @xmath3 .
the _ metric spectrum of @xmath0 relative to @xmath3 _ is defined by @xmath137 the following result shows that @xmath138 is an invariant of the conjugation classes by bi - lipschitz maps .
let @xmath139 be a metric space and @xmath140 be a bi - lipschitz map .
if @xmath141 , @xmath142 and @xmath143 , then @xmath144 and @xmath145 in particular , @xmath138 is independent of equivalent metrics .
since @xmath125 is a bi - lipschitz map , there exist positive constants @xmath146 such that @xmath147 for every @xmath148 .
we have that @xmath149 and thus @xmath150 showing that @xmath151 , since @xmath152 the last assertion follows , since two metrics @xmath20 and @xmath153 are equivalent if and only if the identity map from @xmath154 to @xmath155 is a bi - lipschitz map .
since diffeomorphisms between riemannian manifolds are locally bi - lipschitz maps , we have the following result .
[ corlipschitz ] let @xmath156 be a diffeomorphism between the riemannian manifolds @xmath1 and @xmath139 . if @xmath157 , @xmath158 and @xmath159 , then @xmath144 and @xmath145 in particular , @xmath160 is independent of riemannian metrics .
since @xmath5 is locally compact and @xmath3 is a compact subset , there exists an compact neighborhood @xmath161 of @xmath3 . since @xmath125 is a diffeomorphism , it is a locally bi - lipschitz map and its restriction to @xmath161 is a bi - lipschitz map onto its image .
now we establish the connection between the lyapunov and the metric spectra .
we recall that for a linear flow @xmath6 of a normed vector bundle @xmath162 , we say that @xmath163 is _ lyapunov regular _ if there exists the following limit @xmath164 called the _ lyapunov exponent of @xmath6 starting at @xmath165 _ , measuring the exponential rate of growth of @xmath166 .
the _ lyapunov spectrum of @xmath6 _ is defined by @xmath167 the _ stable lyapunov spectrum of @xmath6 _ is defined by @xmath168 where @xmath169 is the stable set of the null section of @xmath170 .
[ proplyapunovmetric ] let @xmath1 be a riemannian manifold and @xmath171 be a diffeomorphism between some open neighborhood @xmath119 of the null section @xmath172 of the normal bundle @xmath170 of @xmath3 and some open neighborhood @xmath173 of @xmath3 . if @xmath174 and @xmath175 is the restriction to @xmath119 of a linear flow @xmath6 of @xmath170 , then @xmath176 for all @xmath177 , lyapunov regular , and @xmath178 furthermore , we have that @xmath138 is not empty
. there exist @xmath179 such that @xmath180 and @xmath181 .
since @xmath182 and @xmath183 , we can assume that @xmath184 and @xmath185 .
by corollary [ corlipschitz ] , we have that @xmath186 thus it remains to prove that @xmath187 , for all @xmath177 , metric regular .
let @xmath153 be the sasaki metric on the normal bundle @xmath170 ( see section 4.2 of @xcite ) .
we can choose @xmath119 such that @xmath188 , for all @xmath184 .
thus , for all @xmath189 , we have that @xmath165 is metric regular if and only if @xmath165 is lyapunov regular and , in this case , @xmath190 . for the last assertion
, we first observe that @xmath172 is a @xmath6-invariant and isolated , since @xmath3 is a @xmath0-invariant and isolated .
thus the restriction of @xmath169 over some chain transitive component of the flow induced by @xmath6 on the base @xmath3 is a subbundle ( see theorem 2.13 of @xcite ) .
thus we can apply oselec theorem ( see @xcite ) to the restriction of @xmath6 to this subbundle , showing that @xmath191 is not empty . the previous proposition and theorem 1 of @xcite
imply the following result .
if @xmath3 is normally hyperbolic , then @xmath138 is not empty .
in this section , we determine explicitly the metric spectrum of a linear induced flow @xmath7 relative to a morse component @xmath192 .
first we need to introduce some suitable constructions , which are related to the constructions presented in the section [ preliminaries ] .
we denote @xmath193 where @xmath194 now writing @xmath195 for each @xmath196 , we denote @xmath197 and define the family of subspaces given by @xmath198 first we note that the eigenspaces of @xmath207 are invariant by the centralizer @xmath208 of @xmath85 in @xmath9 , since the elements of @xmath208 commute with @xmath207 .
the first claim follows , since @xmath209 , for @xmath200 , where @xmath210 . by corollary 3.6 and 3.9 of @xcite and proposition 5.5 of @xcite ,
if @xmath211 or @xmath212 are verified , then we that @xmath127 .
thus the second claim follows , since @xmath213 . following the proof of proposition 3.2 of @xcite ,
for each @xmath196 , we have that @xmath214 is a differentiable subbundle of the tangent bundle of @xmath215 .
the norm in @xmath118 is the restriction of the norm in the tangent bundle of @xmath8 induced by the riemannian metric introduced in section [ preliminaries ] .
we need to prove an elementary fact about the norm @xmath216 in @xmath118 .
we have that @xmath220 and that @xmath221 , for some @xmath100 .
thus @xmath222 and @xmath223 thus , by the definition of @xmath216 , we have that @xmath224 where we use that @xmath66 acts in @xmath61 by @xmath40-isometries .
using proposition [ proplix ] and the definition of @xmath228 , we have that @xmath229 , for some @xmath230 and some @xmath231 .
writing @xmath232 , with @xmath233 , by the equations ( [ eqlambda ] ) and ( [ eqlix ] ) , we have that @xmath234 on the other hand , by lemma [ lemaspherical ] , we have that @xmath235 be the multiplicative jordan decomposition of the flow @xmath7 .
we have that @xmath82 acts on @xmath61 by @xmath40-isometries , that @xmath84 with @xmath86 a nilpotent operator and that @xmath90 commutes with @xmath236 .
thus we have that @xmath237 since @xmath238 .
hence @xmath239 where @xmath240 is also a nilpotent operator . by equations ( [ eq1 ] ) and ( [ eq2 ] ) , denoting @xmath241 ,
it is sufficient to show that @xmath242 since @xmath243 , we have that @xmath244 denote @xmath245 and @xmath246 where @xmath247 is such that @xmath248 . in order to get equation ( [ eq3 ] )
, we just need to prove that @xmath249 since @xmath250 using equation ( [ eq4 ] ) , we have that @xmath251 where @xmath252 and @xmath253 it is immediate that @xmath254 and thus equation ( [ eq6 ] ) implies the equation ( [ eq5 ] ) .
we denote @xmath256 and , for each @xmath257 , we define @xmath258 where @xmath259 is the map presented in theorem [ teolineariz ] .
it is immediate that the stable set of @xmath192 relative to the flow @xmath7 is given by the following disjoint union @xmath260 the following result determines explicitly the metric spectrum of @xmath7 relative to the morse components . under the above hypothesis
, we have that @xmath118 is invariant and @xmath115 is equivariant by the flow @xmath7 .
hence the corollary is a immediate consequence of propositions [ propmorseflag ] and [ proplyapunovmetric ] and theorems [ teolineariz ] and [ teospherical ] .
the previous result allows us the following definition .
the _ flag spectrum _ of an arbitrary linear induced flow @xmath7 is given by @xmath265 , its metric spectrum relative to the unique attractor component @xmath266 in the maximal flag manifold @xmath26 , where @xmath267 . | we introduce the _ metric spectrum _ , which measures the exponential rate of approximation to an isolated invariant set of points starting in its stable set , and relate it to the lyapunov spectrum .
we determine the metric spectrum of each morse component of the finest morse decomposition of a linear induced flow on a generalized flag manifold . _
ams 2000 subject classification _ : primary : 37b35 , 37d99 , secondary : 14m15 . _
key words : _ lyapunov exponents , stable sets , generalized flag manifolds . |
type ii supernovae ( sne ii ) are thought to arise from the explosion of massive ( @xmath12 ) stars which end their lives with a substantial fraction of hydrogen in their envelopes .
the explosion is triggered by the collapse of the progenitor s core .
the best - understood core - collapse events are the sne iip ( plateau ) , whose light curves are nearly flat during the first @xmath13 months . the long duration of the light curve plateau is believed to be due to the presence of a massive ( @xmath14 ) hydrogen envelope ( e.g. , falk @xmath15 arnett 1977 ) .
after the plateau , the light curves of sne iip are often dominated by radioactive decay of @xmath3co .
the cobalt is the decay product of @xmath3ni , and the luminosity at this epoch can therefore be used to estimate the amount of @xmath3ni ejected at the explosion .
because the nickel is produced in the core region , the mass of ejected @xmath3ni is an important probe of the core - collapse scenario itself , the nucleosynthetic yields of sne ii and their progenitors , and the nature of the compact central object formed at the collapse ( @xcite ; @xcite ) .
a new subclass of core - collapse supernovae has recently emerged : the narrow - line supernovae or sne iin ( schlegel 1990 ; filippenko 1997 ) . like other sne ii they show strong balmer lines , but the lines are narrow ( @xmath16 ) and are seen in emission rather than absorption even at early times .
sn 1994w was discovered on 1994 july 29 in the sbc galaxy ngc 4041 ( @xcite ) .
based on the presence of hydrogen lines it was quickly classified as a sn ii ( bragaglia , munari , & barbon 1994 ) .
subsequent spectroscopy showed that the lines had unusually narrow p - cygni profiles ( @xmath17 ) with broad emission wings , but no broad absorption component ( @xcite ; cumming , lundqvist , & meikle 1994 ) .
reports in iau circulars indicated that the supernova reached a maximum of @xmath18 around 1994 aug 13 .
further photometry by tsvetkov ( 1995 ) showed that the supernova was a sn iip , and had a remarkably sharp drop in the @xmath19 and @xmath2 light curves in early november 1994 .
in this paper we present further photometric and spectroscopic observations of sn 1994w taken on several occasions between 1994 july 31 to 1996 april 20 .
we focus on the low luminosity of the supernova after @xmath20 days , and interpret this as due to an exceptionally low mass of ejected @xmath3ni .
we also discuss the formation of the narrow lines .
our photometric data were obtained mainly in @xmath4 on six occasions from 1994 november 16 to 1996 april 20 , using the jakobus kapteyn telescope ( jkt ) and the nordic optical telescope ( not ) on la palma ( see table 1 ) .
the data were bias subtracted and flat fielded using standard procedures within iraf .
differential magnitudes between the supernova and comparison stars were measured using polyfit ( spnnare 1997 ) , which is a software package developed to do photometry on uneven backgrounds : a point spread function ( psf ) is produced by fitting a moffat function to an isolated star in the frame .
this psf and a polynomial fit to the background are then simultaneously fitted at the position of the object . for the jkt nights a bright comparison star was in the field and
was used for the relative magnitudes . for the not data a fainter comparison star had to be used . up
to seven comparison stars were monitored to avoid variability .
the comparison stars were calibrated against landolt stars ( landolt 1983 , 1992 ) at two nights on jkt and against a standard field in ngc 4147 ( christian et al .
1985 ) for the not observations .
the zero point for the @xmath4 magnitude differs slightly between the 4 photometric nights .
we estimate a total systematic error of 0.15 mag in the @xmath4 magnitude , including the neglect of color transformations and differential atmospheric extinction corrections , which were only applied to the day 646 data . for the @xmath2 magnitude we calibrated relative to local standards measured by tsvetkov ( 1995 ) .
the jkt observations have rather low signal to noise or poor seeing , which gives large errors for the photometry , @xmath210.5 mag . for the not observation ( day 197 )
the error is smaller , and dominated by the error in the polyfit measurement .
we estimated this error by adding to the image ten artificial stars of the same magnitude as the supernova at similar backgrounds using the iraf task addstar .
the rms error in the polyfit measurements of these were @xmath22 mag , which we regard as the measurement error also for the supernova .
a total error of 0.3 mag on day 197 is therefore rather generous and includes errors for the comparison stars . at the last observation ,
day 646 , the supernova was not detected .
we estimated an upper limit by subtracting artificial psfs of different brightness until a hole appeared in the background at the supernova position .
our spectroscopic observations cover five epochs from day 18 to day 202 ( see table 2 ) .
the data were bias subtracted , flat fielded and wavelength calibrated using iraf . the day 31 , 57 and 202
spectra were flux calibrated against a spectrophotometric standard star .
the day 57 spectrum was taken at high airmass and required further corrections for slit losses . a rough flux calibration for the day 18 spectrum was obtained by comparison with contemporary visual photometry and the spectrum on day 31 .
the day 121 spectrum was flux calibrated using the @xmath4 photometry from day 125 .
in figure 1 we present the light curve of sn 1994w . from the early @xmath4 data
, we estimate the date of optical outburst to have been 1994 july 14@xmath23 , and we label observations by number of days after this .
the light curve shows a relatively gentle rise to maximum around 30 days , followed by a slow , plateau - like decline until @xmath24 days .
the slow decline in @xmath19 and @xmath2 is typical of a sn iip . at @xmath24 days , however , the supernova started to fade unprecedentedly fast , dropping by @xmath1 magnitudes in @xmath2 in only 12 days . also on the light curve tail ,
the fading continued at a fast rate , until our last photometric detection on day 197 .
we could not detect the supernova in @xmath2 or @xmath4 on day 646 .
the spectrum of the supernova was also unusual for a sn ii . before the light curve drop , the spectra on days 18 , 31 and 57 ( see figs
. 2 and 3 ) , show the strong p - cygni lines of @xmath25 , typical of a sn ii , but with very narrow profiles ( minima at @xmath26 and maximum blue velocities of only @xmath27 ) .
in addition to the narrow features , the @xmath25 lines had emission wings extending out to @xmath28 , a factor of @xmath29 lower than the maximum velocities normally seen in sne ii at these epochs .
the @xmath25 lines were accompanied by a rich spectrum of narrow lines of low - ionization species ( he i , o i , mg ii , si ii , fe ii ) .
though conspicuous , the emission lines account for only @xmath30 of the total optical flux on days 31 and 57 .
instead , a blue continuum dominates . estimating @xmath31 from the interstellar na i d absorption in our day 31 spectrum ( @xcite ) , we fit it with blackbody spectra with temperatures of @xmath32 k and @xmath33 k , respectively .
the effective temperatures indicated by the blackbody fits are much higher than the @xmath34 k usually observed in sne iip at similar epochs .
the decrease in temperature was accompanied by the disappearance of the @xmath35 lines and increasingly strong @xmath36 features .
a spectrum taken by filippenko ( 1997 ) on day 80 shows that the spectral features had changed only marginally compared to our day 57 spectrum , but the slope of the continuum indicated a decrease in effective temperature to @xmath37 k. the photometric fading at @xmath38110 days coincided with dramatic changes in the spectrum .
the @xmath39 color , previously steady at @xmath40 , increased to 1.75 between days 85 and 111 ( fig . 1 ; @xcite ) .
in addition , the flux in @xmath41 dropped at roughly the same rate as the @xmath2 and @xmath4 light curves .
this coincidence , together with the shift to lower - excitation lines during the plateau phase , suggests that the narrow lines were predominantly excited by the supernova continuum rather than by a circumstellar shock ( see 4.4 ) .
the day 121 spectrum shows narrow ( @xmath42 ) emission in @xmath41 , @xmath43 d @xmath445893 and [ ca ii ] @xmath457291 , 7324 , on top of a continuum - like spectrum .
the flux of the narrow @xmath41 component had dropped by a factor @xmath46 since day 57 .
no broad features can be identified , but the sharp edge present at @xmath47 is a common feature in late - time supernova spectra , usually attributed to the redmost extent of blended fe ii or [ fe ii ] lines .
we suggest that the apparent continuum is in fact made up of overlapping broad emission lines , with a gap around 5650 .
unlike other type ii events at this epoch , there is no sign of strong , broad @xmath41 . on day 202 ,
the spectrum was dominated by a fairly flat , noisy continuum and unresolved @xmath41 , which may contain a contribution from a background @xmath48 region .
the flux in the narrow @xmath41 line had dropped by a factor of @xmath49 compared to the day 121 spectrum , but its width remained constant within the errors .
the flux in the @xmath4 band of our spectrum on day 202 agrees well with our @xmath4 magnitude on day 197 .
our photometric limit from day 646 ( @xmath50 ) demonstrates that the supernova continued to fade after day 202 , and gives a lower limit on the flux contribution of the supernova to the day 202 spectrum of @xmath51 .
we suggest that the continuum at this epoch comes from background stars .
we tentatively identify a broad , @xmath52 , emission feature at @xmath53 as [ ca ii ] @xmath457291 , 7324 , and suggest that the remaining emission left after the continuum is subtracted is , as on day 121 , composed of broad emission lines .
the [ ca ii ] feature has luminosity a factor of @xmath38100 lower than for sn 1987a at the same epoch ; the luminosity of @xmath41 is at least a factor of 2 lower still ( fig .
in figure 5 we compare the absolute light curve of sn 1994w to those of other sne iip , sn 1987a and the luminous sn iil ( linear ) 1979c . at maximum ,
sn 1994w was as luminous as sn 1983k , the most luminous sn iip previously observed ( phillips et al .
evident from figure 5 is the rapid fading and low luminosity of sn 1994w after the plateau phase ends , i.e. , @xmath24 days , and the fast decline after that .
we now investigate whether the unusual light curve and spectra of sn 1994w can be explained in terms of standard supernova theory . for a sn iip , the mass of ejected @xmath3ni , @xmath54 ,
can be estimated from the luminosity after the plateau phase ends . at this point
the optical emission is powered by @xmath55ray energy deposited into the ejecta from the radioactive decay of @xmath3co , the decay product of @xmath3ni .
the e - folding time for the cobalt decay is 111.26 days , so the radioactive decay creates a tail to the bolometric light curve with a slope of 0.00976 mag day@xmath56 , if all the @xmath55rays are trapped .
the luminosity of the tail gives a measure of @xmath54 , when other contributions to the light curve have been accounted for .
the mass of ejected nickel has been determined in this way for a handful of sne iip , for example sne 1969l ( @xmath57 , @xcite ) , 1990e ( @xmath58 , @xcite ) , 1992h ( @xmath59 , @xcite ) and sn 1991 g ( @xmath60 , @xcite ) .
our method to estimate @xmath54 of sn 1994w is to compare the level of its luminosity in @xmath4 after the plateau to that of sn 1987a at similar epochs .
we choose sn 1987a as a template since its distance and extinction are well known , and because its light curve was entirely powered by radioactive decay as soon as a few days after explosion .
sn 1987a also resembled a sn iip in that it had a massive hydrogen envelope ( @xcite ) .
for sn 1987a the value of @xmath54 was @xmath61 ( @xcite ) .
a problem with our method is that the light curve tail of sn 1994w declines substantially faster than the decay rate of @xmath3co
. we must therefore be careful in spelling out the assumptions involved in interpreting the light curve .
we identify two possible scenarios which can lead to such a decline , and which lead to different estimates of @xmath54 .
radioactive decay of @xmath3co could dominate , but be unable for some reason to maintain the standard tail of the @xmath4 light curve .
alternatively , some source other than @xmath3co decay dominates the light curve during this period . in the following sections we investigate both these possibilities and their implications for @xmath54 .
first we consider the possibility that all of the flux on the tail was indeed due to cobalt decay .
the steep decline could then be due to increasing escape of @xmath62-rays , a gradual shift of the spectrum from the optical to the infrared , or increasing dust obscuration . assuming a central radioactive source , the optical depth to @xmath62-rays through the ejecta decreases with time as @xmath63 , where @xmath64 is the time when @xmath65 . in sn 1987a , @xmath66 days ( kozma 1996 ) .
including the fact that @xmath1% of the cobalt decay energy is mediated by positrons and therefore deposited locally , the bolometric luminosity decays by the factor @xmath67 in addition to the decay of @xmath3co .
the observed flux in @xmath4 between days 125 and 197 drops by a factor @xmath68 in addition to cobalt decay . with the assumptions above the maximum value of @xmath69 for the same period is only @xmath70 ( for @xmath71 days ) and thus can not reproduce the fast decline of the light curve .
we therefore believe that @xmath62-ray leakage was not the cause of the exceptionally fast fading of sn 1994w after @xmath20 days . a gradual shift of the spectrum into the infrared might be expected if the mass of @xmath3ni was substantially lower than in sn 1987a , and the reduced radioactive heating lowered the ejecta temperature . in this scenario
our assumption of linear scaling between @xmath54 and the @xmath4-luminosity could break down . to test this numerically , we ran two models of sn 1987a kindly provided for us by c. kozma
. these models do not include dust formation and are similar to her mixed 10h model in kozma ( 1996 ) , but with only @xmath72 @xmath73 , and with other radioactive isotopes decreased accordingly .
we find that on day 150 , more than 60 ( 70)% of the line emission comes out between @xmath74 , @xmath75 and @xmath76 of which in h@xmath77 and [ ca ii ] , respectively .
therefore , if dust formation is unimportant , and the structure similar to that of sn 1987a , linear scaling of @xmath54 with @xmath4 and @xmath2 luminosities seems to be valid down to very low values of @xmath54 .
the situation is different if dust formation is important .
we expect that below some value of @xmath54 , radioactive heating of the ejecta will be unable to prevent the formation of dust .
this could attenuate and shift the supernova emission into the infrared already shortly after the plateau phase .
the effect could be further amplified by increased molecular cooling , since the destruction rate of molecules will decrease with lower @xmath54 .
we know that dust increased the cooling and obscured the line - emitting region in sn 1987a ( kozma 1996 , and references therein ) , albeit at a later epoch ( between @xmath78 days ) than considered here for sn 1994w . the dust covering factor would have increased with time , creating a blueshift of the supernova lines , as observed for sn 1987a ( danziger et al .
our day 202 spectrum is too noisy to reveal such an effect .
thus , we can not rule out dust formation as the cause of the fast fading in @xmath4 . to obtain an estimate of @xmath54 in the case of early dust formation
, we choose the @xmath4 measurement from day 139 , which is the first photometric point on the light curve tail .
this gives @xmath79 .
the estimated error includes the statistical error shown in figure 1 , uncertainty in the distance modulus of sn 1994w ( @xmath800.3 mag ) , the extinction towards the supernova ( @xmath800.15 mag ) and an estimated systematic error of our @xmath4 calibration ( @xmath800.15 mag ) .
this @xmath54 is slightly lower than for the previous record - holder , sn 1991 g , which had @xmath60 .
however , sn 1991 g did not show the very steep decline of the light curve tail ( @xcite ) .
this could mean that even sn 1991 g ejected too much @xmath3ni to allow early dust formation , though the structure of sn 1991 g could also have been less favorable for dust formation to occur than in sn 1994w .
if the steep tail is not primarily powered by radioactive decay we need an additional , fast declining , energy source .
a light echo seems rather improbable as it would produce a spectrum on the tail that is qualitatively similar to the spectrum at the peak .
this is not observed .
a more attractive power source is interaction of the ejecta with circumstellar material at @xmath81 days . in 4.4 we present evidence from the narrow line profiles for a thin , dense circumstellar shell close to the supernova .
the high density we find for the shell may be sufficient to set up the isothermal - shock - wave scenario discussed by grasberg & nadyozhin ( 1986 ) , in which the ejecta s kinetic energy is transformed to uv and optical emission . as the shell
is accelerated , the conversion of kinetic energy becomes less efficient , which could create the steeply - falling light curve . in this case the photometric point on day 197 gives a firm upper limit to @xmath54 . assuming no dust , the @xmath62-rays to be trapped and a linear scaling between @xmath4-luminosity and @xmath54 this gives @xmath82 .
this limit could of course be higher if dust is important also in this scenario . in summary , we suggest that the low luminosity on the light curve tail is due to a very low mass of ejected nickel
. we will now examine whether such a small mass of @xmath3ni can be consistent with the type of progenitor indicated by the light curve up to @xmath24 days .
the very low nickel mass we find , if radioactivity is not the prime power source between days 120 and 197 ( @xmath82 ) , agrees with model calculations only for progenitors in the ranges @xmath7 ( @xcite ) and @xmath8 ( @xcite ) .
if the steep decline of the light curve tail was due instead to early dust formation , our more conservative estimate ( @xmath79 ) is still lower than the range @xmath83 , predicted for progenitors in the range @xmath84 ( @xcite ) , but somewhat higher than @xmath54 predicted for stars with @xmath7 ( @xcite ) .
theory thus favors two kinds of progenitor that eject low masses of @xmath3ni : a low mass progenitor , that produces almost no nickel during the explosion , or a very massive star where the mass - cut ( the location in mass coordinates between ejected matter and that eventually forming the compact , central object ) is situated sufficiently far out for most of the synthesized nickel to be trapped in the compact object . due to the extensive fall - back of material in the latter case
, the mass of the compact object may become large enough to form a black hole ( @xcite ) . as a matter of fact , woosley & timmes ( 1996 ) suggested that the prime observational diagnostic of a supernova forming a black hole would be a bright supernova that plummeted to very low or zero luminosity right after the plateau . to estimate the mass of the progenitor we use the early light curve ( i.e. , at @xmath85 days ) of the supernova . at this epoch
the supernova is powered by the release of stored shock energy from the explosion .
the physics of the plateau phase has been parameterized analytically ( popov 1993 ) and fitted to numerical simulations ( litvinova & nadyozhin 1985 ) in terms of the explosion energy , @xmath86 , the stellar radius , @xmath87 , and the mass of the hydrogen envelope at the time of explosion , @xmath88 .
assuming a canonical value for @xmath86 in the range @xmath89 ergs , we use parameterized expressions to estimate @xmath87 and @xmath88 for given @xmath90 ( the epoch when the plateau ends ) and @xmath2 ( the absolute visual magnitude before turn - down ) . taking @xmath91 days and @xmath92 ,
popov s expressions give @xmath93 and @xmath94 cm , while the numerical models of litvinova & nadyozhin give @xmath95 and @xmath96 cm .
@xmath88 decreases with increasing @xmath87 for given @xmath2 and @xmath86 .
we thus derive a value of @xmath88 which is similar to , or in the case of the numerical models , slightly higher than that of sn 1987a ( @xcite ; @xcite ) .
such high values of @xmath88 strongly favor the higher mass range ( @xmath8 ) for the progenitor of sn 1994w .
to @xmath88 we need to add the mass of the helium- and metal - rich core at the time of explosion . for a progenitor with @xmath97 ,
this is @xmath98 ( @xcite ) . to be consistent with our estimate of @xmath88 for a @xmath97 progenitor
, we must therefore allow for @xmath99 of stellar winds if popov s expressions are used for the plateau , while the numerical models do not require presupernova mass loss .
a low - mass ( @xmath7 ) progenitor is possible for sn 1994w , but only if the explosion energy was lower than the canonical @xmath89 ergs assumed above and the conversion of explosion energy to radiation was very efficient .
model b of falk & arnett ( 1977 ) shows that a high ( @xmath100 ) efficiency is possible if the star is very extended ( @xmath101 cm ) . assuming an explosion energy as low as @xmath102 ergs
, popov s model can give an envelope mass of only @xmath103 for a radius of @xmath104 cm , while the model of litvinova & nadyozhin gives @xmath105 for a radius of @xmath106 cm . both the analytical and numerical estimates are thus consistent with a low mass ( @xmath7 ) progenitor , if the progenitor is very extended and the explosion energy is low . because a @xmath7 supernova only ejects @xmath107 ( @xcite ) ,
this type of explosion is consistent only with our lowest @xmath54 ( see 4.2 ) .
if sn 1994w resulted from the explosion of an @xmath108 star , its compact object should be a neutron star .
if its progenitor was instead in the higher range , the mass of the newly - created compact object could become large enough to form a black hole ( @xcite ) . for a @xmath109 progenitor
, the mass - cut should be @xmath110 ( timmes , woosley & weaver 1996 ) .
the maximum mass of a neutron star is rather uncertain , depending on the adopted equation of state . if it is as low as @xmath111 ( @xcite ) , such a massive progenitor for sn 1994w most likely formed a black hole .
the narrow lines of sn 1994w are not typical for sne iip . in
3 we argued that the flux in the lines was too low to affect the light curve on the plateau , which is why we were able to apply the light curve analysis in 4.3 . here
we discuss the origin of the narrow lines . as already suggested in 3 , the lines
are naturally explained in terms of excitation by the hot supernova continuum .
another possibility could be excitation caused by circumstellar interaction of the ejecta with a clumpy wind , as suggested for sn 1988z by chugai & danziger ( 1994 ) . however , we see no obvious reason why this excitation mechanism should result in an excitation of the lines that closely follows the fading of the supernova continuum as in sn 1994w .
the narrow emission lines were present as early as day @xmath112 ( @xcite , @xcite ) .
h@xmath77 had a roughly constant p - cygni profile from then until at least day 57 , the epoch of our last spectrum on the plateau ( see fig .
the emission part of the p - cygni profile has a rounded rather than flat - topped peak , which immediately indicates that the optical depth in the lines @xmath113
( @xcite ) .
furthermore , the steep red side and relatively broad absorption trough observed in both @xmath25 and @xmath36 indicate that the lines were scattering - dominated , and originated in a rather thin shell ( @xmath114 ) located close to the photosphere ( see fransson 1984 ) .
the optical thickness of the fe ii lines allows us to put a limit on the density in the shell . assuming a gas temperature of @xmath115 k and fe ii / h = fe / h = @xmath116 ( i.e. , the solar value of fe / h )
, we use expressions from mihalas ( 1978 ) to obtain a hydrogen density @xmath117 . here , @xmath118 is the velocity of the shell in units of @xmath119 and @xmath120 the shell radius in units of @xmath121 cm .
we know from the line profiles that @xmath122 , and that the shell radius must be close to that of the photosphere . from our blackbody fits ( see 3 ) we estimate that the photospheric radius on days 31 and 57 was @xmath123 cm and @xmath124 cm , respectively .
this gives a limit on @xmath125 , quite consistent with the absence of forbidden lines in the spectrum during the plateau .
the near - constancy of the shell s velocity up to day 57 and its proximity to the photosphere allow us to locate the starting radius of the shell , @xmath126 , assuming it is the same shell that emits the lines at all epochs . using @xmath127
, we obtain @xmath128 cm from both the day 31 and day 57 data . because the radius of the photosphere is rather uncertain ( @xmath129 ) , so is @xmath126 .
the shell appears to have been quite close to the supernova around the time of explosion .
we identify two possible origins .
one possibility is that the shell was created in the supernova envelope during the explosion .
the formation of such a shell was suggested by falk & arnett ( 1977 ) , who actually also predicted that this should result in narrow p - cygni profiles .
the problem with this interpretation is the low velocity of the shell . even in their low - explosion - energy ,
extended model b , falk & arnett find a shell velocity roughly twice as high as we find for sn 1994w .
it seems more plausible that the shell was circumstellar , perhaps due to an episode of increased mass loss some time @xmath130 prior to the explosion , where @xmath131 is the initial velocity of the shell .
this scenario is similar to the model presented by grasberg & nadyozhin ( 1986 ) for sn 1983k .
we find that the radiation from the supernova can easily accelerate the shell to the observed velocity .
assuming pure electron scattering , a shell initially at rest is accelerated up to @xmath132 , where @xmath133 is the integrated bolometric luminosity of the supernova in units of @xmath134 ergs and @xmath120 the shell radius in units of @xmath121 cm .
the velocity can be substantially higher if line acceleration is important .
we note that the blue edge of the h@xmath77 absorption shifted slightly to the blue between days 18 and 31 , corresponding to an increase in shell velocity from @xmath135 to @xmath136 ( fig .
this is consistent with the supernova radiation continuing to accelerate the shell between these epochs . in the model of grasberg & nadyozhin ( 1986 ) , the shell
is eventually accelerated to the velocity of the supernova ejecta due to circumstellar interaction . however , in reality this may take longer than in grasberg & nadyozhin s 1-d models because the slow shell might break up ( @xcite ) .
in sn 1994w , the disruption of the proposed shell by the ejecta could perhaps power the light curve between @xmath20 and 197 days , as we suggested in 4.2 .
we have observed the luminous sn 1994w .
we find that the low flux of the supernova after @xmath20 days is evidence of very low mass of @xmath3ni ejected at the explosion , in fact the lowest inferred for any sn ii .
the estimated mass of @xmath3ni depends on the interpretation of the unusually rapid fading observed in @xmath4 between days 120 and 197 .
either the supernova ejected @xmath6 of @xmath3ni , or the nickel mass could have been as low as @xmath5 .
the first possibility requires inefficiency in converting radioactive decay energy to @xmath4 luminosity , and we find that early dust formation is the most likely cause for this .
alternatively , the primary power source between @xmath137 days was not radioactive decay , but perhaps interaction with a thin circumstellar shell .
in this case @xmath138 is an upper limit , if dust formation was unimportant .
two types of progenitor seem consistent with the low nickel content and high luminosity of the supernova before the end of its plateau phase .
the first possibility is a low mass progenitor with @xmath139 , which had a very extended envelope and a low explosion energy , @xmath140 ergs . as a low mass progenitor only produces @xmath141 @xmath3ni ( @xcite ) , consistent only with our lowest value for @xmath54 , an additional power source , presumably circumstellar interaction ,
is needed to explain the light curve between @xmath137 days for this type of progenitor .
however , the energy budget seems uncomfortably tight for the low - mass case as sn 1994w was one of the most luminous supernovae observed .
a more natural possibility is therefore a high mass progenitor with @xmath8 , which could have lost a few solar masses as stellar winds , and exploded with the canonical energy @xmath89 ergs .
this type of progenitor is consistent with both our estimates of @xmath54 .
such a supernova should have a mass - cut @xmath110 , sufficient to trap most of the produced @xmath3ni , and could have formed a black hole .
we thank p. rudd , r. rutten , m. breare and e. harlaftis for taking service observations , e. zuiderwijk for giving up patt time to take the day 18 spectrum , and m. nslund for taking part in our last observation of the supernova .
we are especially grateful to c. kozma , b. leibundgut and s. blinnikov for important discussions , and to l. szentasko and m. villi for access to unpublished data .
we also thank m. prez - torres for early analysis of the narrow lines and s. spnnare for help with photometry measurements .
this research was supported by the swedish natural science research council .
and pl also acknowledge support from the swedish national space board .
941116.2 & 125 & jkt / tek4 & v & [email protected] & p. rudd & & & r & [email protected] & p. rudd941130.3 & 139 & jkt / tek4 & r & [email protected] & v. dhillon941222.3 & 161 & jkt / tek4 & r & [email protected] & e. harlaftis950120.3 & 190 & jkt / tek4 & r & [email protected] & e. harlaftis950127.3 & 197 & not / stancam1 & r & [email protected] & j. sollerman , r. cumming960420.0 & 646 & not / brocam2 & r & @xmath142 & j. sollerman , m. nslund 940731.9 & 18 & int / ids & 6507 - 6675 & 0.25 & e. zuiderwijk 940813.9 & 31 & wht / isis & 4290 - 8005 & 2.7 & m. breare 940908.9 & 57 & wht / isis & 3300 - 9360 & 11 & r.
rutten 940908.9 & 57 & wht / isis & 4255 - 5060,6345 - 7150 & 2.7 & r. rutten 941112.3 & 121 & int / ids & 5485 - 7370 & 6 & r. rutten 950201.5 & 202 & not / lds & 4800 - 10000 & 12 & r. cumming , j. sollerman | we present spectroscopic and photometric observations of the luminous narrow - line type iip ( plateau ) supernova 1994w . after the plateau phase ( @xmath0 days ) , the light curve dropped by @xmath1 mag in @xmath2 in only 12 days . between 125 and 197 days
after explosion the supernova faded substantially faster than the decay rate of @xmath3co , and by day 197 it was 3.6 magnitudes less luminous in @xmath4 compared to sn 1987a .
the low @xmath4-luminosity could indicate @xmath5 of @xmath3ni ejected at the explosion , but the emission between 125 and 197 days must then have been dominated by an additional power source , presumably circumstellar interaction .
alternatively , the late light curve was dominated by @xmath3co decay . in this case , the mass of the ejected @xmath3ni was @xmath6 , and the rapid fading between 125 and 197 days was most likely due to dust formation . though this value of the mass is higher than in the case with the additional power source , it is still lower than estimated for any previous type ii supernova . only progenitors with @xmath7 and
@xmath8 are expected to eject such low masses of @xmath3ni . if @xmath7 , the plateau phase indicates a low explosion energy , while for a progenitor with @xmath8 the energy can be the canonical @xmath9 ergs .
as sn 1994w was unusually luminous , the low - mass explosion may require an uncomfortably high efficiency in converting explosion energy into radiation .
this favors a @xmath8 progenitor . the supernova s narrow ( @xmath10 ) emission lines were excited by the hot supernova spectrum , rather than a circumstellar shock
. the thin shell from which the lines originated was most likely accelerated by the radiation from the supernova .
-0.5pc -0.5pc # 1| # 1 | # 110^#1 3 cm^-3 # 1#2 # 1e^-t/#1^d # 1e^-t/#1yr 56m_ej,56 # 120true pt3true cm#1
@xmath11 # 1 # 1d |
the experimental evidence for discontinuities in the sequence of atomic masses , @xmath1 and @xmath2- decay systematics and binding energies of nuclei suggests the existence of a set of magic proton and neutron numbers , which can be described successfully by single particle shell models with a heuristic spin - orbit term @xcite , @xcite .
the most prominent representative is the phenomenological nilsson model @xcite with an anisotropic oscillator potential : @xmath3 although these models are flexible enough to reproduce the experimental results , they lack a deeper theoretical justification , which becomes obvious , when extrapolating the parameters @xmath4 , @xmath5 , which determine the strength of the spin orbit and @xmath6 term to the region of superheavy elements @xcite .
hence it seems tempting to describe the experimental data with alternative methods .
typical examples are relativistic mean field theories @xcite,@xcite , where nucleons are described by the dirac - equation and the interaction is mediated by mesons . although a spin orbit force is obsolete in these models , different parametrizations predict different shell closures @xcite,@xcite .
therefore the problem of a theoretical foundation of magic numbers remains an open question since elsasser @xcite raised the problem 75 years ago .
a fundamental understanding of magic numbers for protons and neutrons may be achieved if the underlying corresponding symmetry of the nuclear many body system is determined .
therefore a group theoretical approach seems appropriate .
group theoretical methods have been successfully applied to problems in nuclear physics for decades .
elliott @xcite has demonstrated , that an average nuclear potential given by a three dimensional harmonic oscillator corresponds to a su(3 ) symmetry .
low lying collective states have been successfully described within the ibm - model @xcite , which contains as one limit the five dimensional harmonic oscillator , which is directly related to the bohr - mottelson hamiltonian . in this paper
we will determine the symmetry group , which generates a single particle spectrum similar to ( [ nilsson ] ) , but includes the magic numbers right from the beginning .
our approach is based on group theoretical methods developed within the framework of fractional calculus .
the fractional calculus @xcite-@xcite provides a set of axioms and methods to extend the coordinate and corresponding derivative definitions in a reasonable way from integer order n to arbitrary order @xmath7 : @xmath8 the definition of the fractional order derivative is not unique , several definitions e.g. the feller , fourier , riemann , caputo , weyl , riesz , grnwald fractional derivative definitions coexist @xcite-@xcite .
a direct consequence of this diversity is the fact , that the solutions e.g. of a one dimensional wave equation differ significantly depending on the specific choice of a fractional derivative definition .
until now it has always been assumed , that the fractional derivative type for an extension of a fractional differential equation to multi - dimensional space should be chosen uniquely .
in contrast to this assumption , in this paper we will investigate properties of higher dimensional rotation groups with mixed caputo and riemann type definition of the fractional derivative .
we will demonstrate , that a fundamental dynamic symmetry is established , which determines the magic numbers for protons and neutron respectively and furthermore describes the ground state properties of nuclei with reasonable accuracy .
we will investigate the spectrum of multi dimensional fractional rotation groups for two different definitions of the fractional derivative , namely the riemann- and caputo fractional derivative .
both types are strongly related . starting with the definition of the fractional riemann integral @xmath9 where @xmath10 denotes the euler @xmath11-function , the fractional riemann derivative is defined as the product of a fractional integration followed by an ordinary differentiation : @xmath12 it is explicitely given by : @xmath13 the caputo definition of a fractional derivative follows an inverted sequence of operations ( [ dr ] ) .
an ordinary differentiation is followed by a fractional integration @xmath14 which results in : @xmath15 applied to a function set @xmath16 using the riemann fractional derivative definition ( [ driemann ] ) we obtain : @xmath17 \ , x^{(n-1)\alpha}\end{aligned}\ ] ] where we have introduced the abbreviation @xmath18 $ ]
. for the caputo definition of the fractional derivative it follows for the same function set : @xmath19 \ , x^{(n-1)\alpha}\end{aligned}\ ] ] where we have introduced the abbreviation @xmath20 $ ] .
both derivative definitions only differ in the case @xmath21 : @xmath22 & = & { _ \textrm{\tiny{r}}}[n ] -\delta_{n0}\ , { _ \textrm{\tiny{r}}}[0 ] \\ & = & { _ \textrm{\tiny{r}}}[n ] - \delta_{n0}\ , \frac{1}{\gamma(1-\alpha)}\end{aligned}\ ] ] where @xmath23 denotes the kronecker-@xmath24 .
we will rewrite equations ( [ rx ] ) and ( [ cx ] ) simultaneously , introducing the short hand notation @xmath25 \ , x^{(n-1)\alpha}\\\ ] ] we now introduce the fractional angular momentum operators or generators of infinitesimal rotations in the @xmath26 plane on the @xmath27-dimensional euclidean space : @xmath28 the commutation relations of the fractional angular momentum operators are isomorph to the fractional extension of the rotational group @xmath29 @xmath30 & = & { _ \textrm{\tiny{r , c}}}\ , i\hbar { f_{ijkl}}^{mn}l_{mn}(\alpha ) \\ & & \qquad\qquad\qquad i , j , k , l , m , n=1,2, .. ,n \nonumber\end{aligned}\ ] ] with structure coefficients @xmath31 .
their explicit form depends on the function set the fractional angular momentum operators act on and on the fractional derivative type used .
the casimir - operators of the fractional rotation group @xmath32 based on the riemann fractional derivative definition have been derived in @xcite and for @xmath33 based on the caputo fractional derivative definition are given in @xcite .
we summarize the major results : spectrum of the casimir operator @xmath34 from ( [ l2 ] ) as a function of the fractional derivative coefficient @xmath7 .
only the @xmath35 state differs for riemann and caputo derivative .
, title="fig:",width=302,height=230 ] + according to the group chain @xmath36 there are two casimir - operators @xmath37 , namely @xmath38 and @xmath39 .
we introduce the two quantum numbers @xmath40 and @xmath41 , which completely determine the eigenfunctions @xmath42 .
it follows @xmath43\ , |lm > \\ & & \qquad\qquad\qquad m = -l ,- l+1, ... ,\pm 0, ...
,l \nonumber \\
\label{l2 } { _ \textrm{\tiny{r , c } } } \ , l^2(\alpha ) |lm > & = & { _ \textrm{\tiny{r , c } } } \ , \hbar^2 [ l][l+1 ] \,|lm > \\ & & \qquad\qquad\qquad
l = 0,1,2 , ... \nonumber\end{aligned}\ ] ] where @xmath44 denotes the absolute value of @xmath41 .
in addition , on the set of eigenfunctions @xmath42 , the parity operator @xmath45 is diagonal and has the eigenvalues @xmath46 in figure [ so3 ] the eigenvalues of the casimir - operator @xmath47 are shown as a function of @xmath7 . only in the case
@xmath35 the spectra differ for the riemann- and caputo derivative .
while for the caputo derivative @xmath48 because @xmath49 = 0 $ ] , using the riemann derivative for @xmath50 there is a nonvanishing contribution @xmath51[1 ] |00 > = \hbar^2 \frac{\gamma(1+\alpha)}{\gamma(1-\alpha)}|00>\ ] ]
we now use group theoretical methods to construct higher dimensional representations of the fractional rotation groups @xmath52 . as an example of physical relevance
we introduce the group @xmath53 with the following chain of subalgebras : @xmath54 the hamiltonian @xmath55 can now be written in terms of the casimir operators of the algebras appearing in the chain and can be analytically diagonalized in the corresponding basis .
the hamiltonian is : @xmath56 with the free parameters @xmath57 and the basis is @xmath58 .
furthermore , we impose the following symmetries : first , the wave functions should be invariant under parity transformations , which according to ( [ parity ] ) leads to the conditions @xmath59 second , we require @xmath60|l_1 m_1 l_2 m_2 l_3 m_3 > \\ { _ \textrm{\tiny{r}}}l_{z_2}(\alpha)|l_1 m_1 l_2 m_2 l_3 m_3 > & = & { _ \textrm{\tiny{r}}}+\hbar[l_2]|l_1 m_1 l_2 m_2 l_3 m_3 > \\ { _ \textrm{\tiny{r}}}l_{z_3}(\alpha)|l_1 m_1 l_2 m_2 l_3 m_3 > & = & { _ \textrm{\tiny{r}}}+\hbar[l_3]|l_1 m_1 l_2 m_2 l_3 m_3 > \end{aligned}\ ] ] which reduces the multiplicity of a given @xmath61 set to 1
. with these conditions , the eigenvalues of the hamiltonian ( [ hamilton ] ) are given as @xmath62[2n_1 + 1]+ \nonumber \\ & & \hbar \omega_2 { _ \textrm{\tiny{r } } } [ 2n_2][2n_2 + 1]+ \hbar \omega_3 { _ \textrm{\tiny{r } } } [ 2n_3][2n_3 + 1 ] \\
\label{e2 } & = & \sum_{i=1}^3 \hbar \omega_i \frac{\gamma(1 + ( 2 n_i + 1 ) \alpha)}{\gamma(1 + ( 2 n_i-1 ) \alpha ) } - \delta_{n_1 0 } \hbar \omega_1 \frac{\gamma(1 + \alpha)}{\gamma(1 - \alpha)}\\ & & \qquad\qquad\qquad\qquad\qquad n_1,n_2,n_3=0,1,2 , .. \nonumber\end{aligned}\ ] ] on a basis @xmath63 .
this is the major result of our derivation .
we call this model the caputo - riemann - riemann symmetric rotor .
what makes this model remarkable is its behaviour near @xmath64 . on the left the energy spectrum @xmath65 from ( [ e2 ] ) for the spherical case ( [ sphere ] ) in units of @xmath66 for the caputo - riemann - riemann symmetric rotor
is presented .
the right diagram shows the neutron energy levels for the spherical nucleus @xmath67 calculated within the framework of the asymmetric two center shell model ( atcsm ) @xcite near the ground state as a function of increasing strength of the spin - orbit term ( @xmath68 ) increasing from @xmath69 to @xmath70 of the recommended @xmath4 value , while the @xmath71 value is kept constant .
the transition from magic numbers of the standard 3-dimensional harmonic oscillator levels ( [ set1 ] ) to the shifted set of magic numbers ( [ set2 ] ) is pointed out with thick lines . left and right figure
therefore show a similar behaviour for the energy levels .
, title="fig:",width=453,height=309 ] + on the left of figure [ fig2 ] we have plotted the energy levels in the vicinity of @xmath72 for the case @xmath73 which we denote as the spherical case . for the idealized case @xmath64 , using the relation @xmath74 the level spectrum ( [ e2 ] ) is simply given by : @xmath75 for @xmath76 this is the well known spectrum of the 3-dimensional harmonic oscillator .
assuming a twofold spin degeneracy of the energy levels , we introduce the quantum number @xmath27 as @xmath77 consequently we obtain a first set @xmath78 of magic numbers @xmath79 @xmath80 which correspond to the standard 3-dimensional harmonic oscillator at energies @xmath81 in addition , for @xmath82 , which corresponds to the @xmath83 states , we obtain a second set @xmath84 of magic numbers @xmath85 at energies @xmath86 which is shifted by the amount @xmath87 compared to the standard 3-dimensional harmonic oscillator values . from figure [ fig2 ]
it follows , that for @xmath88 the second set @xmath89 of energy levels falls off more rapidly than the levels of set @xmath90 . as a consequence for decreasing @xmath7
the magic numbers @xmath90 die out successively . on the other hand , for @xmath91
the same effect causes the magic numbers @xmath90 to survive .
we want to emphasize , that the described behaviour for the energy levels in the region @xmath88 may be directly compared to the influence of a @xmath92-term in phenomenological shell models . as an example , on the right hand side of figure [ fig2 ] a sequence of neutron levels for the superheavy element @xmath67 calculated with the asymmetric two center shell model ( atcsm ) @xcite with increasing strength of the @xmath92-term from @xmath69 to @xmath70 is plotted .
it shows , that the @xmath93 gap breaks down at about @xmath94 and the @xmath95 gap at about @xmath96 of the recommended @xmath4-value for the @xmath92-term .
this corresponds to an @xmath97 value , since in the caputo - riemann - riemann symmetric rotor the @xmath93 gap breaks down at @xmath98 , the @xmath95 gap at @xmath99 and the @xmath100 gap vanishes at @xmath101 .
we conclude , that the caputo - riemann - riemann symmetric rotor predicts a well defined set of magic numbers .
this set is a direct consequence of the underlying dynamic symmetries of the three fractional rotation groups involved .
it is indeed remarkable , that the same set of magic numbers is realized in nature as magic proton and neutron numbers . in the next section
we will demonstrate , that the proposed analytical model is an appropriate tool to describe the ground state properties of nuclei .
we will use the caputo - riemann - riemann symmetric rotor ( [ e2 ] ) as a dynamic shell model for a description of the microscopic part of the total energy @xmath102 of the nucleus .
@xmath103 where @xmath104 and @xmath105 denote the shell- and pairing energy contributions . for the macroscopic contribution we use the finite range liquid drop model ( frldm ) proposed by mller @xcite using the original parameters , except the value for the constant energy contribution @xmath106 .
as the primary deformation parameter we use the ellipsoidal deformation @xmath107 : @xmath108 where @xmath109 are the semi - axes of a rotational symmetric ellipsoid .
consequently a value @xmath110 describes prolate and a value of @xmath111 descibes oblate shapes . in order to relate the ellipsiodal deformation @xmath107 to the quadrupole deformation @xmath112 used by mller ,
we define : @xmath113 furthermore we extend the original frldm - model introducing an additional curvature energy term @xmath114 , which describes the interaction of the nucleus with the collective curved coordinate space @xcite : @xmath115 where @xmath116 is the nucleon number , @xmath117 is the curvature parameter given in @xmath118 $ ] and the relative curvature energy @xmath119 given as : @xmath120 which is normalized relative to a sphere @xmath121 .
therefore the total energy may be splitted into @xmath122 where @xmath123 with two free parameters @xmath124 , which will be used for a least square fit with the experimental data . for calculation of the shell corrections
we use the strutinsky method @xcite,@xcite .
since we expect that the shell corrections are the dominant contribution to the microscopic energy , for a first comparison with experimental data we will neglect the pairing energy term . as a test of the plateau condition @xmath125 for the strutinsky shell correction method , the total shell correction energy @xmath126 is plotted as a function of @xmath127 for different nuclei . , title="fig:",width=453,height=154 ] + comparison of calculated shell corrections @xmath104 from the caputo - riemann - riemann symmetric rotor ( [ e2 ] ) with the parameter set ( [ om0])-([m ] ) ( thick line ) with the tabulated @xmath128 from mller @xcite ( dashed line ) .
upper two rows show values for a given @xmath129 as a function of @xmath27 , lower two rows for a given @xmath27 as a function of @xmath129 .
bars indicate the experimentally known region .
the original @xmath112 values from @xcite are used , which is the main source of error .
, title="fig:",width=453,height=325 ] + comparison of calculated @xmath128 from the caputo - riemann - riemann symmetric rotor ( [ e2 ] ) with the parameter set ( [ om0])-([m ] ) , minimized with respect to @xmath112 with the experimental masses from audi @xcite as a function of @xmath27 . from top to bottom the minimized @xmath112 values , theoretical @xmath128 , experimental microscopic contribution from the difference of experimental mass excess and and macroscopic frldm energy and error in @xmath130 $ ] are plotted .
, title="fig:",width=453,height=472 ] + in order to calculate the shell corrections , we introduce the following parameters : @xmath131\\ \label{om1 } \omega_1 & = & \omega_0 q^{-\frac{1}{3}}\\ \label{om2 } \omega_2 & = & \omega_1\\ \label{om3 } \omega_3 & = & \omega_0 q^\frac{2}{3}\\ \label{az } \alpha_z & = & \cases { 0.46 + 0.000220 \ , z&$z > 50$\\ 0.2469 + 0.00448 \ , z&$28<z \leq 50$\\ 0.2793 + 0.00332 \ , z&$z \leq 28$\\ } \\ \label{an } \alpha_n & = & \cases { 0.41 + 0.000200 \ , n&$n >
50$\\ 0.3118 + 0.00216 \ , n&$28<n \leq 50$\\ 0.2793 + 0.00332 \ ,
n&$n \leq 28$\\ } \\ \label{gam } \gamma & = & 1.1 \ , \hbar \omega_0\\ \label{m } m&= & 4\\ a_0 & = & 2.409 [ { \textrm{mev}}]\\ a_r & = & 15.0 [ { \textrm{mev}}]\end{aligned}\ ] ] input parameters are the number of protons @xmath129 , number of neutrons @xmath27 , the nucleon number @xmath132 , and the ground state quadrupole deformation @xmath112 .
the values obtained include the frequencies ( [ om1]),([om2]),([om3 ] ) , which result from a least square fit and quadratic approximation of equipotential surfaces , the fractional derivative coefficients for protons ( [ az ] ) and neutrons ( [ an ] ) which determine the level spectrum for protons and neutrons for the proton and neutron part of the shell correction energy respectively from a fit of the set of nuclids @xmath133 , @xmath134 , @xmath135 , @xmath136 and from the requirement , that the neutron shell correction for @xmath134 should amount about @xmath137 $ ] , ( [ gam ] ) from the plateau condition @xmath125 ( see figure [ plateau ] ) and ( [ m ] ) the order of included hermite polynomials for the strutinski shell correction method .
finally @xmath138 from a fit of the experimental mass excess given in @xcite .
we compare our results with for the microscopic energy contribution @xmath128 with data from mller et .
@xcite and use their tabulated @xmath112 values .
they have not only listed data for experimental masses but also predictions for regions , not yet confirmed by experiment . in figure [ someshell ]
we compare the calculated @xmath104 values with the tabulated @xmath128 , which is justified for almost spherical shapes ( @xmath139 ) .
the results agree very well within the expected errors ( which are estimated @xmath140 $ ] for the pairing energy and @xmath141 $ ] for @xmath128 ) , especially in the region of experimentally known nuclei .
a remarkable difference between the calculated shell correction and tabulated @xmath128 from mller occurs for superheavy elements ( @xmath142 , last picture in figure [ someshell ] ) .
while phenomenological shell models predict a pronounced minimum in the shell correction energy for @xmath143 @xcite-@xcite the situation is quite different for the rotor model , where two magic shell closures at @xmath144 and @xmath145 are given , but the @xmath144 shell closure is not strong enough to produce a local minimum in the shell correction energy plot as a function of @xmath129 . instead , between @xmath144 and @xmath145 , there emerges a slightly falling energy plateau , which makes the full region promising candidates for stable , long - lived superheavy elements .
while this result contradicts predictions made with phenomenological shell models , it supports recent results obtained with relativistic mean field models @xcite , which predict a similar behaviour in the region of super heavy elements as the proposed rotor model . in figure [ compare ]
we have covered the complete region of available experimental data for nuclids and compare the calculated theoretical microscopic energy contribution minimized with respect to the deformation with the experimental values .
the influence of shell closures is very clear .
the rms - error is about @xmath146 $ ] .
the maximum deviation occures between closed magic shells . therefore in the next section we will introduce a generalization of the proposed fractional rotor model , which not only determines the magic numbers accurately but in addition determines the fine structure of the single particle spectrum correctly .
in the previous section we have demonstrated , that the caputo - riemann - riemann symmetric rotor correctly determines the magic numbers in the single particle spectra for neutrons and protons .
however , there remains a significant difference between calculated and experimental ground state masses for nuclei with nucleon numbers far from magic shell closures .
this is a strong indication for the fact , that the fine structure of the single particle levels is not yet correctly reproduced .
we therefore propose the following generalization of the caputo - riemann - riemann symmetric rotor group : @xmath147 with the casimir operators ( [ lz ] ) and ( [ l2 ] ) it follows for the hamiltonian @xmath55 : @xmath148 with the free parameters @xmath149 , where @xmath150 may be called fractional magnetic field strength in units @xmath151 $ ] . imposing the same symmetries ( [ sym1]),([sym2 ] ) as in the case of the symmetric caputo - riemann - riemann rotor , the eigenvalues of the hamiltonian ( [ hamilton2 ] ) are given as @xmath152[2n_1 + 1]+ b \hbar \omega_0 { _ \textrm{\tiny{c } } } [ 2n_1]+ \nonumber \\ & & \hbar \omega_2 { _ \textrm{\tiny{r } } } [ 2n_2][2n_2 + 1]+ \hbar \omega_3 { _ \textrm{\tiny{r } } } [ 2n_3][2n_3 + 1 ] \\
\label{x2 } & = & \sum_{i=1}^3 \hbar \omega_i \frac{\gamma(1 + ( 2 n_i + 1 ) \alpha)}{\gamma(1 + ( 2 n_i-1 ) \alpha ) } - \delta_{n_1 0 } \hbar \omega_1 \frac{\gamma(1 + \alpha)}{\gamma(1 - \alpha)}\nonumber\\ & & + b \hbar \omega_0 \frac{\gamma(1 + ( 2 n_1 ) \alpha)}{\gamma(1 + ( 2 n_1 - 1 ) \alpha ) } - \delta_{n_1 0 } b \hbar \omega_0 \frac{1}{\gamma(1 - \alpha)}\nonumber\\ & & \qquad\qquad\qquad\qquad\qquad n_1,n_2,n_3=0,1,2 , .. \end{aligned}\ ] ] on a basis @xmath63 .
we call this model the extended caputo - riemann - riemann symmetric rotor .
the additional @xmath153 term yields a level splitting of the harmonic oscillator set of magic numbers @xmath90 ( [ set1 ] ) , while the multiplicity of the @xmath89 set ( [ set2 ] ) remains unchanged , since this set is characterized by @xmath154 .
this is exactly the behaviour needed to describe the experimentally observed fine structure , as can be deduced from the right hand side of figure [ fig2 ] . in order to clearly demonstrate the influence of the additional term
, we first investigate the level spectrum for the spherical ( [ sphere ] ) and idealized case @xmath64 .
the level spectrum ( [ x2 ] ) simply results as : @xmath155 where @xmath156 denotes the double factorial . for @xmath64 , on the left side the level spectrum for the extended caputo - riemann - riemann symmetric rotor ( [ x2 ] )
is plotted as a function of the fractional magnetic field strength @xmath150 .
the levels are labeled according to the corresponding @xmath157 $ ] nilsson scheme and the multiplicity is given .
for @xmath158 the resulting level sequence near @xmath159 is plotted on the right . at @xmath160
the resulting spectrum coincides with the corresponding spherical nilsson level spectrum .
brackets indicate the proposed appropriate combinations of rotor levels . , title="fig:",width=453,height=238 ] + comparison of calculated @xmath128 from the extended caputo - riemann - riemann symmetric rotor ( [ x2 ] ) with the parameter set ( [ om02])-([bbb ] ) , minimized with respect to @xmath112 with the experimental masses from audi @xcite as a function of @xmath27 . from top to bottom the minimized @xmath112 values , theoretical and experimental masses and error in @xmath130 $ ]
are plotted .
, title="fig:",width=453,height=468 ] + on the left side of figure [ gc2 ] this spectrum is plotted in units @xmath151 $ ] .
single levels are labeled according to the nilsson - scheme and multiplicities are given in brackets . for small fractional field strength @xmath150
the resulting spectrum exactly follows the schematic level diagram of a phenomenological shell model with spin - orbit term , as demonstrated e.g. by goeppert - mayer @xcite .
a small deviation from the ideal @xmath161 value reproduces the experimental spectra accurately : for @xmath158 the resulting level spectrum is given on the right hand side of figure [ gc2 ] .
obviously there is an interference of two effects : first , for @xmath162 now the degenerated levels of both magic sets split up and second the fractional magnetic field @xmath150 acts on the subset @xmath90 .
for @xmath160 the spectrum may be directly compared with the spherical nilsson level scheme , which is given for neutrons between @xmath163 as @xmath164 , @xmath165 , @xmath166 , @xmath167 , @xmath168 , @xmath169 , see e.g. results of @xcite , which corresponds to a sequence of sub - shells at @xmath170 .
this sequence is correctly reproduced with the extended caputo - riemann - riemann symmetric rotor . with the parameter set , which is obtained by a fit with the experimental masses of ca- , sn- and pb - isotopes @xmath171\\ \label{az2 } \alpha_z & = & \cases { 0.480 + 0.00022 \ , z&$z > 50$\\ 0.324 + 0.00332 \ , z&$z \leq 50$\\ } \\
\label{an2 } \alpha_n & = & \cases { 0.446 + 0.00022 \ , n&$n > 29$\\ 0.356 + 0.00332 \ , n&$n \leq 29$\\ } \\ \label{bbb } b & = & 0.27 \end{aligned}\ ] ] the experimental masses are reproduced with an rms - error of @xmath172 $ ] .
results are given in figure [ compare2 ] .
the deformation parameters , obtained by minimization of the total energy , are to a large extend consistent with values given in @xcite e.g. for @xmath173 we obtain @xmath174 , which conforms with mllers ( @xmath175 ) and rutzs results @xcite . however , there occur discrepancies mostly for exotic nuclei .
for example our calculations determine the nucleus @xmath176 to be almost spherical , while mller predicts a definitely oblate shape . finally , defining a nucleus with @xmath177 as prolate and with @xmath178 as oblate
the amount of prolate shapes is about @xmath179 of all deformed nuclei .
this is close to the value of @xmath180 @xcite , obtained with the nilsson model using the standard parameters . summarizing the results presented ,
the proposed extended caputo - riemann - riemann symmetric rotor describes the ground state properties of nuclei with reasonable accuracy .
we have demonstrated , that the nuclear shell structure may indeed be successfully described on the basis of a dynamical symmetry model .
the advantages of this model , compared to phenomenological shell and relativistic mean field models respectively are obvious : magic numbers are predicted , they are not the result of a fit with a phenomenological @xmath92-term
. there are no potential - terms or parametrized skyrme - forces involved and finally , all results may be calculated analytically .
the results obtained encourage further investigations in this field .
the next steps should include the pairing energy term and parameters should be determined by a more sophisticated fit procedure . with these additional contributions the model will most probably describe nuclear properties with at least similar accuracy as the models currently used .
based on the riemann- and caputo definition of the fractional derivative we used the fractional extensions of the standard rotation group @xmath0 to construct a higher dimensional representation of a fractional rotation group with mixed derivative types .
we obtained an extended symmetric rotor model , which predicts the sequence of magic proton and neutron numbers accurately .
furthermore we have shown , that the ground state properties of nuclei can be reproduced correctly within the framework of this model .
hence we have demonstrated , that a dynamic symmetry , generated by mixed fractional type rotation groups is indeed realized in nature .
we thank a. friedrich for useful discussions .
[ 10 ] goeppert - mayer m phys . rev .
* 75 * ( 1949 ) 1669 .
haxel f p , jensen j h d and suess h d phys
* 75 * ( 1949 ) 1769 .
nilsson s g kgl .
danske videnskab .
selsk . mat .-
* 29 * , ( 1955 ) 431 .
nilsson s g et al .
a * 131 * , ( 1969 ) 1 .
hofmann s and mnzenberg g reviews of modern physics , * 72 * , ( 2000 ) 733 .
rufa m et al .
c * 38 * ( 1988 ) 390 .
bender m , nazarewicz w , reinhard p - g phys .
b515 * ( 2001 ) 42 .
rutz k et al .
c * 56 * ( 1997 ) 238 .
kruppa a t et al .
c * 61 * ( 2000 ) 034313 .
elsasser w m j. phys .
radium * 5 * ( 1934 ) 635 .
elliott j p proc .
london * a245 * , ( 1958 ) 128 .
iachello f and arima a , 1987 _ the interacting boson model _ cambridge university press , cambridge .
miller k and ross b 1993 _ an introduction to fractional calculus and fractional differential equations _ wiley , new york .
oldham k b and spanier j 2006 _ the fractional calculus _ , dover publications , mineola , new york .
podlubny i 1999 _ fractional differential equations _ , academic press , new york .
herrmann r 2008 _ fraktionale infinitesimalrechnung - eine einfhrung fr physiker _ , bod , norderstedt , germany leibniz g f sep 30 , 1695 _ correspondence with lhospital _ , manuscript .
euler l commentarii academiae scientiarum petropolitanae * 5 * , ( 1738 ) pp . 36 - 57 .
liouville j 1832 j. @xmath181cole polytech .
, * 13 * , 1 - 162 .
riemann b jan 14 , 1847 _ versuch einer allgemeinen auffassung der integration und differentiation _ in : weber h ( ed . ) , _ bernhard riemann s gesammelte mathematische werke und wissenschaftlicher nachlass _ , dover publications ( 1953 ) , 353 . caputo m geophys
. j. r. astr
* 13 * , ( 1967 ) 529 .
weyl h vierteljahresschrift der naturforschenden gesellschaft in zrich * 62 * , ( 1917 ) 296 .
feller w comm .
universite de lund , ( 1952 ) 73 - 81 .
riesz m acta math .
* 81 * , ( 1949 ) 1 .
grnwald a k z. angew . math .
und physik * 12 * , ( 1867 ) 441 .
herrmann r j. phys .
g : nucl . part .
34 * , ( 2007 ) , 607 .
herrmann r arxiv : math - ph/0510099 .
maruhn j and greiner w z. physik * 251 * , ( 1972 ) 431 .
strutinsky v m nucl
a95 * , ( 1967 ) 420 .
strutinsky v m nucl
* a122 * , ( 1968 ) 1 .
mller et .
atomic data nucl .
data tables * 59 * , ( 1995 ) 185 .
herrmann r arxiv : gen - ph/0801.0298 .
myers w d and swiatecki w j nucl .
phys . * 81 * , ( 1966 ) 1 .
sobiczewski a , gareev f a and kalinkin b n phys .
* 22 * ( 1966 ) , 500 .
meldner h ark .
36 * , ( 1967 ) 593 .
mosel u , fink b and greiner w contribution to `` memorandum hessischer kernphysiker '' darmstadt , frankfurt , marburg , ( 1966 ) .
mosel u and greiner z. f. physik * 217 * ( 1968 ) 256 , * 222 * ( 1968 ) 261 .
audi g , wapstra a h and thibault c nucl .
* a729 * ( 2003 ) 337 .
scharnweber d , mosel u and greiner w phys .
* 24 * ( 1970 ) 601 .
tajima n and suzuki n phys .
c * 64 * ( 2001 ) 037301 . | based on the riemann- and caputo definition of the fractional derivative we use the fractional extensions of the standard rotation group @xmath0 to construct a higher dimensional representation of a fractional rotation group with mixed derivative types .
an extended symmetric rotor model is derived , which predicts the sequence of magic proton and neutron numbers accurately .
the ground state properties of nuclei are correctly reproduced within the framework of this model . |
it has been found that very strong magnetic fields are capable of destabilizing the electroweak vacuum by forming a vector boson @xmath0 condensate and restoring the symmetry @xcite .
the required field can only be thought to have existed at the very beginning of the universe and one of the possibilities is that it was generated during the electroweak phase transition @xcite .
this primordial field may have been subsequently the seed of the present galactic magnetic field @xcite .
one may wonder whether the restoration of symmetry caused by this strong magnetic field can delay the electroweak phase transition . in particular ,
if it is of first - order the magnetic field might avoid its completion through the bubble mechanism . the simplest way to see why a strong magnetic field can destabilize
the electroweak vacuum is to consider the energy of a charged spin - one particle interacting with a uniform magnetic field along the 3-axis @xmath1 for the lowest landau level @xmath2 if the gyromagnetic factor @xmath3 is @xmath4 as occurs in the @xmath5 case , it is clear that the effective mass will become zero for @xmath6 this expression is analogous to that of the critical electric field required to create pairs through tunneling .
if one wishes to calculate the decay probability of the vacuum , one must evaluate @xmath7 in euclidean metric the one - loop amplitude for a scalar field depends on @xmath8 with @xmath9 , being @xmath10 .
using the schwinger proper time method @xcite one obtains @xmath11 for constant electromagnetic fields the trace is known to give the vacuum energy density @xcite @xmath12 \quad , \label{e1}\ ] ] where the @xmath13 comes from subtracting @xmath14 .
this integral has a logarithmic divergence for @xmath15 which can be absorbed renormalizing fields and charge @xcite . in eq.([e1 ] ) for @xmath16 the integral has poles in the @xmath17-axis which give origin to an imaginary part corresponding to pair creation
. we will be instead interested in the case of @xmath18 and constant magnetic field for which @xmath19\ ] ] that has no poles .
for the spin-1 @xmath20 case we adopt the view that the only modification to @xmath21 is the interaction of spin with magnetic field @xmath22 in the exponent of eq.([e2 ] ) .
now the trace must be performed on momentum and spin states where the latter involves this added interaction to give @xmath23 since the remaining trace is equal to the scalar case , the relevant part of the vacuum energy density is @xmath24 \quad .\ ] ] this expression has no poles but diverges for @xmath25 when @xmath26 due to the gyromagnetic factor @xmath4 of the @xmath5 boson , which would not occur either for @xmath27 or for the @xmath28 case .
this divergence is an indication of the vacuum instability for large magnetic field .
the decay rate should be evaluated in the more realistic situation of @xmath29 increasing with time , with the consequent generation of an electric field . in our calculation of the next section we will not take into account the evolution with time of the magnetic field but we will consider that when its value is larger than the critical one in the region of a bubble containing the broken - symmetry vacuum , the bubble will be destroyed .
it is known that the phase transition of the standard model is of first order only for higgs masses which are below the experimental bound @xcite .
however we will not attempt to use an extension as the mssm to allow consistency with this bound since our goal is to establish the effect of a strong magnetic field on the first - order transition with a reasonably simple effective potential . in the minimal standard model
this temperature dependent potential for the higgs field @xmath30 is @xcite @xmath31 @xmath32 @xmath33 @xmath34 .
all the parameters of the potential eq.([e3 ] ) are determined by the experimental masses of gauge bosons and top quark , @xmath35 , @xmath36 , @xmath37 and the breaking scale @xmath38 , with the only unknown given by the higgs mass @xmath39 .
the first - order transition will occur in the range from @xmath40 ^{1/2}$ ] to @xmath41 , where @xmath42 corresponds to equal minima and @xmath43 to the disappearance of the barrier between them . in this range of temperatures bubbles of broken - symmetry phase
are formed with a probability per unit volume and time given by the semiclassical approximation @xmath44 where @xmath45 .
@xmath46 will come from the maximization of the bubble energy which , in the thin wall approximation of width @xmath47 is @xmath48 where the surface tension is @xmath49 @xmath50 and @xmath51 are the differences of potential and higgs field between the two phases respectively .
the maximization of eq.([e4 ] ) gives the critical radius @xmath52 and energy @xmath46 @xmath53 the velocity of expansion of the bubble can be estimated by @xcite @xmath54 and the wall width by @xmath55 where @xmath56 , for not too high values of @xmath39 .
we will consider that the phase transition will be completed when the average distance between centres of bubbles @xmath57 is equal to twice their average radius @xmath58 , i.e. when @xmath59 $ ] vanishes ( alternatively , one could determine the moment at which the fraction of volume occupied by bubbles of new phase equals one @xcite ) . the relation between the time @xmath60 and temperature in the radiation regime
is taken as @xmath61 to give the beginning of the phase transition at @xmath62 . in terms of the probability for bubble formation
, its number density will be @xmath63 so that the average radius of the volume per bubble is @xmath64 ^{-1/3}\quad .\ ] ] the average bubble radius comes from @xmath65 where @xmath66 .
now we must include the influence of the magnetic field generated during the expansion of bubbles .
several mechanisms of creation of seed fields ( e.g. charge separation in bubble walls , bubble collisions or fluctuating higgs gradients ) @xcite involve the appearance of non - zero expectation values of the higgs field .
this means that the seed fields will arise in general in the regions where bubbles are present .
once a seed is formed , the magnetic field is quickly amplified due to magnetic turbulence in the conducting fluid , until it reaches in this region the value given by energy equipartition .
therefore we will consider that when a bubble is produced a magnetic field @xmath67 will manifest itself in a region @xmath68 after a time @xmath69 .
the exponent of the denominator is the statistically most favorable one to have sizeable homogeneous fields as the region increases .
the constant @xmath70 has a maximum value of @xmath71 to satisfy the constraint given by nucleosynthesis @xcite .
we will assume @xmath72 which is in reasonable agreement with some mechanisms of magnetic field generation @xcite .
since for @xmath73 the decay rate of the electroweak vacuum , @xmath74 @xcite , is much larger than the bubble nucleation or expansion rates , we can consider that the bubble disappears because the symmetry inside it is restored . in this way @xmath75 for @xmath76 and @xmath77 in ( [ rb ] ) and the completion of the phase transition according to a bubble mechanism will be either delayed or avoided .
our numerical calculation shows that for large values of @xmath39 , e.g. @xmath78 , the first - order phase transition is not completed even without magnetic field ( see fig.1 ) . for a low value of the higgs mass @xmath79 our model
allows the completion of the first - order transition without magnetic field ( see fig.2 ) .
if the maximum possible value of the magnetic field @xmath80 is assumed , the first - order transition is avoided ( fig.2a ) . with a smaller value of the magnetic field , e.g. @xmath81 , the completion of the first - order transition
is only delayed in around @xmath82 ( fig.2b ) , and well below this intensity the effect of the magnetic field is negligible .
an alternative scenario would be to consider that the magnetic field was generated in a prior stage of the universe , e.g. the gut epoch .
its effect on the electroweak phase transition can be studied through an effective potential depending on temperature and constant magnetic field @xcite .
it is interesting that again for @xmath83 the magnetic field delays the first - order transition due to an increase of the free energy of the broken - symmetry phase .
this result agrees with the tendency shown by our calculation but we remark that the two situations are different . in our case we do not have a magnetic field prior to the phase transition but a strong field is formed in small regions around the bubbles , destroying them .
the magnetic field in a subhorizon volume will have a statistical average several orders of magnitude below the critical value , therefore its influence on the effective potencial will be negligible . if the first - order transition is not completed , at temperature @xmath43 our system will be at the maximum of the potential @xmath84 whereas the minimum will be at @xmath85 without any barrier .
now the phase transition will occur through the global rolling down of the field @xmath30 towards the minimum of the potential .
we have seen that the highest possible magnetic field together with the most favorable law for having homogeneous field in regions of increasing size might have cosmological consequences through the non - completion of the first - order electroweak transition through a bubble mechanism .
therefore the usual electroweak baryogenesis due to bubble expansion would be affected .
but one must notice that also the homogeneous increase of @xmath30 can produce a matter - antimatter asymmetry .
this is because there will be a baryonic chemical potential related to the time variation of the @xmath86 violating phase @xmath87 .
the resulting baryonic density will depend on the variation @xmath88 in the interval when the sphalerons are active due to the smallness of @xmath30 . for a weakly first - order transition ,
an advantage of this mechanism compared to the bubble one is that the baryonic density would not be erased in the broken phase because here the value of @xmath89 is larger due to the delay of the phase transition .
one may remind that this problem is also avoided by the baryogenesis in cosmic strings but paying the price of a suppression factor in the active volume . however , it is unlikely that such a strong and large size primordial magnetic field has occurred , and for more acceptable fields the effect would be only a small decrease of the temperature for the completion of a first - order transition .
we have studied the influence of the magnetic field on the phase transition using the easiest model , i.e. the standard model and not the mssm where presumably the first - order phase transition can occur for not too light higgs mass @xcite .
the fact that we obtain the completion of the first - order transition without magnetic field for @xmath79 , not far below the experimental bound is probably due to the definition that it occurs when the bubbles touch each other without taking into account their scattering .
but we believe that the general conclusions on the magnitude of the effect do not depend on the details of the used electroweak model . regarding further developments of this calculation , it would be important to evaluate the vacuum decay rate caused by a time dependent magnetic field in order to consider more carefully its effect on the bubbles instead of taking the simplification of assuming their disappearance as soon as the magnetic field is larger than the critical value .
we are deeply indebted to a. della selva and r. iengo for very stimulating discussions .
l. m. and a. m. thank the kind hospitality of the departments of physics of the universities of napoli and della calabria at cosenza where parts of this research has been performed . b. cheng and a. v. olinto , phys .
* d50 * ( 1994 ) 2451 ; g. sigl , a. v. olinto and k. jedamzik , phys.rev .
* d55 * ( 1997 ) 4582 ; t. w. b. kibble and a. vilenkin , phys . rev .
* d52 * ( 1995 ) 679 ; j. ahonen and k. enqvist , phys.rev . * d57 * ( 1998 ) 664 ; t. vachaspati , phys .
* b265 * ( 1991 ) 258 ; o. tornkvist , fermilab - pub-97/257-a ( hep - ph/9707513 ) , phys . rev . * d * ( to appear ) .
k. enqvist and p. olesen , phys . lett .
* b319 * ( 1993 ) 178 ; m. turner and l. widrow , phys . rev .
* d37 * ( 1988 ) 2743 ; k. enqvist , invited talk at the strong and electroweak matter 97 ( astro - ph/9707300 ) and hip-1998 - 17/ph ( astro - ph/9803196 ) .
k. kajantie , m. laine , k. rummukainen and m. schaposnikov , phys .
* 77 * ( 1996 ) 2887 ; w. buchmuller and o. philipsen , nucl . phys .
* b443 * ( 1995 ) 47 ; w. buchmuller and z. fodor , phys . lett .
* b331 * ( 1994 ) 124 .
d. grasso and h. rubinstein , astrop .
* 3 * ( 1995 ) 95 ; p. j. kernan , g. d. starkman and t. vachaspati , phys .
rev . * d56 * ( 1997 ) 3766 ; b. cheng , a. v. olinto , d. j. schramm and j. w. truran , phys . rev . *
d54 * ( 1996 ) 4714 .
m. carena , m. quiros , a. riotto , i. vilja and c. wagner , nucl .
b503 * ( 1997 ) 404 ; j. m. cline and k. kainulainen , nucl . phys .
* b482 * ( 1996 ) 73 ; k. kajantie , m. laine , k. rummukainen and m. schaposnikov , nucl . phys .
* b446 * ( 1996 ) 189 ; m. losada , phys . rev . *
d56 * ( 1997 ) 2893 . | the broken - symmetry electroweak vacuum is destabilized in the presence of a magnetic field stronger than a critical value .
such magnetic field may be generated in the phase transition and restore the symmetry inside the bubbles .
a numerical calculation indicates that the first - order phase transition is delayed but may be completed for a sufficient low value of the higgs mass unless the magnetic field is extremely high . |
a key step in the modeling of accretion disks is the detailed description of the balance between heating and cooling terms in the energy transport equations ( see pringle @xcite ) . in this context
, two hypotheses have been generally considered . the first states that the only source of heating in the disk is viscous dissipation , so that the available power is provided by the release of gravitational binding energy from the infalling matter .
the second hypothesis is that cooling is essentially radiative .
this latter hypothesis leads to different disk models , depending on whether the disk is optically thick ( as in the `` standard '' model of shakura & sunyaev @xcite ) or optically thin ( as in the model of shapiro , lightman & eardley @xcite ; hereafter sle ) .
one more recent development in the description of the energetics of accretion disks has been the recognition that radiative cooling may sometimes be insufficient . under these circumstances ,
most of the energy remains stored in the disk rather than being radiated away , leading to the so - called _ advection - dominated accretion flows _ ( adaf ; narayan & yi @xcite ) .
this situation is described by a new equation of energy transport relevant to the radial direction , including an advection term ( see sect .
[ sec : standard ] below ) .
once the general characteristics of stationary models are established , one can proceed to address the problem of their stability .
different types of instabilities have been considered , leading to important conclusions about the possibility of applying the models to concrete systems . in one line of research
, the stability analysis is carried out with the goal of identifying microscopic collective mechanisms able to justify the viscosity @xmath1-prescription that is commonly introduced , following shakura & sunyaev ( @xcite ) , in order to bypass the low level of viscosity predicted by classical arguments .
other interesting stability aspects have also been investigated .
in particular , the thermal - viscous instability related to hydrogen ionization ( meyer & meyer - hofmeister @xcite ) has been invoked to explain the variability of dwarf novae .
in contrast , the optically thin sle model has been commonly regarded as inapplicable , because of the thermal instability that is found to occur when optically thin _
bremsstrahlung _ is the dominant cooling mechanism . to be sure
, the thermal stability properties of the models depend strongly on the assumptions about the balance between heating and cooling terms ( for example , it is well known that , in the adaf context , optically thin solutions exist in the inner disks and are thermally stable ; see section [ sec : standard ] and narayan & yi @xcite ) .
indeed , it has been noted ( piran @xcite ) that small variations in the viscosity law can change the stability of accretion disk models drastically .
the self - gravity of the disk can change this picture significantly .
it may have an obvious impact on the radial and on the vertical force relations ( paczyski @xcite ) .
furthermore , it has long been recognized that non - axisymmetric instabilities associated with the self - gravity of the disk can affect the properties of angular momentum transport and hence the relevant viscosity law ( e.g. , see lin & pringle @xcite , laughlin & bodenheimer @xcite ) , although it still remains to be clarified whether such modifications can be reconciled with the standard @xmath1-prescription .
recently we have argued ( bertin @xcite , bertin & lodato @xcite ) that self - gravity should be properly incorporated also in the energy balance equations , which should thus take a more general form than usually considered . in these first investigations we actually _ replaced _ the energy equations by a self - regulation prescription demanding that the axisymmetric stability parameter @xmath0 is maintained close to a value @xmath2 basically dictated by the condition of marginal jeans stability .
the physical justification for the existence of an efficient dynamical thermostat of this kind is based on the presence of suitable competing heating and cooling mechanisms , but was not made explicit in terms of a set of consistent equations for the energy transport .
this simplified approach is adequate for the construction of dynamical models , and has already led to some unexpected results ( in particular , the natural occurrence of flat rotation curves ) .
yet , it deliberately avoids the issue of what is the predicted emission of the accretion disk , which has had such an important role in the confrontation of models with observations in this general research area .
we anticipate that the disk self - gravity should be significant only in the cold outer disk . in this respect
, the identification of a consistent set of energy equations would also be desirable in order to set up a proper matching of solutions across the relevant transition region . in this paper
, we include the effect of jeans - related instabilities in the energy balance equations by means of a term of effective heating .
such heating term is modeled as a function steeply rising when @xmath0 falls below a threshold value @xmath2 and practically vanishing when @xmath0 rises above @xmath2 .
we then show how this more general energy equation leads to the self - gravitating , self - regulated accretion disk solution , characterized by a value of @xmath0 very close to @xmath2 and only weakly dependent on radius .
this stationary solution , although non - advective and optically thin , turns out to be thermally stable .
the paper is organized as follows . in section
[ sec : standard ] we briefly summarize the arguments that lead to the balance of heating and cooling terms in current theories of accretion disks .
in section [ sec : effect ] we propose a way to incorporate the effects of self - gravity in the energy equations by means of an additional heating term . in section
[ sec : equation ] we then provide and discuss a complete set of equations for the construction of the full models of self - gravitating accretion disks . in section [ sec : stab ] we outline the structure of the stationary models and study their thermal stability .
the conclusions of the paper are given in section [ sec : concl ] , where we briefly comment on possible astrophysical applications .
a generally accepted approach to the construction of accretion disk models starts from the assumption that the disk is geometrically thin ( so that vertically integrated and vertically averaged quantities can be usefully introduced ) and adopts a fluid model , specified by the continuity equation , the radial component of the euler equation , the conservation of angular momentum in the presence of viscous torques , a viscosity prescription ( such as the @xmath1-prescription ) , an expression for the vertical thickness of the disk ( to be derived from the solution of the vertical hydrostatic equilibrium equations ) , and the relevant energy equations . in the following
, we will refer to vertically integrated or averaged quantities , such as the surface density @xmath3 , as `` surface '' quantities , while we will refer to other quantities , such as the volume density @xmath4 , as `` volume '' quantities .
the energy equations are generally based on two hypotheses , of purely viscous heating ( with the addition , in some cases , of irradiation from the inner disk or from the central object ) and of purely radiative cooling .
the first hypothesis is thus contained in an expression for the heating rate @xmath5 in terms of `` surface quantities '' : @xmath6 where @xmath7 is the viscosity coefficient and @xmath8 is the differential rotation ( usually taken to be keplerian ) .
the second hypothesis , that cooling is entirely radiative , implies that the cooling rate per unit surface @xmath9 can be calculated from the `` volume '' properties of the disk and the assumed radiation mechanisms : @xmath10 the energy equations are then split into an equation for the radial transport and an equation for the vertical transport . in practice , in systems with given physical properties , one should carefully check the validity of this separation , which is natural only in the limit of thin disks . for the horizontal transport
the equation is the so - called `` advection '' equation : @xmath11 where @xmath12 is the radial velocity , @xmath13 is the midplane temperature , and @xmath14 is the entropy per unit mass .
this equation states that , locally , the unbalance between energy gained from heating and energy lost from cooling is stored in the disk and transported radially along with the infalling matter .
entropy in eq .
( [ adv ] ) is calculated based on the first principle of thermodynamics : @xmath15 where @xmath4 is the volume density , @xmath16 is the pressure , @xmath17 is the specific heat at constant volume and @xmath18 is the `` volume '' entropy . after averaging over the vertical direction one defines the averaged `` surface '' entropy so that : @xmath19 where @xmath20 is the thermal speed on the equatorial plane , @xmath21 is the central volume density , so that @xmath22 ( @xmath23 is the geometrical thickness ) and @xmath24 is the specific heat ratio .
note that in eq .
( [ adv ] ) one neglects the effects of horizontal radiative energy transport ( but see eq . ( 3 ) in narayan & popham @xcite ) . for the vertical transport
one uses the equations of radiative transport , which , integrated over the @xmath25-coordinate , give rise to eq .
( [ qmeno ] ) .
a final remark is in order .
the above equations are a set of differential equations , and hence they should be supplemented by a set of appropriate boundary conditions . however , for standard disks ( see sect .
[ sec : standard1 ] below ) , in which one neglects advection terms , the system reduces to a set of algebraic equations and no boundary conditions are thus required .
in contrast , adaf solutions are based on the full differential problem . even in this case , however , the self - similar solution found by narayan & yi ( @xcite ) ( see sect . [
sec : adaf ] ) avoids the need of specifying the boundary conditions .
other adaf models take such self - similar solution as the approximate solution to be matched at large radii . in this context , it has been noted ( yuan @xcite ) that the boundary conditions are , indeed , important in determining the structure of the accretion disk and should be handled carefully .
the so - called standard disk solutions are those for which , in eq .
( [ adv ] ) , the dominant terms are @xmath5 and @xmath9 , so that the advection equation reduces to : @xmath26 if we express this relation in view of the physical hypotheses mentioned previously , then the energy balance equation for standard disks is : @xmath27 as @xmath28 describes the emission of radiation from the disk , eq .
( [ standa ] ) shows that the emission is related to @xmath29 in a simple way , largely independent of the detailed construction of the model . on the other hand , eq .
( [ standa ] ) , involving all the surface quantities ( @xmath3 , @xmath13 , ... ) , is the closure relation needed to calculate the structure of the accretion model . in some cases
, the output of radiation may be insufficient so that the balance between @xmath5 and @xmath9 in eq .
( [ adv ] ) is not possible .
in this more general case one can not use eq .
( [ sei ] ) ( and hence eq . ( [ standa ] ) ) .
this situation has often been described in terms of a convenient parameter , called @xmath30 , defined as : @xmath31 so that eq .
( [ adv ] ) can be rewritten as : @xmath32 for specified radiation processes , it is clear that eq .
( [ f ] ) and eq .
( [ adve ] ) together can close the model , which is characterized ( for given values of the viscosity parameter @xmath1 and of the mass and angular momentum accretion rates @xmath33 , @xmath34 ) by a profile @xmath35 determined by eq .
( [ f ] ) . in reality
, it has been noted that , at a given radius @xmath36 , by considering the relation between @xmath30 and @xmath33 ( @xmath1 and @xmath34 being fixed ) , there are in general three branches of solutions ( at least for sufficiently small values of @xmath33 ) : ( _ i _ ) one branch corresponding to @xmath37 , so that ( see eq .
( [ f ] ) ) @xmath38 , is similar to the solution of standard disks , ( _ ii _ ) one new branch , called adaf , with @xmath39 , is characterized by the balance between viscous heating and the advection term ( see eq .
( [ adve ] ) ) ; for this branch there is little emission of radiation , as @xmath40 .
finally , a third branch ( _ iii _ ) , for intermediate values of @xmath30 , is thermally unstable and corresponds to the sle solution . treating @xmath30 as a given parameter , narayan & yi ( @xcite )
have shown that models closed only by eq .
( [ adve ] ) ( that is , not taking into account the radiation properties ) admit a simple , self - similar solution .
this solution can be seen as a first approximation to the adaf solution on a wide ( but , in any case , limited ) radial range .
from the solution of the non - self - gravitating disk equations , it is interesting to check the behavior of the parameter @xmath0 , which , as is well known , regulates the stability of the disk against axisymmetric jeans instability . for a fluid one - component disk the parameter @xmath0
is defined as : @xmath41 where @xmath42 is the epicyclic frequency ( equal to @xmath43 in the keplerian case ) .
self - gravity becomes important at large radii .
in fact , @xmath44 is expected to be a decreasing function of radius .
for example , for the optically thick disk dominated by the kramers opacity ( the outer region of the shakura & sunyaev @xcite disk ) it is proportional to @xmath45 .
when the standard model is applied to concrete cases ( for example , the conditions typical of an agn ) @xmath0 is found to plunge well below unity in regions of interest ( kumar @xcite ) .
this behavior is untenable .
it is only a reminder that the equations considered initially to construct the accretion disk model must be reformulated so as to incorporate the important role of the disk self - gravity ( e.g. , see kolykhalov & sunyaev @xcite ) . as a quick , constructive way to address the problem of self - gravitating accretion disks , we have recently argued ( bertin @xcite ) that the energy equations , when self - gravity is dominant , should be replaced by the physically based prescription that @xmath0 is self - regulated to a value @xmath2 close to unity .
this is all we need if we are interested in calculating the main dynamical characteristics of such self - regulated accretion disks . on the other hand , a detailed examination of the physical justification at the basis of the self - regulation process ( see sect . 3 of the article by bertin & lodato @xcite ) readily shows that self - regulation results from heating and cooling mechanisms that go beyond those considered in sect .
[ sec : standard ] . in the self - regulation picture ,
widely considered in the study of spiral galaxies , the jeans instability has the effect of inducing a fast heating of the disk ( on the dynamical time - scale ) , at a rate that depends strongly on @xmath0 .
such heating thus tends to bring @xmath0 towards values of stability . on the other hand ,
if the system is jeans stable to begin with , efficient cooling mechanisms ( that , in some cases , may be just the radiative cooling of eq . ( [ qmeno ] ) ) can cool the disk down to lower values of @xmath0 . in the following sections we will reformulate the energy balance equations starting from the arguments
just presented .
such reformulation of the energy equations will allow us to : ( _ i _ ) demonstrate how self - regulation can be established ; ( _ ii _ ) rediscuss the opacity properties in the self - gravitating regime .
in addition , it will give us ( _ iii _ ) a starting point for the study of the emission properties of self - gravitating accretion disks ; ( _ iv _ ) a tool to analyze the structure of the transition region , where the disk becomes self - gravitating .
a study of the last two points is postponed to a separate investigation .
we suggest that in the expression of the heating rate a term @xmath46 should be added ( so that @xmath47 ) , strongly dependent on the value of @xmath0 , which we express in terms of surface quantities as : @xmath48 here @xmath49 is a function that rises quickly when @xmath0 drops below @xmath2 ( for example , @xmath50 , with @xmath51 ) .
the form of this heating term is argued on dimensional grounds , given the fact that the collective effects associated with self - gravity are expected to heat the disk on the dynamical time - scale .
the general structure of @xmath46 is analogous to that of the heating term introduced in the context of spiral structure in galaxies to demonstrate the time - dependent process of self - regulation ( see bertin & romeo @xcite ) .
the argument is also analogous to that used by lin & pringle ( @xcite ) to incorporate the collective effects associated with self - gravity into an effective viscosity .
in fact , we could write ( if we take @xmath52 ) : @xmath53 with the definition : @xmath54 with this device , we see that the term we are introducing mimics the viscous heating for a system with modified viscosity @xmath55 . in practice , we think that the two issues , of a modified viscosity in the presence of spiral instabilities and of a heating term associated with jeans - related instabilities , should be kept separate . note that in the context of spiral galaxies the emphasis is reversed , and little attention is paid to the problem of viscosity , although it is well known that spiral instabilities eventually favor mass accretion toward the center . in any case , it should be stressed that the present approach remains heuristic .
the issue of energy dissipation requires further clarification ( see also balbus & papaloizou @xcite ) . by referring to the context of galactic disks
, we find another important element that is generally overlooked in the discussion of the energy balance equations for accretion disks .
basically this is the distinction between thermal speed and _ effective _ thermal speed . in practice
, the effective thermal speed entering the dynamical equations may be higher than the thermal speed entering the radiation equations .
the interstellar medium of galaxy disks is a cloudy medium , for which the thermal speed associated with the temperature of the clouds is much smaller than the velocity dispersion of the cold gas clouds .
it is the latter that sets the appropriate effective thermal speed relevant to the dynamics of the disk .
when a fluid model of the interstellar medium is adopted , in studies of disk dynamics , the numbers that are used for the equivalent acoustic speed are then those applicable to the turbulent speed characteristic of the cloudy medium .
therefore , one important cooling mechanism is that associated with the inelastic collisions between gas clouds . mechanisms of this kind may operate also in accretion disks , but they are usually ignored ( but see quataert & chiang @xcite ) . because of these considerations we see that the energy balance equation for a _ non - advective _ disk , should be written as : @xmath56 in the inner disk @xmath0 is likely to be much larger than unity : if we drop the _ caveat _ about the distinction between thermal speed and effective thermal speed , we see that such equation would then reduce back to that of a standard disk .
in the following we will study the problem of self - gravitating accretion disks in a region not far from the transition region from non - self - gravitating to self - gravitating disk ( see fig .
[ fig : sketch ] ) .
the analysis is thus simplified in two respects .
( _ i _ ) on the one hand , the radial balance of forces for a cool , slowly accreting disk will be described by the keplerian relation : @xmath57 where @xmath58 is the mass of the central object . in eq .
( [ kepler ] ) we have neglected the contribution of the disk self - gravity . this approximation is valid for @xmath59 , where @xmath1 is the viscosity parameter ( see eq .
( [ alpha ] ) below ) . for parameters typical of an agn ,
previous studies ( bardou , heyvearts & duschl @xcite ) show that the transition from non - self - gravitating to self - gravitating disk occurs at a radius , which we call @xmath60 , that is much smaller than @xmath61 , so that , at least for a certain parameter range , the approximation in eq .
( [ kepler ] ) is applicable to the inner parts of the disk .
the effects associated with the integral relation between the gravitational field generated by the disk @xmath62 and @xmath3 are described in the paper by bertin & lodato ( @xcite ) .
( _ ii _ ) on the other hand , the energy balance equation is taken to be ( for the definitions of @xmath29 and @xmath46 see eqs .
( [ qpiu ] ) and ( [ accaj ] ) ) : @xmath63 with @xmath28 specified by the relevant radiation mechanism . in the following
we will use the expression for the cooling rate per unit volume appropriate for optically thin _ bremsstrahlung _ ( but see the additional discussion of sect .
[ robustness ] ) : @xmath64 where @xmath65 is the particle density , given by @xmath66 ( @xmath67 is the mean molecular weight ) , and @xmath68 is a gaunt factor averaged over frequencies ( see rybicki & lightman @xcite ) , and where the last expression is evaluated in cgs units ( @xmath69 measured in k ) . in eq .
( [ brems ] ) @xmath20 is the speed of light and should not be confused with the effective thermal speed defined above ; we have assumed that the atomic weight is unity , and that @xmath70 . in the spirit of the thin disk approximation
we will define @xmath71 .
the optically thin description breaks down at small radii , below a scalelength @xmath72 , where the disk is optically thick ( see discussion in sect . [
sec : stab ] ) .
the remaining equations are the equation of continuity : @xmath73 the conservation of angular momentum : @xmath74 and the standard @xmath1-prescription ( shakura & sunyaev @xcite ) for viscosity : @xmath75 where @xmath23 is the vertical thickness of the disk , defined by : @xmath76.\ ] ] the last equation is an interpolation formula for the `` exact '' thickness calculated including both the self - gravity of the disk and the gravitational force of the central object in the vertical hydrostatic equilibrium ( see eq .
( a9 ) in the article by bertin & lodato @xcite and discussion below ) , in the limit of keplerian rotation curve .
note that in the limit of non - self - gravitating disk ( @xmath77 ) this equation leads to the commonly used result @xmath78 . in the following
we will take @xmath79 ; the effects of a non - zero net angular momentum flux modify the disk structure only in the innermost parts of the disk .
the level of viscosity present , as measured by the parameter @xmath1 , should depend also on the development of spiral instabilities , and , therefore , on the parameter @xmath0 . since for the self - gravitating part of the disk @xmath0 will turn out to be self - regulated , we prefer to simplify the analysis by leaving @xmath1 as an assigned parameter .
our reference value ( often considered for agn - type accretion disks ) will be @xmath80 .
the above set of equations is naturally reduced to dimensionless form by the introduction of the length - scale : @xmath81^{2/3}.\ ] ] in practice , this length - scale roughly corresponds to the outermost radius for which our solution exists ( see sect [ sec : branch ] below ) . in the following we will thus refer to the dimensionless radius @xmath82 . based on the scale lengths introduced so far , and on general results that will be better justified in sect .
[ sec : stab ] , we can characterize the structure of the accretion disk , in the region we are focusing on , by the following four zones ( see fig .
[ fig : sketch ] ) : ( _ i _ ) an inner standard disk with the value of @xmath0 decreasing rapidly outwards ; ( _ ii _ ) a transition zone , in which the heating term due to self - gravity becomes important and the optical thickness of the disk decreases until ( _ iii _ ) the optically thin regime is reached ; ( _ iv _ ) at radii of the order of @xmath61 self - gravity affects also the rotation of the disk and at even larger radii the simple set of equations adopted in this section breaks down and possibly other cooling mechanisms come into play . the term self - gravitating " is often used with different meanings , because the quantity @xmath83 may be judged to be important in relation to different quantities : ( a ) the cumulative effect of @xmath83 may make the gravitational force in the plane different from the keplerian field generated by the central point mass ; in this respect , to check whether the disk should be considered to be self - gravitating , we should compare the location @xmath36 of interest with the scale @xmath61 of the problem . in the limit
@xmath84 the gravitational field of the central point mass may be neglected , but at @xmath85 the disk is already self - gravitating .
( b ) the transport properties in the disk may be affected significantly by density waves and other jeans - related phenomena ; here the importance of the disk self - gravity is quantified by comparing @xmath86 to the acceleration @xmath87 and thus it is quantified by @xmath0 .
note that , in this respect , for a disk to be self - gravitating it is sufficient that @xmath0 be close to unity ; in other words , the self - gravitating limit is _ not _ set by the condition @xmath88 , as might be thought at first .
( c ) the vertical scale of the disk may be significantly changed with respect to that of a non - gravitating disk ; as shown in a separate article ( bertin & lodato @xcite ; see fig .
a1 and eq .
( a9 ) in the appendix ) , this depends _ both _ on the local value of @xmath0 _ and _ on the overall mass distribution in the disk .
furthermore , the expression of the disk thickness given in eq . ( a9 ) of bertin & lodato ( @xcite ) is valid uniformly and can be considered as an exact " formula ; the self - gravitating expression @xmath89 is applicable , for example , when @xmath90 and the gravitational field in the disk plane is dominated by the contribution of a disk characterized by @xmath91 . to a large extent ,
the three items above are independent aspects of the role of self - gravity , even though we may argue that at very small radii any accretion disk is likely to be non - self - gravitating from all the three points of view , while a cold , radially extended accretion disk at large radii is bound to be self - gravitating with respect to all aspects ( a ) , ( b ) , and ( c ) .
in this paper we are less interested in the specific impact of item ( a ) ( investigated earlier by us ) and wish to focus on item ( b ) , which had not been properly explored in relation to the energy equations .
fortunately , effects related to jeans instability , which we model by means of the heating term described by eq .
( [ accaj ] ) , become important at a location @xmath92 , which , for models applicable to agn disks , occurs well far in with respect to the location where item ( a ) comes into play , because @xmath93 . as to the disk thickness , we have just recalled that the expression given in eq .
( a9 ) of bertin & lodato ( @xcite ) can be considered as an exact " formula ; thus its keplerian limit , as adopted in eq .
( [ acca ] ) , is legitimate when item ( a ) is unimportant .
therefore , in the transition region we can safely follow the set of equations described in subsection 4.1 .
optically thin , non - advective disks have been generally discarded because of the thermal instability that results from an increasing of the heating rate relative to the cooling rate in response to an increase of the temperature .
this occurs because they are found to violate the condition for thermal stability : @xmath94 where the subscript @xmath3 indicates that the derivative should be taken at constant surface density ( see , for example , pringle @xcite ) . since our discussion of sect .
[ sec : effect ] modifies the basic ingredients of the energy balance equations , we should now check what are the properties of the self - gravitating solutions in relation to the thermal stability condition .
given the fact that the additional heating term @xmath46 given in eq .
( [ accaj ] ) is a function of @xmath0 and that such parameter is proportional to the thermal speed @xmath20 , we can reformulate the thermal stability condition as : @xmath95 in fig .
[ fig : stab ] we show the relation between @xmath96 and @xmath0 at fixed radius , for different values of the radius .
equilibrium corresponds to @xmath97 ( the vertical line in fig . [
fig : stab ] ) and the solution is thermally stable if the slope of the curve is positive . as is easily recognized , there are two branches of solution : one , unstable , is characterized by @xmath77 , for which the dominant heating mechanism is viscous dissipation as in non - self - gravitating disks ( this solution can be regarded as analogous to the sle solution ) .
the other solution is self - gravitating ( it has @xmath90 ) and it is , indeed , thermally stable .
note that such stable equilibrium solution is always characterized by a value of @xmath0 close to unity .
the equilibrium @xmath0 increases slightly with radius ( although this dependence is weak ) ; independently of the physical parameters of the disk , the maximum value of @xmath0 for the stable solution is @xmath98 .
there is a maximum radius for which this kind of solution is possible , located near @xmath99 .
the relative location of the two scale lengths @xmath99 and @xmath61 defines the importance of the deviation from keplerian rotation in our model . for parameters typical of agn configurations ( @xmath100 )
we estimate @xmath101 .
hence , before the scale @xmath99 is reached the rotation curve is expected to depart from the simple keplerian curve , so that the quantitative results of the present simplified work are only approximate when @xmath102 and may even be qualitatively incorrect when @xmath103 is of order of unity .
the lack of solutions at large radii ( @xmath104 ) is due to the fact that radiative cooling is insufficient and unable to balance the heating rate . in this context
, one should keep in mind that , as we indicated in sect .
[ sec : other ] , other kinds of effective cooling could be efficient and play a role in the outer regions , such as inelastic collisions between clouds , which have been neglected here .
we have also calculated the optical depth of the disk , to check that the model of optically thin disk is consistent .
we express the `` effective '' optical depth @xmath105 ( rybicki & lightman @xcite ) for the case @xmath106 ( where @xmath107 and @xmath108 are the opacity due to thomson scattering and _ bremsstrahlung _ , respectively ) , as : @xmath109 where @xmath110 and @xmath111 . in fig .
[ fig : thickness ] we show the optical depth as a function of radius for the thermally stable equilibrium solution .
the disk is optically thin in the outer regions and the optical depth increases inwards . in the innermost regions the disk eventually becomes optically thick
this occurs at a dimensionless radius @xmath112 independent of the central mass and very weakly dependent on the mass accretion rate ( approximately @xmath113 ) ; for @xmath114 the transition to the optically thick region occurs at an inner radius @xmath115 . in fig .
[ fig : thickness2 ] we show the profile of the ratio @xmath116 to check that thomson opacity always dominates over free - free absorption . in the transition region from optically thick to optically thin emission , for a proper matching , a more detailed description would be desired .
in contrast with other cases for which optically thin emission is considered ( i.e. , the inner regions of disks with very low mass accretion rates ) , in our case the temperature of the disk ( for reasonable values of @xmath33 ) is expected to be low , so that pure _ bremsstrahlung _ fails to be the dominant radiative process and line emission and absorption become important . in view of this remark ,
we have checked the consequences of a different cooling function , using a functional interpolation ( sarazin & white @xcite ) of numerical data ( raymond , cox & smith @xcite ) , appropriate for lower temperatures . in this test of robustness , the value of the accretion rate @xmath33 , through the temperature of the outer disk defined as @xmath117 , appears explicitly in the problem .
this is the result of the fact that the various line emissions become efficient at specified temperatures . in practice ,
only minor differences are found for the self - gravitating branch ( even under extreme conditions ; see fig .
[ fig : stabcool ] ) , with respect to the previous analysis , because the cooling function influences only the equilibrium value of @xmath0 but this is , in any case , self - regulated to values close to unity ( on the thermally stable equilibrium solution ) .
self - gravitating accretion disks may be present in many astrophysical systems , ranging from protostellar disks to agn and could also be of interest for the general galactic context . in protostellar disks
it is likely that , especially at the beginning of the accretion processes , during which the mass of the central object is small , self - gravity plays an important role .
the influence of self - gravity has been sometimes invoked to explain the flat infrared spectrum of some t tau stars ( adams , lada & shu @xcite ; duschl , strittmatter & biermann @xcite ) . however , current estimates for the mass accretion rates in these systems ( typically @xmath118 ) so far have discouraged interpretations of the shape of the spectrum in terms of self - gravitating , optically - thick disks . on the other hand
, observational evidence suggests that , in some cases , the currently accepted models may be inadequate to describe the accretion flow of some agn .
for example , in the case of the seyfert galaxy , deviations from keplerian rotation have been observed ; when the standard shakura & sunyaev ( @xcite ) model is applied to this case , the values of the stability parameter @xmath0 in the outer disk turn out to be exceedingly low ( kumar @xcite ) .
in fact , the energy balance in standard disks , for which viscous dissipation is exactly balanced by radiative heating , leads to a @xmath0 profile rapidly decreasing outwards . in this paper
, we have focused on the problem of the energy budget in self - gravitating accretion disks , with the ultimate goal of providing a framework to describe the emission and the spectrum associated with such disks .
we have adopted the view that the effects of self - gravity on the energy balance are not only related to viscous dissipation , but that there are also processes , such as the self - regulation of jeans instability , that should give a significant contribution to the heating of the disk , at least in its outer parts . by referring to conditions applicable to some agn
, we have then shown that a thermally stable , non - advective , optically thin equilibrium solution is possible in the region dominated by self - gravity .
in contrast with other optically thin solutions ( shapiro , lightman & eardley @xcite ) , our solution is thermally stable as a result of the strong dependence of the heating term related to self - gravity on the parameter @xmath0 .
we have also found an unstable solution , which is not self - gravitating ( @xmath77 ) , and we argue that it is a modification of the sle solution . the value of @xmath0 on the thermally stable solution is always close to unity , and it is slightly dependent on radius .
the temperature in the outer disk is constant and depends on the mass accretion rate @xmath33 .
the structure of the disk , as long as the energy balance is dominated by self - gravity , turns out to depend only weakly on the detailed radiation processes involved .
this has been shown by taking into account two different radiation processes : pure _ bremsstrahlung _ and the case of a cooling function dominated by line absorption and emission ( sarazin & white @xcite ) , which is best suited for the low temperature conditions expected in the outer disk . the dependence on the radiation processes is only weak because of the self - regulation mechanism : whatever the radiation process , the disk readjusts , so as to keep @xmath0 close to marginal stability . for simplicity , in this paper we have focused on conditions relevant to the region of the disk where the transition to the self - gravitating regime occurs .
further out , one should include the presence of additional cooling mechanisms , not considered here , and especially the effects related to the deviations from a keplerian rotation curve .
the latter effects have been widely discussed in a previous paper ( bertin & lodato @xcite ) .
adams , f. c. , lada , c. j. , shu , f. h. , 1987 , apj , 312 , 788 balbus , s.a .
, papaloizou , j.c.b . , 1999 , apj , 521 , 650 bardou , a. , heyvaerts , j. , duschl , w.j . , 1998 , a&a , 337 , 966 bertin , g. , 1997 , apj , 478 , l71 bertin , g. , romeo , a. b. , 1988 , a&a , 195 , 105 bertin , g. , lodato , g. , 1999 , a&a , 350 , 694 duschl , w. j. , strittmatter , p. a. , biermann , p. l. , 2000 , a & a , 357 , 1123 kolykhalov , p. i. , sunyaev , r. a. , 1980 , sov .
lett . , 6 , 357 kumar , p. , 1999 ,
apj , 519 , 599 laughlin , g. , bodenheimer , p. , 1994 ,
apj , 436 , 335 lin , d. n. c. , pringle , j. e. , 1987 , mnras , 225 , 607 lin , d. n. c. , pringle , j. e. , 1990 , apj , 358 , 515 meyer , f. meyer - hofmeister , e. , 1981 , a&a , 104 , l10 narayan , r. , popham , r. , 1993 , nat , 362 , 820 narayan , r. , yi , i. , 1994 , apj , 428 , l13 narayan , r. , yi , i. , 1995 , apj , 452 , 710 paczyski , b. , 1978 , acta astronomica , 28 , 91 piran , t. , 1978 , apj , 221 , 652 pringle , j. , 1981 , ara&a , 19 , 137 quataert , e. , chiang , e. i. , 2000 , apj , 543 , 432 raymond , j. c. , cox , d. p. , smith , b. w. , 1976 , apj , 204 , 290 rybicki , g. b. , lightman , a. p. , 1979
, `` radiative processes in astrophysics '' , wiley & sons , new york sarazin , c. l. , white , r. e. , 1987 , apj , 320 , 32 shakura , n.j . ,
sunyaev , r.a . , 1973 ,
a&a , 24 , 337 shapiro , s. l. , lightman , a. p. , eardley , d. m. , 1976 , apj , 204 , 187 ( sle ) yuan , f. , 1999 , apj , 521 , l55 | in the dynamics of accretion disks , the presence of collective effects associated with the self - gravity of the disk is expected to affect not only the momentum transport , but also the relevant energy balance equations , which could differ substantially from the non - self - gravitating case . here
we follow the picture that , when the disk is sufficiently cold , the stirring due to jeans - related instabilities acts as a source of effective heating .
the corresponding reformulation of the energy equations allows us to : ( _ i _ ) demonstrate how self - regulation can be established , so that the stability parameter @xmath0 is maintained close to a threshold value , with weak dependence on radius ; ( _ ii _ ) rediscuss the opacity properties in the self - gravitating regime . in particular ,
we show that , if cooling is dominated by _ bremsstrahlung _ , an optically thin stationary accretion solution is thermally stable , even in the non - advective case , provided the disk is self - gravitating .
the details of the cooling function have little effect on the structure of such accretion disk , which is in any case brought by self - gravity to self - regulate .
this condition of self - gravitating accretion is expected to be appropriate for the outer regions of many disks of astrophysical interest . with the reformulation of the energy equations described in this paper we have also secured : ( _ iii _ ) a starting point for the study of the emission properties of self - gravitating accretion disks ; ( _ iv _ ) a tool to analyze the structure of the transition region , where the disk becomes self - gravitating . |
distant powerful radio sources represent excellent targets to study evolutionary processes related to massive galaxies and their associated surrounding structures at high redshifts ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the most effective method for finding high - redshift radio galaxies is by selecting for ultra steep - spectrum radio sources ( uss , @xmath13 , @xmath1 ) .
a possible explanation of this empirical method is a @xmath2correction induced by a curvature of the radio spectra .
high redshift sources with concave radio spectra will have steeper spectral indices compared to those at low redshift @xcite . comparing the extremely steep spectral index sources associated with galaxies residing closest to the cluster centres , @xcite found an explanation that suggests that the steeper spectra can also be explained by pressure - confined radio lobes which have slow adiabatic expansion losses in high density environments .
high - redshift radio galaxies are extremely luminous , and spatially extended , compared to normal galaxies at similar redshifts .
the tight @xmath8 relation of radio galaxies in the hubble @xmath2 diagram suggest that these objects are amongst the most massive systems at each redshift @xcite . using the @xmath8 diagram and galaxy evolution models
, @xcite found that the typical hosts of radio galaxies correspond to the most massive elliptical galaxies with baryon masses @xmath14@xmath15 .
there is increasing new evidence that distant radio galaxies represent massive forming systems .
@xcite found more than 10 individual clumpy features , possible satellite galaxies in the process of merging , in a region of 50@xmath1640kpc around the radio galaxy mrc 1138262 at @xmath17 . from a large vlt program , searching ly@xmath18 emitters around a sample of @xmath19 radio galaxies , @xcite found that at least six of eight fields studied are overdense in ly@xmath18 emitters by a factor of 3 - 5 compared to the field .
some radio - galaxies , selected with the uss criterion , are surrounded by giant ly@xmath18 haloes .
the size of these gas structures ( 100200kpc ) are similar to the observed size of cd haloes in galaxies in nearby clusters @xcite . from a spatial cross - correlation analysis between uss at @xmath20 and surrounding @xmath21 galaxies
, @xcite found a comoving correlation length comparable to those obtained for clusters of galaxies with masses in the range @xmath22 in a cosmological n - body simulation . in this paper we present @xmath2band observations , high - resolution radio maps and optical spectroscopy for a subsample of southern ultra steep spectrum radio sources selected from the westerbork in the southern hemisphere ( wish , @xcite ) 352mhz catalog .
we also present optical spectroscopy of sources selected from the @xcite and wish uss samples , yielding 12 new redshifts , including four new radio galaxies at @xmath12@xmath233 these provide excellent targets for follow - up observations in this interesting redshift range .
the structure of this paper is organized as follows : section 2 describes the sample analyzed .
we describe the near - ir observations and data reductions in section 3 .
section 4 presents the near - ir identification and source extraction .
we describe optical spectroscopy observations in section 5 and discuss our results in section 6 .
finally , section 7 presents the main conclusions . in this work
we assume a standard @xmath24cdm model universe with cosmological parameters , @xmath25=0.3 , @xmath26=0.7 and a hubble constant of @xmath27100 km s@xmath28mpc@xmath28 .
the uss sample was selected from the 352 mhz westerbork in the southern hemisphere catalogue @xcite .
we used a selection criterion based on the radio properties of the sources , steep radio spectral index ( @xmath29 ) and small angular size ( @xmath30 , as measured on the vla maps ) .
all of the first and most of the second priority targets have been observed .
a detailed description of the radio observations and data reduction can be found in @xcite .
most of the targets selected for optical spectroscopy are selected from the uss samples of @xcite .
in order to obtain high resolution radio maps and accurate positions , we used the very large array ( vla ) of the national radio astronomy observatory in the hybrid bna - configuration at 1.4ghz .
the images were obtained on 1999 october 16 and 20 and consisted of l - band short subscans of 3 minutes each .
we used the standard data reduction recipes in the astronomical image processing software ( aips@xmath31 ) package , including self - calibration for phase corrections .
the uss fields were observed during two runs in 2000 march and 2001 january using the ohio state infrared imager / spectrometer ( osiris ) imager on the 4@xmath32meter v.m .
blanco telescope at the cerro tololo inter - american observatory ( ctio ) .
osiris is a 0.92.4 @xmath33 m camera with a 1024x1024 hawaii hgcdte ccd array .
we used a focal ratio of f/7 which resulted in a respective pixel size of 0161 .
individual frames were obtained as a co - addition of a number of single exposures with different integration times , in this way one effectively removes row / column related defects . for example , for sources with total exposures times of 1920s , we use 12 single exposures of 10s observed with a a 16@xmath32point dithering pattern .
this results in @xmath34 minutes on source .
we used similar procedure for the other images , doing small changes in the number of points of the dithering pattern and the exposure times for single observations during the observations , tip - tilt mirror corrections were made .
this resulted in typical fwhms on the individual pointing frames ( @xmath35 co - added ) of @xmath36 to @xmath37 .
the mean was around @xmath38 for the nights 2000 march 20 to 22 , and around @xmath37 on the 2001 january run .
a subsample of uss sources was observed in 1999 july with the cerro tololo ir imager ( cirim ) camera at the ctio 4@xmath32meter v.m .
blanco telescope .
the ccd detector is a 256@xmath16256 hgcdte nicmos 3 array with a pixel scale of 0414 using a focal ratio of f/7.5 .
for these observations we obtained a mean fwhm of @xmath39 .
seven sources undetected in the ctio images were observed with the infrared spectrometer and array camera ( isaac ; * ? ? ?
* ) on the very large telescope ( vlt ) at paranal , ( chile ) between 2002 april and september . a @xmath40 filter ( 2.0 - 2.3 @xmath41 ) was used.the pixel scale was 0148 per pixel .
all the images were obtained under optimum seeing conditions with fwhm in the range 07 and 10 , and a number of standard stars were observed during the same nights . in table 1 we present the log of the near - ir observations , with the observation dates , telescope / instrument used and the total exposure times for all the uss sample .
data were reduced within iraf , using the standard dimsum @xcite near - ir reduction package , including dark subtraction , flat fielding , sky - substraction , bad pixel masking , bright object masking , registration and summing .
vlt images were reduced with isaac pipeline recipes 5.4.2 using the command line utility esorex , which includes standard routines , including dark correction , flat field calibration , bad pixel cleaning , image correlation and recombination .
the astrometric calibration of the images was performed using the wcstools package @xcite , using the two micron all sky survey ( 2mass ) catalogue of point sources a reference catalog @xcite and the usno - a2.0 catalogue @xcite for fields with a few number of 2mass sources .
we estimate the uncertainty in the relative astrometry to be @xmath1004 - 06 .
the photometry for the ctio observations was calibrated using 2mass point sources detected in these fields . for vlt runs
, magnitudes were calibrated using nicmos near - infrared standards @xcite .
all standard star magnitudes were obtained using sextractor @xcite with an aperture diameter of @xmath42 , which proved to be adequate by monitoring the growth curve of all the measured stars . for the observations made with osiris and cirim at ctio the @xmath2band zero point was [email protected] and [email protected] ( for 1 count / second , integrated over the source ) , respectively . for the isaac / vlt observations we derive the zero - point based on observations of the near - ir photometric standard star s301-d taken from the list of @xcite .
the zero - point obtained was [email protected] .
the zero - points do not include the airmass term , because the airmass dependence in @xmath2band is small and we have observed all our objects with airmasses @xmath44 , with a mean of 1.06 .
because of the large uncertainties in the photometry , we did not correct the @xmath2band magnitudes for galactic extinction .
we have used the ned database , which are based on the @xmath45 values from the extinction maps of @xcite .
we found that most of the sources have values from 0.01 to 0.1 , which are negligible compared to magnitude errors .
we have identified the near - ir counterparts to the radio sources by overlaying the radio contour maps on to the @xmath2band images using the aips@xmath31 task kntr .
we measured the magnitudes of these identifications using sextractor .
the parameters were set such that , to be detected , an object must have a flux in excess of 1 - 1.5 times the local background noise level over at least @xmath46 connected pixels , according to the seeing conditions and image qualities . in order to improve the detections of faint sources , we smoothed the @xmath2band images using a circular gaussian of fwhm=1 - 2 pixels . in some cases
, we find that it is possible that the identification does not fall at the midpoint of the radio lobes , for radio sources with double or multiple components .
we measured magnitudes for all radio source identifications using sextractor and circular apertures with diameters of 20 , 40 and 80 .
we also obtained mag_best magnitudes in order to facilitate comparison with the literature @xcite . this is usually equal to mag_auto , but if the contribution of other source exceeds 10% , it is mag_isocor .
the uncertainties quoted represent our best estimates , which includes the zero - point uncertainty ( see table 2 ) .
table 2 lists our uss sample , the columns are : iau j2000 designation , spectral index obtained between 352 mhz and 1.4 ghz , largest angular size , as measured from radio maps taken with the vla , aperture magnitudes in 20 , 40 and 80 , sextractor mag_best magnitudes , j2000 position coordinates of the radio source and the @xmath2band identification counterpart , the half - light radius and class_star values of the @xmath2band identifications .
sextractor class_star parameter provides an indication of the likelihood of an object being a galaxy or a star . in the ideal cases a galaxy has class_star@xmath47 and a point source has class_star@xmath48 .
we find that all the uss sample analyzed in this work are consistent with extended sources ( galaxies ) rather than point sources ( stars or quasars ) .
we also considered studying the environments of these uss radio galaxy candidates .
the environments of twenty uss sources observed with ctio / osiris were analyzed in @xcite . in those images ,
only 7 radio galaxies were identified , finding a statistically significant signal of clustering . in @xcite ,
a similar analysis was performed on a wider field and with deeper data , finding a stronger result .
we therefore do not analyze the environments of the uss sources in the relatively shallow @xmath6band images from this paper .
during the past 6 years , we have obtained optical spectroscopy of 14 uss sources .
the targets were mostly selected from 4 different uss samples : the 3 sub - samples of de breuck et al .
( 2000 ) and the wish
nvss sample described in this paper . from these samples , we mostly selected the final targets based on the ra range when telescope time was available .
optical spectroscopy of a subsample of uss sources were carried with fors2 installed on the vlt unit telescope 4 yepun between 2000 october and 2001 may .
we used a similar data reduction strategy to @xcite .
we removed cosmic rays using the iraf task szap , after bias and flatfield correction .
we used the iraf task background in order to remove sky variations and we extracted the spectra using a width appropriate to contain all flux in the extended emission lines .
finally we calibrated the one dimensional spectra in wavelength and flux .
we used the procedures described in @xcite and in @xcite to determine the central wavelength , total line flux , deconvolved widths , and rest - frame equivalent widths ( with their related errors ) of all the emission and absorption lines detected in our spectra . for each object , we quote a single redshift , which is an average of all the lines , weighted by the uncertainties due to the line fitting and wavelength calibration . in table 4
we show the measured parameters of the emission / absorption lines in the spectra .
longslit observations were made on ut 2004 january 19 - 20 , using the low resolution imaging spectrometer ( lris ; * ? ? ?
* ) on keck i , with a slit width of 15 , under photometric conditions and 08 seeing .
the spectroscopic setup is described in table 3 .
data reduction was performed in iraf ; data were bias - subtracted , response - corrected , cosmic - ray - zapped using pieter van dokkum s lacosmic routine , sky - subtracted , and spectra extracted in a 15 aperture .
the resultant 1-d spectra were wavelength and flux calibrated using appropriate arcs and standard stars .
on ut 2006 december 14 and 15 , we obtained 3 spectra using the isis double beam spectrograph @xcite at the 4.2 m william herschel telescope ( wht ) at observatorio roque de los muchachos on the canary island of la palma .
conditions were photometric with 05 and 15 seeing during the first and second night , respectively .
we used a dichroic splitting the light at 5300 and a 15 wide slit .
the grisms used were the r300b in the blue arm and the r316r in the red arm .
none of the 3 objects were detected in the blue arm . in the red arm , only wn j0912@xmath321655 was clearly detected in 5400s of observing time .
wn j0604 - 2015 and wn j0610 + 6611 were not detected in 5400s and 7200s , respectively .
we extracted the red spectrum of wn j0912@xmath321655 with an aperture width of 24 .
we have determined redshifts of 12 out of 14 targets observed .
the two targets without redshifts were undetected in 1.5 to 2 hours on a 4 m telescope , and need long integration on 810 m class telescopes .
apart from tn j2009@xmath323040 , which is a @xmath12=3.158 quasar , all 11 targets are radio galaxies with redshifts from @xmath12=0.550 to @xmath12=3.837 .
five of the radio galaxies are at @xmath12@xmath232 , including three at @xmath12@xmath233 .
table 4 lists the emission line parameters of all targets .
the 10 new redshifts from the de breuck et al .
( 2000 ) uss sample raises the number of spectroscopic observations from that sample from 46 to 56 , and the number of known redshifts from 34 to 44 .
in this work we have excluded large radio sources because those are most likely foreground objects and not at high redshift . in figure 1
we present the correlation between largest angular size ( las ) and the @xmath6magnitudes obtained within an aperture of 4@xmath49 .
as can be seen most of the sources have small radio sizes with las @xmath4 8@xmath49 .
the vertical dashed line represent the expected @xmath6band magnitude of a @xmath50 source , indicating that high redshift sources have compact radio morphologies .
several authors have applied the small angular - size criterion in order to find high redshift radio galaxies @xcite .
there are a few cases of radio galaxies at high redshifts with large angular sizes ( 4c 23.56 at z=2.483 has las=53@xmath49 ) .
however , @xcite found that the introduction of a strong angular size bias such as las@xmath51 , increases the efficiency in selecting high redshift radio galaxies by a factor of two .
some radio galaxies have very faint or no @xmath2band counterpart presenting large deviation in the hubble @xmath8 diagram @xcite , and 20% of high - redshift radio galaxies fail to show optical emission lines in spectroscopy observations using the keck telescope @xcite .
this could be an indication that some high redshift radio galaxies contain great amounts of dust .
@xcite found an anti - correlation between dust emission ( measured from 850 @xmath33 m flux ) and the projected linear size , though @xcite did not confirm this using a larger sample of radio galaxies .
it is likely that the compact uss sample analyzed here may also contain a mix of galaxies either at very high redshift , or heavily obscured by dust .
magnitudes ( 4@xmath49 aperture ) . dashed line represent the expected @xmath6band magnitude of a @xmath50 source .
most of the sources have small radio sizes as indicated by the horizontal dotted line.,title="fig:",width=302 ] [ fig1 ] diagram using 64 kpc metric apertures.,title="fig:",width=302 ] [ fig2 ] we have estimated the redshift of the uss sample using linear regression and the hubble @xmath8 diagram of @xcite .
using aperture photometry in a 64kpc metric aperture , @xcite found @xmath524.633@xmath16log@xmath53 .
as we expect most radio galaxies to have @xmath54 , we have assumed @xmath55 , as 8@xmath49 corresponds to @xmath1064kpc at @xmath12=1 , which only a small dependence on redshift @xmath56 in our adopted cosmology .
figure 2 shows the expected redshift distribution for our uss sample .
the median predicted redshift is @xmath9=2.13 , which is higher than the predicted redshift obtained by @xcite for the sumss
nvss sample ( @xmath57 ) and the expected redshift obtained in the 6c@xmath58 ( @xmath59 ; * ? ? ?
* ) and 6c@xmath11 surveys ( @xmath60 ; * ?
* ) . in figure 4
we show the @xmath6band magnitude distribution of uss sources , measured in a 8diameter aperture .
we compare our results with those obtained in the sumss
nvss @xcite , 6c@xmath11 @xcite catalogues and data selected at 74 mhz taken from @xcite .
the mean @xmath6band magnitudes in an 8@xmath49 diameter aperture of the uss is @xmath7=18.58 , which is one magnitude fainter than those obtained by @xcite in the sumss nvss survey ( @xmath7=17.57 ) and by @xcite in the the 6c@xmath11 survey ( @xmath7=17.59 ) .
we find a similar result using best magnitudes .
our uss sample has @xmath61=18.74 , and from a similar study @xcite found @xmath61=17.97 .
the fainter @xmath2-band counterparts of the wish
nvss uss sources could be an indication that the host galaxies are located either at higher redshifts ( i.e. a more efficient selection ) , or in a very dense , dusty medium .
alternatively , the wish nvss uss radio galaxies are intrinsically less massive and hence less luminous than the ones in previous samples . this could be due to the fainter flux limits probed in this sample .
however , using _
spitzer _ rest - frame near - ir phometry , seymour et al .
2007 do not find a strong dependence of host galaxy mass on 3ghz radio luminisity , so we regard this explanation as unlikely given the small difference in radio luminosity between the uss samples . alternatively , the wish
nvss uss sources may be less massive and hence fainted in the observed @xmath6band
. this may be because wish
nvss sources have fainter radio fluxes than most previous uss samples .
such a radio power dependence in the @xmath8 diagram has been suggested by various authors @xcite .
however , using _ rest - frame _ near - ir imaging of a sample of 69 radio galaxies at @xmath56 observed with _
, seymour et al . (
2007 ) do not find a strong dependence of host galaxy mass on 3ghz radio luminisity , so we regard this explanation as unlikely given the small difference in radio luminosity between the uss samples .
band magnitude distribution of uss sources , measured in a 8diameter aperture ( shaded histogram ) .
solid and dashed lines represent the distribution of @xmath6band magnitude , measured in a 8diameter aperture for sources detected in the sumss nvss @xcite and 6c@xmath11 @xcite catalogues . dotted histogram represent uss sources selected at 74 mhz taken from @xcite . ,
title="fig:",width=302 ] [ fig3 ] we use the cats database of the special astronomy observatory @xcite to search for radio measurements in order to study the radio spectral energy distribution at low frequencies .
figure 3 shows the radio two - colour diagram , which compares the spectral indices at 74 - 352 mhz and 352 - 1400 mhz for 12 uss sources of the total sample ( filled circles ) .
fluxes at 74 mhz were obtained from the vla low - frequency sky survey @xcite .
the line indicates the relation for uss sources whose spectra follow a single power law from 74 to 1400 mhz .
open circles represent measurements obtained using the wenss
nvss sample at 325 mhz and 1.4ghz taken from @xcite .
we find a general agreement with those values .
only a single source in our sample is consistent with a straight spectrum , and it is clear that most of the uss sources have substantially flatter spectral indices between 74 and 352 mhz than between 352 and 1400 mhz .
the median spectral index obtained at 74 - 352 mhz is @xmath621.1 , while the median at high frequencies is @xmath631.5 .
the spectral flattening is in the range 0@xmath4@[email protected] .
unfortunately , we do not have redshifts for any of the 12 sources with 74mhz data .
wn j2007 - 1316 at @xmath12=3.837 is detected in the mauritius southern sky survey @xcite with @[email protected] .
this implies @[email protected]@xmath430.19 , compared to @[email protected]@xmath430.04 .
although the uncertainty in the spectral index is larger than for the 74 - 352 mhz data , this result suggests that the highest redshift radio galaxy confirmed in our sample does retain a straight radio spectrum through the lowest frequencies observed . as noted by @xcite @xmath1090% of uss sources selected from the sumss and nvss catalogues
show a straight radio spectra characterised by a single power law beyond @xmath101 ghz . at frequencies below 100 mhz spectral
curvature is much more common than at higher frequencies due principally to synchrotron self - absorption .
this suggests that very low frequency selected uss samples will be likely to be more efficient to find high redshift galaxies .
we present @xmath2band observations obtained with ctio at cerro tololo and with vlt at cerro paranal , high - resolution vla radio maps for a sample of 28 ultra steep spectrum radio sources selected from the wish catalog .
we also present optical spectroscopy from vlt / fors2 , keck / lris and wht / isis for 12 uss sources , mainly from the @xcite samples .
we find that most of our uss sources have small radio sizes and fainter magnitudes .
the mean @xmath6band counterpart magnitude measured in 8@xmath49 diameter aperture is @xmath7=18.58 , which is one magnitude fainter than those obtained on the literature .
the expected redshift distribution estimated using the hubble @xmath8 diagram has a mean of @xmath9=2.13 , which is higher than the predicted redshift obtained for the sumss
nvss sample and the expected redshift obtained in the 6c@xmath58 and 6c@xmath11 surveys . from a radio colour - colour diagram of a subsample of 12 sources ,
we find that all but one radio spectra flatten significantly below 352mhz .
we also find that the highest redshift source from this paper ( at @xmath12=3.84 ) does not show evidence for spectral flattening down to 151mhz . because most low redshift radio galaxies have a flatter spectra below 352 mhz , selecting uss sources at low frequencies is likely to be more efficient selection technique to find high redshift galaxies .
the forthcoming low frequency observatories such as lofar , lwa or ska will be enable to find large samples of low - frequency selected uss sources ( e.g. * ? ? ? * ) and will therefore open new perspectives on the nature and evolution of high redshift radio galaxies . | we present a sample of 40 ultra steep spectrum ( uss , @xmath0 , @xmath1 ) radio sources selected from the westerbork in the southern hemisphere ( wish ) catalog .
the uss sources have been imaged in @xmath2band at the cerro tololo inter - american observatory ( ctio ) and with the very large telescope at cerro paranal .
we also present vlt , keck and willian herschel telescope(wht ) optical spectroscopy of 14 targets selection from 4 different uss samples . for 12 sources
, we have been able to determine the redshifts , including 4 new radio galaxies at @xmath3 .
we find that most of our uss sources have predominantly small ( @xmath46 ) radio sizes and faint magnitudes ( @xmath2@xmath518 ) .
the mean @xmath6band counterpart magnitude is @xmath7=18.6 .
the expected redshift distribution estimated using the hubble @xmath8 diagram has a mean of @[email protected] , which is higher than the predicted redshift obtained for the sumss
nvss sample and the expected redshift obtained in the 6c@xmath11 survey .
the compact uss sample analyzed here may contain a higher fraction of galaxies which are high redshift and/or are heavily obscured by dust . using the 74 , 352 and 1400mhz flux densities of a sub - sample ,
we construct a radio colour - colour diagram .
we find that all but one of our uss sources have a strong tendency to flatten below 352mhz .
we also find that the highest redshift source from this paper ( at @xmath12=3.84 ) does not show evidence for spectral flattening down to 151mhz .
this suggests that very low frequency selected uss samples will likely be more efficient to find high redshift galaxies .
= cmsy9 scaled1 = cmsy6 scaled1 # 1to 0pt#1 surveys radio continuum : general radio continuum : galaxies galaxies : high - redshift |
flavor chaning neutral currents ( fcnc ) are highly suppressed in the standard model ( sm ) due to the gim mechanism @xcite . for this reason ,
the sm predicts very small rates for the branching ratios of top quark fcnc decays to an up type quark and a neutral gauge or a higgs bosons : @xmath0 .
many models beyond the sm predict a huge enhancement on the expected rate of top quark fcnc branching ratios by introducing new heavy particles which can contribute into the loops .
any observation of these decays would indicate new physics . in order to observe a deviation from the sm predictions independent of the specific new physics scenarios , effective lagrangian approach
is used to search for the top quark fcnc signal in all analysis that will be reviewed . following reference @xcite , the most general effective lagrangian describing the top quark fcnc interactions
can be written as follows : @xmath1 the parameters @xmath2 , @xmath3 , @xmath4 , @xmath5 , and @xmath6 define the anomalous coupling constants which are normalized as @xmath7 , @xmath8 , etc .
different measurements use different normalization of coupling constants in @xmath9 , making the comparison of the couplings not straightforward .
therefore , the limits on the anomalous couplings are based on the normalization of the coupling constants in their corresponding publications .
results on the branching ratios are more easily comparable . in the following the latest analyses of top quark fcnc interactions performed at cms @xcite , atlas @xcite , cdf @xcite and d0 @xcite
are summarised .
the discovery of a new boson with a mass around 125 gev in 2012 at the lhc @xcite compatible with the sm higgs boson , opens up the possibility of searching for anomalous top quark decays to an up type quark and a higgs boson . considering the relatively large production cross section of @xmath10 at the lhc ,
the atlas @xcite and cms @xcite collaborations have searched for @xmath11 in @xmath10 events .
the atlas collaboration has chosen the @xmath12 decay mode . despite its small branching ratio ( @xmath130.23 % ) ,
@xmath12 leads to a clean signature .
signal events corresponding to @xmath10 production , with one top quark decaying into a charm quark and a higgs boson were generated using protos 2.2 @xcite . two final states with the hadronic and leptonic decays of w boson from the remaining top quark decay
are analysed .
events are first required to have two well identified and isolated photon with @xmath14 greater than 40 gev and 30 gev for leading and subleading photon candidates . in hadronic channel , events are required to have at least four jets among which at least one is b - tagged . after the selection of four jets , one top - quark candidate is constructed from the two photons and one jet ( @xmath15 ) , and another top - quark candidate
is formed from the three remaining jets ( @xmath16 ) which are required to lie within certain mass windows ( @xmath17 gev and @xmath18 gev ) . in leptonic channel , events with exactly one electron or muon and with two or more jets which at least one is b - tagged are selected . after the selection of two jets , one top quark candidate is constructed from the two photons and one jet ( @xmath15 ) , and another top - quark candidate
is built from the remaining selected jet , the lepton and the neutrino ( @xmath19 ) which are required to lie within certain mass windows ( @xmath17 gev and @xmath20 gev ) .
the contributions of the sm higgs production ( resonant background ) , @xmath10 and w@xmath21 ( non - resonant backgrounds ) are estimated from the simulation .
non - resonant production of diphoton final states with several additional jets which is the dominated background in the hadronic selection is normalised to data .
non - resonant backgrounds are small once a @xmath10-like topology is requested .
the diphoton mass spectrum is fitted using a likelihood function for the individual search channels . in the hadronic channel , the @xmath22 distribution
is fitted using 4.7 fb@xmath23 of 7 tev and 20.3 fb@xmath23 of 8 tev data . in the leptonic channel ,
the analysis is based on the event counting in two the @xmath22 regions using 8 tev data .
no significant signal is observed and an upper limit at the 95% cl is set on the @xmath24 branching ratio to be 0.79% .
the cms search for @xmath24 is a reinterpretation of an inclusive multilepton analysis @xcite and a search involving lepton and photon final states @xcite conducted by cms collaboration in terms of fcnc parameters .
in addition to diphoton decay mode of the higgs boson , leptonic final state decay modes are also considered .
these include h@xmath25 ww@xmath26 @xmath27 , h@xmath28 and h@xmath25 zz@xmath26 @xmath29 .
signal events were generated with madgraph5 event generator @xcite .
multilepton candidate events are categorized into mutually exclusive search channels to maximize the overall search sensitivity .
events are classified by the presence of a hadronic @xmath30 , a b - jet , an opposite - sign same - flavor ( ossf ) dilepton pair in which the invariant mass can be inside , above or bellow z boson mass .
the events are further classified by the missing transverse energy and hadronic activities which is measured by the scalar sum of the selected jets .
diphoton candidate events must have one lepton of any flavor and two well isolated photon .
the events are subdivided into four e@xmath31 bins with lower edges at 0 , 30 , 50 , and 100 gev .
the signal predominantly populates channels that have three leptons ( no hadronic @xmath30 ) , a b - jet , and no ossf pair or an ossf pair off z , as well as diphoton channels with a b - tag .
no significant excess is observed .
a counting experiment is done with several channels to compute the limit .
the upper limit on br(@xmath24 ) obtained from the combination of the multilepton channels with diphoton channels is 0.56% .
the fcnc - induced couplings of the type @xmath32 have been explored at the tevatron for the first time by looking at the anomalous top decays in @xmath10 events @xcite . considering the statistical uncertainty as the main source of systematic uncertainties , the upper limit at 95% cl
is set on the branching fraction br(@xmath33)@xmath34 . in @xmath35 collisions at lep and @xmath36 collisions at hera
, the fcnc @xmath32 have been probed through the single top production with no evidence of any signal @xcite .
r0.4 the cms collaboration conducted a search for fcnc in the production of single top quark in association with a photon @xcite . considering the muonic decay of w boson from top quark decay ,
the signal signature is the presence of a high @xmath37 photon , a b - jet , a muon and missing transverse energy .
the signal samples are generated with protos @xcite .
the backgrounds with a prompt photon arise mainly from w@xmath21+jets and z@xmath21+jets , and the backgrounds with a fake photon are dominated with w+jets and @xmath10 processes
. events with exactly one well isolated photon with @xmath38 gev and muon with @xmath39 gev are selected .
events with two or more b - tagged jets are vetoed to reduce the @xmath10 contributions .
finally a top mass window is defined to reduce the contribution of sm backgrounds without top quark .
a boosted decision tree ( bdt ) is constructed to separate signal from sm background .
the contribution of the w+jets and w@xmath21+jets backgrounds is estimated from data and other from simulation . the data is well described by sm prediction and there is no evidence for the signal .
upper limit on br(@xmath40 ) is 0.0161% .
r0.4 among fcnc top quark decays , @xmath41 is very difficult to distinguish from generic multijet production via quantum chromodynamics ( qcd ) .
it has therefore been suggested to search for fcnc couplings in anomalous single top - quark production .
the anomalous @xmath42 interactions can induce various rare processes at hadron colliders .
the cdf and atlas collaborations have chosen the production of a single top quark without any additional particle @xcite . in figure [ feynman ] ( a ) , the related feynman diagram is shown .
the signature of @xmath43 process is different from sm processes .
first , w and b from top decay tend to be back to back in the azimuthal plane .
secondly , w is boosted and its decay products have smaller opening angle .
lastly , top is produced three times more than antitop .
signal events are generated with toprex @xcite , protos @xcite and metop @xcite event generators for the cdf , atlas 7 tev and atlas 8 tev analysis respectively . for an efficient background rejection which are dominated by w+jets and qcd multijets , bayesian neural - network ( bnn )
is used by both collaborations .
no evidence for such a process is observed
. a binned maximum likelihood fit of the bnn - output distribution is performed to set an upper limit on the fcnc single top quark production cross section .
the d0 and cms experiments searched for fcnc production of single top quark and a light quark or gluon , a topology similar to sm t - channel single top quark production .
a representative feynman diagram is shown in figure [ feynman ] ( b ) .
the single top quark final state without extra jets that was explored by the cdf and atlas collaborations is not r0.62 lllclll exp & @xmath44tev & @xmath45&br & ( q = u)% & ( q = c)% & ref + atlas & 7@xmath468 & 25 & & & @xcite + cms & 8 & 19.5 & & & @xcite + & & & & & + cdf & 1.8 & 0.11 & & & @xcite + cms & 8 & 19.1 & & 0.0161 & 0.182 & @xcite + & & & & & + cdf & 1.96 & 2.2 & & 0.039 & 0.57 & @xcite + d0 & 1.96 & 2.3 & & 0.02 & 0.39 & @xcite + cms & 7 & 4.9 & & 0.56 & 7.12 & @xcite + cms & 7 & 4.9 & & 0.035 & 0.34 & @xcite + atlas & 8 & 14.2 & & 0.0031 & 0.016 & @xcite + cdf & 1.96 & 1.9 & & & @xcite + d0 & 1.96 & 4.1 & & & @xcite + cms & 7 & 4.9 & & 0.51 & 11.40 & @xcite + atlas & 7 & 2.1 & & & @xcite + cms & 7@xmath468 & 24.7 & & & @xcite + considered due to its different final state topology and significantly smaller signal event yield . the presence of exactly one isolated charged lepton ( e or @xmath47 for d0 and @xmath47 for cms ) , one light - flavour jet in the forward region , one b - jet and significant missing transverse momentum due to the presence of a neutrino are the signatures of this final state .
the main backgrounds are from w+jets production , including w+c - quarks and w+b - quarks .
the qcd can also contribute in final selection with a fake lepton .
to suppress qcd background , a dedicated bnn is used in cms analysis .
both experiments have trained a bnn for the discrimination of fcnc signal from the sm processes . since the data are consistent with the background expectation , the 95% upper limits is set on the fcnc cross sections and couplings using bnn distributions .
the obtained upper limits with the experimental conditions are summarised in table [ summaryreport ] .
the limits on the @xmath48 obtained by atlas collaboration is the most stringent limits to date .
the d0 @xcite , cms @xcite and atlas @xcite collaborations made searches for @xmath49 fcnc interactions in @xmath50 events in which either the top or antitop quark has decayed into a z boson and a quark , while the remaining top or antitop quark decayed to wb .
only leptonic decays of the z and w bosons were considered , yielding trilepton final states .
this mode provides a distinct signature with low sm background , albeit suffers from statistical uncertainties .
the backgrounds include events with three real final - state charged leptons mainly from wz and zz processes , as well as events with at least one fake leptons dominated by z+jets and @xmath50 processes .
a similar search is done by cdf collaboration @xcite while the hadronic decays of the z boson is considered .
the signal events are generated with pythia @xcite in cdf and d0 analysis and toprex @xcite and madgraph @xcite for atlas and cms analysis respectively .
the data are consistent with the sm background prediction , and upper limit on the branching fraction @xmath51 are set which are reported in table [ summaryreport ] .
the signature of a single top quark production in association with a z boson is used by cms collaboration to search for @xmath49 and @xmath42 anomalous interactions simultaneously , via diagrams ( c ) and ( d ) in figure [ feynman ] @xcite .
the signal events are discriminated from sm background events using bdt .
no fcnc signal is observed and the upper bounds on the anomalous branching r0.42 fractions are derived which is shown in table [ summaryreport ] .
the @xmath42 upper limits provide an interesting cross - check in a different physics process and the @xmath49 upper limits can be combined with other searches to improve the current limits .
the anomalous production and decay of top quark are two complementary channels which have been considered by d0 , cdf , atlas and cms collaborations in order to search for anomalous fcnc top couplings .
no excess over the sm expectation is found and the strength of the anomalous couplings are limited consequently . although the sm predicts top quark anomalous branching ratios many order of magnitude below the current experimental limits , experiments are closing to the regions which are predicted by some beyond sm models as shown in figure [ 2dplot ] in the @xmath52 vs @xmath51 plane .
s. l. glashow , j. iliopoulos and l. maiani , phys .
d * 2 * , 1285 ( 1970 ) .
cms collaboration , jinst * 3 * , s08004 ( 2008 ) .
atlas collaboration , jinst * 3 * , s08003 ( 2008 ) .
cdf collaboration , phys .
d * 71 * , 032001 ( 2005 ) [ hep - ex/0412071 ] .
d0 collaboration , nucl .
instrum .
meth . a * 565 * , 463 ( 2006 ) [ physics/0507191 [ physics.ins-det ] ] .
atlas collaboration , phys .
b * 716 * , 2 ( 2012 ) cms collaboration , cms - pas - hig-13 - 034 .
j. a. aguilar - saavedra , nucl .
b * 821 * , 215 ( 2009 ) [ arxiv:0904.2387 [ hep - ph ] ] .
cms collaboration , phys .
d * 90 * , 032006 ( 2014 ) [ arxiv:1404.5801 [ hep - ex ] ] .
cms collaboration , cms - pas - hig-13 - 025 .
f. maltoni and t. stelzer , jhep * 0302 * , 027 ( 2003 ) [ hep - ph/0208156 ] .
cdf collaboration , phys .
lett . * 80 * , 2525 ( 1998 ) .
zeus collaboration , phys .
b * 708 * , 27 ( 2012 ) [ arxiv:1111.3901 [ hep - ex ] ] .
h1 collaboration , phys .
b * 678 * , 450 ( 2009 ) [ arxiv:0904.3876 [ hep - ex ] ] .
delphi collaboration , phys .
b * 590 * , 21 ( 2004 ) [ hep - ex/0404014 ] .
cms collaboration , association with a photon , cms - pas - top-14 - 003 .
cdf collaboration , phys .
lett . * 102 * , 151801 ( 2009 ) [ arxiv:0812.3400 [ hep - ex ] ] .
atlas collaboration , atlas - conf-2013 - 063 , atlas - com - conf-2013 - 064 .
s. r. slabospitsky and l. sonnenschein , comput .
commun . *
148 * , 87 ( 2002 ) [ hep - ph/0201292 ] .
r. coimbra , a. onofre , r. santos and m. won , eur .
j. c * 72 * , 2222 ( 2012 ) [ arxiv:1207.7026 [ hep - ph ] ] .
d0 collaboration , phys .
b * 693 * , 81 ( 2010 ) [ arxiv:1006.3575 [ hep - ex ] ] .
cms collaboration , cms - pas - top-14 - 007 .
cms collaboration , cms - pas - top-12 - 021 .
cdf collaboration , phys .
* 101 * , 192002 ( 2008 ) [ arxiv:0805.2109 [ hep - ex ] ] .
d0 collaboration , phys .
b * 693 * , 81 ( 2010 ) [ arxiv:1006.3575 [ hep - ex ] ] .
atlas collaboration , jhep * 1209 * , 139 ( 2012 ) [ arxiv:1206.0257 [ hep - ex ] ] .
cms collaboration , phys .
lett . * 112 * , 171802 ( 2014 ) [ arxiv:1312.4194 [ hep - ex ] ] .
t. sjostrand _ et al .
_ , comput .
phys . commun .
* 135 * , 238 ( 2001 ) [ hep - ph/0010017 ] . | an overview of searches for top quark flavour changing neutral currents ( fcnc ) from the cdf , d0 , atlas and cms collaborations is presented .
the fcnc interactions of top quarks are probed through the anomalous decays and production channels .
no clear evidence for the presence of the signal is observed in none of the searches and the upper limits are set at 95% confidence level ( cl ) on the cross sections , anomalous couplings and the branching ratios . |
very recently , the besiii collaboration reported a charged charmoniumlike structure by studying the @xmath6 invariant mass spectra in @xmath7 at @xmath8 gev @xcite .
this newly observed charged structure around 3.9 gev is referred to @xmath9 , which was confirmed by the belle collaboration @xcite .
later , xiao _ et al . _ also announced the observation of @xmath9 with a 6@xmath10 significance by analyzing 586 pb@xmath11 data collected with the cleo - c detector at @xmath12 @xcite .
as indicated in ref .
@xcite , this charged enhancement structure near the @xmath1 threshold was predicted in refs .
@xcite before the observation of @xmath9 . among these pioneering theoretical investigations ,
the _ initial single pion emission _ ( ispe ) mechanism was applied to study the hidden - charm dipion decays of higher charmonia , where there exist the charged charmoniumlike structures near the @xmath1 and @xmath2 thresholds in the distributions of the @xmath13 , @xmath14 and @xmath15 invariant mass spectra @xcite .
the ispe is a decay mechanism peculiar to the hidden - charm dipion decays of higher charmonia @xcite and the hidden - bottom dipion decays of higher bottomonia @xcite .
the ispe mechanism was first proposed to explain why the charged bottomoniumlike structures @xmath16 and @xmath17 exist in the @xmath18 @xmath19 and @xmath20 ( @xmath21 ) invariant mass spectra of the hidden - bottom decays of @xmath22 , which was reported by belle @xcite .
the authors adopted the ispe mechanism further to study hidden - bottom decays of @xmath23 @xcite , and predicted the charged bottomoniumlike enhancement structures similar to @xmath16 and @xmath17 .
recently , we have developed the ispe mechanism to propose the _ initial single chiral particle emission _
( ische ) mechanism , where the charged charmoniumlike structures with the hidden - charm and open - strange channel were predicted @xcite .
although we have already presented abundant phenomena of charged charmoniumlike / bottomoniumlike structures in the hidden - charm / hidden - bottom dipion decays of higher charmonia / bottomonia , it is not the end of the story .
the fact that besiii confirmed our prediction @xcite again provokes our interest into predicting more novel charged charmoniumlike structures in other processes , which can be accessible at future experiments , especially at besiii , belle , and the forthcoming belleii . in this work ,
we notice the similarity between properties of @xmath0 and @xmath24 . as the first charmoniumlike structure reported in the @xmath25 process ,
@xmath24 was observed by babar @xcite .
later , both the cleo and belle collaborations confirmed @xmath24 in the @xmath25 process @xcite .
@xmath0 was reported by babar @xcite after analyzing the @xmath26 invariant mass spectrum of @xmath27 , which was confirmed by belle @xcite . at present , @xmath9 was observed in the @xmath13 invariant mass spectrum of @xmath28 .
the similarity between @xmath0 and @xmath24 just mentioned above provides us interest in studying whether a similar charged charmoniumlike structure can be found in @xmath29 and other hidden - charm dipion decays of @xmath0 . to answer to this question ,
we need to carry out a more in - depth study of the hidden - charm dipion decays of @xmath0 .
this paper is organized as follows .
after the introduction , we present the hidden - charm dipion decays of @xmath0 discussed above and the calculation of the corresponding decay amplitudes and the differential decay widths . in sec .
[ sec3 ] , the numerical results are given .
the last section devoted into a short summary .
the involved hidden - charm dipion decays of @xmath0 include : @xmath30_{d^{(*)}}^\mp\to \left\ { \begin{array } { l } j/\psi\pi^+\pi^-\\ \psi(2s ) \pi^+\pi^-\\h_c(1p)\pi^+\pi^- \end{array } \right . , \label{eq1}\end{aligned}\ ] ] where @xmath31 is an initial single pion directly emitted from @xmath0 decay .
with the pion emission , @xmath0 transits into intermediate @xmath32 . then , via the exchanged @xmath33 meson , the intermediate @xmath32 can dissolve into a charmonium plus a pion . in eq .
( [ eq1 ] ) , the subscript and superscript of @xmath34_{d^{(*)}}^\mp$ ] denote the total charge of the intermediate charmed mesons and the corresponding exchanged charmed meson .
these processes can be represented by the typical diagrams shown in fig .
[ fig : typical ] .
here , the initial single pion emission plays an important role to these processes , i.e. , because the pion carries off some of the @xmath35 value , the intermediate @xmath33 and @xmath36 with low momenta can easily interact with each other and transit into final states .
these typical diagrams reflect main physical picture of the ispe mechanism @xcite . [ cols="^,^ " , ] the results listed in fig .
[ fig : result ] indicate : \(1 ) for @xmath29 , there are two sharp peaks existing in the @xmath3 invariant mass spectrum .
the one is around the @xmath1 threshold while another one is close to the @xmath2 threshold ( see fig .
[ fig : result ] ( d ) and ( g ) ) .
there are two broad structures , which are the corresponding reflections of these two sharp structures .
\(2 ) the line shapes of @xmath37 invariant mass spectrum of @xmath38 show different behavior from ( 1 ) .
the intermediate @xmath39 can not result in sharp peak structures , instead there is a smooth curve ( see fig .
[ fig : result ] ( e ) ) .
when including only @xmath2 as the intermediate state , we find that a small peak exists although the peak near the @xmath2 threshold and its reflection form one peak as shown in fig .
[ fig : result ] ( h ) .
\(3 ) for the @xmath40 process , one irregular peak exists in the @xmath5 invariant mass spectrum if only considering the intermediate @xmath39 contribution ( see diagram fig .
[ fig : result ] ( f ) ) .
the intermediate @xmath2 contribution can produce a sharp peak near the @xmath2 threshold ( see fig .
[ fig : result ] ( i ) ) .
its reflection is a broad structure , which is around the @xmath1 threshold . comparing with the sharp peak structure shown in fig .
[ fig : result ] ( d ) , the peak in fig . [ fig : result ] ( f ) is more obvious .
in addition , our calculation also shows that the intermediate @xmath41 contribution to @xmath42 can not result in any enhancement structure close to the @xmath41 threshold in the @xmath43 invariant mass spectrum ( see fig .
[ fig : result ] ( a)-(c ) ) .
since we only focus on the line shapes as shown in fig .
[ fig : result ] , the uncertainties of the coupling constants can not result in the changes of our results . as for the @xmath44 process via the intermediate state @xmath1 as an example , a common factor @xmath45 can be extracted from the total decay amplitude , which is due to the constraint from eqs .
( [ couple1])-([couple3 ] ) .
thus , the uncertainties of gauge coupling constants @xmath46 and @xmath47 can not change the corresponding line shapes in the @xmath13 invariant mass spectrum . indeed ,
although in our calculation we use the relations of the coupling constants that are obtained in the heavy quark limit , our qualitative conclusion does not depend on these relations but on the ispe mechanism .
heavy quarkonium physics is an intriguing and active research field full of challenges and opportunities @xcite .
a newly observed charged charmoniumlike structure @xmath9 was reported by besiii @xcite and confirmed by belle @xcite in the @xmath28 process . before the besiii s observation
, we have explored the hidden - charm dipion decays of @xmath24 and have predicted charged charmoniumlike structures by the ispe mechanism @xcite . stimulated by the similarity between @xmath0 and @xmath24 and the recent experimental observation of @xmath9 by the besiii and belle collaborations , we study the hidden - charm dipion decays of charmoniumlike state @xmath0 in the present work .
our calculation shows that there exist charged structures in the @xmath3 invariant mass spectrum , which are near the @xmath1 and @xmath2 thresholds ( see fig .
[ fig : result ] ( a ) and ( d ) ) .
a peak structure around the @xmath2 threshold appears in the @xmath4 invariant mass spectrum ( see fig . [
fig : result ] ( e ) ) .
in addition , we can also find a sharp peak structure in the @xmath5 invariant mass spectrum near the @xmath2 threshold ( see fig .
[ fig : result ] ( f ) ) . experimental search for these charged charmoniumlike structures predicted near the @xmath1 and @xmath2 thresholds is an interesting research topic , which will be helpful to further test our predictions presented in this work .
what is more important is that the ispe mechanism existing in the hidden - charm / hidden - bottom dipion decays of higher charmonium / bottomonium can be tested again .
besiii , belle , and forthcoming belleii will be a good platform to carry out the experimental study on these hidden - charm dipion decays of charmoniumlike state @xmath0 .
we would like to have more experimental progresses in this field .
this project is supported by the national natural science foundation of china under grants no .
11222547 , no . 11175073 , no .
11005129 and no . 11035006 , the ministry of education of china ( fanedd under grant no .
200924 , srfdp under grant no .
20120211110002 , the fundamental research funds for the central universities ) , the fok ying tung education foundation ( grant no .
131006 ) , and the west doctoral project of chinese academy of sciences .
m. ablikim _ et al . _
[ besiii collaboration ] , arxiv:1303.5949 [ hep - ex ] .
z. q. liu _ et al .
_ [ belle collaboration ] , arxiv:1304.0121 [ hep - ex ] .
t. xiao , s. dobbs , a. tomaradze and k. k. seth , arxiv:1304.3036 [ hep - ex ] .
d. -y .
chen and x. liu , phys .
d * 84 * , 034032 ( 2011 ) [ arxiv:1106.5290 [ hep - ph ] ] .
z. -f .
sun , j. he , x. liu , z. -g .
luo , s. -l .
zhu and , phys .
d * 84 * , 054002 ( 2011 ) ; chin .
c * 36 * , 194 ( 2012 ) [ arxiv:1106.2968 [ hep - ph ] ] .
a. ali , c. hambrock and w. wang , phys .
d * 85 * , 054011 ( 2012 ) [ arxiv:1110.1333 [ hep - ph ] ] .
d. -y . chen and x. liu ,
d * 84 * , 094003 ( 2011 ) [ arxiv:1106.3798 [ hep - ph ] ] .
d. -y .
chen , x. liu and t. matsuki , phys .
d * 84 * , 074032 ( 2011 ) [ arxiv:1108.4458 [ hep - ph ] ] .
i. adachi [ belle collaboration ] , arxiv:1105.4583 [ hep - ex ] .
d. -y .
chen , x. liu and t. matsuki , phys .
lett . *
110 * , 232001 ( 2013 ) [ arxiv:1303.6842 [ hep - ph ] ] .
b. aubert _ et al . _ [ babar collaboration ] , phys .
lett . * 95 * , 142001 ( 2005 ) [ arxiv : hep - ex/0506081 ]
. t. e. coan _ et al .
_ [ cleo collaboration ] , phys . rev .
lett . * 96 * , 162003 ( 2006 ) [ arxiv : hep - ex/0602034 ] .
q. he _ et al . _
[ cleo collaboration ] , phys .
d * 74 * , 091104 ( 2006 ) [ arxiv : hep - ex/0611021 ] . c. z. yuan _ et al .
_ [ belle collaboration ] , phys .
lett . * 99 * , 182004 ( 2007 ) [ arxiv:0707.2541 [ hep - ex ] ] .
b. aubert _ et al . _
[ babar collaboration ] , phys .
lett . * 98 * , 212001 ( 2007 ) [ arxiv : hep - ex/0610057 ] . j. p. lees _ et al . _ [ babar collaboration ] , arxiv:1211.6271 [ hep - ex ] .
x. l. wang _ et al . _ [ belle collaboration ] , phys . rev
. lett .
* 99 * , 142002 ( 2007 ) [ arxiv:0707.3699 [ hep - ex ] ] . y. s. oh , t. song and s. h. lee , phys . rev .
c * 63 * , 034901 ( 2001 ) [ arxiv : nucl - th/0010064 ] .
r. casalbuoni , a. deandrea , n. di bartolomeo , r. gatto , f. feruglio and g. nardulli , phys .
rept . *
281 * , 145 ( 1997 ) [ arxiv : hep - ph/9605342 ] .
p. colangelo , f. de fazio and t. n. pham , phys .
b * 542 * , 71 ( 2002 ) [ arxiv : hep - ph/0207061 ] .
j. beringer _
[ particle data group collaboration ] , phys .
d * 86 * ( 2012 ) 010001 .
h. -y .
cheng , c. -k .
chua and a. soni , phys .
d * 71 * ( 2005 ) 014030 [ hep - ph/0409317 ] .
d. -y .
chen , x. liu and t. matsuki , arxiv:1303.6842 [ hep - ph ] .
n. brambilla , s. eidelman , b. k. heltsley , r. vogt , g. t. bodwin , e. eichten , a. d. frawley and a. b. meyer _ et al .
_ , eur .
j. c * 71 * ( 2011 ) 1534 [ arxiv:1010.5827 [ hep - ph ] ] .
n. brambilla _ et al .
_ [ quarkonium working group collaboration ] , hep - ph/0412158 .
n. brambilla , a. pineda , j. soto and a. vairo , rev
. mod .
* 77 * ( 2005 ) 1423 [ hep - ph/0410047 ] . | studying the hidden - charm dipion decays of the charmoniumlike state @xmath0 , we show that there exist charged charmoniumlike structures near @xmath1 and @xmath2 thresholds in the @xmath3 , @xmath4 and @xmath5 invariant mass spectra of the corresponding hidden - charm dipion decays of @xmath0 using the _ initial single pion emission mechanism_. we suggest to do further experimental study on these predicted charged enhancement structures , especially by besiii , belle , and the forthcoming belleii . |
quantum control @xcite and quantum transport ( qt ) @xcite play a prominent role in modern applications of quantum dynamics .
in particular , quantum networks ( qn ) are extensively used to investigate the phenomenon of energy propagation and its relation to quantum coherence .
paradigmatic examples are that of light harvesting complexes @xcite and nano devices @xcite .
this kind of investigation led to the discovery of noise - assisted transport ( nat ) @xcite , where the interaction with the environment helps to enhance transport efficiency when the system parameters are appropriately tuned . in the context of light harvesting complexes ,
there are investigations about the capability of the dynamics in promoting the manifestation of quantum correlations @xcite .
unfortunately , complete control over system state preparation and its evolution is not yet possible in real photosynthetic systems , let alone the use of characterization tools from quantum information science .
a natural step is then to ask how especially arranged qn of _ controlled _ quantum systems can be used to critically assess the role of quantum coherences and correlations in the phenomenon of energy transport @xcite . in this work ,
we are interested in this kind of investigation . in other words , we are interested in the _ active _ role of state preparations and external control over transport efficiency .
these situate our work in the context of quantum technologies where , instead of having a naturally occurring network , such as a photosynthetic complex , one can deliberately engineer a system , prepare its states , and drive it to study its response .
this is precisely the case of setups such as trapped ions , cavity or circuit quantum electrodynamic systems , just to name a few examples . in these systems ,
control is usually achieved by means of interaction with external fields . in particular , it has been recently shown that cleverly chosen external time dependent drivings can assist quantum tunneling between nodes in a qn @xcite . here , we employ this ideia to engineer a qn suitable to our purpose of studying how quantum correlations actively influence energy transport .
this article is organized as follows . in section [ model ] , we present the network model used in this work .
our results are presented in section [ results ] , where we carefully investigate the role of quantum correlations and environment in the efficiency of energy transport through the network . in section [ conc ] ,
we conclude our findings and present our final remarks .
the basic two - dimensional network element we are interested in is depicted in fig.[fig : fig1 ] ( top ) , where one can see the spatial distribution of the sites or nodes .
we say that site @xmath0
( @xmath1 ) is located at @xmath2 , where @xmath3 and @xmath4 are natural numbers ( zero included ) . by closely following the driving mechanism in @xcite , we consider the system hamiltonian to be @xmath5 with the total on - site energies subjected to external driving described by @xmath6\hat{\sigma}_{j}^{+}\hat{\sigma}_{j}^{-},\ ] ] and the coherent hopping given by @xmath7 where @xmath8 with @xmath9 positive integers @xcite , @xmath10 and @xmath11 two - level raising and lowering operators at site @xmath12 , respectively , @xmath13 and @xmath14 are basically the amplitude and angular frequency of the driving field acting on site @xmath12 , and @xmath15 the coherent transfer rate between the sites @xmath12 and @xmath16 .
more precisely , the product @xmath17 gives the strength of the interaction external drive - site @xmath18 and it depends monotonically on the field amplitude and frequency .
also , the external drive frequency is chosen to be site - dependent with the phases @xmath19 all these driving parameters and the energy ladder @xmath20 are controlled externally and provide a variable tool to design effective interactions in the network @xcite . for the sake of simplicity
, we will now choose the driving field controlled parameters such that @xmath21 and @xmath22 , for each site @xmath18 . for the network depicted in fig.[fig : fig1 ] ( top ) , we chose initially resonant sites @xmath23 and @xmath24 , with @xmath25 a positive integer , which forms an energy ladder implemented via the time - independent part of the external field .
in particular , we depicted the case @xmath26 in fig.[fig : fig1 ] ( top ) . from eq.([esc ] ) , it follows that @xmath27 , and @xmath28 .
also , by using eq.([phase ] ) , one finds that @xmath29 the hopping hamiltonian ( [ hop ] ) naturally appears in varied scenarios .
it may represent , for instance , dipole - dipole interaction among two - level atoms or molecules in free space @xcite .
another well known situation where hamiltonian ( [ hop ] ) appears is the dispersive interaction of two - level systems with a common bosonic mode @xcite .
the role of the driving in eq.([h0 ] ) is to suppress the coupling between particular pairs of sites .
our goal is to carefully choose the phases and amplitudes of the external driving field to design an effective regime where the transitions between sites @xmath30 and @xmath31 are suppressed . with this , we can study how energy injected in site @xmath32 arrive at site @xmath33 through indirect pathways @xmath34 and @xmath35 , as depicted in fig.[fig : fig1 ] ( bottom ) .
this is a situation found , for instance , in the description of the conduction of potassium ions in the kcsa channel @xcite .
this suppression is achieved as follows . by transforming the system hamiltonian ( [ ht ] ) to an interaction picture with respect to @xmath36 , and taking into account the condition @xmath24 ( @xmath25 a positive integer ) , a rotating wave approximation ( rwa ) can be performed to obtain @xmath37 with @xmath38 where @xmath39,\end{aligned}\ ] ] @xmath40 and @xmath41 with @xmath42 being the bessel function of first kind and order @xmath43 .
the rotating wave approximation used to obtain ( [ hint ] ) is valid only for @xmath44 @xcite .
now , we carefully look into the content of eq .
( [ hint ] ) . for the sake of simplicity ,
let us consider once again @xmath26 .
the dynamics of energy migration between sites @xmath0 and @xmath18 is ruled by @xmath45 ( see eq.([aux1 ] ) ) .
the magnitude of this quantity for the pair of diagonal sites @xmath32-@xmath33 and @xmath46-@xmath47 is plotted in fig .
[ fig : fig2 ] as a function of the driving parameters
. direct inspection of plots in fig [ fig : fig2 ] and the use of eq.([delta ] ) thus reveal that the choice @xmath48 leads to the sought suppression of hopping along those diagonal sites .
this is true regardless of the driving amplitude @xmath49 for the range of parameters considered .
it is this choice of @xmath50 and @xmath51 that will be used from now on .
we must now emphasize , however , that we will still keep using the full original ( non rwa ) time - dependent hamiltonian ( [ ht ] ) in the simulations .
the rwa argument was just used to understand how the external driving can effectively suppress hopping between particular pair of sites .
the dynamics of the qn depicted in fig.[fig : fig1 ] ( bottom ) , which is our system of interest , is ruled by the following master equation @xcite @xmath52+\mathcal{l}_{s}(\hat{\rho})+\mathcal{l}_{d}(\hat{\rho})+\mathcal{l}_{deph}(\hat{\rho}),\ ] ] where the _ source _ superoperator @xmath53 accounts for the incoherent input of energy into the system through site @xmath32 at a pump rate @xmath54 , the _ drain _
superoperator @xmath55 represents an incoherent loss of energy through site @xmath33 at a rate @xmath56 , and the dephasing superoperator @xmath57 destroys quantum coherence in the network , where @xmath58 is a site - dependent dephasing rate . in all these , @xmath59 denotes the anticommutator @xmath60 .
one of our goals is to study the contribution of each of these terms in the master equation to the dynamics of energy propagation in the qn depicted in fig.[fig : fig1 ] ( bottom ) .
moreover , we want to investigate it taking into account the presence of initial quantum correlations in the control sites @xmath46 and @xmath47 . an important figure of merit is how efficiently energy leaves the system through site @xmath33 which works as a drain ( rate @xmath61 ) .
this is quantified by the integrated population of site @xmath33 @xmath62 where @xmath63 is the occupation of site @xmath33 at instant @xmath64 .
this quantity is proportional to the _ transport efficiency _
@xmath65 through @xmath66 @xcite . in this work ,
we use the entanglement of formation ( eof ) to quantify bipartite entanglement between pairs of sites @xcite .
in addition to that , we also include other form of quantum correlation in our study , the so called quantum discord ( qd ) @xcite .
the former is interesting because it spots quantumness for a set of states which does not necessarily contains entanglement . in this sense
, the qd adds generality to our study .
the just presented formalism allows one to explore site - dependent dephasing scenarios .
however , from now on we will be adopting the same decoherence rates for all sites @xmath67 , a feasible choice for quantum technologies .
nevertheless , if one aims at studying transport in natural systems such as photosynthetic complexes , site - dependent dephasings should necessarily be taken into account in accordance with experimental observations and computational simulations .
we would like to start our investigations by considering two initial states , with same mean energy ( one excitation ) , but with different types of correlations .
the first state is @xmath68 where @xmath69 . for this initial preparation , sites @xmath32 and @xmath33 start in their ground state and they are not correlated with sites @xmath46 and @xmath47 , which share a bipartite maximally entangled state that contains one quantum of excitation .
experimentally , bipartite entangled states such as the one considered here has been generated in a variety of setups ranging from photons @xcite to massive particles @xcite . in the scenario defined by eq.([eq : rho0_init_ent ] ) , sites @xmath46 and @xmath47 are then quantum correlated . the second initial state to be considered in this subsection is @xmath70 where once again the sites @xmath32 and @xmath33 are initially in their ground state , but now sites @xmath46 and @xmath47 are just classically correlated in a maximally mixed state .
one can consider eq.([eq : rho0_init_mix ] ) as the limit of eq.([eq : rho0_init_ent ] ) when previous decoherence @xmath71 on the decoupled model @xmath72 had fully acted and completely destroyed the coherences i.e. , the nondiagonal terms in the basis @xmath73 . in fig .
[ fig : fig3 ] , we present the efficiency quantifier ( [ ind ] ) for an observation time @xmath74 ( interval of integration ) and a dephasing rate @xmath75 what means that the dephasing times are around one order of magnitude longer than the hopping times ( coherent dynamics ) . within this time interval , which can be called a short interaction time , quantum effects have a chance to manifest or to bring some influence on transport . in the long - time regime , to be briefly discussed next , quantum effects usually become irrelevant since dephasing generally kills the coherences . in this scenario , by comparing the plots in fig . [
fig : fig3 ] , it is clear that the initial presence of entanglement helped transport , i.e. , resulted in values of @xmath76 that surpassed those obtained with the initially mixed ( non - entangled ) situation .
in other words , entanglement worked as a _ resource _ for qt .
another interesting feature of fig .
[ fig : fig3 ] is the presence of nat for the minima when entanglement was originally present in sites @xmath46 e @xmath47 . by increasing the dephasing ,
the efficiency increased in those regions .
this is not true for the maxima where dephasing tends to be destructive .
this is expected because dephasing destroys entanglement in our model , and entanglement is precisely the ingredient for the pronounced maxima . for the initial mixed state ,
the phenomenon of nat is practically absent in fig .
[ fig : fig3 ] .
all these interesting features are confirmed by fig .
[ fig : fig3_1 ] where we provide a more general picture of the problem through variation of dephasing over a broad range .
it is interesting to see that for the initial preparation with entanglement ( plot on the left ) , increasing the dephasing @xmath77 is generally beneficial for the minima , and that this feature is practically not manifested for the non - entangled initial situation ( plot on the right ) . on the contrary ,
dephasing acted as a hinderer when no entanglement was initially present .
we now present the long - time behavior of the population @xmath78 of the last site , fixing two values of @xmath79 : the first minimum and the first maximum of @xmath76 with initial entanglement in fig .
[ fig : fig3 ] .
the results are shown in fig .
[ fig : fig4 ] , where one can see that initial entanglement , as expected , gradually looses its capacity to boost transport .
the only enhancer left is dephasing through nat .
this mechanism is clearly manifest in fig .
[ fig : fig4 ] giving the fact that , for times longer than @xmath80 , the curves with non null dephasing @xmath81 are above the ones with null dephasing @xmath82 . before finishing this section
, we would like to go a little deeper in our investigation about the entanglement - assisted transport phenomenon already observed in previous plots .
we will do this in two directions .
first , we would like to see how fig .
[ fig : fig3 ] changes when the input of energy is changed .
this is shown in fig .
the plots are very illustrative because they once again show the competition between entanglement - assisted transport and noise - assisted transport , now for a different scenario where the increase of noise comes from other sources than pure dephasing . for @xmath83 , the beneficial effect coming from initial entanglement is still quite clear , almost like in fig .
[ fig : fig3 ] . to see this , compare , for instance , the maxima arising from the situation with initial maximal entanglement with the situation with and maximal mixedness . when the incoherent input of energy is increased to @xmath84 ,
the initial maximally mixed state and the initial maximally entangled state are practically equivalent in terms of efficiency .
the reason why initial entanglement starts loosing importance when the energy input rate increases is that adding more energy also adds more noise .
this is so because the energy input and output are both incoherent process corresponding to non - unitary terms in the network master equation , see eqs.([s ] ) and ( [ d ] ) .
the second direction we want to explore is the variation of the initial entanglement . up to now
, we worked only with maximal entanglement versus non - entanglement at all , both states with only one excitation shared between sites @xmath46 and @xmath47 .
we now keep considering the one excitation sector , but with the state @xmath85 for sites @xmath46 and @xmath47 . the other two sites are still considered to be initially in the ground state .
the variation of @xmath86 makes the entanglement of formation vary from zero @xmath87 to one @xmath88 . in fig .
[ entvar ] , we once again consider the efficiency indicator @xmath76 in the first maximum @xmath89 [ see fig .
[ fig : fig3 ] ] .
however , we now have it as a function of dephasing and the entanglement in the initial state @xmath90 .
one can see that states with more initial entanglement lead to higher values of @xmath76 .
this confirms the robustness of the entanglement - assisted transport mechanism for a whole class of states ( all pure states with one excitation shared by control sites @xmath46 and @xmath47 ) .
finally , one can see once again that , for initial pure states and the maxima of @xmath76 in fig .
[ fig : fig3 ] , the increase of dephasing @xmath77 is a hindrance to transport .
since the initial presence of entanglement was seen to be beneficial for transport in the short - time behavior , it would be interesting to have a close look at its dynamics . this could help us to better understand the conclusions previously presented about the effect of quantum correlations ( qc ) over transport for our system of interest .
as said before , we will be employing quantifiers of entanglement and quantum discord qd . for the latter ,
it is important to distinguish between the situation where projective measurements are thought to act on one subsystem or another . in our case , the subsystems are sites @xmath46 and @xmath47 , and we will then denote the case where projective measurements are intended to act on @xmath46 by @xmath91 , and when intended on @xmath47 by @xmath92 . the dynamical behavior of quantum correlations for the same parameters considered in fig .
[ fig : fig4 ] is shown in fig .
[ fig : fig5 ] . as expected , at long times the oscillatory behavior of the quantum correlations
is completely damped rendering the system state to be essentially classicality correlated .
it is interesting to see that there are times where quantum discord remains finite in spite of the fact that entanglement goes to zero . in fig .
[ fig : fig6 ] , we present the average correlations over the same time spam consider to evaluate @xmath76 in fig .
[ fig : fig3 ] , for an initially maximally entangled situation . from this plot ,
it is clear that also on average quantum correlations remain finite during the transport for the whole range of driving strengths @xmath79 considered in fig .
[ fig : fig3 ] . from fig .
[ fig : fig6 ] it is also possible to see that entanglement is a bit more sensitive to the choice of the driving strength than discord in the sense that the former oscillates more stronger than the latter as the driving strength is varied .
in this work , we studied the relation between transport efficiency and initial presence of entanglement in a network .
the network used in this work is the simplest one needed to assess the effect of having entanglement in the intermediate sites , i.e. , between two sites not directly connected to energy sources or sinks . in fact , we found that initial entanglement provides a robust enhancer of transport efficiency . for short times ,
entanglement - assisted transport showed up for all possible initial pure states with only one excitation in the network . in this time domain
, we showed that quantum correlations survive dynamically and on average thus rendering the transport to be quantum in essence . on the other hand , for longer times , these correlations progressively vanish and it comes to a point where only noise - assisted transport is available as a transport enhancer .
we also showed that , for short - times , noise - assisted transport is imaterial for the maximally mixed initial situation . a.a.c . acknowledges to `` coordenao de aperfeioamento de pessoal de nvel superior '' ( capes ) .
fls acknowledges partial support from the brazilian national institute of science and technology of quantum information ( inct - iq ) and cnpq under grant no .
307774/2014 - 7 .
we would also like to thank ms .
marcela herrera and prof .
roberto serra for granting access to their computing facilities .
99 h. rabitz , _ focus on quantum control _ , http://iopscience.iop.org/article/10.1088/1367-2630/11/10/105030/meta[new j. phys . * 11 * , 105030 ( 2009 ) ] .
m. mohseni , y. omar , g. engel and m. b. plenio , _ quantum effects in biology _ , cambridge university press , cambridge , 2014 ) .
m. b. plenio and s. f. huelga , _ dephasing - assisted transport : quantum networks and biomolecules _ , http://iopscience.iop.org/article/10.1088/1367-2630/10/11/113019/meta;jsessionid=caa4fb913a437b73a3e57ac360a20071.c2.iopscience.cld.iop.org[new j. phys . * 10 * , 113019 ( 2008 ) ] .
f. caruso , a. w. chin , a. datta , s. f. huelga and m. b. plenio , _ highly efficient energy excitation transfer in light - harvesting complexes : the fundamental role of noise - assisted transport _ , http://scitation.aip.org/content/aip/journal/jcp/131/10/10.1063/1.3223548?ver=pdfcov[j .
phys . * 131 * , 105106 ( 2009 ) ] .
a. w. chin , a. datta , f. caruso , s.f .
huelga and m. b. plenio , _ noise - assisted energy transfer in quantum networks and light - harvesting complexes _ , http://iopscience.iop.org/article/10.1088/1367-2630/12/6/065002/meta;jsessionid=e781be81050b863796e146585349e019.c2.iopscience.cld.iop.org[new j. phys . * 12 * , 65002 ( 2010 ) ] . a. w. chin , s. f. huelga and m. b. plenio , _ coherence and decoherence in biological systems : principles of noise - assisted transport and the origin of long - lived coherences _ , http://rsta.royalsocietypublishing.org/content/370/1972/3638[phil . trans .
a * 370 * , 3638 ( 2012 ) ] .
bao - quan ai and shi - liang zhu , _ complex quantum network model of energy transfer in photosynthetic complexes _ , http://journals.aps.org/pre/abstract/10.1103/physreve.86.061917[phys .
e * 86 * , 061917 ( 2012 ) ] .
f. l. semio , k. furuya and g. j. milburn , _ vibration - enhanced quantum transport _ , http://iopscience.iop.org/article/10.1088/1367-2630/12/8/083033/meta[new j. phys .
* 12 * , 083033 ( 2010 ) ] .
m. mohseni , p. rebentrost , s. lloyd and a. aspuru - guzik , _ enviroment - assisted quantum walks in photosynthetic energy transfer _ , http://scitation.aip.org/content/aip/journal/jcp/129/17/10.1063/1.3002335[j .
129 * , 174106 ( 2008 ) ] .
f. caruso , a. w. chin , a. datta , s. f. huelga , and martin b. plenio , _ entanglement and entangling power of the dynamics in light - harvesting complexes _ , http://journals.aps.org/pra/abstract/10.1103/physreva.81.062346[phys .
a * 81 * , 062346 ( 2010 ) ] .
m. sarovar , a. ishizaki , g. r. fleming , and k. b. whaley , _ quantum entanglement in photosynthetic light - harvesting complexes _ , http://www.nature.com/nphys/journal/v6/n6/abs/nphys1652.html[nat
* 6 * , 462 ( 2010 ) ] .
k. bradler , m. m. wilde , s. vinjanampathy , and d. b. uskov , _ identifying the quantum correlations in light - harvesting complexes _ , http://journals.aps.org/pra/abstract/10.1103/physreva.82.062310[phys .
a * 82 * , 062310 ( 2010 ) ] .
f. nicacio and f. l. semiao , _ transport of correlations in a harmonic chain _ , http://journals.aps.org/pra/abstract/10.1103/physreva.94.012327[phys .
a * 94 * , 012327 ( 2016 ) ] .
f. nicacio and f. l. semiao , _ coupled harmonic systems as quantum buses in thermal environments _ , http://iopscience.iop.org/article/10.1088/1751-8113/49/37/375303/meta[j .
a : math . theor . * 49 * 375303 ( 2016 ] ) .
a. bermudez , t. schaetz and d. porras , _ synthetic gauge fields for vibrational excitations of trapped ions _ , http://journals.aps.org/prl/abstract/10.1103/physrevlett.107.150501[phys .
107 * , 150501 ( 2011 ) ] .
a. bermudez , t. schaetz and d. porras , _ photon - assisted - tunneling toolbox for quantum simulations in ion traps _ , http://iopscience.iop.org/article/10.1088/1367-2630/14/5/053049/meta[new j. phys . * 14 * , 053049 ( 2012 ) ] .
m. grifoni and p. hnggi , _ driven quantum tunneling _ , http://www.sciencedirect.com/science/article/pii/s0370157398000222[phys
304 * , 229 ( 1998 ) ] .
m. glck , a. r. kolovskya and h. c. korsch , _ wannier - stark resonances in optical and semiconductor superlattices _ , http://www.sciencedirect.com/science/article/pii/s0370157302001424[phys
. rep . * 366 * , 103 ( 2002 ) ] .
z. ficek and s. swain , _ quantum interference and coherence : theory and experiments _ ( springer , new york , ny , 2005 ) .
p. munhoz and f. l. semio , _ multipartite entangled states with two bosonic modes and qubits _ , http://link.springer.com/article/10.1140%2fepjd%2fe2010-00188-2[eur .
j. d * 59 * , 509 ( 2010 ) ] .
j. morais - cabral , y. zhou , and r. mackinnon , _
energetic optimization of ion conduction rate by the k@xmath93 selectivity filter _ , http://www.nature.com/nature/journal/v414/n6859/full/414037a0.html[nature , * 414 * 37 ( 2001 ) ] .
s. bernche and b. roux , _ energetics of ion conduction through the k@xmath93 channel _ ,
http://www.nature.com/nature/journal/v414/n6859/full/414073a0.html[nature , * 414 * 73 ( 2001 ) ] .
a. a. cifuentes and f. l. semio , _ quantum model for a periodically driven selectivity filter in a k@xmath93 ion channel _ , http://iopscience.iop.org/article/10.1088/0953-4075/47/22/225503[j .
b : at . mol .
phys . * 47 * , 225503 ( 2014 ) ] .
h. j. carmichael , _ statistical methods in quantum optics 1 : master equations and fokker - planck equations _
( springer - verlag , berlin - heidelberg , 2002 ) .
p. rebentrost , m. mohseni , i. kassal , s. lloyd and a. aspuru - guzik , _ environment - assisted quantum transport _ , http://iopscience.iop.org/article/10.1088/1367-2630/11/3/033003/meta[new j. phys .
* 11 * , 033003 ( 2009 ) ] .
w. k. wootters , _ entanglement of formation of an arbitrary state of two qubits _ , http://journals.aps.org/prl/abstract/10.1103/physrevlett.80.2245[phys .
lett * 80 * , 2245 ( 1998 ) ] . h. ollivier and w. h. zurek , _ quantum discord : a measure of the quantumness of correlations _ , http://journals.aps.org/prl/abstract/10.1103/physrevlett.88.017901[phys .
lett * 88 * , 017901 ( 2001 ) ] .
g. weihs , t. jennewein , c. simon , h. weinfurter , and a. zeilinger , _ violation of bell s inequality under strict einstein locality conditions _ , http://journals.aps.org/prl/abstract/10.1103/physrevlett.81.5039[phys .
* 81 * , 5039 ( 1998 ) ] .
m. steffen , m. ansmann , r. c. bialczak , n. katz , e. lucero , r. mcdermott , m. neeley , e. m. weig , a. n. cleland , and john m. martinis , _ measurement of the entanglement of two superconducting qubits via state tomography _ , http://science.sciencemag.org/content/313/5792/1423.full[science * 313 * , 1423 ( 2006 ) ] .
a. vaziri and martin b. plenio , _ quantum coherence in ion channels : resonances , transport and verification _ , http://iopscience.iop.org/article/10.1088/1367-2630/12/8/085001/meta[new j. phys .
* 12 * , 085001 ( 2010 ) ] .
g. n. watson , _ a treatise on the theory of bessel functions _
( 2@xmath94 ed .
, cambridge university press , cambridge , 1966 ) . | in this work , we investigate how the presence of initial entanglement affects energy transport in a network .
the network have sites dedicated to incoherent input or output of energy and intermediate control sites where initial entanglement can be established . for short times
, we found that initial entanglement in the control sites provides a robust efficiency enhancer for energy transport .
for longer times , dephasing considerably damps the quantum correlations , and the advantage of having initial entanglement tends to disappear in favor of the well known mechanism of noise - assisted transport .
we careful study these two mechanisms in this work , and we believe our findings may be usuful for a better understanding of the relation between nonclassicality and transport , a topic of potential interest for quantum technologies . |
precoding and power control are well studied strategies that support high spectral efficiency in wireless network with multiple antenna transceivers , when channel state information ( csi ) is available at the transmitters .
several system - wide objective functions have been considered in the literature for precoding and power control optimization of broadcast channels ( bcs ) and interference channels ( ics ) .
some of these problems are convex , for example power minimization @xcite or sinr balancing for the multiple - input single - output ( miso ) bc @xcite , and thus solvable with standard techniques in reasonable ( polynomial ) time . however ,
in general , the problems at hand are non - convex .
unlike convex problems , non - convex problems typically do not afford efficient ( i.e. , polynomial - time ) algorithms that are able to achieve global optimality @xcite .
for example , it is known that the weighted sum - rate maximization ( wsrm ) in parallel ic channels , where interference from other users is treated as noise ( a non - convex problem ) is np - hard @xcite ( this result extends also to bc as a special case ) . in this paper
, we address the global minimization of a system - wide , in general non - convex , cost function with respect to the transmit covariance matrices @xmath0 . among global techniques , branch - and - bound ( bb ) algorithms
are methods to solve general non - convex problems @xcite , producing an @xmath1-suboptimal feasible point .
bb methods have been already introduced to solve non - convex power control problems , although so far only multi - user single - input single - output systems have been addressed in @xcite and references therein . in this paper
, we propose a novel bb framework for global optimization of a problem formulation that includes , for instance , miso bc and ic wsrm with general convex power constraint .
the proposed bb approach is based on the observation that a fairly general set of cost functions that arise in communication s problems , albeit non - convex , possess a _ partly convex - monotone _
@xcite structure .
this structure is satisfied whenever one can identify a suitable set of interference functions @xmath2 , for which the following hold : ( _ i _ ) the cost function is _ convex _ in the transmit covariance matrices @xmath0 once the interference functions @xmath3 are fixed ; ( _ ii _ ) the cost function is _
monotone _ in the interference functions @xmath3 .
we design the bb scheme to exploit the _ partly convex - monotone _ structure of the problem .
branching is performed in a reduced space ( of the size of the set of all feasible interference level vectors @xmath4 ) , instead of the original feasible@xmath5space ( of the size of the set of all feasible covariance matrices @xmath0 ) .
bounding is efficiently carried out by solving only convex optimization problems .
in addition to the reduced - space bb method , we propose a suboptimal algorithm that attains quasi - optimal performance with polynomial complexity .
this algorithm reduces to the distributed pricing scheme of @xcite , when applied to sum - rate maximization problems .
numerical results are provided to compare the global optimal solution based on bb , the suboptimal ( pricing ) technique and the nonlinear dirty - paper coding scheme .
_ notation _ : the boldface is used to denote matrices ( uppercase ) and vectors ( lowercase ) ; @xmath6 and @xmath7 denote the transpose and the hermitian transpose , respectively ; @xmath8 denotes the trace of a matrix ; @xmath9 $ ] denotes the expectation operator . moreover , given a vector @xmath10 we address its @xmath11-th component as @xmath12 _ { l } $ ] , and the vector inequality @xmath13 means that @xmath12 _ { l}\leq\left [ \mathbf{y}\right ] _ { l}$ ] @xmath14 . finally , unless otherwise specified , we address the set of covariance matrices @xmath15 as @xmath0 .
we model a multi - user communication system consisting of @xmath16 transmitter - receiver pairs ( or users ) .
the @xmath17-th user has @xmath18 transmit antennas and one receive antenna ( miso system ) .
the signal at the @xmath17-th receiver is given by@xmath19 where @xmath20 is the @xmath17-th transmitter s signal , @xmath21 accounts for the channel response of the miso link between the @xmath17-th transmitter and the @xmath22-th receiver , and @xmath23 models the additive white gaussian noise ( awgn ) at @xmath17-th receiver : @xmath24 .
assuming capacity - achieving gaussian codebooks , we define the correlation matrix of the @xmath17-th transmitted signal as @xmath25 $ ] . while model ( [ eq : model ] )
accounts for an ic , a bc can be obtained as a special case by setting @xmath26 @xmath27 .
due to multi - user interference , the system performance depends on the transmission strategy of every user , i.e. , on the set of covariance matrices @xmath0 .
we consider the minimization with respect to @xmath0 of a system - wide cost function @xmath28 ( to be defined below ) under a general convex set constraints @xmath29:@xmath30 by defining a set of @xmath31 auxiliary variables @xmath32 , problem ( [ eq : f(q_w ) ] ) can be recast in the equivalent form@xmath33 the equivalence means that if @xmath34 is a solution to ( [ eq : f(q_w ) ] ) , then @xmath35 is a solution to ( [ eq : problem_auxiliar_i ] ) .
conversely , if @xmath36 is a solution to ( [ eq : problem_auxiliar_i ] ) , then @xmath37 is a solution to ( [ eq : f(q_w)]).we further make the following assumptions : 1 .
the @xmath31 interference levels are given by the real vector function @xmath2 , _ affine _ with respect to @xmath0 , that is bounded in the @xmath31-dimensional rectangle @xmath38 \subset\mathbb{r}^{l}$ ] ( i.e. , the @xmath11-th component satisfies @xmath39 _ { l}\leq i_{l}^{\max } $ ] for @xmath40 ) .
for instance , we typically have @xmath41 and the interference level at the @xmath17-th receiver reads @xmath42 _ { k}=\sum\nolimits_{j=1,j\neq k}^{k}\mathbf{\mathbf{h}}_{jk}\mathbf{q}_{j}\mathbf{\mathbf{h}}_{jk}^{h}$ ] ; 2 .
the cost function @xmath43 is a real scalar function that is : _ continuous _ in @xmath44 ; _ _ monotonic increasing _ _ results _ monotone increasing _ in @xmath45 and _ monotone decreasing _ in @xmath46 . ] with respect to @xmath47 $ ] for fixed @xmath48 ; _ convex _ with respect to @xmath0 for fixed @xmath49 $ ] ; 3 .
the set @xmath29 is _ closed _ and _
convex_. for example , @xmath29 may be the set of positive semidefinite covariance matrices @xmath50 satisfying the generalized power constraints @xmath51 for @xmath52 , where @xmath53 are positive semidefinite matrices ( possibly @xmath54 if @xmath17-th user does nt belong to @xmath55-th constraint ) and @xmath56 are non - negative coefficients .
this definition includes some important special cases studied in the literature , such as per - antenna , per - group of antennas , the classical sum - power or the interference constraints in cognitive radio scenarios . throughout the paper , we refer to problem ( [ eq : problem_auxiliar_i ] ) as @xmath57 .
we next provide examples of problems that satisfy these assumptions .
an example of cost function included in our framework is the @xmath58-fairness criterion @xcite : @xmath59 , where @xmath60 is a positive constant , @xmath61 is an increasing strictly concave function defined as@xmath62{lc}\log r & \text{if } \alpha=1\\ \left ( 1-\alpha\right ) ^{-1}r^{1-\alpha } & \text{otherwise}\end{array } \text{,}\right.\ ] ] and @xmath63 is the @xmath17-th user s rate , which depends on covariance matrices @xmath0 and on the _ channel scenario_. the @xmath58-fairness criterion reduces , as special cases , to the wsrm problem ( @xmath64 ) or the _ proportional fairness _ problem ( @xmath65 ) . moreover , as @xmath58 becomes large , it converges to the _ max min fairness _
problem @xcite.in the following we present some examples of channel scenario that can be addressed within our framework : * _ parallel miso ic _ : the @xmath17-th transmitter operates over @xmath66 parallel subcarriers , it has power constraint @xmath67 and has knowledge of channels @xmath68 for @xmath69 and @xmath14.the minimization of the @xmath70-fairness cost function reads@xmath71{ll}r_{k}\leq\sum\limits_{l=1}^{l_{c}}\log\left ( 1\mathbf{+}\frac { \mathbf{\mathbf{h}}_{kkl}\mathbf{q}_{kl}\mathbf{\mathbf{h}}_{kkl}^{h}}{\sigma_{kl}^{2}\mathbf{+}{\textstyle\sum_{j=1,j\neq k}^{k } } \mathbf{\mathbf{h}}_{jkl}\mathbf{q}_{jl}\mathbf{\mathbf{h}}_{jkl}^{h}}\right ) & \forall k\\ \operatorname{tr}\left ( \sum_{l=1}^{l_{c}}\mathbf{q}_{kl}\right ) \leq p_{k } & \forall k \end{array } \right .
\nonumber\end{gathered}\ ] ] defining @xmath72 _ { l+l_{c}\left ( k-1\right ) } = \sum\nolimits_{j=1,j\neq k}^{k}\mathbf{\mathbf{h}}_{jkl}\mathbf{q}_{jl}\mathbf{\mathbf{h}}_{jkl}^{h}$ ] @xmath73 and @xmath74 , problem ( [ eq : ic wsrm ] ) is recast into @xmath57 . also ,
@xmath75 $ ] where @xmath76 is a proper upper bound on interference , always available since @xmath29 is bounded ( finite power constraints ) . * _ parallel miso bc _ : this scenario is obtained from ( [ eq : ic wsrm ] ) by setting @xmath77 @xmath78 and imposing a sum - power constraint @xmath79 .
in this section we show that , adopting standard bb techniques ( see @xcite and @xcite ) , problem @xmath80 can be optimally solved by means of an efficient bb that exploits the structure dictated by assumptions ( a1-a3 ) .
the bb algorithm is fully characterized by two procedures : _ branching _ and _ bounding_. these are iteratively performed until the solution s suboptimality falls below some prescribed accuracy @xmath1 . in the following
we explicitly tailor those procedures to problem @xmath80 and we show convergence of the proposed bb algorithm to the global optimal solution of @xmath57 . for readability s sake
, we define @xmath81 stands for @xmath48 . ] and we address an interval as @xmath82 $ ] , meaning that @xmath83 _ { l}\leq\left [ \mathbf{\mathbf{c}}\right ] _ { l}\leq\left [ \mathbf{b}\right ] _ { l}$ ] for @xmath40 . a partition set @xmath84 of rectangles @xmath85 in the space @xmath86 , each labeled with a _ lower _
@xmath87 and _ upper _ @xmath88 bounds , is given . by splitting a rectangle that satisfies @xmath89 in @xmath90 non - overlapping sub - rectangles @xmath91 ( i.e. , @xmath92 and @xmath93 ) ,
the enhanced partition @xmath94 is obtained .
lower and upper bounds for each sub - rectangle in @xmath95 are then obtained via the following _ bounding procedure_. exploiting the _ partly convex - monotone _ structure of problem @xmath80 , for every rectangle @xmath96 \in\mathcal{p}_{t}$ ] , a _ lower bound _
@xmath87 is evaluated by solving the following problem:@xmath97 thanks to assumptions ( a1-a3 ) two fundamental results can be verified : ( _ i _ ) problem ( [ eq : pbeta ] ) is _ convex _ since the cost function @xmath98 is convex for a fixed @xmath32 and the constraints form a convex set , ( _ ii _ ) using standard convex optimization arguments , it can be shown that this bounding procedure satisfies the natural condition:@xmath99 moreover , denoting with @xmath100 the optimal solution of problem ( [ eq : pbeta ] ) , a valid upper bound @xmath88 is obtained by evaluating the function at @xmath100 , i.e. , @xmath101.finally , the algorithm checks if the prescribed accuracy is met ( i.e. , if @xmath102 ) otherwise it goes back to the _ branching procedure_. here we proves convergence of the proposed bb algorithm .
[ lemma 1 ] the proposed bb algorithm ( which is performed in the reduced space spanned by interference levels / variable @xmath32 ) , is _ convergent _ to a global optimal solution of problem @xmath103 .
as explained above , since the chosen bounding procedure satisfyies ( [ eq : bound condition ] ) , the bb algorithm generates a sequences of partition sets @xmath104 collapsing to a point @xmath105 ( recall that @xmath106 is the rectangle selected for splitting at the @xmath107-th branching iteration ) . in order to prove convergence we need to show that , as the size of rectangle @xmath106 gets smaller , @xmath108 is also sufficiently small .
the proof follows standard arguments @xcite .
this is shown in appendix .
considering the bc wsrm scenario ( see sec.[marker : example utility ] ) , for a given interval @xmath109 $ ] , the evaluation of a lower bound @xmath110 results in the following convex problem@xmath111{lc}\operatorname{tr}\left ( \sum_{k=1}^{k}\mathbf{q}_{k}\right ) \leq p_{tot } & \\ i_{k}^{\min}\leq\mathbf{\mathbf{h}}_{k}\left ( { \textstyle\sum\nolimits_{j\neq k } } \mathbf{q}_{j}\right ) \mathbf{\mathbf{h}}_{k}^{h}\leq i_{k}^{\max } & \forall k \end{array } \text{.}\right . \nonumber\end{aligned}\ ] ] defining @xmath34 the optimal solution of ( [ eq : lb ] ) , a valid upper bound is given by @xmath112 where @xmath113 @xmath114 .
while the proposed bb algorithm always converges to the global optimal solution and has reduced complexity with respect to a general - purpose implementation of bb , it is still feasible only for offline simulation . in this section we propose a suboptimal algorithm with polynomial complexity that extends the distributed pricing schemes of @xcite to the more general class of problem @xmath115.exploiting the _ partly convex - monotone _ structure , problem ( [ eq : problem_auxiliar_i ] ) can be equivalently reformulated as the non - convex problem:@xmath116 -\mathbf{\lambda}^{t}\mathbf{i}\right ] \text{.}\label{eq : convex opt g}\ ] ] where @xmath117 is the lagrange multiplier associated to the affine constraint @xmath118 . building on ( [ eq :
convex opt g ] ) , in the following table we formalize the proposed suboptimal algorithm .
_ 0 _ : _ set _ @xmath119 _ 1 _ : _ initialize _ @xmath120 _ 2 _ : _ initialize _ @xmath121 _ 3 _ : _ evaluate _ @xmath122 _ _ 4 : _ _ * * if * * @xmath123 _ 5 : update _ @xmath124 _ 6 : go back to step 3 _ _ 7 _ : * elseif * @xmath125 _ 8 : update _ @xmath126 _ 9 : go back to step 2 _ _ 10 _ : * end * @xmath127since a stationary point of this algorithm fulfills the _ necessary _ karush - kuhn - tucker ( kkt ) conditions of problem ( [ eq : problem_auxiliar_i ] ) , if the algorithm converges , it attains a local optimal point of problem ( [ eq : problem_auxiliar_i]).it is worth noticing that , by specializing our framework to the case when the cost function @xmath98 is the wsrm ( i.e. , @xmath128 @xmath114 in ( [ eq : ic wsrm ] ) ) the lagrangian multiplier @xmath117 plays the role of the _ interference prices _ defined in the distributed pricing algorithm @xcite . thus algorithm [ algo1 ] can be seen as a generalization of distributed pricing technique with an arbitrarily cost function and arbitrary _ interference _ functions ( satisfying assumptions a1-a3).finally , since the problem at hand is non - convex , initialization of the parameters @xmath117 and @xmath32 results crucial for performances and convergence . in sec.[sec : num res ] we assess the performances of this technique in relation to the global optimal solution evaluated via bb algorithm .
we assess the performance of the two proposed techniques : a multi - user linear precoder optimized via ( _ i _ ) the efficient bb algorithm ( bb - lb and ub ) ; ( _ ii _ ) the suboptimal algorithm [ algo1].we consider the sum - rate utility function ( i.e. , in ( [ eq : ic wsrm ] ) , @xmath129 @xmath114 and @xmath130 @xmath114 ) . in bb algorithm ,
the solution s accuracy is @xmath131 , while , in algorithm [ algo1 ] , we run two different price initializations ( @xmath132 @xmath114 and @xmath133 @xmath114 ) and for both we initialize @xmath134 @xmath114 , selecting @xmath135 at each receive antenna .
[ ptb ] sum_capacity_h_randn_4x4_004_v4_now.eps fig.[fig .
1 ] shows the sum - rate versus the transmitting power for a single - carrier ( @xmath136 ) bc channel channels realization is available at : ` http://web.njit.edu/mr227/papers/paper_bb_h_bc.mat ` ] where a @xmath137 transmit antennas base - station serves @xmath138 single - antennas users , subject to a sum - power constraint , @xmath139 .
the sum - capacity achieving _ non - linear _ technique dirty paper coding ( dpc ) is also plotted as a reference.it can be noticed that the suboptimal algorithm [ algo1 ] , while showing near - optimal performance at several power levels , happens to be quite sensitive to initialization .
for instance , initialization @xmath132 @xmath114 yields a suboptimal slope in high power regime , as observed for @xmath140 , and , at @xmath141 both initializations lead to highly suboptimal performances .
a last observation pertains to the significant gains of _ non - linear _ dpc with respect to _ linear _ precoding at high power regime.finally , not to confuse the reader , since a utility function ( sum - rate ) instead of a cost function is plotted , in fig.[fig .
1 ] , the lower bound results as the maximum feasible value while the upper bound is the maximum upper bound among bb partitions .
this work presents a global optimization framework for the minimization of non - convex cost functions in miso bc and ic channels .
examples are given for the general @xmath58-fairness optimization considering parallel ic and bc channels .
knowing the global optimal solution , even if impractical for real - time implementation , allows to assess the quality and to fine - tune ( e.g. , initialize ) suboptimal schemes .
in addition to the global optimal bb , we have proposed a real - time , hence suboptimal , algorithm that generalizes the pricing scheme of @xcite .
extensions to mimo networks are the subject of future work .
we need to prove that , as the maximum length of the edges of @xmath142 , denoted by @xmath143 , goes to zero , the difference between upper and lower bounds uniformly converges to zero , i.e. , @xmath144 @xmath145.for each @xmath146 $ ] , we define the function @xmath147 as the result of the following constraint optimization problem:@xmath148 using this notation , the lower bound in ( [ eq : pbeta ] ) is given by @xmath149 , while an upper bound is given by @xmath150.from jointly - continuity of the function @xmath151 with respect to @xmath152 ( assumption a2 ) and from the definition of @xmath147 , we have that @xmath147 is continuous in the norm of @xmath153 ( i.e. , @xmath154 ) .
it follows that also @xmath110 and @xmath155 will result continue in @xmath156 , thus it holds@xmath157 concluding the proof .
9 f. rashid - farrokhi , k. r. liu , and l. tassiulas , transmit beamforming and power control for cellular wireless systems , in _ ieee journal on selected areas of communications _ , vol .
14371450 , oct .
1998 .
v. balakrishnan , s. boyd , and s. balemi . branch and bound algorithm for computing the minimum stability degree of parameter - dependent linear systems .
journal of robust and nonlinear control _
, 1(4):295317 , october
december 1991 .
y. xu , t. le - ngoc , and s. panigrahi , global concave minimization for optimal spectrum balancing in multi - user dsl networks , _ ieee trans . on signal processing _ ,
56 , no . 7 , pp .
. 28752885 , jul . | utility ( e.g. , sum - rate ) maximization for multiantenna broadcast and interference channels ( with one antenna at the receivers ) is known to be in general a non - convex problem , if one limits the scope to linear ( beamforming ) strategies at transmitter and receivers . in this paper , it is shown that , under some standard assumptions , most notably that the utility function is decreasing with the interference levels at the receivers , a global optimal solution can be found with reduced complexity via a suitably designed branch - and - bound method . although infeasible for real - time implementation , this procedure enables a non - heuristic and systematic assessment of suboptimal techniques .
in addition to the global optimal scheme , a real - time suboptimal algorithm , which generalizes the well - known distributed pricing techniques , is also proposed .
finally , numerical results are provided that compare global optimal solutions with suboptimal ( pricing ) techniques for sum - rate maximization problems , affording insight into issues such as the robustness against bad initializations in real - time suboptimal strategies .
nonconvex optimization , branch - and - bound , interference channel , multiple - input single - output channel |
the existence of massive dark matter halos around galaxies is widely accepted , based on different lines of evidence , such as flat rotation curves of spiral galaxies ( e.g. , van albada & sancisi 1986 ) and strong lensing systems ( e.g. , keeton , kochanek & falco 1998 ) . however , relatively little is known about the properties of dark matter halos .
strong lensing only probes the gravitational potential on small ( projected ) scales , whereas the lack of visible tracers at large radii hamper dynamical methods . to date , only satellite galaxies have provided some information ( e.g. , zaritsky & white 1994 ; mckay et al .
2002 ; prada et al .
2003 ) .
a promising approach to study the galaxy dark matter halos is weak gravitational lensing .
the tidal gravitational field of the dark matter halo introduces small coherent distortions in the images of distant background galaxies .
the weak lensing signal can be measured out to large projected distances from the lens , and hence provides a unique probe of the gravitational potential on large scales .
the applications of this approach are numerous : one can infer masses of galaxies and compare the results to their luminosities ( e.g. , mckay et al .
2001 ; wilson et al .
2001 ) , or one can attempt to constrain the halo mass profile ( e.g. , brainerd et al .
1996 ; hudson et al . 1998 ; fischer et al .
2000 ; hoekstra et al .
also , weak lensing can be used to constrain the shapes of halos by measuring the azimuthal variation of the lensing signal .
unfortunately , one can only study ensemble averaged properties , because the weak lensing signal induced by an individual galaxy is too low to be detected .
a successful measurement of the lensing signal requires large samples of both lenses and background galaxies .
the first attempt to detect the lensing signal by galaxies was made by tyson et al .
( 1984 ) using photographic plates .
it took more than a decade and ccd cameras before the first detections were reported ( brainerd et al .
1996 ; griffiths et al . 1996 ; dellantonio & tyson 1996 ; hudson et al .
these early results were limited by the small areas covered by the observations .
the accuracy with which the galaxy - galaxy lensing signal can be measured depends on the area of sky that is observed , and on the availabitity of redshifts for the lenses ( as it allows for a proper scaling of the lensing signal ) .
photometric redshifts were used by hudson et al .
( 1998 ) to scale the lensing signal of galaxies in the hubble deep field , and by wilson et al .
( 2001 ) who measured the lensing signal around early type galaxies as a function of redshift . furthermore , several lensing studies targeted regions covered by redshift surveys .
smith et al .
( 2001 ) used 790 lenses from the las campanas redshift survey ; hoekstra et al .
( 2003 ) used 1125 lenses from the canadian network for observational cosmology field galaxy redshift survey ( cnoc2 ) .
the areas covered by these surveys are relatively small .
the sloan digital sky survey ( sdss ) combines both survey area and redshift information .
its usefulness for galaxy - galaxy lensing was demonstrated clearly by fischer et al .
( 2000 ) . more recently , mckay et al .
( 2001 ) used the available sdss redshift information to study the galaxy - galaxy lensing signal as a function of galaxy properties ( also see guzik & seljak 2002 ; seljak 2002 ) .
the data used in this paper currently lacks redshift information for the lenses . however , compared to previous work , the combination of large area and depth of our observations allow us to measure the galaxy lensing signal with great precision .
we use 45.5 deg@xmath0 of @xmath1-band imaging data from the red - sequence cluster survey ( rcs ) .
these data have been used previously for several weak lensing studies .
hoekstra et al .
( 2002a ; 2002b ) placed joint constraints on @xmath11 and @xmath12 by measuring the lensing signal caused by large scale structure .
related to the subject of this paper is the study of the bias parameters as a function of scale by hoekstra et al .
( 2001b ) and hoekstra et al .
( 2002c ) .
the latter studies made use of the galaxy - mass cross - correlation function measured from the rcs data .
here we use the galaxy - mass cross - correlation function for a different purpose : we effectively deconvolve the cross - correlation function to study the properties of dark matter halos surrounding galaxies at intermediate redshifts .
the structure of the paper is as follows . in 2 we briefly discuss the data and the redshift distributions of the lenses and the sources .
the ensemble averaged tangential shear around the lenses ( galaxy - mass cross - correlation function ) is presented in 3 . in
4 we use a maximum likelihood analysis to derive constraints on the extent of dark matter halos . the measurement of the projected shapes of the halos is presented in 5 .
we use the @xmath1-band imaging data from the red - sequence cluster survey ( yee & gladders 2001 ; gladders & yee 2003 ) .
the complete survey covers 90 deg@xmath0 in both @xmath1 and @xmath13 , spread over 22 widely separated patches of @xmath14 degrees . in this paper
we use data from the northern half of the survey , which consists of 10 patches , observed with the cfh12k camera on the cfht .
these data cover 45.5 deg@xmath0 on the sky , but because of masking the effective area is somewhat smaller . in the lensing analysis we use a total of 42 deg@xmath0 .
a detailed description of the data reduction and object analysis can be found in hoekstra et al .
( 2002a ) , to which we refer for technical details .
here we present a short description of the various steps in the analysis .
we use single exposures in our analysis , and consequently cosmic rays have not been removed . however , cosmic rays are readily eliminated from the photometric catalogs : small , but very significant objects are likely to be cosmic rays , or artifacts from the ccd .
the object finder gives fair estimates of the object sizes , and we remove all objects smaller than the size of the psf .
some faint cosmic rays may hit galaxies , and consequently might not be recognized as cosmic rays .
based on the number of cosmic ray hits , and the area covered by galaxies we find that less than @xmath15 of the galaxies might be affected .
also , cosmic rays only introduce additional noise in the shape measurement , but do not bias the result .
consequently we conclude that remaining cosmic rays have a negligble effect on our results . the objects in this cleaned catalog
are then analysed , which yield estimates for the size , apparent magnitude , and shape parameters ( polarisation and polarisabilities ) .
the objects in this catalog are inspected by eye , in order to remove spurious detections .
these objects have to be removed because their shape measurements are affected by cosmetic defects ( such as dead columns , bleeding stars , halos , diffraction spikes ) or because the objects are likely to be part of a resolved galaxy ( e.g. , hii regions ) .
to measure the small , lensing induced distortions in the images of the faint galaxies it is important to accurately correct the shapes for observational effects , such as psf anisotropy , seeing and camera shear ; psf anisotropy can mimic a cosmic shear signal , and a correction for the seeing is required to relate the measured shapes to the real lensing signal . to do so , we follow the procedure outlined in hoekstra et al .
we select a sample of moderately bright stars from our observations , and use these to characterize the psf anisotropy and seeing .
we fit a second order polynomial to the shape parameters of the selected stars for each chip of the cfh12k camera .
these results are used to correct the shapes of the galaxies for psf anisotropy and seeing .
the effect of the psf is not the only observational distortion that has to be corrected .
the optics of the camera stretches the images of galaxies ( i.e. , it introduces a shear ) because of the non - linear remapping of the sky onto the ccd .
we have used observations of astrometric fields to find the mapping between the sky and the ccd pixel coordinate system , and derived the corresponding camera shear , which is subsequently subtracted from the galaxy ellipticity ( see hoekstra et al .
1998 ) .
the findings presented in hoekstra et al .
( 2002a ) suggest that the correction for psf anisotropy has worked well . the absence of a `` b''-mode on large scales in the measurements of the cosmic shear ( hoekstra et al .
2002b ) provides additional evidence that systematics are well under control ( the small scale `` b''-mode is attributed to intrinsic alignments ) .
furthermore , cosmic shear studies are much more sensitive to systematics than galaxy - galaxy lensing measurements ( e.g. , see hoekstra et al .
2003 ) . in galaxy - galaxy
lensing one measures the lensing signal that is perpendicular to the lines connecting many lens - source pairs .
these connecting lines are randomly oriented with respect to the psf anisotropy , and hence suppress any residual systematics .
for the analysis presented here , we select a sample of `` lenses '' and `` sources '' on the basis of their apparent @xmath1 magnitude .
we define galaxies with @xmath2 as lenses , and galaxies with @xmath3 as sources which are used to measure the lensing signal .
this selection yields a sample of @xmath16 lenses and @xmath17 sources . for a singular isothermal sphere ,
the amplitude of the lensing signal depends on @xmath18 , the average ratio of the angular diameter distances between the lens and the source , @xmath19 , and the distance between the observer and the source @xmath20 .
more general , the signal also depends on @xmath21 , the distance between the observer and the lens . hence , to interpret the measurements , such as size and mass
, one needs to know the redshift distributions of both lenses and sources . the cnoc2 field galaxy redshift survey ( e.g. , lin et al . 1999 ; yee et al .
2000 ; carlberg et al .
2001 ) has measured the redshift distribution of field galaxies down to @xmath22 , which is ideal , given our limits of @xmath2 .
the derived redshift distribution gives a median redshift @xmath23 for the lens sample .
in addition , we use the redshifts and the colors of the galaxies observed in the cnoc2 survey to compute their rest - frame @xmath24 luminosity . compared to studies using sdss data ( mckay et al .
2001 ) we have the disadvantage that we do not have ( spectroscopic ) redshifts for the individual lenses . as shown by schneider & rix ( 1997 ) and hoekstra et al .
( 2003 ) this limits the accuracy of the measurements
. we can derive useful constraints on the masses and extent of dark matter halos , but we have to assume scaling relations between these parameters and the luminosity of the galaxies .
multi - color data for the northern part of the rcs will be available in the near future , allowing us to select a sample of lenses based on their photometric redshifts , and constrain the scaling relations .
nevertheless , the large area covered by the rcs allows us to derive interesting information about the properties of the lenses . in
4 we use different cuts in apparent magnitude to study the properties of dark matter halos using a maximum likelihood analysis . for the source galaxies
the situation is more complicated .
these galaxies are generally too faint for spectroscopic surveys , although recently cohen et al .
( 2000 ) measured spectroscopic redshifts around the hubble deep field north down to @xmath25 .
cohen et al .
( 2000 ) find that the spectroscopic redshifts agree well with the photometric redshifts derived from multi - color photometry .
because of likely field - to - field variations in the redshift distribution , we prefer to use the photometric redshift distributions derived from both hubble deep fields ( fernndez - soto et al .
1999 ) , which yields a median redshift of @xmath26 for the source galaxies .
the adopted source and lens redshift distributions result in an average value of @xmath27 ( average for the full sample of lenses and sources ) , where the error bar is based on the field - to - field variation and the finite number of galaxies in the hubble deep fields .
the uncertainty in @xmath18 affects predominantly our estimates for the galaxy masses ( and velocity dispersions ) , but is negligible for the other model parameters .
we note that for the galaxy - galaxy lensing analysis presented here the relevant parameter is @xmath28 , as opposed to simply @xmath18 , but we have verified that results in a negligible change in the adopted systematic error . throughout the paper we indicate the systematic error in the masses and velocity dispersions by a second error bar .
the galaxy - mass cross - correlation function provides a convenient way to present the measurements .
it is obtained from the data by measuring the tangential alignment of the source galaxies with respect to the lens as a function of radius .
its use for studying the halos of galaxies is limited , because the clustering of galaxies complicates a direct interpretation of the signal : on small scales the signal is dominated by the mass distribution of the lens , but on larger scales one measures the superposition of the contributions from many lenses .
the observed galaxy - mass cross - correlation function as a function of angular scale is presented in figure [ gtprofall]a .
a significant signal is detected out to one degree from the lens .
if the signal presented in figure [ gtprofall]a is caused by gravitational lensing , no signal should be present when the phase of the distortion is increased by @xmath29 ( i.e. , when the sources are rotated by 45 degrees ) .
the results of this test , shown in figure [ gtprofall]b , suggest that residual systematics are negligible . before we can interpret the results we need to examine the contribution of foreground galaxies
. some of the source galaxies will be in front of the lenses , and lower the lensing signal independent of radius .
this is absorbed in the value of @xmath18 .
these galaxies decrease the lensing signal independent of angular scale .
some sources , however , are physically associated with the lenses .
these galaxies cluster around the lenses , affecting the lensing signal more on small scales .
we need to account for this source of contamination .
to do so , we measure a fractional excess of sources around lenses which decreases with radius as @xmath30 ( r in arcseconds ) , similar to what was found by fischer et al .
( 2000 ) . under the assumption that the orientations of these galaxies are random ( the tidal interaction with the lens has not introduced an additional tangential or radial alignment )
, the observed lensing signal has to be increased by a factor @xmath31 .
this assumption is supported by the findings of bernstein & norberg ( 2002 ) who examined the tangential alignment of satellite galaxies around galaxies , extracted from the 2df galaxy redshift survey .
the measurements presented in figures [ gtprofall ] have been corrected for this decrease in signal .
we note that the correction for the presence of satellite galaxies is small , and has no significant effect on our results .
the signal on small angular scales is dominated by a single lens galaxy , and can be used to obtain an estimate of the mass weighted velocity dispersion of the sample of lenses ( although such a mass estimate can still be slightly biased because of the clustering of the lenses ) . figure
[ gtprofzoom]a shows the ensemble averaged tangential shear on small scales .
the measurements presented in figure [ gtprofzoom]b show no evidence for residual systematics .
we fit a singular isothermal sphere ( sis ) model to the tangential shear at radii smaller than 2 arcminutes ( which corresponds to @xmath32 kpc at the mean redshift of the lenses ) .
the best fit model is indicated by the dashed line in figure [ gtprofzoom]a . for the einstein radius @xmath33
we obtain a value of @xmath34 .
if we extend the fit to much larger radii , the inferred value for the einstein radius increases systematically . with our adopted redshift distributions for the lenses and the sources
, the value of @xmath35 corresponds to a value of @xmath36 km / s .
the corresponding circular velocity can be obtained using @xmath37 ( for a spherical halo ) .
the derived value of @xmath38 depends on the selection of the sample of lens galaxies , which hampers a direct comparison with other studies .
it is more useful to compute the velocity dispersion of a galaxy with some fiducial luminosity , for which we take .
to do so , we have to adopt a scaling relation between the velocity dispersion and the luminosity .
we assume @xmath39 ( see 4 ) .
furthermore , the luminosity might evolve with redshift , and we will consider two cases . under the assumption that the luminosity does not evolve with redshift , we obtain a velocity dispersion @xmath40 km / s for a galaxy with . hoekstra et al .
( 2003 ) measured @xmath41 km / s for a galaxy with a luminosity of using a sample of galaxies with redshifts from cnoc2 . scaling our measurements to
this luminosity implies a velocity dispersion of @xmath42 km / s , in excellent agreement with hoekstra et al .
if we assume that the luminosity scales with redshift as @xmath43 , we find a velocity dispersion of @xmath44 km / s for a galaxy with .
the galaxy - mass cross - correlation function is the convolution of the galaxy distribution and the galaxy dark matter profiles . to examine the ensemble average properties of the dark matter halos properly
, we need to deconvolve the galaxy - mass cross - correlation function ( i.e. , we need to account for the clustering of the lenses ) .
this is done naturally in a maximum likelihood analysis , where a model for the mass distribution of individual galaxies is compared to the observations .
fisher et al . ( 2000 ) and mckay et al . ( 2001 ) used a model with the velocity dispersion and `` extent '' of the halo as free parameters , and computed the resulting galaxy - mass correlation function for a range of model parameters and compared the results to the observed galaxy - mass correlation function .
they obtained good constraints for the velocity dispersion , but were not able to constrain the extent of the dark matter halos .
the galaxy - mass correlation function ignores additional information contained in the data , whereas the maximum likelihood analysis allows one to gauge how much signal might arise from nearby galaxies .
hoekstra et al .
( 2003 ) compared their mass model to the observed polarisation field , making use of both components of the polarisation ( also see brainerd et al .
1996 ; schneider & rix 1997 ; hudson et al .
this greatly improved the constraints on the extent of the halo , and we will use the same method here . in our analysis we make an important assumption : all clustered matter is associated with the lenses .
if the matter in galaxy groups ( or clusters ) is associated with the halos of the group members ( i.e. , the halos are indistinguishable from the halos of isolated galaxies ) our results should give a fair estimate of the extent of galaxy halos .
however , if a significant fraction of the dark matter is distributed in common halos , a simple interpretation of the results becomes more difficult .
hoekstra et al .
( 2003 ) examined how known groups in their observed fields affected the lensing results , and found that the masses and sizes might be overestimated by at most @xmath45 .
guzik & seljak ( 2002 ) found similar results from their analysis of the galaxy - galaxy lensing signal in the context of halo models .
their approach allows one to separate the contribution from groups to the lensing signal .
as expected , guzik & seljak ( 2002 ) found that the effect depends on galaxy type : early type galaxies are found in high density regions , and are affected more .
alternatively , a comparison with numerical simulations which include a prescription for galaxy formation ( e.g. , kauffmann et al .
1999a , 1999b ; guzik & seljak 2001 ) can be used to quantify this effect .
another complication is the fact that we can not separate the lenses in different morphology classes with the current data .
early and late type galaxies of a given luminosity have different masses , etc .
( e.g. , guzik & seljak 2002 ) .
hence , it is important to keep in mind that the results presented here are ensemble averages over all galaxy types .
this is where the sdss can play an important role , although we can significantly improve the rcs results with upcoming multi - color data .
the variance in the polarisations is approximately constant with apparent magnitude and we approximate the distribution by a gaussian distribution . with the latter assumption ,
the log - likelihood follows a @xmath46 distribution with the number of degrees of freedom equal to the number of free model parameters , and the determination of confidence intervals is straightforward .
the log - likelihood is given by the sum over the two components of the polarisation @xmath47 of all the source galaxies @xmath48 where @xmath49 are the model distortions , @xmath50 is the shear polarisability , @xmath51 are the observed image polarisations for the @xmath52th galaxy , and @xmath53 is the dispersion in the polarisation of the @xmath52th galaxy ( which is the combination of the intrinsic shape of the galaxy and the measurement error caused by shot noise in the images ) . in order to minimize the contribution of the baryonic ( stellar ) component of the galaxies we compare our model to the observations at radii larger than 10 arcseconds ( which corresponds to @xmath54 kpc at the redshift of the lenses ) , where the dark matter halo should dominate the lensing signal . in our maximum likelihood analysis
we ignore the contribution from lenses outside the field of view ( e.g. , hudson et al .
1998 ) . for small fields of view
this tends to slightly lower the resulting halo masses and sizes .
the area covered by our observations is much larger than the hdf north studied by hudson et al .
( 1998 ) , and the effect on our estimates is negligible . to infer the best estimates for the model parameters , one formally has to perform a maximum likelihood analysis in which the redshift of each individual galaxy is a free parameter , which has to be chosen such that it maximizes the likelihood .
this approach is computationally not feasible , and instead we create mock redshift catalogs , using the observed redshift distributions from the cnoc2 survey ( hoekstra et al .
2003 ) , which allows us to find estimates for the model parameters . for each lens galaxy in the rcs catalog , with a given apparent @xmath1 magnitude
, we randomly draw a galaxy from the cnoc2 survey ( with a similar @xmath1 , but allowing for a small range in magnitude ) .
the lens galaxy is then assigned the redshift and restframe @xmath24-band luminosity of that cnoc2 galaxy .
we take the incompleteness of the survey into account when the redshifts are drawn from the survey .
the mock catalogs are then analysed as if the redshifts of the lenses were known .
although this procedure does not provide the formal maximum likelihood parameter estimation , it does yield an unbiased estimate of the parameters .
the procedure is repeated 10 times and enables us to properly account for the uncertainty introduced by the lack of precise redshifts for the lenses .
future multi - color catalogs will improve the results presented here significantly .
in particular , in the analysis presented here we have to adopt scaling relations between the velocity dispersion ( or mass ) and the luminosity of the lens and the extent of the halo and the luminosity . the use of photometric redshifts for the lenses will allow us to actually contrain these relations . in 4.1 we consider the truncated isothermal sphere ( tis ; e.g. , brainerd et al .
1996 ; schneider & rix 1997 ; hoekstra et al .
2003 ) , which has been used previously in galaxy - galaxy lensing analyses . in
4.2 we compare the data to the popular nfw profile ( navarro et al . 1995 , 1996 , 1997 ) .
a simple model is the truncated isothermal sphere proposed by brainerd et al .
its density profile is given by @xmath55 where @xmath56 is the line - of - sight velocity dispersion , and @xmath57 is a truncation scale , i.e. the radius where the profile steepens . on small scales
@xmath58 the model behaves as an singular isothermal sphere ( sis ) model , whereas for @xmath59 the density decreases @xmath60 .
the mass contained within a sphere of radius @xmath61 is given by @xmath62 which results in a finite total mass of @xmath63 the projected surface density for this model is given by @xmath64 the corresponding expressions for the shear can be found in brainerd et al . (
1996 ) and schneider & rix ( 1997 ) .
we use this model to compute the model shear field and compare it to the data .
the lenses , however , span a range in masses and we need to account for that using scaling relations , which allow us to relate the halo properties of the lenses to those of a fiducial galaxy . for the fiducial galaxy
we take a luminosity of .
dynamical studies provide evidencence of a power law scaling relation between the velocity dispersion and the luminosity ( e.g. , tully - fisher relation for spiral galaxies and faber - jackson relation for early type galaxies ) .
we assume @xmath65 , which is based on the observed slope of the @xmath24-band tully - fisher relation ( e.g. , verheijen 2001 ) .
little is known , however , about the relation of the extent of dark matter halos with other ( observable ) parameters .
using sdss data , guzik & seljak ( 2002 ) find that @xmath66 .
motivated by this result , we adopt @xmath67 , which gives a total mass @xmath68 . we note , however , that we probe lower luminosities than guzik & seljak ( 2002 ) ( also see mckay et al .
2001 ) and as a result , the adopted scaling relation might differ from the actual ones .
figure [ size ] shows the joint constraints on the velocity dispersion @xmath56 and truncation parameter @xmath57 for a fiducial galaxy with , under the assumption that the luminosity does not evolve with redshift .
conveniently , for this particular luminosity , the inferred values of the velocity dispersion and truncation parameter depend only marginally on the adopted scaling relations . for the velocity dispersion
we obtain a value of @xmath69 km / s ( 68% confidence , marginalizing over all other model paremeters ) .
hoekstra et al .
( 2003 ) find a velocity dispersion of @xmath70 for a galaxy with a luminosity of .
for a galaxy with that luminosity , the adopted scaling relations imply a velocity dispersion of @xmath71 km / s , in excellent agreement with hoekstra et al .
mckay et al . ( 2001 ) who found a best fit value of @xmath72 km / s ( 95% confidence ) for a galaxy with @xmath73 .
our results correspond to a velocity dispersion of @xmath74 km / s for a galaxy of that luminosity , in good agreement with the sdss result .
we derive tight constraints on the truncation parameter , i.e. the extent of dark matter halos .
we find a value of @xmath75 kpc ( 68% confidence ) , and a total mass @xmath76 .
the results presented in hoekstra et al .
( 2003 ) imply a value of @xmath77 kpc ( 68% confidence ) for their fiducial galaxy . for a galaxy with a luminosity of
, we obtain @xmath78 kpc , marginally consistent with hoekstra et al .
( 2003 ) .
if we assume that @xmath43 we obtain a velocity dispersion of @xmath79 km / s and a truncation size of @xmath80 kpc for a galaxy with .
numerical simulations of collisionless cold dark matter ( cdm ) reproduce the observed structure in the universe remarkably well .
furthermore these simulations suggest that cdm gives rise to a specific density profile , which fits the radial mass distribution for halos with a wide range in mass ( e.g. , dubinski & carlberg 1991 ; navarro et al .
1995 , 1996 , 1997 ) .
the nfw density profile is characterized by 2 parameters , a density contrast @xmath81 and a scale @xmath82 @xmath83 where @xmath84 is the critical surface density at the redshift of the halo .
the `` virial '' radius @xmath85 is defined as the radius where the mass density of the halo is equal to @xmath86 , and the corresponding mass @xmath87 inside this radius is given by @xmath88 with a corresponding rotation velocity @xmath89 of @xmath90 the concentration parameter is defined as @xmath91 , which yields an expression for the overdensity of the halo @xmath81 in terms of @xmath92 @xmath93 it is important to note that the density profile on small scales remains controversial as other groups find different results ( e.g. , moore et al .
1999 ; ghigna et al .
in addition , realistic simulations should include the effect of baryons , which complicate matters even further ( see e.g. , mo , mao & white 1998 ; kochanek & white 2001 ) .
unfortunately the current rcs data do not allow us to constrain the slope of the inner mass profile .
however , future , deep lensing surveys , such as the cfht legacy survey , will be well suited for such a study . on the other hand , there is good agreement for the density profile on large scales , and a comparison of the profiles of real objects with the predictions provides an important test of the assumption that structures form through dissipationless collapse .
the predicted profiles agree well with the observed mass distribution in clusters of galaxies ( e.g. , hoekstra et al .
2002d ) , but the situation is less clear for galaxy mass halos .
rotation curves can provide some constraints , but typical values for @xmath82 for galaxy mass halos are @xmath94 kpc , comparable to the outermost point for which rotation curves have been measured .
galaxies that are thought to be dark matter dominated , such as low surface brightness galaxies , potentially might be more suitable to test the cdm predictions .
studies of the rotation curves of low surface brightness galaxies suggest that , at least for a fraction of them , the observed rotation curves rise more slowly than the cdm predictions ( e.g , de blok , mcgaugh & rubin 2001 ; mcgaugh , barker & de blok 2002 ) .
it is not clear , however , whether such studies provide a good test of cdm , because low surface brightness galaxies are peculiar ( zwaan & briggs 2000 ) , and their formation is not well understood .
hence , it is not obvious that their halos should be described by an nfw profile .
recently ricotti ( 2003 ) has suggested that the inner slope might depend on halo mass , with low mass systems having shallow cores , whereas massive galaxies are well described by the nfw profile . in this section
we compare the nfw profile to the observations , with @xmath81 and @xmath82 as free parameters .
the equations describing the shear for the nfw profile have been derived by bartelmann ( 1996 ) and wright & brainerd ( 2000 ) . as before we have to adopt scaling relations .
we assume that the maximum rotation velocity scales @xmath95 ( the b - band tully - fisher relation ; verheijen 2001 ) . if we also assume that @xmath96 , as motivated by the findings of guzik & seljak ( 2002 ) , we obtain that @xmath97 and @xmath98 .
figure [ size_nfw ] shows the joint constraints on @xmath89 and @xmath82 for a galaxy with a luminosity of .
in addition , the right axis indicates the corresponding values for @xmath87 .
we derive a best fit value of @xmath99 km / s , or @xmath9 ( 68% confidence ) , and a corresponding value of @xmath100 kpc .
it is useful to compare this result with the mass from the tis model .
the tis model yields @xmath101 , which is slightly larger than the nfw value . as shown by wright & brainerd ( 2000 ) ,
isothermal models give higher masses compared to nfw models .
hence , the results derived from both models are consistent . from their galaxy - galaxy lensing analysis of the sdss ,
guzik & seljak ( 2002 ) find @xmath102 for a galaxy of @xmath103 , in good agreement with our results . for the scale @xmath82
we find @xmath10 kpc ( 68% confidence ) , and the best fit density contrast is @xmath104 ( 68% confidence ; confidence interval from monte carlo simulation ) .
the tis model provides a slightly better fit to the data , but the difference is not significant , and consequently the data are not sufficient to distinguish between the nfw and tis model . as before , we also calculated the results under the assumption that the luminosity evolves @xmath105 . in this case
we find a value of @xmath106 kpc and @xmath107 km / s for a galaxy . in our maximum likelihood analysis we considered @xmath82 and @xmath89 ( or equivalently the mass @xmath87 ) as free parameters .
numerical simulations of cdm , however , show that the parameters in the nfw model are correlated , albeit with some scatter . as a result
, the nfw model can be considered as a one - parameter model : given the cosmology , redshift , and one of the nfw parameters , the values for all other parameters can be computed using the routine charden made available by julio navarrojfn / charden ] . hence , the simulations make a definite prediction for the value of @xmath89 as a function of @xmath82 . the dotted line in figure [ size_nfw ] indicates this prediction .
if the simulations provide a good description of dark matter halos , the dotted line should intersect our confidence region , which it does .
this result provides important support for the cdm paradigm , as it predicts the correct `` size '' of dark matter halos .
it is important to note that this analysis is a direct test of cdm ( albeit not conclusive ) , because the weak lensing results are inferred from the gravitational potential at large distances from the galaxy center , where dark matter dominates .
most other attempts to test cdm are confined to the inner regions , where baryons are , or might be , important .
the average shape of dark matter halos can provide important information about the nature of dark matter .
numerical simulations of cold dark matter yield triaxial halos , with a typical ellipticity of @xmath108 ( e.g. , dubinski & carlberg 1991 ) . hence
, in the context of collisionless cold dark matter , the theoretical evidence for flattened halos is quite strong .
if the dark matter is interacting , it tends to produce halos that are more spherical ( compared to cold dark matter ) .
this difference is more pronounced in the central parts of the halo , where the density is high . on the large scales probed by weak lensing , the different types of dark matter ( for reasonable interaction cross - sections ) produces halos with similar shapes . nevertheless , a measurement of the average shape of dark matter halos is important , because the observational evidence is still limited .
dynamical measurements are limited by the lack of visible tracers , and therefore only probe the vertical potential on scales @xmath109 15 kpc . although the spread in inferred values for the axis ratio @xmath110 ( where @xmath110 is the ratio of the shortest to longest principle axis of the halo ) is large , the results suggest an average value of @xmath111 ( sackett 1999 ) .
weak gravitational lensing is potentially the most powerful way to derive constraints on the shapes of dark matter halos .
the amount of data required for such a measurement , however , is large ( e.g. , brainerd & wright 2000 ; natarajan & refregier 2000 ) : the galaxy - galaxy lensing signal is small , and now one needs to measure an even smaller azimuthal variation
. we also have to assume that the galaxy and its halo are aligned . an imperfect alignment between light and halo
will reduce the amplitude of the azimuthal variation detectable in the weak lensing analysis .
hence , weak lensing formally provides a lower limit to the average halo ellipticity .
brainerd & wright ( 2000 ) and natarajan & refregier ( 2000 ) proposed to study the azimuthal variation in the tangential shear around the lenses .
on very small scales , the lensing signal is dominated by the lens , but on larger scales , the clustering of the lenses will lower the signal one tries to measure ( the two point function is axisymmetric ) .
we therefore use the maximum likelihood approach used in the previous section . to maximize the signal - to - noise ratio of the measurement one has to assign proper weights to the lenses : edge - on galaxies
have maximal weight , whereas the lensing signal around face - on galaxies contains no information about the shape of the halo .
we adopt a simple approach , and assume that the ( projected ) ellipticity of the dark matter halo is proportional to the shape of the galaxy : @xmath112 .
the measurement of the azimuthally averaged tangential shear around galaxies is robust against residual systematics ( e.g. , imperfect correction for psf anisotropy ) : contributions from a constant or gradiant residual shear cancel .
this is no longer the case for the quadrupole signal , and imperfect correction for the psf anisotropy can mimick the signal from a flattened halo .
if the lens galaxy is oriented randomly with respect to the residual shear , the average over many lenses will cancel the contribution from systematics . in real data , however , the uncorrected shapes of the lenses are aligned with the systematic signal . hence , an imperfect correction can give rise to a small quadrupole signal , although we note that the lenses used in our analysis are large compared to the psf .
we estimate the amplitude of this effect in appendix a , and show that it is negligible for the measurements presented here .
we also examined the robustness of our results by splitting the data into two samples and comparing the results .
we use an elliptical tis model to compute the model shear field , and compare this to the data .
figure [ flat_halo ] shows the resulting @xmath113 as a function of @xmath4 .
we find a best fit value of @xmath5 ( 68% confidence ) .
this suggests that , on average , the dark matter distribution is rounder than the light distribution .
as discussed above , our analysis formally provides only a lower limit on the halo ellipticity , and the true ellipticity might be higher if some of the halos are misaligned with the light .
nevertheless , the fact that we detect a significant flattening implies that the halos are well aligned with the light distribution .
also note that the lensing signal is caused by a range of different galaxy types , for which our simple relation between the halo ellipticity and light distribution might not be valid .
consequently the interpretation of the results is difficult , although a simple interpretation actually yields sensible results .
for instance , the average ellipticity of the lens galaxies is @xmath114 .
hence , the measured value of @xmath4 implies an average projected halo ellipticity of @xmath115 ( 68% confidence ) , which corresponds to an projected axis ratio of @xmath116 ( 68% confidence ; where we have used @xmath117 ) . although the weak lensing yields a projected axis ratio , the result is in fair agreement with the results from numerical simulations .
a robust outcome of our analysis is that spherical halos ( @xmath118 ) are excluded with 99.5% confidence . as we demonstrate below , this poses serious problems for alternative theories of gravity , which attempt to explain the observations without dark matter . in this section
we examine the implications of our measurement of the anisotropy in the lensing signal around galaxies for theories of gravity without dark matter .
we focus on one particular approach : modified newtonian dynamics ( mond ; milgrom 1983 ; sanders 1986 ; sanders & mcgaugh 2002 ) , which has been shown to describe rotation curves rather well ( e.g. , begeman et al . 1991 ; sanders & verheijen 1998 ) . in principle
weak lensing can be used as a powerful test of mond , but unfortunately no relativistic description of mond has been found . consequently one can not compute the lensing signal in this theory . however , even in the absence of an appropriate description of lensing , we can use the observed anisotropy in the lensing signal around galaxies to test mond .
in any reasonable alternative theory of gravity , the anisotropy in the lensing signal of an _ isolated _ galaxy is caused by the distribution of light and gas in that galaxy . in order to explain the flat rotation curves , these alternative theories typically need an effective force law @xmath119 .
hence , on small scales one expects an anisotropic signal , but at large radii ( where there are no stars and gas ) we assume that the anisotropy in the lensing signal decreases @xmath120 . as a result , these theories predict an almost isotropic weak lensing signal on the scales probed by our analysis , which is not observed .
galaxies , however , are not isolated and the external field effect ( milgrom 1986 ) might complicate the interpretation of our measurements . in mond , if a galaxy is embedded in an external field , this field dominates the dynamics if its acceleration is larger than the acceleration of the galaxy . as a result , the effective gravitational field is non - spherical , even if the potential of the galaxy is isotropic ( as is the case for an isolated galaxy ) .
this effect is important if galaxies would be aligned with this external field .
we know , however , that the intrinsic alignments of galaxies are small ( e.g. , lee & pen 2001 , 2002 ) and for the measurements presented here , it is safe to assume that the lenses have random orientation with respect to any external field .
consequently , the observed anisotropy in the lensing signal can not be caused by the external field effect .
hence , our findings provide strong support for the existence of dark matter , because alternative theories of gravity predict an almost isotropic lensing signal .
better constraints can be derived from future weak lensing surveys , which will allow us to study the anisotropy a function of projected distance from the galaxy .
we have analysed the weak lensing signal caused by a sample of lenses with @xmath2 using 45.5 deg@xmath0 of @xmath1 band imaging data from the red - sequence cluster survey ( rcs ) .
we have studied the average mass profile around the lenses using a maximum likelihood analysis .
to this end , we considered two models for the halo mass profile : a truncated isothermal sphere and an nfw profile .
we have assumed ( observationally motivated ) scaling relations between the luminosity of the lens and the velocity dispersion and the extent of the halo .
the tis model yields a best fit velocity dispersion of @xmath121 km / s and a truncation radius @xmath75 kpc for a galaxy with a fiducial luminosity of .
alternatively , the best fit nfw model yields a mass @xmath122 and a scale radius @xmath10 kpc .
this value for the scale radius is in excellent agreement with predictions from numerical simulations for a halo of this mass .
we also present the first detection of the flattening of galaxy dark matter halos from weak lensing .
we use a simple model in which the ellipticity of the halo is @xmath4 times the observed ellipticity of the lens . we find a best fit value of @xmath5 ( 68% confidence ) , suggesting that the dark matter halos are somewhat rounder than the light distribution .
the fact that we detect a significant flattening implies that the halos are aligned with the light distribution .
given the average ellipticity of the lenses , this implies a halo ellipticity of @xmath115 ( 68% confidence ) , in fair agreement with results from numerical simulations of cdm .
this result provides strong support for the existence of dark matter , as an isotropic lensing signal is excluded with 99.5% confidence .
we are grateful to the anonymous referee whose comments have significantly improved the quality of this paper . the rcs project is partially supported by grants from the natural science and engineering science council of canada and the university of toronto to hkcy .
in this appendix we examine how residual systematics affect the measurement of the flattening of dark matter halos . a schematic overview of the situation is presented in figure [ res_fig ] .
the thin lines indicate the direction of residual systematics .
the residual shear has an amplitude @xmath123 and a position angle @xmath124 with respect to the major axis of the lens .
the tangential shear @xmath125 observed at a position @xmath126 is the sum of the lensing signal @xmath127 and the contribution from systematics @xmath128 .
the latter is given by hence , the azimuthally averaged tangential shear is not affected by systematics as @xmath130 , and the weak lensing mass estimate is very robust . for a flattened halo ,
the lensing signal @xmath127 is given by we consider the worst case scenario , and demonstrate that even in this situation the results are robust .
one way to estimate the flattening of the halo is to measure the shears @xmath134 ( at @xmath135 and @xmath136 ) , and @xmath137 ( at @xmath138 and @xmath139 ) .
the observed ratio @xmath140 is if @xmath124 is uncorrelated with the lens , the observed ratio , averaged over many lenses , is unaffected by systematics because @xmath142 .
however , in real data , the psf anisotropy affects both the lens and the source galaxies .
although the lenses used in this paper are large compared to the psf , any ( small ) residual in the correction will introduce a correlation in the position angle of the lens and the direction of the psf anisotropy .
the effect is maximal for @xmath143 ( aligned with the major axis of the lens ) or @xmath144 ( perpendicular to major axis of the lens ) .
the latter situation corresponds to an overcorrection of the psf anisotropy , whereas the former occurs when the correction for psf anisotropy is too small .
for both ( extreme ) situations , only the observed value of @xmath145 is affected by systematics ( @xmath146 is left unchanged ) .
the observed value @xmath147 for the lens can be written as @xmath148 , where @xmath149 is a measure of the correlation between the position angle of the lens and the direction of the systematic shear of the background galaxies . in our case @xmath150 , because the lenses are large compared compared to the psf .
for a lens with with an observed @xmath151 , we find this introduces a relatively large signal because face - on lenses ( @xmath155 ) align easily with the psf anisotropy .
such lenses , however , contain no information about the shape of the halo . in our analysis
we assign a weight @xmath156 to each lens . hence
in our case we are less sensitive to systematics as the correct estimate is given by we can now estimate how robust the measurement of the average halo shape is .
we take a conservative estimate of @xmath160 , and @xmath161 ( see hoekstra et al .
2002a for a discussion of the residuals in our data ) .
the average separation of stars in our data is @xmath162 and therefore the residual psf anisotropy is likely to change direction on scales larger than the separation of the stars used to measure the anisotropy .
the average tangential shear at 2 arcminutes is @xmath163 . for a spherical halo ( @xmath164 )
we would observe a ratio @xmath165 .
this corresponds to an ellipticity of 2.5% ( with the halo oriented perpendicular to the lens ) , which is small ( @xmath166 10% of the observed ellipticity ) . in this very conservative estimate
, we have assumed that for each lens , the residual psf anisotropy is aligned with the lens ( i.e. , the psf anisotropy was underestimated ) .
there is , however , an equal probability of overestimating the psf anisotropy .
hence , the estimate is a very conservative upper limit for the importance of systematics .
such a systematic underestimate for psf anisotropy would give rise to very large `` e '' and `` b''-modes in the cosmic shear measurements , which are not observed . hence , in reality the change in the ellipticity of the halo , caused by systematics , is much smaller than our conservative estimate . | we present the results of a study of weak lensing by galaxies based on 45.5 deg@xmath0 of @xmath1 band imaging data from the red - sequence cluster survey ( rcs ) .
we define a sample of lenses with @xmath2 , and a sample of background galaxies with @xmath3 .
we present the first weak lensing detection of the flattening of galaxy dark matter halos .
we use a simple model in which the ellipticity of the halo is @xmath4 times the observed ellipticity of the lens . we find a best fit value of @xmath5 , suggesting that the dark matter halos are somewhat rounder than the light distribution .
the fact that we detect a significant flattening implies that the halos are well aligned with the light distribution .
given the average ellipticity of the lenses , this implies a halo ellipticity of @xmath6 , in fair agreement with results from numerical simulations of cdm .
we note that this result is formally a lower limit to the flattening , since the measurements imply a larger flattening if the halos are not aligned with the light distribution .
alternative theories of gravity ( without dark matter ) predict an isotropic lensing signal , which is excluded with 99.5% confidence .
hence , our results provide strong support for the existence of dark matter .
we also study the average mass profile around the lenses , using a maximum likelihood analysis .
we consider two models for the halo mass profile : a truncated isothermal sphere ( tis ) and an nfw profile .
we adopt observationally motivated scaling relations between the lens luminosity and the velocity dispersion and the extent of the halo .
the tis model yields a best fit velocity dispersion of @xmath7 km / s ( all errors are 68% confidence limits ; the first error bar indicates the statistical uncertainty , whereas the second error bar indicates the systematic error ) and a truncation radius @xmath8 kpc for a galaxy with a fiducial luminosity of ( under the assumption that the luminosity does not evolve with redshift ) .
alternatively , the best fit nfw model yields a mass @xmath9 and a scale radius @xmath10 kpc .
this value for the scale radius is in excellent agreement with predictions from numerical simulations for a halo of this mass .
= -15 mm |
instead of giving a table of contents , let me just say that the organizational plan of the talk is to explain the words in my title , the first of which is _
symmetries_. let us start at the very beginning and recall what is meant by a symmetry in the context of quantum mechanics .
following wigner ( who was awarded the 1963 physics nobel prize for his foundational work on symmetry principles and their application to nuclear and particle physics ) a symmetry in quantum mechanics is primarily a transformation on the rays of hilbert space with the property that all transition probabilities are preserved .
now one corner stone of our subject is a theorem ( attributed to wigner ) stating that any quantum mechanical symmetry lifts to hilbert space as a linear operator which is either unitary , or anti - unitary . in the former case the hermitian scalar product of hilbert space
is preserved , in the latter case it is preserved up to complex conjugation .
i wish to make two remarks here .
first comes the obvious statement that symmetries always form a group , @xmath1 .
indeed , if two operators @xmath2 and @xmath3 are symmetries , then so is their composition @xmath4 .
the second remark is that in order for an operator @xmath5 to be a symmetry of a quantum system with hamiltonian @xmath6 , we require that @xmath5 commutes with @xmath6 .
thus so - called chiral symmetries , which _
anti_-commute with the hamiltonian a prominent example is the chirality operator @xmath7 which anti - commutes with the massless dirac operator do _ not _ qualify as symmetries , at least not in the sense of this talk .
( i must emphasize this point because my work on symmetry classification is often cited in an abridged form that overlooks or ignores this aspect . ) now , why should we care about any symmetries ? we all know
, of course , that symmetries in spectroscopy lead to selection rules and relations among the transition rates ; and in the case of integrable systems they are the key feature responsible for integrability .
however , here our interest is in the exact opposite , namely disorder and quantum chaos .
let me give two examples to illustrate that symmetries do matter even in that case .
first , i would like to return to the theme of universal conductance fluctuations and focus on its symmetry aspects , referring to the cartoon in fig .
2a for illustration . by assuming the single - electron approximation and using that the electrical conductance @xmath8 is essentially a sum of squares of transmission amplitudes
, we can make a picture of it as a sum over semi - classical paths crossing the disordered metal . taking ucf to be a statement about
the _ variance _ of the fluctuations , we consider two pairs of paths for the transmission amplitudes and their conjugates . we also arrange them in a way that makes our graph phase - insensitive and thus immune to the process of disorder averaging .
the leading contribution to the variance is due to a form of ` promiscuity ' of our electron paths : after passing from the leads into the disordered region , the electron paths exchange partners and diffuse with the alien conjugate for a while before they rejoin their proper mate in preparation for exit . to compute the variance
, we integrate over the promiscuous diffusion legs and the two positions of adultery and re - marriage .
this amounts to doing a momentum integral involving two diffusion propagators @xmath9 , and as is well known , the result is of the order of unity , independent of the system size @xmath10 and other system - specific features : @xmath11 to pinpoint the role of symmetries , let us review a little bit of theory . on very elementary grounds ,
the green operator @xmath12 satisfies the resolvent identity @xmath13 .
now if our single - electron hamiltonian is hermitian and conserves particle number ( a property that can be viewed as a consequence of a global @xmath14 gauge symmetry ) then the complex conjugate of the retarded green s function is equal to the advanced green s function with the initial and final states exchanged : @xmath15 . by closure and the resolvent identity
, this relation results in a sum rule for the sum of squares : @xmath16 and the fourier transform of the disorder - averaged squared amplitude has the form of a massless diffusion propagator .
thus we see that the @xmath14 symmetry of particle number conservation gives rise to a massless mode , namely the `` diffuson '' . what happens if the system has more symmetries ?
well , then we should expect more massless modes ! for example
, if magnetic fields are absent and the disordered system is invariant under the operation of reversing the time direction , then the green s function remains the same when we exchange the initial state with the time - reversed final state . by the line of reasoning of before , this symmetry relation gives rise to another sum rule and
hence another massless mode called the `` cooperon '' .
returning to our graphical illustration , in the presence of time - reversal symmetry there exists a second type of graph ( fig .
2b ) , which is different from the earlier one but still immune to phase cancelations due to disorder averaging . in this
ucf graph , the promiscuous legs are cooperons , as the two arrows indicating the direction of motion point the opposite way , whereas previously they pointed the same way .
its numerical contribution to the variance is the same as that of the other graph featuring the diffuson .
thus the outcome is that the symmetry under time reversal makes the variance double .
let me move on to a second example , namely that of disordered ` symplectic ' wires , meaning electrons in a quasi - one dimensional geometry with spin - orbit scattering and , again , time - reversal invariance . here
the symmetries of the disordered quantum system turned out to be of truly spectacular consequence .
for concreteness , let us consider a model of dirac fermions in one space dimension with coordinate @xmath17 .
the model is built from @xmath18 right - moving modes ( or channels ) coupled by a hermitian random matrix @xmath19 , @xmath18 left - moving channels coupled by the transpose of @xmath20 , and a skew - symmetric random matrix @xmath21 causing backscattering , i.e. , coupling between the right- and left - movers : @xmath22 where @xmath23 and @xmath24 .
the particular form and arrangement of the random matrices @xmath20 and @xmath25 places this model in what i call `` class @xmath20ii '' .
a hallmark of this class is kramers degeneracy of the energy eigenvalues .
now , for gaussian random variables and large channel number @xmath18 , the disorder averaged observables of this model can be computed by the mapping to a nonlinear sigma model due to wegner and efetov . in a letter published in 1992 @xcite ,
i solved the sigma model to derive an exact expression for the average electrical conductance @xmath26 of a wire of length @xmath10 .
the striking feature of my result , which was controversial ( and in fact considered to be unphysical ) at that time , is that the conductance does not go to zero with increasing length as one might have expected from the standard scenario of exponential localization in one dimension , but approaches a constant ( @xmath27 in units of @xmath0 ) .
one decade later a japanese colleague @xcite made perfect sense of this result by pointing out that the large-@xmath18 limit i had computed was a mixture of contributions from even and odd channel number @xmath18 , and that one really should separate even from odd .
thus the good way of presenting and interpreting my result is in a split form : @xmath28 where for brevity i am not showing the exact result but just the leading behavior for a long wire . for even @xmath18 exponential localization does hold , but for odd @xmath18 there is one perfectly conducting channel ( or zero mode ) .
let me emphasize that the existence of this mode is _ not _ an artifact of the dirac approximation used , or the limit of large channel number , or any other specific feature .
rather , it is a universal and inevitable consequence of time - reversal symmetry for spin-1/2 electrons in combination with the channel number being odd . as a matter of fact ,
the zero mode that showed up in my super - fourier analysis is nothing but the celebrated edge state of the quantum spin hall insulator !
we nowadays understand that this mode is robust on stringent topological grounds and , in particular , is stable with respect to disorder .
( sadly , i did not know about the topological significance of odd channel number at the time when my work was done . )
so much for symmetries and how in disordered systems they give rise to massless modes with observable consequences for transport .
on to the next word in my title : _
symmetry classes_. here i must warn you that i need to be moving rapidly and will be rather brief . for a more comprehensive and leisurely account of this story , you may want to consult my web pages where you will find a colloquium style talk with slides and text included . in short , by `` symmetry class ''
i mean an organizational principle which is relatively new and was _ not _ part of the vocabulary of mesoscopic physics fifteen years ago .
at that time , one spoke of _ universality classes_. to establish the physical context , let me recall a striking statement referred to as the random - matrix universality conjecture for spectral fluctuations .
it is this : take any of the linear equations of wave mechanics , be it the schrdinger equation , the dirac equation or , for that matter , the wave equation of maxwell electrodynamics , and look at the spectrum of energy levels .
if there is enough disorder or chaos , then you will see level fluctuations which obey the laws predicted by random matrix theory for the appropriate symmetry class and in a certain limit called the ergodic regime .
this finding has been made in numerous examples , beginning with the neutron resonances observed in compound nucleus scattering . to single out one example
, numerical diagonalization shows that the energy levels of the sinai billiard , a quantum chaotic system with two degrees of freedom , exhibit the level correlations of the so - called gaussian orthogonal ensemble , i.e. , a real symmetric random matrix .
this example was of great importance for the historical development , as it led bohigas , giannoni , and schmit in 1984 to state the said universality conjecture .
@xcite the classification principle behind this kind of universality had been formulated some 20 twenty years earlier by dyson in a famous paper @xcite called `` threefold way '' .
dyson proved that , in his own words , `` the most general kind of matrix ensemble , defined with a symmetry group which may be completely arbitrary , reduces to a direct product of independent irreducible ensembles each of which belongs to one of the three known types . ''
dyson had asked the following question . given a finite - dimensional hilbert space and a general symmetry group acting on it , how do the hamiltonians look that commute with all the symmetries ?
his answer for the irreducible blocks was that there exist but 3 possibilities : the matrix of the hamiltonian is either complex hermitian , or real symmetric , or has entries that are quaternions . it should be emphasized that this organization by what we nowadays call the wigner - dyson symmetry classes is very coarse and relies on nothing but linear algebra .
in fact , a symmetry class is simply a type of real vector space , or linear space , for the hamiltonian matrix to be in .
it is only by putting probability measures on these matrix spaces the simplest ones are the so - called gaussian measures or gaussian ensembles and analyzing the statistical correlations that one discovers the universality mentioned earlier .
thus the notion of _ symmetry class _ is a notion more basic and more primitive than the notion of _ universality class_. now around the middle of the 1990s , there was an accumulation of evidence that some extension or variant of dyson s threefold way was called for .
first of all , it was found that the eigenvalue statistics of the massless dirac operator which , for example , is relevant for quantum chromodynamics , could be modeled by so - called chiral ensembles of a type beyond dyson s classification .
@xcite second , in mesoscopic physics the process of andreev reflection at the boundary between a normal metal and a superconductor was found to give rise to quantum interference phenomena not present in the wigner - dyson symmetry classes .
@xcite third , mathematicians had placed the riemann zeta function in an ensemble of similar functions and found statistical phenomena which did not fit into the wigner - dyson scheme .
@xcite let me single out the case of superconductors for a little bit of detail .
the theoretical background here is that the gorkov - green operator satisfies a set of relations ( even when there are no physical symmetries at all ) which are just due to the canonical anti - commutation relations for fermions . by similar reasoning as before , these relations give us a sum rule which in turn gives rise to what altland and i called the @xmath29-type diffuson .
@xcite the ` mass ' of this mode goes to zero as the energy of the quasi - particle approaches the chemical potential , and its strength is given by the low - energy density of states .
let us come back , once more and now for the final time , to the subject of universal conductance fluctuations and our cartoon , featuring this time the @xmath29-type diffuson ( fig .
the novel feature here is that the promiscuous legs are formed by pairing retarded with retarded and advanced with advanced green s functions . to achieve experimental verification of the @xmath29-type diffuson mode ,
it is not a good idea to look for its effects in the transport of electric charge , which is shunted in a superconductor by the condensate .
however , for a spin - singlet superconductor the effect can be detected by measuring the spin transport , and if the condensate carries spin , then one can still look at the heat transport due to the quasi - particles .
after this example , let us get on to the systematics .
based on the classification of massless modes , altland and myself argued that in addition to the 3 classes of wigner - dyson and the 3 chiral classes of verbaarschot , @xcite there were 4 classes which can be realized in superconductors , or in metals in proximity with superconductors , thus raising the count to a grand total of ten .
we also claimed that this was it , and `` no further classes will be found '' .
later , in joint work with two colleagues from bochum mathematics , this conjecture was given a precise formulation and proved .
here is an outline .
first we have to specify the rules of the game .
we refine the setting of the threefold way by replacing dyson s general hilbert space by the more elaborate structure of a fock space for fermions . following dyson
, we adopt the setting of a symmetry group acting on fock space by unitary and anti - unitary operators .
we require the group of unitary symmetries to be defined on the single - particle space @xmath30 and extend it to fock space @xmath31 in the natural way .
( this requirement , i should say , excludes yangian and other quantum group symmetries , which arise at the many - particle level . )
there is no further restriction ; in dyson s words : the group of unitary symmetries may be completely arbitrary . as for the anti - unitaries , we allow for the presence of time - reversal symmetry , which is defined on the single - particle hilbert space and extends to fock space in the usual way .
moreover , the structure of fock space opens the possibility for another anti - unitary operation to be a symmetry ; this is particle - hole conjugation , transforming the particle vacuum into the fully occupied state , and in general , a state of @xmath32 particles into a state of @xmath32 holes .
we can now formulate the problem to be solved .
let there be any fermionic fock space carrying a @xmath33-action , where @xmath33 is an arbitrary symmetry group made from generators as described above ; i.e. , with the most general element @xmath34 being any combination of unitary symmetry operations and/or time reversal and/or particle - hole conjugation . our object of interest is the set @xmath35 of all hamiltonians which are @xmath33-invariant one - body operators , i.e. , operators that are _ quadratic _ in the particle creation and annihilation operators and _ commute _ with all operators from the symmetry group @xmath33 .
the question then is : what can we say about the structure of the set @xmath35 ?
can we classify the types of irreducible block which occur in this setting ? after toying with this question for a number of years , the final and definitive answer was given in a paper with heinzner and huckleberry : @xcite every irreducible block which occurs in this setting corresponds a classical irreducible symmetric space , and conversely , every classical irreducible symmetric space arises in this way .
( more precisely , the time - evolution operators are in a compact symmetric space , while the hamiltonian are in a euclidean symmetric space that arises by linearization . ) before explaining what is meant by a symmetric space , let us do a quick example , namely that of class @xmath29 , where one has no symmetries at all .
if we introduce a basis of majorana operators @xmath36 and @xmath37 ( @xmath38 ) , we can express our general one - body hamiltonian in the simple form @xmath39 with imaginary coefficients @xmath40 that constitute a skew - symmetric matrix . by exponentiating the hamiltonian one gets time evolution operators @xmath41 in the even - dimensional orthogonal group , @xmath42 , which is called a symmetric space of type @xmath29 ( hence the name class @xmath29 ) .
experimental realizations of this class occur in disordered superconductors with spin - triplet pairing and @xmath43-breaking @xmath44-wave order a much debated candidate is @xmath45 .
a non - charged realization is the @xmath20-phase of @xmath46 .
moreover , a number of other realizations have recently been proposed , @xcite motivated by the far goal of engineering a topological quantum computer .
finally , to give some meaning to the statement of our theorem , let me explain the last word of the title , namely _
symmetric spaces_. recall that in riemannian geometry there exists an object called the riemann curvature tensor . by definition ,
a symmetric space is a riemannian manifold @xmath47 with a riemann curvature tensor which is covariantly constant : @xmath48 .
the simplest example of a such a space is the round two - sphere @xmath49 with line element induced by the euclidean distance of three - dimensional space ; in spherical polar coordinates this is @xmath50 .
another example is the set @xmath51 of all complex @xmath32-dimensional subspaces in @xmath52 dimensions .
such a space is called a grassmann manifold , @xmath53 , or grassmannian for short .
note that any two @xmath32-dimensional subspaces can be mapped into each other by a unitary transformation of the full space .
since nothing changes when we just transform within a fixed subspace and its orthogonal complement , one has the identification @xmath54 of the grassmannian with a quotient of unitary groups .
here are a few facts about symmetric spaces .
for one , they were completely classified by the french geometer elie cartan .
apart from a finite number of exceptional spaces , they come in 10 large families , which cartan called @xmath20 , @xmath20i , @xmath20ii , @xmath20iii , @xmath55 , @xmath55i , @xmath8 , @xmath8i , @xmath8ii , and @xmath29iii . for
another fact , let me note that any @xmath33-invariant symmetric @xmath56-tensor on a symmetric space @xmath57 must be a constant multiple of the metric tensor . there exist several places in physics where symmetric spaces appear .
for one , they are natural candidates for order parameter spaces .
( indeed , when the symmetry group @xmath33 of a physical system spontaneously breaks down to a subgroup @xmath58 , the order parameter space is the quotient @xmath57 . ) for another , symmetric spaces serve as target spaces for a certain type of field theory , the so - called nonlinear sigma models .
these models have the important property of being one - parameter renormalizable in two dimensions , i.e. , under a change of short - distance cutoff of the field theory the metric tensor of the symmetric space ( defining the nonlinear sigma model ) changes only by a multiplicative constant .
this follows from the mentioned fact that the metric is the only @xmath33-invariant tensor of its kind .
third , i have already spoken about the correspondence between symmetric spaces and symmetry classes . and
last but not least , many - fermion ground states in the mean - field approximation of hartree - fock - bogoliubov organize into symmetric spaces .
let me now elaborate on the last point , as this will lead us to the final theme of the talk : topological insulators . as before , let @xmath59 denote the fock operator creating a single particle in the state labeled by @xmath60 , and let @xmath61 be the corresponding annihilation operator .
the fock vacuum is uniquely characterized by the property of being annihilated by all of the annihilation operators : @xmath62 ( @xmath63 ) .
now define the notion of _ quasi - particle vacuum _
@xmath64 by an analogous property : @xmath65 where the new operators @xmath66 are the result of making a bogoliubov transformation : @xmath67 .
such vacua are also referred to as many - body ground states in the hartree - fock - bogoliubov mean - field approximation .
note in particular that by specializing to @xmath68 we get an @xmath18-particle slater determinant state . in the presence of a group @xmath33 of symmetries
we require the quasi - particle vacuum state to be invariant under the @xmath33-action .
( the context determines whether or not this is a reasonable condition to impose . in any case , it is part of the physical setting and mathematical statement i am driving at . )
let us now fix some symmetry group @xmath33 , but keep the parameters of the system variable .
as the mean - field hamiltonian varies , so does its ground state , the quasi - particle vacuum @xmath69 . omitting the irrelevant overall phase factor
, we introduce the notation @xmath70 for the ray of the ground state .
the space of quasi - particle vacua can be made into a riemannian manifold @xmath51 by taking the geodesic distance between two such vacua @xmath71 and @xmath72 to be the positive number @xmath73 if the ray of one is obtained from the ray of the other by applying the exponential of an anti - hermitian particle - hole type operator @xmath20 of norm @xmath73 : @xmath74 for example , let us take @xmath33 to be the @xmath14 symmetry group underlying conservation of particle number . for that choice ,
quasi - particle vacua are @xmath18-particle slater determinants , and the riemannian manifold of such states is a grassmannian , @xmath75 .
indeed , an @xmath18-particle slater determinant is completely defined by specifying just its @xmath18-dimensional subspace of occupied states inside the single - particle hilbert space @xmath30 . moreover
, the geometry of @xmath76 leads to the geodesic distance above .
one may now pose the question : what can be said about the structure of @xmath51 in general ?
well , as a corollary of the theorem proved in @xcite , the answer is that for any symmetry group @xmath33 the manifold of @xmath33-invariant quasi - particle vacua is a symmetric space ( or more precisely : a product of classical irreducible symmetric spaces ) .
this is the punch line of my talk .
now , to make a connection with some of the physics discussed at this march meeting , let us see how the quantum spin hall insulator emerges from our picture .
we again assume that particle number is conserved and take the symmetry group @xmath33 to be generated by time reversal ( for spin-1/2 electrons ) and by a group @xmath78 of translations .
the fourier dual of the translation group , namely the ( first ) brillouin zone , is denoted by @xmath79 . in the two - dimensional case at hand ,
our quasi - particle vacua are hartree - fock mean - field ground states .
because momentum @xmath80 is conserved due to translation symmetry , we may describe these ground states by assigning to each value @xmath81 the corresponding vector space , @xmath82 , of occupied states ( or , using the language of solid state physics , the vector space spanned by the valence band states at momentum @xmath80 ) .
if the system is a band insulator , i.e. , the fermi energy sits in a gap , this assignment @xmath83 depends continuously on @xmath80 .
it is an example of what is called a `` vector bundle '' in mathematics . in the presence of time - reversal symmetry
we require that the time - reversal operator @xmath43 applied to a valence state of momentum @xmath80 yields a valence state at momentum @xmath84 .
for the special case of a @xmath43-invariant momentum @xmath85 it follows that the vector space of valence states is @xmath43-invariant : @xmath86 . by the properties of the anti - unitary operator @xmath43
this implies that the vectors of @xmath87 organize into so - called kramers pairs .
now there exists a mathematical tool called @xmath58-theory , which was developed in the 1950s ( in the general context of the atiyah - singer index theorem ) for the purpose of classifying vector bundles up to isomorphism . in our case ( and for @xmath88 , the case of a fluid where @xmath89 are the only @xmath43-invariant momenta ) @xmath58-theory tells us that there exist exactly two isomorphism classes of such vector bundles . physically speaking , we nowadays say that the topology of the situation gives us two distinct phases : the trivial band insulator phase and the quantum spin hall phase ( and every one of our hartree - fock mean - field fluid ground states belongs to one of the two ) . to distinguish between the two cases , one looks at a topological invariant that measures the twisting of the vector bundle
following kane and mel @xcite one associates with @xmath82 a certain pfaffian , and the argument of this pfaffian winds either an even or an odd number of times along a circle through @xmath90 and @xmath91 in @xmath80-space .
you might now be confused if not irritated .
i announced a story about symmetric spaces , but then i turned it into a story about vector bundles .
well , i am sorry , but unfortunately it is not always possible to get a full understanding from just one narrow perspective .
so let us change our perspective now and go to an alternative viewpoint leaning on symmetric spaces . recall that a choice of @xmath32-dimensional subspace @xmath92 determines a point @xmath93 . in the vector bundle picture we associated with each value
@xmath81 a vector space @xmath82 , but we may equivalently associate with each @xmath80-value the corresponding point in @xmath51 . thus instead of a vector bundle
we get a mapping , say @xmath94 .
the condition @xmath95 due to time - reversal symmetry translates into a condition @xmath96 . at @xmath43-invariant points @xmath97
this condition constrains @xmath98 to take values in a symplectic grassmannian : @xmath99 here you have it ; that s the symmetric space formulation . quite generally , in this mean - field picture , topological phases are classified by equivalence classes ( so - called homotopy classes ) of mappings @xmath98 from momentum space @xmath79 into a symmetric space @xmath51 subject to a condition @xmath100 for every element of the reduced group of symmetries which do not commute with the translations . in 2009
most if not all of the examples of topological insulators and superconductors known at that time were put into a `` periodic table '' by ludwig and co - workers , @xcite following kitaev .
@xcite related to a mathematical result called bott periodicity , this table seems to have caused quite a stir in the community ; in any case , it has been reprinted in numerous articles including a major review by hasan and kane .
@xcite let me finish with a few remarks on this .
first , the table , as wonderful as it is , does not give the complete story , as there exist symmetry groups of physical relevance which are not present in the scheme of @xcite .
second , the @xmath58 theory formalism used to compute it may miss some of the finer points of topology .
in fact some of the table s zero entries are not zero in all possible cases .
my final remark is that , to the best of my knowledge , there exists no mathematical proof of the table up to now ( but i believe that such will soon become available ) .
in this talk i tried to cover quite a lot of ground .
starting from the role of symmetries in systems with disorder or chaos , i introduced the notion of symmetry classes and symmetric spaces , and then described a one - to - one correspondence between these , the tenfold way . while conceived for the purpose of classifying disordered fermions , this scheme turned out to have the rewarding feature of being relevant for a topic of current interest , namely topological insulators .
altshuler , jetp letters * 41 * ( 1985 ) 648 p.a .
lee and a.d .
stone , phys .
lett . * 55 * ( 1985 ) 1622 p.a .
lee , a.d .
stone , and h. fukuyama , phys .
b * 35 * ( 1987 ) 1039 m.r .
zirnbauer , phys .
* 69 * ( 1992 ) 1584 y. takane , j. phys .
japan * 73 * ( 2004 ) 1430 o. bohigas , m.j .
giannoni , and c. schmit , phys .
* 52 * ( 1984 ) 1 f.j .
dyson , j. math .
* 3 * ( 1962 ) 1199 j.j.m .
verbaarschot and i. zahed , phys .
* 70 * ( 1993 ) 3852 a. altland and m.r .
zirnbauer , phys . rev .
b * 55 * ( 1997 ) 1142 n.m .
katz and p. sarnak , bull .
* 36 * ( 1999 ) 1 j.j.m .
verbaarschot , phys .
* 72 * ( 1994 ) 2531 p. heinzner , a. huckleberry , and m.r .
zirnbauer * 257 * ( 2005 ) 725 see j. alicea , arxiv:1202.1293 , for an overview . c.l . kane and e.j .
mele , phys .
* 95 * ( 2005 ) 146802 s. ryu , a.p .
schnyder , a. furusaki , and a.w.w .
ludwig , new j. phys .
* 12 * ( 2010 ) 065010 a. kitaev , aip conf .
proc . 1134 ( 2009 )
hasan and c.l .
kane , rev .
* 82 * ( 2010 ) 3045 | quantum mechanical systems with some degree of complexity due to multiple scattering behave as if their hamiltonians were random matrices . such behavior , while originally surmised for the interacting many - body system of highly excited atomic nuclei , was later discovered in a variety of situations including single - particle systems with disorder or chaos .
a fascinating theme in this context is the emergence of universal laws for the fluctuations of energy spectra and transport observables . after an introduction to the basic phenomenology
, the talk highlights the role of symmetries for universality , in particular the correspondence between symmetry classes and symmetric spaces that led to a classification scheme dubbed the `` tenfold way '' .
perhaps surprisingly , the same scheme has turned out to organize also the world of topological insulators .
let me begin by expressing that i feel greatly honored to be this year s recipient of the max - planck medal .
i also appreciate the opportunity to give a talk on some of the work that may have earned me this distinction . to set the stage and give you a flavor of what is to come , let me remind you of the old but still fascinating story of universal conductance fluctuations ( ucf ) .
predicted theoretically in the middle of the 1980s by altshuler @xcite and by lee and stone @xcite , ucf was investigated in a large number of experiments .
it was found that in a great variety of different mesoscopic systems such as a small gold ring for example , or an even smaller silicon mosfet the electrical conductance displays characteristic fluctuations of the order of one when expressed in units of the conductance quantum @xmath0 ( fig . 1 ) .
what is most remarkable is that the size of the fluctuations in a broad range of parameters does _ not _ depend on the system dimension , the disorder strength , etc . , but only on a few fundamental symmetries .
it was realized early on that there exists a close connection with the fluctuations that had been observed decades earlier in the scattering cross section of slow neutrons on atomic nuclei .
this far reaching connection is at the very root of what i have to say .
it led , among other things , to the development of a broad framework in which to model and calculate mesoscopic effects such as ucf . |
m giants are long - period variables requiring years of continuous observations for their study .
much of our recent knowledge was gained from microlensing surveys of the magellanic cloud and the galactic bulge , such as macho @xcite , ogle @xcite and eros @xcite . while analysing photometric data of red giants in the macho survey of the lmc , @xcite found several sequences in the period - luminosity ( p - l ) plane , which were labelled as a , b , c , e and d , representing shorter to longer periods .
subsequent studies have shown that these structures of the sequences is rich , with over a dozen features that are related to luminosity ( below or above the tip of the red giant branch - see , e.g. , @xcite ) and chemical composition ( carbon - rich vs. oxygen - rich , * ? ? ? * ) or might have a dependency on the wavelength range of the luminosity indicator @xcite .
the most distinct parallel sequences a and b represent the radial overtone modes of semiregulars ( sr ) .
these stars are numerous and most of them have multiple periods .
the miras lie on sequence c , which corresponds to the fundamental mode @xcite .
the power spectra of semiregulars that are observed for a large number of pulsation periods show modes with solar - like lorentzian envelopes @xcite .
this suggests that stochastic excitation and damping take place . with decreasing luminosities the pulsations decrease in amplitude and become more difficult to detect .
however , these also have shorter periods , making them good candidates for high - quality space photometry from _ corot _ and _ kepler_. in addition to the mentioned sequences of the srs and miras , there are two sequences in the p - l plane : sequence e and d representing the eclipsing binaries and the long secondary periods , respectively @xcite .
the latter remains unexplained @xcite .
although there is significant improvement in the understanding of m giant variability , there remain many question regarding the excitation and damping of the pulsations , and the expected crossover from mira - like to solar - like excitation @xcite which must take place in m giants . while the period - luminosity relations seem to be universal , regardless of the galactic environment ( see * ? ? ?
* ) , the full potential of these stars as tracers of the galactic structure is yet to be fully explored .
the presence of many modes of oscillation is expected to enable the application of asteroseismology for the most luminous giants @xcite , which may be affected by the mass - loss in the upper parts of the giant branch .
there has also been some controversy on the short - period microvariability of mira - like stars @xcite , and _ kepler _ might prove to be ideal for resolving this issue .
the complex light curves have also been interpreted in terms of stochasticity and chaos @xcite .
working group 12 ( hereafter wg12 ) of the kepler asteroseismic consortium @xcite was formed for the purpose of studying mira and semiregular pulsations in the _ kepler _ data .
here we present the first results obtained from the analysis of the wg12 sample .
the paper is organised as follows .
[ section : wg12sample ] presents a detailed description of the wg12 stars , which indclude the selection criteria .
[ section : dataanalysis ] describes the data analysis .
[ section : discussion ] presents the comparison of our results with ground - based photometry , the study of frequencies and amplitudes of light - curve variations and time - frequency analyses .
a brief summary is given in sect .
[ section : summary ] .
are plotted .
the bright k giants are confined between the two red vertical dashed lines , while the m giants were selected from the upper right region.,width=302 ] m giants are the longest period variable stars in the kasc program .
the typical time - scale of variability is of the order of one or two kepler quarters , which means removing instrumental drifts is potentially difficult . on the other hand ,
their amplitudes are above the usual instrumental effects , so correcting the m giant light curves should be a relatively simple task ( and involves essentially neglecting every systematic effect that goes beyond a constant vertical shift in the light curves from quarter to quarter ) . after having combined three years of _ kepler _ data
( we used the quarters q0 @xmath0 q12 ) , we can characterize m giant variability in a homogeneous and meaningful way .
the total sample includes over 300 m giant stars .
we had two lists of targets : one was initially selected from a dedicated northern asas3 variability survey of the kepler field @xcite , while the second one was selected from the kepler input catalogue @xcite .
the first set of m giant targets have been selected by combining @xmath1 and @xmath2 values from kic , the j@xmath0k colour from 2mass and the variability information from the asas3 survey .
we adopted @xmath3 k , @xmath4 and restricted the sample to kepler magnitude @xmath5 ( see in fig .
[ or an asas3 variability that is incompatible with a red giant ) . that resulted in 200 targets with variability information . since the asas3 variables are all cool and large amplitude stars , we created a second list of further 200 target candidates with warmer m giants randomly selected from kic10 with the same limits ( @xmath3 k and @xmath4 ) .
these stars are expected to show small - amplitude pulsations that were not detectable with the asas3 survey .
the final list of targets that was approved for observations by _ kepler _ contained 198 stars from the asas3 variable list and 119 from the other list .
most of these 317 stars have uninterrupted long - cadence ( one point per every 29.4 points and short gaps between the quarters ) coverage throughout q0 to q12 and their data were analysed using the process described below .
the _ kepler _ space telescope rolls 90 degrees every quarter of a year , and consequently , variability of the majority of target stars is measured by a different ccd camera and using a slightly different aperture every quarter in a cycle of a year . for m giants with variability time - scales comparable to a yearly quarter ,
there is great difficulty distinguish quarter - to - quarter variations from the intrinsic stellar variability @xcite .
[ shows an illustration of why identical fitting parameters were not used : while 20 points were usually enough for getting a smooth transition between two quarters , stars with rapid fluctuations must be treated differently .
figure [ . after correcting for the jumps .
the program also divided the data by the mean flux level . in the case of kic 4908338 ,
kic 6838420 and kic 8840004 we used 2000 points for the fits , 20 points were used for kic 2986893 ( except for the jump around bjd 55500 ) , kic 6279696 , kic 7274171 and kic 11759262 , while the dataset for kic 11768249 was corrected with linear fits to 200 points . note that each of these light curves contains approximately 48,000 points .
all the corrected wg12 data analysed here are available for download through the electronic version of this paper .
[ section : methods ] to characterise m giant variability with _ kepler _ , we have performed several simple analyses .
we compared the data to ground - based observations where available , then studied the amplitudes and periodicities with standard approaches .
finally , we looked into the time - dependent changes of the periods and amplitudes using the time - frequency distributions . to demonstrate the potential and properties of the data
, we carried out several comparisons with ground - based photometric observations such as those of the american association of variable star observers ( aavso ) or the all sky automated survey ( asas ) .
amplitude and periods were determined from the fourier transform of the time series with the program period04 @xcite . in order to study
the time dependent phenomena , e.g. amplitude and frequency modulation , mode switching , etc .
we calculated time - frequency distributions for all stars using the weighted wavelet - z transform code ( wwz , @xcite ) .
we compared our _ kepler _ measurements with ground - based photometry provided by the asas and aavso databases . in this section
we use aavso data ( visual , photoelectric @xmath6 and rgb - band digital single - lens reflex ( dslr ) data ) to study how well the ground- and space - based data can be cross calibrated . for three well - known long - period variables we compared the aavso and _ kepler _ light curves and their frequency spectra . in fig . [ .
the two sets of light curves are very similar even though the photometric bands were different .
we noticed that in the case of rwlyr a large amount of flux has been lost due to saturation .
this might be the cause of the imperfect agreement between the _ kepler _ and aavso visual data .
to remedy the situation we checked the target pixel files of this object and verified that the assigned optimal apertures were incapable of retaining all the flux , especially around maxima ( with the approx .
500-day period maxima occurred in q3 , q8 and q13 ) . around the maximum light as much as 50 - 60% of the flux was lost . in quarters where time intervals with no flux - loss were available
we applied the method developed in @xcite , namely we used the ratio of the saturated and neighbouring column integrated flux values to correct for the lost flux . in q3 , q8 and q13
however , this method failed , because flux leaked out of the downloaded pixel area during the whole quarter , therefore no reference values could be found . in these cases
we simply scaled the heavily saturated columns to follow the neighbouring non - saturated columns or sum of these .
this process indeed helped to match the space- and ground - based data in some quarters but not in others , hence we decided not to include the corrected light curves in fig .
[ ) are also very similar , with only minor differences in the amplitude distribution .
the plots are organised in such a way that the wavelet map , in which the amplitude is colour - coded and normalized to unity , is surrounded by the light curve on the top and the corresponding fourier spectrum on the left . this way we can see the temporal behaviour of the peaks in the spectrum and also in some cases the effects of gaps in the data . to characterise periodicities we performed a fourier analysis for all the corrected light curves . with iterative pre - whitening steps we determined the first 50 frequencies with period04 . in many cases ,
there were only a couple of significant peaks ( like for af cyg ) , while for the lower - amplitude stars even 50 frequencies may not include every significant peak .
the general conclusion , based on the various comparisons to ground - based data , is that there is a good correspondence between the two data sources in terms of the dominant periods and the shape of the light curves for the high - amplitude long - period variables .
_ kepler _ s superior precision allows for the determination of more periods for the lower - luminosity stars , but in cases when the frequency content is simple ( like for a mira star or a high - amplitude semiregular variable ) , 1,100 days of _ kepler _ data are still too short for revealing meaningful new information .
however , the uninterrupted _ kepler _ light curves should allow for the detection of microvariability with time - scales much shorter than those of the pulsations .
we note here that we found no star with flare - like events that would resemble those reported from the hipparcos data by @xcite .
the light curves of m giant variables can be very complex , seemingly stochastic with one or few dominant periodicities .
the complexity is inversely proportional to the overall amplitude .
the large - amplitude miras are known to be single periodic variables with coherent and stable light curves .
the lower amplitudes are typically associated with complicated light curve shapes that can be interpreted as a superposition of multiple pulsation modes .
the _ kepler _ wg12 sample ( sec .
[ section : wg12sample ] ) shows many features in the distribution of amplitudes and periods that were found previously in ground - based surveys . here
, we attempt to characterise the systematic distinction between different groups of stars . based on the complexity in the time and the frequency domains , we sorted the stars into three groups .
stars in group 1 have a wide range of periods between a few days and 100 days ( e.g. kic 4908338 or kic 11759262 in fig .
[ ) , occasionally supplemented by slow changes that may be related to rotational modulation or instrumental drifts .
a close inspection showed that all of them belong to the non - asas sample ( sec .
[ section : wg12sample ] ) .
stars with light curves containing only a few periodic components ( miras and srs ) compose group 3 ( e.g. kic 7274171 in fig . [ similarly to previous studies ( e.g. * ? ? ?
* ; * ? ? ?
* ) , we made a petersen diagram ( period ratios vs. periods ) for all wg12 stars to check any regularity in the frequency spacing and ratios . to make the diagram
, we took the five most significant periods and selected every possible pair to plot their ratio as a function of the longer of the two . in fig .
[ , a ratio that is known to belong to the upper red giant branch stars , one to two magnitudes below the absolute magnitude of the tip of the red giant branch ( trgb , see @xcite ) .
the ratio is in excellent agreement with the analysis of period ratios for a similar - sized sample of bright southern pulsating red giants by @xcite and those in ogle observations by @xcite .
this ratio has been interpreted as due to pulsations in the first and second radial overtone modes in theoretical models ( e.g. * ? ? ?
recently , @xcite performed detailed modelling of ogle small amplitude red giants ( osargs ) from the ogle - iii catalogue @xcite , for which they found that the rich structure in the petersen diagram can be explained by radial overtone modes , as well as by non - radial dipole and quadropole modes .
our plot in fig .
[ , it may seem surprising that stars in group 2 , i.e. objects characterised by their rapid variability , hardly appear in the lower left corner of the plot , where the short - period variables should fall . instead , the green points are scattered in the long - period end of fig . [ ( except the vertical concentration at @xmath7 and 2.87 resulting from to the _ kepler_-year variability ) confirms the noise - like behaviour for group 2 . for further investigations the nature of the three groups , we studied the amplitudes and their distributions .
first we made the period - amplitude plot for the whole sample , using all the 50 frequencies and amplitudes calculated by period04 .
this is shown in the upper panel of fig . [ and @xmath8 with a very well defined upper envelope pointed to the upper right corner of the plot . to the left of this upper envelope
there is a distinct feature that lies between @xmath9 ; @xmath10 and
@xmath11 ; @xmath12 which shows strong correlation between the period and amplitude .
this correlation resembles the @xmath13-amplitude scaling relation that has been extensively studied with _
data from the main sequence to red giants ( e.g. * ? ? ?
* ) . in this section we present typical power spectra for the three groups of the wg12 sample .
the reason for this is to illustrate the rich variety of the spectra , which indicates the complexity of m giant variability .
we show spectra for four stars in each group in the three columns of fig .
[ calculator of period04 to quantify the significance of each peak , following the method of @xcite . as can be seen in the middle column of fig .
[ and a distinct set of significant peaks .
note that the four panels in the middle column were sorted from top to bottom by the increasing average frequency of the acoustic signal ( roughly corresponding to the classical @xmath13 ) . here
every spectrum contains very high peaks ( in relative sense ) between 0 and 0.046 @xmath29 .
peaks near 0.031 @xmath29 correspond to the _ kepler_-year and they appear in almost every spectrum with smaller or higher amplitude . often there is another high peak near 0.008 @xmath29 , which is near to @xmath30 , where @xmath31 is the length of the full dataset .
the acoustic signal emerges at higher frequencies in form of some sort of structured peaks at lower amplitudes but still significant compared to the local noise in the spectrum .
spectra for group 3 are shown in the right column of fig . [ and 0.069 @xmath29 , i.e. periods between 150 and 400 days . in some cases ,
we see the integer harmonics of the dominant peak , which is caused by the strong departure from the pure sinusoidal shape , characteristic for most of the mira stars . as a last step we surveyed systematically the time - frequency distributions for all stars in the three groups .
we searched for phenomena that are more difficult to detect with the traditional methods of time - series analysis , such as mode switching , amplitude modulation or frequency modulation . while a detailed study of all these phenomena is beyond the scope of the paper , we can demonstrate the typical cases in each group . in fig .
13 we show three examples from each group .
the three groups were ordered in the subsequent rows from top to bottom .
group 1 stars ( top row in fig .
[ ) , the wavelet maps were calculated from 0.12 @xmath29 frequency ( without the long - period trends ) in order to get a clearer picture of the shorter and smaller amplitude variations between 1.16 - 5.74 @xmath29 , where the acoustic signal is dominant . here
the random changes of the frequencies are very apparent , a behaviour that naturally arises from the stochastic excitation of the solar - like oscillations . for group 3 stars ,
the available time - span only allows measuring the stability of the dominant peak , in good agreement with the mira character .
in this paper we studied the global characteristics of light variations for 317 red giant stars in the _ kepler _ database , containing 198 already known variable stars observed by the asas north survey and 117 stars in a control sample selected based on their estimated physical parameters .
the main results of this study : 1 .
the study of m giants with _ kepler _ poses new challenges because of the time - scale of variability that is comparable to the length of the _ kepler _ quarters .
most of the usual methods for correcting the trends and jumps are thus not applicable .
after extensive testing we ended up with a simple light curve stitching method , which is based on linear fits at the edges of the quarters and then matching the quarter - to - quarter shifts for creating the smoothest possible light curves .
we developed a software with a user - friendly gui , which made it easy to set the fitting parameters and stitch together each quarter .
when the data contain missing quarter(s ) , no unique solution is possible .
three years of observations revealed a so far unnoticed systematic fluctuation in the data , at the levels of up to 1 - 3% .
we found that the period equals to one _ kepler_-year and the phase behaviour is clearly correlated with the position in the whole _
field - of - view .
it is not yet clear if a more sophisticated pixel photometry would be able to remove the artefact .
we compared the data with various ground based photometries ( visual , asas ccd , aavso dslr , etc . ) and concluded that for the large - amplitude stars , _ kepler _ light curves can be matched very well with the ground - based data , but the amplitudes require a significant scaling by about a factor of two .
_ kepler _ s main advantage for these slow variables is the uninterrupted observations at high - precision and high - cadence ( relative to the pulsation periods ) .
we studied the distributions of periods , period ratios and amplitudes .
there are several regular patterns in these distributions that can be explained by the presence of several pulsation modes , some possibly non - radial dipole or quadropole modes .
we find evidence of a distinction between the solar - like oscillations and those larger amplitude pulsations characteristic for mira / sr stars in the period - amplitude plane .
this may show the transition between two types of oscillations as a function of luminosity .
the power spectra and wavelets reveal very complex structures and rich behaviour .
peaks in the spectra are often transient in terms of time - dependent amplitudes revealed by the wavelet maps .
the overall picture is that of random variations presumably related to the stochasticity of the large convective envelope . with this paper we highlighted the global characteristics of m giant stars seen with _
kepler_. there are several possible avenues to follow in subsequent studies . given the time - span and the cadence of the data , an interesting avenue of investigation is to perform a systematic search for rapid variability that can be a signature of mass - accreting companions .
one of the archetypal types of such systems , the symbiotic binary ch cyg , has been both kasc target and guest observer target , and its data can be used as a template to look for similar changes in the full _ kepler _ red giant sample .
another possibility is to quantify the randomness of the amplitude changes using detailed statistical analysis of the time - frequency distributions .
this project has been supported by the hungarian otka grants k76816 , k83790 , k104607 and human mb08c 81013 grant of mag zrt . , esa pecs c98090 , ktia urkut_10 - 1 - 2011 - 0019 grant , the lendlet-2009 young researchers program of the hungarian academy of sciences and the european community s seventh framework programme ( fp7/2007 - 2013 ) under grant agreement no .
269194 ( irses / ask ) .
ad , rsz and gymsz have been supported by the jnos bolyai research scholarship of the hungarian academy of sciences .
ad was supported by the hungarian etvs fellowship .
rsz acknowledges the university of sydney irca grant .
funding for this discovery mission is provided by nasa s science mission directorate . the kepler team and the kepler guest observer office
are recognized for helping to make the mission and these data possible . | m giants are among the longest - period pulsating stars which is why their studies were traditionally restricted to analyses of low - precision visual observations , and more recently , accurate ground - based data . here
we present an overview of m giant variability on a wide range of time - scales ( hours to years ) , based on analysis of thirteen quarters of _ kepler _ long - cadence observations ( one point per every 29.4 minutes ) , with a total time - span of over 1000 days .
about two - thirds of the sample stars have been selected from the asas - north survey of the _ kepler _ field , with the rest supplemented from a randomly chosen m giant control sample .
we first describe the correction of the light curves from different quarters , which was found to be essential .
we use fourier analysis to calculate multiple frequencies for all stars in the sample . over 50 stars show a relatively strong signal with a period equal to the kepler - year and a characteristic phase dependence across the whole field - of - view .
we interpret this as a so far unidentified systematic effect in the _ kepler _ data .
we discuss the presence of regular patterns in the distribution of multiple periodicities and amplitudes . in the period - amplitude plane
we find that it is possible to distinguish between solar - like oscillations and larger amplitude pulsations which are characteristic for mira / sr stars .
this may indicate the region of the transition between two types of oscillations as we move upward along the giant branch .
stars : variables : other stars : agb and post - agb techniques : photometric |
bloch oscillations , landau zener ( lz ) tunneling , and wannier stark ladders @xcite , are fundamental quantum effects occurring in a system of electrons moving in a periodic potential and subjected to a constant electric field . due to complications such as impurities , lattice vibrations , and multiparticle interactions ,
clean observations of these effects have been difficult @xcite . in recent years , ultra - cold atoms and bose
einstein condensates in optical lattices have been increasingly used to simulate solid state systems and the above mentioned phenomena @xcite .
optical lattices are nowadays easy to realize in the laboratory , and their parameters can be perfectly controlled both statically and dynamically , which makes them attractive model systems for crystal lattices .
more complicated potentials can be realized by adding further lattice beams @xcite .
in fact , by superimposing laser beams from different directions and with slightly different wave - lengths , it is possible to generate many different three - dimensional lattice geometries @xcite .
the question arises of how to control the dynamics of particles by quasiperiodic potentials ( possibly time - dependent or even stochastic ones ) .
+ in this paper , we present results on the wannier stark system realized with ultracold atoms , forming a bose
einstein condensate , in an optical lattice @xcite .
we compute the time dependence of the tunneling probability of the bose
einstein condensate atoms out of the ground band in which they were originally prepared . by changing the initial condition and the system parameters and introducing atom - atom interactions into the system , we are able to control the tunneling rate of the bose einstein condensate to higher bands .
finally a controlled noise added to the system will be shown to be a further handle to engineer the interband tunneling .
we study the temporal evolution of ultracold atoms loaded into a quasi one - dimensional optical lattice which can be a spatially periodic potential or a time - dependent stochastic potential , subjected to an additional static force , in the presence of weak atom - atom interactions @xcite .
we use the following general form of three - dimensional gross pitaevskii equation to model the temporal evolution of the atoms @xmath0 with lattice potentials along the longitudinal direction , @xmath1 , as @xmath2 @xmath3 is the mass of condensate atoms and @xmath4 is the stark force . @xmath5 and @xmath6 are spatially periodic potentials with incommensurate lattice spacings @xmath7 and @xmath8 , respectively . as will be shown below around eq .
, the noise has a tendency to average over the second lattice , and therefore the amplitudes of the two lattices should be comparable . for convenience we chose equal amplitudes @xmath9 and @xmath10 and @xmath11 for the lattice constants .
@xmath6 is a time - dependent stochastic potential with a time - dependent stochastic phase @xmath12 , which we will characterize further down in section [ subsec:2 ] .
the renormalization factor @xmath13 is introduced to be able to compare the dynamics in the presence of the potential given by eqs . and to the dynamics of the reference system " , i.e. the dynamics in the potential @xmath14 .
@xmath13 will be chosen in such a way that the following standard deviations are equal : @xmath15 where the effective potential @xmath16 will be defined in section [ subsec:2 ] .
the average @xmath17 is an integral over space for a sufficiently large @xmath18 , i.e. , @xmath19 .
the third term in the hamiltonian is the nonlinearity , which makes the equation different from the schrdinger equation .
@xmath20 represents the condensate wave function and @xmath21 the local atomic density .
@xmath22 is the coupling constant which is proportional to the scattering length @xmath23 and determines the strength of atom - atom interactions , where @xmath24 , with @xmath25 and @xmath26 for rubidium 87 .
@xmath27 is the number of atoms in the condensate .
the recoil energy @xmath28 and the recoil momentum @xmath29 are the characteristic energy and momentum scales for our system .
moreover , we define @xmath30 and @xmath31 as dimensionless quantities in this energy unit . experimentally , the initial state is the relaxed condensate wave function prepared in the confining potential given by a harmonic trap @xmath32 , with @xmath33 for a quasi 1d situation , and then loaded adiabatically into the optical lattice when the stark force @xmath4 equals zero . then @xmath34 is either switched off or relaxed to a small value @xmath35 and the stark force @xmath4 is simultaneously switched on to induce the dynamics . in the following sections we study the system in the presence / absence of different terms of the above hamiltonian . + , and a narrow width of the initial momentum distribution of the bec cloud , @xmath36 ( trap frequency @xmath37 ) , and various stark forces . for non - ret condition : @xmath38 ( solid line ) and @xmath39 ( dash - dotted line ) ; for ret condition : @xmath40 ( dashed line ) .
inset : height of one step ( of data as shown in the main panel ) for fixed @xmath41 and various @xmath42 .
the step height as predicted by eq .
( solid line ) in comparison with the numerically obtained values ( filled circles ) .
a significant deviation from the lz prediction can be observed at values of @xmath43 corresponding to ret conditions and marked by the vertical dashed lines . ] for zero or small @xmath35 ( c.f .
, e.g. , @xcite ) and negligible atom - atom interactions , when @xmath44 and @xmath45 , the above hamiltonian in eq
. describes the dynamics of atoms in a tilted periodic potential which is the well - known single - particle wannier stark problem . without the nonlinearity term
, we can use the 1d version of eq . for our simulations .
in the presence of @xmath4 , the quasimomentum of a condensate ( initially prepared at the center of the brillouin zone in the ground band ) scans the lower band in an oscillating motion periodically
so - called bloch oscillations with the bloch period @xmath46 . at the edge of the brillouin zone ,
where the gap between the ground and the first excited band @xmath47 of the @xmath48 system ( increasing with @xmath49 @xcite ) , acquires its minimum value , a tunneling of the condensate to the first excited energy band may occur .
the tunnelled atoms escape from the system through successive tunneling events across the much smaller band gaps between the upper bands .
this phenomenon is known as the lz tunneling .
lz theory predicts a decay rate @xmath50 where @xmath42 is the adiabaticity parameter and @xmath51 @xcite . in order to study the lz prediction for our system
we need to access the decay rate of the population from the ground band . in that respect , we compute the time - dependence of the probability of the condensate to remain in the ground band in which it has been initially prepared . such a survival probability is best measured in momentum space . from the time - dependent momentum distribution we can determine @xmath52 by projection of the evolved state @xmath53 on to the support of the initial state @xcite @xmath54 where @xmath55 is an ad hoc cut - off momentum . in our calculation we chose @xmath56 .
. starts to measure the wave packet leaving the ground - band with an artificial delay of one bloch period , since by the acceleration theorem the average momentum is proportional to time ( more details can be found in @xcite ) .
a very nice step structure local deviation from the overall exponential decay can be seen in the survival probability ( see figures [ fig:1 ] and [ fig:2 ] ) .
such step structures reflect the above mentioned phenomena , i.e. , bloch oscillations and lz tunneling .
one can see that the tunneling events from the ground band to the next band occur after each bloch period when the wave packet is at the edge of the brillouin zone .
it is possible to control the lz tunneling of the bose
einstein condensate from the ground band in the linear system and in the absence of atom - atom interactions by * changing the system parameters , such as external force @xmath57 , and the amplitude of the optical lattice @xmath49 ( see figure [ fig:1 ] ) , * exploiting resonantly enhanced tunneling ( ret ) between degenerate wannier
stark states at @xmath58 with @xmath59 being an integer number ( see figure [ fig:1 ] ) @xcite , * changing the initial condition by changing the trap frequencies and hence preparing the bose
einstein condensate with different widths @xmath60 of its initial momentum distribution ( see figure [ fig:2 ] ) . , a constant stark force @xmath40 , and various widths of the initial momentum distribution of the bec cloud , @xmath36 ( solid line ) ; @xmath61 ( dashed line ) ; @xmath62 ( dash - dotted line ) .
inset : step width @xmath63 for several @xmath60 ( filled circles ) .
the step width @xmath63 is the distance between the two vertical dashed lines shown in the main panel for the case of @xmath36 . ]
as seen in figure [ fig:1 ] , increasing the tilting force leads to more and more tunneling of the atoms from the ground band . on the other hand , depending on the system parameters , one can tune into a special condition for which the rate of the tunneling is enhanced .
this occurs when an integer multiple of the energy scale of the tilting force @xmath64 matches the energy difference between the initial state and the final state , i.e. , approximately the band gap @xmath47 .
this phenomenon has been observed experimentally @xcite and is called resonantly enhanced tunneling ( ret ) .
a deviation from the lz prediction is expected in this case . in order to see whether the tunneling probability given by the standard lz tunneling probability correctly predicts the height of a step corresponding to a single tunneling event , we fit a step function to our step - like survival probability and extract the height of each step .
the result and the comparison to the lz prediction is shown in the inset of figure [ fig:1 ] .
when the system parameters are in the ret condition ( e.g. , @xmath65 and @xmath66 ) , the height of the steps of the survival probability increases and shows a significant deviation from the lz prediction given in eq . .
this behavior is seen in figure [ fig:1 ] where the survival probability for @xmath40 decays much faster than for the other two cases . since the derivation of eq .
@xcite does not take into account the actual wannier
stark level structure , which is necessary to describe the ret condition , eq . or the survival probability derived from it can not describe the enhancement of the tunneling probability due to ret . + the other parameter which can affect the survival probability is the width @xmath60 of the initial momentum distribution . by changing the trap frequencies
@xmath67 and @xmath68 one can prepare the initial distribution with different widths .
figure [ fig:2 ] demonstrates the dependence of the width of the steps @xmath63 on @xmath60 .
the steps are smooth and partly washed out since the wave packet reaches the edge of the brillouin zone earlier when it has a broader initial momentum distribution .
still the remnants of the steps cause a local deviation from the exponential decay of the tunneling probability .
nevertheless , the survival probability exhibits an exponential decay globally in time , i.e. , on large time scales , for all the mentioned cases .
+ in the regime of weak atom - atom interactions the effect of interactions is studied in the mean - field regime based on the gross pitaevskii equation . the 3d gross
pitaevskii equation can describe the dynamics of the entire bose
einstein condensate in terms of an equation of motion for a single particle wave function .
therefore , the following nonlinear equation describes the dynamics of interacting bose
einstein condensate atoms in a tilted periodic potential ( @xmath44 ) with @xmath45 : @xmath69 as mentioned in section [ sec:1 ] , @xmath70 is the coupling constant calculated from the s - wave scattering wavelength @xmath23 and the number of atoms in the condensate @xmath27 .
@xmath71 is the local atomic density . as an estimate for the nonlinear term in eq .
, we define @xmath72 at the peak density of the initial state .
the following effects can be seen by increasing the strength of a repulsive nonlinearity ( @xmath73 ) in the system : * enhancement of the tunneling rate , * deviation from mono - exponential decay , * washed out steps ( corresponding to damped bloch oscillations ) . according to our results shown in figure [ fig:3 ] for a ret case and the experimental results in @xcite ,
the temporal behavior of the survival probability depends on the strength of atom - atom interactions .
as can be found with more detail in @xcite , the enhancement of decay rate is generic for repulsive interactions .
the scaling of the decay rate as a function of nonlinearity is yet more interesting in the ret case ( see the inset of figure [ fig:3 ] ) .
we can quantify the decay rate @xmath74 of the survival probability by globally fitting an exponential decay function to the step - like curves of the survival probability .
such rates , for various nonlinearities , are depicted in the inset of figure [ fig:3 ] .
a repulsive interaction initially enhances the interband tunneling probability of the ultracold atoms @xcite .
since the tunneling events occurring at different integer multiples of the bloch period are correlated by the presence of the nonlinearity , a clear deviation from the mono - exponential decay is observed making the definition of a global decay rate @xmath74 somewhat problematic @xcite .
a continuous change in the density of the condensate in time due to escaped particles from the system leads to a decreasing impact of the nonlinearity .
therefore , the time local rate of decay systematically decreases as the time increases .
additionally , the nonlinearity leads to a dephasing and damping of the bloch oscillations , not discussed here , but a discussion of this phenomenon can be found in @xcite .
+ , @xmath40 ( ret ) , a narrow width of the initial momentum distribution of the bec cloud , @xmath36 , and composed of the following number of atoms prepared in a trap with @xmath37 and @xmath75 : @xmath76 , @xmath77 ( open circles ) ; @xmath78 , @xmath79 ( filled squares ) ; @xmath80 , @xmath81 ( open triangles ) as compared to the linear case @xmath82 , ( solid line ) .
inset : decay rate of the survival probability @xmath74 at ret condition vs. @xmath27 ( filled circles ) . ]
going back to the linear system , @xmath82 , the case of @xmath83 leads to a time - dependent stochastic potential where we claim that we are able to control the dynamics of the bose
einstein condensate atoms by changing the characteristic parameters of the time - dependent stochastic phase @xmath12 .
@xmath13 is no longer 1 and we can calculate it as defined above by eq . , using the effective potential introduced below in eq . .
we use correlated noise for the time - dependent phase @xmath12 for the second lattice .
a standard example is exponentially correlated noise , which is characterized by a single correlation time .
such a noise can be obtained from linearly filtered white noise as follows @xmath84 where @xmath85 is a gaussian white noise with zero mean and standard deviation equal to one .
@xmath86 is the correlation time of the noise and the strength of the noise is given by the parameter @xmath87 .
the introduced noise has a zero mean ( @xmath88 ) and an exponential correlation function @xmath89 .
@xmath86 describes the time - scale of the fluctuations and @xmath90 its variance .
the power spectrum of the exponentially correlated noise is given by a lorentzian function as follows @xmath91 figure [ fig:4 ] demonstrates the survival probability of the atoms in the ground band for various correlation times @xmath86 and a fixed value of the variance @xmath92 of the noise .
the system parameters are @xmath93 and @xmath94 ( non - ret condition ) .
the solid line shows the time evolution of the survival probability for the reference system introduced by eq . with @xmath82 , @xmath44 and @xmath45 . a nice step structure similar to the ones depicted in figures [ fig:1 ] and [ fig:2 ]
is observed for this case .
the other curves in figure [ fig:4 ] exhibit the time evolution of the survival probability for the temporally disordered system introduced in eq .
with @xmath82 , @xmath95 and @xmath96 . as seen for all the cases the step structure is washed out .
considering the characteristic parameters of the exponentially correlated noise , different regimes of the noise are introduced .
the regime of slowly varying noise is recovered when the noise has a large correlation time compared with the time scales of the system ( e.g. , @xmath97 ) .
the tunneling probability of the atoms in such a regime is suppressed ( e.g. , for @xmath98 shown by the dashed line in figure [ fig:4 ] ) .
decreasing the correlation time the rate of tunneling increases but , nevertheless , for @xmath99 the noise suppresses the tunneling compared with the reference system ( see the data for @xmath100 as shown by the dot - dashed line in figure [ fig:4 ] ) . for correlation
times smaller than @xmath101 , the noise recovers the regime of fast noise ( e.g. , for @xmath102 ) and causes an enhancement in the tunneling rate ( dot - dot - dashed line ) compared with the reference system ( thick solid line ) .
surprisingly a further decrease of the correlation time decreases the tunneling rate of the atoms ( e.g. , for @xmath103 depicted by the thin solid line ) . + in order to understand the effect of the time - dependent stochastic potential on the system , we renormalize the potential .
the time - dependent potential can be replaced by a suitable static effective potential in the limit of small @xmath86 .
such an effective potential can be calculated integrating over all possible phases giving : @xmath104 , \label{eqno9}\ ] ] with a renormalization factor for the second lattice @xmath105 .
. provides a time - independent potential which can be used to compute @xmath13 as stated in eq . .
as seen in figure [ fig:4 ] the survival probability for a small @xmath103 ( thin solid line ) shows perfect agreement with the results achieved using the effective potential of eq .
( filled circles ) .
+ , @xmath94 , and @xmath106 ; the thick solid line is the decay of the reference system defined by eq . with @xmath82 , @xmath44 and @xmath45 ;
in the presence of exponentially correlated noise with @xmath92 and correlation times : @xmath98 ( dashed line ) , @xmath100 ( dot - dashed line ) , @xmath102 ( dot - dot - dashed line ) , @xmath103 ( thin solid line ) ; results for the effective potential of eq .
are shown by the filled circles . ] keeping the noise parameters constant , we study the decay rate of the survival probability , @xmath74 , of the condensate for various system parameters
. the results of a scan over the stark force are shown in figure [ fig:5](@xmath107 ) .
the simulations have been done for three values of the correlation time of the noise @xmath108 , keeping the variance of the noise constant ( here @xmath92 ) .
the three correlation times are chosen from the left shoulder , the peak point and the right shoulder of curves in figure [ fig:5](@xmath109 ) .
the solid line depicts the decay rate of the reference system ( given by eq . with @xmath82 , @xmath44 and @xmath45 ) .
noise in the system leads to a washing out of the ret peaks ( present in the solid line of figure [ fig:5](@xmath107 ) ) in the decay rate , and , dependent on the correlation time of the noise , a suppression ( for very large and very small values of @xmath86 ) or an enhancement of the tunneling ( e.g. , for @xmath110 ) can be obtained .
+ we also ran a scan over the correlation time of the noise @xmath86 , again keeping the variance of the noise constant ( @xmath92 ) .
the decay rate of the survival probability for three amounts of the stark force ( @xmath111 ) is shown in figure [ fig:5](@xmath109 ) .
comparing the symbols to the lines ( which specify the decay rate in the corresponding reference system ) one can realize that the decay rate is enhanced when the system parameters do _ not _ fulfill the ret condition , i.e. , in this case for @xmath112 and @xmath113 .
for the case of @xmath114 ( ret ) the tunneling rate is suppressed , and the symbols lie always below the reference line ( dot - dot - dashed line ) .
the enhancement is pronounced in the range of @xmath115 , corresponding to an energy scale equal or larger to / than the band gap @xmath47 . on the other hand , for small values of the correlation time
, the noise recovers the regime of white noise , where the effective potential describes the dynamics of the system very well , and for large values of @xmath86 , the system is in the regime of a slowly varying noise corresponding to energy scales that do not help to enhance the tunneling of atoms either .
+ , and @xmath92 : ( a ) scan over @xmath116 for @xmath117 ( open squares ) , @xmath102 ( open circles ) , @xmath118 ( open triangles ) , and for the reference system ( solid line ) ; ( b ) scan over @xmath86 for @xmath94 ( filled circles ) and its reference system ( dot - dot - dashed line ) , @xmath119 ( filled squares ) and its reference system ( dot - dashed line ) , @xmath120 ( filled triangles ) and its reference system ( dashed line ) . ]
it is possible to control the landau
zener ( lz ) tunneling probability of the ultracold atoms from the ground band in tilted optical lattices .
this control is possible by changing the system parameters such as the lattice depth and the stark force , or by changing the initial condition which is given by the initial width of the momentum distribution of the bec .
all the mentioned parameters can be easily tuned experimentally .
furthermore , our calculations showed that atom - atom interactions affect the lz tunneling probability and a repulsive interaction typically leads to an enhancement of the lz tunneling of the ultracold atoms from the ground band .
most interestingly , our results demonstrate that it is also possible to control the tunneling by adding noise to the system . by changing the noise parameters ,
the tunneling probability can be enhanced or suppressed .
the noise can particularly enhance the tunneling probability when the system parameters are chosen far from the ret condition .
+ we acknowledge funding by the excellence initiative by the german research foundation ( dfg ) through the heidelberg graduate school of fundamental physics ( grant number gsc 129/1 ) and the global networks mobility measures .
s. w. is grateful to the heidelberg academy of sciences and humanities for the academy award 2010 and to the hengstberger foundation for support by the klaus - georg and sigrid hengstberger prize 2009 .
g. t. thanks the landesgraduiertenfrderung baden - wrttemberg for support . | several ways are discussed how to control the landau zener tunneling in the wannier stark system .
we focus on a realization of this system with interacting and noninteracting ultracold bosons .
the tunneling from the ground band to the continuum is shown to depend crucially on the initial condition and system parameters and , more interestingly , on added time - dependent disorder
noise on the lattice beams . |
remarkbale progress has been made in understanding the physics of heavy mesons composed of a heavy quark and a light quark with the discovery of the heavy quark symmetry . to obtain detailed predictions
one has to employ some specific nonperturbative methods , among which qcd sum rules @xcite is very useful .
the heavy quark effective theory ( hqet ) @xcite provides a systematic expansion of qcd in terms of @xmath4 , where @xmath5 is the heavy quark mass .
the spectrum of the ground state heavy meson has been studied with the qcd sum rules in hqet in @xcite . in @xcite
the mass of the lowest excited heavy meson doublets @xmath6 and @xmath7 were studied with qcd sum rules in the heavy quark effective theory ( hqet ) up to the order of @xmath8 .
qcd sum rules has been used to analyse the couplings of the heavy hadrons with pions [ 5 - 15 ] .
the widths for pionic decays of the lowest two excited doublets @xmath6 and @xmath7 is calculated with conventional qcd sum rules in the leading order of hqet in @xcite .
light - cone qcd sum rules ( lcqsr ) with the full qcd lagrangian is first employed to derive the strong coupling constants : @xmath9 , @xmath10 @xcite .
recently the strong coupling constants : @xmath11 , @xmath12 and @xmath13 were calculated using lcqsr with finite heavy quark mass @xcite .
the couplings of heavy baryons with soft pions have been estimated with qcd sum rules in an external axial field @xcite . in this work
we employ the lcqsr in hqet to calculate the on - shell couplings of the pions with heavy meson doublets @xmath0 , @xmath1 and @xmath2 to the leading order of @xmath4 .
the lcqsr is quite different from the conventional qcd sum rules , which is based on the short - distance operator product expansion ( ope ) .
the lcqsr is based on the ope on the light cone , which is the expansion over the twists of the operators .
the main contribution comes from the lowest twist operator .
matrix elements of nonlocal operators sandwiched between a hadronic state and the vacuum defines the hadron wave functions . when the lcqsr is used to calculate the coupling constant , the double borel transformation
is always invoked so that the excited states and the continuum contribution can be treated quite nicely . moreover
, the final sum rule depends only on the value of the wave function at a specific point like @xmath14 , which is much better known than the whole wave function @xcite .
one difficult problem encountered in studying the decay widths of excited heavy mesons with qcd sum rules is the following .
except for the lowest states @xmath15 , @xmath16 , the spectra contains a pair of states for any spin - parity @xmath17 with close values in their masses but quite different in magnitudes of their decay widths . in our case ,
one of the two @xmath3 states is a narrow resonance decaying mainly by emitting a @xmath18 wave pion , while the other one is a very wide resonance decaying by emitting a @xmath19 wave pion .
an interpolating current used for the narrow @xmath3 state with a small coupling to the other @xmath20 state may cause sizable error in the result of calculation .
it is only in the @xmath21 limit , there is a conserved quantum number @xmath22 , the angular momentum of the light component , which can be used to differentiate the two states .
therefore , hqet has important and unique advantage for this purpose .
these are the motivation for our approach of using lcqsr in hqet .
the proper interpolating current @xmath23 for the states with the quantum number @xmath24 , @xmath25 , @xmath22 in hqet was given in @xcite .
they were proved to satisfy the following conditions @xmath26&&\times\ : \int \,dt\delta(x - vt)\:\pi_{p , j_{\ell}}(x)\end{aligned}\ ] ] in the @xmath21 limit , where @xmath27 is the polarization tensor for the spin @xmath24 state , @xmath28 is the velocity of the heavy quark , @xmath29 is the transverse metric tensor , @xmath30 denotes symmetrizing the indices and subtracting the trace terms separately in the sets @xmath31 and @xmath32 , @xmath33 and @xmath34 are a constant and a function of @xmath35 respectively which depend only on @xmath25 and @xmath36 . because of equations ( [ decay ] ) and ( [ corr ] ) , the sum rule in hqet for decay widths derived from a correlator containing such currents receive no contribution from the unwanted states with the same spin - parity as the states under consideration in the @xmath21 . starting from the calculations in the leading order , the decay amplitudes for finite @xmath5
can be calculated unambiguously order by order in the @xmath37 expansion in hqet .
in the present work we shall confine ourselves to the lowest lying states in the leading order of @xmath37 expansion . denote the doublet @xmath38 with @xmath39 by @xmath40 and the doublet @xmath1 with @xmath41 by @xmath42 .
there are many combinations for the coupling constant @xmath43 , where @xmath44 , @xmath45 belongs to the three doublets @xmath0 , @xmath1 and @xmath2 . due to
the heavy quark symmetry and chiral symmetry there exist only six independent coupling constants , i.e. , three independent coupling constants when @xmath44 , @xmath45 belongs to different doublets and another three when @xmath44 , @xmath45 belongs to the same doublet .
for example , from covariance and conservation of the angular momentum of the light component in the @xmath46 limit , the amplitudes for the decay of @xmath47 , @xmath48 to the ground states @xmath49 , @xmath50 in the doublet @xmath51 with @xmath41 have the following forms @xmath52 where @xmath53 , @xmath54 and @xmath55 are polarization tensors for states @xmath56 , @xmath3 and @xmath57 respectively .
@xmath59 , @xmath60 for charged and neutral pion respectively .
as shown in @xcite the constants in ( [ coup1])-([coup3 ] ) satisfy @xmath61 we calculate the following six independent coupling constants : @xmath62 , @xmath63 , @xmath64 , @xmath65 , @xmath66 and @xmath67 .
other coupling constants are related to them .
there are related results for @xmath68 , @xmath69 and @xmath70 in [ 6 - 15 ] . in @xcite @xmath71 and @xmath68
is calculated with qcd sum rules in the short distance expansion . for the sake of comparison and completeness
, we also present the sum rules for @xmath68 , @xmath69 and @xmath70 . for deriving the sum rules for the coupling constants we consider the correlators @xmath72\label{7b }
\int d^4x\;e^{-ik\cdot x}\langle\pi(q)|t\left(j_{0,+,\frac{1}{2}}(0 ) j^{\dagger}_{0,-,\frac{1}{2}}(x)\right)|0\rangle= i\ ; g_{b_0^{\prime } b } ( \omega,\omega')\;,\hspace{1.4cm}\\[2mm]\label{7c } \int d^4x\;e^{-ik\cdot x}\langle\pi(q)|t\left(j_{0,+,\frac{1}{2}}(0 ) j^{\dagger\alpha}_{1,+,\frac{1}{2}}(x)\right)|0\rangle= q_t^{\alpha } i\;g_{b_1^{\prime}b_0^{\prime } } ( \omega,\omega')\ ; , \\[2mm]\label{7d}\nonumber \int d^4x\;e^{-ik\cdot x}\langle\pi(q)|t\left(j^{\alpha}_{1,+,\frac{3}{2}}(0 ) j^{\dagger\sigma\beta}_{2,+,\frac{3}{2}}(x)\right)|0\rangle= \\ \left(q_t^{\alpha}q_t^{\sigma}q_t^{\beta}- \frac{1}{6}q^2_t(g_t^{\alpha\beta}q_t^{\sigma } + g_t^{\alpha\sigma}q_t^{\beta } + { 4\over 3}g_t^{\sigma\beta}q_t^{\alpha } ) \right ) i\ ; g_{b_2^*b_1 } ( \omega,\omega')\;,\\[2mm]\label{7e } \int d^4x\;e^{-ik\cdot x}\langle\pi(q)|t\left(j^{\alpha}_{1,+,\frac{3}{2}}(0 ) j^{\dagger\beta}_{1,+,\frac{1}{2}}(x)\right)|0\rangle= i\epsilon^{\alpha\beta\sigma\delta } q^t_{\sigma}v_{\delta } i\ ; g_{b_1^{\prime}b_1 } ( \omega,\omega')\;,\\[2mm]\label{7f } \int d^4x\;e^{-ik\cdot x}\langle\pi(q)|t\left(j^{\alpha}_{1,-,\frac{1}{2}}(0 ) j^{\dagger}_{0,-,\frac{1}{2}}(x)\right)|0\rangle = q_t^{\alpha } i\;g_{b^*b}(\omega,\omega')\;,\end{aligned}\ ] ] where @xmath73 , @xmath74 , @xmath75 and @xmath76 .
the forms of the right hand side of ( [ 7a])-([7f ] ) are determined by the conservation of angular momentum of the light component and the fact that @xmath77 , @xmath78 , @xmath79 and @xmath80 are transverse indices , @xmath81 on the heavy quark propagator .
the interpolationg currents are given in @xcite as @xmath82 @xmath83 where @xmath84 is the heavy quark field in hqet and @xmath85 .
let us first consider the function @xmath86 in ( [ 7a ] ) . as a function of two variables , it has the following pole terms from double dispersion relation @xmath87 where @xmath88 are constants defined in ( [ decay ] ) , @xmath89 .
as explained in section [ sec1 ] , only one state with @xmath90 contributes to ( [ pole ] ) as the result of equation ( [ decay ] ) .
this would not be true if the last term in ( [ curr1 ] ) is absent . for deriving qcd sum rules
we calculate the correlator ( [ 7a ] ) by the operator expansion on the light - cone in hqet to the leading order of @xmath91 .
the expression for @xmath92 reads @xmath93 the pion wave function is defined as the matrix elements of nonlocal operators between the vacuum and pion state . up
to twist four they are @xcite : @xmath94 the wave function @xmath95 is associated with the leading twist two operator , @xmath71 and @xmath68 correspond to twist four operators , and @xmath96 and @xmath97 to twist three ones . due to the choice of the gauge @xmath98 , the path - ordered gauge factor
@xmath99 has been omitted . expressing ( [ lam5 ] ) with the pion light - cone wave functions ,
we arrive at @xmath100 where @xmath101gev , @xmath102mev for neutral pions . for large euclidean values of @xmath103 and @xmath104 this integral
is dominated by the region of small @xmath105 , therefore it can be approximated by the first a few terms .
similarly , we have : @xmath106 \}\ ; .\ ] ] @xmath107 @xmath108 @xmath109 \}\ ; , \ ] ] @xmath110 after performing wick rotation and double borel transformation with the variables @xmath103 and @xmath104 the single - pole terms in ( [ pole ] ) are eliminated .
subtracting the continuum contribution which is modeled by the integral in region @xmath111 , we arrive at : @xmath112 where @xmath113 , @xmath114 , @xmath115 , @xmath116 are the borel parameters , @xmath117 and @xmath118 .
the presence of the factor @xmath119 is the result of subtracting the integral @xmath120 as a contribution of the continum . in obtaining ( [ final - a ] )
we have used the borel transformation formula : @xmath121 .
the integral in the function @xmath122 in ( [ final - a ] ) arises from the factor @xmath123 in ( [ q1 ] ) .
here we have used integration by parts to absorb the factor @xmath123 . in this way
we arrive at the simple form after double borel transformation . in the following
we shall use the same technique to deal with the factor @xmath124 in other sum rules .
similarly we have : @xmath125 where @xmath126 , @xmath127 are the first derivatives of @xmath128 , @xmath129 at @xmath130 .
the appearance of derivatives is due to the factor @xmath131 .
@xmath132 where @xmath133 .
@xmath134 where @xmath135 .
@xmath136 @xmath137
in order to obtain the coupling constants from ( [ final - a])-([final - f ] ) we need to use the mass parameters @xmath138 s and the coupling constants @xmath139 s of the corresponding interpolating currents as input . @xmath140 and
@xmath141 can be obtained from the results in @xcite as @xmath142 gev and @xmath143 gev@xmath144 at the order @xmath145 .
notice that the coupling constant @xmath146 defined in the present work is a factor @xmath147 smaller than that defined in @xcite .
@xmath148 and @xmath149 are given in @xcite .
@xmath150 and @xmath151 can be determined from the formulas ( 34 ) , ( 27 ) and ( 28 ) of reference @xcite derived from sum rules for two point correlators .
the results are @xmath152 we use the wave functions adopted in @xcite to compute the coupling constants .
moreover , we choose to work at the symmetric point @xmath153 , i.e. , @xmath154 as traditionally done in literature @xcite .
we adopt the scale @xmath155gev , at which the values of the various functions appearing in ( [ final - a])-([final - f ] ) , at @xmath156 , are : @xmath157 , @xmath158 , @xmath159 , @xmath160gev@xmath161 , @xmath162gev@xmath161 , @xmath163gev@xmath161 , @xmath164gev@xmath161 , @xmath165 , @xmath166 and @xmath167
we now turn to the numerical evaluation of the sum rules for the coupling constants .
the lower limit of @xmath168 is determined by the requirement that the terms of higher twists in the operator expansion is less than one third of the whole sum rule .
this leads to @xmath169 gev for the sum rules ( [ final - a])-([final - f ] ) .
in fact the twist - four terms contribute only a few percent to the sum rules for such @xmath168 values .
the upper limit of @xmath168 is constrained by the requirement that the continuum contribution is less than @xmath170 .
this corresponds to @xmath171gev . with the values of pion wave functions at @xmath154 given above
we obtain the left hand side of the sum rules ( [ final - a])-([final - f ] ) as functions of @xmath168 .
the results are shown in fig .
stability develops for the sum rule ( [ final - a])-([final - f ] ) in the region @xmath172 gev @xmath173@xmath168@xmath173@xmath174 gev .
numerically we have : @xmath175 where the first error refers to the variations with @xmath168 in this region and the second error takes into account the uncertaity in @xmath176 .
and the central value corresponds to @xmath177gev and @xmath178gev for ( [ final - a])-([final - e ] ) . for the sum rule ( [ final - f ] )
we use @xmath179gev . with the central values of f s in ( [ fvalue ] ) we get the absolute value of the coupling constants :
@xmath180 we have not included the uncertainties due to f s here .
note that the sum rules ( [ final - a ] ) and ( [ final - c])-([final - f ] ) are also sensitive to the value of @xmath181 . for example , @xmath182 with @xmath183 which is one of the values used in @xcite .
the couplings of pions with heavy mesons obey symmetry relations in the leading order of hqet . only in hqet the ambiguity due to the presence of two distinct @xmath3 states can be resolved .
we use light - cone qcd sum rules to calculate all the six independent strong coupling constants @xmath71-@xmath70 for the lowest three doublets .
the sum rules in hqet for @xmath69 and @xmath70 are consistent with those for @xmath184 and @xmath185 in @xcite in the limit of @xmath186 with @xmath187 and @xmath188 .
the sum rule ( [ final - b ] ) for @xmath68 can be compared to that for @xmath189 in @xcite with @xmath190 from definition . in terms of @xmath189
our value is @xmath191 , which is only one half of that given in @xcite .
our result @xmath192 is slightly smaller than that from the short - distance qcd sum rule @xcite .
the calculation for @xmath193 and @xmath194 is new . with the cleo measurement of @xmath195 @xcite ,
a model independent extraction of @xmath184 has been performed recently @xcite .
two possible solutions of @xmath184 ( @xmath196 in our notation ) and @xmath78 are found from fitting to the experimental data , either @xmath197 , @xmath198gev@xmath199 or @xmath200 , @xmath201gev@xmath199 .
the @xmath202 solution is excluded by the experimental limit @xmath203mev @xcite .
the coupling @xmath184 from the lcqsr approach is @xmath204 @xcite , which is consistent with the experimentally favored solution @xmath205 . in a recent work @xcite the mixing effect of the two @xmath3 states in the lowest excited heavy meson doublets @xmath206 and @xmath207 and the @xmath8 correction to the leading order coupling @xmath71
was calculated .
the resulting pionic decay widths of @xmath208 and @xmath209 are in agreement with the experimental data @xcite .
b. grinstein , nucl .
b339 * ( 1990 ) 253 ; e. eichten and b. hill , phys . lett .
* b234 * ( 1990 ) 511 ; a. f. falk , h. georgi , b. grinstein and m. b. wise , nucl . phys .
* b343 * ( 1990 ) 1 ; f. hussain , j. g. krner , k. schilcher , g. thompson and y. l. wu , phys .
lett . * b249 * ( 1990 ) 295 ; j. g. krner and g. thompson , phys . lett .
* b264 * ( 1991 ) 185 .
p. colangelo , g. nardulli , a. a. ovchinnikov and n. paver , phys .
* b269 * ( 1991 ) 204 ; p. colangelo , g. nardulli and n. paver , phys . lett .
* b293 * ( 1992 ) 207 ; p. colangelo , f. de fazio , g. nardulli , n. di bartolomeo and r. gatto , phys . rev .
* d 52 * ( 1995 ) 6422 . | the couplings of pions with lowest three doublets @xmath0 , @xmath1 and @xmath2 of heavy mesons are studied with light - cone qcd sum rules in the leading order of heavy quark effective theory .
the ambiguity due to presence of two distinct @xmath3 states are solved .
= -5 mm @sh[#1]#2 fmsl@sh fmsl@sh fmsl@sh fmsl@sh#1#2#3@th * couplings of pions with excited heavy mesons from + light - cone qcd sum rules + in the leading order of hqet * yuan - ben dai and shi - lin zhu + pacs number : 12.39.hg , 14.40.nd , 12.38.lg |
ngc3077 is a prominent example of an interacting dwarf galaxy .
together with the prototypical starburst galaxy m82 and the spiral galaxy m81 it forms the famous m81 triplet .
although an interaction between the three galaxies is not at all obvious at optical wavelengths ( but see arp 1965 , getov & georgiev 1988 ) it has been long known through observations in the line that the 3 galaxies are interacting : tidal arms and tails connect the three galaxies over a projected distance of more than 70 kpc ( cottrell 1977 , van der hulst 1979 , yun _ et al . _ 1994 )
. the outstanding star forming activity in m82 is usually attributed to the gravitational interaction with m81 and/or ngc3077 .
although the molecular and starburst properties of m82 have been the subject of innumerable studies , little is known about its starforming neighbor , ngc3077 .
this is surprising since it has been long known that ngc3077 is experiencing intense star formation ( e.g. , barbieri _ et al .
as noted by price & gullixson ( 1989 ) it is difficult to classify ngc3077 on basis of its optical morphology .
emission line studies of ngc3077 ( e.g. , price & gullixson 1989 ; thronson , wilton & ksir 1991 ( twk91 ) ; martin 1997 , 1998 , 1999 ) show that streamers and supershells of ionized gas surround a bright core of h@xmath5 emission .
this core is also associated with soft x
ray emission , consistent with intense star formation in the center ( bi et al.1994 ) .
molecular gas in the center was detected by becker et al.(1989 ) using the iram 30 m telescope .
first interferometric maps of the molecular component of ngc3077 using the ovro millimeter array have been presented by thronson & carlstrom ( 1992 ) and meier , turner & beck ( 2001 ) .
some general information about ngc3077 is compiled in tab . 1 .
the in ngc3077 is strongly disrupted by the interaction with its two neighbors m81 and m82 : most of the neutral gas is located in a prominent tidal arm east of ngc3077 ( e.g. , van de hulst 1979 , yun 1994 , walter 1999 ) .
walter & heithausen ( 1999 , see also heithausen & walter 2000 ) recently discovered a massive molecular complex in this tidal feature .
numerical simulations suggest that the gas in the tidal arm has been stripped off the outskirts of ngc3077 during the latest close encounter with m81 , some 3 @xmath8 yr ago ( brouillet _ et al .
_ 1990 , thomasson & donner 1993 , yun et al.1993 ) .
this is in contrast to the suggestion of some authors ( price & gullixson 1989 , bi et al .
1994 , twk91 ) , that ngc3077 accreted ( ` stole ' ) from m81 in this recent interaction . in the following
, we will assume a distance to ngc3077 of 3.2mpc ( tammann & sandage 1968 ) .
we note that this value has recently refined to be 3.9mpc ( sakai & madore 2001 ) , i.e. , 20@xmath9 higher , however we decided to use the old distance to make it easier to compare our results to earlier studies . in this paper , we attempt to shed more light on the atomic and molecular gas as well as the starburst properties of ngc3077 and present a multi wavelength study of this fascinating galaxy . in section 2
we describe the observations of the atomic gas ( , vla ) , molecular gas ( co , ovro ) , optical broad and narrow band ( kpno ) and the nir ( nicmos hst ) .
the different wavelengths are then compared and discussed in detail in section 3 .
section 4 summarizes the most important results of our study and relates the gas and star formation properties of ngc3077 to those of its neighbor , m82 .
lc object & ngc3077 ( ugc5398 ) right ascension ( j2000.0 ) & 10@[email protected] declination ( j2000.0 ) & 68@xmath12 43@xmath13 59@xmath14adopted distance & 3.2 mpc scale & 15.5 pc/@xmath14 systemic velocity & 5 apparent magnitude ( b ) & 10.6 magdistance modulus & m@xmath15-m@xmath15=27.6 corrected absolute magnitude ( b ) & -17.0 mag blue luminosity l@xmath15 & 9.3@xmath16 l@xmath17 mass m@xmath18 & @xmath19 m@xmath20 mass to blue light ratio m@xmath18/l@xmath15 & 0.030.1 m@xmath20/l@xmath17h@xmath5 luminosity&@xmath21ergs@xmath6pa@xmath5 luminosity&@xmath22ergs@xmath6star formation rate & 0.05 m@xmath2yr@xmath6 h@xmath4 mass & @xmath23 m@xmath2
we observed ngc3077 in the co(1@xmath240 ) transition using the owen s valley radio observatory s mm array ( ovro ) in c , l and h configurations . in total
, 44 hours were spent on source ; the observational details are listed in table [ ovro_obs ] .
data were recorded using two simultaneous correlator setups resulting in velocity resolutions of 5 and 1.3 ( after hanning smoothing ) with a total bandwidth of 320 and 80 , respectively .
unless otherwise mentioned , the results presented in this paper were derived using the 1.3 resolution data .
flux calibration was determined by observing 1328 + 307 ( 3c286 ) and neptune for approximately 20 minutes during each observing run .
these calibrators and an additional noise source were used to derive the complex bandpass corrections .
the nearby calibrators 1031 + 567 ( 1.1jy ) was used as secondary amplitude and phase calibrator . the data for each array were edited and calibrated separately with the mma and the aips packages .
the _ uv_data were inspected and bad data points due to either poor atmospheric coherences or shadowing were removed , after which the data were calibrated .
two sets of datacubes were produced using the task imagr in aips , each of them cleaned to a level of two times the rms noise ( hgbom 1974 , clark 1980 ) . the first data set was made with natural weighting , leading to a resolution of @xmath25 ( @xmath26pc ) and emphasizing large scale structures ( noise : 16mjybeam@xmath6 ; 0.13k in a 5wide channel , 35mjybeam@xmath6 ; 0.29k in a 1.3 channel ) .
a second cube was produced using the robust weighting scheme ( briggs 1995 ) .
eventually , we chose a value of robust=0.25 , resulting in a beamsize of @xmath27 ( @xmath28pc ) and an rms noise of 22 mjy beam@xmath6 ( 0.44k , 5wide channels ) .
this latter resolution starts to resolve individual giant molecular clouds ( gmcs ) in ngc3077 . to separate real emission from noise when deriving moment maps we applied the following procedure : the natural weighted data cube was convolved to a circular beam with a fwhm of @xmath29 .
the smoothed map was then tested at the @xmath30 level ; if a pixel fell below this level , the counterpart in the cube was blanked .
after that , the remaining peaks were inspected .
emission that was present in 3 consecutive channels was considered to be real while all other remaining spikes were considered to be noise and blanked .
the final result was named the master cube .
this mask was used to blank the original natural and robust data cubes .
this method ensures that the same regions are included when inspecting cubes at different resolutions and with different signal to noise ratios . to derive physical properties ,
the co data were primary beam corrected .
lccc ovro configuration & c & l & h baselines & @xmath31 & @xmath32 & @xmath33 & @xmath34 & @xmath35 & @xmath36date of observation & 1999 sep . 22 & 1999 nov . 10 & 1999 dec . 16 & 1999 sep . 24 & 1999 nov
. 13 & 1999 dec .
17 total time on source & 800 min & 900 min & 950 min total bandwidths & & 124 mhz , 31 mhz & no . of channels & & 62 , 62 & velocity resolution & & 5 , 1.3 & central velocity & & 7 & angular resolution & & @xmath37 arcseclinear resolution & & @xmath38 pcrms noise & & 22 mjybeam@xmath6 ( 0.44 k ) ngc 3077 was observed with the nrao very large array ( vla ) in b , c and d configuration .
part of the d
array observations were affected by solar interference , which was removed by deleting _
uv_spacings shorter than typically 0.4 k@xmath39 . in total
, 17 hours were spent on source which were divided into 4 hours for d array , 2.5 hours in c array and 10.5 hours in b array ( see table [ vla_obs ] for a detailed description of the vla observations ) .
the flux calibration was determined by observing 1328 + 307 ( 3c286 ) for approximately 20 minutes during each observing run , assuming a flux density of 14.73 jy according to the baars _ et al . _
( 1977 ) scale .
this calibrator was also used to derive the complex bandpass corrections .
the nearby calibrators 1031 + 567 and 0945 + 664 were used as secondary amplitude and phase calibrators and their fluxes were determined to be 1.80 jy and 2.22 jy , respectivey . since the systemic velocity of ngc 3077 ( @xmath40 ) overlaps in velocity with galactic emission ,
each calibrator observation was observed at velocities shifted by + 300 and 300 to avoid galactic contamination . in the course of the calibration these observations were then averaged to give interpolated amplitude and phase corrections .
ngc3077 was observed using a 1.56 mhz bandwidth centered at a heliocentric velocity of 38 .
this band was divided into 128 channels resulting in a velocity resolution of 2.58 after online hanning smoothing . in editing and analysing the data we followed the same procedure as outlined in the previous section ( including the production of a master cube ) . obtaining fluxes for extended emission in multi - array observations is non trivial here we followed the procedure described in walter & brinks ( 1999 ) .
we also calculated a radio continuum image of ngc3077 at 21 cm by using the line free channels of the vla observations ; here we boosted the resolution to 5@xmath14 ( @xmath41pc ) by employing uniform weighting ( rms : 0.4 mjybeam@xmath6 ) .
lccc vla configuration & b & c & d baselines & @xmath42 & @xmath43 & @xmath44 & @xmath45 & @xmath46 & @xmath47date of observation & 1992 jan .
18 & 1992 apr . 20 & 1991 mar .
03 & 1992 jan .
19 & & 1992 sep .
03 total time on source & 647 min & 133 min & 245 min total bandwidth & & 1.56 mhz & no . of channels & & 128 & velocity resolution & & 2.58 & central velocity & & 14 & angular resolution & & @xmath48 arcseclinear resolution & & @xmath49 pcrms noise & & 1.0 mjybeam@xmath6 ( 3.7 k ) narrowband images were obtained january 2 - 10 , 2000 at the kitt peak national observatory 2.1 m telescope at @xmath50 ( 67fwhm ) and @xmath51 ( fwhm 84 ) .
the ccd frames were corrected for zero level bias offsets , pixel - to - pixel bias variations , and pixel - to - pixel sensitivity fluctuations . after constructing cosmic ray masks for each frame , and registering the frames , the images in each bandpass were co - added to obtain on - band and off - band images .
these images were flux calibrated using observations of spectrophotometric standards ( massey _ et al . _
the off - band continuum was scaled iteratively to match the stellar continuum in the on - band image and then subtracted from it .
the net emission - line image contains both [ nii ] @xmath52 and h@xmath5 emission .
the spectroscopic observations discussed here are described in martin ( 1998 ) .
we retreived nicmos broad and narrow band images from the hst archive ( pi : sparks , h : dataset n4k40jp3q , p@xmath5 : n4k40jp2q , both images were obtained on june 16 , 1998 ) . in calibrating the images
, we followed the approach given in bker et al .
the continuum was subtracted from the pa@xmath5 images and the flux in the pa@xmath5 line was derived using:@xmath53 where @xmath54 , @xmath55erg@xmath56@xmath6dn ( the latest value given in the nicmos handbook ) and cr is the countrate ( dns@xmath6 ) .
we have also obtained near infrared broad band j , h and k images from the 2mass project .
the distribution of neutral hydrogen at high angular resolution ( moment 0 map , corrected for primary beam attenuation ) is presented in fig 1 , right .
most of the in the region is situated in the prominent eastern tidal arm , little is associated with the optical counterpart of ngc3077 ( contours ) . to obtain accurate masses for ngc3077 we defined 3 different apertures as indicated in fig . 1 .
the largest aperture ( i ) accounts for most of the in the field of view .
aperture ii only includes which is roughly associated with the optical body of ngc3077 the smallest aperture ( iii ) only accounts for the central nucleus .
the total fluxes ( and masses , adopting a distance to ngc3077 of 3.2mpc ) for the apertures are : ( i ) : 192.3 jy ( @xmath57 m@xmath2 ) , ( ii ) : 42.13 jy ( @xmath58 m@xmath2 ) and ( iii ) : 11.30 ( @xmath59 m@xmath2 ) .
the content associated with the optical body of ngc3077 is therefore somewhere between @xmath60 m@xmath2 ( depending on the aperture chosen ) .
as mentioned in the introduction , the massive complex east of ngc3077 has presumably been stripped off the outskirts of ngc3077 during the recent interaction with m81 .
evidence for this comes from numerical simulations ( e.g. , yun et al .
1993 ) which show that the material in the tidal arm east of ngc3077 originally belonged to ngc3077 itself .
this scenario is supported by the high resolution observations : figure 2 shows the velocity field of ngc3077 and the region around it . the mean velocity of the tidal feature is @xmath61 and close to the one of ngc3077 ( @xmath62 ) . note that the systemic velocity of m81 is some @xmath63 with velocities reaching @xmath64 in the spiral arms pointing towards ngc3077 ( westpfahl & adler 1996 ) .
we also find no evidence for mass accretion and/or infall onto the centre of ngc3077 in our data .
all this renders the possibility that the material originally belonged to m81 unlikely .
an additional ( but somewhat weaker ) argument comes from the fact that the mass to blue light ratio in centre of ngc3077 is only @xmath65 m@xmath20/l@xmath17 ( see tab . 1 ) .
if the in the tidal arm previously belonged to ngc3077 the mass to blue light ratio is more like @xmath66m@xmath20/l@xmath17 , more typical for dwarf irregular galaxies ( this argument obviously only holds if ngc3077 was not a dwarf elliptical before the interaction ) .
we conclude that the tidal gas around ngc3077 originated from ngc3077 itself .
we note that since the total mass of the tidal system is similar in mass to ngc3077 itself ( heithausen & walter , 2000 ) , a complete detachment of this system ( i.e. , the creation of a true tidal dwarf system ) is likely ( but see meier , turner & beck 2001 ) .
the lines in fig .
2 indicate the orientation of 3 position velocity ( pv ) diagrams through the distribution around ngc3077 as shown in fig 3 .
multiple velocity components are visible where the emission in the tidal feature is strongest .
the cuts along the tidal arms towards m81 and m82 show a rather smooth velocity gradient away from the the maximum in the tidal feature . the first pv diagram ( fig . 3 , i ) cuts through the center of ngc3077 and the prominent tidal feature . the bright emission at offset 2.3@xmath13 is ngc3077 itself .
multiple velocity components are visible in the tidal feature ( around offset -2@xmath13 ) .
the second ( ii ) and third ( iii ) pv diagrams are oriented along the prominent tidal arms which roughly point towards m82 ( north , cut ii ) and m81 ( west , cut iii ) .
they show a rather smooth velocity gradient away from the maximum in the tidal feature .
the co(10 ) channel maps as obtained with ovro are shown in fig . 4 ( based on the natural weighted data at 1.3resolution ) .
clearly , the co emission in ngc3077 originates from many different sub structures .
these sub clumps are even better visible in position velocity diagrams as the one presented in fig
. 5 ( at high spatial and velocity resolution the orientation of this cut is indicated in fig . 6 , left ) .
the numbers in this plot refer to the individual subclumps discussed in sec .
3.2.2 . the co intensity distribution ( moment 0 ) is presented in fig
. 6 ( left : natural weighting , emphazising diffuse emission , right : robust weighting , focussing on the small scale structure ) .
globally , we detected 4 molecular regions which are labeled r1
. 6 .
r1 r3 have a complex velocity structure and consist of multi velocity components . to analyze the substructure of these molecular complexes we decomposed
the robust weighted ( @xmath67 , high velocity ( 1.3 ) resolution data cube with a gaussian clump fitting code ( stutzki & gsten 1990 ) as well as by eye . the clump fitting code ( ` gaussclump ' ) determines the maximum in the data cube and subtracts a 3d gaussian fit to the data ( ra , dec , velocity ) from the cube .
the procedure works iteratively on the residual until the specified rms level ( 5 sigma in this case ) is reached . in this procedure the spatial and velocity resolution is taken into account ( for further detail see stutzki & gsten 1990 ) . each clump determined by ` gaussclump ' was checked by eye using pv - cuts and by inspecting spectra at individual positions . in case of doubt we reran gaussclump with different parameters for contrast , minimum spatial and velocity structures and other parameters controlling the gaussian fit .
the parameters for our final clump decomposition of the data set are given in tab .
parameters regarding the spatial and velocity information as well as the flux ( column 2 to 7 ) refer to the fit results . in total
we identified 16 co complexes within ngc3077 ( tab .
4 ) . to emphasize the complex substructure
we show a pv - diagram ( along the line shown in fig . 6 , left ) in fig .
5 . nine out of the 16 components are at least partly visible in this diagram and are labeled by their respective number .
the size of most of the clumps is comparable to our beam implying that most of the co emission is not resolved at a linear resolution of @xmath68pc .
the radius and the velocity width shown in tab .
4 are deconvolved for the beam size as well as the velocity width .
for measured values less than 1.1 times the resolution , the estimate is listed as an upper limit .
the approximate errors for the measured quantities are given in the footnote of tab .
we derived cloud masses assuming that the clouds are virialized using @xmath69 ^ 2 \times r[{\rm pc}]$ ] ( rohlfs & wilson , 1996 ; see tab . 4 column 8) .
we also derived molecular masses assuming a galactic conversion factor ( @xmath70 , strong _
et al . _
1988 ) using @xmath71 ( column 9 ) .
complexes 1 to 5 are the components associated with region r1 , complexes 6 to 11 belong to r2 , complexes 12 to 15 belong to r3 and complex 16 is the diffuse co component in the north west of ngc3077 ( r4 ) .
the virial masses for clouds in r1r3 are significantly lower than the masses derived from converting co luminosities to column densities ( using the galactic conversion factor )
. in column 10 we present the derived conversion factor x@xmath72 based on the assumption that the complexes are indeed virialized . on average ,
we get values below @xmath73@xmath74kkms@xmath75 ( see the discussion in sec . 4 ) .
we note that meier , turner & beck ( 2001 ) derived a galactic conversion factor for ngc3077 ( see their paper for details ) . in the following
we will adopt a conversion factor of @xmath76 for ngc3077 ( roughly half of the galactic value ) , primarily for convenience . using this conversion factor ,
the total mass of ngc3077 is @xmath77 ( this is counting the complexes discussed above as well as diffuse emission that has not been catalogued ) .
becker et al.(1989 ) observed the center of ngc3077 with the iram 30 m telescope in the co(@xmath78 ) and co(@xmath79 ) transition at resolutions of 21@xmath14 and 13@xmath14 , respectively .
they only present the co(@xmath79 ) data in their paper and conclude that a giant molecular complex with a velocity width of 30 and a diameter of 320pc ( fwhm ) is located near the center . using a conversion factor of @xmath80
they derive a total molecular mass of @xmath81m@xmath2 and a virial mass of @xmath82m@xmath2 .
obviously , our higher resolution observations show that their complex breaks up into many smaller complexes .
the linewidth used by becker _
_ to derive the virial mass therefore measured the velocity dispersion of the co clumps rather than the intrinsic linewidth of a gravitational bound complex their virial mass is therefore too high .
nevertheless our mass estimate based on the measured co intensity of @xmath83 is about a factor of 1.5 less than the mass derived by becker _
et al . _
( taking the different conversion factors into accout ) .
this indicates that we may miss some extended emission in our interferometer data .
+ lllccccccc & ra ( 2000 ) & dec ( 2000 ) & v@xmath84 & dv@xmath85 & r@xmath86 & @xmath87 & m@xmath88 & m@xmath89 & x@xmath90&&&[]&[]&[pc ] & [ jy kms@xmath6 ] & [ @xmath92 ] & [ @xmath92 ] & [ @xmath93 1&10:03:18.95 & 68:43:57.37 & 24.5 & 4.5 & @xmath94 & 1.15 & @xmath95 & 14.5 & @xmath96 2&10:03:18.67 & 68:43:55.38 & 21.9 & 1.2 & @xmath94 & 0.42 & @xmath97 & 5.2 & @xmath98 3&10:03:19.32 & 68:44:00.36 & 19.2 & 2.1 & @xmath94 & 0.62 & @xmath99 & 7.8 & @xmath100 4&10:03:18.86 & 68:43:56.88 & 15.5 & 6.2 & 24 & 3.36 & 22.1 & 42.4 & 1.2 5&10:03:19.14 & 68:44:00.37 & 11.6 & 3.4 & @xmath94 & 1.07 & @xmath101 & 13.5 & @xmath102 6&10:03:18.77 & 68:43:56.38 & 6.4 & 3.1 & @xmath94 & 0.92 & @xmath103 & 11.6 & @xmath102 7&10:03:18.95 & 68:44:01.37 & -2.8 & 1.9 & 18 & 0.62 & 1.5 & 7.8 & 0.4 8&10:03:18.78 & 68:44:00.38 & -8.0 & 2.5 & @xmath94 & 0.56 & @xmath104 & 7.0 & @xmath102 9&10:03:18.86 & 68:43:59.88&-13.0 & 1.6 & @xmath94 & 0.55 & @xmath105 & 7.0 & @xmath98 10&10:03:20.33 & 68:44:01.82 & 3.8 & 1.4 & @xmath94 & 0.34 & @xmath102 & 4.3 & @xmath106 11&10:03:20.24 & 68:44:04.33 & 3.7 & 2.2 & @xmath94 & 0.55 & @xmath107 & 6.9 & @xmath108 12&10:03:20.24 & 68:44:03.83 & 2.5 & 1.9 & @xmath94 & 0.60 & @xmath109 & 7.6 & @xmath100 13&10:03:19.96 & 68:44:02.34 & 2.5 & 2.1 & @xmath94 & 0.62 & @xmath99 & 7.8 & @xmath100 14&10:03:20.15 & 68:44:03.83 & -2.8 & 2.0 & @xmath94 & 0.54 & @xmath110 & 6.8 & @xmath97 15&10:03:20.89 & 68:44:04.30 & -4.0 & 1.8 & @xmath94 & 0.51 & @xmath111 & 6.4 & @xmath100 16&10:03:14.10 & 68:44:09.05 & -21.3 & 7.9 & 103 & 18.4 & 153 & 232 & 1.5 + a ) deconvolved + b ) this assumes a standard co - to - h@xmath4 conversion factor of @xmath112@xmath56(kkms@xmath6)@xmath6 , strong et al .
( 1988 ) including a correction for helium .
if the recalibrated conversion factor of @xmath113@xmath56(kkms@xmath6)@xmath6 , hunter et al.(1997 ) is used , all co - based masses have to be reduced by @xmath114% . + c ) assuming m@xmath115 = m@xmath88 + note on errors : the measured quantities have the following errors : @xmath116(v@xmath84,dv)@xmath00.6 , @xmath116(r)@xmath0 10pc , @xmath116(s@xmath117 ) @xmath118 + the co and distribution are compared in fig . 7 ( left ) .
the greyscale and the thin countours represent the surface brightness .
the thick contours represent the co distribution as shown in fig . 4 ( right ) .
note the apparent asymmetry between the and the co distribtion : the maxima is located east of the co maximum . most of the co emission is located in an depression this suggests that atomic material in this region has largely turned molecular .
the total mass in the area shown is @xmath119m@xmath2 .
we created a total gas column density map by combining the and co moment maps ( fig .
10 , right ) .
the co data were first convolved to the resolution of the map ( @xmath120 ) and the h@xmath4 column density derived from employing a conversion factor ( sec .
3.2.2 ) was multiplied by two to get the proton density .
the molecular gas has a clear inpact on the total gas column density map ( reaching maximum surface densities of 5.2@xmath121 @xmath56 ) .
the total gas mass ( atomic and molecular ) in this plot is @xmath122m@xmath2 .
the thick contours represent the orientation of the j band image ( 2mass ) of ngc3077 m image presented in price & gullixson ( 1989 ) is off by @xmath123 .
the j band image mostly shows the distribution of the old stellar population of ngc3077 ( see sec .
it is evident that the apparent asymmetry of the gas discussed above is weaker if one considers the total gas column density .
nevertheless the gas distribution in the center of ngc3077 is clearly displaced towards the south east compared to the stellar distribution ( tracing the potential of the galaxy ) .
this implies that not only the gas in the outer parts of ngc3077 has been affected by the tidal forces within the triplet but also the gas in the center .
a similar displacement of the gaseous component with respect to the nucleus has been found in the center of m82 ( wei _ et al . _
2001 ) .
the optical appearance of ngc3077 is heavily influenced by dust absorption near the nucleus ( see the study by price & gullixson 1989 ) . in fig .
8 we present a b band image of ngc3077 .
the contours again represent the j band image of ngc3077 as obtained from the 2mass survey .
a nicmos h band image obtained on board hst of the same region is presented on the right side of this figure .
the near infrared image mainly shows the distribution of the older stellar population in ngc3077 .
9 ( left ) shows the same broad band image as fig .
the contours in the left panel of fig .
9 are the ovro co contours as presented in fig .
the dust feature in the south coincides very well with the regions r1 and r2 where we detected molecular gas .
this finding implies that co is a good tracer for dust in ngc3077 .
price & gullixson ( 1989 ) derive an mass of @xmath124m@xmath2 for the absorbing cloud .
our co data implies a total virial mass of @xmath125m@xmath2 ( only summing up the complexes c1 to c11 ) .
the difference might indicate that not all clumps ( 1 - 11 ) are associated with the foreground dust extinction .
this view is supported by the fact that this particular region shows a large co velocity gradient , suggesting a large distance between individual clouds along the line of sight ( see sec .
the optical images of ngc3077 also reveal the presence of radial absorption features ( indications can be already seen in fig . 9 , left ) .
the dust lanes look similar to the spokes of a cartwheel ; they are aligned radially in the north
west , however they look more ` curved ' in the south east ( see the discussion in sec . 4 )
. some of these dust lanes are associated with weak h@xmath5 emission as far out as 500pc from the center . to reproduce this situation we present 5
intensity profile cuts parallel to the right ascension in b band as well as in h@xmath5 emission .
the intensity cuts are indicated as 5 lines in fig .
11 and are plotted in fig . 9 ( right ) .
the lines showing the absorption represent the b band emission along one cut ; the h@xmath5 intensity profiles show emission .
major parts of the ongoing star formation in ngc3077 as traced by h@xmath5 emission is hidden by the central dust cloud ( sec .
3.3 ) . to reveal the true location of the starburst we analyzed pa@xmath5 imaging ( nicmos , hst ) .
the pa@xmath5 image is presented in fig .
10 ( right ) our ground based h@xmath5 image of the same region is shown on the left side of the figure . as discussed in martin ( 1998 ) the central starburst region
is surrounded by expanding h@xmath5 shells which are breaking out from the center ( these shells are also visible in pa@xmath5 ) .
the brightest emission is located just north of the molecular complexes r1 and r2 .
et al . _
( thr891998 ) suggested that this morphology might be indicative for a confinement of the ionized gas to the south and east where the molecular complexes and the column density peak are located .
however , from the pa@xmath5 imaging it is obvious that major parts of the starburst are hidden by the dust feature discussed in sec .
the contours in the right panel of this fig .
10 again represent the co emission from our high resolution ovro data ( see fig . 13 for a color composite ) .
the pa@xmath5 emission peaks right in between the regions r1 and r2 where only a minor fraction of is detected .
this implies that due to absorption the emission in ngc3077 only partly traces the ongoing star formation .
this view is supported by our observations in the radio continuum at 21 cm which are shown as contours in the left panel of fig . 10 .
although our resolution of 5@xmath14 ( @xmath126pc ) is too poor to resolve the emission , the radio continuum emission clearly shows that the most active starforming region is associated with the molecular complexes r1 and r2 . a comparison between the co kinematics and our slit spectroscopy ( fig . 12 ,
the orientation of the slit is indicated in fig .
10 , left ) in this particular region reveals another interesting aspect : the center velocity of the emission is between the velocities of the molecular complexes r1 and r2 .
it is suggestive that this region of strong star formation has disrupted its parental molecular cloud which was located between r1 and r2 .
note that even the faint diffuse co complex r4 in the north west ( sec .
3.2.1 ) coincides with h@xmath5 emission at the same velocity . a quantitative analysis of the narrow band images shown in fig .
10 yields a total pa@xmath5 flux of ngc3077 of f(pa@xmath5)=@xmath127erg@xmath56s@xmath6 ( consistent with the value derived by bker _
et al . _
the total h@xmath5 flux is f(h@xmath5)=@xmath128 erg@xmath56s@xmath6 ( l(h@xmath5)=@xmath129ergs@xmath6 , using @xmath130 ) .
the pa@xmath5/h@xmath5 intensity ratio in the outer parts of ngc3077 ( where the absorption of h@xmath5 is negligible ) is about 0.17 .
this is close to the ratio of the emission coefficients for @xmath131k and case b recombination of @xmath132pa@xmath133ha@xmath134 ( osterbrok 1989 , table 4.4 ) .
the highest absorption in the galaxy is found close to the pa@xmath5 peak ( intensity ratio ( pa@xmath5/ha@xmath5 ) = 1.30 ) .
this corresponds to a local extinction for this particular region of about an order of magnitude ( @xmath02.5mag ) . if there was no extinction we would expect an h@xmath5 flux of 5.2@xmath135erg@xmath56s@xmath6 based on the observed pa@xmath5 flux ( l(h@xmath5)=@xmath136ergs@xmath6 ) .
we used sfr(m@xmath2yr@xmath6)=l(h@xmath5)/(@xmath137)ergs@xmath6 ( kennicutt _ et al . _
1994 ) to derive a star formation rate ( sfr ) based on the extinction corrected h@xmath5 emission of 0.05m@xmath2yr@xmath6 .
this is some 25% larger than the value obtained from the h@xmath5 emission alone and is a measure for the total extinction within ngc3077 .
note that this is a factor of a few less than the value derived by meier , turner & beck ( 2001 ) who used 2.6 mm radio continuum emission to derive the sfr in ngc3077 .
the outflow in ngc3077 ( as traced by h@xmath5 and pa@xmath5 emission ) reaches galactocentric distances of up to 1.5kpc .
given the intrinsic small size of ngc3077 this is remarkable ( for a detailed discussion see martin 1997 ) . at larger scale ,
an anti correlation between diffuse and the h@xmath5 outflow is evident : in fig .
11 we plot the in blue and the h@xmath5 in red ( for comparison , the optical broad band image is shown in green as well ) . at large galactocentric radii , the is predominantly situated @xmath138 the expanding shells .
typical surface densities at these galactocentric radii are @xmath139@xmath56 .
one explanation for these feature may be that the pressure of the interior of the h@xmath5 shells is high enough to push the out to larger radii .
the wealth of data presented here give exciting new insights on the interplay between the atomic gas , molecular gas and ongoing star formation in ngc3077 .
the total ( atomic and molecular ) gas content in the centre of ngc3077 is clearly displaced from the stellar component of ngc3077 this implies that not only the gas at large galactocentric radii is affected by the interaction within the triplet .
the mass associated with ngc3077 is @xmath19m@xmath2(depending on the aperture chosen , sec .
it was suggested in numerical studies of the triplet that the tidal material around ngc3077 originally belonged to ngc3077 itself ; the high resolution data of the morphology and kinematics show observational evidence for this scenario .
based on the ovro interferometer data we estimate a molecular mass of @xmath140 for ngc3077 ( sec .
3.1.2 ) . in total
, 16 complexes have been detected with our 2@xmath14 ( @xmath114pc ) beam .
all of the complexes have virial masses which are lower than the masses calculated from the co luminosity ( assuming a galactic conversion factor ) .
this implies that the conversion factor for ngc3077 is lower than the galactic value . at first glance ,
this seems surprising since the conversion factor is usually believed to be higher in dwarf galaxies as compared to spirals such as our own galaxy ( arimoto _ et al .
_ 1996 , wilson 1995 ) .
however , the metallicity of ngc3077 is high ( o / h around solar ) for a dwarf ( martin 1998 ) , so this objection may not apply .
the higher co emissivity and lower conversion factor can also be understood by an increased kinetic temperature and cosmic ray heating near the starburst powering the outflow in ngc3077 .
e.g. , in the case of m82 , wei et al .
( 2001 ) have shown that x@xmath72 may be as low as @xmath141@xmath74kkms@xmath75 .
the prominent star forming regions as well as the outflow visible in h@xmath5 and pa@xmath5 , as well as the radio continuum source near the starburst already suggest that the state of the molecular gas in ngc3077 may be affected by the starburst ( similar to the case in m82 ) . indeed , as already pointed out by becker _
et al . _
( 1989 ) , a co(2 - 1)/co(1 - 0 ) line intensity ratio of 0.82 ( on a t@xmath142 scale ) indicates that a substantial fraction of the gas is heated to kinetic temperatures @xmath143k . in a more recent study meier _
et al . _
2001 , using observations in the co(3 - 2 ) transition , estimated a kinetic gas temperature in ngc3077 of t@xmath144k using lvg calculations .
we note that the co(2 - 1)/co(1 - 0 ) line intensity ratio of 0.82 as reported by becker _
et al . _
( 1989 ) refers to antenna temperatures . taking antenna efficiencies into account
the co(2 - 1)/co(1 - 0 ) line intensity ratio is about 1.3 ( t@xmath145 scale , efficiencies as in guelin
_ et al . _
1995 ) indicative for even larger temperatures . using the ratio above and co(3 - 2)/co(1 - 0)=0.7 ( meier _ et al . _ 2001
; corrected for efficiencies at 115ghz ) we find t@xmath146k using lvg models .
this supports the idea that the molecular gas in ngc3077 is heated by the starburst resulting in a lower conversion factor ( cf .
weiss et al .
2001 in the case of m82 ) .
we find a striking correspondence between the co distribution and the dust cloud seen in the optical . from the pa@xmath5 and h@xmath5 imaging
we estimate the extinction in the central dust complex to be an order of magnitude ( @xmath147mag ) . surprisingly , co emission is also coincident in position and velocity with a faint star forming region located 500pc west of ngc3077 , far off the central star formation .
a comparison of the ionized outflow with the shows that the distribution of around ngc3077 s center is affected by the outflow .
a puzzle is the presence of the ` radial ' dust fingers which reach lengths up to @xmath148kpc .
it may be that these dust lanes originally belonged to rudimentary spiral arms which have been ripped apart by the interaction .
the dust lanes may have also been blown out of the ngc3077 s center by the violent sf ( e.g. , by radiation pressure ) .
somewhat counter
intuitively , the most prominent lanes are associated with h@xmath5 emission .
one way to interpret the apparent correlation of dust with h@xmath5 is that the surface of the dust filaments is exposed to the strong radiation field of the central starforming region which causes the weak emission on the rim of the filament .
correcting for extinction , we derive a star formation rate ( sfr ) of 0.05m@xmath2yr@xmath6 for ngc3077 .
although this value does not look dramatic one has to keep in mind that the star formation in ngc3077 is very concentrated towards the center .
the effective area where massive star formation is found is @xmath149pc@xmath150 resulting in a sfr per surface area of @xmath151m@xmath2yr@xmath6pc@xmath152 , much higher than what is usually found in spiral galaxies ( e.g. martin & kennicutt 2001 ) .
we will now compare the neutral hydrogen , molecular gas and starburst properties of ngc3077 to the corresponding parameters of the prototypical starburst galaxy m82 ( situated at a projected distance of only @xmath153kpc ) .
about 10% of the total gas mass ( including the tidal arm ) of ngc3077 is associated with the optical galaxy ; the corresponding fraction for m82 is @xmath15470% ( yun et al .
this suggests that ngc3077 lost most of its gas during the interaction whereas m82 was able to hold it .
the central starbursting region of m82 still has a reservoir of about @xmath155 of neutral hydrogen ( more than @xmath04 times the corresponding value of ngc3077 ) ; the molecular gas content of m82 is @xmath156m@xmath2(@xmath157 times higher than in ngc3077 , wei et al .
2001 , d@xmath158=3.2mpc ) .
the ratio of molecular to atomic gas in m82 s center is therefore of order unity , significantly higher than the same ratio in ngc3077 ( @xmath159 ) .
this implies that the efficiency to form molecular out of atomic gas is much higher in m82 as compared to ngc3077 .
interestingly the star formation efficiencies ( sfe ) are similar in ngc3077 and m82 : our and co observations yield a sfe for ngc3077 of l@xmath160/m ( ) @xmath161 ( = @xmath162ergs@xmath6 .
the corresponding value for m82 is about @xmath163 ( adopting l@xmath164ergs@xmath6 ; heckman _ et al . _ 1990 ,
the true h@xmath5 flux in the centre of m82 is presumably somewhat higher due to the edge on orientation of m82 ) . in that sense ngc3077
can be regarded as a scaled down version of m82 : it has a similar star formation efficiency and shows a prominent outflow .
the ` only ' differences may be that the overall total masses are different and that ngc3077 lost a larger fraction of its atomic gas during the interaction as m82 did . in any case
the starburst activity in both galaxies was presumably triggered by the redistribution of atomic and molecular gas in their centers due to the gravitational interaction within the m81 triplet .
fw acknowledges nsf grant ast9613717 .
aw acknowledges dfg grant sfb494 .
clm acknowledges support from a sherman fairchild fellowship .
research with the owens valley radio telescope , operated by caltech , is supported by nsf grant ast9613717 .
support of this work was also provided by a grant from the k.t . and e.l .
norris foundation .
kpno / noao is operated by the association of universities for research in astronomy ( aura ) , inc . under cooperative agreement with the national science foundation .
the national radio astronomy observatory ( nrao ) is operated by associated universities , inc .
, under cooperative agreement with the national science foundation .
this research has made use of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , caltech , under contract with the national aeronautics and space administration ( nasa ) , nasa s astrophysical data system abstract service ( ads ) , and nasa s skyview .
the 2mass project is a collaboration between the university of massachusetts and the infrared processing and analysis center ( jpl / caltech ) .
young j.s .
, kleinmann d.e . ,
allen l.e . , 1988 ,
apj 334 , 63 + yun m.s . , 1997
, tidal interactions in m81 group , in _ proceedings of iau symposium 186 : + ` galaxy interactions at low and high redshift ' _ , ed . | we present a comprehensive multi wavelength study of the nearby interacting dwarf galaxy ngc3077 ( member of the m81 triplet ) .
high resolution vla observations show that most of the atomic gas ( @xmath090% ) around ngc3077 is situated in a prominent tidal arm with a complex velocity structure .
little ( @xmath1m@xmath2 ) is associated with ngc3077 itself .
high resolution ovro observations of the molecular component ( co ) reveal the presence of 16 molecular complexes near the center of ngc3077 ( total mass : @xmath3m@xmath2 ) . a virial mass analysis of the individual complexes yields a lower co to h@xmath4 conversion factor in ngc3077 than the galactic value - a surprising result for a dwarf galaxy .
the lower conversion factor can be explained by extreme excitation conditions and the metallicity of the molecular gas .
the total ( atomic and molecular ) gas content in the centre of ngc3077 is displaced from the stellar component of ngc3077 this implies that not only the gas at large galactocentric radii is affected by the interaction within the triplet but also the center .
we speculate that the starburst activity of ngc3077 was triggered by this redistribution of gas in the center : h@xmath5 as well as pa@xmath5 images show the presence of violent central star formation as well as dramatic ionized supershells reaching galactocentric distances of @xmath01kpc .
some of these supershells are surrounded by neutral hydrogen . in a few cases ,
the rims of the ionized supershells are associated with dust absorption .
the most prominent star forming region in ngc3077 as probed by pa@xmath5 observations is hidden behind a dust cloud which is traced by the molecular complexes .
correcting for extinction we derive a star forming rate of 0.05 m@xmath2year@xmath6 , i.e. given the reservoir in atomic and molecular gas in ngc3077 , star formation may proceed at a similar rate for a few 10@xmath7 years .
the efficiency to form stars out of molecular gas in ngc3077 is similar to that in m82 . |
to benefit from the recent progress on the calculation of field - theoretic bound states in light - cone quantization @xcite , we explore methods by which form factors and scattering amplitudes can be extracted nonperturbatively . in the case of form factors , this is relatively straightforward ; well - known formulas @xcite yield the form factors as overlap integrals of fock - state wave functions . for scattering amplitudes ,
the way is less certain .
one possible method @xcite is discussed briefly here .
others have been considered by krger @xcite , ji and surya @xcite , and fuda @xcite .
the formulations given are in terms of light - cone coordinates @xcite , where @xmath1 plays the role of time and the conjugate variable @xmath2 is the light - cone energy .
the light - cone three - momentum is @xmath3 .
an eigenstate @xmath4 of the light - cone hamiltonian operators @xmath5 , @xmath6 and helicity @xmath7 is written as a fock - state expansion @xmath8[d^2k_\perp ] \psi_{p,\sigma}^{(n)}(x,{\bf k}_\perp ) |n:\underline{p}_i \rangle\ , , \nonumber\ ] ] with @xmath9[d^2k_\perp]=\int\delta(1-\sum_i x_i ) \prod_i \frac{dx_i}{\sqrt{x_i } } 16\pi^3\delta(\sum_i { \bf k}_{\perp i } ) \prod_i \frac{d^2k_{\perp i}}{16\pi^3}\ ] ] and where the @xmath10 are wave functions for @xmath11 particles , @xmath12 are longitudinal momentum fractions , and @xmath13 are relative transverse momenta .
use of light - cone coordinates brings several advantages , including boost invariance of the wave functions . the eigenvalue problem @xmath14 for fixed @xmath7 determines the wave functions as solutions of a coupled set of integral equations .
a method frequently applied to these equations is discrete light - cone quantization ( dlcq ) @xcite , which approximates the integrals by the trapezoidal rule and computes the wave functions on an equally spaced momentum grid .
any bound - state property can then , in principle , be calculated from these wave functions .
the grid is parameterized by a longitudinal resolution @xmath15 and transverse resolution @xmath16 , such that longitudinal momentum fractions are multiples of @xmath17 and transverse momenta have as many as @xmath18 values in each direction .
the value of @xmath16 is associated with a cutoff @xmath19 on the invariant mass of each constituent and with the choice of transverse momentum scale @xmath20 .
for a spin-@xmath21 fermion , the two form factors can be obtained from matrix elements of the plus component of the electromagnetic current @xmath22 @xmath23 these can be reduced to overlap integrals @xcite @xmath24[d^2k_\perp ] \psi_{p+q,1/2}^{(n)*}(x,{\bf k}_\perp^\prime ) \psi_{p,1/2}^{(n)}(x,{\bf k}_\perp)\,,\ ] ] @xmath25[d^2k_\perp ] \psi_{p+q,1/2}^{(n)*}(x,{\bf k}_\perp^\prime ) \psi_{p,-1/2}^{(n)}(x,{\bf k}_\perp)\ , , \nonumber\ ] ] in the frame where the photon momentum @xmath26 is written @xmath27 and @xmath28 for the model studied by brodsky , hiller , and mccartor @xcite , an explicit calculation of @xmath29 has been done @xcite . in this model , a bare fermion acts as a source and sink for bosons of mass @xmath30 .
the lowest massive eigenstate is a fermion dressed by a boson cloud .
the theory is regulated by a pauli villars boson @xcite with an imaginary coupling , and renormalized by fits of physical quantities to `` data . '' because no spin - flip interactions are included , @xmath31 is zero .
results for @xmath29 are shown in fig .
[ jhiller : fig : formfactor ] .
the large - momentum - transfer value of @xmath29 is the bare fermion probability and therefore is not zero . =
the center - of - mass cross section for two - body scattering ( @xmath32 ) is @xcite @xmath33 where @xmath34 is the invariant amplitude obtained from the @xmath35 matrix @xmath36 with @xmath37 .
the @xmath0 matrix for scattering of composites is given by @xcite @xmath38 here @xmath39 and @xmath40 are composite - particle eigenstates of the light - cone hamiltonian @xmath41 , and the operator @xmath42 is defined by @xmath43 -\frac{m_b^2+p_{b\perp}^2}{p_b^+/p^+}b^\dagger\,,\ ] ] with @xmath44 the creation operator for the @xmath45 particle , _ i.e. _ @xmath46 .
this construction generalizes one presented some time ago by wick @xcite .
details can be found in ref .
given numerical solutions for the composite - particle eigenstates , obtained with dlcq , the most difficult remaining task is the estimation of the matrix element of @xmath47 . for this type of matrix element , the recursion method of haydock
@xcite has worked well .
a nonrelativistic application is described in @xcite ; an application to the field - theoretic model studied in @xcite is in progress .
this work was supported in part by the minnesota supercomputing institute through grants of computing time and by the department of energy , contract de - fg02 - 98er41087 .
hiller , j.r .
, `` nonperturbative calculation of scattering amplitudes , '' to appear in the proceedings of the fourth workshop on continuous advances in qcd , minneapolis , minnesota , may 12 - 14 , 2000 , hep - ph/0007231 .
hiller , j.r .
, `` on the use of discrete light - cone quantization to compute form factors , '' in _ the transition from low to high q form factors _ , edited by g. strobel and d. mack , tjnaf workshop proceedings , 1999 , pp . 193 - 199 , hep - ph/9909471 . | methods are described for the nonperturbative calculation of wave functions and scattering amplitudes in light - cone quantization .
form factors are computed from the boost - invariant wave functions , which appear as coefficients in a fock - state expansion of the field - theoretic eigenstate .
a technique is proposed for calculating scattering amplitudes from matrix elements of a @xmath0 operator between such composite - particle eigenstates .
umn - d-00 - 4 + july 2000 |
the transition from an insulating to metallic behavior in doped manganites , observed as a function of temperature or filling , is one of the outstanding problems in the field of strongly correlated systems .
the theoretical challenge is to understand a rich variety of metallic , insulating , magnetically ordered and orbital ordered phases of doped manganites in terms of the dynamics of correlated @xmath1 electrons which involves their orbital degrees of freedom @xcite .
magnetic interactions : spin superexchange ( se ) and double exchange ( de ) , as well as orbital se , frustrate each other and their competition leads to several magnetic phases , including phase separation @xcite .
while orbital order supports @xmath2-type antiferromagnetic ( af ) phase in lamno@xmath3 @xcite , one expects that @xmath1 orbitals play an important role in the insulating ferromagnetic ( fm ) phase at finite doping , below the crossover to a metallic fm phase @xcite .
here we investigate a one - dimensional ( 1d ) model derived from the realistic spin - orbital model for manganites @xcite , showing an interplay between spin and orbital order .
we investigate the 1d orbital @xmath4-@xmath5 model for manganites for a given configuration of core spins @xmath6 , @xmath7 which consists of the hopping term @xmath8 , the orbital superexchange @xmath9 for the @xmath1 electrons , superexchange @xmath10 for the core spins formed by @xmath0 electrons , and an alternating jahn - teller ( jt ) potential @xmath11 .
as we consider 1d chains in @xmath12-direction , only electrons within @xmath13 orbitals are mobile , while those within @xmath14 orbitals are localized : @xmath15 where @xmath16 and @xmath17 are the annihilation and creation operators restricted to the hilbert space without double occupancies . due to strong hund s coupling between @xmath1 and @xmath0 electrons ,
the hopping amplitude , @xmath18 , depends of the relative angle @xmath19 of the two classical core spins at sites @xmath20 and @xmath21 , see e.g. ref .
the complex phase @xmath22 can be neglected for a 1d chain @xcite .
it is convenient to introduce orbital pseudospin operators @xmath23 , with eigenstates @xmath24 and @xmath25 . with the electron number operator @xmath26 restricted to the hilbert space without double occupancies , the superexchange term which follows from refs . @xcite and
@xcite becomes @xmath27 .
\end{split}\ ] ] like the kinetic energy , it depends on the direction of core spins via the factors @xmath28 .
the first term @xmath29 favors fm spin order for ion pairs mn@xmath30mn@xmath30 , while the second and third ones @xmath31 favor af order for mn@xmath30mn@xmath30 pairs [ @xmath32 , and for mn@xmath30mn@xmath33 pairs [ @xmath34 .
the se parameter @xmath35 is given by the high - spin excitation energy @xmath36 and the prefactors of the above terms were chosen to preserve the relative magnitude of the different excitations for realistic parameter values @xcite . throughout this work ,
we use @xmath37 and choose @xmath4 as unit of energy . the third term in eq .
( 1 ) is an af heisenberg interaction between the core @xmath0 spins , @xmath38 as we used classical core spins @xmath39 of unit length , their physical value @xmath40 was compensated by a proper increase of @xmath41 .
finally , we include an alternating jt potential of the form @xmath42 we employ a markov chain monte carlo ( mcmc ) algorithm for the classical core spins ( see e.g. ref .
@xcite ) combined with exact diagonalization for the electron degrees of freedom @xcite . in this method , the many - particle hamiltonian determined by @xmath43 , which are in turn given by the core spin configuration @xmath6 , is solved by exact diagonalization taking into account its block - diagonal structure .
the lowest states of each block are used to evaluate the trace over the fermionic degrees of freedom , @xmath44 , which gives the statistical weight for @xmath6 and which is sampled by the mcmc .
autocorrelation analysis was used in order to ensure that enough configurations were skipped between measurements .
we begin with analyzing spin and orbital correlations at increasing doping .
figure [ ssk ] shows the core spin structure factor , @xmath45 , for an @xmath46 chain and doping @xmath47 ( @xmath48 , where @xmath49 is electron filling ) near the system filled by @xmath50 electron per site ( as in lamno@xmath3 ) , and for three different parameter sets at low temperature .
the orbital structure factor , @xmath51 , which is renormalized by the squared electron density for better comparability of results obtained at different fillings , for the same parameters is shown in fig .
altogether , the orbital and magnetic order obtained at @xmath52 and @xmath53 _ complement each other _ polarized ( ferro ) @xmath24 orbitals at @xmath54 induce af spin order [ figs .
[ ssk](a ) and [ ttk](a ) ] , but if the occupied @xmath1 orbitals alternate , fm spin order is supported instead [ figs . [ ssk](b ) and [ ttk](b ) ] . increasing doping induces a gradual crossover towards fm order due to the de mechanism , being the most efficient in the metallic phase when only @xmath24 orbitals are occupied .
this insulator - metal transition is hindered by the jt potential which favors orbital alternation [ fig .
[ ttk](c ) ] and suppresses fm correlations [ fig .
[ ssk](c ) ] . for a finite se between core spins @xmath54 and in absence of jt potential ( @xmath55 ) , a transition from af to fm spin order is induced by doping [ fig .
[ ssk_jse0.02_ejt0 ] ] , and predominantly the mobile @xmath24 orbitals are occupied at all doping levels [ see fig . [ ttk_jse0.02_ejt0 ] ] . in the spectral density depicted in fig .
[ ak_b100_jse0.02_0h ] , one sees signals with low weight for the @xmath25 orbitals and a rather broad incoherent peak due to the @xmath24 excitations .
the broad peak contains dispersionless states stemming from @xmath24 electrons next to the few @xmath25 electrons .
additionally , a dispersive band is found which comes from @xmath24 electrons moving in the af background , showing a remarkable similarity to a one - orbital model @xcite . at the higher temperature @xmath56 ,
more @xmath25 orbitals are occupied ( @xmath57% vs. @xmath58 at @xmath59 ) , and their weight in the spectral density increases [ fig .
[ ak_b30_jse0.02_0h ] ] .
the dispersionless broad feature becomes stronger , as the motion of @xmath24 electrons is hindered by @xmath25 electrons .
both effects are further enhanced when spin order weakens at very high temperature @xmath60 .
+ a chain doped with three holes @xmath61 has fm order of core spins [ see fig .
[ ssk_jse0.02_ejt0 ] ] and practically only mobile @xmath24 electrons [ fig .
[ ttk_jse0.02_ejt0 ] ] , due to the dominating de mechanism which favors both fm order and ferro orbital polarization . the spectral density at low temperature @xmath59 is shown in fig . [ ak_b100_jse0.02_3h ] and
one sees a coherent band with almost unrenormalized dispersion @xmath62 due to @xmath1 electrons moving in @xmath24 orbitals , and only a weak signal above the fermi energy for the immobile excitations in @xmath25 orbitals . for @xmath56
[ see fig .
[ ak_b30_jse0.02_3h ] ] , the peaks of the @xmath24-band are broadened , the bandwidth is slightly reduced and a small spectral weight is induced in @xmath25 states below the fermi energy ( @xmath63 for the filling by 9 electrons ) . again , these effects become more pronounced for the high temperatures @xmath60 , see fig .
[ ak_b10_jse0.02_3h ] , where @xmath64 electrons occupy @xmath25 orbitals .
the broadening of the peaks and the reduction of the bandwidth result from core spin fluctuations weakening fm correlations , while the additional ( incoherent ) signals besides the tight - binding band are caused by the @xmath25 electrons .
the reduction of fm correlations and the increasing electron density within @xmath25 orbitals occur simultaneously with rising temperature .
yet another situation arises ( at @xmath54 ) in presence of a small alternating jt potential @xmath65 .
it favors alternating orbital pattern [ fig .
[ ttk_jse0.02_ejt0.1 ] ] , which in turn destroys the core spin af order in the undoped chain [ fig .
[ ssk_jse0.02_ejt0.1 ] ] . upon doping ,
the initially weak fm correlations increase , @xmath25 orbitals are depleted and the number of @xmath24 electrons even increases slightly to @xmath66 for 2 and 3 holes . for the undoped chain there are just three peaks in the spectral function @xmath67 [ fig .
[ ak_b100_ejt01_0h ] ] : a central peak from the @xmath24 electrons trapped between two occupied @xmath25 orbitals , and two side peaks due to the @xmath25 electrons .
very similar results ( not shown ) are obtained for @xmath68 and @xmath69 , where the orbitals also alternate , see above . doping with one hole
removes an @xmath25 electron and a three - site well is created , leading to three @xmath24 signals in the spectral density depicted in fig .
[ ak_b100_ejt01_1h ] .
the spectral function @xmath67 for doping with three holes [ fig .
[ ak_b100_ejt01_3h ] ] shows new features on top of the tight - binding - like band at @xmath69 ( fig .
[ ak_b100_jse0.02_3h ] ) , that arise from the two remaining @xmath25 electrons and from the alternating on - site potential seen by @xmath24 electrons .
note that the fm is weaker at this doping level than when the jt potential is absent , because de is hindered . from doping @xmath70 onward ,
the chain is fully polarized and fm order is enhanced , see figs .
[ ssk_jse0.02_ejt0.1 ] and [ ttk_jse0.02_ejt0.1 ] .
a strong jt potential @xmath71 delays a crossover to the metallic fm phase to very high doping @xmath72 . finally , we determined the de constant for the present 1d chain , @xmath73 , where @xmath74 and @xmath75 @xcite . while the de increases with @xmath76 and
is simply frustrated by the af se between core spins ( @xmath77 does not depend on @xmath41 ) , increasing jt potential reduces @xmath77 in the insulating regime .
for instance , at @xmath61 one finds @xmath78 , 0.027 and @xmath79 for @xmath53 , 0.1 and @xmath80 .
however , at doping @xmath81 the fm phase is already metallic for the entire range of @xmath82 , with @xmath83 .
even in one dimension , the orbital @xmath4-@xmath5 model shows an interesting competition of fm and af se terms . at low temperatures and for the undoped chain , a relatively small af se @xmath41 between
the @xmath0 core spins already suffices for a transition from fm spin order and alternating orbital order ( @xmath68 ) to af spin order and ferro orbital polarization ( @xmath84 ) .
these different situations , both encountered qualitatively in lamno@xmath3 , result in quite distinct one - particle spectra .
because of the close interplay of orbital- and spin - degrees of freedom , the change between these two scenarios can also be triggered by an alternating jt potential ( @xmath85 ) . at higher temperatures ,
the competing interactions lead to further changes in the spectra as both the spin and the orbital order are destroyed . for the doped chain ( @xmath86 ) without jt
potential , the kinetic energy favors fm order and orbital polarization .
this metallic fm phase may be destroyed either by finite jt potential , or by increasing temperature , when both the fm order and the orbital polarization are reduced , leading in both cases to incoherent spectra with reduced bandwidth .
this behavior indicates that an insulating phase may be induced either by increasing temperature , or by enhancing the jt distortions by chemical substitution .
summarizing , the present study highlights : ( @xmath20 ) a complex interplay between spin , orbital and change degrees of freedom , and ( @xmath87 ) the importance of the coupling to the lattice all these all these aspects decide about spin and orbital correlations in doped manganites and have to be included in their realistic microscopic models@xcite . | starting from a spin - orbital model for doped manganites , we investigate a competition between ferromagnetic and antiferromagnetic order in a one - dimensional model at finite temperature .
the magnetic and orbital order at half filling support each other and depend on a small antiferromagnetic superexchange between @xmath0 spins and on an alternating jahn - teller potential .
the crossover to a metallic ferromagnetic phase found at finite doping is partly suppressed by the jahn - teller potential which may localize @xmath1 electrons .
[ published in phys .
. sol . * 242 * , 311 ( 2005 ) ] |
though quantum chromodynamics ( qcd ) is believed to be the underlying theory of the strong interactions , low energy phenomena such as static properties of hadrons defy solutions based on qcd because of formidable mathematical complexities .
the pertinacity of qcd in the low energy regime have led to a great deal of efforts to construct an effective theory for the strong interactions . in pursuit of this aim , the chiral quark - soliton model ( @xmath0qsm ) ,
also known as the semibosonized nambu - jona - lasinio model , emerged as a simple and successful effective theory to describe the low energy phenomena without loss of important properties of qcd such as chiral symmetry and its spontaneous breaking .
originally , the idea of finding the soliton in a model with quarks coupled to pions was realized by kahana , ripka and soni @xcite and birse and banerjee @xcite .
the bound states of the valence quarks were well explored in the model while it suffered from the vacuum instability .
having studied the qcd instanton vacuum in the low - momenta limit , diakonov and petrov obtained an effective lagrangian resembled the nambu - jona - lasinio type model @xcite with @xmath1-quark vertices .
they showed that the resulting bosonized low - momenta theory is equivalent to the @xmath0qsm free from the vacuum instability .
in fact , the effective model by diakonov and petrov has several important virtues : the mechanism of the spontaneous breaking of chiral symmetry is well explained in a natural way .
it provides also a renormalization scale by the inverse of the average size of the instanton @xmath2 mev .
in addition , the @xmath0qsm throws a bridge between the nonrelativistic constituent quark model and the skyrme model .
the baryon in this model is regarded as @xmath3 valence quarks coupled to the polarized dirac sea bound by a nontrivial chiral field configuration in the hartree approximation @xcite .
the identification of the soliton as the baryon is acquired by the semiclassical collective quantization @xcite which is performed by integrating over zero - mode fluctuations of the pion field around the saddle point .
the model enables us to describe quantitatively a great deal of static properties of the nucleon such as baryon octet - decuplet mass splittings @xcite , magnetic moments @xcite , axial constants @xcite , electromagnetic form factors @xcite and so on ( see for example a recent review @xcite ) .
let me first sketch the @xmath0qsm . because one can find the detailed formalism elsewhere , i want to present the general idea of the model briefly .
the @xmath0qsm in su(3 ) is characterized by a low - energy partition function in euclidean space given by the functional integral over pseudoscalar meson and quark fields : @xmath4 where @xmath5 stands for the dirac differential operator @xmath6 with the pseudoscalar chiral field @xmath7 the @xmath8 denotes the matrix element of the current quark mass given by @xmath9 @xmath10 represents the usual gell - mann matrices normalized as @xmath11 .
@xmath12 designates the dynamical quark mass arising from the spontaneous breaking of chiral symmetry , which is in general momentum - dependent @xcite . for convenience ,
we shall regard @xmath12 as a constant and introduce the ultraviolet cutoff via the proper time regularization .
the @xmath13 , @xmath14 and @xmath15 are respectively defined by @xmath16 we assume isospin symmetry , _ i.e. _ @xmath14 is taken to be zero .
the operator @xmath5 is expressed in euclidean space in terms of the euclidean time derivative @xmath17 and the dirac one - particle hamiltonian @xmath18 @xmath19 with @xmath20 @xmath21 and @xmath22 are the well - known dirac hermitian matrices .
the @xmath23 is defined by @xmath24 , which is introduced to avoid certain divergences in some observables such as the isovector electric charge radius in the chiral limit ( @xmath25 ) .
note that the effective chiral action given by eq.([eq : z ] ) contains the wess
zumino and four - derivative gasser leutwyler terms with correct coefficients in the gradient expansion .
therefore , at least the first four terms in the gradient expansion of eq.([eq : z ] ) are correctly reproduced and chiral symmetry arguments do not leave much room for further modifications . in order to calculate an observable in the @xmath0qsm , we consider a correlation function : @xmath26 at large euclidean time @xmath27 .
@xmath28 stands for the corresponding spin operator , while @xmath29 is the corresponding flavor operator .
the @xmath30 denotes the nucleon current consisting of @xmath3 quark fields .
the corresponding matrix element can be represented by the euclidean functional integral : @xmath31}. \label{eq : ev } \end{aligned}\ ] ] with the quark fields being integrated out , eq.([eq : ev ] ) can be divided into two separate contributions : @xmath32 schematically , the valence and sea contributions are shown in fig.1 . the integral over chiral bosonic fields @xmath33 can be carried out by the saddle point method in the large @xmath3 limit with the following ansatz : @xmath34 where @xmath35 is the su(2 ) chiral background field @xmath36 } .
\label{eq : profile}\ ] ] @xmath37 denotes the profile function satisfying the boundary condition @xmath38 and @xmath39 .
since the angular velocity @xmath40 ( @xmath41 ) and the strange quark mass @xmath42 are regarded as small parameters in our model , we expand the propagator @xmath43 with respect to the @xmath40 and @xmath42 up to the first order : @xmath44 figure 2 shows the rotational @xmath45 corrections and @xmath42 ones diagrammatically .
the rotational @xmath45 corrections are of particular importance in the present model .
the time - ordering of collective operators has to be taken into account in the quantization , since they do not not commute in principle .
the rotational @xmath45 corrections made it possible to solve the long - standing problem of underestimation of axial charges and magnetic moments in the @xmath0qsm @xcite .
in addition , there have been arguments that the rotational @xmath45 corrections violate charge conjugation @xcite
. however , christov _ et al . _
@xcite and wakamatsu @xcite proved that the rotational @xmath45 corrections possess correct properties under charge conjugation .
moreover , it was shown by praszaowicz _
et al . _
@xcite that with the rotational @xmath45 corrections considered the @xmath0qsm reproduces the result of the axial charge in the nonrelativistic quark model , _
i.e. _ @xmath46 , while it gives @xmath47 without them .
the collective su(3 ) hamiltonian is no longer su(3 ) symmetric when the @xmath42 corrections are taken into account .
hence , the eigenstates of the hamiltonian are neither in a pure octet nor in a pure decuplet but in mixed state with higher representations . dealing with the @xmath42 as a perturbation
, we can obtain the mixed su(3 ) baryon states : @xmath48 the coefficients @xmath49 and @xmath50 can be found elsewhere ( see for example the recent review @xcite ) .
i want to remark the parameters of the model for the numerical calculation , before i start to discuss the results .
the present su(3 ) @xmath0qsm contains four free parameters .
two of them are fixed in the meson sector by adjusting them to the pion mass , @xmath51 , the pion decay constant , @xmath52 , and the kaon mass , @xmath53 . as for the fourth parameter , _
i.e. _ the constituent mass @xmath12 of up and down quarks , values around @xmath54 have been used because they have turned out to be the most appropriate one for the description of nucleon mass splittings and other observables of baryons ( see ref .
hence , we fix the parameter @xmath12 to 420 mev . for the description of the baryon sector , the method of ref.@xcite
is chosen , modified for a finite meson mass .
the resulting strange current quark mass comes out around @xmath55 . with this set of fixed parameters ,
all the results which will be presented from now on have been calculated .
setting @xmath56 and @xmath57 , we can calculate the electromagnetic form factors @xcite .
figure 3 shows the electric form factors of the nucleon . in the case of the proton
, the result agrees well with the empirical data @xcite . as for the neutron , the result seems to be smaller than the empirical data by platchkov @xcite .
however , compared to the recent experiment conducted in mainz @xcite , the su(3 ) result is in good agreement with it .
the electric charge radii of the proton and the neutron are @xmath58 and @xmath59 , respectively .
the corresponding experimental data are @xmath60 and @xmath61 @xcite . in dotted curves in fig .
1 , the prediction of the su(2 ) model is shown .
as for the electric form factor of the proton , it is comparable to the su(2 ) , whereas a great discrepancy is observed in the case of that of the neutron .
it is partly because of the absence of @xmath42 and terms appearing only in su(3 ) and partly because of the different expectation values of the collective operators .
figure 4 displays the magnetic form factors of the nucleon . as we can see , the momentum dependence of the magnetic form factors are well reproduced , compared to the empirical data .
the @xmath42 corrections enhance the magnetic form factors about @xmath62 in the case of the neutron , which is not negligible to improve the prediction .
the magnetic moments of the proton and the neutron are , respectively , @xmath63 and @xmath64 , while experimental data are @xmath65 and @xmath66 .
the magnetic charge radii of the proton and the neutron are obtained as follows : @xmath67 @xmath68 and @xmath69 @xmath68 .
the corresponding experimental data are @xmath70 @xmath68 and @xmath71 @xmath68 . on the whole
, they are in remarkable agreement with the experimental data within around @xmath72 . in fact , one can show that any model with _ hedgehog _
symmetry can not reproduce the experimental data of baryonic magnetic moments better than the error of @xmath73 @xcite . in that sense
, the @xmath0qsm lies in the upper limit of accuracy which can be attained in any hedgehog model in the case of the magnetic properties .
it is of great interest to study the scalar form factor of the nucleon @xcite , since it provides also a clue of strangeness in the nucleon .
the analysis of the @xmath74 term the momentum dependence of the scalar form factor was carried out by gasser , leutwyler and sainio @xcite .
the results of ref .
@xcite are summarized as @xmath75 mev and @xmath76 mev , and @xmath77 which means a share of @xmath78 in the @xmath74 term .
the results indicate that the strangeness content of the nucleon in the scalar channel is not negligible , while a couple of recent theoretical works insist that there is no need to introduce a portion of strange quarks to explain the @xmath79 term @xcite . however , though it might be small , it is still important to consider the contribution of the strange quarks to the @xmath79 term in accordance with the recent experimental indication that strange quarks might play an important role of explaining the properties of the nucleon .
figure 5 draws the scalar form factor of the nucleon .
the error bar presented in fig.5 stands for the empirical analysis due to gasser , leutwyler and sainio .
as shown in fig .
5 our theoretical prediction is in good agreement with ref .
the @xmath42 corrections seem to be negligible in the scalar form factor . however , they play an important role of suppressing the strangeness contribution : the value of @xmath80 with the @xmath42 corrections is 0.27 while @xmath81 without the @xmath42 corrections .
it implies that though the @xmath42 corrections have a tiny effect on the magnitude of the scalar form factor , it leads to a large suppression of the @xmath80.the difference @xmath82 we have obtained is @xmath83 mev .
this value is very close to what ref.@xcite extracted , _
i.e. _ @xmath84 mev .
the prediction of the @xmath0qsm for the scalar radius @xmath85 is @xmath86 which is also comparable to the empirical value obtained by ref .
@xcite @xmath87 .
the strangeness content of the nucleon in the vector channel is one of hot issues these days .
while a great deal of theoretical works pile up , there is no clear theoretical consensus and no evident experimental judgment yet . encouraged by successful results of the scalar , electromagnetic and axial properties in the su(3 ) @xmath0qsm , it is of great interest to investigate the strange vector form factors in the same framework @xcite .
however , before we pursue the study of the strange vector form factors , we need to take into account the kaonic effect properly in line with the recent theoretical calculations incorporating the kaonic loops @xcite . from this theoretical point of view
, the strangeness in the nucleon can be interpreted in terms of the @xmath88 or @xmath89 components .
figure 6 displays a schematic diagram of explaining how the strangeness arises in the nucleon .
the diagram shown in the left - hand side of fig .
6 can be redrawn in terms of quarks .
in fact , the nucleon is known to consist of three valence quarks , _ i.e. _ @xmath90 . when one of the @xmath91 quarks is hit by the external strange vector current as shown in fig .
6 , the @xmath92 pair is created and rearranged so that we may have quark compositions @xmath93 and @xmath94 .
they correspond respectively to @xmath95 and @xmath96 .
this explains that we have the contribution from the valence part as well as the sea part to the strange vector form factors , though the nucleon itself does not include any strange valence quark .
as is explained above , kaons play a dominant role in describing the strange vector form factors .
the @xmath0qsm discussed up to now does not include the kaon cloud which is believed to be of little importance in the former calculation except for the neutron electric form factor @xcite .
however , to treat the strange vector form factors , we have to take the effect of the kaon cloud into account properly . to do so
, we have incorporated the kaonic tails selfconsistently in the profile @xmath37 in eq.([eq : profile ] ) at the expense of the pion tails .
figure 7 shows the strange vector form factors .
the effect of the kaon cloud is in particular prominent in the case of the strange electric form factor . as shown in fig .
7 , the replacement of the pion cloud by the kaon one brings about a sizable decrease of the strange electric form factor almost by a factor of three .
this remarkable result is in line with the recent investigation of the kaonic effects on the neutron electric form factor @xcite .
such a drastic reduction of the strange electric form factor can be easily understood explicitly by evaluating the strange electric radii . in any hedgehog model
the strange radii depend on the inverse of the meson mass @xmath97 which suppresses the tail of the profile , _
i.e. _ @xmath98 . from such a behavior of the @xmath99
, we can derive the relation @xmath100 eq.([eq : radius ] ) explains the decrease of the @xmath101 with @xmath102 . though the strange magnetic form factor is not changed as much as the strange electric one , the effect of the kaon cloud is still noticeable .
in contrast to the strange electric form factor , the kaon cloud enhances the magnitude of the strange magnetic one almost @xmath103 . as a result , by replacing the pion cloud by the kaon one , the strange magnetic moment @xmath104 is brought from @xmath105 @xmath106 to @xmath107 @xmath106 .
in addition , i want to mention a preliminary experimental result of the strange magnetic moment announced by mckeown @xcite in this workshop .
surprisingly , his finding is _
positive_. it implies that it conflicts with almost all of the theoretical models including the present model .
if the experimental result turns out to be correct , the present simple picture will not be enough to describe the strange vector form factors .
a more sophisticated and higher order corrections should be considered .
however , we still anticipate more compiled experimental data to enlighten us on it .
finally , i want shortly to present a recent calculation of the tensor charges of the nucleon @xcite .
the tensor charge @xmath108 is related to the transverse quark distribution @xmath109 .
however , @xmath110 is not measurable in inclusive deep - inelastic scattering .
that is the reason why it has not been extensively studied for long .
in fact , ralston and soper @xcite proposed the @xmath110 which can be measured in polarized drell - yan processes almost 20 years ago .
recently , it was also suggested that the @xmath110 can be measured in other exclusive hard reactions @xcite .
jaffe and ji demonstrated that the first moment of the @xmath110 is related to the tensor charge of the nucleon : @xmath111 where @xmath112 is an antiquark transversity distribution .
setting @xmath113 and @xmath114 in eq.([eq : corr ] ) , we can evaluate the tensor charges @xmath115 in the @xmath0qsm , which are linearly related to @xmath108 .
note that the tensor charges depend on the renormalization scale , though their dependence on it is very weak .
the normalization point pertinent to the @xmath0qsm is not determined from the first principle . however , as mentioned in section i , the renormalization point in the @xmath0qsm has be chosen by @xmath116 mev , but there may be a factor of order unity .
in tables i - ii the results of the tensor charges are summarized . as shown in table i the rotational @xmath45 corrections are of great significance numerically , while the @xmath42 corrections are relatively small .
in contrast to the axial charges , the tensor ones in the @xmath0qsm are closer to their values in the nonrelativistic constituent quark model and in particular the strangeness contribution to the tensor charge @xmath117 is compatible with zero , while that to the axial ones @xmath118 in the @xmath0qsm is negative and distinctive from zero @xcite .
this difference between the tensor charges and the axial ones can be explained by calculating them in the gradient expansion and scrutinizing their dependence on the size of the soliton @xcite .
my aim in this talk has been to review the recent investigation on various form factors of the nucleon in the su(3 ) chiral quark - soliton model(@xmath0qsm ) .
the rotational @xmath45 and linear @xmath42 corrections were taken into account .
the only parameter we have in the model is the constituent quark mass @xmath12 which is fixed to 420 mev by the mass splitting of the su(3 ) baryon octet and decuplet .
the results for the electromagnetic and scalar form factors are in remarkable agreement with experimental and empirical data within about 15 @xmath119 .
the strange vector form factors were also discussed .
the effect of kaon cloud turns out to be very large . by taking into account this kaonic effect ,
the strange charge radius is reduced to be almost 3 times smaller than that without the kaon cloud , while the magnitude of the strange magnetic moment is increased .
the tensor charges of the nucleon are presented in the same framework .
the @xmath0qsm predicts the number of the transversely polarized strange quarks @xmath117 in the transversely polarized nucleon compatible with zero , whereas it yields the negative nonzero number of the polarized strange quarks @xmath118 in the longitudinally polarized nucleon , which is consistent with the corresponding experimental value .
the works presented in this talk were done in collaboration with a. blotz , m.v .
polyakov , c. schneider , t. watabe and k. goeke .
the work is supported by the bmbf , the dfg , and the cosy project ( jlich ) .
i am grateful to o. benhar and a. fabrocini for the warm hospitality in elba .
s. kahana , g. ripka and v. soni , _ nucl .
phys . _ * a415 * ( 1984 ) 351 .
birse and m.k .
banerjee , _ phys .
lett . _ * 136b * ( 1984 ) .
d. diakonov and v. petrov , _ nucl .
* b272 * ( 1986 ) 457 . y. nambu and g. jona - lasinio , _ phys .
_ * 122 * ( 1961 ) 345 ; _ phys . rev._*124 * ( 1961 ) 246 .
d. diakonov , v. petrov and p. pobylitsa , _ nucl .
phys . _ * b272 * ( 1988 ) 809 .
h. reinhardt and r. wnsch , _ phys .
_ * 215b * ( 1988 ) 577 . th .
meiner , f. grmmer and k. goeke , _ phys .
lett . _ * 227b * ( 1989 ) 296 . m. wakamatsu and h. yoshiki , _ nucl .
phys . _ * a524 * ( 1991 ) 561 .
adkins , c.r .
nappi and e. witten , _ nucl .
phys . _ * b228 * ( 1983 ) 552 .
a. blotz , d. diakonov , k. goeke , n.w .
park , v. petrov and p.v .
pobylitsa , _ phys .
lett . _ * b287 * ( 1992 ) 29 .
kim , m.v .
polyakov , a. blotz and k. goeke , _ nucl .
phys . _ * a598 * ( 1996 ) 379 . m. wakamatsu and t. watabe , _ phys .
lett . _ * 312b * ( 1993 ) 184 .
christov , a. blotz , k. goeke , p. pobylitsa , v. petrov , m. wakamatsu , and t. watabe , _ phys .
_ * 325b * ( 1994 ) 467 .
christov , a.z .
g ' orski , k. goeke and p.v .
pobylitsa , _ nucl .
phys . _ * a592 * ( 1995 ) 513 .
kim , a. blotz , m.v .
polyakov , and k. goeke , _ phys .
rev . _ * d53 * ( 1996 ) 4013 .
christov , a. blotz , h .- c .
pobylitsa , t. watabe , th .
meissner , e. ruiz arriola and k. goeke , _ prog .
* 37 * ( 1996 ) 91 .
j.schechter and h. weigel , _ mod .
lett . _ * a10 * ( 1995 ) 885 ; _ phys .
* d51 * ( 1995 ) 6296 .
christov , k. goeke , and p.v .
pobylitsa , _ phys .
* c52 * ( 1995 ) 425 .
m. wakamatsu , _ phys .
lett . _ * 349b * ( 1995 ) 204 .
m. praszalowicz , a. blotz and k. goeke , _ phys .
lett . _ * 354 b * ( 1995 ) 415 . g. h " ohler _ et al .
_ , _ nucl .
_ * b114 * ( 1976 ) 505 .
s. platchkov _
et al . _ , _ nucl .
phys . _ * a510 * ( 1990 ) 740
. m. meyerhoff _
_ , _ phys .
* b327 * ( 1994 ) 201 .
s. kopecky , p. riehs , j.a .
harvey and n.w .
hill , _ phys .
* 74 * ( 1995 ) 2427 .
bruins _ et al .
* 75 * ( 1995 ) 21 .
kim , a. blotz , c. schneider , and k. goeke , _ nucl .
phys . _ * a596 * ( 1996 ) 415 . j. gasser , h. leutwyler , and m.e .
sainio , _ phys .
lett . _ * 253b * ( 1991 ) 252 ; _ phys .
lett . _ * 253b * ( 1991 ) 260 .
bass , _ phys .
_ * 329b * ( 1994 ) 358 .
ball , s. forte , and j. tigg , _ nucl phys . _ * b428 * ( 1994 ) 485 .
kim , t. watabe , and k. goeke , rub - tpii-11/95 [ hep - ph/9606440 ] ( 1996 ) .
musolf and m. burkardt , _ z. phys . _ * c61 * ( 1994 ) 433 .
t. watabe , h .- c .
kim , and k. goeke , rub - tpii-17/95 ( 1995 ) .
mckeown , map-202 [ hep - ph/9607340 ] and see also the present proceedings .
kim , m.v .
polyakov , and k. goeke , _ phys .
* b * in print ( 1996 ) .
j. ralston and d. soper , _ nucl .
_ * b152 * , 109 ( 1979 ) .
jaffe and x. ji , _ phys .
lett . _ * 71 * , 2547 ( 1993 ) .
j. collins , _ nucl .
phys . _ * b394 * ( 1993 ) 169 .
c. bourrely and j. soffer , _ nucl .
_ * b243 * , 329 ( 1994 ) ; _ nucl .
phys . _ * b445 * , 341 ( 1995 ) .
a. blotz , m. praszaowicz , and k. goeke , _ phys . rev . _ * d53 * ( 1996 ) 485 .
kim , m.v .
polyakov , and k. goeke , _ phys . rev . _
* d53 * ( 1996 ) r4715 .
* figure 1 * : the valence and sea contributions to an observable .
the left panel draws a schematic diagram for the valence contribution , while the right one is for the sea contribution .
the solid lines denote the @xmath3 valence quarks , whereas the loops designate the polarized dirac sea .
the wiggled line stands for the corresponding external current @xmath121 to the observable @xmath122 .
* figure 2 * : the rotational @xmath45 and @xmath42 corrections .
the upper panel shows the rotational @xmath45 corrections , while the lower panel is for the @xmath42 corrections .
the solid lines denote the @xmath3 valence quarks , whereas the loops designate the polarized dirac sea .
the wiggled line stands for the corresponding external current @xmath121 to the observable @xmath122 .
* figure 3 * : the electric form factors of the proton and the neutron .
the left panel displays that of the proton , while the right panel is for the neutron .
the solid curve corresponds to the strange quark mass @xmath123 , while the dashed curve draws that without @xmath42 .
the dotted curve shows the case of the su(2 ) model .
@xmath124 is chosen for the constituent quark mass .
the empirical data for the proton are taken from hhler _ et al . _
@xcite while for the neutron they are taken from platchkov @xcite .
* figure 4 * : the magnetic form factors of the proton and the neutron .
the left panel displays that of the proton , while the right one is for the neutron .
the solid curve corresponds to the strange quark mass @xmath123 , while the dashed curve draws that without @xmath42 .
the dotted curve shows the case of the su(2 ) model .
@xmath124 is chosen for the constituent quark mass .
the empirical data are taken from hhler _ et al . _
the experimental data for the neutron magnetic form factor ( open triangle ) are due to bruins _ et al . _
@xcite .
* figure 5 * : the scalar form factor of the nucleon @xmath125 .
the solid curve displays the case of @xmath126 mev , while the dashed curve shows that without @xmath42 .
the error bar denotes the empirical value from gasser _ et al . _
@xcite .
* figure 6 * : a schematic diagram of the kaon effect on the strange vector form factors .
the left panel displays the @xmath127 component of the strange vector form factors , while the right one shows the same component in a quark language .
* figure 7 * : the strange vector form factors of the nucleon .
the left one displays the strange electric form factor , while the right panel draws the strange magnetic form factor .
the solid curve corresponds to the @xmath128 , while dashed curve draws @xmath129 .
the constituent quark mass @xmath12 and @xmath42 are 420 mev and 180 mev , respectively . | the recent investigation on various form factors of the nucleon is reviewed in the framework of the su(3 ) chiral quark - soliton model .
the results for the electromagnetic and scalar form factors are in remarkable agreement with experimental and empirical data .
the strange vector form factors are also discussed with the effect of the kaon cloud being considered .
in addition to the form factors , the recent calculation of the tensor charges is presented . |
to write a field theoretical model which has nonzero values for the coupling constants at zeroes of the beta function of the renormalization group is an endeavor which is still continuing in particle physics .
the @xmath0 theory is a `` laboratory '' where different methods in quantum field theory are first applied .
after it was shown that this model became a trivial theory when the cutoff was removed , @xcite , it was clear that analyzing the terms in the perturbation series was not sufficient to decide whether one had a truly interacting theory .
work in this field was also given by wilson and others , @xcite .
renormalization group methods , first introduced by wilson for this purpose @xcite , are the most commonly used technique in studying whether one has a trivial theory or not . since
a nontrivial fixed point is not yet found for qcd , there are attempts to study alternative models for this purpose is given in @xcite .
a very popular model is the nambu - jona - lasinio model , hereafter njl , @xcite .
this model is written in terms of spinor fields only , and is used as an effective theory extensively in high energy physics @xcite .
the njl model was also shown to be trivial @xcite .
recent attempts to gauge this model to obtain a nontrivial theory are given in references @xcite .
both functional and diagram summing methods were used in these papers .
exact renormalization group methods proposed by wilson and polchinski , @xcite , are often employed for this purpose . a very recent paper on this method is given by sonoda @xcite .
another model , which uses only spinors is the g " ursey model @xcite .
we have worked on different forms of the g " ursey model @xcite .
our starting point was both our earlier work , @xcite , where g " ursey model lagrangian was attempted to be written in a polynomial form , and recent work , @xcite , which suggested that the gauged form of the njl model can be interpreted as a nontrivial theory .
in @xcite we reinterpreted our earlier work , @xcite , and showed that rather than finding a trivial theory , as claimed in @xcite , we ended up in a model where composite particles took part in physical processes .
the constituent fields , however , did not interact with each other when perturbation theory was applied to the model , as already shown in @xcite . in @xcite , we showed that , when this model is coupled to a constituent @xmath1 gauge field , we were mimicking a gauge higgs - yukawa ( ghy ) system , which had the known problems of the landau pole , with all of its connotations of triviality .
the essential point of our earlier work was the fact that the propagator of the composite scalar field was equal to @xmath2 .
since @xmath3 goes to zero as the cutoff is removed , many of the diagrams , where the scalar field propagator takes part as an internal line , become convergent .
we could show that there was no breaking of the chiral symmetry , thus no mass generation , for the fermion fields in our model in higher orders of perturbation theory . here
we will study our original model @xcite , coupled to a @xmath4 gauge field and use solely renormalization group techniques .
we start with the description of our starting model without the gauge field @xcite .
then we derive the renormalization group equations ( rges ) in one loop , and try to derive the criteria for obtaining non - trivial fixed points for the coupling constants of the theory . here
we closely follow the line of discussion followed in our reference @xcite . in our model
, however , there is a composite scalar field with a propagator completely different from a constituent scalar field used in this reference .
this gives rise to rge s in our case which are different from those given by harada et al .
since our starting models are different the motivation of our work is different from that of this reference .
we show that the renormalization group equations point to the non - triviality of the model when it is coupled to an @xmath4 gauge field .
we end up with a few remarks in the last section .
our initial model is given by the lagrangian @xmath5 here the only terms with kinetic part are the spinors .
@xmath6 is a lagrange multiplier field , @xmath7 is a scalar field with no kinetic part , @xmath8 and @xmath9 are coupling constants .
this expression contains two constraint equations , obtained from writing the euler - lagrange equations for the @xmath6 and @xmath10 fields .
hence , it should be quantized by using the dirac constraint analysis as performed in reference @xcite .
the lagrangian given above is just an attempt in writing the original g " ursey lagrangian @xmath11 in a polynomial form .
we already showed how the @xmath12 invariance of the grsey lagrangian , which prevents the fermion field from acquiring a finite mass in higher orders , is retained in our model , and the fact that our model is equivalent to the original grsey model only classically in @xcite .
to quantize the latter system consistently we proceed via the path integral method .
this procedure is carried out in reference @xcite . at the end of these calculations
we find out that we can write the constrained lagrangian given in equation ( [ cl ] ) as @xmath13\psi-{{a}\over{4}}(\phi^{4}+2\phi^{3}\xi-2\phi\xi^{3}- \xi^{4})+{i\over{4}}c^*(\phi^{2}+2\phi\xi+\xi^{2 } ) c , \label{eff1}\ ] ] where the effective lagrangian is expressed in terms of scalar fields @xmath14 , and @xmath15 , ghost fields @xmath16 , @xmath17 and spinor fields only .
the fermion propagator is the usual dirac propagator in lowest order , as can be seen from the lagrangian . after integrating over the fermion fields in the path integral
, we obtain the effective action .
the second derivative of the effective action with respect to the @xmath14 field gives us the induced inverse propagator for the @xmath14 field , with the infinite part given as @xmath18= { { g^2 q^2}\over { 4\pi \epsilon}}.\ ] ] here dimensional regularization is used for the momentum integral and @xmath19 .
we see that the @xmath14 field propagates as a massless field .
when we study the propagators for the other fields , we see that no linear or quadratic term in @xmath15 exists , so the one loop contribution to the @xmath15 propagator is absent . similarly the mixed derivatives of the effective action with respect to @xmath15 and @xmath14 are zero at one loop , so no mixing between these two fields occurs .
we can also set the propagators of the ghost fields to zero , since they give no contribution in the one loop approximation .
the higher loop contributions are absent for these fields . in reference @xcite we also studied the contributions to the fermion propagator at higher orders and we found , by studying the dyson - schwinger equations for the two point function , that there were no new contributions .
we had at least one phase where the mass of the spinor field was zero .
here we couple an @xmath20 gauge field to the model .
we also take spinors with different flavors , up to @xmath21 .
the new lagrangian reads : @xmath22 . \label{newlag}\ ] ] upon performing constraint analysis similar to the one performed in @xcite , we see that we have to satisfy @xmath23 after calculating the constraint matrix , raising the result to the exponential by using ghost field , and performing the transformations @xmath24 and @xmath25 we get similar equations as given in equation ( [ eff1 ] ) .
we see that both the @xmath15 and the ghost fields coming from the compositeness constraint decouple from our model . at this point
we have to note that there are two kinds of ghost contributions in the new model .
the ghosts coming from the gauge condition on the vector field do not decouple , and contribute to the renormalization group equations in the usual way .
we impose these constraints on equation ( [ newlag ] ) . after these steps we start with the effective lagrangian given as @xmath26-\frac{a}{4}\phi^4+\sum_{i=1}^{n_{f}}\overline{\psi}_{i } id\!\!\!\!/\psi_{i}-\sum_{i=1}^{n_{f}}g\phi\overline{\psi}_{i}\psi_{i } + l_{\mbox{ghost}}+ l_{\mbox{gauge fixing}}.\ ] ] here
@xmath21 is the number of flavors .
the gauge field belongs to the adjoint representation of the color group @xmath27 where @xmath28 is the color covariant derivative .
@xmath29 are the yukawa , quartic scalar and gauge coupling constants , respectively .
we take @xmath30 in the same order as @xmath31 . in the one loop approximation
, the renormalization group equations read as 16 ^ 2e(t ) & = & -be^3(t),[yenimodelerge ] + 16 ^ 2g(t ) & = & -cg(t)e^2(t),[yenimodelgrge ] + 16 ^ 2a(t ) & = & -ug^4(t ) , [ yenimodelarge ] where @xmath32,@xmath33 and @xmath34 are positive constants given as @xmath35 here @xmath36 is a second casimir , @xmath37 and @xmath38 is the fundamental representation with @xmath39 .
we take @xmath40 as a reference scale at low energies , @xmath41 , where @xmath42 is the renormalization point . in the rge
we see that the diagrams , where scalar propagators take part , are down by powers of @xmath3 .
hence we do not have contributions proportional to @xmath43 , @xmath44 and @xmath45 , as one would have in the ghy system as described in the work of @xcite .
since the diagrams , omitted in @xcite via a @xmath46 analysis , are down by an order of @xmath47 in our analysis , we do not need a relation between @xmath48 and the coupling constants at this point .
the solution for the first rg equation ( [ yenimodelerge ] ) can be obtained easily as e^2(t)=e_0 ^ 2(1+t)^-1 , [ e_nin_cozumu]where @xmath49 .
define ( t ) , where @xmath50 which is the initial value at the reference scale @xmath40 .
for the solution of the second rg equation ( [ yenimodelgrge ] ) we can define a rg invariant @xmath51 as h(t)=(c - b)^-1+(t ) .
since @xmath51 is a constant , we call it @xmath52 . then , the solution of the gauge coupling constant can be written as g^2(t)=^(t ) .
[ g_nin_cozumu]the solution of the last rg equation ( [ yenimodelarge ] ) can be defined by another rg invariant @xmath53 , given as k(t)=-u^-1+(t).we can then write a(t)=g^2(t ) [ a(t1)].here @xmath54 is the value of the rg invariant .
we can rewrite equation ( [ a(t1 ) ] ) as a(t)= [ a(t2)].when we check the ultraviolet limit now , we find ( t ) ll + 0 , & + ^(t ) ll + 0 , & + and ^-1+(t ) \ { lll + 0 , & + + 0 , & + + , & + .
we see that the constants @xmath52 and @xmath54 play important roles on the behavior of solutions of coupling equations ( [ e_nin_cozumu]),([g_nin_cozumu]),([a(t2 ) ] ) . for @xmath55
, @xmath52 should be positive ; for @xmath56 , @xmath52 should be negative to have the yukawa coupling take a real value .
this is necessary to have a unitary theory .
also for a region @xmath57 , with @xmath58 , the unitarity condition is satisfied for all coupling constants .
@xmath59 condition is also needed for stability of the vacuum . if @xmath60 , we get @xmath61 , which raises the problem of the vacuum instability .
next we study the different limits our parameters can take : we find e^2(t)&=&e^2_0 , + g^2(t)&=&exp(-t ) , + a(t)&=&.here @xmath64 and @xmath65 .
this means that when we set the @xmath32 term to zero , the yukawa running coupling constant decreases exponentially to zero . for this limit the gauge and the quadratic coupling constants go just to a constant . if @xmath33 approaches @xmath32 , the limit depends on the value of @xmath52 .
if @xmath52 is non zero , @xmath67 goes to zero .
if @xmath52 goes to zero as a constant times @xmath68 , i.e. @xmath69 , we find that @xmath67 and @xmath70 are both proportional to @xmath71 as follows g^2(t)=h_1 e^2(t ) , h_1>0 ; a(t)=. [ h1 ] when @xmath73 approaches @xmath32 , the behavior of @xmath70 changes .
if we set @xmath74 , then @xmath70 goes as @xmath75 a(t)=.this behaviour is not allowed since @xmath70 diverges as @xmath76 .
in this section we use the preceding results to investigate the non - triviality of the system with several criteria such as : all the running coupling constants : * should not diverge at finite t@xmath770 ( no landau poles ) ; * should not vanish identically ; * should not violate the consistency of the theory such as unitarity and/or vacuum stability . since the composite scalar field is the novel feature of our model , we will not consider the case when the scalar field is completely decoupled from the theory . we derive the expressions given below from the rge equations .
8 ^ 2=(b - c ) e^2(t ) , [ c1 ] 8 ^ 2=(2c - b ) e^2(t ) [ c2].for the fixed point solution , @xmath32 equals @xmath33 in equation ( [ c1 ] ) . for this value , there is a single solution which satisfies both equations ( [ c1 ] ) and ( [ c2 ] ) .
this solution is given as , = , where @xmath78 is a constant , and = .if we take @xmath79 approaching zero as @xmath33 approaches to @xmath32 , while @xmath80 in equation ( [ a(t2 ) ] ) , then we find g^2(t)=h_1e^2(t ) , a(t)=g^2(t ) .since @xmath81 and @xmath82 are constants , the behavior of the yukawa and quartic scalar couplings are completely determined by the gauge coupling .
this corresponds to `` coupling constant reduction '' in the sense of kubo , sibold and zimmermann @xcite . in the context of the rge
, it corresponds to the pendleton - ross fixed point @xcite . as seen from the previous sections the behavior of the yukawa coupling depends on whether @xmath55 or @xmath56 .
the point where @xmath83 needs a special care .
moreover the sign of the @xmath52 is important . in this case
@xmath52 should not equal to zero .
then we find in the uv limits g^2(t ) \ { ll + 0 , & + -0 , & + .
so the yukawa coupling is asymptotically free .
as it is seen , the sign of the rg invariant is important .
it should be chosen positive not to cause the violation of stability of the vacuum . in figure 1
we plot @xmath84 vs. @xmath85 for @xmath86 , @xmath87 .
both coupling constants approach the origin as @xmath63 goes to infinity .
thus , our model fulfills the condition required by the asymptotic freedom criterion . + in this case with a non zero value of @xmath52 g^2(t ) \ { ll -0 , & + + 0 , & + . for @xmath58 , our system satisfies the asymptotic freedom condition .
our system does not have a landau pole . in this respect
it differs from the ghy system @xcite .
as shown below , there is a restriction on the value of @xmath32 in this case .
this is the fixed point solution analyzed above .
g^2(t)=h_1e^2(t ) .
@xmath70 can be analyzed with four non trivial limits of the yukawa coupling . * @xmath55 with @xmath90 , * @xmath57 with @xmath58 , * @xmath91 with @xmath58 , * @xmath83 with @xmath92 .
for @xmath55 case , we should have @xmath90 , whereas in the @xmath57 case we have @xmath58 . in both cases
@xmath54 should be greater or equal to zero for the stability of the vacuum . in the third case , @xmath91 with @xmath58 , for all the real values of @xmath54 , @xmath70 diverges in the uv limit .
this means that there is no chance for a nontrivial theory in that region .
finally the @xmath83 case with @xmath92 have already be shown in equation ( [ h1 ] ) .
it is clear that in the uv limits @xmath54 should not take negative values . as seen above these constraints
give different relations between numbers of color and flavor .
note that in all the cases studied , if we take @xmath93 , one can deduce from equation ( [ a(t2 ) ] ) that @xmath70 can be made equal to zero for a finite value of @xmath63 , a situation which should not be allowed .
therefore , we can use only the option with @xmath59 .
the standard model with three colors and six flavors satisfies the @xmath55 case . for @xmath80 at the uv limit , the equation ( [ a(t2 ) ] ) a(t)= ^-1+(t ) + 0
, shows that the coupling constant is asymptotically free . also for a non zero @xmath54
, we find in the uv limit a(t ) .
then the sign of the @xmath54 is crucial for the stability of the vacuum .
although for @xmath94 we do not violate unitarity , we see that the asymptotic freedom criterion is not satisfied .
the requirement of this criterion fixes @xmath54 at the value zero . in figure 2
, we plot the rg flows in @xmath95 plane for different values of @xmath52 higher than zero while the gauge coupling @xmath96 is fixed to @xmath97 .
the origin is the limit where @xmath63 goes to infinity , there both coupling constants approach zero when @xmath80 .
here we write the @xmath4 gauge version of the polynomial lagrangian inspired by the g " ursey model . in @xcite
we had found an interacting model , where only the composites take part in scattering processes , if only perturbative calculations are done . gauging it with a constituent @xmath1 field resulted in a model which looked like the ghy system , with all the problems associated with the landau pole @xcite . when a @xmath4 gauge field is coupled , instead , we find that the renormalization group equations for the three coupling constants indicate that this model is nontrivial .
all the coupling constants go to zero asymptotically as the cutoff parameter goes to infinity , exhibiting the behavior dictated by asymptotic freedom .
* acknowledgement * : we thank ferhat takin for many discussions while preparing this manuscript .
this work is supported by the itu bap project no : 31595 .
the work of m.h . is supported by tuba , the academy of sciences of turkey .
this work is also supported by tubitak , the scientific and technological council of turkey . | we show that when a model , which is equivalent to the g " ursey model classically , is gauged with a su(n ) field , we get indications of a nontrivial field theory . |
chemical networks in interstellar clouds consist of gas - phase and grain - surface reactions @xcite .
reactions that take place on dust grains include the formation of molecular hydrogen @xcite as well as reaction networks producing ice mantles and various organic molecules . unlike gas phase reactions in cold clouds that mainly produce unsaturated molecules , surface processes are dominated by hydrogen - addition reactions that result in saturated , hydrogen - rich molecules , such as h@xmath0co , ch@xmath1oh , nh@xmath1 and ch@xmath2 .
in particular , recent experiments show that methanol can not be efficiently produced by gas phase reactions @xcite . on the other hand ,
there are indications that it can be efficiently produced on ice - coated grains @xcite .
therefore , the ability to perform simulations of the production of methanol and other complex molecules on grains is of great importance @xcite . unlike gas - phase reactions , simulated using rate equation models @xcite , grain - surface reactions require stochastic methods such as the master equation @xcite , or monte carlo ( mc ) simulations @xcite .
this is due to the fact that under interstellar conditions , of extremely low gas density and sub - micron grain sizes , surface reaction rates are dominated by fluctuations which can not be accounted for by rate equations @xcite .
a significant advantage of the master equation over mc simulations is that it consists of differential equations , which can be easily coupled to the rate equations of gas - phase chemistry .
furthermore , unlike mc simulations that require the accumulation of statistical information over long times , the master equation provides the probability distribution from which the reaction rates can be obtained directly .
however , the number of equations increases exponentially with the number of reactive species , making the simulation of complex networks infeasible @xcite .
the recently proposed multi - plane method dramatically reduces the number of equations , by breaking the network into a set of fully connected sub - networks @xcite , enabling the simulation of more complex networks .
however , the construction of the multi - plane equations for large networks turns out to be difficult . in this letter
we introduce a method based on moment equations which exhibits crucial advantages over the multi - plane method . the number of equations is further reduced to the smallest possible set of stochastic equations , including one equation for the population size of each reactive specie ( represented by a first moment ) and one equation for each reaction rate ( represented by a second moment ) .
thus , for typical sparse networks the complexity of the stochastic simulation becomes comparable to that of the rate equations .
unlike the master equation ( and the multi - plane method ) there is no need to adjust the cutoffs - the same set of equations applies under all physical conditions . unlike the multi - plane equations , the moment equations are linear and for steady state conditions
can be easily solved using algebraic methods . moreover , for any given network the moment equations can be easily constructed using a diagrammatic approach , which can be automated @xcite .
to demonstrate the method we consider a simple network , shown in fig . [ fig:1 ] , that involves three reactive species : h and o atoms and oh molecules @xcite . for simplicity
we denote the reactive species by @xmath3 h , @xmath4 o , @xmath5 oh , and the non - reactive product species by @xmath6 h@xmath0 , @xmath7 o@xmath0 , @xmath8 h@xmath0o .
the reactions that take place in this network include h + o @xmath9 oh ( @xmath10 ) , h + h @xmath9 h@xmath0 ( @xmath11 ) , o + o @xmath9 o@xmath0 ( @xmath12 ) , and h + oh @xmath9 h@xmath0o ( @xmath13 ) . consider a spherical grain of diameter @xmath14 , exposed to fluxes of h and o atoms and oh molecules .
the cross - section of the grain is @xmath15 and its surface area is @xmath16 .
the density of adsorption sites on the surface is denoted by @xmath17 ( sites @xmath18 ) .
thus , the number of adsorption sites on the grain is @xmath19 .
the desorption rates of atomic and molecular species from the grain are given by @xmath20 $ ] , where @xmath21 is the attempt rate ( standardly taken to be @xmath22 s@xmath23 ) , @xmath24 is the activation energy for desorption of specie @xmath25 and @xmath26 ( k ) is the grain temperature .
the hopping rate of adsorbed atoms between adjacent sites on the surface is @xmath27 $ ] , where @xmath28 is the activation energy for hopping of @xmath25 atoms ( or molecules ) .
here we assume that diffusion occurs only by thermal hopping , in agreement with experimental results @xcite .
for small grains it is convenient to define the scanning rate , @xmath29 , which is approximately the inverse of the time it takes an @xmath25 atom to scan the surface of the entire grain .
the master equation provides the time derivatives of the probabilities @xmath30 that on a random grain there will be @xmath31 adsorbed atoms / molecules of the reactive specie @xmath25 .
it takes the form @xmath32 \nonumber \\ & & + \sum_{i=1}^3 w_i \left [ ( n_i+1 ) p( .. ,n_i+1 , .. ) - n_i p(n_1,n_2,n_3 ) \right ] \nonumber \\ & & + \sum_{i=1}^2 a_i [ ( n_i+2)(n_i+1 ) p( .. ,n_i+2 , .. ) - n_i(n_i-1 ) p(n_1,n_2,n_3 ) ] \nonumber\\ & & + ( a_1+a_2 ) [ ( n_1 + 1)(n_2 + 1 ) p(n_1 +
1,n_2 + 1,n_3 - 1 ) - n_1 n_2 p(n_1,n_2,n_3 ) ] \nonumber \\ & & + ( a_1+a_3 ) [ ( n_1 + 1)(n_3 + 1 ) p(n_1 + 1,n_2,n_3 + 1 ) - n_1 n_3 p(n_1,n_2,n_3 ) ] .
\label{eq : master } \end{aligned}\ ] ] the terms in the first sum describe the incoming flux , where @xmath33 ( atoms s@xmath23 ) is the flux _ per grain _ of the specie @xmath25 .
the second sum describes the effect of desorption .
the third sum describes the effect of diffusion mediated reactions between two atoms of the same specie and the last two terms account for reactions between different species .
the rate of each reaction is proportional to the number of pairs of atoms / molecules of the two species involved , and to the sum of their scanning rates .
the moments of @xmath30 are given by @xmath34 , where @xmath35 are integers .
in particular , @xmath36 is the average population size of the specie @xmath25 on a grain .
the production rate per grain , @xmath37 ( molecules s@xmath23 ) , of @xmath38 molecules produced by the reaction @xmath39 is given by @xmath40 , or by @xmath41 in case that @xmath42 .
in numerical simulations the master equation must be truncated in order to keep the number of equations finite .
this can be done by setting upper cutoffs @xmath43 , @xmath44 on the population sizes , where @xmath45 is the number of reactive species .
however , the number of coupled equations , @xmath46 , grows exponentially with the number of reactive species .
this severely limits the applicability of the master equation to interstellar chemistry @xcite . to reduce the number of equations one tries to use the lowest possible cutoffs under the given conditions . in any case , to enable all reaction processes to take place , the cutoffs must satisfy @xmath47 for species that form homonuclear diatomic molecules ( h@xmath0,o@xmath0 , etc . ) and @xmath48 for other species .
the average population sizes of the reactive species and the reaction rates are completely determined by all the first moments and selected second moments of the distribution @xmath30 .
therefore , a closed set of equations for the time derivatives of these first and second moments could provide complete information on the population sizes and reaction rates . for the simple network considered here one needs equations for the time derivatives of the first moments @xmath49 , @xmath50 and @xmath51 and of the second moments @xmath52 , @xmath53 , @xmath54 and @xmath55 [ nodes and edges , respectively , in the graph shown in fig .
[ fig:1 ] ] .
such equations are obtained by taking the time derivative of each moment and using eq .
( [ eq : master ] ) to express the time derivatives of the probabilities @xcite . here
we show two of the resulting moment equations : @xmath56 in these equations , the time derivative of each moment is expressed as a linear combination of several other moments .
however , the right hand sides of these equations include third order moments for which we have no equations . in order to close the set of moment equations
we must express the third order moments in terms of first and second order moments @xcite .
this can be done by imposing the following constraint on the master equation : at any given time , at most two atoms or molecules can be adsorbed simultaneously on the surface .
furthermore , these two atoms or molecules must be from species that react with each other .
the resulting cutoffs allow only eight non - vanishing probabilities , namely , @xmath57 , @xmath58 , @xmath59 , @xmath60 , @xmath61 , @xmath62 , @xmath63 and @xmath64 .
the third moments in eq .
( [ eq : moment1 ] ) can now be expressed in terms of these non - vanishing probabilities , giving rise to the following rules : ( a ) @xmath65 ; ( b ) @xmath66 ; ( c ) @xmath67 . using these rules , which are general and apply to any network of binary reactions
, one can modify eqs .
( [ eq : moment1 ] ) , and obtain a closed set of the form @xmath68 where @xmath69 , @xmath70 and @xmath71 if @xmath42 and 0 otherwise .
this set includes one equation that accounts for the population size of each reactive species and one equation that accounts for the rate of each reaction . although these equations were derived using strict cutoffs , that are expected to apply only in the limit of very small grains and low flux , they provide accurate results for a very broad range of conditions .
the point is that once the set of moment equations is derived , the probabilities do not appear anymore , so the constraint is not explicitly enforced .
in fact , the equations maintain their accuracy even when the populations sizes of the reactive species are well beyond the constraints imposed above . in fig .
[ fig:1](a ) we present the population sizes of h ( @xmath72 ) and o ( @xmath73 ) atoms on a grain , vs. grain diameter , obtained from the moment equations . in fig .
[ fig:1](b ) we present the production rates of h@xmath0 ( squares ) , o@xmath0 ( triangles ) and h@xmath0o ( circles ) vs. grain diameter , obtained from the moment equations .
the results are in excellent agreement with the master equation ( solid lines ) . in the limit of large grains
they also coincide with the rate equations ( dashed lines ) .
note that the moment equations apply even when there are as many as 10 hydrogen atoms on a grain .
the parameters used in the simulations are @xmath74 ( sites @xmath18 ) , @xmath75 ( atoms s@xmath23 ) , @xmath76 and @xmath77 . the activation energies for diffusion and desorption were taken as @xmath78 , @xmath79 , @xmath80 , @xmath81 , @xmath82 and @xmath83 mev .
the grain temperature was @xmath84k .
the parameters used for hydrogen are the experimental results for low density amorphous ice @xcite . for the other species
there are no concrete experimental results , and the values reflect the tendency of heavier species to bind more strongly .
the fluxes and grain temperatures are suitable for dense molecular clouds .
consider the case in which a flux of co molecules is added to the network .
this gives rise to the network shown in fig .
[ fig:2 ] , which includes the following sequence of hydrogen addition reactions @xcite : h + co @xmath9 hco , h + hco @xmath9 h@xmath0co , h + h@xmath0co @xmath9 h@xmath1co and h + h@xmath1co @xmath9 ch@xmath1oh .
two other reactions that involve oxygen atoms also take place : o + co @xmath9 co@xmath0 and o + hco @xmath9 co@xmath0 + h. this network was studied before using the multiplane method , which required about a thousand equations compared to about a million equations in the master equation with similar cutoffs @xcite .
the moment equations include one equation for each node and one for each edge , namely , the network shown in fig . [ fig:2 ] requires only 17 equations .
we have performed extensive simulations of this network using the moment equations and found that they are in excellent agreement with the master equation . in fig .
[ fig:2](a ) we present the moment - equation results for the population sizes of h , o and co on a grain vs. grain diameter for the methanol network . in fig .
[ fig:2](b ) we present the moment - equation results for the production rates per grain of some of the final products of the network .
the results are in excellent agreement with the master equation ( solid lines ) , and coincide with the rate equations ( dashed lines ) for large grains .
the activation energies for diffusion and desorption of co ( @xmath85 ) , hco ( @xmath86 ) , h@xmath0co ( @xmath87 ) and h@xmath1co ( @xmath88 ) were taken as @xmath89 , @xmath90 , @xmath91 , @xmath92 , @xmath93 , @xmath94 , @xmath95 and @xmath96 mev .
the flux of co molecules was taken as @xmath97 .
note that experiments on co desorption from ice surfaces indicate that @xmath98 should be slightly higher than the value used here @xcite .
however , it turns out that using a higher value would compromise the feasibility of the master equation simulations , which are required here in order to establish the validity of the moment equation method .
when simulating highly complex networks one encounters the problem of obtaining the equations themselves .
the ability to automate the construction of the equations becomes important .
a crucial advantage of the moment equations is that they can be easily obtained using a diagrammatic approach .
a detailed presentation of the diagrammatic method will be given in barzel & biham ( 2007 ) .
in summary , we have introduced a method , based on moment equations , for the simulation of chemical networks taking place on dust - grain surfaces in interstellar clouds .
the method provides highly efficient simulations of complex reaction networks under the extreme conditions of low gas density and sub - micron grain sizes , in which the reaction rates are dominated by fluctuations and stochastic simulations are required .
the number of equations is reduced to one equation for each reactive specie and one equation for each reaction , which is the lowest possible number for such networks .
this method enables us to efficiently simulate networks of any required complexity without compromising the accuracy .
it thus becomes possible to incorporate the complete network of surface reactions into gas - grain models of interstellar chemistry . to fully utilize the potential of this method ,
further laboratory experiments are needed , that will provide the activation energy barriers for diffusion , desorption and reaction processes not only for hydrogen but for all the molecules involved in these networks . ) and o ( @xmath73 ) atoms on a grain vs. grain diameter , @xmath14 ( and the number of adsorption sites , @xmath99 ) , obtained from the moment equations for the network shown above , where nodes represent reactive species , edges represent reactions and the products are indicated near the edges ; ( b ) the production rates of h@xmath0 ( squares ) , o@xmath0 ( triangles ) and h@xmath0o ( circles ) on a grain vs. grain diameter , obtained from the moment equations .
the results are in excellent agreement with the master equation ( solid line ) . in the limit of large grains
they also coincide with the rate equations ( dashed line ) .
[ fig:1 ] , width=480 ] ) , o ( @xmath100 ) and co ( @xmath73 ) for the methanol network shown above , vs. grain diameter , @xmath14 ( and the number of adsorption sites , @xmath99 ) , obtained from the moment equations .
( b ) the production rates of several final products of the methanol network per grain vs. grain diameter . obtained from the moment equations .
the results are in excellent agreement with the master equation ( solid line ) . in the limit of large grains
they also coincide with the rate equations ( dashed line ) .
[ fig:2 ] , width=480 ] | networks of reactions on dust grain surfaces play a crucial role in the chemistry of interstellar clouds , leading to the formation of molecular hydrogen in diffuse clouds as well as various organic molecules in dense molecular clouds . due to the sub - micron size of the grains and the low flux
, the population of reactive species per grain may be very small and strongly fluctuating . under these conditions
rate equations fail and the simulation of surface - reaction networks requires stochastic methods such as the master equation .
however , the master equation becomes infeasible for complex networks because the number of equations proliferates exponentially .
here we introduce a method based on moment equations for the simulation of reaction networks on small grains .
the number of equations is reduced to just one equation per reactive specie and one equation per reaction .
nevertheless , the method provides accurate results , which are in excellent agreement with the master equation .
the method is demonstrated for the methanol network which has been recently shown to be of crucial importance . |
thick disks in lenticular and spiral galaxies are defined as disk - like components with a scaleheight larger than that of the thin or ` canonical ' disk .
they are typically detected as exponential excesses of light a few thin disk scaleheights above the galaxy mid - plane in edge - on galaxies .
they have been first detected by tsikoudi ( 1979 ) and recognized as a distinct galaxy structural component by burstein ( 1979 ) .
a thick disk component was soon found in the milky way ( gilmore & reid 1983 ) and it was afterwards proved to be made of old and metal - poor stars when compared to the thin disk stellar population ( reid & majewski 1993 ; chiba & beers 2000 ) . later studies found thick disks to be nearly ubiquitous ( yoachim & dalcanton 2006 ; comern et al .
2011a ) and to be systematically older than their thin counterparts ( yoachim & dalcanton 2008a ) .
the origin of thick disks is still a matter of debate , with four formation mechanisms being discussed ( see the introduction of yoachim & dalcanton 2008a and references therein ) .
the first two possibilities consider two different mechanisms to dynamically heat an originally thin disk , meaning that its vertical stellar velocity dispersion is increased . in the first possibility
this dynamical heating could be a consequence of internal evolution due to gravitational interaction with thin disk overdensities like giant molecular clouds or spiral arms ( villumsen 1985 ; hnninen & flynn 2002 ; schnrich & binney 2009 ; bournaud et al .
2009 ) . in the second the heating
is considered to be due to galaxy - galaxy or dark matter subhalo - galaxy interactions and mergers ( quinn et al .
1993 ; robin et al . 1996 ; walker et al .
1996 ; velzquez & white 1999 ; chen et al . 2001 ; benson et al . 2004 ;
hayashi & chiba 2006 ; kazantzidis et al . 2008 ; read et al .
2008 ; villalobos & helmi 2008 ; qu et al .
the third possibility is that the thick disk is a consequence of _ in situ _ star formation ( brook et al .
2004 ; robertson et al . 2006 ; elmegreen & elmegreen 2006 ; brook et al . 2007 ; richard et al .
2010 ) or star formation with a high initial velocity dispersion in very massive star clusters ( kroupa 2002 ) .
possibilities one and three are related if thick disk star formation occurs in gas that has a high turbulent velocity dispersion from its own clump stirring .
the fourth possibility is that the thick disk is formed by accretion of stars from disrupted small satellite galaxies after the initial build - up of the galaxy ( statler 1988 ; gilmore et al .
2002 ; abadi et al . 2003 ; navarro et al . 2004
; martin et al . 2004 ; read et al . 2008 ) .
another topic of debate is that of `` missing '' or `` lost '' baryons ( persic & salucci 1992 ) . the fraction of baryons detected at high redshift ( see , e.g. , wmap predictions by spergel et al .
2003 ) , does apparently not match the present - day fraction of baryons .
part of the missing baryons are expected to be found in hot gas in the warm - hot intergalactic medium ( whim ; see , e.g. , fukugita et al .
1998 , simulations from cen & ostriker and whim emission detection / constraining by hickox & markevitch 2007 and zappacosta et al .
additionally , dark matter haloes in which galaxies lie are also found to be very baryon deficient compared to the cosmic baryon fraction ( mcgaugh 2008 ) .
this apparent lack of baryons inside the galaxy potential well is what has been called the `` local missing baryon '' problem .
part of these baryons have probably escaped due to winds caused by supernova feedback , but the local baryon budget is far from being accounted for ( mcgaugh 2008 and references therein ) . [
cols="^ " , ] if we ignore galaxies hosting an x - shaped bulge , our fraction of galaxies in which the thick disk contains more mass than the thin disk is 8 out of 23 ( @xmath3 ) for @xmath4 , and 15 out of 23 ( @xmath5 ) for @xmath6 .
this remains in sharp contrast with the 20% found by yoachim & dalcanton ( 2006 ) .
one possibility is that this difference is due to the choice of the fitting function .
yoachim & dalcanton s ( 2006 ) results for each galaxy come from doing the median of the results of six fits with different mid - plane maskings , different smoothings to account for the seeing and different combinations of @xmath7 and @xmath8 accounting for the thin and the thick disk . of these six fitting models , three make use of a sum of two @xmath8 functions and are thus likely to dominate the median of the six models .
we tested whether fitting the sum of two @xmath8 functions gives more weight to the thin disk than the functions we used .
we produced @xmath8 fits over our luminosity profiles using the same procedure than when fitting the equilibrium solutions , except for the fact that , for simplicity , we have not used a psf convolution and we have not taken into account the effect of dust extinction which we show to be very small in most cases ( see table [ table2 ] ) .
we found that the shape of the thick disk close to the mid - plane appears rounded in the case of a @xmath8 fit , but it is significantly more peaked when making a fit using the equations of equilibrium ( see figure [ comparison ] for an example ) . from a linear regression crossing the origin ( figure [ sech2 ] )
, we found that @xmath9 , where @xmath10 denotes the fits using two @xmath8 functions , and @xmath11 stands for the fits using the equations of equilibrium .
this implies that , if we assume our approach to be accurate , and yoachim & dalcanton ( 2006 ) mass measurements to be dominated by the models with a sum of two @xmath8 functions , they might have underestimated @xmath12 by a factor of @xmath13 .
when multiplying by 1.7 the @xmath12 values in fig .
22 of yoachim & dalcanton ( 2006 ) , we find that 12 out 34 galaxies ( @xmath3 ) would have @xmath14 , a value close to that we obtain with @xmath4 .
the gradient in @xmath15 with @xmath16 discussed in section 3.3.4 is likely to have some effect in the trend seen in figure [ vmax ] .
as @xmath15 is larger for galaxies with a small @xmath16 , it would cause the trend to be strengthened . for further comparison with the results of yoachim & dalcanton ( 2006 )
we should decompose galaxies from their sample with our methodology using s@xmath2 g imaging .
there is a four galaxy overlap between the s@xmath2 g and their sample but , unfortunately , at the time of the sample selection of this paper ( december 2010 ) those galaxies have not gone through the automated mosaicking / cleaning process yet .
we plan to study this overlap in detail in a follow - up paper . in summary , we find that our thick to thin disk mass ratios ( @xmath12 ) are higher than those found in literature due to the large fraction of mass attributed to the thick disk at low @xmath17 when fitting luminosity profiles with the equations of coupled isothermal disks in equilibrium .
these results have been obtained from data which can reproduce yoachim & dalcanton s ( 2006 ) results when fitting a sum of two @xmath8 functions . as our model
is based in physically motivated functions and we are using images less affected by dust than in previous studies , our results are likely to be the most accurate so far for such a large sample .
we will now explore the implications of our results , in particular higher thick disk masses .
the main result of this paper is that the total stellar mass of thick disks is similar to that of thin disks and that in some cases , the thick disk mass can be up to 2 3 times larger than that in the thin disk . these surprisingly massive thick disks , together with the fact that we have found no significant flares , put some constraints on their formation mechanisms .
the first formation mechanism to be discussed is that of a dynamical heating of an originally thin disk by disk overdensities .
secular heating due to giant molecular clouds ( gmcs ) has sometimes been suggested as a reason of disk heating .
however , this mechanism has been recognized as insufficient to explain the highest velocity dispersion stars ( villumsen 1985 ; even a fourfold increase in the number of gmcs would not be enough to explain the presence of the hottest populations of stars according , to hnninen & flynn 2002 ) .
this mechanism is thus not likely to create very massive thick disks .
alternatively , the disk may have been mostly heated in a short interval of time at the beginning of the history of the galaxy , an epoch in which there is a large population of clumpy disks whose clumps formed through disk gravitational instabilities and had a mass of a few @xmath18 ( elmegreen & elmegreen 2005 ) .
these massive clumps have a stellar scattering power much larger than present - day gmcs .
bournaud et al .
( 2009 ) found that this mechanism plus further secular evolution could cause a thick to thin disk mass ratio @xmath19 in simulations designed to produce galaxies with milky way mass .
thus , bournaud s et al . (
2009 ) milky way - like simulated galaxies , although a little bit short in thick disk mass , would be compatible with the @xmath20 relationship for @xmath4 . however , as these simulations have only been run for milky way - like galaxies , the effect caused in a less massive galaxy in unknown .
thick disks created through star scattering and radial mixing by spiral arms has been discussed and modeled by schnrich & binney ( 2009 ) who also succeed to reproduce a reasonable thick disk ( although again light when put in the @xmath20 relationship due to the fact that they fitted the milky way which we have shown to have often been measured to be very thin disk - dominated ) .
another possibility is that the thick disk has been created by the heating of a thin disk by satellite galaxies or dark matter subhaloes .
this mechanism is more efficient for galaxies with orbits far from coplanar with the disk ( read et al .
according to simulations , these interactions may heat the disk to form a thick disk , but they are thought to also cause a significant flare whose exact properties are dependent on the encounter parameters ( quinn et al . 1993 ; walker et al .
1996 ; kazantzidis et al . 2008 ;
villalobos & helmi 2008 ; bournaud et al . 2009 ) .
most of these authors suggest the flare to be at galactocentric radii small enough to be detectable in some cases .
the exception is the work by villalobos & helmi ( 2008 ) , where the flare would only affect the thick disk at very large galactocentric radii , which would need deeper photometry than ours to be detected .
kazantzidis et al . (
2008 ) suggested that the flares in edge - on galaxies are what we interpret as truncations , as flares would be so substantial that the stars would be scattered over a large vertical range , causing the disk in those regions to have a very low surface brightness when seen edge - on .
this would also imply that truncations are more often detected in edge - on galaxies , which is likely to be the case ( see discussion in pohlen et al .
we have found that strong flares are not common among our galaxies , implying that thick disks are unlikely to form through the heating of a disk by satellites .
even if the flares were so strong to appear as truncations , the heating through encounters would fail to explain the existence of thick disks in some galaxies , since thick disks are ubiquitous and truncations , although very common , do not appear in all galaxies ( kregel et al .
2002 ; van der kruit 2007 ) .
minor mergers indeed occur in disk galaxies , but they are likely to have a small effect at high redshift , where they are more frequent , due to the presence of a high gas fraction in the disk ( moster et al . 2010 ; previous simulations usually do not consider gas effects ) .
the effect of gas for avoiding heating by interactions with satellite galaxies is still significant for galaxies with a gas fraction similar to that of present - day milky way , although it allows part of the thin disk to be heated into a flared thick disk ( qu et al .
2011 ) .
some authors theorize that a thick disk may be a consequence of an _ in situ _ star formation . according to the simulations of brook et al .
( 2004 ) , the thick disk star formation is triggered by the accretion of gas - rich protogalactic fragments during hierarchical clustering at high redshift .
the gas disk would still be dynamically hot at the moment of the thick disk creation , which would explain the difference in scaleheight between the thick disk which formed early and the thin disk formed from a cooled gas disk and from additional gas coming from cold flows .
brook et al .
( 2004 ) found that thick disk scalelengths are larger than those of thin disks , which is in contradiction with observations ( yoachim & dalcanton 2006 ) .
later models using disk - disk gas - rich mergers ( robertson et al . 2006 ; brook et al . 2007 ; richard et al .
2010 ) are able to produce reasonable thick disks from stars from the progenitor galaxies plus stars created during the merger , with the stars which are created later settling in a thin disk .
however , these simulations do not include posterior accretion of cold gas , which will settle in the thin disk , increase its mass , and reduce the scaleheight of the thick disk .
another _ in situ _ formation model is that from elmegreen & elmegreen ( 2006 ) , who study chain and spiral galaxies in the _ hubble space telescope _
ultra deep field ( udf ) and find that the scaleheight of those galaxies is similar to that of present - day thick disks , with a mass column density of 4 to 40 @xmath21 .
recent studies ( e.g. , moni bidin et al . 2010 ) found the milky way column mass density in the solar neighborhood to be @xmath22 , thus , if we assume udf thick disks to be progenitors of present - day thick disks , thick - to - thin disk column mass density fractions up to @xmath23 would be compatible with the udf thick disk mass estimates for milky way - like galaxies .
however , interpreting these galaxies as direct progenitors of thick disks for which a thin disk has been added after accreting gas through cold flows is problematic , as discussed in elmegreen & elmegreen ( 2006 ) as the udf thick disks are likely to shrink to lower scaleheights when further material is accreted , making them much thinner than observed present - day thick disks .
elmegreen & elmegreen ( 2006 ) calculated that in order to create a milky way - like thick disk with @xmath24pc , the equivalent udf thick disk should have @xmath25kpc , which is far larger than the observed average udf thick disk .
however , as we found that thick disks are much more massive than previously thought , the fraction of added gas needed to generate a thin disk is lower than that estimated by elmegreen & elmegreen ( 2006 ) , making the decrease in @xmath26 of the udf thick disks lower than what they estimated . in a @xmath27cdm scenario ,
at least a fraction of the thick disk stars is expected to be accreted from satellite galaxies dragged into the disk plane by dynamical friction .
this process would be effective for galaxies with an impact angle @xmath28 , and larger angle encounters produce spherical - like distributions as well as a thick disk by thin disk heating , as discussed before ( read et al .
2008 ; but again these simulations do not include gas and thus the effect of disk heating may be overestimated ) . however , according to these authors , this mechanism applied to the milky way would produce a disk @xmath29 10 times less massive than observed , and thus needs to be combined with other thick disk formation mechanisms .
scannapieco et al .
( 2011 ) found that their simulated thick disks , which have been defined to be made of stars older than 9gyr , have `` relatively high fractions '' of non - accreted stars although the exact value is not specified . a strong indication that this mechanism is likely to be at the origin of a significant fraction of material in at least some galaxies is the discovery of a thick disk with a significant fraction of counterrotating material by yoachim & dalcanton ( 2008b ) .
we now summarize our comments on the various proposed thick disk formations scenarios : * heating of a thin disk by its own overdensities , in the form of , e.g. , spiral arms , gmcs or giant clusters : this mechanism may have difficulties to produce massive enough thick disks .
* heating of a thin disk by satellite galaxies or dark matter subhaloes crossing it : not likely to cause a significant effect as it would cause noticeable flares .
its effect is highly reduced when the disk contains a significant fraction of gas . * _ in situ _ thick disk formation : this formation mechanism had the problem of the shrinking of the thick disk scaleheight due to the accretion of external material after the thick disk formation , which would make thick disks indistinguishable from their thin counterparts .
however , our results indicate that thick disks of many galaxies are much more massive than previously thought , lessening the impact of the newer material in the thick disk scaleheight .
* accretion of stars of satellite galaxies ( tidal stripping ) : this mechanism occurs necessarily for some galaxies , as indicated by their kinematical or morphological signatures .
however , simulations have shown that this effect can not explain all the mass found in thick disks .
thus , our results plus data from previous studies tend to favor an _ in situ _
origin for most of the stars in the thick disk , where the thin disk created afterwards is not massive enough to reduce @xmath26 by too much .
in addition the thick disk may contain a significant amount of stars coming from satellites accreted after the initial build - up of the galaxy and an extra fraction of stars coming from the secular heating of the thin disk .
thick disk - dominated galaxies may play some role in what has been called the `` missing baryon problem '' ( persic & salucci 1992 ) .
wmap cosmic microwave background observations have constrained the fraction of baryons in the universe ( spergel et al .
2003 ) , but most of this baryonic matter is apparently not found in the `` obvious '' reservoirs which are stars and cold gas in galaxies ( bell et al .
2003 ) . at a galactic scale ,
several studies indicate that observations are missing a large fraction of the baryons relative to the dark matter content in the milky way halo , and this problem appears to be even more severe for galaxies with smaller masses ( hoekstra et al .
2005 ; mcgaugh 2007 ; mcgaugh 2008 ; bregman 2009 ) .
rough counts by mcgaugh ( 2008 ) for the milky way show that we are missing at least @xmath30 , which is comparable or larger than the mass of all the baryonic mass thought to be contained in the disk .
two mechanisms have been proposed to explain the `` missing baryons '' in galaxy dark matter haloes .
the first is that there is a quantity hot gas in the halo which has either failed to cool down , or was blown away by supernova feedback .
this hot gas in the halo has been measured to have a mass of @xmath31 ( bregman & lloyd - davies 2007 ) , which falls way short of solving the `` local missing baryon '' problem .
the second possibility is that supernova feedback has blown away halo gas so efficiently that this material has been lost into the intergalactic medium .
the amount of baryons lost in this way is difficult to estimate but it is also apparently not enough to explain the low fraction of baryons in galaxies ( mcgaugh 2010 ) . even though they remain insufficient , both explanations have the merit of explaining satisfactorily why lower - mass galaxies have lower baryon fractions , as supernova feedback is more effective in shallower potential wells . for external galaxies ,
the stellar masses inside the disks are calculated by assuming a given mass - to - light ratio . however , ignoring the presence of an unexpectedly high fraction of stars in the thick disk may have led to an underestimate of the value of @xmath32 .
some numbers may be easily obtained from the data plotted in figure [ vmax ] .
let us consider the more thin disk - dominated galaxies ( @xmath33 ) for the case @xmath4 .
if we were ignoring a thick disk , and assign a thin disk @xmath32 for all the light emitted by the galaxy , we would be underestimating the stellar mass of the galaxy by @xmath34 . on the other hand ,
if we consider the most thick disk - dominated galaxy ( @xmath35 ) , and @xmath6 , we would underestimate the stellar disk mass by a factor of two .
these numbers are calculated for the 3.6@xmath36m - band , which implies that the underestimation of the disk stellar mass would be different when measured at shorter wavelengths .
we stress that these numbers only account for the baryons in stellar disks , and that if extended gas disks beyond the optical radius were included in the thin disk mass as done by yoachim & dalcanton ( 2006 ) @xmath12 would go down .
thus , our results suggest that a least a fraction of the `` local missing baryons '' are located in the thick disk component of galaxies and that they have been `` lost '' until now due to the underestimation of @xmath12 .
this explanation is complementary , and not an alternative , to those which have already been suggested to solve the problem .
the nature of the mechanisms creating the thick disks has been a topic of discussion for a long time .
basically four pictures have been suggested , each leaving some distinct observational signature .
the first one is that the thick disk is the consequence of the thin disk have been dynamically heated by its own overdensities .
the second possibility is that thin disks being dynamically heated by the disk crossing of satellite galaxies or dark matter subhaloes .
the third possibility is _ in situ _ star formation during or shortly after the build - up of the galaxy .
the fourth is that the thick disk could be made of stars which have been accreted from infalling satellites after the formation of the thin disk . in order to study the problem of thick disk formation we have used 3.6@xmath36 m images of edge - on galaxies from the s@xmath2 g .
the selection criterion used , maximum isophote ellipticity @xmath37 , has led to a sample of 46 late - type galaxies with types ranging from @xmath0 to @xmath1 .
we have fitted the luminosity profiles of these galaxies in radial bins with @xmath38 and @xmath39 , using solutions of the equations of equilibrium for two stellar and one gaseous coupled isothermal disks .
these functions have the advantage of being physically motivated .
the largest uncertainty of the fit comes from the assumed ratio of mass - to - light ratios of the thick and the thin disk , @xmath15 .
we studied cases with @xmath4 and @xmath6 in order to account for the range of fitted star formation histories for the solar neighborhood .
the low value of @xmath15 arises when the stars in thick disks form over a long period of time and the high value arises when the thick disk stars formed entirely in a burst in the early universe .
the studied interval in @xmath15 is in broad agreement with that derived from the study by yoachim & dalcanton ( 2006 ) .
we found that the stellar mass of the thick disk component ranges from one third to two times that of the thin disk if @xmath4 and from two thirds to more than three times if @xmath6 .
we thus found that thick disks are much more massive than generally assumed so far . in order to check the reasons of this difference
, we compared our results in detail with those from yoachim & dalcanton ( 2006 ) .
we found that the choice of the fitting function is responsible for the high fraction of stellar light we detect in the thick disk .
we also found that disks do not flare significantly within the range of galactocentric radii we have studied , and that the ratio of thick to thin disk scaleheights is higher for earlier - type galaxies in our sample .
the lack of flares discards the kinematical heating of a thin disk by satellite galaxies or dark matter subhaloes as the main thick disk formation mechanism . as a result of these findings
we suggest the following tentative formation mechanism for thick disks .
first , a thick stellar disk is created during and soon after the build - up of the galaxy by the stirring action from massive disk clumps and the instabilities that form those clumps . a thin disk subsequently forms from gas which has not initially been spent in stars , and from gas that arrives later at a lower rate so that , when combined with the hot disk and bulge that are already present , it is relatively stable and can not heat itself much .
the newly accreted gas makes the thick disk scaleheight , @xmath26 , shrink , and if this shrinkage is a relatively small amount , then the final thick disk remains a distinct component of the galaxy , with a younger thin disk inside of it . in the case of thick - disk dominated galaxies , the thick disk forms partly at high redshift by internal disk stirring , as above , and then increases its mass over a hubble time by the heating of the thin disk by internal disk substructure , such as large star clusters and giant molecular clouds , and the accretion of satellite galaxies . as our results suggest that thick disks have masses comparable to those of thin disks and in addition they have a large mass - to - light ratio due to the absence of young stars
, we suggest that ignoring them has led to an underestimate of the baryonic mass located in disk galaxies .
we thus suggest that thick disks contain part of what has been called the `` local missing baryons '' .
we have calculated this underestimate to be between 10% and 50% of the stellar mass of a disk , depending on the ratio of the thick versus thin disk masses and mass - to - light ratios , @xmath12 and @xmath15 .
the authors wish to thank the entire s@xmath2 g team for their efforts in this project .
this work is based on observations and archival data made with the spitzer space telescope , which is operated by the jet propulsion laboratory , california institute of technology under a contract with nasa .
we are grateful to the dedicated staff at the spitzer science center for their help and support in planning and execution of this exploration science program .
we also gratefully acknowledge support from nasa jpl / spitzer grant rsa 1374189 provided for the s@xmath2 g project .
ks , j - cmm , tkim and tmizusawa acknowledge support from the national radio astronomy observatory , which is a facility of the national science foundation operated under cooperative agreement by associated universities , inc .
ea and ab thank the centre national dtudes spatiales for financial support . this research has made use of observations made with the nasa / esa hubble space telescope and obtained from the hubble legacy archive , which is a collaboration between the space telescope science institute ( stsci / nasa ) , the space telescope european coordinating facility ( st - ecf / esa ) and the canadian astronomy data centre ( cadc / nrc / csa ) .
funding for the sdss and sdss - ii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , the u.s .
department of energy , the national aeronautics and space administration , the japanese monbukagakusho , the max planck society , and the higher education funding council for england .
the sdss web site is http://www.sdss.org/. this research has made use of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
this research has made use of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
pohlen , m. , beckman , j. e. , httemeister , s. , knapen , j. h. , erwin , p. , & dettmar , r .- j . , 2004b , in penetrating bars through masks of cosmic dust , ed .
d. l. block et al .
( cambridge : cambridge univ . press ) , 713 lcccccc|cccc|cccc|cccc|cccc & & & & & & & @xmath40 & v & v & @xmath41 & & 0.33&0.30 & & & 2.28 & 2.51 & & & 1.00 & 1.00 & + & & & & & & & @xmath40 & v & v & @xmath41 & & 0.33 & 0.29 & & & 2.19 & 2.37 & & & 1.00 & 1.00 & + & & & & & & & @xmath40 & v & v & @xmath41 & & 0.61 & 0.56 & & & 2.41 & 2.65 & & & 1.00 & 1.00 & + & & & & & & & @xmath40 & v & v & @xmath41 & & 0.61 & 0.54 & & & 2.32 & 2.51 & & & 1.00 & 1.00 & + & & & & & & & @xmath42 & @xmath40 & v & v & & & 0.50 & 0.60 & & & 2.57 & 2.30&&&1.00&1.00 + & & & & & & & @xmath42 & @xmath40 & v & v&&&0.51&0.61&&&2.47&2.19&&&1.00&1.00 + & & & & & & & @xmath42 & @xmath40 & v & v&&&0.96 & 1.08 & & & 2.76 & 2.41 & & & 1.00 & 1.00 + & & & & & & & @xmath42 & @xmath40 & v & v&&&0.91&1.11&&&2.59&2.32&&&1.00&1.00 + & & & & & & & @xmath41 & v & * & @xmath41 & & 0.16&&&&2.19&&&&1.00 & & + & & & & & & & @xmath41 & v & * & @xmath41 & & 0.15 & & & & 2.14&&&&1.00 & & + & & & & & & & @xmath41 & v & * & @xmath41 & & 0.30 & & & & 2.28 & & & & 1.00 & & + & & & & & & & @xmath41 & v & * & @xmath41 & & 0.27 & & & & 2.24 & & & & 1.00 & & + & & & & & & & @xmath41 & @xmath42 & v & @xmath41 & & & _ 1.29 _ & & & & _ 1.73 _ & & & & _ 1.00 _ & + & & & & & & & @xmath41 & @xmath42 & v & @xmath41&&&_1.29_&&&&_1.64_&&&&_1.00 _ & + & & & & & & & @xmath41 & @xmath42 & v & @xmath41&&&_2.25_&&&&_1.76_&&&&_1.00 _ & + & & & & & & & @xmath41 & @xmath42 & v & @xmath41&&&_2.25_&&&&_1.70_&&&&_1.00 _ & + & & & & & & & @xmath41 & v & v & @xmath42&&0.42&0.37&&&2.14&2.12&&&1.00&1.00 & + & & & & & & & @xmath41 & v & v & @xmath42&&0.42&0.21&&&2.05&1.97&&&1.00&1.00 & + & & & & & & & @xmath41 & v & v & @xmath42&&0.78&0.69&&&2.26&2.24&&&1.00&1.00 & + & & & & & & & @xmath41 & v & v & @xmath42&&0.78&0.39&&&2.17&2.05&&&1.00&1.00 & + & & & & & & & @xmath41 & v & v & @xmath41 & & 0.57&0.66&&&2.24&2.66&&&0.98&1.00 & + & & & & & & & @xmath41 & v & v & @xmath41&&0.57&0.69&&&2.14&2.59&&&0.97&1.00 & + & & & & & & & @xmath41 & v & v & @xmath41&&1.03&1.17&&&2.35&2.79&&&0.99&1.00 & + & & & & & & & @xmath41 & v & v & @xmath41&&1.05&1.22&&&2.26&2.70&&&0.98&1.00 & + & & & & & & & @xmath41 & v & v & v & & 0.44 & 0.12 & 0.10 & & 1.97 & 2.14 & 2.47 & & 1.00 & 1.00 & 1.00 + & & & & & & & @xmath41 & v & v & v & & 0.09 & 0.09&0.10&&2.14&2.17&2.39&&1.00&1.00&1.00 + & & & & & & & @xmath41 & v & v & v & & 0.80 & 0.24 & 0.19 & & 2.05 & 2.21 & 2.57 & & 1.00 & 1.00 & 1.00 + & & & & & & & @xmath41 & v & v & v & & 0.15 & 0.18 & 0.19 & & 2.28 & 2.24 & 2.49 & & 1.00 & 1.00 & 1.00 + & & & & & & & @xmath41 & v & v & @xmath41&&0.15&0.12&&&2.45&2.55&&&1.00&1.00 & + & & & & & & & @xmath41 & v & v & @xmath41 & & 0.13 & 0.12&&&2.41 & 2.47&&&1.00&1.00 & + & & & & & & & @xmath41 & v & v & @xmath41 & & 0.27 & 0.21 & & & 2.57&2.68&&&1.00&1.00 & + & & & & & & & @xmath41 & v & v & @xmath41&&0.25&0.21&&&2.51&2.59&&&1.00&1.00 & + & & & & & & & @xmath41 & v & v & @xmath41 & & 0.61 & 0.66&&&1.92 & 2.32&&&1.00&0.99 & + & & & & & & & @xmath41 & v & v & @xmath41 & & 0.61&0.69 & & & 1.84 & 2.26 & & & 1.00&0.98 & + & & & & & & & @xmath41 & v & v & @xmath41&&1.14&1.18 & & & 2.02 & 2.43&&&1.00&0.99 & + & & & & & & & @xmath41 & v & v & @xmath41&&1.15&1.23&&&1.95&2.37&&&1.00&0.99 & + & & & & & & & @xmath42 & v & v & v&&0.53&0.51&0.19&&2.21&2.17&2.35&&1.00&1.00&1.00 + & & & & & & & v & v & v & v & 0.61 & 0.54 & 0.53 & 0.19 & 2.05 & 2.12 & 2.07 & 2.28 & 1.00 & 1.00&1.00&0.99 + & & & & & & & @xmath42 & v & v & v&&0.96&0.93&0.37&&2.32&2.28&2.47&&1.00 & 1.00 & 1.00 + & & & & & & & @xmath42 & v & v & v&&0.98&0.95&0.36&&2.24&2.19&2.39&&1.00&1.00&1.00 + & & & & & & & v&v&@xmath41&@xmath41&_0.93_&_1.02_&&&_1.95_&_2.07_&&&_1.00_&_0.96 _ & & + & & & & & & & v&v&@xmath41&@xmath41&_0.95_&_1.08_&&&_1.84_&_2.00_&&&_1.00_&_0.96 _ & & + & & & & & & & v&v&@xmath41&@xmath41&_1.65_&_1.80_&&&_2.00_&_2.14_&&&_1.00_&_0.96 _ & & + & & & & & & & v&v&@xmath41&@xmath41&_1.71_&_1.88_&&&_1.95_&_2.10_&&&_1.00_&_0.96 _ & & + & & & & & & & v & v & v & v & 0.29 & 0.29 & 0.33 & 0.33 & 2.17 & 2.26 & 2.37 & 2.10 & 1.00 & 1.00 & 1.00 & 1.00 + & & & & & & & v&v&v&v&0.25&0.27&0.32&0.32&2.07&2.14&2.21&2.02&1.00&1.00&1.00&1.00 + & & & & & & & v&v&v&v&0.69&0.60&0.60&0.61&2.35&2.45&2.47&2.21&1.00&1.00&1.00&1.00 + & & & & & & & v&v&v&v&0.50&0.48&0.59&0.60&2.17&2.24&2.35&2.12&1.00&1.00&1.00&1.00 + & & & & & & & @xmath41&v & @xmath43 & @xmath43 & & 0.25&&&&2.14&&&&1.00 & & + & & & & & & & @xmath41&v & @xmath43 & @xmath43 & & 0.10 & & & & 2.39 & & & & 1.00 & & + & & & & & & & @xmath41&v & @xmath43 & @xmath43&&0.48&&&&2.24&&&&1.00 & & + & & & & & & & @xmath41&v & @xmath43 & @xmath43&&0.23&&&&2.43&&&&1.00 & & + & & & & & & & @xmath42&v&v&@xmath41&&0.46&0.50&&&2.51&2.47&&&1.00&1.00 & + & & & & & & & @xmath42&v&v&@xmath41&&0.46&0.51&&&2.39&2.37&&&1.00&1.00 & + & & & & & & & @xmath42&v&v&@xmath41&&0.83&0.89&&&2.65&2.61&&&1.00&1.00 & + & & & & & & & @xmath42&v&v&@xmath41&&0.86&0.91&&&2.55&2.51&&&1.00&1.00 & + & & & & & & & @xmath42&v&v&@xmath42&&0.56&0.80&&&2.17&1.87&&&1.00&1.00 & + & & & & & & & @xmath42&v&v&@xmath42&&0.57&0.81&&&2.07&1.79&&&1.00&1.00 & + & & & & & & & @xmath42&v&v&@xmath42&&1.00&0.44&&&2.28&1.95&&&1.00&1.00 & + & & & & & & & @xmath42&v&v&@xmath42&&1.02&0.46&&&2.19&1.87&&&1.00&1.00 & + & & & & & & & @xmath41&v&v&@xmath41&&0.16&_1.23_&&&1.90&_2.07_&&&1.00&_1.00 _ & + & & & & & & & @xmath41&v&v&@xmath41&&0.12&_1.29_&&&1.92&_2.00_&&&1.00&_1.00 _ & + & & & & & & & @xmath41&v&v&@xmath41&&1.18&_2.16_&&&1.97&_2.12_&&&0.97 & _ 1.00 _ & + & & & & & & & @xmath41&v&v&@xmath41&&0.21&_2.25_&&&2.00&_2.07_&&&1.00&_1.00 _ & + & & & & & & & @xmath40&v&v&@xmath41&&0.45&0.53&&&1.82&1.87&&&1.00&1.00 & + & & & & & & & @xmath40&v&v&@xmath41&&0.39&0.53&&&1.73&1.79&&&1.00&1.00 & + & & & & & & & @xmath40&v&v&@xmath41&&0.84&0.96&&&1.90&1.95&&&1.00&1.00 & + & & & & & & & @xmath40&v&v&@xmath41&&0.78&0.95&&&1.82&1.87&&&1.00&1.00 & + & & & & & & & @xmath42&v&v&@xmath42&&0.48&0.53&&&2.68&2.59&&&1.00&1.00 & + & & & & & & & @xmath42&v&v&@xmath42&&0.48&0.56&&&2.57&2.49&&&0.98&1.00 & + & & & & & & & @xmath42&v&v&@xmath42&&0.86&0.93&&&2.85&2.72&&&1.00&1.00 & + & & & & & & & @xmath42&v&v&@xmath42&&0.86&0.98&&&2.74&2.63&&&0.98&1.00 & + & & & & & & & @xmath42&v&v&v&&0.29&0.30&0.25&&2.85&2.90&3.00&&1.00&1.00&1.00 + & & & & & & & @xmath42&v&v&v&&0.29&0.29&0.30&&2.70&2.76&2.85&&1.00&1.00&1.00 + & & & & & & & @xmath42&@xmath42&@xmath42&@xmath42 & & & & & & & & & & & & + & & & & & & & @xmath42&v&v&v&&0.54&0.56&0.50&&2.93&2.95&3.15&&1.00&1.00&1.00 + & & & & & & & v&v&v&v&0.48&0.33&0.32&0.33&2.43&2.17&2.19&2.32&1.00&1.00&1.00&1.00 + & & & & & & & v&v&v&v&0.50&0.32&0.30&0.33&2.30&2.07&2.10&2.26&1.00&0.99&1.00&0.97 + & & & & & & & v&v&v&v&0.87&0.61&0.74&0.61&2.57&2.28&2.43&2.47&1.00&1.00&1.00&1.00 + & & & & & & & v&v&v&v&0.90&0.60&0.56&0.61&2.47&2.19&2.19&2.37&1.00&1.00&1.00&0.99 + & & & & & & & v&v&v&v&0.21&0.12&0.10&0.12&1.87&2.00&2.10&2.07&0.94&0.99&1.00&1.00 + & & & & & & & v&v&v&v&0.15&0.12&0.09&0.09&1.87&1.95&2.07&2.12&0.94&0.94&1.00&1.00 + & & & & & & & v&v&v&v&0.44&0.25&0.19&0.24&1.92&2.05&2.17&2.12&0.94&1.00&1.00&1.00 + & & & & & & & v&v&v&v&0.35&0.24&0.18&0.19&1.90&2.00&2.12&2.14&0.94&0.95&1.00&1.00 + & & & & & & & v&v&v&v&0.68&0.06&0.06&0.19&2.26&2.32&2.43&2.19&1.00&1.00&1.00&1.00 + & & & & & & & v&v&v&v&0.70&0.06&0.06&0.15&2.14&2.26&2.37&2.19&1.00&1.00&1.00&1.00 + & & & & & & & v&v&v&v&1.20&0.12&0.10&0.36&2.35&2.37&2.55&2.28&1.00&1.00&1.00&1.00 + & & & & & & & v&v&v&v&1.24&0.10&0.10&0.30&2.26&2.37&2.47&2.26&1.00&1.00&1.00&1.00 + & & & & & & & v&v&v&v&0.29&0.33&0.25&0.30&2.45&2.32&2.24&2.59&1.00&1.00&1.00&1.00 + & & & & & & & v&v&v&v&0.27&0.32&0.16&0.30&2.35&2.21&2.17&2.49&0.97&1.00&1.00&1.00 + & & & & & & & v&v&v&v&0.53&0.60&0.64&0.56&2.57&2.45&2.43&2.76&1.00&1.00&1.00&1.00 + & & & & & & & v&v&v&v&0.53&0.59&0.42&0.56&2.49&2.32&2.24&2.63&1.00&1.00&1.00&1.00 + & & & & & & & v&v&v&v&0.12&0.08&0.15&0.53&2.12&2.21&2.05&2.12&1.00&1.00&0.99&1.00 + & & & & & & & v&v&v&v&0.10&0.08&0.15&0.54&2.10&2.14&2.00&2.05&0.99&1.00&0.96&1.00 + & & & & & & & v&v&v&v&0.27&0.21&0.27&0.96&2.14&2.14&2.12&2.26&1.00&0.99&1.00&1.00 + & & & & & & & v&v&v&v&0.23&0.13&0.27&0.98&2.12&2.24&2.07&2.17&1.00&0.99&0.96&1.00 + & & & & & & & @xmath41&@xmath42&v&@xmath41&&&0.53&&&&2.49&&&&1.00 & + & & & & & & & @xmath41&@xmath42&v&@xmath41&&&0.54&&&&2.39&&&&1.00 & + & & & & & & & @xmath41&@xmath42&v&@xmath41&&&0.93&&&&2.61&&&&1.00 & + & & & & & & & @xmath41&@xmath42&v&@xmath41&&&0.96&&&&2.53&&&&1.00 & + & & & & & & & @xmath41&v&v&@xmath41&&0.61&0.50&&&2.30&2.32&&&1.00&1.00 & + & & & & & & & @xmath41&v&v&@xmath41&&0.63&0.51&&&2.21&2.24&&&1.00&1.00 & + & & & & & & & @xmath41&v&v&@xmath41&&1.10&0.90&&&2.39&2.45&&&1.00&1.00 & + & & & & & & & @xmath41&v&v&@xmath41&&1.12&0.91&&&2.32&2.37&&&1.00&1.00 & + & & & & & & & v&v&v&v&_0.83_&0.13&0.13&0.33&_1.92_&2.07&2.05&1.92&_1.00_&1.00&1.00&1.00 + & & & & & & & v&v&v&v&_0.84_&0.12&0.10&0.30&_1.84_&2.03&2.05&1.84&_1.00_&1.00&1.00&1.00 + & & & & & & & v&v&v&v&_1.47_&0.24&0.24&0.63&_2.00_&2.14&2.14&2.00&_1.00_&1.00&1.00&1.00 + & & & & & & & v&v&v&v&_1.50_&0.25&0.24&0.57&_1.92_&2.07&2.07&1.92&_1.00_&1.00&1.00&1.00 + & & & & & & & @xmath41&v&v&@xmath41&&0.42&0.68&&&1.92&2.05&&&0.99&1.00 & + & & & & & & & @xmath41&v&v&@xmath41&&0.42&0.69&&&1.87&1.97&&&0.96&1.00 & + & & & & & & & @xmath41&v&v&@xmath41&&0.78&1.22&&&2.00&2.14&&&1.00&1.00 & + & & & & & & & @xmath41&v&v&@xmath41&&0.78&1.24&&&1.95&2.07&&&0.97&1.00 & + & & & & & & & @xmath41&v&v&@xmath41&&0.25&0.18&&&2.28&2.39&&&1.00&1.00 & + & & & & & & & @xmath41&v&v&@xmath41&&0.25&0.18&&&2.21&2.32&&&1.00&1.00 & + & & & & & & & @xmath41&v&@xmath42&@xmath41&&0.46&&&&2.39&&&&1.00 & & + & & & & & & & @xmath41&v&v&@xmath41&&0.45&0.33&&&2.30&2.43&&&1.00&1.00 & + & & & & & & & @xmath41&@xmath41&v&@xmath41&&&_0.59_&&&&_1.73_&&&&_1.00 _ & + & & & & & & & @xmath41&@xmath41&v&@xmath41&&&_0.46_&&&&_1.64_&&&&_1.00 _ & + & & & & & & & @xmath41&@xmath41&v&@xmath41&&&_1.07_&&&&_1.79_&&&&_1.00 _ & + & & & & & & & @xmath41&@xmath41&v&@xmath41&&&_0.89_&&&&_1.70_&&&&_1.00 _ & + l|cccc|cccc|cccc|cccc|cccc & & 200 & 180 & & & 850 & 860 & & & 0.95 & 0.99 & & & 24.27 & 23.19 & & & 5.5 & 4.5 & + & & 210&200 & & & 850&910 & & & 0.95&0.92 & & & 24.27&23.19 & & & 5.5&4.5 & + & & 180 & 170 & & & 820 & 840 & & & 1.83 & 1.87 & & & 24.27 & 23.19 & & & 5.5 & 4.5 & + & & 200&180 & & & 830&870 & & & 1.83&1.80 & & & 24.27&23.19 & & & 5.5&4.5 & + & & & 180 & 220 & & & 870 & 920 & & & 1.63 & 1.69 & & & 25.21 & 25.28 & & & 5.5 & 4.5 + & & & 190&230 & & & 880&930 & & & 1.67&1.70 & & & 25.21&25.28 & & & 5.5&4.5 + & & & 160 & 220 & & & 820 & 900 & & & 3.26 & 3.11 & & & 24.71 & 25.28 & & & 5.0 & 4.5 + & & & 180&220&&&860&900&&&3.02&3.17&&&25.21&25.28&&&5.5&4.5 + & & 190 & & & & 850 & & & & 0.47 & & & & 25.07 & & & & 5.0 & & + & & 210&&&&900&&&&0.44&&&&25.07&&&&5.0 & & + & & 190 & & & & 840 & & & & 0.87 & & & & 25.07 & & & & 5.0 & & + & & 210 & & & & 890 & & & & 0.81 & & & & 25.07 & & & & 5.0 & & + & & & _ 130_&&&&_340_&&&&_2.50_&&&&_25.67_&&&&_4.5 _ & + & & & _ 140 _ & & & & _ 340 _ & & & & _ 2.40 _ & & & & _ 25.67 _ & & & & _ 4.5 _ & + & & & _ 120 _ & & & & _ 340 _ & & & & _ 4.39 _ & & & & _ 25.67 _ & & & & _ 4.5 _ & + & & & _ 130_&&&&_340_&&&&_4.29_&&&&_25.67_&&&&_4.5 _ & + & & 180&160&&&690&590&&&1.11&0.98&&&25.65&24.91&&&5.0&5.0 & + & & 200&210&&&700&720&&&1.09&0.54&&&25.65&25.41&&&5.0&5.5 & + & & 170&140&&&670&580&&&2.12&1.86&&&25.65&24.91&&&5.0 & 5.0 & + & & 180&200&&&680&720&&&2.09&1.02&&&25.65&25.41&&&5.0&5.5 & + & & 170&120&&&680&610&&&1.56&2.22&&&25.45&25.42&&&4.5&4.5 & + & & 180&130&&&690&610&&&1.54&2.34&&&25.45&25.42&&&4.5&4.5 & + & & 160&110&&&670&600&&&2.89&3.98&&&25.45&25.42&&&4.5&4.5 & + & & 170&120&&&670&600&&&2.90&4.12&&&25.45&25.42&&&4.5&4.5 & + & & 200 & 220 & 240&&680 & 910&1280&&1.03 & 0.33 & 0.36 & & 24.23 & 24.49 & 25.59&&4.5&5.5&5.5 + & & 310&240&250&&1320&1040&1280&&0.27&0.27&0.37&&25.23&25.49&25.59&&5.5&6.5&5.5 + & & 190&210&230&&660&870&1250&&1.92&0.68&0.69&&24.23 & 24.49 & 25.59&&4.5&5.5&5.5 + & & 310&230&250&&1400&1020&1270&&0.48&0.55&0.70&&24.73&25.49&25.59&&5.0&6.5&5.5 + & & 190&180 & & & 990&1030 & & & 0.50&0.43 & & & 24.49&25.00&&&5.0&5.5 & + & & 220&200&&&1070&1050&&&0.47&0.44&&&24.49&25.00&&&5.0&5.5 & + & & 190 & 180 & & & 990 & 1040&&&0.93&0.78&&&24.49&25.00&&&5.0&5.5 & + & & 200&200&&&1030&1050&&&0.91&0.79&&&24.49&25.00&&&5.0&5.5 & + & & 160&120&&&490&490&&&1.39&1.87&&&25.59&25.73&&&4.5&4.5 & + & & 170 & 120 & & & 500 & 490 & & & 1.36 & 1.97 & & & 25.59 & 25.73 & & & 4.5&4.5 & + & & 150&110&&&490&480&&&2.67&3.43&&&25.59&25.73&&&4.5&4.5 & + & & 160&110&&&490&480&&&2.66&3.56&&&25.59&25.73&&&4.5&4.5 & + & & 230&220&370&&930&840&1700&&1.42&1.35&0.60&&24.65&24.86&25.92&&4.5 & 5.0 & 5.0 + & 270 & 250 & 230 & 390 & 990&940&850&1720 & 1.56 & 1.44&1.37&0.62&25.35&24.65&24.86&25.92&4.5&4.5&5.0&5.0 + & & 220&210&340&&910&830&1640&&2.67&2.53&1.20&&24.65&24.86&25.92&&4.5&5.0&5.0 + & & 230&230&370&&910&840&1680&&2.68&2.54&1.16&&24.65&24.86&25.92&&4.5&5.0&5.0 + & _ 170 _ & _ 140 _ & & & _ 530 _ & _ 470 _ & & & _ 2.10 _ & _ 2.47 _ & & & _ 26.00 _ & _ 25.53 _ & & & _ 4.5 _ & _ 4.5 _ & & + & _ 190 _ & _ 140 _ & & & _ 530 _ & _ 470 _ & & & _ 2.05 _ & _ 2.57 _ & & & _ 26.00 _ & _ 25.53 _ & & & _ 4.5 _ & _ 4.5 _ & & + & _ 160 _ & _ 130 _ & & & _ 530 _ & _ 470 _ & & & _ 3.75 _ & _ 4.42 _ & & & _ 26.00 _ & _ 25.53 _ & & & _ 4.5 _ & _ 4.5 _ & & + & _ 170_&_130_&&&_530_&_480_&&&_3.86_&_4.59_&&&_26.00_&_25.53_&&&_4.5_&_4.5 _ & & + & 230 & 190 & 160 & 220 & 930 & 810 & 730 & 850 & 0.78 & 0.82 & 1.00 & 0.85 & 25.08 & 24.75 & 24.13 & 25.51 & 5.0 & 5.5 & 5.0 & 5.5 + & 260&210&190&240&980&830&760&880&0.69&0.77&0.92&0.82&25.08&24.75&24.13&25.51&5.0&5.5&5.0&5.5 + & 190&160&150&210&800&740&710&840&1.98&1.83&1.85&1.65&24.58&24.25&24.13&25.51&4.5&5.0&5.0&5.5 + & 240&200&170&220&950&830&740&850&1.37&1.39&1.77&1.59&25.08&24.75&24.13&25.51&5.0&5.5&5.0&5.5 + & & 450&&&&1760&&&&0.69&&&&25.18&&&&4.5 & & + & & 580&&&&2890&&&&0.37&&&&25.68&&&&5.0 & & + & & 420&&&&1710&&&&1.32&&&&25.18&&&&4.5 & & + & & 540&&&&2640&&&&0.77&&&&25.68&&&&5.0 & & + & & 120&110&&&580&530&&&1.49&1.55&&&24.90&25.48&&&4.5&5.0 & + & & 130&120&&&590&530&&&1.47&1.58&&&24.90&25.48&&&4.5&5.0 & + & & 120&110&&&570&510&&&2.70&2.83&&&24.90&25.48&&&4.5&5.0 & + & & 130&110&&&570&520&&&2.79&2.91&&&24.90&25.49&&&4.5&5.0 & + & & 100&120&&&370&350&&&1.46&1.71&&&24.61&24.80&&&5.0&5.0 & + & & 100&130&&&380&350&&&1.47&1.70&&&24.61&24.80&&&5.0&5.0 & + & & 90&110&&&370&340&&&2.72&3.20&&&24.61&24.80&&&5.0&5.0 & + & & 90&120&&&370&340&&&2.72&3.15&&&24.61&24.80&&&5.0&5.0 & + & & 180&_90_&&&600&_330_&&&0.38&_2.95_&&&25.75&_25.54_&&&5.0&_4.5 _ & + & & 210&_100_&&&700&_330_&&&0.30&_3.04_&&&25.75&_25.54_&&&5.0&_4.5 _ & + & & 120&_90_&&&370&_330_&&&2.69&_5.21_&&&25.25&_25.54_&&&4.5&_4.5 _ & + & & 200&_90_&&&710&_330_&&&0.55&_5.39_&&&25.75&_25.54_&&&5.0&_4.5 _ & + & & 210&180&&&610&560&&&0.95&1.15&&&25.70&25.64&&&5.5&5.5 & + & & 240&200&&&650&560&&&0.81&1.12&&&25.70&25.64&&&5.5&5.5 & + & & 210&170&&&600&550&&&1.84&2.16&&&25.70&25.64&&&5.5&5.5 & + & & 220&180&&&630&560&&&1.67&2.09&&&25.70&25.64&&&5.5&5.5 & + & & 290&300&&&1470&1520&&&1.67&1.73&&&24.41&25.01&&&5.0&5.5 & + & & 300&330&&&1510&1540&&&1.66&1.82&&&24.41&25.01&&&5.0&5.5 & + & & 270&290&&&1460&1510&&&3.05&3.12&&&24.41&25.01&&&5.0&5.5 & + & & 280&310&&&1470&1510&&&3.04&3.26&&&24.41&25.01&&&5.0&5.5 & + & & 120&110&180&&720&680&1190&&1.12&1.20&1.09&&22.99&22.93&24.67&&4.5&4.5&4.5 + & & 130&120&190&&730&700&1210&&1.11&1.20&1.08&&22.99&22.93&24.67&&4.5&4.5&4.5 + & & & & & & & & & & & & & & & & & & & & + & & 120&110&180&&700&680&1140&&2.19&2.26&2.22&&22.99&22.93&24.67&&4.5&4.5&4.5 + & 150&190&170&190&700&730&710&820&1.47&0.84&0.87&0.98&25.54&25.65&25.44&25.93&5.0&5.5&5.5&5.5 + & 170&200&190&200&710&760&740&830&1.48&0.84&0.82&0.99&25.54&25.65&25.44&25.93&5.0&5.5&5.5&5.5 + & 150&180&140&170&680&710&620&800&2.74&1.71&2.19&1.89&25.54&25.65&24.94&25.93&5.0&5.5&5.0&5.5 + & 150&190&180&190&690&730&710&820&2.81&1.66&1.55&1.88&25.54&25.65&25.44&25.93&5.0&5.5&5.5&5.5 + & 330&300&300&350&1040&1110&1200&1400&0.47&0.30&0.28&0.32&25.98&25.97&25.98&25.87&5.0&6.0&6.0&5.0 + & 380&320&330&390&1200&1120&1300&1630&0.36&0.31&0.25&0.26&25.98&25.97&25.98&25.87&5.0&6.0&6.0&5.0 + & 310&280&290&350&990&1050&1190&1330&0.99&0.65&0.54&0.64&25.98&25.97&25.98&25.87&5.0&6.0&6.0&5.0 + & 340&300&320&370&1100&1080&1250&1500&0.81&0.62&0.51&0.56&25.98&25.97&25.98&25.87&5.0&6.0&6.0&5.0 + & 240&420&450&380&930&2080&2290&1600&1.85&0.19&0.21&0.55&25.94&25.70&25.74&25.84&4.5&6.0&6.0&4.5 + & 250&460&460&430&950&2140&2350&1850&1.88&0.20&0.21&0.45&25.94&25.70&25.74&25.84&4.5&6.0&6.0&4.5 + & 220&410&430&360&930&2010&2360&1560&3.33&0.39&0.38&1.04&25.94&25.70&25.74&25.84&4.5&6.0&6.0&4.5 + & 240&440&450&410&930&2200&2400&1740&3.40&0.36&0.39&0.91&25.94&25.70&25.74&25.84&4.5&6.0&6.0&4.5 + & 90&70&90&90&430&330&380&460&0.92&0.98&0.73&1.03&25.58&24.64&24.70&25.99&5.0&5.0&5.0&5.0 + & 100&90&110&100&440&350&460&470&0.87&0.92&0.49&1.04&25.58&24.64&25.70&25.99&5.0&5.0&6.0&5.0 + & 80&70&70&80&420&330&310&440&1.72&1.83&1.94&1.97&25.58&24.64&24.20&25.99&5.0&5.0&4.5&5.0 + & 90&80&90&90&420&330&390&450&1.73&1.75&1.23&1.95&25.58&24.64&24.70&25.99&5.0&5.0&5.0&5.0 + & 310&270&240&190&1230&1190&880&700&0.33&0.22&0.39&1.35&25.80&25.87&25.93&25.70&5.0&6.5&6.5&5.0 + & 330&280&250&200&1310&1200&890&710&0.30&0.23&0.40&1.38&25.80&25.87&25.93&25.70&5.0&6.5&6.5&5.0 + & 290&250&220&170&1110&970&880&690&0.73&0.57&0.72&2.58&25.80&25.37&25.93&25.70&5.0&6.0&6.5&5.0 + & 310&280&240&180&1220&1210&890&700&0.63&0.42&0.73&2.58&25.80&25.87&25.93&25.70&5.0&6.5&6.5&5.0 + & & & 130&&&&620&&&&1.65&&&&25.41&&&&4.5 & + & & & 110&&&&640&&&&1.68&&&&25.41&&&&4.5 & + & & & 120&&&&620&&&&2.97&&&&25.41&&&&4.5 & + & & & 130&&&&620&&&&3.07&&&&25.41&&&&4.5 & + & & 130&130&&&540&560&&&1.73&1.43&&&25.70&25.66&&&4.5&4.5 & + & & 140&140&&&540&570&&&1.76&1.47&&&25.70&25.66&&&4.5&4.5 & + & & 120&120&&&530&550&&&3.12&2.67&&&25.70&25.66&&&4.5&4.5 & + & & 130&130&&&540&550&&&3.20&2.71&&&25.70&25.66&&&4.5&4.5 & + & _ 180_&250&240&230&_570_&960&950&750&_1.84_&0.36&0.35&0.76&_25.40_&24.93&25.44&26.00&_4.5_&5.5&6.0&5.5
+ & _ 200_&270&280&260&_580_&1010&1070&790&_1.83_&0.32&0.29&0.69&_25.40_&24.93&25.94&26.00&_4.5_&5.5&6.5&5.5
+ & _ 180_&250&250&220&_570_&970&970&740&_3.36_&0.65&0.65&1.49&_25.40_&24.93&25.44&26.00&_4.5_&5.5&6.0&5.5
190_&250&260&240&_580_&950&980&770&_3.35_&0.69&0.65&1.34&_25.40_&24.93&25.44&26.00&_4.5_&5.5&6.0&5.5 + & & 280&230&&&870&800&&&0.96&1.64&&&25.79&25.93&&&4.5&4.5 & + & & 280&240&&&880&810&&&0.96&1.66&&&25.79&25.93&&&4.5&4.5 & + & & 250&210&&&860&770&&&1.83&3.03&&&25.79&25.93&&&4.5&4.5 & + & & 280&230&&&860&790&&&1.83&3.07&&&25.79&25.93&&&4.5&4.5 & + & & 230&270&&&1030&1310&&&0.75&0.57&&&25.58&25.71&&&5.5&5.5 & + & & 260&300&&&1050&1340&&&0.76&0.59&&&25.58&25.71&&&5.5&5.5 & + & & 230&&&&1010&&&&1.40&&&&25.58&&&&5.5 & & + & & 240&280&&&1040&1300&&&1.36&1.10&&&25.58&25.71&&&5.5&5.5 & + & & & _ 250_&&&&_680_&&&&_1.16_&&&&_25.89_&&&&_4.5 _ & + & & & _ 290_&&&&_720_&&&&_0.89_&&&&_25.89_&&&&_4.5 _ & + & & & _ 240_&&&&_670_&&&&_2.16_&&&&_25.89_&&&&_4.5 _ & + & & & _ 280_&&&&_720_&&&&_1.74_&&&&_25.89_&&&&_4.5 _ & + | most , if not all , disk galaxies have a thin ( classical ) disk and a thick disk . in most models thick disks
are thought to be a necessary consequence of the disk formation and/or evolution of the galaxy .
we present the results of a study of the thick disk properties in a sample of carefully selected edge - on galaxies with types ranging from @xmath0 to @xmath1 .
we fitted one - dimensional luminosity profiles with physically motivated functions the solutions of two stellar and one gaseous isothermal coupled disks in equilibrium which are likely to yield more accurate results than other functions used in previous studies .
the images used for the fits come from the spitzer survey of stellar structure in galaxies ( s@xmath2 g ) .
we found that thick disks are on average more massive than previously reported , mostly due to the selected fitting function .
typically , the thin and the thick disk have similar masses .
we also found that thick disks do not flare significantly within the observed range in galactocentric radii and that the ratio of thick to thin disk scaleheights is higher for galaxies of earlier types .
our results tend to favor an _ in situ _
origin for most of the stars in the thick disk .
in addition the thick disk may contain a significant amount of stars coming from satellites accreted after the initial build - up of the galaxy and an extra fraction of stars coming from the secular heating of the thin disk by its own overdensities .
assigning thick disk light to the thin disk component may lead to an underestimate of the overall stellar mass in galaxies , because of different mass to light ratios in the two disk components . on the basis of our new results
, we estimate that disk stellar masses are between 10% and 50% higher than previously thought and we suggest that thick disks are a reservoir of `` local missing baryons '' . |
all four solar neutrino experiments ( homestake , kamiokande , gallex , sage ) observe a deficit of solar neutrinos compared to the predictions of the standard solar model . additionally , any two of the three classes of solar neutrino experiments indicate that the largest suppression is in the middle of the solar neutrino spectrum ( the be-7 line and the lower part of the b-8 spectrum)@xcite . among these four experiments
the homestake experiment is taking data for almost 25 years @xcite .
the reliability of the radiochemical method for detecting solar neutrinos has been tested recently by the gallex experiment @xcite .
all efforts to resolve the solar neutrino problem by improving solar , nuclear , and neutrino physics have not proven successful so far@xcite .
this may also mean that the average solar neutrino flux extracted from the four experiments may not be the proper quantity to explain the production of neutrinos in the gravitationally stabilized solar fusion reactor @xcite .
this conjecture is supported by the fact that the pattern of the detected solar neutrino flux in the four solar neutrino experiments hints at possible variation of the solar neutrino flux over time , which has been generated interest to look at time dependent phenomena @xcite related to solar physics and neutrino physics .
figs . 1 and 2 show the ar-37 production rate data which davis and collaborators @xcite have acquired and analyzed in more than 25 years . for the analysis of the radioactive decay of argon with a small number of counts
the method of maximum likelihood was employed which brings in a wide range of error .
these error bars are omitted in figs . 1 and 2 and
will not be taken into account in the following analysis .
thus , figs .
1 and 2 show only the mean values for the ar-37 production rate for the various individual experiments .
the data are unevenly spaced and the two data sets contain only low numbers of data .
recently , the ar-37 production rate data , shown in fig.1 , have been reviewed , taking into account counter efficiencies , chemical yields , and background effects @xcite .
the new analysis has yielded an overall rate approximately ten percent higher than the earlier analysis .
the time distribution of the data in figs .
1 and 2 is essentially unchanged . the combined maximum likelihood rate for the 99 experiments in fig .
2 is @xmath0 ar-37 atoms per day in 615 tons of perchlorethylen or @xmath1 snu @xmath2 captures / target atom / sec ) .
this experimental result , compared with the theoretical predictions of @xmath3 snu for the standard solar model , has become known as the solar neutrino problem @xcite .
the following provides the fourier spectrum analysis of the earlier ar-37 production data @xcite and , for comparison , of the new data .
the analysis shows that the review of the earlier data did not change essentially the resulting power spectral density that reveals a strong harmonic content in the data . to remove some of the high frequency noise from the data in figs . 1 and 2 , the five - point moving average
is calculated , which adjusts each point to be the average value of the five points around it .
the results in figs .
3 and 4 are longer than original by one point .
this causes a slight phase shift in the display .
the fast fourier transform of a real series results in a complex series where the second half of the result is the mirror image of the first half .
the power spectral density is the magnitude squared of the first half of the fast fourier transform and factors out the length of the original series .
the power spectral density in figs .
5 and 6 is useful for comparing the frequency spectrum of different series and can be thought of as the power of a series at a particular frequency @xcite .
the investigation of the variation of the ar-37 production rate over time in the homestake solar neutrino experiment reveals several periodicities shorter than or almost equal to 11 years . a number of harmonics contained in the ar-37 data set may have a non - random origin .
it is a striking fact that the power spectrum density did not change significantly in the various analysis undertaken over the years while gradually accumulating data in the homestake experiment .
it can not be excluded at this point of time that the discovered harmonic content of the data reflects solar activity through a not yet known physical phenomena , seated in the deep interior of the sun .
attention is drawn to the existence of distinct peaks in the power spectrum density of the ar-37 production rate . extending and improving of the data
could be achieved through taking into account the results of the other three solar neutrino experiments ( kamiokande , gallex , sage ) if they continue operating over a sufficient period of time . in conclusion ,
the present results confirm the existence of periodicities in the new set of homestake data , previously reported for the earlier data .
fig . 1 and 2 exhibit periods of high and low ar-37 production rates indicating that there is an apparent change in the signal .
in figs . 3 and 4 one notes that the five - point moving average in the periods 1977 to 1980 and 1987 to 1990 shows a strongly suppressed signal .
the highest peakes in figs .
5 and 6 appear at about 10.0 yr and 0.5 yr periodicities .
shorter periods of 1.66 yr and 0.71 yr can also be observed . | the paper provides results of the fourier spectrum analysis of the new ar-37 production rate data of the homestake solar neutrino experiment and compares them with results for earlier data , revealing the harmonic content in the ar-37 production in the homestake experiment . |
le premier cas intressant de la thorie de patterson - sullivan , est ltude des sous - groupes du groupe @xmath0 .
tant donn un sous - groupe @xmath1 de @xmath2 , on sintresse une donne dynamique , lexposant critique , et une donne gomtrique , lensemble limite , qui sont dfinis de la faons suivante : l*exposant critique * est le rel not @xmath3 , dfinit par @xmath4 cela correspond la vitesse exponentielle laquelle crot le groupe .
l*ensemble limite * est la partie de @xmath5 note @xmath6 , qui est le plus petit ferm invariant non vide pour laction de @xmath1 sur @xmath7 par homographie . les ensembles limites de sous - groupes discrets de @xmath0 ressemblent en gnral des ensembles de cantor , et leur exposant critique est un nombre entre @xmath8 et @xmath9 . on a alors le trs joli rsultat suivant , qui relie cette donne dynamique et cette donne gomtrique , qui navaient priori rien voir .
si @xmath1 est un sous - semi - groupe de type fini et non lmentaire de @xmath2 , alors on a lgalit @xmath10 bien que lnonc ne fasse pas intervenir de gomtrie hyperbolique , la preuve utilise de faon cruciale le fait que @xmath2 agisse par isomtrie sur le plan hyperbolique @xmath11 .
sur le disque de poincar ] pour un sous - semi - groupe @xmath1 du groupe disomtries @xmath12 dun espace gromov - hyperbolique @xmath13 propre , nous dfinissons une notion dentropie ( par analogie avec lentropie volumique ) , qui gnralise la notion dexposant critique des groupes discrets .
lentropie est une faon de mesurer lespace occup par lorbite dun semi - groupe dans lespace @xmath13 , tandis que lexposant critique mesure la quantit dlments .
lnonc suivant donne une caractrisation de lentropie qui justifie son intrt : [ thm0 ] soit @xmath13 un espace gromov - hyperbolique dadhrence compacte , et soit @xmath1 un semi - groupe disomtries de @xmath13 do nt lensemble limite contient au moins deux points , alors on a @xmath14 autrement dit , lentropie @xmath15 est la borne suprieure des exposants critiques des sous - semi - groupes de schottky du semi - groupe @xmath1 ( cest - - dire des sous - semi - groupes ayant la dynamique la plus simple ( voir [ schottky ] pour une dfinition prcise ) ) .
nous supposons ici que lespace est propre et de bord compact ) , mais nous verrons que cette hypothse nest pas beaucoup plus forte que de demander seulement la propret de lespace @xmath13 ( voir le paragraphe [ scompacite ] ) .
lorsque @xmath1 est un groupe , si lon remplace les semi - groupes de schottky par des _ groupes _ de schottky au sens classique , le rsultat devient faux daprs un thorme de p.doyle ( voir @xcite ) .
le thorme [ thm0 ] permet dtudier la dimension linfini du semi - groupe , puisque lon obtient en corollaire une gnralisation dun rsultat de f.paulin ( voir @xcite ) : [ cpaulin ] soit @xmath13 un espace gromov - hyperbolique propre bord compact , et soit @xmath1 un semi - groupe disomtries de @xmath13 do nt lensemble limite contient au moins deux points .
alors la dimension visuelle de lensemble limite radial du semi - groupe @xmath1 est gale lentropie du semi - groupe : @xmath16 la dimension visuelle @xmath17 est une gnralisation naturelle de la dimension de hausdorff au bord dun espace hyperbolique .
et lensemble limite radial @xmath18 ( appel aussi ensemble limite conique ) du semi - groupe @xmath1 est lensemble des points @xmath19 du bord qui sont limite dune quasi - godsique de lorbite @xmath20 .
ceci nous permet de calculer la dimension de hausdorff de certains ensemble auto - similaires pour lesquels on ne savait pas encore faire ma connaissance : si @xmath21 est un nombre de salem et si @xmath1 est le sous - semi - groupe du groupe affine de @xmath22 engendr par les applications @xmath23 o @xmath24 pour @xmath25 une partie finie de @xmath26 , alors on a lgalit @xmath27 le rsultat tait connu pour un nombre de pisot , mais semble nouveau pour un nombre de salem .
une telle galit est toujours ltat de conjecture pour les semi - groupes de kenyon ( voir sections [ dev - beta ] et [ semi - ken ] ) .
voici dautres corollaires du thorme [ thm0 ] : [ lim ] soit @xmath13 un espace gromov - hyperbolique propre bord compact et soit @xmath1 un semi - groupe discret disomtries de @xmath13 do nt lensemble limite contient au moins deux points .
la limite suprieure dans la dfinition de lexposant critique du semi - groupe @xmath1 est une vraie limite : @xmath28 [ cont - inf ] lentropie est semi - continue infrieurement en les semi - groupes do nt lensemble limite contient au moins deux points .
voir corollaire [ semi - cont ] pour un nonc plus prcis .
ce rsultat gnralise celui que donne f. paulin la fin de son article @xcite , puisque lon ne fait ni lhypothse que le semi - groupe soit un groupe , ni quil soit discret , ni quil soit de type fini , ni que lespace soit godsique ou quasi - godsique , et cela fonctionne aussi bien pour la convergence algbrique que pour la convergence gomtrique .
nos rsultats sur les semi - groupes permettent dobtenir un rsultat sur les groupes : [ sous - groupes ] soit @xmath13 un espace gromov - hyperbolique propre bord compact , et soit @xmath1 un groupe discret et sans torsion disomtries de @xmath13 ne fixant pas de point au bord , alors on a @xmath29 o la borne suprieure est prise sur lensemble des sous - groupes de schottky du groupe @xmath1 .
dans la section [ cadre ] , nous donnons les dfinitions et outils qui serviront dans la suite . la section [ contract ] est consacre dfinir et donner des proprits sur les semi - groupes contractants . on y
dmontre par exemple que lensemble limite dun semi - groupe contractant de type fini est toujours radial .
on construit dans la section [ pthm1 ] une grosse partie contractante dun semi - groupe disomtries .
cest ltape principale pour dmontrer le thorme [ thm0 ] .
dans la section [ schottky ] , on dfinit ce quest un semi - groupe de schottky , et lon dmontre lexistence de gros sous - semi - groupes de schottky .
cela permettra davoir une preuve du thorme [ thm0 ] .
les sections [ dim - vis ] [ semi - continuite ] sont consacres des corollaires du thorme .
dans la premire , nous obtenons une gnralisation dun rsultat de f. paulin ( corollaire [ cpaulin ] , voir section [ paulin ] ) , que nous appliquons ltude des semi - groupes de dveloppement en base @xmath21 ( voir section [ dev - beta ] ) .
dans la deuxime , nous voyons un rsultat sur les sous - groupes de schottky dun groupe discret ( corollaire [ sous - groupes ] , voir section [ groupes ] ) .
nous montrons ensuite dans la section [ car - ent ] que lexposant critique est une vraie limite .
et pour finir nous montrons la semi - continuit infrieure de lentropie des semi - groupes ( corollaire [ cont - inf ] , voir section [ semi - continuite ] ) .
nous dfinissons ici les objets que nous manipulerons dans tout larticle , en commenant par les espaces gromov - hyperboliques et leur bord .
nous verrons en particulier la dfinition et des proprits de lentropie . on notera @xmath30 la distance entre deux point @xmath31 dun espace mtrique @xmath13 .
tant donn un espace mtrique , on peut dfinir le produit de gromov , qui permet de mesurer le dfaut dgalit triangulaire de trois points @xmath32 , @xmath33 et @xmath34 : soit @xmath13 un espace mtrique de point base @xmath34 . on appelle * produit de gromov * de deux points @xmath31 le rel @xmath35 on peut alors dfinir la gromov - hyperbolicit : on dit quun espace @xmath13 est * @xmath36-hyperbolique * pour un rel @xmath37 , si cest un espace mtrique vrifiant lingalit @xmath38 pour tous @xmath39 et @xmath40 @xmath41 . on dit
quun espace @xmath13 est * gromov - hyperbolique * sil existe un rel @xmath36 tel que lespace @xmath13 est @xmath36-hyperbolique .
un espace gromov - hyperbolique est un espace qui ressemble , vu de loin , un arbre .
les arbres sont dailleurs des espaces @xmath8-hyperboliques .
dans un espace hyperbolique , un grand produit de gromov caractrise des points qui sont proches vus de @xmath34 . dfinissons lexposant critique .
il sagit de la vitesse exponentielle de croissance dune partie de @xmath13 . on appelle
* exposant critique * dune partie @xmath42 dun espace mtrique @xmath13 de point base @xmath34 , le rel ( ventuellement infini ) @xmath43 par ingalit triangulaire , lexposant critique ne dpend pas du point @xmath34 choisi .
[ delta - ann ] on peut aussi considrer seulement les lments dun anneau . on a @xmath44 si @xmath45 est une partie non borne . on va maintenant dfinir lentropie de nimporte quelle partie @xmath45 de @xmath13 .
cela correspond la vitesse exponentielle de croissance du point de vue de lespace occup , et non plus du point de vu du comptage comme pour lexposant critique .
soit @xmath13 un espace mtrique . on dit quune partie @xmath46 est * spare * sil existe un rel @xmath47 tel que la partie @xmath45 soit * @xmath48-spare * , cest - - dire tel que @xmath49 pour tous @xmath50 @xmath51 .
+ on dit quune partie @xmath46 est une partie * couvrante * de @xmath52 sil existe un rel @xmath47 telle que la partie soit * @xmath48-couvrante * de @xmath53 , cest - - dire telle que pour tout @xmath54 , il existe @xmath55 tel que @xmath56 on appelle * entropie * dune partie @xmath46 dun espace mtrique @xmath13 , le rel ( ventuellement infini ) @xmath57 o la borne suprieure est prise sur les parties @xmath58 spares de @xmath13 .
+ on pose @xmath59 par convention .
la remarque suivante justifie le nom dentropie : quand @xmath60 , lentropie dune partie @xmath46 est gale lentropie volumique de lensemble @xmath61 .
on peut trouver une partie @xmath58 @xmath48-spare et @xmath62-couvrante de la partie @xmath45 pour un rel @xmath63 assez petit et un rel @xmath62 assez grand ( voir [ epsc ] ) , et on peut montrer que son exposant critique est alors gale lentropie de @xmath45 .
soit @xmath34 un point de @xmath60 . on a les ingalits @xmath64 @xmath65 en passant la limite aprs avoir pris le @xmath66 et divis par @xmath67 , on obtient le rsultat annonc .
dans lespace @xmath68 muni de la mtrique usuelle , lentropie de @xmath11 vaut @xmath9 et lentropie dune horoboule vaut @xmath69 .
voici quelques proprits de lentropie .
* proprits de lentropie * [ pptex ] + soit une partie @xmath70 dun espace mtrique @xmath13 de point base @xmath34 .
1 . lentropie ne dpend pas du point base @xmath71 choisi .
2 . on a lingalit @xmath72 .
si la partie @xmath25 est spare , alors on a lgalit @xmath73 .
4 . lentropie est croissante : si @xmath74 , alors on a @xmath75 1 .
cela dcoule du fait que lexposant critique ne dpende pas du point base @xmath34 choisi , ce qui dcoule de lingalit triangulaire .
2 . pour toute partie spare @xmath58 , on a @xmath76 .
do le rsultat en passant la borne suprieure .
si la partie @xmath25 est spare , on a @xmath77 .
4 . pour toute partie spare @xmath58 on a @xmath78 , donc @xmath79 .
do le rsultat en passant la borne suprieure .
[ propre ] on dit quun espace mtrique @xmath13 est * propre * si ses boules fermes sont compactes .
nous allons voir qu chaque espace hyperbolique on peut ajouter un bord sur lequel le produit de gromov stend naturellement .
lespace hyperbolique muni de son bord est en quelque sorte une compactification , puisque nous verrons que cette union est compacte si lespace est propre et vrifie une proprit supplmentaire .
soit @xmath13 un espace gromov - hyperbolique .
[ def_bord ] on dit quune suite @xmath80 est * convergente * si lon a @xmath81 on dfinit le * bord * @xmath82 dune partie @xmath25 lespace hyperbolique @xmath13 comme quotient de lensemble des suites convergentes @xmath83 par la relation dquivalence @xmath84 dfinie par @xmath85 si lon navait pas suppos lespace @xmath13 gromov - hyperbolique , la relation @xmath84 dfinie ci - dessus ne serait pas ncessairement transitive .
par exemple elle ne lest pas pour @xmath86 . on appelle *
adhrence de gromov * dune partie @xmath87 lensemble @xmath88 o @xmath89 est ladhrence de @xmath25 dans @xmath13 . 1 . le groupe @xmath12 agit naturellement sur le bord @xmath90 .
2 . le produit de gromov stend naturellement ladhrence de gromov : @xmath91 pour @xmath92 .
un espace mtrique @xmath13 est propre bord compact si et seulement si ladhrence @xmath93 est compacte .
le premier point ci - dessus permet de gnraliser la notion de suite convergentes toute suite de @xmath93 .
cela dfinit la topologie que lon considre sur lespace @xmath93 .
voir aussi dans la partie compacit de ladhrence ci - dessous pour une caractrisation de la topologie de ladhrence de gromov de @xmath13 .
soit @xmath13 un espace mtrique .
tant donn un rel @xmath94 , on dit que deux parties @xmath25 et @xmath95 de @xmath96 sont * @xmath97-disjointes * si elles sont dadhrence dans @xmath13 disjointes et que lon a @xmath98 on dit que les parties @xmath25 et @xmath95 sont * gromov - disjointes * sil existe une constante @xmath94 telle que les parties @xmath25 et @xmath95 sont @xmath97-disjointes .
ces notions se gnralisent de faon claire des partie de lespace @xmath99 quand lespace @xmath13 est gromov - hyperbolique .
si lespace @xmath13 est propre et bord compact , alors des parties @xmath25 et @xmath95 de @xmath99 sont gromov - disjointes si et seulement si leur adhrences de gromov @xmath100 et @xmath101 sont disjointes , mais ceci est faux en gnral .
dire que deux partie @xmath25 et @xmath95 sont gromov - disjointes revient dire quelles sont disjointes en tant quensembles , et disjointes en linfini vues du point base @xmath34 .
cette notion permettra de dfinir ce quest un semi - groupe contractant ( voir [ contract ] ) et ce quest un semi - groupe de schottky ( voir [ schottky ] ) .
nous aurons besoin de la compacit de ladhrence @xmath93 pour contrler la faon do nt lentropie part linfini ( voir la sous - section [ supp ] ) . on muni ladhrence de gromov @xmath93 dun espace gromov - hyperbolique @xmath13 , de la base de voisinages forme des boules ouvertes @xmath102 de @xmath13 et des boules
@xmath103 pour les points @xmath104 .
cela dfinit bien la mme topologie que celle donne par les suites convergentes .
voici une condition sur lespace @xmath13 qui sert avoir la compacit de ladhrence : soit @xmath13 un espace mtrique de point base @xmath34 . on dit
que lespace @xmath13 est @xmath97-toil si pour tout point @xmath105 , il existe une @xmath97-godsique de @xmath34 @xmath32 , cest - - dire une suite finie @xmath106 dlments de @xmath13 telle que @xmath107 , @xmath108 , @xmath109 @xmath110 pour tous @xmath111 et @xmath112 .
un espace godsique est @xmath97-toil pour tout @xmath94 .
@xmath113 muni de la mtrique euclidienne est @xmath114-toil .
dans cet exemple , il suffit de considrer les points @xmath115 de @xmath113 qui sont distance infrieure ou gale @xmath116 du segment @xmath117 $ ] dans @xmath118 .
[ compacite ] soit @xmath13 un espace gromov - hyperbolique , propre , et @xmath97-toil , alors son adhrence de gromov @xmath93 est compacte .
il suffit de suivre la preuve du lemme 7.3 dans @xcite . lensemble @xmath119 muni de la mtrique @xmath120 est gromov - hyperbolique , propre , et dadhrence non compacte ( voir figure [ fig_ex - noncompact ] ) .
sur la figure ci - dessus , les artes grises ne font pas parties de lespace et indiquent des distances .
tant donn un ensemble disomtries dun espace mtrique @xmath13 , on dfinit son exposant critique , respectivement son entropie , qui correspondent aux vitesses auxquelles crot lorbite dun point sous laction de lensemble disomtries , dun point de vue de comptage , respectivement dun point de vue de lespace occup .
+ sur le demi - plan de poincar @xmath11 ] soit @xmath13 un espace mtrique de point base @xmath34 . on appelle * exposant critique * dune partie @xmath122 , lexposant critique de lorbite @xmath123 : @xmath124 on dit quune partie @xmath125 est * spare * ( respectivement * @xmath48-spare * ) si lorbite @xmath123 est spare ( respectivement @xmath48-spare ) , et si pour toutes les isomtries @xmath126 , on a @xmath127 . en gnral , lexposant critique dpend du point @xmath34 choisi ( par exemple pour un groupe elliptique dense ) , mais lexposant critique dune partie spare de @xmath12 nen dpend pas .
si une partie @xmath125 est spare et que lorbite @xmath128 est infinie , alors lexposant critique de la partie @xmath25 est aussi lexposant critique de la srie de poincar @xmath129 de @xmath25 : @xmath130 cest - - dire que la srie diverge pour @xmath131 et converge pour @xmath132 .
voir @xcite pour plus de dtails . on dit
quune partie @xmath125 est une partie * couvrante * ( respectivement * @xmath48-couvrante * ) de @xmath133 si lorbite @xmath123 est une partie couvrante ( respectivement @xmath48-couvrante ) de @xmath134 . on appelle *
entropie * dune partie
@xmath125 lentropie de lorbite @xmath123 : @xmath135 quand @xmath60 , lentropie dune partie @xmath125 est gale lentropie volumique de lorbite @xmath136 dune boule @xmath137 .
voici quelques proprits de lentropie .
* proprits de lentropie * [ ppteg ] + soit @xmath13 un espace mtrique de point base @xmath34 , et soit une partie @xmath125 . on a les proprits : 1 .
lentropie ne dpend pas du point base @xmath34 choisi .
2 . on a lingalit [ hmaj ] @xmath138 .
si la partie @xmath139 est spare , alors on a lgalit @xmath73 .
+ [ sep ] ceci est en particulier le cas si @xmath25 est un groupe discret .
4 . lentropie est croissante : si @xmath140 , alors on a [ hc ] @xmath141 5 .
si lespace @xmath13 est propre , alors pour toute partie @xmath58 couvrante de @xmath25 , on a lingalit [ couv ] @xmath142 .
si @xmath25 , @xmath95 et @xmath97 sont des parties de @xmath12 telles que lon ait @xmath143 , [ hu ] alors on a @xmath144
pour toute isomtrie @xmath145 , on a [ hm ] @xmath146 .
si lespace @xmath13 est propre , alors en posant @xmath147 , on a [ hf ] @xmath148 [ epsc ] pour toute partie @xmath52 dun espace mtrique @xmath13 , il existe une partie @xmath58 @xmath48-spare et @xmath149-couvrante de @xmath53 .
ainsi , quand lespace @xmath13 est propre , pour calculer lentropie dune partie @xmath25 de @xmath12 , on est ramen calculer lexposant critique dune partie spare de @xmath25 .
les points 2 , 3 et 4 sont vidents partir des proprits [ pptex ] .
le point 1 dcoule du fait que lexposant critique dune partie spare ne dpende pas du point base @xmath34 choisi .
montrons le point 5 .
soit @xmath58 une partie couvrante de @xmath25 et soit @xmath150 une partie spare de @xmath25 .
montrons que lon a @xmath151 .
cela va dcouler du lemme suivant .
[ isc ] soit @xmath13 un espace mtrique propre de point base @xmath34 ,
et @xmath152 une partie du groupe disomtries .
alors , pour tous rels @xmath153 et @xmath154 , il existe une constante @xmath155 telle que pour toute partie @xmath62-couvrante @xmath58 de @xmath53 , pour toute partie @xmath156-spare @xmath150 de @xmath53 , et pour toute partie @xmath157 de @xmath53 , on ait lingalit @xmath158 o @xmath159 est le @xmath62-voisinage ferm de @xmath157 .
soit @xmath160 une partie @xmath156-couvrante et spare de la boule @xmath161 .
posons @xmath162 son cardinal , qui est fini par propret .
pour toute partie @xmath156-spare @xmath150 de @xmath13 on a alors @xmath163 et comme pour toute isomtrie @xmath164 , la partie @xmath165 est encore @xmath156-spare , on a aussi @xmath166 on a donc , pour toute partie @xmath58
@xmath62-couvrante de @xmath53 , et @xmath150 partie @xmath156-spare de @xmath53 , @xmath167 daprs le lemme ci - dessus , il existe une constante @xmath97 telle que @xmath168 ainsi , on a @xmath151 en passant la limite . on obtient alors
lingalit souhaite @xmath169 en passant la borne suprieure sur les parties spares @xmath150 de @xmath25 .
montrons maintenant le point [ hu ] .
si @xmath58 est une partie spare de @xmath25 , alors on a @xmath170 pour tout rel @xmath171 .
donc @xmath172 ds que @xmath173 , et donc @xmath174 .
montrons le point [ hm ] .
soit @xmath175 , et soit @xmath58 une partie @xmath62-spare et couvrante de @xmath25 , pour un rel @xmath176 .
pour tout @xmath177 , lingalit triangulaire donne @xmath178 .
ainsi la partie @xmath179 est encore une partie spare et couvrante de @xmath180 , et on a @xmath181 do @xmath182 lautre ingalit @xmath183 sobtient par symtrie , en remplaant llment @xmath184 par @xmath185 et la partie @xmath25 par @xmath180 .
lgalit @xmath186 sobtient de la mme faon .
le point [ hf ] sobtient en remarquant que la partie @xmath187 est borne et que donc son intersection avec toute partie spare est finie par propret .
voici un exemple de semi - groupe dexposant critique strictement suprieur son entropie cause dun phnomne de chevauchements qui nexiste pas pour les groupes
. dans le disque de poincar . ] [ exee ] le sous - semi - groupe de @xmath2 engendr par les matrices @xmath188 et @xmath189 et agissant sur le disque de poincar @xmath190 , a pour exposant critique @xmath191 et a pour entropie @xmath9 . en particulier
il nest pas spar .
lexposant critique du semi - groupe lexemple ci - dessus sobtient facilement puisque le semi - groupe est libre , et parce - que cest un semi - groupe de dveloppement @xmath21-adique , ce qui permet davoir que la norme dun lment est comparable sa longueur en les gnrateurs : @xmath192 le fait que lentropie vaille @xmath9 ( le maximum pour un semi - groupe disomtries
de @xmath190 ) dcoule du thorme [ cpaulin ] , parce - quil est facile de montrer que lensemble limite de ce semi - groupe est un segment de longueur non nulle ( vu dans @xmath193 ) , et parce - que lensemble limite radial est gal lensemble limite tout entier par la proposition [ pradial ] , puisque le semi - groupe est contractant et de type fini .
les isomtries dun espace @xmath13 gromov - hyperbolique agissent naturellement sur le bord @xmath90 .
pour tudier cette action sur le bord , commenons par dfinir lensemble limite du semi - groupe .
soit @xmath13 un espace gromov - hyperbolique de point base @xmath34 , et soit @xmath1 un semi - groupe disomtries de @xmath13 .
on appelle * ensemble limite * du semi - groupe @xmath1 , et on note @xmath6 le bord de lorbite @xmath20 : @xmath194 .
lensemble limite ne dpend pas du point base @xmath34 choisi . on va
maintenant dfinir une partie de lensemble limite appele ensemble limite radial , do nt on saura mieux contrler la dimension visuelle .
on dit quune partie @xmath70 dun espace mtrique @xmath13 est une * sous - quasi - godsique * sil existe une constante @xmath97 telle que lon ait @xmath195 pour tous @xmath32 , @xmath33 et @xmath40 dans @xmath25 tels que @xmath196 .
si lespace @xmath13 est godsique , alors dire quune partie @xmath25 est une sous - quasi - godsique revient dire quil existe une godsique do nt tout point de @xmath25 est distance borne .
soit @xmath13 un espace gromov - hyperbolique de point base @xmath34 et @xmath1 un semi - groupe disomtries de @xmath13 .
on appelle * ensemble limite radial * ( ou * ensemble limite conique * ) du semi - groupe @xmath1 , et on note @xmath197 lensemble : @xmath198 o @xmath84 est la relation dquivalence vue dans la dfinition du bord dun espace gromov - hyperbolique .
lensemble limite radial est une partie de lensemble limite qui ne dpend pas non plus du point base @xmath34 choisi .
[ auto - sim ] si @xmath1 est un semi - groupe de type fini disomtries dun espace gromov - hyperbolique , alors son ensemble limite est auto - similaire : @xmath199 on a clairement linclusion @xmath200 .
montrons lautre inclusion .
soit @xmath201 et soit @xmath202 une suite dlments de @xmath1 telle que lon ait @xmath203 ( ceci est une notation qui signifie @xmath204 ) + comme le semi - groupe @xmath1 est de type fini , il existe un gnrateur @xmath205 tel que lon ait une sous - suite @xmath206 . on a alors
@xmath207 on va associer chaque isomtrie @xmath209 dun espace mtrique @xmath13 , une partie @xmath208 de lespace @xmath13 qui correspond un domaine dans lequel llment @xmath209 contracte .
ceci nous servira dans toute la suite .
soit @xmath13 un espace mtrique .
pour @xmath145 , on dfinit un domaine @xmath210 par @xmath211 ] [ remxg ] lingalit @xmath212 est quivalente @xmath213 les lments de @xmath208 sont donc les points de @xmath13 qui sont plus prs de @xmath214 que de @xmath34 . le lemme suivant indique que lensemble @xmath208 est un domaine dans lequel @xmath209 contracte , et quil est de taille petite vue de @xmath34 quand llment @xmath209 est de grande norme .
[ lxg ] soit @xmath13 un espace mtrique .
pour tout @xmath145 , on a 1 .
@xmath215 , 2 .
si lespace @xmath13 est @xmath36-hyperbolique de point base @xmath34 , alors pour tout @xmath216 , on a @xmath217 1 . on a @xmath218 2 .
soient @xmath32 et @xmath33 dans @xmath219 .
par @xmath36-hyperbolicit , on a @xmath220 or , par dfinition de @xmath208 on a pour tout @xmath32 dans @xmath219
@xmath221 do le rsultat . on dit quune isomtrie @xmath145 dun espace mtrique @xmath13 est * contractante
* si les domaines @xmath208 et @xmath222 sont gromov - disjoints .
si une isomtrie @xmath209 est contractante , alors la partie @xmath223 est un domaine fondamental pour llment @xmath209 .
voici un critre de contraction . [ non_cont ] [ seph ] soit @xmath13 un espace @xmath36-hyperbolique de point base @xmath34 , et @xmath209 une isomtrie de @xmath13 . sil existe deux points @xmath32 et @xmath224 dans @xmath225 , tels que @xmath226 alors lisomtrie @xmath209 est contractante .
par contrapose , supposons que lisomtrie @xmath209 nest pas contractante .
les ensembles @xmath208 et @xmath227 ne sont alors pas gromov - disjoints .
soit alors @xmath228 et @xmath229 tels que @xmath230 .
soient aussi @xmath32 et @xmath224 deux points de lunion @xmath225 .
si les points @xmath32 et @xmath224 taient tous les deux dans @xmath219 ou tous les deux dans @xmath231 , on aurait lingalit @xmath232 par le lemme [ lxg ] . on peut donc supposer que lon a par exemple @xmath233 et @xmath234 .
par @xmath36-hyperbolicit et par le lemme [ lxg ] , on a alors @xmath235 voici une proprit des isomtries contractantes .
[ hn ] soit @xmath13 un espace mtrique propre , et soit @xmath236 une isomtrie contractante .
alors on a @xmath237 de plus , si lespace @xmath13 est @xmath36-hyperbolique , alors il existe un entier @xmath238 tel que pour tout @xmath239 , les parties @xmath240 et @xmath241 soient @xmath242-disjointes , o @xmath243 .
montrons que lon a @xmath244 .
pour cela , choisissons un rel @xmath47 tel que lon ait @xmath245 cela est possible puisque la partie @xmath246 est ouverte et contient @xmath34 .
pour tous entiers @xmath247 , les boules @xmath248 et @xmath249 sont disjointes .
en effet , quitte tout composer gauche par @xmath250 , on se ramne @xmath251 . on a alors @xmath252 .
ainsi , lorbite @xmath253 est une partie @xmath48-spare de @xmath13 .
par propret , on a donc bien @xmath244 .
montrons que les parties @xmath240 et @xmath241 sont @xmath242-disjointes .
par @xmath36-hyperbolicit , pour @xmath254 et @xmath255 , on a @xmath256 or , on a @xmath257 et @xmath258 , donc on a @xmath259 .
dautre part , par dfinition de @xmath240 , on a @xmath260 et de mme @xmath261 .
comme on a @xmath244 , quitte choisir @xmath67 assez grand on a @xmath262 on obtient donc lingalit @xmath263 .
les ensembles @xmath240 et @xmath241 sont alors disjoints , puisque sinon on aurait pour @xmath264 , par le lemme [ lxg ] , labsurdit @xmath265
nous voyons ici la dfinition et des proprits des parties contractantes . il sagit
densemble disomtries qui contractent toutes dans la mme direction , de faon contrle .
cette proprit est trs pratique , puisquelle est stable par produit , et permet de contrler la vitesse laquelle les produits disomtries tendent vers linfini .
cela permet par exemple de garantir que lensemble limite dun semi - groupe de type fini qui a cette proprit est radial ( proposition [ pradial ] ) .
soit @xmath13 un espace mtrique de point base @xmath34 . on dit
quune partie @xmath125 est * contractante * sil existe deux domaines gromov - disjoints @xmath266 et @xmath267 de @xmath13 tels que lon ait @xmath268 , et que lon ait @xmath269 .
cette dfinition dpend priori du point base @xmath34 choisi .
dans cette partie , nous donnons quelques rsultats sur les parties contractantes qui nous seront utiles par la suite .
voici quelques proprits simples des parties contractantes .
* proprits des parties contractantes * [ pptc ] + soit @xmath13 un espace mtrique . 1
. si @xmath209 est une isomtrie contractante , alors le singleton @xmath270 est contractant . [ cont_imp_cont ] 2 .
si une partie @xmath125 est contractante , alors le semi - groupe engendr ( cest - - dire lensemble des produits _ non vides _ dlments de @xmath25 ) lest aussi .
si une partie @xmath125 est contractante , alors linverse @xmath271 lest aussi .
1 . il suffit de prendre @xmath272 et @xmath273 .
si @xmath209 et @xmath274 sont deux lments dune partie contractante @xmath25 , alors on a @xmath275 3 .
si @xmath25 est une partie contractante pour des domaines @xmath267 et @xmath266 , alors la partie @xmath276 est contractante pour les domaines @xmath266 et @xmath267 respectivement .
en effet , on a pour tout @xmath177 , linclusion @xmath277 , qui se dduit de linclusion @xmath278 par contrapose .
la rciproque du point [ cont_imp_cont ] est fausse : une isomtrie @xmath209 telle que lensemble @xmath279 est contractant nest pas ncessairement contractante .
voici un critre de contraction . [ crit_cont ] soit @xmath13 un espace @xmath36-hyperbolique de point base @xmath34 , et soit une partie @xmath125 telle que lon ait @xmath280 alors la partie @xmath25 est contractante .
montrons que les domaines @xmath281 conviennent . on a dj bien
pour toute isomtrie @xmath177 , linclusion @xmath282 montrons maintenant que les domaines @xmath267 et @xmath266 sont gromov - disjoints .
soient @xmath283 et @xmath284 .
il existe alors deux isomtries @xmath209 et @xmath274 de @xmath13 tels que lon ait @xmath285 et @xmath286 . on a ensuite , par @xmath36-hyperbolicit , @xmath287 o @xmath288 . or , on a @xmath289 et @xmath290 par dfinition de @xmath227 et @xmath291 .
de plus , on a @xmath292 et @xmath293 par hypothse . on en
dduit que lon a @xmath294 pour vrifier que les domaines @xmath267 et @xmath266 sont bien gromov - disjoints , il ne reste donc plus qu montrer quils sont dadhrences dans @xmath13 disjointes . on a @xmath295
puisque lon a @xmath296 par le lemme [ lxg ] pour @xmath285 , et de mme @xmath297 pour @xmath286 .
les domaines @xmath266 et @xmath267 sont donc bien gromov - disjoints .
pour finir , lensemble @xmath298 contient bien le point base @xmath34 , puisquil contient la boule ouverte @xmath299 pour un rel @xmath47 assez petit .
voici un lemme qui permet de contrler la taille de limage du domaine @xmath267 par les lments dun semi - groupe contractant .
cela sera utilis dans la preuve du thorme [ thm2 ] .
[ cont1 ] soit @xmath13 un espace mtrique de point base @xmath34 , soit une partie @xmath300 et soit une isomtrie @xmath145 telles que lon ait @xmath301 alors pour tous @xmath32 et @xmath284 , on a lingalit @xmath302 soient @xmath32 et @xmath224 des lments de @xmath267 . on a @xmath303
\\ & = \frac{1}{2 } [ d(x , o ) + d(\gamma^{-1 } o , o ) - 2(\gamma^{-1 } o | x ) \\ & \qquad + d(\gamma^{-1 } o , o ) + d(o , x ' ) - 2(\gamma^{-1 } o | x ' ) - d(x , x ' ) ] \\ & = ( x | x ' ) - ( \gamma^{-1 } o | x ) - ( \gamma^{-1 } o | x ' ) + d(o , \gamma o ) \\ & \geq ( x | x ' ) + d(o , \gamma o ) - 2c \\ & \geq d(o , \gamma o ) - 2c .
\end{aligned}\ ] ] le lemme qui suit dit que lon a lgalit triangulaire une constante prs dans un semi - groupe contractant .
[ cet ] soit @xmath13 un espace mtrique de point base @xmath34 .
si @xmath25 est une partie contractante de @xmath12 , alors pour tous lments @xmath209 et @xmath274 de @xmath25 on a @xmath304 o @xmath305 . on a lgalit @xmath306 .
voici une proprit intressante des parties contractantes qui permettra de montrer que lensemble limite dun semi - groupe contractant de type fini est radial . [ cqg ] si @xmath25 est une partie contractante du groupe disomtries @xmath12 dun espace mtrique @xmath13 de point base @xmath34 et si @xmath307 est une suite de @xmath308 , alors lensemble @xmath309 est une sous - quasi - godsique
. montrons que pour @xmath310 , on a @xmath311 o @xmath305 . on se ramne @xmath312 , et lon conclut avec le lemme [ cet ] avec @xmath313 et @xmath314 .
le lemme qui suit dit que les isomtries dun semi - groupe contractant tendent vers linfini avec leur longueur quand lespace est propre .
[ cdv ] si @xmath25 est une partie contractante du groupe disomtries @xmath12 dun espace mtrique propre @xmath13 de point base @xmath34 et si @xmath307 est une suite de @xmath308 , alors on a @xmath315 lensemble @xmath316 est une partie spare de @xmath13 . en effet
, il existe un rel @xmath47 tel que la boule @xmath317 soit incluse dans @xmath318 , et ses images par les lments @xmath319 sont disjointes , puisque pour toute isomtrie @xmath209 de @xmath25 , on a linclusion @xmath320 .
le rsultat dcoule alors de la propret de lespace @xmath13 . [ pradial ]
soit @xmath13 un espace mtrique propre et @xmath1 un semi - groupe disomtries de @xmath13 , contractant et de type fini .
alors lensemble limite de @xmath1 est radial : @xmath321 soit @xmath19 un point de lensemble limite @xmath6 .
daprs la remarque [ auto - sim ] , il existe un gnrateur @xmath322 du semi - groupe @xmath1 tel que @xmath323 .
par rcurrence , il existe une suite @xmath324 telle que pour tout @xmath325 , on ait @xmath326 puis par le lemme [ cdv ] , on a @xmath327 . on a donc @xmath328 par le lemme [ cont1 ] appliqu lisomtrie @xmath329 et lensemble @xmath330 . or , le lemme [ cqg ] affirme que lensemble @xmath331 est une sous - quasi - godsique , donc le point @xmath19 est dans dans
lensemble limite radial : @xmath332 . on a montr linclusion @xmath333 .
lautre inclusion est claire . le fait que le semi - groupe @xmath1 soit de type fini est une hypothse ncessaire , comme le montre le contre - exemple suivant : soit @xmath1 le sous - semi - groupe de @xmath334 engendr par les matrices @xmath335 .
alors le semi - groupe @xmath1 est contractant , mais les ensembles limite et limite radial diffrent : @xmath336 .
en effet , lenveloppe convexe @xmath337 de lintervalle @xmath338 $ ] dans @xmath11 ( en identifiant le bord de @xmath11 @xmath193 de faon usuelle ) est envoye dans celle de lintervalle @xmath339 $ ] , et donc le semi - groupe est contractant
. mais le point @xmath8 est dans lensemble limite , et nest pas radial .
le but de cette section est de dmontrer le thorme : [ thm1 ] soit @xmath13 un espace gromov - hyperbolique propre bord compact , et soit @xmath1 un semi - groupe disomtries de @xmath13 do nt lensemble limite @xmath6 contient au moins deux points .
alors il existe un sous - semi - groupe @xmath340 de @xmath1 , contractant et de mme entropie : @xmath341 ce thorme est ltape principale de la preuve du thorme [ thm0 ] .
la preuve du thorme [ thm1 ] repose sur ltude du support du semi - groupe @xmath1 , qui est une partie ferme et @xmath342-invariante de @xmath343 dcrivant les directions dans lesquelles il y a beaucoup disomtries qui contractent et dilatent .
on montre que ce support , sil nest pas rduit deux points , fournit un sous - semi - groupe contractant dentropie @xmath15 .
et pour cela , on commence par montrer quil existe une isomtrie contractante dans le semi - groupe @xmath1 , et on effectue un ping - pong entre cet lment et une grosse partie du semi - groupe @xmath1 .
puis lon montre que dans le cas o le support est rduit deux points , le semi - groupe fixe un doublet de points au bord , ce qui permet de conclure rapidement si le semi - groupe ne fixe pas de point au bord .
enfin , on termine par le cas o le semi - groupe fixe un point au bord .
dans ce cas , on ne peut pas effectuer de ping - pong , mais lexistence dune isomtrie contractante nous permet de dmontrer quil existe une proportion suffisante des lments du semi - groupe qui contractent loin du point fixe . on fera
cela en utilisant le lemme des tiroirs sur une partition en copies dun domaine fondamental pour lisomtrie contractante .
+ nous allons montrer quun semi - groupe disomtries dun
espace gromov - hyperbolique do nt lensemble limite contient au moins deux points possde toujours une isomtrie contractante . on a la proposition : [ econt ] soit @xmath13 un espace gromov - hyperbolique , et @xmath1 un semi - groupe disomtries de @xmath13 do nt lensemble limite nest pas rduit un point .
alors @xmath1 contient une isomtrie @xmath344 contractante . de plus , si @xmath345 est un point du bord de @xmath13 , alors on peut choisir @xmath344 tel que @xmath346 .
lide de la preuve est de considrer deux grands lments de @xmath1 qui contractent en deux endroits distincts ( prs de points distincts de lensemble limite ) .
le produit de ces deux lment est alors contractant parce - que ces deux lments ntant pas contractants ( sinon il ny a rien dmontrer ) et ayant une grande norme , leur inverses contractent nouveau prs des mmes points , et ainsi , le produit contracte depuis lun des endroits vers lautre .
( voir figure [ fig_contract ] . )
et @xmath274 non contractantes . ]
supposons pour commencer que lensemble limite contienne
au moins trois points .
soit @xmath347 , et soient @xmath348 deux points distincts de lensemble limite @xmath6 , et qui sont distincts du point @xmath345 , et soient @xmath209 et @xmath274 deux isomtries de @xmath1 telles que lon ait @xmath349 on a alors galement @xmath350 et @xmath351 . + si une des isomtries @xmath209 ou @xmath274 tait contractante , alors elle conviendrait , et la preuve serait finie .
en effet , on a @xmath352 et @xmath353 , puisque par @xmath36-hyperbolicit on a @xmath354 et donc @xmath355 , et de mme avec @xmath274 .
on peut donc supposer que les lments @xmath209 et @xmath274 ne sont pas contractants . on a alors le lemme suivant .
les ensembles @xmath227 et @xmath291 sont @xmath356-disjoints .
soient @xmath285 et @xmath286 .
par @xmath36-hyperbolicit , on a @xmath357 or , par le lemme [ non_cont ] , on a @xmath358 , et de mme pour @xmath359 .
on conclut donc que lon a @xmath360 pour finir la preuve du lemme , il reste montrer que les parties @xmath227 et @xmath291 sont dadhrences dans @xmath13 disjointes .
par labsurde , supposons quil existe @xmath361 . on a alors dune
part @xmath362 par le lemme [ lxg ] , et dautre part @xmath363 par lingalit ci - dessus .
contradiction .
montrons alors que lisomtrie @xmath364 est contractante . on a @xmath365 et de mme @xmath366 , donc par le lemme ci - dessus @xmath367 or , on a @xmath368 . en effet , on a @xmath369 or , en utilisant le lemme ci - dessus avec @xmath209 et @xmath274 permuts , on obtient @xmath370 , et dautre part , on a @xmath371 .
on obtient donc bien @xmath372 et donc @xmath373 donc llment @xmath374 est contractant par le critre de contraction ( lemme [ seph ] ) .
pour finir , on a @xmath375 daprs ce que lon a fait ci - avant .
supposons maintenant que lensemble limite @xmath6 soit rduit deux points : @xmath376 par @xmath1-invariance de lensemble limite , les isomtries de @xmath1 fixent le doublet @xmath377 .
distinguons alors deux cas : 1 .
si toutes les isomtries de @xmath1 fixent chacun des points @xmath19 et @xmath345 , considrons une isomtrie @xmath209 telle que @xmath378 celle - ci existe bien puisque lon a @xmath201 . on a alors le lemme : + on a @xmath379 et @xmath380 .
+ commenons par montrer que lon a @xmath381 .
par @xmath36-hyperbolicit , on a @xmath382 et donc @xmath383 , ce qui prouve la premire partie du lemme .
+ on a ensuite @xmath384 , puisque lisomtrie @xmath209 fixe le point @xmath345 .
+ pour terminer la preuve de la proposition dans le cas o toutes les isomtries fixent chacun des points @xmath19 et @xmath345 , il ne reste plus qu dmontrer le lemme : + lisomtrie @xmath209 est contractante .
+ le point fixe @xmath19 est dans lunion @xmath225 , puisque si lon avait @xmath385 , alors on aurait @xmath386 , ce qui est absurde .
+ le fait que lisomtrie @xmath209 soit contractante dcoule alors du critre de contraction [ seph ] , puisque lon a @xmath387 2 .
sil existe une isomtrie @xmath388
qui change @xmath19 et @xmath345 , cest - - dire @xmath389 et @xmath390 .
+ la premire partie de la preuve permet dobtenir une isomtrie contractante @xmath391 , mais avec @xmath392 .
en outre , on peut choisir lisomtrie @xmath344 aussi grande que lon veut , et donc demander avoir lingalit @xmath393 + montrons qualors lisomtrie @xmath394 convient .
pour cela , nous allons utiliser le critre de contraction [ seph ] avec les points
@xmath19 et @xmath345 . on a le lemme suivant .
+ on a @xmath395 et @xmath396 .
+ montrons que @xmath395 . on a @xmath397 et par @xmath36-hyperbolicit , on a @xmath398 or ,
le point @xmath399 nest pas dans lensemble @xmath400 daprs le lemme [ lxg ] , puisque lon a @xmath401 . on a donc @xmath402 .
comme on a aussi @xmath392 , le lemme [ lxg ] donne @xmath403 ainsi , on obtient @xmath404 ce qui prouve que @xmath345 nest pas dans @xmath208 .
+ on montre de manire semblable que @xmath19 nest pas dans @xmath227 .
+ ainsi , on a @xmath405 et de mme @xmath380 .
le critre de contraction [ seph ] sapplique donc , puisque lon a bien lingalit @xmath406 dfinissons le support dun ensemble disomtries , qui correspond aux couples de points du bord pour lesquels il y a beaucoup disomtries qui contractent dans un voisinage du premier point du couple et dilatent depuis un voisinage du deuxime point du couple .
le support dun semi - groupe est relier au support de la mesure de patterson - sullivan , mais il est plus simple dfinir et manipuler .
soit @xmath13 un espace gromov - hyperbolique de point base @xmath34 .
on appelle * support * dune partie @xmath125 lensemble @xmath407 o lon a pos @xmath408 et @xmath409 on peut montrer que le support ne dpend pas du point base @xmath34 choisi .
[ nonvide ] soit @xmath13 un espace gromov - hyperbolique propre bord compact .
alors le support de toute partie @xmath410 est non vide
. soit @xmath411 tel que lespace @xmath13 soit @xmath36-hyperbolique .
construisons par rcurrence une suite @xmath412 de couples de points du bord telle que lon ait lgalit @xmath413 , et que lon ait @xmath414 pour @xmath415 , tous les points @xmath416 et @xmath417 conviennent puisque lon a @xmath418 .
supposons @xmath419 et @xmath420 construits tels que @xmath413 .
le carr @xmath343 du bord tant compact , il en va de mme du pav ferm @xmath421 , et lon peut extraire un recouvrement fini du recouvrement par les pavs ouverts @xmath422 quand @xmath423 dcrit @xmath424 .
le complmentaire dans @xmath424 de ce recouvrement est alors born . or , lentropie dune partie borne de @xmath13 est nulle : on a @xmath425 pour tout @xmath426 , par propret
. par le point [ hu ] des proprits [ ppteg ] de lentropie , il existe alors un couple @xmath427 de @xmath424 vrifiant @xmath428 la suite @xmath429 ainsi obtenue des pavs grossis de @xmath430 est alors dcroissante .
en effet , soit @xmath431 et soit @xmath432 .
par @xmath36-hyperbolicit on a alors @xmath433 ce qui donne bien linclusion @xmath434 . en faisant de mme avec @xmath345 , on obtient bien la dcroissance souhaite .
la suite converge donc vers un point @xmath423 .
ce point est bien dans le support @xmath435 , puisque pour tout @xmath47 , on peut trouver par gromov - hyperbolicit un entier @xmath67 assez grand tel que lon ait linclusion @xmath436 et donc tel que lon ait lingalit @xmath437 .
cest un des seuls endroits de la preuve o lon utilise la compacit de ladhrence de lespace @xmath13 ( lautre endroit o lon utilise cette compacit est pour dfinir la mesure de patterson - sullivan , voir [ paulin ] [ patt - sull ] ) .
cette hypothse nest pas beaucoup plus forte que de demander seulement la propret de lespace @xmath13 ( voir [ sbord ] ) .
voici une proprit de @xmath1-invariance du support : [ ginv ] soit @xmath13 un espace gromov - hyperbolique et soit @xmath1 un semi - groupe disomtries de @xmath13 .
le support @xmath435 est @xmath342-invariant pour laction de @xmath438 sur @xmath343 donne par @xmath439 pour toutes isomtries
@xmath209 et @xmath440 et @xmath441 .
cest - - dire que lon a @xmath442 pour tout @xmath443 et tout @xmath444 .
soient @xmath443 et @xmath444 .
montrons que lon a @xmath442 .
soit @xmath47 .
par le point [ hm ] des proprits [ ppteg ] , on a @xmath445 montrons que lon a linclusion @xmath446 soit @xmath447 . on a alors dune part @xmath448 et dautre part
@xmath449 de la mme faon on obtient les ingalits @xmath450 et donc @xmath451 . on a donc @xmath452 ,
et on a lautre ingalit @xmath453 par linclusion @xmath454 .
ceci prouve bien que le point @xmath455 est dans le support @xmath435 puisque le rel @xmath456 parcours @xmath457 quand @xmath48 parcours @xmath457 . on a linclusion @xmath458 ,
mais a - ton lgalit ? on va maintenant montrer que si le support
@xmath435 nest pas rduit certains sous - espaces de @xmath343 , alors on a la conclusion du thorme [ thm1 ] , cest - - dire lexistence dun sous - semi - groupe contractant de @xmath1 qui est dentropie totale @xmath15 . [ diag ] soit @xmath13 un espace gromov - hyperbolique , et soit @xmath1 un semi - groupe disomtries de @xmath13 .
si le support de @xmath1 nest pas inclus dans la diagonale de @xmath343 , alors il existe un sous - semi - groupe contractant @xmath340 de @xmath1 tel que @xmath459 .
soit @xmath460 avec @xmath461 . on peut alors
trouver un rel @xmath47 assez petit pour que le produit @xmath462 soit gromov - disjoint de la diagonale .
daprs le critre de contraction ( proposition [ crit_cont ] ) , lensemble @xmath463 est alors contractant pour @xmath67 assez grand , o lon a pos @xmath464 or , par dfinition du support et par le point [ hf ] des proprits [ ppteg ] de lentropie , on a @xmath465 le semi - groupe engendr par @xmath463 convient donc . [ hyp_cont ] soit @xmath13 un espace gromov - hyperbolique propre , et @xmath1 un semi - groupe disomtries de @xmath13 .
si le semi - groupe @xmath1 contient une isomtrie contractante @xmath344 , et que le support @xmath435 contient au moins trois points , alors il existe un sous - semi - groupe contractant @xmath340 de @xmath1 tel que @xmath459 .
lide de la preuve est de montrer quil y a beaucoup dlments du semi - groupe avec lesquels llment @xmath344 joue un ping - pong .
cela permet alors de faire contracter ces lments dun endroit prcis vers un endroit prcis en les composant gauche et droite avec h. on obtient alors un semi - groupe contractant en ne gardant que les lments assez grands .
pour une isomtrie contractante @xmath344 dun espace @xmath36-hyperbolique @xmath13 , dfinissons une partie @xmath466 de @xmath467 par @xmath468 montrons que si le support @xmath435 nest pas inclus dans @xmath469 , alors il existe un sous - semi - groupe contractant @xmath340 de @xmath1 , avec @xmath470 . lensemble @xmath466 contient @xmath471 .
il est un peu plus gros que @xmath471 afin de garantir que les lments @xmath472 assez grands et tels que lon ait @xmath473 soient tels que les parties @xmath227 et @xmath471 sont disjointes ( et idem dans lautre sens ) .
ceci permettra alors de faire le ping - pong .
[ ih ] soit @xmath13 un espace gromov - hyperbolique , soit @xmath344 une isomtrie contractante de @xmath13 et soient @xmath466 et @xmath474 les parties dfinies ci - dessus .
pour toute isomtrie @xmath145 telle que @xmath475 et @xmath476 , on a @xmath477 et pour toute isomtrie @xmath145 telle que @xmath475 et @xmath478 , on a @xmath479 soit @xmath145 tel que @xmath475 et @xmath476 .
montrons linclusion @xmath480 .
soit @xmath481 .
par @xmath36-hyperbolicit , on a @xmath482 or , par dfinition de @xmath483 et de @xmath208 , on a les ingalits @xmath484 do lingalit @xmath485 qui prouve que le point @xmath32 nest pas dans lensemble @xmath400 .
la deuxime partie du lemme dcoule de la premire en remplaant @xmath344 par @xmath486 et @xmath209 par @xmath487 .
posons @xmath488 lensemble des isomtries satisfaisant les conditions du lemme ci - dessus : @xmath489 faisons alors un ping - pong entre les isomtries de @xmath488 et lisomtrie @xmath344 pour obtenir une partie contractante
. soit @xmath13 un espace gromov - hyperbolique , soit @xmath344 une isomtrie contractante de @xmath13 et soit @xmath488 la partie de @xmath13 dfinie ci - dessus .
alors lensemble @xmath490 est une partie contractante de @xmath12 .
si @xmath209 est un lment de @xmath488 , alors par le lemme [ ih ] on a @xmath491 et lensemble @xmath492 contient bien le point base @xmath34 .
la partie @xmath493 est donc contractante pour les domaines @xmath494 et @xmath495 . .
] continuons la preuve de la proposition [ hyp_cont ] .
@xmath391 un lment contractant .
le semi - groupe engendr par @xmath496 est alors un sous - semi - groupe contractant de @xmath1 daprs le lemme ci - dessus . si le support @xmath435 nest pas inclus dans @xmath469 , alors lentropie de @xmath497 est gale @xmath15 par le point [ hf ] des proprits [ ppteg ] . or , par le point [ hm ] des proprits [ ppteg ] , lentropie de @xmath498 est gale celle de @xmath497 , et donc lentropie du semi - groupe engendr est aussi gale @xmath15 . on est donc
ramen ce que le support @xmath435 soit inclus dans @xmath469 .
par les lemmes [ hn ] et [ seph ] , comme lespace @xmath13 est propre , il existe un entier @xmath238 tel que pour tout @xmath239 , lisomtrie @xmath499 soit aussi une isomtrie contractante du semi - groupe @xmath1 .
on est alors mme ramen ce que le support @xmath435 soit inclus dans @xmath500 , pour tout @xmath239 . or , le diamtre des ensembles @xmath501 et @xmath502 tend vers @xmath8 : on a @xmath503 on est donc ramen ce que le support @xmath435 soit inclus dans un ensemble @xmath504 o @xmath505 et @xmath506 sont des points du bord @xmath90 . de plus , en utilisant la proposition [ diag ] , on est ramen au cas o le support @xmath435 est inclus dans la diagonale de @xmath343 , donc on est ramen ce que le support soit inclus dans le doublet @xmath507 .
ainsi , on a bien dmontr lexistence dun sous - semi - groupe contractant dentropie @xmath15 ds que le support @xmath435 contient au moins 3 points .
ceci termine la preuve de la proposition [ hyp_cont ] . ] on a le lemme : soit @xmath13 un espace gromov - hyperbolique et soit @xmath1 un semi - groupe disomtries de @xmath13 .
si le support @xmath435 est un singleton @xmath508 , alors le semi - groupe @xmath1 fixe le point @xmath19 .
cela dcoule de la proprit de @xmath342-invariance du support ( proposition [ ginv ] ) .
ainsi , on est ramen ce que le semi - groupe fixe un point au bord + ( voir sous - section [ ptfixe ] ) .
supposons maintenant que le support @xmath435 soit un doublet de points du bord @xmath509 . on a alors
le lemme suivant .
soit @xmath13 un espace gromov - hyperbolique et soit @xmath1 un semi - groupe disomtries de @xmath13 .
si le support @xmath435 est un doublet de points @xmath509 , alors le semi - groupe @xmath1 fixe le doublet @xmath510 .
cela dcoule de la proprit de @xmath342-invariance du support ( proposition [ ginv ] ) .
montrons maintenant que quitte multiplier par un lment qui change les deux points du doublet , si lon se restreint aux lments de norme assez grande et qui contractent dans un des deux sens ( dun des points du doublet vers lautre ) , alors on obtient un ensemble contractant . et lun de ces deux sous - semi - groupes engendrs sera forcment dentropie totale ( cest - - dire dentropie @xmath15 ) . + le lemme qui suit dit quun point fixe dune isomtrie est toujours dans son domaine de contraction ou dans son domaine de dilatation .
[ fixe ] soit @xmath13 un espace gromov - hyperbolique et @xmath209 une isomtrie de @xmath13 .
si @xmath511 est un point fixe pour lisomtrie @xmath209 ( i.e. @xmath512 ) , alors on a @xmath513 supposons que lon ait @xmath514
. alors on a @xmath515 .
en faisant de mme avec linverse @xmath487 , on conclut .
le lemme suivant dit quune isomtrie assez grande qui fixe deux points contracte de lun des points vers lautre ou bien change les deux points .
[ disj ] si une isomtrie @xmath209 dun espace @xmath36-hyperbolique @xmath13 de point base @xmath34 fixe un doublet de points du bord @xmath516 sans les changer ( i.e. @xmath517 ) et vrifie lingalit @xmath518 , alors on a @xmath519 comme llment @xmath209 fixe le doublet @xmath520 , et nchange pas @xmath505 et @xmath506 , les points
@xmath505 et @xmath506 sont fixes par @xmath209 .
le lemme [ fixe ] nous donne donc linclusion @xmath521 or , on ne peut pas avoir @xmath522 , puisque par le lemme [ lxg ] on a @xmath523 en faisant de mme avec @xmath208 , on obtient donc bien ce qui tait annonc .
dmontrons donc le thorme [ thm1 ] dans le cas o le semi - groupe @xmath1 fixe un doublet @xmath524 , mais ne fixe pas de point au bord .
considrons les parties suivantes du semi - groupe @xmath1 : @xmath525 @xmath526 on a alors le lemme suivant . [ cont+- ] les ensembles @xmath527 et @xmath528 sont contractants , + pour tout entier @xmath529 , o lon a pos @xmath530 pour une partie @xmath531 .
montrons que @xmath527 est contractant .
soient @xmath209 et @xmath274 deux isomtries de @xmath527 .
par @xmath36-hyperbolicit , on a @xmath532 or , par dfinition de @xmath291 , on a @xmath533 et on a de mme @xmath534 , puisque @xmath535 et @xmath536 . on en dduit
lingalit @xmath537 ainsi , on obtient @xmath538 et on a bien pour tout @xmath539 , @xmath540 donc par le critre [ crit_cont ] , lensemble @xmath527 est contractant .
de la mme faon , lensemble @xmath528 est contractant .
pour terminer la preuve du thorme [ thm1 ] dans ce cas , il ne reste donc plus qu dmontrer que lentropie dune de ces deux parties contractantes est @xmath15 : [ h+- ] on a @xmath541 , pour @xmath542 .
il y a deux cas : 1 .
il nexiste pas dlment qui change @xmath505 et @xmath543 .
dans ce cas , par le lemme [ disj ] , on a @xmath544 2 .
il existe un lment @xmath388 qui change @xmath505 et @xmath543 .
dans ce cas , on peut crire @xmath545 le rsultat dcoule alors des point [ hu ] , [ hm ] et [ hf ] des proprits [ ppteg ] de lentropie
. les lemmes [ cont+- ] et [ h+- ] donnent un semi - groupe contractant dentropie @xmath15 parmi lun des deux semi - groupes suivants : lun engendr par @xmath527 et lautre engendr par @xmath528 , pour @xmath67 assez grand .
ceci termine la preuve du thorme [ thm1 ] dans le cas o le semi - groupe @xmath1 ne fixe pas de point au bord .
supposons que le semi - groupe @xmath1 fixe un point @xmath104 mais ait un ensemble limite @xmath6 contenant au moins deux points .
par la proposition [ econt ] , il existe alors un lment contractant @xmath344 tel que le point fixe @xmath19 ne soit pas dans @xmath546 . on a alors la proposition suivante .
[ ptf ] soit @xmath13 un espace mtrique propre , soit @xmath1 un semi - groupe disomtries de @xmath13 , et soit @xmath344 une isomtrie contractante .
si lon pose @xmath547 alors on a lgalit @xmath459 .
ceci permettra de conclure grce au lemme suivant .
[ rcontf ] soit @xmath13 un espace gromov - hyperbolique , soit @xmath1 un semi - groupe disomtries
fixant un point @xmath104 , et soit @xmath344 une isomtrie contractante telle que @xmath548 .
alors il existe un rel @xmath62 tel que lensemble @xmath549 soit contractant .
le semi - groupe engendr sera alors encore contractant et aura encore pour entropie @xmath15 par les points [ hc ] et [ hf ] des proprits [ ppteg ] de lentropie , donc on aura bien obtenu la conclusion du thorme [ thm1 ] .
soit @xmath550 montrons que les rels @xmath551 conviennent .
pour toute isomtrie @xmath552 , le point @xmath19 est dans @xmath227 .
en effet , on a lingalit @xmath553 qui donne que @xmath19 nest pas dans @xmath554 , et donc on a @xmath555 par @xmath36-hyperbolicit , pour toutes isomtries @xmath209 et @xmath556 on a lingalit @xmath557 or , on a @xmath558 , donc on obtient lingalit @xmath559 par le critre de contraction [ crit_cont ] , on obtient donc que la partie @xmath560 est contractante .
pour dmontrer que lensemble @xmath340 est dentropie @xmath15 ( i.e. la proposition [ ptf ] ) , nous allons dcouper les boules @xmath561 en morceaux , selon les copies dun domaine fondamental pour llment @xmath344 , et montrer que quitte appliquer chaque morceau une puissance de llment @xmath344 , on peut ramener chaque morceau dans une partie proche de @xmath471 , tout en restant dans la boule .
le lemme suivant permettra de ramener chaque morceaux .
soit @xmath13 un espace mtrique , et soit @xmath344 une isomtrie contractante de @xmath13 .
alors il existe une constante @xmath562 telle que pour tout entier @xmath325 et pour tout rel @xmath426 on ait linclusion @xmath563 comme llment @xmath344 est contractant , il existe une constante @xmath97 telle que lon ait @xmath564 en particulier , pour tout entier @xmath565 et tout point @xmath566 , on a @xmath567 on a alors , @xmath568 do linclusion souhaite .
[ minxh ] soit @xmath13 un espace mtrique propre , soit @xmath1 un semi - groupe disomtries de @xmath13 , soit @xmath344 une isomtrie contractante de @xmath1 , et soit @xmath58 est une partie @xmath48-spare et couvrante de @xmath1 .
alors il existe une constante @xmath94 , telle que pour tout rayon @xmath426 il existe une partie @xmath48-spare @xmath569 telle que lon ait
lingalit @xmath570 ce lemme dit que lon a une proportion non ngligeable des lments de la boule @xmath571 qui sont dans le domaine @xmath471 .
pour le dmontrer , nous allons utiliser le lemme des tiroirs pour trouver un morceau de la boule qui contient beaucoup dlments , et ramener ce morceau par le lemme prcdent dans le domaine @xmath471 .
posons @xmath572 alors @xmath573 est un domaine fondamental pour llment @xmath344 qui est inclus dans @xmath400 .
majorons le nombre de morceaux du dcoupage de la boule @xmath574 par ce domaine fondamental .
il existe une constante @xmath575 telle que pour tout @xmath576 et tout @xmath577 , on ait @xmath578 si @xmath579 , par le lemme [ cont1 ] ( pour @xmath580 et @xmath581 ) , on a @xmath582 or , par les lemmes [ hn ] et [ cet ] , il existe des constantes @xmath583 et @xmath584 telles que @xmath585 il existe donc une constante @xmath575 telle que pour tout @xmath586 et @xmath587 on ait @xmath588 do @xmath589 pour @xmath590 avec @xmath587
. laide dun domaine fondamental pour une isomtrie contractante @xmath344 . ]
daprs ce sous - lemme , on peut donc partitionner la boule @xmath574 en @xmath591 morceaux : @xmath592 le lemme des tiroirs nous donne alors que lun des morceaux du dcoupage que lon obtient pour @xmath593 est de cardinal au moins @xmath594 .
si le morceau en question est @xmath595 ou @xmath596 , alors le rsultat est clair : @xmath597 convient .
sinon , le morceau est @xmath598 pour un entier @xmath599 .
par le lemme prcdent , on a alors @xmath600 pour une constante @xmath601 assez grande . et par ingalit triangulaire , on a @xmath602 comme on a linclusion @xmath603 ,
les ingalits prcdentes donne lingalit annonce @xmath604 avec @xmath605 , pour tout @xmath586 , pour une constante @xmath97 assez grande .
voici maintenant un lemme dinversion des quantificateurs .
[ ich ] soit @xmath13 un espace mtrique propre , soit @xmath606 , et soient @xmath607 , @xmath94 et @xmath47 des rels .
si pour tout rel @xmath586 , il existe une partie @xmath48-spare @xmath608 telle que lon ait lingalit @xmath609 alors on a @xmath610 .
soit @xmath611 une partie spare et @xmath612-couvrante . on a les ingalits @xmath613 soit @xmath58 une partie @xmath48-spare et couvrante du semi - groupe @xmath1 . en utilisant le lemme [ minxh ] , on obtient lingalit @xmath614 pour une partie @xmath615 @xmath48-spare de @xmath340 , pour tout @xmath576 et pour une constante @xmath97 .
comme @xmath58 est une partie couvrante de @xmath1 , son exposant critique est minor par @xmath616 , ce qui donne @xmath617 avec @xmath618 .
le lemme [ ich ] nous donne alors lingalit @xmath619 , puis on obtient lingalit @xmath620 en faisant tendre @xmath621 vers linfini . lautre ingalit @xmath622 se dduit de linclusion @xmath623 .
ceci termine la preuve du thorme [ thm1 ] : tous les cas ont t traits , puisque le support est non vide par la proposition [ nonvide ] .
sans lhypothse @xmath624 , les thormes [ thm1 ] et [ thm0 ] sont faux en gnral .
[ contre - ex ] pour @xmath68 muni de sa mtrique usuelle , le sous - semi - groupe de @xmath2 engendr par les matrices @xmath625 et @xmath626 a pour entropie @xmath69 , mais ne contient que des sous - semi - groupes contractants dentropie nulle , donc en particulier ne contient que des sous - semi - groupes de schottky dexposant critique nul .
mme chose avec le groupe parabolique engendr par la matrice @xmath625 . . ]
les semi - groupes de schottky sont les semi - groupes ayant la dynamique la plus simple , puisque par dfinition leurs gnrateurs jouent au ping - pong . en particuliers ils
sont libres et spars .
le thorme principal de cette section , que nous dmontrons ici ( thorme [ thm2 ] ) , affirme que lentropie dun semi - groupe est approche aussi prs que lon veut par celle de ses sous - semi - groupes de schottky .
soient @xmath13 un espace mtrique , et @xmath1 un semi - groupe disomtries de @xmath13 .
on dit que le semi - groupe @xmath1 est * de schottky * pour une partie @xmath627 , sil admet une partie gnratrice finie @xmath628 , telle que les parties @xmath629 , @xmath630 , ... , @xmath631 et @xmath632 soient deux deux gromov - disjointes , et que lensemble @xmath633 soit dintrieur non vide .
et @xmath634 soit @xmath13 un espace mtrique et soit @xmath1 un semi - groupe de schottky disomtries de @xmath13 . on a alors : 1 .
le semi - groupe @xmath1 est libre .
le semi - groupe @xmath1 est spar ( donc en particulier dorbite discrte ) .
3 . le semi - groupe @xmath1 est contractant .
4 . si lespace @xmath13 est propre , on a @xmath635 .
soit @xmath95 une boule incluse dans lensemble @xmath633 , alors ses images par les isomtries de @xmath1 sont toutes deux deux disjointes .
donc le semi - groupe est libre et spar .
le semi - groupe @xmath1 est contractant pour les domaines @xmath632 et @xmath636 , en choisissant @xmath637 .
pour obtenir la finitude de lexposant critique , il suffit dutiliser le lemme [ cdv ] pour obtenir un entier @xmath638 tel que pour tous gnrateurs @xmath639 , ... ,
@xmath640 , on ait @xmath641 o @xmath642 est la constante de gromov - disjonction du semi - groupe contractant @xmath1 .
puis on utilise le lemme [ cet ] pour obtenir la minoration @xmath643 pour tout @xmath67 , et pour des gnrateurs @xmath639 , ... , @xmath644 .
et on obtient alors la majoration @xmath645 , o @xmath646 est le nombre de gnrateurs du semi - groupe @xmath1 . [ thm2 ] soit @xmath13 un espace gromov - hyperbolique propre , et soit @xmath1 un semi - groupe disomtries de @xmath13 .
si le semi - groupe @xmath1 est contractant , alors pour tout @xmath47 , il existe un sous - semi - groupe schottky @xmath623 tel que @xmath647 . pour construire un gros sous - semi - groupe de schottky , nous considrerons une partie suffisamment spare de ce semi - groupe contractant , et montrerons que les lments de norme donne assez grande , contractent des endroits suffisamment carts les uns des autres pour jouer au ping - pong et donc engendrer un semi - groupe de schottky .
+ [ sc ] soit @xmath13 un espace gromov - hyperbolique , et soient @xmath266 et @xmath267 des parties de @xmath13 .
alors il existe un rel @xmath62 et un entier @xmath238 tels que pour tout semi - groupe contractant @xmath1 disomtries de @xmath13 pour les parties @xmath267 et @xmath266 , pour toute partie @xmath58 @xmath62-spare de @xmath1 , et pour tout entier @xmath239 , lensemble @xmath648 engendre un semi - groupe de schottky pour le domaine @xmath267 , o lon a pos @xmath649 supposons quil existe un semi - groupe contractant pour les parties @xmath266 et @xmath267 ( sinon il ny a rien dmontrer ) . montrons que le rel @xmath650 convient , o @xmath651 et @xmath36 est un rel tel que lespace @xmath13 soit @xmath36-hyperbolique .
soit @xmath1 un semi - groupe contractant pour les parties @xmath266 et @xmath267 , et soient @xmath209 et @xmath274 deux isomtries de @xmath652 qui vrifient lingalit @xmath653 montrons qualors les domaines @xmath654 et @xmath655 sont gromov - disjoints . soient @xmath656 et @xmath657 .
par gromov - hyperbolicit , on a alors @xmath658 or , on a lingalit @xmath659 et daprs le lemme [ cont1 ] , on a les ingalits @xmath660 et de mme @xmath661 . on obtient donc lingalit @xmath662 les ensembles @xmath654 et @xmath655 sont alors disjoints , puisque si lon avait @xmath663 , on aurait labsurdit @xmath664 on a donc montr que les ensembles @xmath654 et @xmath655 sont @xmath665-disjoints . pour finir , lintersection @xmath666 est dinterieur non vide si lentier @xmath67 est assez grand , et elle est incluse dans @xmath667 .
ainsi , pour @xmath67 assez grand et pour le rel @xmath62 donn par le lemme , si lon considre une partie @xmath58 @xmath62-spare et couvrante de @xmath1 , alors lensemble @xmath668 engendre un semi - groupe de schottky . et par les points [ sep ]
et [ couv ] des proprits [ ppteg ] et la remarque [ delta - ann ] on a @xmath669 pour tout @xmath47 , on peut donc trouver un entier @xmath67 tel que @xmath670 montrons alors que le semi - groupe de schottky @xmath340 engendr par @xmath668 a un exposant critique suprieur ou gal @xmath671 .
pour cela , on va utiliser le lemme suivant .
[ mnec ] soit @xmath13 un espace mtrique . si une partie @xmath58 du groupe disomtries
@xmath12 engendre un semi - groupe libre @xmath1 , alors on a la minoration de lexposant critique : @xmath672 o @xmath673 .
cela dcoule de lingalit triangulaire .
pour des gnrateurs @xmath674 , on a @xmath675 donc , par libert du semi - groupe @xmath1 pour la partie @xmath58 , on obtient @xmath676 on a donc @xmath677 en appliquant le lemme [ mnec ] la partie @xmath668 , qui engendre un semi - groupe qui est de schottky et qui est donc libre , on obtient lingalit @xmath678 or , lentier @xmath67 pouvait tre choisi arbitrairement grand , et le rel @xmath47 est arbitraire .
ceci achve la preuve du thorme [ thm2 ] .
+ on peut maintenant facilement dmontrer le thorme [ thm0 ] .
on dduit aisment des thormes [ thm1 ] et [ thm2 ] lingalit @xmath679 lautre ingalit sobtient en remarquant quun semi - groupe de schottky est spar .
dans cette section , nous voyons une application du thorme [ thm0 ] ltude au bord dun semi - groupe .
nous obtenons le corollaire [ cpaulin _ ] ci - aprs qui est une gnralisation dun rsultat de f. paulin ( voir @xcite ) qui gnralise lui - mme un rsultat de coornaert ( voir @xcite ) .
+ soit @xmath13 un espace gromov - hyperbolique .
pour dfinir ce quest la dimension visuelle dune partie @xmath680 du bord @xmath90 , introduisons quelques notations .
on dfinit la * boule @xmath681 * de centre @xmath19 et de rayon @xmath62 sur le bord @xmath90 par @xmath682 on appelle * mesure visuelle * de dimension @xmath171 dune partie @xmath683 du bord dun espace @xmath13 gromov - hyperbolique , le rel @xmath684 o @xmath685 est la borne infrieure des sommes @xmath686 sur tous les recouvrements @xmath687 de lensemble @xmath680 par des boules de rayons @xmath688 . on appelle * dimension visuelle * dun ensemble @xmath683 le rel @xmath689 on a aussi @xmath690 la mesure visuelle est une mesure .
la notion de dimension visuelle gnralise celle de dimension de hausdorff . on a le rsultat suivant .
[ cpaulin _ ] soit @xmath13 un espace gromov - hyperbolique propre bord compact et soit @xmath1 un semi - groupe disomtries de @xmath13 do nt lensemble limite contient
au moins deux points .
alors on a lgalit @xmath691 f. paulin a nonc ce rsultat pour les groupes discret , et sa preuve semble sadapter aux semi - groupes .
cependant , il fait des hypothses supplmentaires par rapport notre preuve , qui sont le fait que lespace @xmath13 soit godsique , quil soit quasi - godsique , que le semi - groupe soit spar , et quil ne fixe pas de point au bord ( voir @xcite ) .
+ voici lingalit facile entre entropie et dimension visuelle de lensemble limite radial . [ if ] soit @xmath13 un espace gromov - hyperbolique propre , et soit @xmath1 un semi - groupe disomtries de @xmath13 . on a lingalit @xmath692
soit @xmath58 une partie spare et couvrante de @xmath1 .
montrons que lon a lingalit @xmath693 .
comme on a les galits @xmath694 et @xmath695 , ceci donnera bien lingalit souhaite .
dfinissons l*ombre * dune boule @xmath102 par @xmath696 on a alors linclusion @xmath697 o @xmath698 .
en effet , si un lment @xmath19 est dans lensemble limite radial @xmath699 , alors il existe un rel @xmath153 et une partie @xmath25 de @xmath700 qui est une @xmath62-sous - quasi - godsique telle que @xmath701 . on a alors pour tout @xmath702 , @xmath703 , et pour tout @xmath325 , @xmath704 .
posons alors @xmath705 pour un rel @xmath706 et montrons que pour @xmath707 , on a @xmath708 . on peut recouvrir chaque ombre par une boule de rayon @xmath709 .
en effet , soient @xmath19 et @xmath710 deux points de lombre @xmath711 . on a alors @xmath712 par @xmath36-hyperbolicit ,
et par lingalit @xmath713 et de mme avec @xmath710 .
ainsi , pour @xmath47 , en considrant un recouvrement de lensemble @xmath714 par des boules de rayon @xmath715 qui recouvrent les ombres @xmath716 pour @xmath717 assez grand , on obtient @xmath718 o @xmath719 est la srie de poincar de @xmath58 . on a @xmath720 ds que @xmath721 , do @xmath722 on a ensuite @xmath723 pour tout @xmath721 , puis @xmath724 ainsi , on a @xmath725 pour tout @xmath721 ,
do lingalit @xmath726 . pour obtenir le corollaire [ cpaulin _ ] partir du thorme [ thm0 ] , il suffit de dmontrer le rsultat dans le cas des semi - groupes de schottky : [ pschottky ] soit @xmath13 un espace gromov - hyperbolique propre bord compact , et soit @xmath1 un semi - groupe de schottky disomtries de @xmath13 .
alors on a lgalit @xmath727 par les propositions [ if ] et [ pradial ] , on a dj lingalit @xmath728 montrons lingalit @xmath729 .
pour cela , on va utiliser le lemme suivant , d frostman , qui ramne le problme construire une mesure convenable sur lensemble limite @xmath6 du semi - groupe @xmath1 .
[ mnd ] soit @xmath13 un espace gromov - hyperbolique et soit @xmath730 une probabilit porte par une partie @xmath680 du bord @xmath90 .
sil existe un rel @xmath171 et une constante @xmath94 tels que lon ait @xmath731 pour toute boule @xmath681 du bord @xmath90 , alors on a lingalit @xmath732 soit @xmath47 , et soit @xmath621 un recouvrement de lensemble @xmath680 par des boules de tailles infrieures @xmath48 . on a alors
les ingalits @xmath733 on en dduit , en passant la borne infrieure sur tous ces recouvrements que lon a lingalit @xmath734 , puis en passant la limite quand @xmath48 tend vers 0 , que lon a @xmath735 . on obtient donc bien lingalit souhaite .
la mesure laquelle nous appliquerons ce lemme pour conclure est la mesure @xmath730 de patterson - sullivan , que nous allons dfinir maintenant .
soit @xmath13 un espace gromov - hyperbolique propre bord compact et soit @xmath1 un semi - groupe discret disomtries de @xmath13 , avec @xmath635 .
dfinissons des probabilits @xmath736 sur lespace @xmath13 , pour des rels @xmath737 , par @xmath738 o @xmath739 est la srie de poincar de @xmath1 , et @xmath740 est le dirac en @xmath32 .
la srie de poincar @xmath129 diverge pour @xmath741 et converge pour @xmath737 .
pour dfinir la mesure @xmath730 , nous aurons besoin que la srie de poincar soit divergente en @xmath3 ( i.e. @xmath742 ) . on la rend
divergente grce au lemme suivant .
[ lpat ] soit @xmath743 un rel et @xmath744 une suite de rels positifs .
si la srie de dirichlet @xmath745 est divergente pour @xmath746 et convergente pour @xmath747 , alors il existe une fonction croissante @xmath748 telle que la srie @xmath749 converge pour @xmath747 et diverge pour @xmath746 et pour @xmath750 , et avec de plus la proprit : pour tout @xmath47 , il existe un rel @xmath751 tel que pour @xmath752 et @xmath753 , on ait @xmath754 voir @xcite pour une preuve .
pour rendre la srie de poincar divergente en @xmath755 , il suffit de la remplacer par : @xmath756 o @xmath638 est la fonction fournie par ce lemme , et lon fait de mme pour la dfinition des mesures @xmath736 .
les mesures @xmath736 sont des mesures de probabilits . or , par hypothse , ladhrence @xmath93 est compacte .
et lensemble de probabilits @xmath757 sur le compact @xmath93 , muni de la convergence vague , est alors compact ( voir par exemple @xcite ) . il existe donc une suite de rels @xmath758 , avec pour tout @xmath638 , @xmath759 , telle que la suite de mesures @xmath760 converge vaguement vers
une mesure de probabilit @xmath730 : @xmath761 { } \delta_\gamma \quad \text{et } \quad \mu_{s_k } \xrightharpoonup[k \to \infty ] { } \mu.\ ] ] la mesure @xmath730 est alors porte par le bord @xmath90 , puisque @xmath1 est une partie discrte , que lespace @xmath13 propre et que lon a @xmath762 .
avant de majorer la mesure @xmath730 sur toutes les boules , majorons l sur les ensembles @xmath654 .
[ mmpn ] soit @xmath13 un espace gromov - hyperbolique propre bord compact , et soit @xmath1 un semi - groupe disomtries de @xmath13 , de schottky pour un domaine @xmath267 .
alors il existe un point @xmath34 tel que si @xmath730 est la mesure de patterson - sullivan dfinie ci - dessus ( pour ce point @xmath34 ) , alors il existe une constante @xmath94 telle que pour toute isomtrie @xmath472 , on ait lingalit @xmath763 on a le lemme suivant . [ pto ] soit @xmath13 un espace mtrique , et @xmath1 un sous - semi - groupe de schottky de @xmath12 , pour une partie @xmath764 . pour un point @xmath765 , on a les quivalences @xmath766 @xmath767 pour toutes isomtries @xmath209 et @xmath768 .
montrons la premire quivalence .
soient @xmath209 et @xmath274 deux isomtries de @xmath1 telles que lon ait @xmath769 .
soient @xmath205 et @xmath770 les gnrateurs tels que @xmath771 et @xmath772 .
tant donn que les ensembles @xmath773 et @xmath774 sont gromov - disjoints si @xmath775 et que lon a les inclusions @xmath776 et @xmath777 , on a ncessairement @xmath778 .
par rcurrence , on a bien obtenu que @xmath779 .
la rciproque est claire .
montrons la deuxime quivalence .
@xmath209 et @xmath274 deux isomtries de @xmath1 telles que lon ait @xmath780 . on a @xmath781 puisque @xmath782 .
par lquivalence prcdente , on a donc @xmath783 ou @xmath784 .
supposons que lon ait @xmath785 . on peut alors
trouver un gnrateur @xmath205 tel que lon ait @xmath786 . on a ensuite linclusion @xmath787 ,
donc @xmath788 , puis @xmath789 , ce qui contredit lhypothse .
donc on a bien @xmath783 .
la rciproque est claire .
choisissons un point @xmath765 ( i.e. comme dans le lemme [ pto ] ) .
le semi - groupe @xmath1 soit divergent .
pour tout @xmath737 et pour toute isomtrie @xmath472 , on a alors @xmath790 or , le lemme [ cet ] nous donne lingalit @xmath791 o @xmath601 est la constante de contraction du semi - groupe @xmath1 pour le point @xmath34 : @xmath792 on obtient alors @xmath793 do lingalit @xmath794 avec @xmath795 . en passant la limite , on obtient lingalit voulue @xmath796 o @xmath797 , puisque lensemble @xmath798 est isol ( cest - - dire ouvert et ferm ) dans lensemble limite @xmath6 .
si le semi - groupe nest pas divergent , on modifie le calcul prcdent en consquent en utilisant lastuce de patterson , et on conclut de la mme faon .
montrons maintenant que la mesure @xmath730 est majore pour toutes les boules .
[ mmcg ] soit @xmath13 un espace gromov - hyperbolique propre bord compact , et soit @xmath1 un semi - groupe disomtries de @xmath13 , de schottky pour un domaine @xmath267 .
soit @xmath34 le point donn par le lemme [ pto ] , et soit @xmath730 la mesure de patterson - sullivan correspondante .
alors il existe une constante @xmath94 telle que pour toute boule @xmath681 du bord @xmath90 , on ait lingalit
@xmath799 soit @xmath601 la constante donne par le lemme [ mmpn ] .
soit @xmath681 une boule .
si la boule nintersecte pas lensemble limite @xmath6 , on a @xmath800 , et il ny a rien dmontrer .
supposons donc que la boule rencontre lensemble limite . on peut alors
supposer que lon a @xmath201 , quitte recouvrir la boule @xmath681 par une boule de rayon @xmath801 centre en un point de lensemble limite @xmath6 , et multiplier la constante @xmath97 par @xmath802 .
notons @xmath803 et @xmath804 .
soit @xmath67 un entier tel que lon ait @xmath805 @xmath806 cet entier existe bien , puisque pour @xmath807 on a @xmath808 , et puisque lon a @xmath809 en effet , si lon a @xmath810 { } \xi$ ] pour une suite @xmath307 dlments de @xmath1 , alors on a @xmath811 partir dun certain rang par le lemme [ cont1 ] et par gromov - hyperbolicit .
soit @xmath812 tel que @xmath813 .
comme la mesure @xmath730 est porte lensemble limite @xmath6 , on a @xmath814 par le lemme [ mmpn ] et par le choix de @xmath67 .
il reste donc majorer la quantit @xmath815 en fonction de @xmath62 .
pour tout @xmath816 tel que @xmath817 , on a @xmath783 .
soient alors @xmath775 deux gnrateurs tels que lon ait @xmath818 soit @xmath819 une constante telle que les parties @xmath773 pour @xmath205 parcourant les gnrateurs , et @xmath267 , soient deux deux @xmath820-gromov disjointes .
limage @xmath821 de la boule @xmath681 par lisomtrie @xmath487 rencontre les ensembles @xmath822 et @xmath823 , donc il existe des points @xmath345 et @xmath824 de @xmath821 tels que lon ait lingalit @xmath825 on a alors @xmath826 do lingalit @xmath827 on obtient donc lingalit escompte avec @xmath828 .
les lemmes [ mmcg ] et [ mnd ] donnent lingalit @xmath829 ce qui termine cette preuve de la proposition [ pschottky ] .
on peut maintenant facilement retrouver la gnralisation du rsultat de paulin .
soit @xmath47 .
par le thorme [ thm0 ] , il existe un sous - semi - groupe @xmath340 de schottky de @xmath1 tel que lon ait @xmath647 .
par les propositions [ pschottky ] et [ pradial ] , on a donc les ingalits @xmath830 ceci tant vrai pour tout @xmath47 , on en dduit lingalit
@xmath831 lautre ingalit est donne par la proposition [ if ] .
dans cette sous - section , nous obtenons une application du corollaire [ cpaulin _ ] aux semi - groupes de dveloppement en base @xmath21 .
le * semi - groupe de dveloppement en base @xmath832 avec ensemble de chiffres @xmath25 * est le semi - groupe engendr par les applications affines : @xmath833 o @xmath24 , pour une partie finie @xmath25 de @xmath22 .
on peut voir ce semi - groupe comme un sous - semi - groupe de @xmath834 . en effet
, lapplication @xmath835 , on peut associer la matrice @xmath836 , o @xmath837 est une racine carre de @xmath21 . on a donc une action par isomtrie du semi - groupe sur lespace @xmath838 ( vu comme partie de lensemble des quaternions ) ,
do nt le bord @xmath90 sidentifie @xmath839 .
laction de lapplication @xmath840 sur @xmath841 est donne par @xmath842 lensemble limite du semi - groupe est alors exactement lensemble des nombres complexes qui admettent un dveloppement @xmath21-adique nayant quun seul chiffre avant la virgule et avec ensemble de chiffres @xmath25 .
tout ceci fonctionne galement en remplaant le corps @xmath22 par @xmath193 .
on appelle * nombre de salem * gnralis un entier algbrique @xmath832 de module strictement suprieur @xmath9 , do nt tous les conjugus sont de modules infrieurs ou gaux @xmath9 , sauf ventuellement son conjugu complexe . on appelle * nombre de pisot * gnralis un entier algbrique @xmath832 de module strictement suprieur @xmath9 , do nt tous les conjugus sont de modules _ strictement _ infrieurs @xmath9 , sauf ventuellement son conjugu complexe .
dans la dfinition classique de nombres de pisot et de salem , on demande ce que le nombre soit un rel @xmath843 , mais tout ce que lon verra est valable pour cette dfinition plus gnrale .
[ psalem ] soit @xmath1 le semi - groupe engendr par les applications @xmath844 o @xmath24 pour une partie finie @xmath845 .
si @xmath21 est un nombre de salem gnralis , alors on a lgalit @xmath846 ( le nombre dor , qui est un nombre de pisot ) , avec ensemble de chiffres @xmath847 . ]
avant de dmontrer la proposition , on a le lemme suivant .
+ le lemme qui suit dit que si lon regarde lorbite dune boule par le semi - groupe , alors le nombre de chevauchements en un point donn nest pas trop grand par rapport la distance au point base @xmath34 .
[ poly ] sous les hypothses de la proposition [ psalem ] , il existe un entier @xmath62 tel que lon ait @xmath848 pour @xmath849 .
[ non - sep ] si lentier algbrique @xmath21 est de pisot , alors le semi - groupe @xmath1 est mme spar ( voir la condition de sparation de lalley @xcite ) , et donc on peut prendre @xmath850
. le resultat suivant permet de majorer le paramtre de translation des isomtries du semi - groupe @xmath1 .
[ tb ] si @xmath1 est un sous - semi - groupe de @xmath851 engendr par des application @xmath840 , pour @xmath24 , avec @xmath25 partie finie de @xmath22 et @xmath832 tel que @xmath852 , alors il existe une constante @xmath97 telle que pour toute application @xmath853 , on ait @xmath854 . un lment du semi - groupe @xmath1 scrit @xmath855 avec @xmath856 . on a alors la majoration @xmath857 notons @xmath858 on a alors les resultats suivants .
[ lm ] il existe une constante @xmath97 telle que pour tout @xmath849 , toute isomtrie de @xmath859 est de longueur au moins @xmath860 et au plus @xmath861 en les gnrateurs . en effet , on a lingalit triangulaire @xmath862 par ailleurs , si lon crit @xmath863 , on a @xmath864 o @xmath67 est la longueur de @xmath209 .
ensuite , daprs le sous - lemme [ tb ] , la quantit @xmath865 est borne par une constante @xmath601 indpendante de @xmath209 et de @xmath32 .
et on vrifie que lon a @xmath866 pour
la mtrique usuelle de @xmath841 . on obtient alors
lencadrement @xmath867 do lencadrement sur la longueur @xmath67 de @xmath209 avec @xmath868 .
posons @xmath869 la plus grande longueur possible des lments de @xmath859 . on a alors
le resultat suivant .
[ bplus ] il existe une constante @xmath97 telle que pour tout @xmath849 on ait @xmath870 o @xmath8 est le point du bord @xmath871 .
lapplication @xmath872 tant une isomtrie , on a @xmath873 pour tout @xmath874 .
dautre part , si lon crit @xmath875 , alors on a @xmath876 o @xmath67 est la longueur de @xmath209 . or , par le sous - lemme [ lm ] , il existe une constante @xmath97 telle que pour @xmath874 , on ait @xmath877 , o @xmath67 est la longueur de @xmath209 .
la distance @xmath878 est donc borne indpendamment de @xmath32 et @xmath874 , par ingalit triangulaire .
quitte multiplier la partie @xmath25 par les dnominateurs ( ce qui ne change pas la conclusion du sous - lemme ) , on peut supposer que lon a @xmath879 $ ] .
la quantit @xmath880 est alors un polynme en @xmath21 coefficients entiers , pour tout lment @xmath874 .
construisons alors un espace @xmath881 ( indpendant du point @xmath32 ) , dans lequel lanneau @xmath882 $ ] sera discret .
soit @xmath883 lensemble des valeurs absolues archimdiennes du corps @xmath884 , quivalence prs .
lensemble @xmath883 est fini ( de cardinal major par le degr de @xmath21 ) , et lon peut poser @xmath885 o @xmath886 est le complt du corps @xmath638 pour la valeur absolue @xmath887 . on a alors @xmath888 , o @xmath62 est le nombre conjugus rels de @xmath21 ,
et @xmath889 est son nombre de conjugus complexes . on a maintenant
le rsultat suivant , [ discret ] lanneau @xmath882 $ ] est discret dans lespace @xmath881 .
qui dcoule de la formule du produit , qui est un rsultat classique de thorie des nombre ( voir par exemple @xcite ) . on peut
mme montrer que @xmath882 $ ] est un rseau co - compact de lespace @xmath881 , mais nous nen aurons pas besoin .
soit @xmath638 un corps de nombres .
pour tout @xmath890 , on a @xmath891 o @xmath892 est lensemble des valeurs absolues de @xmath638 quivalence prs , o lon a choisis les valeurs absolues standards dans chaque classe dquivalence .
il suffit de montrer que le point @xmath893 $ ] est isol . on aura alors
bien la discrtude puisque lensemble @xmath882 $ ] est un groupe additif .
soit @xmath95 la boule de @xmath881 de centre @xmath8 et de rayon @xmath894 .
si un point @xmath32 est dans @xmath895 $ ] , alors on a @xmath896 puisque pour toute valeur absolue ultramtrique @xmath887 , on a @xmath897 et donc @xmath898 , tant donn que @xmath21 est un entier algbrique .
daprs la formule du produit , on a donc @xmath899 .
do la discrtude de lanneau @xmath882 $ ] dans lespace @xmath881 .
lensemble de valeurs absolues @xmath883 peut scrire @xmath900 o * @xmath901 est lensemble des valeurs absolues @xmath902 telles que @xmath903 , * @xmath904 est lensemble des valeurs absolues @xmath902 telles que @xmath905 , * @xmath906 est lensemble des valeurs absolues @xmath902 telles que @xmath907
. on peut alors dcomposer cet espace @xmath881 dans lequel lanneau @xmath882 $ ] est discret en 3 morceaux : @xmath908 o @xmath909 , @xmath910 , et @xmath911 .
le nombre @xmath21 tant de salem gnralis , il existe une unique valeur absolue @xmath887 telle que @xmath903 . on a donc @xmath912
ou @xmath22 selon que le nombre @xmath21 est rel ou complexe .
montrons maintenant que la partie @xmath913 est suffisamment borne dans lespace @xmath881 .
il existe une constante @xmath94 telle que pour tout point @xmath849 , il existe des compacts @xmath914 , @xmath915 et @xmath916 respectivement de @xmath917 , @xmath918 et @xmath919 , de diamtres majors par @xmath97 , tels que lon ait linclusion @xmath920 \cap k_+ \times ( ( n_x + 1 ) k_0 ) \times k_-.\ ] ] soit un point @xmath849 et une isomtrie @xmath874 .
lexpression @xmath921 est un polynme en @xmath21 coefficients dans @xmath922 , mais cest aussi un polynme en @xmath21 , de degr au plus @xmath923 , coefficient dans @xmath25 .
pour obtenir le compact @xmath916 , il suffit alors de remarquer que si @xmath209 est un nombre rel ou complexe avec @xmath924 , alors pour toute suite @xmath925 , on a @xmath926 si maintenant @xmath209 est un nombre de module @xmath9 , alors on a @xmath927 pour toute suite @xmath925 , ce qui nous donne le compact @xmath915 .
pour finir , le lemme [ bplus ] permet dobtenir le compact @xmath914 do nt le diamtre est indpendant de @xmath32 , et les compacts @xmath915 et @xmath916 ne dpendent pas du point @xmath849 .
finissons la preuve de la proposition [ poly ] .
le groupe additif @xmath882 $ ] tant discret dans lespace @xmath881 , il existe un rel @xmath47 tel que les boules de @xmath881 centres aux points de @xmath882 $ ] et de rayons @xmath48 sont disjointes .
la quantit @xmath928 \cap \left ( k_+ \times ( ( n_x + 1 ) k_0 ) \times k_- \right ) \cdot { \operatorname{vol}}(b(j , \epsilon))\ ] ] est donc majore par le volume dun @xmath48-voisinage du compact @xmath929 .
on obtient donc la majoration @xmath930 \cap \left ( k_+ \times ( ( n_x + 1 ) k_0 ) \times k_- \right ) \\ & \leq \frac{c ( n_x + 1)^{p_0}}{{\operatorname{vol}}(b(j , \epsilon ) ) } , \end{aligned}\ ] ] pour une constante @xmath97 , et pour @xmath931 le nombre de conjugus de @xmath21 de module @xmath9 ( en comptant bien les conjugus complexes ) .
dautre part , on a @xmath932 .
par le sous - lemme [ lm ] , on a alors , pour une constante @xmath97 , @xmath933 ceci termine la preuve du lemme [ poly ] .
le cas o @xmath25 est de cardinal infrieur ou gal @xmath9 est clair : on a facilement @xmath934 .
supposons donc que lensemble @xmath25 est de cardinal au moins @xmath935 .
lensemble limite contient alors au moins deux points .
daprs la proposition [ pradial ] , lensemble limite du semi - groupe est radial : @xmath936 . or , daprs le corollaire [ cpaulin _ ] , lensemble limite radial a une dimension de hausdorff gale @xmath15 ( puisque la dimension de hausdorf concide avec la dimension visuelle , voir @xcite pour plus de dtails ) .
il suffit donc de montrer lgalit @xmath937 pour conclure .
soit @xmath58 une partie spare et @xmath9-couvrante de @xmath1 .
daprs le lemme [ poly ] , il existe alors un entier @xmath62 et une constante @xmath97 tels que pour tout rel @xmath621 assez grand on ait @xmath938 on a alors @xmath939 puisque @xmath940 .
dautre part , on a @xmath695 puisque @xmath58 est une partie spare et couvrante de @xmath1 . on a donc obtenu lingalit @xmath941 , et lautre ingalit @xmath942 est claire .
cela termine la preuve de la proposition [ psalem ] .
soit @xmath1 un sous - semi - groupe de type fini de @xmath0 do nt lensemble limite nest pas rduit un seul point .
alors on a lgalit @xmath943 la conjecture originale posait plutt la question de lgalit @xmath944 . lavantage de cette formulation , qui est quivalente grce au corollaire [ cpaulin _ ] quand le semi - groupe est contractant , est quelle ne fait plus intervenir le bord .
kenyon attribue cette conjecture furstenberg , dans le cas particulier du semi - groupe engendr par les trois applications @xmath945 o @xmath946 est un rel .
cette question , sur cet exemple particulier , est toujours ouverte ma connaissance , bien que lon sache dire pas mal de choses ( voir @xcite ) .
sur cet exemple , la conjecture de furstenberg se rsume dterminer si lon a lgalit @xmath947 quand @xmath946 est irrationnel , puisque dans ce cas le semi - groupe est libre , ce qui donne @xmath948 . le cas o @xmath946 est rationnel a t rsolu par kenyon ( et est aussi consquence de mes rsultats puisque dans ce cas le semi - groupe est spar ) .
avec mes travaux , la conjecture de furstenberg se ramne dterminer si lon a lgalit @xmath949 pour tout @xmath946 irrationnel . ,
avec ensemble de chiffres @xmath950 . ]
les rsultats sur les semi - groupes permettent dobtenir un rsultat sur les groupes : voir corollaire [ ss - groupes ] ci - dessous . plus prcisment , nous parvenons construire des sous - groupes de schottky partir de sous - semi - groupes de schottky , et ceci nous donne des groupes de schottky ayant un gros exposant critique . les groupes de schottky sont dfinit de faon similaire aux semi - groupes de schottky , il sagit des groupes
de type fini do nt les gnrateurs jouent au ping - pong : soit @xmath13 un espace mtrique .
on dit quun ensemble @xmath951 disomtries de @xmath13 engendre un * groupe de schottky * si cest un ensemble fini et quil existe des parties @xmath952 et @xmath953 pour tout @xmath954 qui sont toutes deux - - deux gromov - disjointes , et telles que pour tout @xmath954 on ait @xmath955 , et telles que lensemble @xmath956 soit dintrieur non vide . et @xmath344 . ]
[ ss - groupes ] soit @xmath13 un espace gromov - hyperbolique propre et soit @xmath1 un groupe discret et sans torsion disomtries de @xmath13 ne fixant pas de point au bord , alors on a @xmath957 o la borne suprieure est prise sur lensemble des sous - groupes de schottky du groupe @xmath1 .
ce rsultat est relier une question do nt parle m. kapovich dans son article @xcite ( voir problem 10.27 , the gap problem ) .
avec ses notations , mon rsultat donne lingalit @xmath958 .
la question est de savoir si lon peut atteindre @xmath959 ou non .
le thorme [ thm1 ] permet de trouver un sous - semi - groupe @xmath960 du groupe @xmath1 qui soit contractant pour des parties @xmath267 et @xmath961 , et dexposant critique @xmath962 . on a alors
le lemme suivant .
[ scic ] soit @xmath13 un espace mtrique de point base @xmath34 , et soit @xmath960 un sous - semi - groupe dun groupe discret et sans torsion @xmath1 disomtries de @xmath13 , qui soit contractant pour des parties @xmath267 et @xmath266 de @xmath13 .
pour tout @xmath47 , et pour un entier @xmath67 arbitrairement grand , il existe une partie @xmath963 de @xmath960 telle que * @xmath963 engendre un semi - groupe de schottky pour la partie @xmath267 , * @xmath964 engendre un semi - groupe de schottky pour la partie @xmath266 , * @xmath965 , * @xmath966 , o @xmath649 daprs le lemme [ sc ] , il existe un rel @xmath62 et un entier @xmath238 , tels que pour toute partie @xmath62-spare @xmath58 de @xmath960 et pour tout @xmath239 , le semi - groupe engendr par @xmath648 soit de schottky pour la partie @xmath267 . en faisant de mme pour le semi - groupe inverse @xmath967 , on obtient un rel @xmath156 et un entier @xmath968 .
soit @xmath58 une partie @xmath62-spare et couvrante du semi - groupe contractant @xmath960 .
la partie @xmath969 est spare , puisque faisant partie du groupe discret et sans torsion @xmath1 . par le lemme des tiroirs , il existe donc une constante @xmath94 ( dpendant de la sparation du groupe @xmath1 et du rel @xmath156 ) , et une partie @xmath156-spare @xmath964 de @xmath969 , telles que lon ait @xmath970 pour @xmath47 fix , on peut alors trouver un entier @xmath67 arbitrairement grand , pour lequel on a lingalit @xmath971 puisque lon a @xmath972 , par sparation du sous - semi - groupe @xmath960 du groupe discret @xmath1 .
et par le lemme [ sc ] les parties @xmath963 et @xmath964 engendrent chacune un semi - groupe de schottky respectivement pour les parties @xmath267 et @xmath266 , puisquelles sont respectivement @xmath62-spares et @xmath156-spare .
construisons alors un groupe de schottky de la faon suivante : [ scgc ] soit @xmath13 un espace mtrique , et soit @xmath184 une isomtrie de @xmath13 telle que lensemble @xmath973 soit contractant pour des domaines @xmath266 et @xmath267 .
soit @xmath58 les gnrateurs dun semi - groupe de schottky pour la partie @xmath267 , tel que linverse @xmath974 engendre un semi - groupe de schottky pour la partie @xmath975 .
alors @xmath951 engendre un groupe de schottky .
pour toute isomtrie @xmath717 , on a linclusion @xmath977 et les parties @xmath978 pour @xmath209 dcrivant @xmath58 , et @xmath655 pour @xmath274 dcrivant @xmath58 , sont toutes deux deux gromov - disjointes .
ainsi , en appliquant le lemme [ scgc ] avec la partie @xmath963 donne par le lemme [ scic ] et avec un lment @xmath979 quelconque , on obtient une partie @xmath980 qui engendre un groupe @xmath981 de schottky .
il reste maintenant minorer lexposant critique du groupe obtenu .
lingalit triangulaire donne @xmath982 .
par le lemme [ mnec ] , on a donc la minoration @xmath983 puisque lensemble @xmath984 engendre un semi - groupe de schottky ( donc libre ) qui est un sous - semi - groupe du groupe @xmath981 .
lentier @xmath67 pouvait tre choisi arbitrairement grand , et le rel @xmath47 tait arbitraire , donc on obtient bien lingalit annonce , ce qui termine la preuve du corollaire [ ss - groupes ] .
la minoration de la borne suprieure des exposants critiques des sous - groupes de schottky par @xmath985 nest pas optimale . en pratique ,
les groupes de schottky construits dans cette preuve ont des exposants critiques qui se rapprochent mieux que cela de lexposant critique total @xmath3 .
le corollaire suivant du thorme [ thm0 ] donne en particulier que lexposant critique dun semi - groupe spar , qui tait dfinit comme une limite suprieure dune certaine quantit , est en fait une vraie limite .
ceci gnralise un rsultat que roblin a tablit pour un groupe discret disomtries dun espace cat(-1 ) ( mais avec une conclusion plus forte ) , voir @xcite .
[ lim _ ] soit @xmath13 un espace gromov - hyperbolique propre , et soit @xmath1 un semi - groupe disomtries de @xmath13 do nt lensemble limite contient au moins deux points .
alors on a @xmath986 pour toute partie @xmath58 spare et couvrante de @xmath1 .
soit @xmath47 .
par dfinition de la limite suprieure , il existe un entier @xmath238 tel que pour tout @xmath239 , on ait @xmath987 montrons lautre sens .
daprs le thorme [ thm1 ] , il existe un sous - semi - groupe @xmath960 de @xmath1 qui est contractant et dexposant critique @xmath962 , puisque les semi - groupes @xmath1 et @xmath960 sont spars .
soit @xmath13 un espace mtrique et @xmath70 une partie .
soit @xmath58 une partie spare et couvrante de @xmath25 et @xmath153 un rel .
alors il existe une partie @xmath988 telle que @xmath150 est une partie @xmath62-spare et couvrante de @xmath25 .
par rcurrence ordinale , on construit une suite
@xmath989 dlments de @xmath58 indexe par les ordinaux , en choisissant un lment @xmath990 tant que lensemble est non vide .
cela termine ncessairement puisque la suite ainsi construite ne peut pas avoir un cardinal strictement suprieur celui de @xmath58 .
la partie @xmath150 constitue des lments de la suite est alors @xmath62-spare , et elle est @xmath991-couvrante de @xmath25 , o @xmath97 est telle que @xmath58 est @xmath97-couvrante de @xmath25 .
en effet , si @xmath702 , il existe un lment @xmath992 tel que @xmath993 . et par construction ,
il existe un lment @xmath994 tel que @xmath995 . on a alors bien trouv @xmath994 tel que @xmath996 .
daprs le lemme [ sc ] , quitte remplacer la partie @xmath58 par une sous - partie suffisamment spare et encore couvrante de @xmath960 , il existe un entier @xmath997 tel que pour tout @xmath998 , lensemble @xmath999 engendre un semi - groupe de schottky @xmath1000 , o @xmath1001 pour tout @xmath1002 , en utilisant lingalit triangulaire et le fait que le semi - groupe @xmath1000 soit libre , on obtient lingalit @xmath1003 on a donc lingalit @xmath1004 pour tout @xmath1005 . or , il existe un entier @xmath638 tel que @xmath1006 et comme @xmath58 est une partie couvrante de @xmath960 , on a @xmath1007 .
de plus , on a @xmath1008 , ce qui nous donne lingalit @xmath1009 on peut alors trouver un entier @xmath1010 tel que pour tout @xmath1011 on ait @xmath1012 ainsi , on a bien montr que lon a @xmath1013
nous allons voir que lentropie est semi - continue infrieurement en un semi - groupe .
pour cela , commenons par donner une notion de convergence sur lensemble des semi - groupes disomtries dun espace mtrique @xmath13 .
soit @xmath13 un espace mtrique . on dit quune suite disomtries
@xmath1014 * converge * vers une isomtrie @xmath145 si pour tout compact @xmath1015 de @xmath13 , la suite @xmath1016 des isomtries restreintes
@xmath1015 converge uniformment vers @xmath1017 , et si de mme la suite @xmath1018 converge uniformment vers @xmath1019 .
ici , la convergence uniforme des inverses @xmath1018 sur tout compact @xmath1015 est automatique
partir de la convergence uniforme de la suite @xmath1016 pour tout compact @xmath1015 , puisque ce sont des isomtries .
* pour toute isomtrie @xmath472 , il existe une suite disomtries @xmath307 , avec pour tout @xmath67 , @xmath1021 et telle que @xmath644 converge vers @xmath209 . *
pour toute partie infinie @xmath1022 et toute suite disomtries @xmath1023 qui converge vers une isomtrie @xmath145 , avec @xmath1021 pour tout @xmath1024 , on a @xmath472 .
[ semi - cont ] soit @xmath1020 une suite de semi - groupes disomtries dun espace gromov - hyperbolique propre bord compact @xmath13 qui converge vers un semi - groupe @xmath1 do nt lensemble limite contient au moins deux points .
alors on a lingalit @xmath1025 autrement dit , lentropie est semi - continue infrieurement en les semi - groupes do nt lensemble limite contient au moins deux points .
lide de la preuve du corollaire [ semi - cont ] , est de montrer que si lon a une suite de semi - groupes qui converge , alors on peut approcher un sous - semi - groupe de schottky du semi - groupe limite par des sous - semi - groupes des semi - groupes de la suite , en trouvant des lments qui sapprochent des gnrateurs .
ces semi - groupes seront alors des semi - groupes de schottky do nt les exposants critiques seront proches de celui du semi - groupe de schottky de dpart , et ainsi on obtiendra lingalit voulue .
daprs le thorme [ thm1 ] , il existe un sous - semi - groupe contractant @xmath960 de @xmath1 pour des parties @xmath266 et @xmath764 , avec @xmath1026 .
de plus , on peut supposer que les parties @xmath266 et @xmath267 de @xmath13 sont ouvertes , quitte les remplacer chacune par un @xmath48-voisinage ouvert , pour @xmath47 assez petit .
daprs le critre de contraction ( proposition [ crit_cont ] ) , il existe un rel @xmath238 tel que lensemble disomtries @xmath1027 soit contractant , pour des domaines @xmath1028 et @xmath1029 .
daprs le lemme [ sc ] , il existe alors un rel @xmath62 et un entier @xmath1030 , tels que pour toute partie @xmath62-spare @xmath58 de @xmath1031 et pour tout @xmath1032 , le semi - groupe engendr par @xmath1033 soit de schottky pour la partie @xmath1029 , o @xmath1034n , n+1 [ \}.\ ] ] soit @xmath58 une partie @xmath62-spare et couvrante du semi - groupe contractant @xmath960 .
lensemble @xmath1033 tant fini , il existe une suite densembles disomtries @xmath1035 de @xmath1000 , avec @xmath1036 , qui converge uniformment vers @xmath1033 .
les ensembles @xmath1037 et @xmath1038 tant ouverts , il existe un entier @xmath997 tel que pour tout @xmath998 , on ait @xmath1039 et @xmath1040 . pour @xmath1041 et @xmath998 , on a donc linclusion @xmath1042 .
la condition pour une partie @xmath58 dtre @xmath62-spare est galement une condition ouverte , donc il existe un entier @xmath1043 tel que pour tout @xmath1044 la partie @xmath1035 soit r - spare .
ainsi , pour @xmath67 assez grand et pour tout entier @xmath638 assez grand en fonction de @xmath67 , la partie @xmath1035 est incluse dans le semi - groupe contractant @xmath1031 , et est @xmath62-spare .
elle engendre donc un semi - groupe de schottky .
daprs le corollaire [ semi - cont ] , lapplication @xmath1049 est semi - continue infrieurement .
en particulier , pour trouver un contre exemple la conjecture de furstenberg ( i.e. un rel @xmath946 pour lequel on a @xmath1050 ) , il suffit de trouver une suite de rationnels @xmath1051 qui converge vers un irrationnel et avec pour tout @xmath67 , @xmath1052 .
mais bien que lon sache calculer lexposant critique @xmath1053 du semi - groupe @xmath1047 pour tout rationnel @xmath946 , on ne sait pas sil existe de telles suites .
voir @xcite pour plus de dtails .
k. falconer _ dimensions and measures of quasi self - similar sets _
, proc . of ams , vol .
106 , n2 , p. 543 - 554 , jun . 1989 .
p. doyle _ on the bass note of a schottky group _
e.ghys , p. de la harpe , et al .
_ sur les groupes hyperboliques daprs mikhael gromov _ ,
birkhuser , 1990 . t. haettel _ compactification de chabauty des espaces symtriques de type non compact _ , 2010 .
s. lalley _
@xmath21-expansions with deleted digits for pisot numbers @xmath21 _ , trans .
ams 349 , p. 4355
- 4365 , november 1997 .
p. mercat _ un thorme de patterson et sullivan _ , mmoire de magistre , 2009 . + http://www.math.u-psud.fr/~mercat/memoire.pdf f. paulin _ on the critical exponent of a discrete group of hyperbolic isometries _ , 1995 . | nous gnralisons aux semi - groupes convexes co - compacts un trs joli thorme de patterson - sullivan , donnant lgalit entre exposant critique ( cest - - dire la vitesse exponentielle de croissance ) et dimension de hausdorff de lensemble limite ( cest - - dire la taille du plus petit ferm invariant non vide ) , dun groupe discret disomtrie dun espace hyperbolique .
nous dmontrons ce rsultat dans le cadre gnral des semi - groupes disomtries dun espace gromov - hyperbolique propre bord compact .
pour cela , nous introduisons une notion dentropie , qui gnralise la notion dexposant critique des groupes discrets , et nous montrons que celle - ci est gale la borne suprieure des exposants critiques des sous - semi - groupes de schottky ( cest - - dire les semi - groupes ayant la dynamique la plus simple ) .
nous obtenons ainsi plusieurs autres corollaires tels que la semi - continuit infrieure de lentropie , le fait que lexposant critique dun semi - groupe spar , qui est dfinit comme une limite suprieure , soit en fait une vraie limite , et enfin lexistence de gros sous - groupes de schottky dans les groupes discrets disomtries . 18./dp -init |
the isocam instrument ( cesarsky et al . 1996 ) on board of the _ infrared space observatory _ ( _ iso _ ; kessler et al .
1996 ) has been able to unveil at 15 @xmath0 a population of mir galaxies with two orders of magnitude fainter flux densities than observed by the _ infrared astronomical satellite _ ( _ iras _ ) .
these galaxies result to be ten times more numerous than expected if there were no evolution from @xmath12 up to @xmath13 ( see elbaz et al .
these results are confirmed by the detection of a substantial diffuse cosmic infrared background in the 140 @xmath14 m - 1 mm range ( puget et al .
1996 ; hauser et al . 1998 ; fixsen et al . 1998
; lagache et al .
1999 ) , which suggests a strong evolution of the galaxies in the ir ( stronger than observed at any other wavelengths ) .
the suggested evolution can either imply a larger fraction of galaxies experiencing an ir luminous phase in the past , or that mir galaxies were more luminous in the past ( see e.g. chary & elbaz 2001 ) .
so far , the nature of the sources responsible for the strong evolution observed in the mir band has only been studied in small fields , containing sources well known at other wavelengths : the hubble deep field north ( hdf - n ; aussel et al .
1999a , b ) and south ( hdf - s ; oliver et al .
2002 ; mann et al . 2002 ; franceschini et al .
2003 ) , the canada - france redshift survey ( cfrs ) 1452 + 52 field ( flores et al . 1999 ) and the elais - s2 field ( pozzi et al . 2003 ) .
this has yielded a small but meaningful sample of sources sufficient for most multiwavelength studies .
most of these sources can easily be identified with relatively bright optical counterparts ( @xmath15 ) with a median redshift of @xmath16 imposed by the mir _ k_-correction ( aussel et al .
1999a , b ; flores et al .
1999 ) . however , to obtain reliable source counts and statistically significant information about the nature and evolution of the mir extragalactic source population , large - area surveys covering a wide range of flux densities are required .
in particular , it is very important to bridge the gap between the _ iras _ surveys ( @xmath17 mjy ) and the deep isocam surveys ( @xmath18 mjy ) .
elais ( oliver et al . 2000 ) , with its broad flux range ( 0.5@xmath19150 mjy ) and several thousand sources detected over a @xmath312 deg@xmath20 area ( rowan - robinson et al .
2004 ) , is the crucial survey to investigate the nature and evolution of the mir sources in the region where the source counts start diverging from the non - evolution predictions ( a few mjy at 15 @xmath14 m ; elbaz et al .
1999 ; gruppioni et al .
2002 ) and with data not affected by cosmic variance ( which might be a strong effect in the narrow area surveys ) . in order to better understand the nature of the mir extragalactic population on the basis of a larger identification statistics than the ones obtained so far , we performed photometric and spectroscopic identifications for a source sample selected at 15 @xmath14 m in the elais southern field s1 ( @xmath34 deg@xmath20 ) . here
we present the results of the identification of 406 objects , which form a complete and highly reliable sub - sample drawn from the 15 @xmath0 catalog of lari et al .
the areal coverage , mir and optical completeness are discussed in detail to allow statistical and evolutionary analyses with this sample .
the counts down to 0.6 mjy for each class of source have been derived . the multiwavelength ( radio , mir , near - infrared ( nir ) , optical ) properties of the 15 @xmath14 m sources are investigated in order to study the nature and evolution of the infrared extragalactic population .
the infrared luminosity function and the evolution of 15 @xmath14 m galaxies and agn are discussed in two companion papers ( pozzi et al .
2004 ; matute et al .
2004 , in preparation ) .
a preliminary analysis of the agn evolution has been presented in matute et al .
( 2002 ) this paper is structured as follows : in section 2 we describe the 15 @xmath0 sample ; in section 3 we present the optical ccd observations , the source identification and the cross - correlation with catalogs in other bands ; in section 4 we describe the spectroscopic observations , the classification and the spectroscopic catalog ; in section 5 we present the counts ; in section 6 we discuss the nature of the mir galaxies . throughout this paper
we will assume @xmath21 km s@xmath22 mpc@xmath22 , @xmath23 and @xmath24 .
the primary _ iso _ survey in the elais regions was restricted to two bands : 15 @xmath14 m ( using isocam ) , and 90 @xmath14 m ( using isophot ; lemke et al .
additional observations were performed in some parts of the survey area at 6.7 @xmath14 m and 175 @xmath14 m .
the main survey area , consisting of @xmath312 deg@xmath20 , is divided into four fields distributed across the sky in order to decrease the biases due to cosmic variance : three in the northern hemisphere ( n1 , n2 and n3 ) and one in the southern hemisphere ( s1 ) . details of the survey can be found in oliver et al .
( 2000 ) . the final band merged catalog ( 6.7 , 15 , 90 and 175 @xmath0 ) of the whole elais survey is in rowan - robinson et al .
( 2003 ) .
the southern field , s1 , centered at @xmath25(2000 ) = 00@xmath26 34@xmath27 44.4@xmath28 , @xmath29(2000 ) = -43@xmath30 28@xmath31 12@xmath32 , covers an area of the sky of about @xmath33@xmath34@xmath33 . in this field , a new method of data reduction ( _ lari technique _ ) , especially developed for the detection of faint sources , has been successfully applied , producing a catalog ( complete at the 5@xmath35 level ) of 462 sources in the flux density range 0.5 150 mjy .
details about the data reduction technique and the source catalog have been presented in lari et al .
( 2001 ) , while the sample completeness and the source counts at 15 @xmath14 m obtained from that catalog have been presented and discussed by gruppioni et al .
( 2002 ) .
the whole s1 area has been surveyed in the radio with the australia telescope compact array down to @xmath36 0.4 mjy ( 5@xmath35 , gruppioni et al .
the radio catalog in s1 consists of 652 sources detected at the @xmath37 level at 1.4 ghz .
the radio - mir correlation for starburst galaxies has been discussed by gruppioni et al .
about 40% the area has been observed by _ bepposax _ down to a 2 - 10 kev sensitivity of @xmath310@xmath38 erg s@xmath22 @xmath39 by alexander et al .
( 2001 ) . in this work we restricted our analysis to a highly reliable subsample of 406 sources out of a total 462 15 @xmath14 m sources .
the elais - s1 area has been observed by isocam at 15 @xmath14 m with 9 separate rasters ( see figure [ field ] ) .
20 out of the 462 sources were excluded as they were detected twice at the edges of overlapping rasters ( in this case the sources with the highest s / n ratio have been selected ) .
moreover , 36 sources were excluded as we decided to use only those sources with the highest reliability ( rel=0 ) after a visual inspection of their pixel histories .
the isocam observations of the central ( s1_5 ) raster have been repeated three times ( instead of one ) and thus in this part the survey reaches fainter fluxes . as a consequence , we have divided the sample into two regions : a ) s1_5 ( i.e. raster 5 ) which covers 0.55 deg@xmath20 and reaches a 20@xmath40 completeness at @xmath30.7 mjy , and b ) s1_rest ( i.e. all the remaining rasters ) which covers 3.65 deg@xmath20 and reaches a 20@xmath40 completeness at @xmath31.1 mjy .
note that in this work , as well as in the counts paper ( gruppioni et al .
2002 ) , we used the 15 @xmath0 fluxes of the elais - s1 sources reported by lari et al .
( 2001 ) , corrected following the detailed recipe described by gruppioni et al .
the completeness functions of the two regions as a function of the corrected 15 @xmath0 flux are shown in figure [ fig_compl ] and are reported in table [ tab_compl ] , while the corrected 15 @xmath0 fluxes are reported in table [ tab_id ] and table [ tab_lines ] .
in order to obtain r - band optical photometry of the elais - s1 region down to magnitudes fainter than the typical schmidt plates limits ( e.g. cosmos ) , we have carried out a ccd imaging campaign at the european southern observatory ( eso ) .
the whole s1 region has been covered by 168 overlapping ccd exposures obtained with dfosc mounted on the 1.54 danish / eso telescope at la silla ( chile ) .
the observing runs were carried out during 12 nights ( almost all photometric ) during september 1996 , october 1996 and september 1997 .
the ccd had 2052@xmath342052 pixel array , with a pixel scale of 0.39@xmath41 , and a field of view of [email protected] arc - minutes .
the filter used was the bessel r ( eso # 452 ) .
about 97% the elais - s1 area has been observed in the r - band ( see figure [ field ] ) .
photometric calibration was carried out by frequently observing ( every @xmath31 - 1.5 hours ) the standard field stars `` t phe '' ( @xmath25(2000 ) = 00@xmath26 27@xmath27 49.0@xmath28 , @xmath29(2000 ) = -46@xmath30 48@xmath31 02@xmath32 ; landolt 1992 ) . those fields ( about 10@xmath40 ) observed without photometric conditions were later calibrated using shorter integration images taken during photometric weather .
flat fields were obtained from blank fields in the sky at sunset and sunrise every night .
the integration time for all science frames was 20 minutes .
standard data reduction was carried out with the midas and iraf reduction packages .
every science frame was subtracted for bias , and corrected for flat field .
source searching on the final reduced images was carried out with sextractor ( bertin & arnouts 1996 ) the `` mag_best '' source magnitudes have been used .
the r - band magnitudes were computed after correction for airmass extinction , and without a color term , i.e. assuming a @xmath42 color for all the objects . due to the long observing campaign the characteristics of the ccd observations ( e.g. seeing and magnitude limits ) varied from field to field .
the magnitude limits vary from 22.5 to 23.0 at 95% completeness level .
the astrometry has been computed by using the usno2 catalog as reference , with an internal accuracy of 0.26@xmath41 ( rms ) , and an absolute radial accuracy ( compared to usno2 ) of 0.38@xmath41 ( rms ) . for the optical identifications of the isocam sources we used the @xmath43 magnitudes and positions given in the ccd catalog .
ten sources were not observed by the ccd imaging . for 7 out of these 10 sources
the r - band usno2 magnitudes have been used .
the remaining 3 sources ( elaisc15_j003559 - 430232 , elaisc15_j003912 - 442051 ,
elaisc15_j004023 - 433459 ) have to be considered fainter than r=20 ( see table [ tab_id ] ) .
two sources ( elaisc15_j003034 - 433428 and elaisc15_j003857 - 431719 ) were too close to bright stars and the optical identification was not possible . to estimate the identification reliability
, we adopted the likelihood ratio technique as described by sutherland & saunders ( 1992 ) ( see ciliegi et al .
( 2003 ) for a detailed description ) . when possible counterparts fainter than r=23.0 were found , the identifications were not statistically discussed , and presented only as tentative .
taking into account all of the optical identifications , 332 ( @xmath382% ) out of the 406 elais - s1 sources have a reliable counterpart down to r@xmath323 ( only 7 sources have 23.0@[email protected] ) , 2 are impossible to identify , and 72 ( @xmath318@xmath40 ) are to be considered fainter than [email protected] ( see figure [ fig_histf ] ) . in the following
we will call these sources `` empty '' .
the sample has been cross - correlated with the k - band catalog from the two micron all sky survey ( 2mass ) , which reaches a k - band limit of @xmath315.6 ( at 95% completeness level ) . in total 211 sources have been detected in the k - band . since the whole s1 field has been surveyed in the radio ( 1.4 ghz ) with the australia telescope compact array to an average 1@xmath35 level of 0.08 mjy ( gruppioni et al .
1999 ) , at each isocam position , we have searched for detection in the radio map down to 3 @xmath35 , finding 71 likely radio counterparts ( 61 with measured redshifts ; see gruppioni et al .
2003 for details ) .
the results of the multi - band identifications are reported in table [ tab_id ] .
spectroscopic observations of the optical counterparts of the isocam s1 sources were carried out at the 2df / aat and eso danish 1.5-m , 3.6-m and ntt telescopes during september and october 1998 , september and november 2001 , september and october 2002 .
a total of 15 eso nights and 2 uk ( from the panel for the allocation of telescope time ; patt ) nights were allocated .
the 2df / aat multifiber feed spectrograph was used with 200 fibers equipped with the 300b grism and the remaining 200 fibers with the 270r grism ( about 10 instrumental resolution ) . because of poor weather conditions only half an hour of integration was possible during each night , reaching a magnitude limit of [email protected] .
the eso spectrographs ( dfosc , efosc2 and emmi ) were all used in single slit low resolution mode ( 10 - 20 resolution ) in the wavelength range 3500 - 9500 , and reaching an average s / n ratio per resolution element of @xmath315 - 30 in the wavelength range 4000 - 7000 .
the reduction process used standard midas and iraf facilities .
the raw data were sky - subtracted and corrected for pixel - to - pixel variations by division with a suitably normalized exposure of the spectrum of an incandescent source ( flat field ) .
wavelength calibrations were carried out by comparison with exposures of he , ar and ne lamps .
relative flux calibration was carried out by observations of spectrophotometric standard stars ( oke 1990 , hamuy 1992 , 1994 ) .
we have spectroscopically classified 290 ( @xmath387% ) out of the 332 sources optically identified ( @xmath371% of the whole 15 @xmath14 m sample ) .
the sample has been cross - correlated with the tycho2 ( hg et al .
2000 ) and ned catalogs in order to identify known stars and extragalactic objects .
sixty - two tycho2 stars were found .
the remaining objects brighter than r=15 were visually inspected in the r band images and those suspected to be extended / extragalactic were spectroscopically observed .
fifteen sources with r@xmath115 were considered to be bona fide stars on the basis of the visual inspection , while 14 were spectroscopically identified as stars at the telescope . in total
we found 91 stars ( all of them brighter than r=15 ) and 199 extragalactic sources . in order to classify the extragalactic sources we have first identified type 1 agn ( agn1 ) when permitted lines with rest frame full width half maximum ( fwhm ) larger than 1200 km / s were found .
the remaining objects were classified according to two methods .
the first method was used to identify type 2 agn ( agn2 ) , and to separate them from other emission line galaxies such as hii / starburst and low ionization narrow emission regions ( liner ) .
all the objects with narrow h@xmath44 or [ oiii]@xmath455007 emission lines having ew ( measured as the ratio between the integrated flux in the line , as resulting from a gaussian fit , and the flux of the continuum at the line wavelength ) larger than @xmath33 - 5 ( our line detection limit ) , were classified according to three diagnostic diagrams : 1 ) [ oiii]@xmath455007/h@xmath44 ratio versus [ nii]@xmath456583/h@xmath25 ratio ; 2 ) [ oiii]@xmath455007/h@xmath44 ratio versus [ sii]@xmath456725/h@xmath25 ratio ; 3 ) [ oiii]@xmath455007/h@xmath44 ratio versus [ oii]@xmath453727/h@xmath44 ratio ( e.g. veilleux & osterbrock 1987 ; tresse et al . 1996 ; see figure [ fig_diagn ] for an example ) . the h@xmath25 and [ nii]@xmath456583 lines were deblended using the iraf task splot , which assumes gaussian line profiles and a linear background .
the above method is efficient in separating agn from starburst activity , but does not provide information on the nature of a large fraction of our galaxies , which do not show [ oiii ] and h@xmath44 emission lines .
we have thus applied the classification scheme of dressler et al .
( 1999 ) ( see also poggianti et al .
1999 and poggianti & wu 2000 ) to all the `` non agn '' galaxies .
this method is based primarily on two lines , [ oii]@xmath453727 in emission and h@xmath29 in absorption , which are good indicators of ( respectively ) current and recent star formation ( sf ) in distant galaxy spectra . under this scheme
the spectra have been divided in : 1 .
e(a ) ( having 3@xmath1ew([oii])@xmath140 and ew(h@xmath29)@xmath464 ) , spectra of dust - enshrouded starburst galaxies ; 2 .
e(b ) ( having ew([oii])@xmath4740 ) , spectra with very strong emission lines of galaxies which are undergoing strong star formation ; 3 .
e(c ) ( having 3@xmath1ew([oii])@xmath140 and ew(h@xmath29)@xmath14 ) , typical spectra of spirals which have been forming stars in a continuous fashion ; 4 .
k ( absent [ oii ] and ew(h@xmath29)@xmath13 ) , spectra of passive elliptical - like galaxies with neither ongoing nor recent star formation ; 5 .
k+a / a+k ( absent [ oii ] and ew(h@xmath29)@xmath463 ) , spectra of poststarburst galaxies with no current star formation which were forming stars at a vigorous rate in the recent past ( last 1.5 gyr ) .
k(e ) , spectra similar to the k - type but with signs of at least one emission line .
this class was introduced by us following duc et al .
( 2002 ) . following this classification
, we found 25 agn1 , 23 agn2 , 37 e(a ) , 11 e(b ) , 68 e(c ) , 32 k(e ) and 3 k galaxies .
no k+a / a+k galaxy was found .
thirteen objects could not be classified on the basis of this scheme because the wavelength range of [ oii ] and h@xmath29 lines was not observed ( they were observed with the red grism of the 2df spectrograph ) .
two of them were classified e(b ) ( starburst ) galaxies according to the [ oiii]@xmath455007/h@xmath44 and [ nii]@xmath456583/h@xmath25 ratios , while the remaining 11 galaxies were classified k(e ) because no [ oiii]@xmath455007 and h@xmath44 line was detected , while the h@xmath25 emission line was present . in figure [ fig_spettri ]
we show the composite spectra of each type of galaxy found under this classification scheme .
we should stress that the number of identified agn2 has to be considered a lower limit .
_ xmm _ and _ chandra _ surveys have revealed significant numbers of ( mostly absorbed ) agn with l(2 - 10 kev)@xmath47 10@xmath48 erg s@xmath22 ( i.e. agn2 ) whose optical spectra do not show sign of agn activity ( see e.g. fiore et al . 2000 ; 2003 ) .
recent ( june - july 2003 ) , 100 ksec long , _
observations of s1 - 5 have revealed indeed that about 10 - 20% of the counterparts of our _ iso _ sources classified as no - agn galaxies on the basis of their optical spectra , can harbor an x - ray emitting agn ( the _ bepposax _ observations presented in alexander et al .
( 2001 ) are too shallow to provide tighter agn constraints ) . under the reasonable assumption
that all these objects are agn2 , the number of agn2 in the elais - s1 sample should be increased by at least a factor of 2 .
a more detailed discussion on the x - ray / mir relationship of agn and starburst galaxies is beyond the scope of this paper and will be presented by la franca et al .
( in preparation ) . in table [ tab_id ]
we list the multiband associations to the 15 @xmath0 sources with the likelihood and reliability of the r band identifications , plus the redshift , the spectroscopic classification and the @xmath49 luminosities at 15 @xmath0 and in the r - band . in table
[ tab_lines ] we present the line measurements of the spectra which include the calcium break ( d@xmath50 ; measured from the fluxes in the intervals 3750 - 3950 and 4050 - 4250 ; see dressler et al .
1999 ) and the ew for the [ oii]@xmath453727 , h@xmath29 , h@xmath44 , [ oiii]@xmath51 , h@xmath25 , [ nii]@xmath456583 and [ sii]@xmath456725 lines ( positive values correspond to absorption lines ) .
figure [ fig_r15_rk ] shows the distribution of the 211 elais - s1 sources detected both in the r and k bands , in the @xmath52 versus @xmath53 plane , where we have defined @xmath54 .
this diagram shows a clear separation of the locus of the stars from the locus of the galaxies : for any given @xmath52 color , stars populate a region with a two order of magnitude lower mir to optical luminosity ratio .
note that only one early - type galaxy and two stars are located in the intermediate area of the diagram between those occupied by stars and galaxies .
this means that stars can easily be separated from galaxies based purely on their infrared to optical flux ratio ( see also visnen et al .
2002 ) .
the s1_5 area is 99@xmath40 spectroscopically complete ( 72/73 ) down to r=21.6 .
the s1_rest area is 97@xmath40 ( 216/222 ) complete down to r=20.5 ( see figures [ fig_histr ] and [ fig_sr ] ) .
these magnitude limits have been used for the statistical analysis of the evolution of agns and galaxies by matute et al .
( in preparation ) and pozzi et al . ( 2004 ) respectively . both figures [ fig_histr ] and [ fig_sr ]
show a separation between stars and galaxies , the stars being optically brighter than galaxies , and also the appearance of a separate population of galaxies at faint magnitudes ( @xmath55 ) .
see 6.3 for a detailed discussion on this point .
in the following we will call @xmath56 all the 116 elais - s1 sources which are without spectroscopic identification ( i.e. , including also 72 empty sources and 2 sources with no optical identification becouse too close to bright stars ; see 3.2 ) .
we have computed the contribution of each class of sources to the 15 @xmath14 m source counts in the mjy flux region .
the integral counts for star - forming galaxies and agns ( as defined on the basis of the optical spectra ) are reported in table [ tab_counts ] ( see figure [ fig_counts ] ) . as already discussed in section
4.2 , the counts of agn should be considered as lower limits .
all elais - s1 extragalactic sources ( including the @xmath56 ) , with 15 @xmath14 m fluxes brighter than 0.6 mjy , produce a 15 @xmath14 m background of @xmath57i(@xmath57 ) = 0.34 nw m@xmath58sr@xmath22 .
this corresponds to @xmath312@xmath40 of the total extragalactic background estimated by metcalfe et al .
( 2003 ) of @xmath57i(@xmath57)=2.7 nw m@xmath58 sr@xmath22 .
the sources spectroscopically identified with galaxies ( no agn ) produce a flux of @xmath57i(@xmath57)=0.16 nw m@xmath58 sr@xmath22 .
the empty fields ( mainly objects fainter than r@xmath323 ) produce @xmath57i(@xmath57)=0.14 nw m@xmath58 sr@xmath22 , while the agns produce @xmath57i(@xmath57)=0.04 nw m@xmath58 sr@xmath22 .
if we assume that the @xmath59 sources correspond mainly to starburst galaxies , we derive a @xmath390@xmath40 contribution of starburst galaxies and a @xmath310@xmath40 contribution of agns to the 15 @xmath14 m background at fluxes brighter than 0.6 mjy .
if our preliminary estimate of the fraction of of hidden x - ray emitting agn2 among the galaxies is taken into account , these figures may well be consistent with the higher estimates ( @xmath31520% ) for the contribution of agn derived by alexander et al .
( 2002 ) , and fadda et al .
in figure [ fig_histz ] the redshift distribution of the spectroscopically identified sources is shown . while agn1 reach redshifts up to @xmath11@xmath33 , all agn2 and galaxies have redshifts lower than 1 , with an average of @xmath1@xmath11@xmath47@xmath60 .
a possible hint of structure at @xmath61 and @xmath62 is present .
we computed the @xmath57l@xmath63 luminosities of our sources at 15 @xmath0 ( l@xmath64 ) and in the r - band ( l@xmath65 ) . for each class of objects a k - correction was applied according to the spectral classification ( see figure [ fig_kcorr ] ) . in the infrared the k - corrections were based on : a ) the sed from elvis et al .
( 1994 ) for agn1 , b ) the sed of ngc1068 for agn2 ( sturm et al .
2000 ) and , c ) the sed of m82 for all of the remaining sources , except the absorption line galaxies ( for which m51 was considered ) . in the optical , the k - correction from natali et al .
( 1997 ) was used for agn1 , while for agn2 the k - corrections were computed by using the average of our corresponding optical spectra . for the galaxies the seds of m82 and m51
were used , consistently with what has been done for the 15 @xmath0 flux .
the model seds of both m82 and m51 were taken from the grasil library of models ( silva et al .
1998 ) . for m82 , the mir ( 5
18 @xmath0 ) has been replaced with the observed isocam cvf spectrum ( frster schreiber et al . 2003 ) . in figure [ fig_zl ]
the @xmath66 luminosities are shown as a function of redshift
. a luminosity dependence of the classes of objects is evident .
agn1 are the most luminous and most distant sources with @xmath67@xmath47@xmath68 , while agn2 are less luminous and less distant with @xmath67@xmath47@xmath69 .
finally , the least luminous and least distant sources are the galaxies with @xmath67@xmath47@xmath70 . in figure [ fig_l15_lopt ] , @xmath71 versus @xmath66
is shown for all of the elais - s1 15 @xmath14 m galaxies plus sources from the spectroscopic identifications in the hdf - s from franceschini et al .
( 2003 ) , the hdf - n ( isocam data from aussel et al . 1999 ; spectroscopic data from cohen et al .
2000 ; magnitudes from hogg et al .
2000 ) , and the local sample of _ iras _ 12-@xmath14 m galaxies of rush , malkan & spinoglio ( 1993 ; rms ;
after conversion to 15 @xmath14 m by using the m51 and m82 sed for normal and starburst galaxies respectively ) . a clear relation which is valid for all of the considered surveys is found , although it involves only objects with optical identification .
the best fit to the data is @xmath72 - 5.02.\ ] ] the 1@xmath35 dispersion of the relation is 0.32 . note that type 1 agn have been excluded because of the different mechanism responsible for their luminosities , placing them on a different @xmath71 versus @xmath66 plane location with respect to galaxies . on the contrary ,
type 2 agn follow the same relation as all the other galaxies .
this relation implies that the average optical - ir spectral energy distribution of galaxies is luminosity dependent , with more luminous galaxies having larger @xmath71 ratios .
a two order of magnitude increase in luminosity implies a 10 times larger @xmath71 ratio .
this relation is also discussed and taken into account for the analyses of the evolution of starburst galaxies and agns presented in two companion papers by pozzi et al .
( 2004 ) and matute et al .
( in preparation ) . as already shown by gruppioni et al .
( 2002 ) , the 15-@xmath0 extragalactic counts show an upturn at fluxes fainter than @xmath31 my ( see also elbaz et al .
1999 ) , which is interpreted as due to the contribution of the strong cosmological evolution of the starburst population .
this interpretation is discussed in detail by pozzi et al .
anyway the counts ( see figure [ fig_counts ] ) show that the upturn is mostly due to the contribution of the @xmath56 sources ( i.e. sources fainter than our spectroscopic or photometric limits ) , which represent about half of the extragalactic sources at fluxes @xmath30.6 - 1 mjy .
these sources could be either optically less luminous ( absorbed ? ) galaxies at redshift similar to those of the already spectroscopically identified sample , or higher redshift sources .
we favor the second hypothesis on the basis of the following discussion .
we have tried to estimate a redshift for the @xmath56 sources in order to distinguish between the two different scenarios described above .
we have combined our sample with the fainter hdf - s and hdf - n samples .
then we have tried to derive a flux - redshift relationship applicable to the elais - s1 unidentified sources in order to estimate their redshifts .
after several trials using multiple polynomial relations between @xmath73 and both 15 @xmath14 m flux ( @xmath74 ) and magnitude ( @xmath43 ) , we have assigned a redshift to the 116 @xmath75 sources on the basis of their @xmath43 magnitudes only ( @xmath76 was assumed for all the unidentified 76 ( @xmath77 ) sources ) , by considering the following empirical relation and its spread : @xmath78 where @xmath79 (= 0.15 ) is the 1@xmath35 dispersion of the relation and @xmath80 is a gaussian distribution with center 0 and width @xmath79 ( see figure [ fig_fitz ] , @xmath81 ) .
note that the fit is non - linear in @xmath73 vs. @xmath43 , since we found that a linear fit to the data extrapolated to faint magnitudes was providing too high redshifts , inconsistent with the k - correction limit of isocam ( @xmath82 ) and with the median redshifts of @xmath83 objects from the hdf - s and hdf - n . using the estimated redshifts , we have computed mir and optical luminosities for the unidentified sources ( by assuming all of them being star - forming galaxies , thus using m82 sed ) and we have placed also these objects in the @xmath71 vs. @xmath66 diagram ( see the @xmath84 panel of figure [ fig_fitz ] ) . according to our estimates ,
the unidentified objects in our survey are likely to be rather luminous mir objects ( @xmath66 ranging between @xmath85 l@xmath7 and a few @xmath86 l@xmath7 , in the lig / ulig region ) , and their optical and mir estimated luminosities appear to be consistent with ( and extending to higher luminosities ) the @xmath71 vs. @xmath66 correlation found for the identified sources . the estimated redshift distribution for the unidentified sources is shown in figure [ fig_z_empty ] , where it is compared with the distribution found for the spectroscopically identified sources and with the model fitting derived from our data ( i.e. the 15 @xmath14 m luminosity function in s1 ; pozzi et al .
the agreement between the model and the data ( either observed and estimated ) is good .
it results that the unidentified sources are expected to fill the secondary high-@xmath11 peak of the model - predicted redshift distribution .
the double peaked redshift distribution ( see figure [ fig_z_empty ] ) is caused by the 15 @xmath0 k - correction . as shown in figure [ fig_kcorr ] , at @xmath11=0.5 the 15 @xmath0 observed fluxes decrease by @xmath330% . for our flux
limited sample this implies that the elais - s1 galaxies at @xmath11=0.5 correspond to 30% higher intrinsic luminosities .
but at @xmath11=0.5 the elais - s1 galaxies sample the bright steep slope of the luminosity function ( log(l@xmath87@xmath4711 @xmath47 l@xmath88 ) , where 30% higher intrinsic luminosities correspond to an order of magnitude lower density of galaxies .
this decrease of density of galaxies is also evident in the r - magnitude distribution ( see figures [ fig_histr ] and [ fig_sr ] ) , where , due to the k - correction effects , a lack of sources with r@xmath320 - 21 ( and @[email protected] ) is evident .
the r - band histogram ( figure [ fig_histr ] ) shows a triple peaked distribution : stars populate the brightest peak at r@xmath310 - 12 , @[email protected] galaxies populate the intermediate peak at r@xmath317 - 19 , and higher redshift ( @[email protected] ) galaxies populate the fainter peak that can be observed at r@xmath4721 - 22 . under the alternative hypothesis
that elais - s1 @xmath56 were instead lower redshift ( @[email protected] ) sources suffering from dust extinction , it would be more appropriate to consider the 15-@xmath0 flux
@xmath11 relation rather than the @xmath89 relation ( because affected by dust ) to estimate their redshifts . in this case , we would obtain a redshift distribution for the @xmath56 sources similar to that of the optically identified , but they would be located in a separate region of the @xmath71 vs. @xmath66 diagram ( i.e. at similar @xmath66 , but with at least one order of magnitude higher @xmath71 ) .
this would imply the existence of a absolutely different population of objects , with much higher level of obscuration ( without continuity ) than the identified ones , but at similar flux density and redshifts . the first hypothesis is also supported by the faint spectroscopic identifications in the hdf - n ( aussel et al .
1999 ) and cfrs 1452 + 52 ( flores et al .
1999 ) fields which have mainly found a population of @xmath11=0.4 - 1.0 galaxies .
for all these reasons we conclude that the vast majority of the faint ( [email protected] ) elais - s1 @xmath56 objects should belong to the same population as the identified sources , but at higher redshifts ( @xmath11=0.5 - 1.5 ) , and that consequently the present sample is virtually spectroscopically complete up to @xmath11=0.5 , as far as the star forming galaxies are concerned . as we have seen , our sample is virtually complete up to @xmath90 . up to this redshift
we have identified 167 extragalactic objects : 3 agn1 , 20 agn2 , 33 e(a ) , 9 e(b ) , 67 e(c ) , 32 k(e ) , and 3 k galaxies .
no k+a / a+k galaxy was found .
the average redshift of the sample is @xmath1@xmath11@xmath47=0.20 .
the number of e(a ) galaxies rises to 37 if we classify also the agn2 according to the classification scheme of dressler et al .
( 1999 ) ( see 4.2 ) . the majority of the sources with @[email protected] is composed by normal star forming spirals ( e(c ) , 41% ) , and by absorbed star forming galaxies ( e(a ) , 20% ; 23% including the agn2 ) .
the total fraction of optical spectra which show emission lines which are indicators of star formation ( e(b)+e(c)+e(a)+k(e ) ) is 84% .
14% of the sources are agns , and only 2% of the sources do not show emission lines ( k galaxies ) .
this confirms that the mir and the optical data agree on revealing both ongoing star formation and accretion driven ( agn ) activities .
the lack of k+a / a+k galaxies in the elais - s1 sample suggest that these galaxies are not strong mir emitters .
this result is consistent with the lack of emission lines in their optical spectrum and with their post - starburst interpretation ( see poggianti and wu 2002 ) .
the e(a ) galaxies are believed to be associated with dust enshrouded starburst galaxies .
following poggianti et al .
( 1999 ) and poggianti & wu ( 2000 ) , the spectra of e(a ) galaxies are explained as the result of selective dust extinction which affects the youngest , most massive , stars in hii regions much more than the older stellar population responsible for the continuum flux .
a search for similar spectra in the local universe ( poggianti et al . 1999 ) revealed that they are frequent among merging / strongly interacting galaxies ( @xmath340% ) , while they are scarce in optically selected surveys of field galaxies at low @xmath11 : 7 - 8% in the las campanas redshift survey and in kennicut ( 1992 ) .
dressler et al . ( 1999 ) and poggianti et al . ( 1999 ) found that e(a ) galaxies constitute about 10% of their sample of cluster and field galaxies at @[email protected] - 0.5 . at 15 @xmath14
m most ( 71% ) of the sources with optical spectroscopy in the cfrs are classified as e(a ) galaxies and have a mean redshift @xmath1@xmath11@xmath47=0.76 , while at @xmath11=0.18 , in the cluster of galaxies abell 1869 , duc et al .
( 2002 ) find 25% of 15 @xmath14 m galaxies classified as e(a ) .
the observed 20 - 23% fraction of e(a ) galaxies in the elais - s1 sample at @xmath1@xmath11@xmath47=0.20 , is thus higher than observed in optically selected samples and in agreement with the estimates at similar redshifts by duc et al .
( 2002 ) in a mir selected sample .
this confirms that e(a ) galaxies are dusty absorbed star forming galaxies which are preferentially selected at longer wavelengths ( mir ) , and whose frequency increases with increasing redshift , as the cosmic star formation rate .
we have compared the @xmath52 colors of our sources to those of the field galaxies and we did not find evidences of reddening .
becouse of the bright k - band limit magnitude we limited our comparison in the magnitude range r=15 - 17.5 , and consequently the statistic are poor .
the average r - k color is 2.04 and 2.27 for elais and field galaxies respectively , with a 1@xmath35 dispersion of @xmath30.50 for both samples . on the contrary
, dust extinction seems to affect the spectral lines of our objects . in figure [ fig_alphaoii ] the ew of the [ oii ] line
is compared to the ew of the h@xmath25+[nii ] line .
the @xmath91)/ew(h\alpha+[nii])$ ] ratio is related to the differential extinction due to dust , the @xmath92 $ ] emission being more affected than @xmath93 emission due to its shorter wavelength .
the dashed line in figure [ fig_alphaoii ] is the relation found by kennicutt ( 1992 ) for local field galaxies ( @xmath91)/ew(h\alpha+[nii])$]@xmath940.4 ) .
most of our sources lie below that relation , and the presence of upper limits ( i.e. [ oii ] non - detections ) would even lower the average value of the @xmath91)/ew(h\alpha+[nii])$ ] , implying a larger amount of extinction in our data than in the local sample of kennicutt ( 1992 ) .
the average ratio for the galaxies having both [ oii ] and h@xmath25+[nii ] lines is [email protected] , [email protected] , [email protected] for e(b ) , e(c ) and e(a ) galaxies respectively .
thus , as expected , at least for the normal star forming ( e(c ) ) and absorbed star forming ( e(a ) ) galaxies the difference from the local observed ratio ( 0.4 ) is significant . in order to disentangle a possible dependence on redshift or luminosity of the average extinction of the galaxies
, we have combined our sample with a subsample of 39 galaxies from the local sample of rms for which ew measures for [ oii ] and h@xmath25+[nii ] were available in the literature ( see alexander & aussel ( 2000 ) for a discussion of this sample ) .
the data show that the extinction affecting the lines is an increasing function of luminosity while a correlation with redshift is not statistically significant : a partial correlation kendall test gives in fact a 10@xmath96 probability that the ew is not dependent on luminosity , and a 0.04 probability that the ew is not dependent on redshift .
figure [ fig_linetrend ] ( @xmath97 ) shows the distribution of the rms and elais galaxies in the luminosity - redshift plane .
we have then selected the galaxies with @xmath98@xmath1@xmath11@xmath1@xmath99 in order to show the dependence on luminosity ( figure [ fig_linetrend ] , @xmath100 ) , and the galaxies in a small luminosity interval ( @xmath101@xmath1@xmath102@xmath1@xmath103 ) in order to show the not significant trend with the redshift ( figure [ fig_linetrend ] , @xmath84 ) .
these results suggest an increase of dust extinction with luminosity .
similar results , based on a smaller sample , were found also by pozzi et al .
( 2003 ) in one of the smaller but deeper elais fields , s2 .
a quantitative estimate of the amount of optical extinction affecting the elais - s1 mir galaxy sample can be obtained by the measure of the observed balmer decrement in the h@xmath25 and h@xmath44 lines .
we have made this measure on two composite spectra for each class of galaxies : one including only spectra with detected h@xmath25 and h@xmath44 emission lines , and one including also spectra with no detected h@xmath44 line .
spectra showing h@xmath44 lines in absorption were excluded .
this method was adopted in order to minimize the effects of underling h@xmath44 absorption which causes the observed h@xmath25/h@xmath44 ratio to be larger .
for this reason the extinction estimates we derive have to be considered upper limits . in table
[ tab_ext ] our estimates in the case of b recombination with a density of 100 @xmath104 and a temperature of 10000@xmath105k ( osterbrock 1989 ) are shown .
the higher values of e(b - v ) were always obtained when the spectra without h@xmath44 detection were included in the composite spectra .
these values correspond to an extinction @xmath106=3.2@xmath34@xmath107 of about 1.5 , 3.5 , 4.0 , 4.5 mag for e(b ) , e(c ) , e(a ) and k(e ) galaxies respectively .
these value refers to the absorption of the lines emission ; typically the absorption on the continuum is a factor of two smaller ( see also pozzi et al . 2003 , and poggianti & wu 2000 for similar results ) .
as expected , the larger extinction values are found for e(a ) and k(e ) galaxies , reinforcing the interpretation that the e(a ) signature is associated with dust extincted starbursts .
we have estimated the star formation rate ( sfr ) of the elais - s1 sources from their 15 @xmath0 infrared luminosity .
the sfr estimator has been computed as : @xmath108= 11.25\times10^{-23}l_{15}[w hz^{-1}],\ ] ] assuming a salpeter initial mass function ( imf , salpeter 1955 ) between 0.1 and 100 @xmath109 , a dependence of sfr from l@xmath110 as given by mann et al .
( 2002 ; @xmath111= 0.9\times10^{-23}l_{60\mu m}[w hz^{-1}]$ ] ) , and a mean value of the ratio @xmath112 / l_{60\mu m}[w hz^{-1 } ] > \simeq 0.08 $ ] , typical of starburst galaxies such as m82 . in figure [ fig_zsfr ] the sfrs as a function of redshift are shown .
type 1 agn have been excluded as their mir luminosities are not attributed to star forming activity , but to reprocessed higher frequency luminosity driven by the accretion on a super - massive black hole ( 10@xmath113 - 10@xmath114 m@xmath7 ) . for the sake of comparison the estimates from the spectroscopic identifications in the hdf - s , and
the hdf - n are shown .
the isocam observations of the hdf - s and hdf - n reach 15 @xmath0 fluxes down to @xmath30.1 mjy ( i.e. , more than six times fainter than our survey ) and probe a different region of the sfr vs. @xmath11 plane .
the shallower wide - area elais survey is complementary to the deep surveys and is required to make a more accurate estimate of the sfr history of the universe ( see pozzi et al .
the elais - s1 sources have mean sfrs of 13 m@xmath10/yr , 45 m@xmath10/yr and 242 m@xmath10/yr , in the redshift ranges @[email protected] , 0.2@xmath1@[email protected] and @[email protected] respectively .
obviously this should not be attributed to an evolutionary trend , but to the luminosity - redshift relation which is present in any flux limited sample like elais - s1 .
we have found a good correlation between the sfrs derived from the 15 @xmath0 luminosity and those derived from the 1.4-ghz luminosity ( computed as @xmath115= 1.1\times10^{-21}({\nu\over ghz})^{\alpha}l_{1.4ghz}[{w \over hz}]$ ] , assuming a salpeter imf between 0.1 and 100 @xmath109 , and a power - law spectrum with a spectral index @xmath1160.7 ; @xmath117 ) .
the correlation between the sfrs obtained through the two different indicators reflects the radio - mir luminosity correlation found for elais - s1 galaxies by gruppioni et al .
( 2003 ; see also elbaz et al .
we have presented multiband ( r , k , 1.4 ghz ) and optical spectroscopic identifications of a reliable subsample of 406 15-@xmath0 sources from the elais - s1 survey over the flux density range 0.5@xmath19150 mjy .
these observations substantially contribute to the understanding of the evolution of star - forming galaxies and agn in the mir , which is presented in two companion papers by pozzi et al .
( 2004 ) and matute et al .
( in preparation ) respectively .
this paper has described the details of the ground - based follow - up observations and data reduction , and has presented optical identifications for @xmath382@xmath40 of the 15 @xmath0 sources down to r@xmath323 , and spectroscopic identification for @xmath388@xmath40 of the optically identified sample .
the areal coverage and completeness of the survey has been described in detail , in order to allow statistical analyses .
the source counts for star - forming galaxies and agn have been derived : the elais extragalactic sources down to s@xmath118=0.6 mjy have a density of 370@xmath9514/deg@xmath20 , and produce a 15 @xmath0 background of @xmath57i(@xmath57 ) = 0.34 nw m@xmath58sr@xmath22 .
this corresponds to @xmath312@xmath40 of the total extragalactic background estimated by metcalfe et al .
taking into account the optical based classification , we estimate an agn contribution of at least @xmath36% to the mir sources
brighter than s@xmath118=0.6 mjy .
a larger fraction is expected if the classification will be based on the x - ray luminosity .
the line ratios of the optical spectra of the mir galaxies suggest that they are affected by substantial reddening , whith the more mir - luminous sources having larger dust extinction
. a general trend is found in the mir - optical seds of _ iras _ and _ iso _ galaxies whith the more mir - luminous sources having larger @xmath71 ratios . according to our analyses we conclude that most of the elais galaxies fainter than r@xmath323 should have redshift greater than 0.5 and are likely to be lig - ulig galaxies . as a consequence , the present sample of star - forming galaxies has to be considered virtually complete up to @xmath119 .
based on observations collected at the european southern observatory , chile , eso n@xmath30 : 57.a-0752 , 58.b-0511 , 59.b-0423 , 61.b-0146 , 62.p-0457 , 67.a-0092(a ) , 68.a-0259(a ) , 69.a-0538(a ) , 70.a-0362(a ) ; patt proposal p076 ) .
this paper is based on observations with the infrared space observatory ( iso ) .
iso is ane esa project with instruments funded by esa member states ( especially the pi countries : france , germany , the netherlands and the united kingdom ) and with the partecipation of isa and nasa .
this research has made use of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
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 / california institute of technology , funded by the national aeronautics and space administration and the national science foundation .
this research has been partially supported by asi contracts ars-99 - 75 , asi i / r/107/00 , asi i / r/113/01 , murst grants cofin-98 - 02 - 32 , cofin-00 - 02 - 36 .
i m acknowledges a phd grant from cnaa / inaf .
dma acknowledges the royal society for generous support .
hraudeau acknowledges the financial support provided through the european community s human potential programme under contract hprn - ct-2002 - 00316 , sisco .
alexander , d.m . ,
aussel , h. 2000 , lecture notes in physics , 548 , 113 alexander , d.m . , et al .
2001 , , 554 , 18 alexander , d.m . ,
aussel , h. , bauer , f.e . , brandt , w.n .
, hornschemeier , a. e. , vignali , c. , garmire , g.p . ,
schneider , d.p 2002 , , 568 , 85 aussel , h. , csarsky , c.j . , elbaz , d. , & starck j .- l .
1999a , a&a , 342 , 313 aussel , h. , elbaz d. , csarsky , c.j . ,
& starck j .- l .
1999b , the universe as seen by _ iso _ , p. cox & m.f .
kessler , esa special publication series ( sp-427 ) , p. 1023
bertin , e. , arnouts , s. 1996 , a&as , 117 , 393 bruzual , a.g .
, charlot , s. 1993 , , 405 , 538 cesarsky , c.j . ,
1996 , a&a , 315 , l32 chary , r. , elbaz , d. , 2001 , , 556 , 562 ciliegi , p. , zamorani , g. , hasinger , g. , lehmann , i. , szokoly , g. , & wilson g. 2003 , a&a , 398 , 901 cohen , j.g . ,
hogg , d.w .
, blandford , r. , cowie , l.l . , hu , e. , songaila , a. , shopbell , p. , & richberg k. 2000 , apj , 538 , 29 dressler , a. et al .
1999 , apjs , 122 , 51 duc , p.a . , et al .
2002 , , 382 , 60 elbaz , d. , et al . 1999 ,
a&a , 351 , l37 elbaz , d. , et al .
2002 , a&a , 384 , 848 elvis , m. , et al .
1994 , , 95 , 1 fadda , d.flores , h. , hasinger , g. , franceschini , a. , altieri , b. , cesarsky , c.j . , elbaz , d. , ferrando , ph .
2002 , , 383 , 838 fixsen , d.j . ,
dwek , e. , mather , j.c . ,
bennett , c.l . , &
shafer r.a .
1998 , apj , 508 , 123 fiore f. , et al .
2000 , new ast . , 5 , 143 fiore f. , et al .
2003 , , 409 , 79 flores , h. , et al . 1999 ,
apj , 517 , 148 frster schreiber , n.m . , sauvage , m. , charmanfaris , v. , laurent , o. , gallais , p. , mirabel , i.f .
, & vigoux , l. 2003 , , 399 , 833 franceschini , a. , et al .
2003 , a&a , 403 , 501 hamuy , m. , et al .
1992 , pasp , 104 , 533 hamuy , m. , et al .
1994 , pasp , 106 , 566 hauser , m.g . ,
arendt , r.g .
, & kelsall , t. 1998 , apj , 508 , 25 hg , e. , et al . 2000 , , 355 , l27 kennicutt , r.c .
1992 , , 388 , 310 kessler , m.f . , et al .
1996 , , 315 , 27 lagache , g. , abergel , a. , boulanger , f. , dsert , f.x . , &
puget , j .- l .
1999 , a&a , 344 , 322 lari , c. , et al .
2001 , mnras , 325 , 1173 mann , r.g . ,
2002 , mnras , 332 , 549 mazzei , p. , aussel , h. , xu , c. , salvo , m. , de zotti , g. , & franceschini , a. 2001 , new astronomy , 6 , 265 metcalfe , l. , et al . 2003 , , 407 , 791 natali , f , , giallongo , e. , cristiani , s. , & la franca , f. 1998 , , 115 , 397 oke , j.b .
1990 , aj , 99 , 1621 oliver , s. , et al .
2000 , mnras , 316 , 749 oliver , s. , et al .
2002 , mnras , 332 , 536 osterbrock d.e .
1989 , astrophysics of gaseous nebulae and active galactic nuclei , uni .
books poggianti , b.m . ,
smail , i. , dressler , a. , couch , w. , barger , a.j . , butcher , h. , ellis , r.s . , &
oemler , a. 1999 , , 529 , 157 poggianti , b.m . , &
wu , h. 2000 , , 529 , 157 pozzi , f. , et al .
2003 , mnras,343 , 1148 pozzi , f. , et al .
2004 , apj , in press [ astro - ph/0403242 ] rowan - robinson m. , et al .
2004 , mnras , in press [ astro - ph/0308283 ] rush , b. , malkan , m.a . , & spinoglio , l. 1993 , , 89 , 1 salpeter , e.e .
1955 , vistas in astronomy , 1 , 283 silva , l. , granato , g.l . , bressan , a. , & danese , l. 1998 , apj , 509 , 103 sturm , e. , lutz , d. , tran , d. , feuchtgruber , h. , genzel , r. , kunze , d. , moorwod , a.f.m . ,
thornley , m.d . , 2000 , , 358 , 481 sutherland , w. , & saunders w. 1992 , mnras , 259 , 413 visnen , p. , et al .
2002 , , 337 , 1043 veilleux , s. , & osterbrock , d.e .
1987 , apjs , 63 , 295 ccc 0.45 & 0.001 & 0.000 + 0.50 & 0.007 & 0.000 + 0.55 & 0.023 & 0.000 + 0.60 & 0.053 & 0.000 + 0.65 & 0.110 & 0.001 + 0.70 & 0.167 & 0.004 + 0.75 & 0.239 & 0.009 + 0.80 & 0.312 & 0.019 + 0.85 & 0.379 & 0.031 + 0.90 & 0.442 & 0.049 + 0.95 & 0.501 & 0.076 + 1.00 & 0.545 & 0.107 + 1.05 & 0.585 & 0.147 + 1.10 & 0.625 & 0.189 + 1.15 & 0.668 & 0.232 + 1.20 & 0.708 & 0.278 + 1.25 & 0.743 & 0.323 + 1.30 & 0.776 & 0.373 + 1.35 & 0.807 & 0.415 + 1.40 & 0.841 & 0.465 + 1.45 & 0.866 & 0.506 + 1.50 & 0.889 & 0.559 + 1.60 & 0.929 & 0.646 + 1.70 & 0.957 & 0.699 + 1.80 & 0.977 & 0.743 + 1.90 & 0.991 & 0.781 + 2.00 & 1.000 & 0.813 + 2.10 & 1.000 & 0.841 + 2.20 & 1.000 & 0.867 + 2.30 & 1.000 & 0.888 + 2.40 & 1.000 & 0.906 + 2.50 & 1.000 & 0.921 + 2.60 & 1.000 & 0.934 + 2.70 & 1.000 & 0.945 + 2.80 & 1.000 & 0.954 + 2.90 & 1.000 & 0.962 + 3.00 & 1.000 & 0.969 + 3.10 & 1.000 & 0.974 + 3.20 & 1.000 & 0.979 + 3.30 & 1.000 & 0.983 + 3.40 & 1.000 & 0.986 + 3.50 & 1.000 & 0.989 + 3.60 & 1.000 & 0.991 + 3.70 & 1.000 & 0.993 + 3.80 & 1.000 & 0.995 + 3.90 & 1.000 & 0.996 + 4.00 & 1.000 & 0.997 + 4.10 & 1.000 & 0.999 + 4.20 & 1.000 & 1.000 + [ tab_compl ] ccrcccccrrrcrrrc elaisc15_j003957 - 432013&8 & 3.01&00:39:57.85&-43:20:13.4 & 1.2 & 17.21&15.32&363.16 & 0.999 & @xmath1 0.24 & 0.128&5 & 9.78&10.27 & + elaisc15_j003958 - 441511&9 & 25.39&00:39:58.67&-44:15:10.6 & 2.0 & 10.41 & 7.39&999.99 & 1.000 & @xmath1 0.30 & 0.000&8 & & & + elaisc15_j004000 - 431325&8 & 1.01 & & & & & & & & @xmath120 & & 0 & & & + elaisc15_j004004 - 433710&9 & 1.24 & & & & & & & & @xmath120 & & 0 & & & + elaisc15_j004007 - 432516&8 & 3.11&00:40:06.88&-43:25:18.2 & 2.5 & 12.21 & 9.46&999.99 & 1.000 & @xmath1 0.30 & 0.000&8 & & & + elaisc15_j004009 - 434424&9 & 2.62&00:40:09.26&-43:44:25.3 & 0.9 & 19.13 & & 97.33 & 0.997 & @xmath1 0.30 & 0.188&2&10.15 & 9.87 & + elaisc15_j004010 - 432012&8 & 1.03 & & & & & & & & @xmath120 & & 0 & & & + [ tab_id ] crccrrrrrrrr elaisc15_j002904 - 432415 & 1.61&0.207&5 & 1.20 & -8.9 & 3.1 & -3.7 & 0.0 & -56.7 & -10.6 & -12.0 + elaisc15_j002915 - 430333 & 2.04&0.417&5 & 1.23 & -39.2 & 3.4 & -10.3 & -13.9 & & & + elaisc15_j002924 - 432233 & 2.34&0.374&5@xmath121 & 1.17 & -39.4 & 0.0 & -3.4 & 0.0 & -36.4 & -47.3 & -46.3 + elaisc15_j002939 - 430625 & 3.73&0.071&5 & 1.22 & -11.8 & 3.6 & -1.9 & 0.0 & -22.7 & -8.6 & -8.4 + elaisc15_j002949 - 430703 & 1.33&0.302&2 & 1.40 & -7.4 & 0.0 & 0.0 & -25.0 & -15.8 & -23.8 & -12.9 + elaisc15_j003001 - 432202 & 1.47&0.274&5 & 1.33 & -3.4 & 3.4 & 0.0 & 0.0 & -28.1 & -8.0 & 0.0 + elaisc15_j003011 - 432947 & 1.35&0.084&6 & 1.87 & -2.6 & 0.0 & -1.1 & 0.0 & -4.8 & -3.4 & 0.0 + [ tab_lines ] rcccccccc 0.60 & 370@xmath9514 & 349 @xmath9513 & [email protected] & 13.6 @xmath952.6 & 7.9 @xmath952.0 + 0.76 & 242@xmath9511 & 221 @xmath9510 & [email protected] & 13.6 @xmath952.6 & 7.9 @xmath952.0 + 0.95 & [email protected] & 122.2 @xmath957.8 & [email protected] & 10.7 @xmath952.3 & 7.9 @xmath952.0 + 1.20 & [email protected] & 57.6 @xmath955.4 & [email protected] & 6.1 @xmath951.7 & 6.4 @xmath951.8 + 1.51 & [email protected] & 32.7 @xmath954.0 & [email protected] & 4.8 @xmath951.6 & 4.9 @xmath951.6 + 1.90 & [email protected] & 20.7 @xmath953.2 & [email protected] & 3.2 @xmath951.3 & 3.9 @xmath951.4 + 2.39 & [email protected] & 13.3 @xmath952.6 & [email protected] & 2.9 @xmath951.2 & 3.1 @xmath951.2 + 3.01 & [email protected] & 9.3 @xmath952.2 & [email protected] & 1.9 @xmath951.0 & 2.9 @xmath951.2 + 3.79 & [email protected] & 5.9 @xmath951.7 & [email protected] & 1.2 @xmath950.8 & 2.4 @xmath951.1 + 4.77 & [email protected] & 4.3 @xmath951.5 & [email protected] & 1.0 @xmath950.7 & 2.1 @xmath951.0 + 6.00 & [email protected] & 2.8 @xmath951.2 & [email protected] & 1.0 @xmath950.7 & 1.8 @xmath951.0 + 7.55 & [email protected] & [email protected] & [email protected] & ... & 1.4 @xmath950.8 + 9.51 & [email protected] & [email protected] & [email protected] & ... & 1.2 @xmath950.8 + 11.97 & [email protected] & [email protected] & [email protected] & ... & ... + 15.07 & [email protected] & [email protected] & ... & ... & ... + 18.98 & [email protected] & [email protected] & ... & ... & ... + [ tab_counts ] | we present a multi - wavelength catalog ( 15 @xmath0 , r , k - band , 1.4 ghz flux ) plus spectroscopic identifications for 406 15 @xmath0 sources detected in the european large area _ iso _ survey ( elais ) region s1 , over the flux density range 0.5@xmath1@xmath2@xmath1150 mjy . 332 ( @xmath382% ) sources are optically identified down to [email protected]
. spectra or bona fide stellar identifications are obtained for 290 objects ( @xmath388% of the optically identified sources ) .
the areal coverage , mid - infrared ( mir ) and optical completeness of the sample are discussed in order to allow statistical and evolutionary analyses .
two main spectroscopic classes have been found to dominate the mir extragalactic population : @xmath4 star - forming galaxies ( from absorbed to extreme starbursts : @xmath5@xmath6 l@xmath7 ) , which account for @xmath375% of the sources , and active galactic nuclei ( agn ; both type 1 and 2 ) , which account for @xmath325% of the sources .
about 20% of the extragalactic sources are dust - enshrouded starburst galaxies [ e(a ) spectra ] , and all the starburst galaxies appear more dust extincted in the optical than nearby normal galaxies .
we also identified 91 stellar objects ( @xmath322% of the mir sources ) .
the counts for starburst galaxies and agn down to 0.6 mjy have been derived .
a general trend is found in the optical - mir spectral energy distribution ( sed ) of the galaxies , where the mir - luminous objects have larger mir to optical luminosity ratios . based on a variety of analyses
, we suggest that the elais sources fainter than r@xmath323 are luminous and ultra - luminous mir galaxies ( lig - uligs ; @xmath5=10@xmath8 - 10@xmath9 l@xmath10 ) at intermediate redshifts ( @xmath11=0.51.5 ) , and that consequently the present sample is virtually 100% spectroscopically complete up to @xmath11=0.5 . |
in the past few years , there has been a large volume of literature regarding the applications of the nonlinear schrdinger equation ( nls ) in the presence of the so - called _ nonlinearity management _
( often referred to also as feshbach resonance management ) .
the nls is a prototypical dispersive nonlinear wave equation of the form : @xmath0 where @xmath1 is a complex envelope field , @xmath2 is an external potential , @xmath3 is a time - periodic nonlinearity coefficient , and @xmath4 is the laplacian operator with @xmath5 , @xmath6 .
nonlinearity management arises in applications in optics for transverse beam propagation in layered optical media @xcite , as well as in atomic physics for the feshbach resonance of the scattering length of inter - atomic interactions in bose - einstein condensates ( becs ) @xcite . in the latter case ,
the periodic variation of @xmath3 through an external magnetic field has been used as a means of producing robust matter - wave breathers in quasi - one - dimensional becs @xcite .
it has also been suggested that the nonlinearity management may prevent collapse - type phenomena in higher dimensions @xcite .
theoretical studies of the nonlinearity management were performed with a reduction of the time - periodic pde problem ( [ eqn1 ] ) to a time - periodic ode problem using a variational method @xcite and a moment method @xcite .
the physical relevance of the time - periodic nls equation , as evidenced by the above works , led to further developments in analysis of the pde problem ( [ eqn1 ] ) .
as an example , the success of the averaging theory @xcite for optical solitons in the presence of _ strong dispersion management _ , led to an analogous development for _ strong nonlinearity management _ produced originally in @xcite and systematized in @xcite .
the time - periodic nls equation ( [ eqn1 ] ) is averaged in the limit @xmath7 , where @xmath8 measures a short period and the large variation of @xmath3 in the decomposition @xmath9 where @xmath10 , @xmath11 has a unit period and zero mean .
after the averaging procedure , the time - periodic pde problem ( [ eqn1 ] ) is governed by the averaged nls equation , @xmath12 where @xmath13 and @xmath14 is the mean - zero anti - derivative of @xmath10 .
derivation and local well - posedness of solutions of the averaged nls equation ( [ eqn0 ] ) in function space @xmath15 are reviewed for @xmath16 in @xcite . in the present paper
, we consider global well - posedness of solutions of the averaged nls equation ( [ eqn0 ] ) in the energy space @xmath17 for @xmath6 .
the use of @xmath18 and @xmath6 seems more appropriate for physical applications of the averaged model ( [ eqn0 ] ) . in particular
, we address the question whether the averaged nls equation with a nonlinearity management @xmath19 arrests the blowup of solutions of the nls equation in two and three dimensions that would occur if no nonlinearity management was applied @xmath20 ) .
we show that the averaged nls equation is globally well - posed and no blowup of solutions occurs for @xmath21 .
this is demonstrated initially , in section 2 , from the point of view of scaling arguments .
the statement is subsequently proved using rigorous estimates in section 3 . in section 4
we compare the above conclusion and earlier works where possibilities of blowup of solutions of the full time - periodic nls equation ( [ eqn1 ] ) have been reported .
section 5 summarizes our findings .
the averaged nls equation ( [ eqn0 ] ) has a standard hamiltonian form ( see @xcite ) with the hamiltonian functional : @xmath22 due to the gauge invariance , the averaged nls equation ( [ eqn0 ] ) also conserves the squared @xmath23 norm : @xmath24 solitary wave solutions of the averaged nls equation ( [ eqn0 ] ) are critical points of @xmath25 at the level set of fixed values of @xmath26 . using formal scaling arguments @xcite
( see also the review of ref .
@xcite ) , we consider a two - parameter family of dilatations : @xmath27 where @xmath28 are parameters and @xmath29 is a suitable function of @xmath30 .
the squared @xmath23 norm ( [ eqn2a ] ) is preserved by the dilatations ( [ eqn4a ] ) whenever @xmath31 .
the hamiltonian ( [ eqn2 ] ) at the dilatations ( [ eqn4a ] ) is scaled as a function of parameter @xmath32 : @xmath33 where @xmath34 let us consider the case of no nonlinearity management and no external potential , when @xmath35 and @xmath36 .
it follows from ( [ eqn4 ] ) that the hamiltonian function @xmath37 is positive definite in the defocusing case , when @xmath38 . in the focusing case , when @xmath39 , the hamiltonian function @xmath37 is bounded from below for @xmath16 and @xmath40 , @xmath41 and is unbounded from below for @xmath40 , @xmath42 and @xmath43 , where @xmath44 when @xmath37 is unbounded from below as @xmath45 , the critical points of @xmath25 at a fixed value of @xmath26 ( i.e. , solitary wave solutions ) can not be stable for small width @xmath46 and instability of solitary waves implies a blowup of localized initial data in the time evolution of the cubic nls equation ( see @xcite for details ) . when the nonlinearity management is applied , the last term in the decomposition ( [ eqn4 ] ) always dominates and it preserves the boundness of @xmath37 from below for any @xmath47 .
this indicates on the level of formal scaling arguments that the blowup of solutions is arrested by the nonlinearity management term in the averaged nls equation ( [ eqn0 ] ) .
we shall prove this conjecture with rigorous analysis of well - posedness of solutions .
we also note that the first term in the decomposition ( [ eqn4 ] ) does not change the conclusions above if @xmath48 is a smooth non - negative potential , such that @xmath49 . typical examples of @xmath48 are parabolic magnetic traps , when @xmath50 , and periodic optical lattices , when @xmath51 .
the rigorous analysis of the local well - posedness of the averaged nls equation is not a trivial task .
in fact , to the best of our knowledge , one can not verify even the local existence and uniqueness of solutions to this problem .
the problem has been considered in one dimension @xmath16 by poppenberg where a local well - posedness result with data in @xmath52 was established @xcite . in higher dimensions
@xmath53 , one needs to require that the initial data be in @xmath54 for sufficiently large @xmath55,@xmath56 , that is @xmath57 ( see @xcite ) .
in addition , one needs to assume a `` non - trapping '' condition on the symbol of the second order operator .
this is a geometric condition , which depends on the profile of the initial data ( see @xcite for details ) . to summarize
, we can not state a precise condition under which the averaged nls equation has a ( local ) solution that preserves values of @xmath58 and @xmath59 constant in time .
we will however show , that whenever such a local solution exists for a short time @xmath60 , it can be extended globally for all @xmath61 . to that end
, we represent the hamiltonian @xmath25 in the form : @xmath62 where @xmath63 we consider the focusing case @xmath39 and prove that @xmath64 and @xmath65 are bounded by the two conserved quantities @xmath25 and @xmath26 .
first , we quote a variant of the gagliardo - nirenberg inequality ( see ( 1.1.16 ) on p. 15 in @xcite ) . for all @xmath66 , @xmath67 and @xmath68 , it is true for every function @xmath69 on @xmath70 that @xmath71 next , we define and use the fourier transform and its inverse for a function @xmath69 on @xmath70 : @xmath72 the plancherel s formula gives @xmath73 and the inequality @xmath74 is immediate from the definition .
let @xmath75 be a @xmath76 function with a compact support on @xmath77 $ ] , such that @xmath78 on @xmath79 $ ] .
let @xmath80 be a @xmath81 function , such that @xmath82 .
the ( smooth ) fourier multiplier @xmath83 is defined for every positive @xmath84 by @xmath85 .
equivalently , @xmath86 let @xmath87 .
then both @xmath88 are self - adjoint , bounded on @xmath23 operators .
the following statement is a modification of the sobolev embedding theorem .
[ le:1 ] there exists a constant @xmath89 , which depends only on the dimension @xmath6 , so that it is true for every function @xmath69 on @xmath70 that @xmath90 since @xmath91 , it will suffice to show that @xmath92 the bound follows simply by the plancherel s theorem and @xmath93 .
indeed , @xmath94 by duality , the bound follows from the estimate @xmath95 . that is trivial as well , since
@xmath96 the positive constant @xmath97 can be taken to be the square root of the volume of the ball in @xmath98 with radius @xmath99 .
the central result of our analysis is the following theorem .
[ theo:1 ] there exist @xmath100 and @xmath101 , so that for every @xmath102 , it is true for every function @xmath103 on @xmath70 that @xmath104 let @xmath105 with @xmath106 in and obtain @xmath107 next , we set @xmath108 and obtain @xmath109 for some positive constant @xmath110 .
we have used here that @xmath111 by the gagliardo - nirenberg inequality ( [ gn ] ) , we have @xmath112 , such that the last inequality is rewritten in the form : @xmath113 where in the last line we have used the cauchy - schwartz inequality : @xmath114 let @xmath115 , where @xmath116 .
then , the term @xmath117 can be estimated from the bound as follows : @xmath118 for some constants @xmath119 and @xmath120 . by the gagliardo - nirenberg inequality ( [ gn ] )
, we have @xmath121 , such that the upper bound for @xmath122 follows from ( [ cs ] ) and : @xmath123 which is the desired upper bound . as a corollary of the main theorem , we pick @xmath124 and immediately obtain the following upper bounds .
there exists constants @xmath125 and @xmath126 that depend on @xmath127,@xmath128,@xmath129 , so that @xmath130 and @xmath131 since @xmath25 and @xmath26 are conserved in the time evolution , the cauchy problem for the averaged nls equation ( [ eqn0 ] ) has global solutions in the energy space @xmath17 , if the initial data @xmath132 gives rise to local solutions in the same energy space .
therefore , the blowup of solutions of the nls equation with @xmath35 in @xmath53 is arrested by the nonlinearity management for any @xmath47 .
we have proven that the blowup does not occur in the averaged nls equation ( [ eqn0 ] ) in higher dimensions @xmath53 for @xmath21 .
this result raises the question whether the blowup of solutions is arrested in the full nls equation ( [ eqn1 ] ) for any non - zero variance of the time - periodic nonlinearity coefficient @xmath3 .
we address this question within the ode reduction of the time - periodic problem , which was considered recently with a variational method @xcite and a moment method @xcite . in both cases ,
the time evolution of the radially symmetric localized solutions of the full nls equaiton ( [ eqn1 ] ) is approximated by a time - dependent , generalized ermakov - pinney @xcite equation : @xmath133 where @xmath134 is an effective width of a localized solution , while @xmath135 are constants found from an initial data , such that @xmath136 ( see @xcite for details ) .
we shall consider the critical case @xmath40 and rewrite the ode ( [ eqn5 ] ) with the nonlinearity coefficient @xmath3 in ( [ eqn2aa ] ) in the explicit form : @xmath137 where @xmath138 , @xmath139 , and @xmath140 is a mean - zero @xmath8-periodic function of @xmath141 .
conditions for blowup and existence of bounded oscillations in solutions of the ode ( [ eqn5a ] ) were recently reviewed in @xcite ( see also references therein ) .
the sufficient condition for the blowup ( when @xmath142 in a finite time @xmath143 ) is @xmath144 the sufficient condition for the unbounded growth of the solution width ( when @xmath145 as @xmath146 ) is @xmath147 the necessary condition for the bounded oscillations of @xmath148 for any @xmath149 is @xmath150 numerical simulations in @xcite showed that the condition ( [ suf-2 ] ) was also sufficient in the case @xmath151 for the blowup arrest for any @xmath149 .
it is obvious from the explicit scaling that @xmath152 in the asymptotic limit @xmath7 .
therefore , the second condition ( [ suf-2 ] ) is satisfied for @xmath151 and solutions of the time - periodic ode ( [ eqn5a ] ) do not collapse , in agreement with our results derived for the averaged nls equation ( [ eqn0 ] ) .
previous work @xcite ( see also the review in @xcite ) has also addressed the averaged nls equation ( [ eqn0 ] ) in the limit of _ weak _ nonlinearity management , when @xmath10 is rescaled as @xmath153 and the parameter @xmath154 is small in the limit @xmath7 as @xmath155 .
solutions of the averaged nls equation ( [ eqn0 ] ) at the leading order @xmath156 blow up in a finite time , but the small @xmath157-terms formally stabilize the blow - up for any @xmath158 .
although this approximation of the critical nls equation has been considered in many applications of nonlinear optics ( see sections 4 - 5 in @xcite and references therein ) , it is clearly insufficient for a correct identification of the domain , where the blowup of solutions occurs . indeed , while the averaged nls equation ( [ eqn0 ] ) with small @xmath154 predicts no blowup of solutions , it is clear that the weak nonlinearity management corresponds to the case @xmath159 ) ) collapse in the domain ( [ suf-1 ] ) .
following the work @xcite , we address the failure of the averaging procedure for weak nonlinearity management in a simple time - periodic ode problem : @xmath160 where @xmath151 , @xmath161 and @xmath162 are order of @xmath163 in the limit @xmath7 . by using the formal asymptotic multi - scale expansion method ( see @xcite for details ) , we construct an asymptotic solution to the problem ( [ eqn5b ] ) : @xmath164 where @xmath165 and @xmath166 are recursively found from the set of linear inhomogeneous problems , @xmath167 the mean - value term @xmath168 satisfies an extended dynamical equation that excludes secular growth of the correction terms of the series ( [ series ] ) in @xmath169 : @xmath170 the averaged ode problem ( [ ode ] ) is an equation of motion for an effective particle with a coordinate @xmath168 in the potential field with an effective potential energy : @xmath171 when @xmath151 and @xmath172 , the particle with @xmath173 reaches @xmath174 at a finite time @xmath175 , that indicates the blowup of a localized solution . when the next @xmath176-term is taken into account ( as in the approximation of weak nonlinearity management @xcite ) , the blow - up is arrested and the mean - value term @xmath168 oscillates in an effective minimum of the potential energy @xmath177 , truncated at @xmath176-terms . when the next @xmath178-term is taken into account ( beyond the approximation of weak nonlinearity management ) , the potential energy @xmath177 with @xmath151 does not prevent the blowup of the localized solution depending on the initial data @xmath179 .
existence versus non - existence of blowup depends on the ratio of parameters @xmath162 but the difference can only be detected in the averaging method if convergence of the power series ( [ potential ] ) is established in a closed analytical form .
similarly , the averaged nls equation ( [ eqn0 ] ) can not be used in the limit of weak nonlinearity management for an accurate prediction of existence versus non - existence of blowup of solutions . in order to illustrate this point ,
we have performed numerical simulations of the full nls equation ( [ eqn1 ] ) in @xmath40 with @xmath180 we have observed that collapse of localized initial data does occur ( see solid curve on fig .
[ qfig1 ] ) by monitoring the half - width ( of one dimensional slices along @xmath181 ) of the wavefunction ( for radially symmetric gaussian initial data ) , @xmath182 ) with the same parameters show that the half - width @xmath55 never decreased below @xmath183 ( see dashed curve on fig .
[ qfig1 ] ) indicating the absence of collapse in accordance with the rigorous results presented above .
therefore , the averaged nls equation ( [ eqn0 ] ) can only be used for modeling of the blowup arrest in the limit of strong nonlinearity management of the full nls equation ( [ eqn1 ] ) when @xmath184 and @xmath8 is small .
in conclusion , we have studied the global well - posedness of solutions of the averaged nls equation that describe strong nonlinearity management of the time - periodic nls equation .
we have showed with formal scaling arguments and rigorous analysis that the blowup of solutions in higher dimensions is arrested within the averaged nls equation .
we have also discussed the non - applicability of the averaged nls equation to the weak nonlinearity management , where the blowup of solutions can occur beyond the weak management limit .
it is an open problem to study well - posedness of the full time - periodic nls equation , depending on parameters of the nonlinearity management and profile of initial data .
rigorous results on the latter problem are only available within the ode approximation ( [ eqn5 ] ) , when the pde model is reduced to a dynamical system with one degree of freedom .
it would be particularly interesting to study mathematically and to examine numerically whether the theoretical prediction from the method of moments provides an optimal bound for the full pde model with arbitrary initial data .
* acknowledgements*. the first author is supported in part by nsf - dms 0204585 , an nsf career award and the eppley foundation for research .
the second author supported in part by nsf - dms 0300511 .
the third author is supported by nserc and prea .
the authors are thankful to prof .
frantzeskakis for numerous enlightening discussions on the subject .
s. inouye _
et al . _ , nature * 392 * , 151 ( 1998 ) ; j. stenger _
et al . _ ,
. lett . * 82 * , 2422 ( 1999 ) ; j.l .
roberts _ et al .
lett . * 81 * , 5109 ( 1998 ) ; s.l .
et al . _ ,
lett . * 85 * , 1795 ( 2000 ) ; e.a .
et al . _ ,
nature * 412 * , 295 ( 2001 ) . | in the present short communication , we revisit nonlinearity management of the time - periodic nonlinear schrdinger equation and the related averaging procedure .
we prove that the averaged nonlinear schrdinger equation does not support the blow - up of solutions in higher dimensions , independently of the strength in the nonlinearity coefficient variance .
this conclusion agrees with earlier works in the case of strong nonlinearity management but contradicts those in the case of weak nonlinearity management .
the apparent discrepancy is explained by the divergence of the averaging procedure in the limit of weak nonlinearity management . |
the two - stage protocol for physical layer network coding for the two - way relay channel first introduced in @xcite , exploits the multiple access interference occurring at the relay so that the communication between the end nodes can be done using a two stage protocol . the works in @xcite ,
@xcite deal with the information theoretic studies for bidirectional relaying . in @xcite , modulation schemes to be used at the nodes for uncoded transmission for the two - way relaying were studied .
the work done for the relay channels with three or more user nodes is given in @xcite .
in @xcite , authors have proposed a two stage operation for three - way relaying called joint network and superposition coding , in which the three users transmit to the relay node one - by - one in the first phase , and the relay node makes two superimposed xor - ed packets and transmits back to the users in the bc phase .
the packet from the node with the worst channel gain is xor - ed with the other two packets .
the protocol employs four channel uses , three for the ma phase and one for the bc phase .
it is claimed by the authors that this scheme can be extended to more than three users as well . in the work by pischella and ruyet in @xcite a lattice - based coding scheme combined with power control , composed of alternate ma and bc phases , consisting of four channel uses for three - way relaying
is proposed .
the relay receives an integer linear combination of the symbols transmitted by the user nodes .
it is stated that the scheme can be extended to more number of users .
these two works essentially deal with the information theoretic aspects of multi - way relaying .
an ` opportunistic scheduling technique ' for physical network coding is proposed by authors jeon et al . in @xcite , where using a channel norm criterion and a minimum distance criterion , users in the ma as well as the bc phase are selected on the basis of instantaneous snr .
this approach utilizes six channel uses in case of three - way relaying and it is mentioned that the approach can be extended to more number of users . in @xcite , a ` latin square - like condition ' for the three - way relay channel network code
is proposed and cell swapping techniques on these latin cubes are suggested in order to improve upon these network codes .
the protocol employs five channel uses , and the channel gains associated with the channels are not considered in the construction of this network coding map .
-way relay channel , height=132 ] we consider the @xmath0-way wireless relaying scenario shown in fig .
1 , where @xmath0-way data transfer takes place among the nodes @xmath7 , @xmath8 , ... , @xmath9 with the help of the relay r assuming that the @xmath0 nodes operate in half - duplex mode .
the relaying protocol consists of two phases , _ multiple access _ ( ma )
phase , consisting of one channel use during which @xmath7 , @xmath8 , ... , @xmath9 transmit to r ; and _ broadcast _ ( bc ) phase , in which r transmits to @xmath7 , @xmath8 , ... , @xmath9 in a single channel use .
network coding is employed at r in such a way that @xmath10 can decode @xmath11 s message for @xmath12 and @xmath13 , given that @xmath10 knows its own message .
latin cubes have been explored as a tool to find the network coding map used by the relay , depending on the channel gain in @xcite .
the throughput performance of the two stage protocol for three - way relaying given in @xcite is better than the throughput performance of the ` opportunistic scheduling technique ' given in @xcite at high snr , as can be observed from the plots given in @xcite .
the work in @xcite further extends the approach used in @xcite to four - way relaying and employs two channel uses for the entire information exchange amongst the four users , which makes the throughput performance of the scheme better than the other existing schemes .
this scheme that utilizes two channel uses for the entire information exchange between three and four users using a relay in @xcite and @xcite respectively , is extended to @xmath0 users in this paper , for the case when @xmath5-psk is used at the end nodes .
for our physical layer network coding strategy we use the mathematical structure called a latin hyper - cube defined as follows : an _ n - fold latin hyper - cube l of r - th order of side @xmath5 _ @xcite is an @xmath14 array containing @xmath15 entries , @xmath16 of each of @xmath17 kinds , such that each symbol occurs at most once for each value taken by each dimension of the hyper - cube . for our purposes
, we use only @xmath0-fold latin hyper - cubes of side @xmath5 on the symbols from the set @xmath18 , @xmath19 .
* _ multiple access ( ma ) phase : _ * + suppose user node @xmath20 for @xmath21 wants to send a @xmath22-bit binary tuple to all user nodes @xmath23 for @xmath24 and @xmath25
. the symmetric @xmath5-psk @xmath26 constellation , given by @xmath27 is used at @xmath7 , @xmath8 , ... , @xmath9 , and @xmath28 denotes the map from bits to complex symbols used at @xmath7 , @xmath8 , ... ,
@xmath9 where @xmath29 .
let @xmath30 denote the complex symbols transmitted by @xmath7 , @xmath8 , ... , @xmath9 respectively , where @xmath31 . here , we assume that the channel state information ( csi ) is not available at the transmitting nodes and perfect csi is available at the receiving nodes .
the received signal at r in the ma phase is given by , @xmath32where @xmath33 , @xmath34 , ... , @xmath35 are the fading coefficients associated with the @xmath7-r , @xmath8-r , ... , @xmath9-r link respectively .
the additive noise @xmath36 , where @xmath37 denotes the circularly symmetric complex gaussian random variable with variance @xmath38 .
the effective constellation seen at the relay during the ma phase , denoted by @xmath39 , is given by , + @xmath40 the minimum distance between the points in the constellation @xmath41 denoted by @xmath42 is given in ( [ dist ] ) on the next page . from ( [ dist ] ) , it is clear that there exists values of @xmath43 , for which @xmath44 .
a fade state @xmath45 is defined to be a _ singular fade state _ for the ma phase of @xmath0-way relaying , if @xmath44 .
let @xmath46 denote the set of all singular fade states . for singular fade states , @xmath47 .
let the maximum likelihood ( ml ) estimate of @xmath48 be denoted by @xmath49 at r based on the received complex number @xmath50 , i.e. , @xmath51where @xmath52 $ ] and @xmath53^t.$ ] @xmath54 * _ broadcast ( bc ) phase : _ * + during the bc phase , the received signals at @xmath55 are respectively given by , @xmath56 where @xmath57 denotes the complex number transmitted by r and @xmath58 @xmath59 ... , @xmath60 respectively are the fading coefficients corresponding to the links r-@xmath7 , r-@xmath8 , ... , r-@xmath9 .
the additive noises @xmath61 @xmath62 , ... , @xmath63 are @xmath64 . during the _ bc _ phase
, r transmits a point from a signal set @xmath65 given by a many to one map @xmath66 chosen by r , depending on the values of @xmath67 , @xmath68 , ... , @xmath69 .
the cardinality of @xmath70 , since @xmath71 bits about the other @xmath3 users needs to be conveyed to each of @xmath7 , @xmath8 , ... , @xmath9 .
a _ cluster _ is the set of elements in @xmath72 which are mapped to the same signal point in @xmath65 by the map @xmath73 .
let @xmath74 denote the set of all such clusters . the _ cluster distance _ between a pair of clusters @xmath75 ,
as given in ( [ cl1 ] ) on the next page , is the minimum among all the distances calculated between the points @xmath76 and @xmath77 in the effective constellation seen at the relay node r. the minimum among all the cluster distances among all pairs of clusters of a clustering @xmath78 is the _ minimum cluster distance _ of the clustering , as given in ( [ cl2 ] ) on the next page . during the ma phase , the performance depends on the minimum cluster distance , while during the bc phase ,
the performance is dependent on the minimum distance of the signal set @xmath65 .
_ distance shortening _ , a phenomenon given in @xcite ,
is described as the significant reduction in the value of @xmath79 for values of @xmath80 in the neighborhood of the singular fade states .
if the clustering used at the relay node r in the bc phase is chosen such that @xmath81 is non zero , then the effect of distance shortening can be avoided .
a clustering @xmath82 is said to remove a singular fade state @xmath83 if @xmath84 , i.e. , any two message sequences @xmath85 that coincide in the effective constellation received at the relay during the ma phase is in the same cluster of @xmath82 .
so , removing singular fade states for a @xmath0-way relay channel can alternatively be defined as : a clustering @xmath82 is said to _ remove the singular fade state _ @xmath86 , if any two possibilities of the messages sent by the users @xmath87 that satisfy @xmath88 are placed together in the same cluster by the clustering .
we denote the clustering which removes the singular fade state @xmath89 by @xmath90 ( selecting one randomly if there are multiple clusterings which remove the same singular fade state @xmath89 ) .
let the set of all such clusterings be denoted by @xmath91 , i.e. , @xmath92 .
the minimum cluster distance of the clustering @xmath93 for @xmath94 , when the fade state @xmath43 occurs in the ma phase , denoted by @xmath95 , is the minimum among all its cluster distances . if @xmath96 the clustering @xmath97 is chosen to be @xmath98 , that satisfies , + @xmath99 + @xmath100 + @xmath101 .
the clustering used by the relay is indicated to @xmath7 , @xmath8, ... ,@xmath9 using overhead bits . in order to ensure that @xmath102 is able to decode the message sent by @xmath103 , the clustering @xmath104 should satisfy the exclusive law , as given in ( [ mel ] ) . we explain exclusive law in more detail in the next section .
the contributions of this paper are as follows : * we propose a scheme that enables the exchange of information in the wireless @xmath0-way relaying scenario when @xmath5-psk is used at the @xmath0 user nodes with totally two channel uses while attempting to remove the harmful effects of fading , extending the schemes given in @xcite , @xcite , @xcite for @xmath105 respectively . * for this scheme , the singular fade spaces are identified , enumerated and classified based on their removability in the given scenario . the remaining content is organized as follows : section iii demonstrates how a @xmath0-fold latin hyper - cube of side @xmath5 can be utilized to represent the network code that satisfies the exclusive law for @xmath0-way relaying when @xmath5-psk is used at the end nodes . in section
iv we describe and enumerate the singular fade subspaces for the given scenario and in section v , focus in on the removal of such singular fade subspaces using @xmath0-fold latin hyper - cube of side @xmath5 .
section vi provides some insights using simulations and section vii concludes the paper .
the clustering @xmath104 that represents the map used at the relay should satisfy the exclusive law @xcite in order to ensure that @xmath106 is able to decode the message sent by @xmath107 , where we assume that the nodes @xmath108 transmit symbols from the @xmath5-psk constellation . consider a @xmath14 array , containing @xmath15 entries indexed by @xmath109 , i.e. , the @xmath0 symbols sent by @xmath108 in the ma phase . for @xmath21 , fixing the @xmath110 dimension of this @xmath14 array , the @xmath5 @xmath3 dimensional arrays obtained , denoted by say @xmath111 , are indexed by the @xmath5 values taken by @xmath112 . for fixed values of @xmath113 and @xmath114 , the repetition of a symbol in @xmath115 results in the failure of the @xmath110 exclusive law given by ( [ mel ] ) .
thus , for the exclusive law to be satisfied , the cells of this array should be filled such that the @xmath14 array so obtained , is a @xmath0-fold latin hyper - cube of side @xmath5 , with entries from @xmath116 for @xmath117 ( definition 1 ) .
the symbol @xmath118 of a particular clustering @xmath119 denotes the cluster obtained by putting together all the tuples @xmath120 such that the entry in the @xmath121-th slot is the same entry @xmath122 from @xmath123 .
the adjoining figures fig .
[ fig : three ] and fig .
[ fig : four ] show the exclusive law condition for the three - way and four - way relaying scenario when 4-psk is used at end nodes .
a set @xmath124 consisting of all the possibilities of @xmath125 that must be placed in the same cluster of the clustering used at relay node r in the bc phase in order to remove the singular fade state @xmath89 , is referred to as a _ singularity removal constraint _ for the singular fade state @xmath89 for @xmath0-way relaying scenario . at the end of the ma phase ,
the relay node receives a complex number , given by ( [ yr ] ) . using
the ml estimate of this received complex number , r transmits a point from the constellation @xmath126 with cardinality at most @xmath127 .
instead of r transmitting a point from the @xmath127 point constellation resulting from all the possibilities of @xmath109 , depending on the fade states , the relay r can choose to group these possibilities into clusters represented by a smaller constellation , so that the minimum cluster distance is non - zero , as well as all the users receive the messages from the other @xmath3 users , i.e. , the clustering satisfies the exclusive law .
we provide one such clustering for the case of @xmath0-way relaying in the following .
suppose the fade coefficient in the ma phase , @xmath128 , is a singular fade state , and @xmath129 is a singularity removal constraint corresponding to the singular fade state @xmath130 .
then there exist @xmath131 , @xmath132 such that , @xmath133 \right\rangle ^{\bot } % \vspace{-.5 cm } \label{nullspacen } \vspace{-.2 cm}\end{aligned}\]]where for a @xmath134 non - zero vector @xmath135 over @xmath136 , @xmath137 note that @xmath138 is the dot product over @xmath136 ( and not an inner product over @xmath136 ) . for a @xmath134 non - zero vector @xmath135 over @xmath136
, @xmath139 is a @xmath3-dimensional vector subspace of @xmath140 over @xmath136 .
let @xmath141 where @xmath142 .
then , by definition , @xmath143 proving that the following subset of @xmath140 + @xmath144 + is a linearly independent set over @xmath136 of cardinality @xmath3 would be sufficient to prove that @xmath139 is a @xmath3-dimensional vector subspace of @xmath140 .
let @xmath145 such that @xmath146comparing the @xmath147 components of the lhs and rhs , we get , @xmath148 . since @xmath149 , where @xmath150 is finite , there are only finitely many possibilities for the right - hand side of ( [ nullspacen ] ) .
thus the uncountably infinite singular fade states @xmath43 , are points in a finite number of @xmath3-dimensional vector subspaces of @xmath140 over @xmath136
. we shall refer to these finite number of vector subspaces as the _ singular fade subspaces _ @xcite .
we now give a detailed description of all the possibilities of singular fade subspaces for @xmath0-way relaying scenario when @xmath5-psk is used at the end nodes .
for the proof of the following theorem , we extensively use the following lemmas given with proofs in @xcite . let @xmath151 denote the difference constellation of the signal set @xmath150 , i.e. , @xmath152 .
then , for any @xmath5-psk signal set , @xmath151 is of the form , @xmath153 where @xmath154 and @xmath155 . as a result of the above lemma ,
the non - zero points in @xmath151 lie on @xmath156 circles of radius @xmath157 with each circle containing @xmath5 points .
the phase angles of the @xmath5 points on each circle is @xmath158 , if @xmath114 is odd and @xmath159 if @xmath114 is even , where @xmath160 .
@xcite let @xmath161 be the ordered indices corresponding to the non - zero components in @xmath162 and @xmath163 ( the location of non - zero components is the same in the vectors @xmath162 and @xmath163 ) . for @xmath5-psk signal set , @xmath164 , for some @xmath165 , if and only if the magnitudes of the non - zero components in @xmath166 are equal and the magnitudes of the non - zero components in @xmath163 are equal , i.e. , @xmath167 and @xmath168 . from ( [ nullspacen ] ) and lemma 2 , the singular fade subspaces are given by
, @xmath169 \right\rangle ^{\bot } = \left\langle \left [ { \begin{array}{cc } 2 \sin ( \frac { \pi l_1}{m } ) e^{j m_1 } \\ 2 \sin ( \frac { \pi l_2}{m } ) e^{j m_2 } \\ . \\. \\ 2 \sin ( \frac { \pi l_n}{m } ) e^{j m_n } \\ \end{array } } \right ] \right\rangle ^{\bot}\ ] ] @xmath170 \right\rangle ^{\bot}\]]where @xmath171 if @xmath172 is odd and @xmath173 if @xmath172 is even , where @xmath174 for @xmath175 .
there are @xmath176 m^{k-1}$ ] singular fade subspaces for @xmath0-way relaying when @xmath5-psk constellation is used at the end nodes .
the singular fade subspaces are of the form @xmath177\right\rangle^{\bot } $ ] where @xmath178 .
let @xmath113 be the number of non - zero @xmath179 .
we fix the relative phase vector of the vector @xmath180 = w ~(say)$ ] .
the points in @xmath181 lie on @xmath156 circles .
so there are @xmath182 possibilities for absolute values of the non - zero components of @xmath183
. there are @xmath156 possibilities for the case that the absolute values of all the components of @xmath183 that are non - zero , are equal . from lemma 3 ,
the singular fade subspaces resulting from all of these @xmath156 cases are the same , and hence account for 1 out of the @xmath182 cases .
so for a fixed relative phase vector , there are @xmath184 $ ] singular fade subspaces . from lemma 3 , fixing the absolute values of the non - zero components of @xmath183 , each distinct relative phase vector corresponds to a distinct singular fade subspace .
there are @xmath185 distinct possibilities for the relative phase vector .
so , there are @xmath184m^{k-1}$ ] singular fade subspaces when @xmath183 has @xmath113 non - zero components . here
, @xmath113 can take values from 1 to @xmath0 .
summing over all possible values of @xmath113 , we have @xmath176 m^{n-1}$ ] singular fade subspaces for @xmath0-way relaying when @xmath5-psk constellation is used at the end nodes
. the above theorem coincides with the results given for @xmath186 in @xcite , and the results obtained using explicit enumeration for @xmath187 in @xcite and for @xmath188 in @xcite . for illustration , we discuss the case when @xmath189 . in five - way
relaying , user nodes ( say ) a , b , c , d and e transmit @xmath190 , @xmath191 , @xmath192 , @xmath193 and @xmath194 @xmath195 respectively in the first channel use .
suppose the fade coefficient in the ma phase , @xmath196 , is a singular fade state .
then there exist @xmath197 , @xmath198 such that , @xmath199 \right\rangle ^{\bot}. % \vspace{-.5 cm } \label{nullspace } \vspace{-.2 cm}\end{aligned}\ ] ] the adaptive network coding for five - way relaying attempts at removing the singular fade subspaces for the case given by , @xmath200 \right\rangle ^{\bot } $ ] . in the second channel use
, relay node r transmits @xmath201 using a network coding map that depends on the values of @xmath202 , @xmath203 , @xmath204 , @xmath205 and @xmath206 . as explained in section
iii , using a network coding map represented by a 5-fold hyper latin - cube of side @xmath5 ( when @xmath5-psk is used at the end nodes a , b , c , d and e ) ensures that exclusive law is satisfied .
it can be shown using explicit enumeration , that when the end nodes use 4-psk , there are @xmath207 singular fade subspaces for five - way relaying , which coincides with theorem 1 , for @xmath208 .
we cluster the possibilities of @xmath209 into a clustering that can be represented by an @xmath0-fold latin hyper - cubes of side @xmath5 , to obtain a clustering that removes the singular fade subspaces , and also attempts to minimize the size of the constellation used by r. this clustering is represented by a constellation given by @xmath126 , which is utilized by the relay node r in the bc phase .
this is done by first constraining the @xmath15 possibilities of @xmath209 transmitted at the ma phase , to remove the singular fade subspaces , and then using these constraints , filling the entries of an empty @xmath210 array representing the map to be used at the relay .
this partially filled array is completed so as to form a @xmath0-fold latin hyper - cube of side @xmath5 .
the mapping to be used at r can be obtained from the complete latin hyper - cube keeping in mind the equivalence between the relay map that satisfies the exclusive law with the @xmath0-fold latin hyper - cube of side @xmath5 as shown in section iii .
start with the constrained @xmath211 array @xmath212 initialize all empty cells of @xmath212 to 0 the @xmath213 cell of @xmath212 is the @xmath214 transmission of @xmath7 , the @xmath215 transmission of @xmath8 , ... , the @xmath216 transmission of @xmath9 . during the ma phase for
the @xmath0-way relaying scenario , nodes @xmath7 , @xmath8 , ... , @xmath9 transmit to the relay r. let the fade state @xmath43 denote a point in one of the @xmath176 m^{n-1}$ ] singular fade subspaces ( section iv ) .
the constraints on the @xmath15 array representing the map at the relay node r during bc phase for a singular fade state , can be obtained using the vectors of differences , viz .
, @xmath217 $ ] contributing to this particular singular fade state .
so , if @xmath218 \right\rangle^{\bot}$ ] , then , for @xmath219 , @xmath220 . for a clustering to remove the singular fade state @xmath43 , i.e. , for the minimum distance of the clustering to be greater than 0 ( section ii ) , the pair @xmath221 must be kept in the same cluster .
alternatively , we can say that the entry corresponding to @xmath109 in the @xmath15 array must be the same as the entry corresponding to @xmath222 . similarly , every other such pair in @xmath223 contributing to this same singular fade subspace must be kept in the same cluster .
apart from all such pairs in @xmath223 being kept in the same cluster of the clustering , in order to remove this particular fade state , there are no other constraints .
consider the ordering given as follows on the entries of the constrained @xmath15 array : @xmath224 if @xmath225 where @xmath122 is the first component among the @xmath0 components , where @xmath226 .
this constrained @xmath15 array can then be completed by simply filling the first empty cell in this order , with @xmath227 in the increasing order of @xmath122 such that the completed array is an @xmath0-fold latin hyper - cube of side @xmath5 ( algorithm 1 ) .
the above clustering scheme , however , can not be utilized to remove all the singular fade subspaces , as shown in the following lemma .
! width 1.ptc!width 0.9ptp0.3cm|p0.3cm|p0.3cm|p0.3cm!width 1.ptc!width 0.9ptp0.3cm|p0.32cm|p0.32cm|p0.32cm!width 1.ptc!width 0.9ptp0.32cm|p0.32cm|p0.32cm|p0.32cm!width 1.ptc!width 0.9ptp0.32cm|p0.32cm|p0.32cm|p0.32cm!width
1.pt @xmath228 & & & & & @xmath228 & & & & & @xmath228 & & & & & @xmath228 & & & & + @xmath229 & & & & & @xmath230 & & & & & @xmath231 & & & & & @xmath232 & & & & + 0 & @xmath233 & @xmath234 & @xmath235 & @xmath236 & 0 & @xmath237 & @xmath238 & @xmath239 & @xmath240 & 0 & @xmath241 & @xmath242 & @xmath243 & @xmath244 & 0 & @xmath245 & @xmath246 & @xmath247 & @xmath248 + 1 & @xmath249 & @xmath250 & @xmath251 & @xmath252 & 1 & @xmath253 & @xmath254 & @xmath255 & @xmath256 & 1 & @xmath257 & @xmath258 & @xmath259 & @xmath260 & 1 & @xmath261 & @xmath262 & @xmath263 & @xmath264 + 2 & @xmath265 & @xmath266 & @xmath267 & @xmath268 & 2 & @xmath269 & @xmath270 & @xmath271 & @xmath272 & 2 & @xmath273 & @xmath274 & @xmath275 & @xmath276 & 2 & @xmath277 & @xmath278 & @xmath279 & @xmath280 + 3 & @xmath281 & @xmath282 & @xmath283 & @xmath284 & 3 & @xmath285 & @xmath286 & @xmath287 & @xmath288 & 3 & @xmath289 & @xmath290 & @xmath291 & @xmath292 & 3 & @xmath293 & @xmath294 & @xmath295 & @xmath296 + @xmath297 & & & & & @xmath297 & & & & & @xmath297 & & & & & @xmath297 & & & & + @xmath229 & & & & & @xmath230 & & & & & @xmath231 & & & & & @xmath232 & & & & + 0 & @xmath298 & @xmath299 & @xmath300 & @xmath301 & 0 & @xmath302 & @xmath303 & @xmath304 & @xmath305 & 0 & @xmath306 & @xmath307 & @xmath308 & @xmath309 & 0 & @xmath310 & @xmath311 & @xmath312 & @xmath313 + 1 & @xmath314 & @xmath315 & @xmath316 & @xmath317 & 1 & @xmath318 & @xmath319 & @xmath320 & @xmath321 & 1 & @xmath322 & @xmath323 & @xmath324 & @xmath325 & 1 & @xmath326 & @xmath327 & @xmath328 & @xmath329 + 2 & @xmath330 & @xmath331 & @xmath332 & @xmath333 & 2 & @xmath334 & @xmath335 & @xmath336 & @xmath337 & 2 & @xmath338 & @xmath339 & @xmath340 & @xmath341 & 2 & @xmath342 & @xmath343 & @xmath344 & @xmath345 + 3 & @xmath346 & @xmath347 & @xmath348 & @xmath349 & 3 & @xmath350 & @xmath351 & @xmath352 & @xmath353 & 3 & @xmath354 & @xmath355 & @xmath356 & @xmath357 & 3 & @xmath358 & @xmath359 & @xmath360 & @xmath361 + @xmath362 & & & & & @xmath362 & & & & & @xmath362 & & & & & @xmath362 & & & & + @xmath229 & & & & & @xmath230 & & & & & @xmath231 & & & & & @xmath232 & & & & + 0 & @xmath363 & @xmath364 & @xmath365 & @xmath366 & 0 & @xmath367 & @xmath368 & @xmath369 & @xmath370 & 0 & @xmath371 & @xmath372 & @xmath373 & @xmath374 & 0 & @xmath375 & @xmath376 & @xmath377 & @xmath378 + 1 & @xmath379 & @xmath380 & @xmath381 & @xmath382 & 1 & @xmath383 & @xmath384 & @xmath385 & @xmath386 & 1 & @xmath387 & @xmath388 & @xmath389 & @xmath390 & 1 & @xmath391 & @xmath392 & @xmath393 & @xmath394 + 2 & @xmath395 & @xmath396 & @xmath397 & @xmath398 & 2 & @xmath399 & @xmath400 & @xmath401 & @xmath402 & 2 & @xmath403 & @xmath388 & @xmath404 & @xmath405 & 2 & @xmath406 & @xmath407 & @xmath408 & @xmath409 + 3 & @xmath410 & @xmath411 & @xmath412 & @xmath413 & 3 & @xmath414 & @xmath415 & @xmath416 & @xmath417 & 3 & @xmath418 & @xmath419 & @xmath420 & @xmath421 & 3 & @xmath422 & @xmath423 & @xmath424 & @xmath425 + @xmath426 & & & & & @xmath426 & & & & & @xmath426 & & & & & @xmath426 & & & & + @xmath229 & & & & & @xmath230 & & & & & @xmath231 & & & & & @xmath232 & & & & + 0 & @xmath427 & @xmath428 & @xmath429 & @xmath430 & 0 & @xmath431 & @xmath432 & @xmath433 & @xmath434 & 0 & @xmath435 & @xmath436 & @xmath437 & @xmath438 & 0 & @xmath439 & @xmath440 & @xmath441 & @xmath442 + 1 & @xmath443 & @xmath444 & @xmath445 & @xmath446 & 1 & @xmath447 & @xmath448 & @xmath449 & @xmath450 & 1 & @xmath451 & @xmath452 & @xmath453 & @xmath454 & 1 & @xmath455 & @xmath456 & @xmath457 & @xmath458 + 2 & @xmath459 & @xmath460 & @xmath461 & @xmath462 & 2 & @xmath463 & @xmath464 & @xmath465 & @xmath466 & 2 & @xmath467 & @xmath468 & @xmath469 & @xmath470 & 2 & @xmath471 & @xmath472 & @xmath473 & @xmath474 + 3 & @xmath475 & @xmath476 & @xmath477 & @xmath478 & 3 & @xmath479 & @xmath480 & @xmath481 & @xmath482 & 3 & @xmath483 & @xmath484 & @xmath485 & @xmath486 & 3 & @xmath487 & @xmath488 & @xmath489 & @xmath490 + !
width 1.ptc!width 0.9ptp0.32cm|p0.32cm|p0.32cm|p0.32cm!width 1.ptc!width 0.9ptp0.3cm|p0.32cm|p0.32cm|p0.32cm!width 1.ptc!width 0.9ptp0.32cm|p0.32cm|p0.32cm|p0.32cm!width 1.ptc!width 0.9ptp0.32cm|p0.32cm|p0.32cm|p0.32cm!width
1.pt @xmath228 & & & & & @xmath228 & & & & & @xmath228 & & & & & @xmath228 & & & & + @xmath229 & & & & & @xmath230 & & & & & @xmath231 & & & & & @xmath232 & & & & + 0 & @xmath491 & @xmath318 & @xmath492 & @xmath493 & 0 & @xmath494 & @xmath314 & @xmath317 & @xmath495 & 0 & @xmath496 & @xmath466 & @xmath342 & @xmath343 & 0 & @xmath323 & @xmath322 & @xmath497 & @xmath498 + 1 & @xmath499 & @xmath302 & @xmath500 & @xmath501 & 1 & @xmath347 & @xmath298 & @xmath502 & @xmath503 & 1 & @xmath504 & @xmath482 & @xmath505 & @xmath359 & 1 & @xmath307 & @xmath306 & @xmath506 & @xmath507 + 2 & @xmath319 & @xmath350 & @xmath508 & @xmath509 & 2 & @xmath315 & @xmath346 & @xmath510 & @xmath511 & 2 & @xmath416 & @xmath310 & @xmath512 & @xmath513 & 2 & @xmath355 & @xmath354 & @xmath514 & @xmath515 + 3 & @xmath516 & @xmath334 & @xmath517 & @xmath518 & 3 & @xmath519 & @xmath330 & @xmath333 & @xmath520 & 3 & @xmath401 & @xmath326 & @xmath327 & @xmath521 & 3 & @xmath339 & @xmath338 & @xmath522 & @xmath523 + @xmath297 & & & & & @xmath297 & & & & & @xmath297 & & & & & @xmath297 & & & & + @xmath229 & & & & & @xmath230 & & & & & @xmath231 & & & & & @xmath232 & & & & + 0 & @xmath254 & @xmath253 & @xmath524 & @xmath525 & 0 & @xmath250 & @xmath265 & @xmath526 & @xmath527 & 0 & @xmath262 & @xmath261 & @xmath528 & @xmath529 & 0 & @xmath274 & @xmath530 & @xmath531 & @xmath403 + 1 & @xmath238 & @xmath237 & @xmath532 & @xmath239 & 1 & @xmath234 & @xmath281 & @xmath533 & @xmath534 & 1 & @xmath246 & @xmath245 & @xmath535 & @xmath536 & 1 & @xmath290 & @xmath291 & @xmath537 & @xmath418 + 2 & @xmath286 & @xmath285 & @xmath538 & @xmath287 & 2 & @xmath282 & @xmath539 & @xmath540 & @xmath363 & 2 & @xmath294 & @xmath293 & @xmath541 & @xmath542 & 2 & @xmath243 & @xmath244 & @xmath543 & @xmath372 + 3 & @xmath270 & @xmath269 & @xmath544 & @xmath545 & 3 & @xmath266 & @xmath546 & @xmath547 & @xmath379 & 3 & @xmath278 & @xmath277 & @xmath548 & @xmath549 & 3 & @xmath550 & @xmath260 & @xmath551 & @xmath388 + @xmath362 & & & & & @xmath362 & & & & & @xmath362 & & & & & @xmath362 & & & & + @xmath229 & & & & & @xmath230 & & & & & @xmath231 & & & & & @xmath232 & & & & + 0 & @xmath337 & @xmath447 & @xmath552 & @xmath553 & 0 & @xmath444 & @xmath443 & @xmath554 & @xmath555 & 0 & @xmath556 & @xmath345 & @xmath557 & @xmath472 & 0 & @xmath452 & @xmath451 & @xmath558 & @xmath559 + 1 & @xmath353 & @xmath431 & @xmath560 & @xmath433 & 1 & @xmath428 & @xmath427 & @xmath561 & @xmath562 & 1 & @xmath563 & @xmath361 & @xmath564 & @xmath488 & 1 & @xmath436 & @xmath435 & @xmath565 & @xmath437 + 2 & @xmath432 & @xmath479 & @xmath434 & @xmath481 & 2 & @xmath476 & @xmath475 & @xmath289 & @xmath477 & 2 & @xmath313 & @xmath439 & @xmath566 & @xmath567 & 2 & @xmath484 & @xmath483 & @xmath568 & @xmath485 + 3 & @xmath448 & @xmath463 & @xmath450 & @xmath465 & 3 & @xmath460 & @xmath459 & @xmath273 & @xmath461 & 3 & @xmath329 & @xmath455 & @xmath569 & @xmath570 & 3 & @xmath468 & @xmath467 & @xmath571 & @xmath572 + @xmath426 & & & & & @xmath426 & & & & & @xmath426 & & & & & @xmath426 & & & & + @xmath229 & & & & & @xmath230 & & & & & @xmath231 & & & & & @xmath232 & & & & + 0 & @xmath384 & @xmath383 & @xmath573 & @xmath574 & 0 & @xmath380 & @xmath395 & @xmath382 & @xmath397 & 0 & @xmath392 & @xmath391 & @xmath394 & @xmath575 & 0 & @xmath576 & @xmath577 & @xmath578 & @xmath579 + 1 & @xmath368 & @xmath367 & @xmath580 & @xmath581 & 1 & @xmath364 & @xmath410 & @xmath582 & @xmath412 & 1 & @xmath376 & @xmath375 & @xmath583 & @xmath303 & 1 & @xmath419 & @xmath405 & @xmath584 & @xmath299 + 2 & @xmath415 & @xmath414 & @xmath417 & @xmath585 & 2 & @xmath411 & @xmath586 & @xmath413 & @xmath242 & 2 & @xmath423 & @xmath422 & @xmath587 & @xmath588 & 2 & @xmath589 & @xmath390 & @xmath590 & @xmath591 + 3 & @xmath400 & @xmath399 & @xmath402 & @xmath592 & 3 & @xmath396 & @xmath593 & @xmath398 & @xmath258 & 3 & @xmath407 & @xmath406 & @xmath409 & @xmath335 & 3 & @xmath374 & @xmath594 & @xmath595 & @xmath596 + !
width 1.ptc!width 0.9ptp0.32cm|p0.32cm|p0.32cm|p0.3cm!width 1.ptc!width 0.9ptp0.3cm|p0.32cm|p0.32cm|p0.32cm!width 1.ptc!width 0.9ptp0.3cm|p0.3cm|p0.3cm|p0.32cm!width 1.ptc!width 0.9ptp0.32cm|p0.32cm|p0.32cm|p0.32cm!width
1.pt @xmath228 & & & & & @xmath228 & & & & & @xmath228 & & & & & @xmath228 & & & & + @xmath229 & & & & & @xmath230 & & & & & @xmath231 & & & & & @xmath232 & & & & + 0 & @xmath335 & @xmath387 & @xmath597 & @xmath467 & 0 & @xmath331 & @xmath572 & @xmath468 & @xmath563 & 0 & @xmath593 & @xmath379 & @xmath410 & @xmath396 & 0 & @xmath461 & @xmath353 & @xmath598 & @xmath400 + 1 & @xmath372 & @xmath371 & @xmath599 & @xmath589 & 1 & @xmath421 & @xmath485 & @xmath600 & @xmath556 & 1 & @xmath586 & @xmath363 & @xmath601 & @xmath411 & 1 & @xmath477 & @xmath337 & @xmath602 & @xmath415 + 2 & @xmath388 & @xmath418 & @xmath603 & @xmath577 & 2 & @xmath559 & @xmath438 & @xmath604 & @xmath391 & 2 & @xmath412 & @xmath366 & @xmath605 & @xmath427 & 2 & @xmath581 & @xmath370 & @xmath606 & @xmath431 + 3 & @xmath303 & @xmath403 & @xmath607 & @xmath435 & 3 & @xmath299 & @xmath454 & @xmath313 & @xmath376 & 3 & @xmath397 & @xmath382 & @xmath380 & @xmath443 & 3 & @xmath574 & @xmath386 & @xmath608 & @xmath447 + @xmath297 & & & & & @xmath297 & & & & & @xmath297 & & & & & @xmath297 & & & & + @xmath229 & & & & & @xmath230 & & & & & @xmath231 & & & & & @xmath232 & & & & + 0 & @xmath482 & @xmath519 & @xmath373 & @xmath416 & 0 & @xmath470 & @xmath592 & @xmath609 & @xmath516 & 0 & @xmath398 & @xmath596 & @xmath610 & @xmath460 & 0 & @xmath402 & @xmath465 & @xmath611 & @xmath612 + 1 & @xmath466 & @xmath401 & @xmath613 & @xmath273 & 1 & @xmath486 & @xmath585 & @xmath614 & @xmath407 & 1 & @xmath413 & @xmath591 & @xmath615 & @xmath476 & 1 & @xmath417 & @xmath481 & @xmath616 & @xmath617 + 2 & @xmath258 & @xmath494 & @xmath618 & @xmath452 & 2 & @xmath575 & @xmath491 & @xmath619 & @xmath455 & 2 & @xmath562 & @xmath430 & @xmath620 & @xmath367 & 2 & @xmath480 & @xmath433 & @xmath621 & @xmath622 + 3 & @xmath242 & @xmath437 & @xmath623 & @xmath375 & 3 & @xmath588 & @xmath378 & @xmath624 & @xmath471 & 3 & @xmath555 & @xmath446 & @xmath625 & @xmath383 & 3 & @xmath464 & @xmath553 & @xmath626 & @xmath627 + @xmath362 & & & & & @xmath362 & & & & & @xmath362 & & & & & @xmath362 & & & & + @xmath229 & & & & & @xmath230 & & & & & @xmath231 & & & & & @xmath232 & & & & + 0 & @xmath291 & @xmath550 & @xmath628 & @xmath338 & 0 & @xmath357 & @xmath542 & @xmath629 & @xmath326 & 0 & @xmath546 & @xmath268 & @xmath630 & @xmath266 & 0 & @xmath545 & @xmath272 & @xmath631 & @xmath334 + 1 & @xmath530 & @xmath243 & @xmath632 & @xmath354 & 1 & @xmath549 & @xmath280 & @xmath633 & @xmath342 & 1 & @xmath539 & @xmath284 & @xmath634 & @xmath282 & 1 & @xmath287 & @xmath288 & @xmath635 & @xmath350 + 2 & @xmath260 & @xmath292 & @xmath306 & @xmath307 & 2 & @xmath536 & @xmath529 & @xmath499 & @xmath358 & 2 & @xmath534 & @xmath236 & @xmath636 & @xmath298 & 2 & @xmath239 & @xmath240 & @xmath637 & @xmath238 + 3 & @xmath244 & @xmath276 & @xmath322 & @xmath323 & 3 & @xmath638 & @xmath248 & @xmath310 & @xmath311 & 3 & @xmath527 & @xmath252 & @xmath639 & @xmath314 & 3 & @xmath525 & @xmath256 & @xmath640 & @xmath254 + @xmath426 & & & & & @xmath426 & & & & & @xmath426 & & & & & @xmath426 & & & & + @xmath229 & & & & & @xmath230 & & & & & @xmath231 & & & & & @xmath232 & & & & + 0 & @xmath523 & @xmath289 & @xmath641 & @xmath277 & 0 & @xmath296 & @xmath521 & @xmath359 & @xmath327 & 0 & @xmath520 & @xmath333 & @xmath330 & @xmath269 & 0 & @xmath351 & @xmath509 & @xmath642 & @xmath643 + 1 & @xmath515 & @xmath341 & @xmath644 & @xmath293 & 1 & @xmath343 & @xmath645 & @xmath646 & @xmath647 & 1 & @xmath511 & @xmath349 & @xmath648 & @xmath285 & 1 & @xmath518 & @xmath305 & @xmath649 & @xmath650 + 2 & @xmath507 & @xmath309 & @xmath245 & @xmath246 & 2 & @xmath264 & @xmath651 & @xmath361 & @xmath652 & 2 & @xmath503 & @xmath301 & @xmath653 & @xmath237 & 2 & @xmath501 & @xmath321 & @xmath654 & @xmath655 + 3 & @xmath498 & @xmath325 & @xmath261 & @xmath262 & 3 & @xmath513 & @xmath656 & @xmath657 & @xmath658 & 3 & @xmath495 & @xmath317 & @xmath315 & @xmath253 & 3 & @xmath493 & @xmath659 & @xmath660 & @xmath661 + !
0.9ptp0.32cm|p0.32cm|p0.32cm|p0.3cm!width 1.ptc!width 0.9ptp0.32cm|p0.32cm|p0.32cm|p0.32cm!width 1.ptc!width 0.9ptp0.32cm|p0.32cm|p0.32cm|p0.32cm!width 1.ptc!width 0.9ptp0.32cm|p0.32cm|p0.32cm|p0.32cm!width
1.pt @xmath228 & & & & & @xmath228 & & & & & @xmath228 & & & & & @xmath228 & & & & + @xmath229 & & & & & @xmath230 & & & & & @xmath231 & & & & & @xmath232 & & & & + 0 & @xmath361 & @xmath486 & @xmath662 & @xmath422 & 0 & @xmath570 & @xmath650 & @xmath663 & @xmath483 & 0 & @xmath481 & @xmath417 & @xmath647 & @xmath479 & 0 & @xmath664 & @xmath413 & @xmath665 & @xmath475 + 1 & @xmath409 & @xmath470 & @xmath666 & @xmath406 & 1 & @xmath627 & @xmath643 & @xmath667 & @xmath668 & 1 & @xmath465 & @xmath402 & @xmath669 & @xmath463 & 1 & @xmath670 & @xmath398 & @xmath671 & @xmath459 + 2 & @xmath378 & @xmath394 & @xmath672 & @xmath487 & 2 & @xmath437 & @xmath661 & @xmath673 & @xmath436 & 2 & @xmath553 & @xmath450 & @xmath674 & @xmath675 & 2 & @xmath477 & @xmath555 & @xmath676 & @xmath444 + 3 & @xmath345 & @xmath567 & @xmath677 & @xmath392 & 3 & @xmath405 & @xmath655 & @xmath452 & @xmath451 & 3 & @xmath433 & @xmath434 & @xmath678 & @xmath679 & 3 & @xmath430 & @xmath562 & @xmath680 & @xmath428 + @xmath297 & & & & & @xmath297 & & & & & @xmath297 & & & & & @xmath297 & & & & + @xmath229 & & & & & @xmath230 & & & & & @xmath231 & & & & & @xmath232 & & & & + 0 & @xmath485 & @xmath425 & @xmath681 & @xmath484 & 0 & @xmath390 & @xmath478 & @xmath682 & @xmath683 & 0 & @xmath684 & @xmath490 & @xmath685 & @xmath414 & 0 & @xmath686 & @xmath412 & @xmath687 & @xmath688 + 1 & @xmath572 & @xmath689 & @xmath690 & @xmath468 & 1 & @xmath577 & @xmath462 & @xmath691 & @xmath692 & 1 & @xmath693 & @xmath474 & @xmath694 & @xmath399 & 1 & @xmath695 & @xmath397 & @xmath696 & @xmath289 + 2 & @xmath454 & @xmath421 & @xmath697 & @xmath456 & 2 & @xmath698 & @xmath699 & @xmath700 & @xmath419 & 2 & @xmath386 & @xmath574 & @xmath701 & @xmath384 & 2 & @xmath382 & @xmath702 & @xmath703 & @xmath410 + 3 & @xmath438 & @xmath559 & @xmath704 & @xmath440 & 3 & @xmath705 & @xmath589 & @xmath706 & @xmath576 & 3 & @xmath370 & @xmath581 & @xmath707 & @xmath368 & 3 & @xmath366 & @xmath708 & @xmath709 & @xmath395 + @xmath362 & & & & & @xmath362 & & & & & @xmath362 & & & & & @xmath362 & & & & + @xmath229 & & & & & @xmath230 & & & & & @xmath231 & & & & & @xmath232 & & & & + 0 & @xmath710 & @xmath711 & @xmath712 & @xmath651 & 0 & @xmath341 & @xmath515 & @xmath713 & @xmath355 & 0 & @xmath321 & @xmath518 & @xmath714 & @xmath715 & 0 & @xmath349 & @xmath511 & @xmath716 & @xmath346 + 1 & @xmath594 & @xmath296 & @xmath717 & @xmath657 & 1 & @xmath718 & @xmath523 & @xmath719 & @xmath339 & 1 & @xmath509 & @xmath720 & @xmath721 & @xmath722 & 1 & @xmath333 & @xmath520 & @xmath723 & @xmath330 + 2 & @xmath327 & @xmath724 & @xmath656 & @xmath659 & 2 & @xmath325 & @xmath498 & @xmath249 & @xmath261 & 2 & @xmath725 & @xmath493 & @xmath726 & @xmath318 & 2 & @xmath317 & @xmath495 & @xmath727 & @xmath728 + 3 & @xmath729 & @xmath264 & @xmath343 & @xmath730 & 3 & @xmath309 & @xmath507 & @xmath233 & @xmath245 & 3 & @xmath305 & @xmath501 & @xmath731 & @xmath318 & 3 & @xmath301 & @xmath503 & @xmath732 & @xmath733 + @xmath426 & & & & & @xmath426 & & & & & @xmath426 & & & & & @xmath426 & & & & + @xmath229 & & & & & @xmath230 & & & & & @xmath231 & & & & & @xmath232 & & & & + 0 & @xmath280 & @xmath549 & @xmath734 & @xmath294 & 0 & @xmath292 & @xmath735 & @xmath260 & @xmath736 & 0 & @xmath288 & @xmath287 & @xmath350 & @xmath239 & 0 & @xmath284 & @xmath534 & @xmath737 & @xmath401 + 1 & @xmath542 & @xmath357 & @xmath738 & @xmath278 & 1 & @xmath276 & @xmath739 & @xmath740 & @xmath741 & 1 & @xmath272 & @xmath516 & @xmath742 & @xmath270 & 1 & @xmath268 & @xmath527 & @xmath743 & @xmath519 + 2 & @xmath248 & @xmath563 & @xmath326 & @xmath744 & 2 & @xmath745 & @xmath746 & @xmath323 & @xmath290 & 2 & @xmath256 & @xmath525 & @xmath747 & @xmath748 & 2 & @xmath252 & @xmath749 & @xmath750 & @xmath281 + 3 & @xmath529 & @xmath536 & @xmath622 & @xmath751 & 3 & @xmath752 & @xmath753 & @xmath307 & @xmath274 & 3 & @xmath240 & @xmath239 & @xmath238 & @xmath491 & 3 & @xmath236 & @xmath754 & @xmath755 & @xmath494 + the clustering map used at the relay node r can not remove the singular fade spaces @xmath756\right\rangle $ ] where at least one of @xmath757 for some @xmath21 and simultaneously satisfy the mutually exclusive law .
let @xmath758\right\rangle^ { \bot } $ ] be a singular fade state where for some @xmath759 , @xmath760 .
then , in order to remove @xmath150 , @xmath761 and @xmath762 that must be kept in the same cluster .
this would imply user @xmath20 not being able to distinguish between the messages @xmath763 and @xmath764 for some @xmath765 sent by user @xmath23 .
this will clearly violate the mutually exclusive law , since in order to satisfy the mutually exclusive law , for the same value of @xmath766 , all possible @xmath0-tuples of messages must be kept in different clusters .
these two statements can not be satisfied at the same time , hence such a singular fade subspace can not be removed if the mutually exclusive law has to be satisfied by the relay map used in the bc phase .
we refer to the singular fade subspaces whose harmful effects can not be removed by a proper choice of the clustering , as the _ non - removable singular fade subspaces _ also talked about in @xcite .
there are @xmath767 m^{n-1}$ ] removable and @xmath768 m^{k-1}$ ] non - removable singular fade subspaces for @xmath0-way relaying when @xmath5-psk constellation is used at the end nodes .
the number of non - removable singular subspaces is @xmath769 while the number of removable singular fade subspaces is @xmath770 .
thus , the number of non - removable singular fade subspaces is a small fraction of the total number of singular fade subspaces .
for the five - way relaying scenario described in the previous section , there are @xmath207 singular fade subspaces for five - way relaying , out of which the scheme given in this paper removes @xmath771 singular fade subspaces using @xmath772-fold latin hyper - cubes of side @xmath773 .
this can be done , as described above , by first marking the singularity removal constraints in the empty @xmath774 array and then completing the array to form a @xmath772-fold latin hyper - cubes of side @xmath773 using algorithm 1 .
we now illustrate the removal of a singular fade state for five - way relaying with the help of the following example .
consider the case for five - way relaying scenario where 4-psk is used at end nodes a , b , c , d and e , and one of the singular fade subspace to be removed is given by , @xmath775\right\rangle ^{\bot}\hspace{-0.3 cm } ~=~\hspace{-0.1cm}\left\langle \left [ { \begin{array}{c } 1-j \\ 2 \\ 2 \\ 1+j \\ -1+j \\ \end{array } } \right]\right\rangle ^{\bot}\hspace{-0.3cm}\ ] ] @xmath776\right\rangle ^{\bot}\hspace{-0.3 cm } ~=~\hspace{-0.1cm}\left\langle \left [ { \begin{array}{c } -1+j \\ -2 \\ -2 \\ -1-j \\ 1-j \\ \end{array } } \right]\right\rangle ^{\bot}.\ ] ] the first vector is @xmath777 $ ] .
now , @xmath778 can be obtained either as a difference of @xmath779 and @xmath780 or as a difference of @xmath781 and @xmath782 ; @xmath783 can be obtained only as a difference of @xmath784 and @xmath785 ; @xmath786 can be obtained as a difference of @xmath230 and @xmath787 or as a difference of @xmath788 and @xmath789 ; @xmath790 can be obtained as a difference of @xmath791 and @xmath792 or as a difference of @xmath793 and @xmath794 .
thus , the entries corresponding to @xmath795 , @xmath796 , @xmath797 , @xmath798 , @xmath799 , @xmath800 , @xmath801 , @xmath802 must lie in the same clustering representing the network coding map used at the relay node in the bc phase .
similarly the constraints resulting from the other three vectors above can be obtained . replacing @xmath803 with @xmath804
, we get constraints in the form of entries that must be kept the same in the @xmath774 array representing the clustering . for instance , corresponding to @xmath795 , the entries @xmath805 are kept in the same cluster @xmath496 as shown in bold in tables i , ii , iii , iv .
the constrained hyper - cube of side 4 is completed using algorithm 1 to form a 4-fold latin hyper - cube of side 4 .
the completed latin hyper cube is as shown in tables i , ii , iii and iv , where the constraints are marked in bold .
similarly , a ( removable ) singular fade subspace can be removed by first constraining the array representing the relay map , and then completing the constrained array using the provided algorithm .
simulation results presented in this section identify the cases where the proposed scheme outperforms the naive approach that uses the same map for all fade states and vice verse . here , the channel states are distributed according to rician distribution and channel variances equal to 0 db and the frame length is 256 bits .
[ fig : comparison_plot ] compares the snr vs bit - error - rate curves for three - way , four - way and five - way relaying scenario for ( a ) the adaptive network coding scheme presented in this paper with ( b ) the non - adaptive network coding using two channel uses , in which the same array is used by the relay as an encoder for all channel conditions .
the details of three - way and four - way relaying can be found in @xcite and @xcite .
the non - adaptive network coding for three - way , four - way and five - way relaying utilizes the same @xmath806-fold latin hyper - cubes of side @xmath773 , @xmath773-fold latin hyper - cube of side @xmath773 and @xmath772-fold latin hyper - cube of side @xmath773 ( respectively ) for all channel conditions . for any communication system ,
the effects of additive noise are predominant at low snr and the effects of noise due to fading dominate at high snr . since adaptive network coding scheme attempts at reducing the effects of fading for the @xmath0-way relaying scenario by removing a part of singular fade states , ( a ) performs better than ( b ) at high snr .
this also implies that the performance of the scheme in the ma phase ( removal of singular fade spaces takes place in the ma phase ) is predominant at higher snrs , and the performance of the scheme in the bc phase dominates at lower snrs . also , as the number of user nodes @xmath0 increases , the snr at which the performance of adaptive network coding improves over the performance of non - adaptive network coding increases , as can be seen in the plot , since the size of the received constellation in the ma phase and hence the size of the constellation used in the bc phase increases with increasing values of @xmath0 .
we consider the @xmath0-way wireless relaying scenario , where @xmath0 nodes operate in half - duplex mode and transmit points from the same @xmath5-psk constellation .
information exchange is made possible using just two channels uses , unlike the existing work done for the case , to the best of our knowledge .
the relay node clusters the @xmath15 possible transmitted tuples @xmath209 into various clusters depending on the fade states such that _ the exclusive law _ is satisfied and some of the singular fade subspaces are removed .
this necessary requirement of satisfying the exclusive law is shown to be the same as the clustering being represented by a @xmath0-fold latin hyper - cube of side @xmath5 .
the size of the clustering utilizing modified clustering may not be the best that can be achieved , and it might be possible to fill the array with lesser symbols
. the first author would like to thank her colleague vijayvaradharaj t muralidharan at the department of electrical and communication engineering , indian institute of science for his feedback and suggestions related to this paper .
t. koike - akino , p. popovski and v. tarokh , `` optimized constellation for two - way wireless relaying with physical network coding '' , ieee journal on selected areas in comm . ,
vol.27 , pp .
773787 , june 2009 .
srishti shukla , vijayvaradharaj t. muralidharan and b. sundar rajan , `` wireless network - coded three - way relaying using latin cubes '' , 23rd ieee personal indoor mobile radio communications ( pimrc)-2012 , sydney , australia .
mylene pischella and didier le ruyet , `` lattice based coding scheme for mimo bi - directional relaying with three nodes '' , 22nd ieee personal indoor mobile radio communications , toronto , canada , pp .
14591463 , sept .
2011 .
youngil jeon , young - tae kim , moonseo park , inkyu lee , `` opportunistic scheduling for three - way relay systems with physical layer network coding '' , ieee 73rd vehicular technology conference ( vtc - spring ) 2011 , budapest , hungary , pp .
15 , may 2011 .
vijayvaradharaj t muralidharan , vishnu namboodiri and b. sundar rajan , `` channel quantization for physical layer network - coded two - way relaying '' , available online at arxiv:1109.6101v2 [ cs.it ] , 16 nov . | physical layer network - coding for the @xmath0-way wireless relaying scenario is dealt with , where each of the @xmath0 user nodes @xmath1 @xmath2 wishes to communicate its messages to all the other @xmath3 nodes with the help of the relay node r. the given scheme , based on the denoise - and - forward scheme proposed for two - way relaying by popovski et al . in @xcite , employs two phases : multiple access ( ma ) phase and broadcast ( bc ) phase with each phase utilizing one channel use and hence totally two channel uses .
physical layer network - coding using the denoise - and - forward scheme was done for the two - way relaying scenario in@xcite , for three - way relaying scenario in @xcite , and for four - way relaying scenario in @xcite .
this paper employs denoise - and - forward scheme for physical layer network coding of the @xmath0-way relaying scenario illustrating with the help of the case @xmath4 not dealt with so far .
it is observed that adaptively changing the network coding map used at the relay according to the channel conditions reduces the impact of multiple access interference which occurs at the relay during the ma phase .
these network coding maps are chosen so that they satisfy a requirement called _ exclusive law_. we show that when the @xmath0 users transmit points from the same @xmath5-psk @xmath6 constellation , every such network coding map that satisfies the exclusive law can be represented by a @xmath0-fold latin hyper - cube of side @xmath5 .
the singular fade subspaces resulting from the scheme are described and enumerated for general values of @xmath0 and @xmath5 and are classified based on their removability in the given scenario . a network code map to be used by the relay for the bc phase aiming at reducing the effect of interference at the ma stage
is obtained . |
our first task is to make the allusive definitions of the introduction precise . for @xmath33 a metric space and @xmath34 a non - expanding ( _ i.e. _
, @xmath35-lipschitz ) embedding of @xmath36 into @xmath4 , the distortion @xmath37 of @xmath34 is defined to be the supremum of the compression ratio between the distance of two points in @xmath33 and that of their images in @xmath4 : @xmath38 by construction , @xmath37 is at least @xmath35 , and the larger it is , the bigger the deformation of distances caused by @xmath34 .
let us turn to @xmath39 , _
i.e. _ , let us consider images of triangles .
in the denominator of , the length of the segment @xmath40 $ ] is replaced with the area of the triangle @xmath41 $ ] . as for the numerator , the area makes no sense in the source space @xmath33 , but we observe that , at least in good cases , @xmath42 is the sup of the lengths @xmath43 for @xmath44 a non - expanding embedding of @xmath36 to @xmath4 ( provided @xmath45 ) .
this naturally leads to defining @xmath46 to be the sup of @xmath47)$ ] for @xmath44 a non - expanding embedding of @xmath36 to @xmath4 ( provided @xmath28 ) , and to defining the @xmath2-distortion of @xmath34 to be @xmath48 ) } ; p , q , r \in v\big\}.\ ] ] we shall be interested in the minimal possible value of @xmath49 , _ i.e. _ , in the configurations that minimalize the distortion of triangles .
we are thus led to the following notion : the _ @xmath2-distortion _ @xmath50 is defined to be the infimum of @xmath49 over all non - expanding embeddings @xmath34 of @xmath36 into @xmath4 .
the definition for @xmath6-tuples would be similar , with volume replacing area .
( 110,27)(0 , 0 ) ( 0,5)-distortion of a non - expanding embedding @xmath51 : @xmath52 generic case : the area is @xmath53 , whence @xmath54 , @xmath55 an optimal case : @xmath56 ; @xmath57 a worst case : the isometrical embedding ; then the image of @xmath34 is a flat triangle of area @xmath58 , hence @xmath59.,title="fig : " ] ( 14,0)@xmath52 ( 53,0)@xmath55 ( 96,0)@xmath57 ( -6,5)@xmath60 ( 20,5)@xmath61 ( 31,25)@xmath62 ( 9,5)@xmath63 ( 26,12)@xmath64 ( 18,12)@xmath65 ( 55,7)@xmath35 ( 66,15)@xmath35 figure [ 3dis ] describes the situation for the graph @xmath66 . in this
( very simple ) case , there exist embeddings with @xmath2-distortion equal to @xmath35 , namely the ones of figure [ 3dis]@xmath55 , and , therefore , we find @xmath67 . in the general case
, we always have @xmath68 by construction , and , the flatter the triangles in the image of @xmath34 , the larger @xmath49 .
for instance , when @xmath34 is an isometrical embedding of @xmath69 in @xmath4 , all triangles are flat , as in figure [ 3dis]@xmath57 , and the distortion @xmath49 is infinite . thus the @xmath2-distortion is a measure of the inevitable flattening of triangles that occurs when a ( large ) metric space is embedded in some fixed euclidean space : then , it is impossible that all triples of vertices are embedded so as to form a rectangular triangle as in figure [ 3dis]@xmath55 , and the question is to evaluate how far from that one must lie . the reader can check that , even in the case of embeddings of @xmath70 into @xmath1 , it is not so easy to prove that the minimal @xmath2-distortion is @xmath71 , corresponding to a u - shape with length @xmath35 edges and @xmath72 angles , and obtaining an exact value in the general case of @xmath69 seems out of reach .
this contributes to making asymptotic bounds desirable . in the specific case of the space @xmath13 ,
_ i.e. _ , of @xmath0 equidistant points at distance @xmath35 on the real line , the definition of @xmath2-distortion can be given a more simple form .
indeed , if @xmath44 is a non - expanding embedding of @xmath69 into @xmath4 , we have @xmath73 and therefore , for @xmath74 , we find @xmath75 ) \le ( j - i)(k - j)/2 $ ] ; on the other hand , provided @xmath28 , we can always find @xmath44 such that the latter inequality is an equality as in figure [ 3dis]@xmath55 .
hence , for @xmath76 , we have @xmath77 so , for @xmath34 a non - expanding embedding of @xmath69 into @xmath4 , takes the form @xmath78 ) } ; 0 \le
i < j < k \le n\big\}.\ ] ] in the sequel , we shall forget about embeddings and only work inside the target space @xmath4 .
a finite sequence of points @xmath79 in @xmath4 is said to be _ tame _ if , for each @xmath80 , we have @xmath81 . in this case
, we put @xmath82 ) } ; 0 \le i < j < k \le n\big\}.\ ] ] if @xmath34 is an embedding of @xmath69 into @xmath4 , then the sequence @xmath83 is tame , and , conversely , each tame sequence determines a unique embedding .
now , translating gives for @xmath84 and the notation is consistent .
then the @xmath2-distortion of @xmath13 can be expressed in terms of tame sequences of points : for all @xmath85 , we have @xmath86 thus , from now on , our aim is to study the possible values of the quantity @xmath25 of .
in order to prove proposition [ borneinf ] , we shall consider an arbitrary tame sequence in @xmath1 , and prove that some triangle is much distorted , _
i.e. _ , flattened . to this end
we observe that points in convex position provide a triangle with large @xmath2-distortion .
say that a sequence @xmath87 of points in the plane is _ convex _ if the boundary of the convex hull of @xmath88 , , @xmath89 is exactly the polygon with vertices @xmath90 in this order .
[ lconvexe ] assume that @xmath87 is a convex sequence with @xmath91
. then there exists @xmath80 such that the @xmath2-distortion of the triangle @xmath92where indices are taken modulo @xmath93is at least @xmath94 .
the sum of angles @xmath95 is @xmath96 . as all angles are positive and less than @xmath97 , one of them is at least @xmath98 .
the @xmath2-distortion of the corresponding triangle is then at least @xmath94 .
( 40,35)(0 , 0 ) ( 0,0 ) ( -5,18)@xmath99 ( -3,26)@xmath100 ( 7,32)@xmath101 ( 38,31)@xmath102 ( -10,7)@xmath103 [ distdroite ] ( 100 , 50)(0 , 0 ) ( 0,0 ) to the line @xmath104,title="fig : " ] ( 13,12)@xmath105 ( 30,22)@xmath106 ( 83,24)@xmath107 ( 79,34)@xmath108 [ lsoussuite ] assume that @xmath79 is a tame sequence in @xmath1 , and that @xmath109 is an integer greater than or equal to @xmath110 . then the sequence @xmath111 is convex
. let @xmath112 . for all @xmath113
, we have @xmath114 .
since for @xmath115 the area of the triangle @xmath116 $ ] is at least @xmath117 , hence _ a fortiori _ @xmath118 , the distance between the point @xmath107 and the line @xmath119 is at least @xmath120 ( figure [ distdroite ] ) .
therefore , for @xmath121 , the points @xmath107 and @xmath122 lie on the same side of the line @xmath119 : otherwise , the distance between @xmath107 and @xmath122 would be at least @xmath123 , contrary to the tameness hypothesis . hence , for @xmath124 , the point @xmath107 lies on the same side of the line @xmath119 as @xmath125 . for a contradiction , assume that , for some @xmath80 , the sequence @xmath126 , @xmath127 , @xmath128 , @xmath129 is not convex . then either the four points are not in convex position , or they are in convex position but they do not appear in the right order on the border of their convex hull . in the first case ( figure [ cas1 ] ) , one point lies in the convex hull of the three others .
but this contradicts the hypothesis that adjacent points lie on the same side of each line @xmath130 . in the second case ( figure [ cas2 ] ) ,
the points are in convex position , but the segment @xmath131 , @xmath132 $ ] crosses the line @xmath133 .
then there exists @xmath134 with @xmath135 such that the distance from @xmath106 to @xmath133 is at most @xmath136 .
the area of the triangle @xmath137 $ ] is therefore at most @xmath138 . on the other side , by definition of @xmath139 , this area is at least @xmath140 , hence _ a fortiori _ @xmath141 . since @xmath142 ,
the latter quantity is at least @xmath109 , a contradiction .
( 80 , 34)(0 , 0 ) ( 0,0 ) and @xmath143,title="fig : " ] ( 15,10)@xmath144 ( 24,18)@xmath145 ( 59,10)@xmath146 ( 40,31)@xmath143 ( 65,22)@xmath147 ( 110 , 24)(0 , 0 ) ( 0,0 ) and @xmath146,title="fig : " ] ( 16,7)@xmath144 ( 29,16)@xmath145 ( 59,-2)@xmath146 ( 77,11)@xmath143 ( 33,-4)@xmath147 let @xmath79 be a tame sequence in @xmath1 , and let @xmath109 be @xmath148 . if we have @xmath149 , then we have @xmath150 , hence @xmath151 _ a fortiori_. assume now @xmath152 . then by lemma [ lsoussuite ] , the sequence @xmath153 is convex , and by lemma [ lconvexe ] there is a triangle whose distortion is least @xmath154 . by definition ,
this quantity is at most @xmath109 , hence we have @xmath155 .
so in any case , @xmath29 lies in @xmath3 .
the proof of lemma [ lsoussuite ] gives many constraints for the sequence @xmath156 @xmath157 .
here we use these constraints to construct a convex subsequence of size @xmath158 , but it is likely that larger subsequences with properties slightly weaker than convexity could be constructed as well .
so we think that the result of proposition [ borneinf ] is not optimal .
now we turn to dimension @xmath24 and we wish to establish the lower bound result stated as proposition [ bornesup ] . our aim is to construct for each @xmath0 a tame sequence of length @xmath0 in @xmath4 with a small @xmath2-distortion , _ i.e. _ , such that all extracted triangles are not too much flattened . a natural idea would be to construct the @xmath0th sequence @xmath159 by taking more and more points on a single curve @xmath160 of length @xmath35 , and rescaling .
but then a small @xmath2-distortion would require a complicated curve @xmath160 .
indeed , assume that @xmath160 is an immersion of class @xmath161 .
as @xmath160 is compact , the infimum @xmath162 of the radii of the osculating circles of @xmath160 is reached at some point , and therefore it is non - zero . for any @xmath0 ,
there exists @xmath80 such that the curvilinear distance between @xmath163 and @xmath164 is lower than @xmath165 before rescaling .
then the distances between @xmath163 and @xmath166 , and between @xmath166 and @xmath164 are lower than @xmath165 too .
therefore the sine of the angle between the lines @xmath167 and @xmath168 is at most @xmath169 , and the distortion of the triangle @xmath170 is at least @xmath171 .
this leads to a @xmath2-distortion in @xmath31 for @xmath172 .
so , in order to construct sequences of points with small @xmath2-distortion , we have either to use curves depending on @xmath0 , or to use a non-@xmath161 curve ( typically a fractal curve ) . in the following construction
we choose the first option . for simplicity
, we assume @xmath173 for some @xmath93 .
we recursively construct a family of curves @xmath174 in @xmath4 , and , on each of them , we mark @xmath175 points @xmath176 in such a way that @xmath177 lies in @xmath178 for each fixed @xmath24 .
when @xmath179 points lie at mutual distance 1 on an arc of circle , the @xmath2-distortion is in @xmath180 .
the idea of our construction is to use this fact and to recursively put circles one above the others .
let @xmath181 be the sixth of a circle whose radius @xmath182 will be chosen later .
on @xmath181 we put points @xmath183 with regular angular distance @xmath184 .
then we replace the arc between @xmath185 and @xmath186 with a coplanar arc of radius @xmath187 lying between the original arc and the chord connecting @xmath185 to @xmath186 .
we rescale the figure so that the curvilinear coordinate of @xmath185 becomes @xmath80 for each @xmath80 .
we let @xmath188 be the resulting curve ( oriented from @xmath99 to @xmath189 ) and @xmath190 be the marked points on @xmath188 .
( 130 , 100)(0 , 0 ) ( 0,0 ) and the points @xmath191 .
on the left : three points @xmath192 with at least one @xmath193 between them yield an angle @xmath194,title="fig : " ] ( 23.5,15)@xmath195 ( 17,29.5)@xmath196 ( 18,59)@xmath197 ( 2,43)@xmath193 ( 69,20)@xmath198 ( 69,31)@xmath199 ( 69,67)@xmath200 the main remark for the proof is that , for all triples @xmath201 taken in increasing order on @xmath202 ( not necessarily some @xmath193 s ) and not all lying on some arc @xmath203 , we have @xmath204 . by construction ,
the euclidean distance between two points of @xmath188 is at least @xmath205 times their curvilinear distance , and therefore the 3-distortion of the triangle @xmath206 is in @xmath178 .
the idea for the induction is to add a copy of @xmath202 between @xmath207 and @xmath208 , orthogonally to the hyperplane in which @xmath209 lies .
more precisely , we construct @xmath174 and @xmath210 from @xmath211 and @xmath212 so that the following induction hypothesis is preserved : the construction of @xmath174 is as follows .
we identify @xmath4 with @xmath223 , where @xmath224 is the space containing @xmath211 .
next we work in the cylinder @xmath225 defined by @xmath226 with the induced metric .
note that this cylinder @xmath227 is orthogonal to the hyperplane containing @xmath211 .
for each @xmath80 between @xmath58 and @xmath228 , we insert in @xmath225 a rescaled copy of @xmath202 from @xmath229 to @xmath208 . in this way
, we obtain a curve on which @xmath175 points are marked : the @xmath207 s from @xmath211 plus @xmath230 new points between @xmath207 and @xmath208 for @xmath231 . we denote them by @xmath210 according to the linear ordering .
we then rescale the figure so that the curvilinear distance between consecutive points @xmath232 s is 1 .
we call @xmath213 the resulting curve . for @xmath52
, we observe that the angle between any chord of @xmath236 and the hyperplane containing @xmath209 is lower than @xmath237 . therefore , when going from @xmath209 to @xmath236 , no distance is decreased by more than a factor @xmath238 .
first case : there exists @xmath134 such that @xmath201 lie between @xmath239 and @xmath240 .
this means that @xmath201 lie on some copy of @xmath202 in @xmath241 inserted in the last step of the inductive construction . in the case of @xmath202
, we know that the @xmath2-distortion is at most @xmath242 . here there is an additional @xmath2-distortion due to the fact that the copy was made on the cylinder @xmath241 .
the projection of @xmath174 on @xmath224 is @xmath209 , and not a line as in the @xmath222 case . by induction hypothesis ,
the distances on @xmath209 ( compared with the euclidean distances ) are not contracted by more than @xmath243 , hence the distortion of the triangle @xmath219 $ ] is bounded by @xmath244 .
second case : there exists @xmath134 such that @xmath195 lies before @xmath245 and @xmath197 lies after @xmath246 .
then , when @xmath201 are projected from @xmath213 on @xmath209 along @xmath241 , the area of the triangle @xmath219 $ ] decreases by a multiplicative factor at most @xmath247 . by the induction hypothesis the projection of the triangle has 3-distortion at most @xmath248 ,
therefore the original triangle @xmath219 $ ] has 3-distortion at most @xmath220 .
@xmath52 the choice of the curve @xmath202 may look strange , in particular the choice of an arc of radius @xmath187 between @xmath249 and @xmath250 rather than an arc of radius @xmath182 or a chord .
the reason is that , in both cases , the key property , namely that the triangle @xmath219 $ ] has @xmath2-distortion @xmath178 if @xmath201 do not all lie on some arc @xmath251 , fails . with arcs of radius @xmath182
, if we take @xmath201 close to some @xmath252 , then the 3-distortion of @xmath219 $ ] can be arbitrary large . with chords ,
if we take @xmath253 strictly between @xmath254 and @xmath255 and @xmath197 just after @xmath255 , then the 3-distortion is not bounded either .
@xmath55 our construction uses @xmath256 pairwise orthogonal directions to draw the curves @xmath213 one above the other .
we could use other fixed directions as well , the point being that the projections preserve the convexity of the specific patterns we consider .
alternatively we could replace cylinders by cones , as central projection also preserves the needed convexity .
but it seems difficult to use more than one cylinder , and therefore more than one curve , for each new dimension , because no projection preserves the needed convexity for several sufficiently distinct directions simultaneously . | we prove that , for embeddings of a path of length @xmath0 in @xmath1 , the @xmath2-distortion is an @xmath3 , and that , when embedded in @xmath4 , the @xmath2-distortion is an @xmath5 .
the general context of this paper is the study of the distortion that appears when a metric space is embedded into a euclidean space .
such a study plays an important role in algorithmic geometry and its applications .
in particular , significant memory gains can be achieved when a metric space is embedded into a low dimensional euclidean space , and , therefore , the study of such embeddings is directly connected with the construction of efficient computer representations of ( finite ) metric spaces .
the price to pay for such memory gains is the inevitable deformations that result from the embedding , and it is therefore quite important to control them , typically to understand their asymptotic behaviour when the size of the metric space increases . a standard parameter for controlling the deformation is the distortion , that takes into account pairs of points and compares their distances in the source and the target spaces see precise definition below .
the distortion is rather well understood , and , in particular , precise bounds for its values in the case of general finite metric spaces are known @xcite .
now , other parameters may be associated with an embedding naturally . typically , for each @xmath6
, one can introduce the notion of a _
@xmath6-distortion _ by taking into account @xmath6-tuples of points rather than just pairs , and measuring the way the volume of the associated polytope is changed .
this is what u.feige does in @xcite in order to construct an algorithm minimizing the bandwidth of a graph , _
i.e. _ , finding a numbering @xmath7 of the vertices for which the supremum of @xmath8 over all pairs @xmath9 such that @xmath10 is an edge is as small as possible .
the idea of @xcite is to consider volume - respecting embeddings of the graph into a euclidean space .
the point is to show that , among all projections of such an embedding on a line , a positive proportion has a minimal bandwidth of the expected size , and the main step is to investigate the @xmath6-distorsion .
owing to the above applications and connections , understanding @xmath6-distortion for every @xmath6 seems to be a quite natural goal .
now , in contrast to the case @xmath11 , very little is known so far about @xmath6-distortion for @xmath12 .
the aim of this paper is to establish some results about @xmath2-distortion , in the most simple case of a metric space consisting of equidistant points on a line .
so , we denote by @xmath13 the set @xmath14 equipped with the distance @xmath15
. then , for each @xmath16 , there exists a real parameter @xmath17 that measures the deformation of triangles when @xmath13 is embedded in @xmath4 .
the intuition is that , the bigger @xmath18 , the flatter the triangles the precise definition is given in section 1 below .
as @xmath4 isometrically embeds in @xmath19 , the inequality @xmath20 immediately follows from the precise definition , implying in particular @xmath21 for @xmath22 .
the meaning is that , when we have more space , we can more easily embed with small distortion . for @xmath23 ( the planar case ) , hence for every @xmath24 , it is easy to see that @xmath25 is at most linear in @xmath0 , so the question is to compare @xmath25 with the polynomial functions @xmath26 , @xmath27 .
what we do below is to prove one lower bound result for @xmath23 , and one upper bound result for @xmath28 : [ borneinf ] the @xmath2-distortion @xmath29 is an @xmath3 .
[ bornesup ] for each fixed @xmath24 , the @xmath2-distortion @xmath25 is an @xmath30 .
the results are likely not to be optimal : we conjecture that @xmath29 might be an @xmath31 , and that @xmath25 might be lower than polynomial , typically polylogarithmic , for @xmath32 .
this would mean that the behaviour of the @xmath2-distortion radically differs from the standard distortion which is polynomial in @xmath0 for each dimension @xmath24 . |
finitely many real polynomials @xmath9={\mathbb r}[x_1,\ldots , x_n]$ ] define a basic closed semi - algebraic set @xmath10 one is interested in finding characterizations of @xmath11 , the set of all polynomials that are nonnegative on @xmath7 .
obvious candidates for such nonnegative polynomials are the elements of @xmath12 ^ 2\right\},\ ] ] the so called _ preordering _ generated by @xmath13 .
many works in real algebra and real algebraic geometry deal with the question how @xmath2 relates to @xmath11 . in dimension one , equality occurs often , at least if the @xmath14 are the so called _ natural generators _ for @xmath7 ( see @xcite ) . in dimension two , equality is a much more uncommon phenomenon .
for example , not every globally nonnegative polynomial in two variables is a sum of squares of polynomials .
however , scheiderer has given two local global principles that yield equality between @xmath2 and @xmath11 under certain conditions , in the case that @xmath7 is compact and two - dimensional ( see @xcite and also @xcite ) .
the results require a good behavior of the curves bounding @xmath7 .
noncompact two - dimensional examples where equality holds are even more rare .
one of the few known examples is the preordering generated by @xmath15 in @xmath16 $ ] , see @xcite . from dimension three upwards , equality between @xmath2 and @xmath11
is never possible .
this is one of the main results from @xcite .
of course one can ask questions beside equality . for example , schmdgen s famous result from @xcite says that in case @xmath7 is compact , @xmath2 contains every polynomial which is _ strictly _ positive on @xmath7 , no matter what generators @xmath13 of @xmath7 are chosen , and independent of the dimension of @xmath7
. however , this result fails in general if @xmath7 is not compact . if @xmath7 is very big , for example if it contains a full dimensional cone , then the preordering is indeed far from containing every nonnegative or positive polynomial ( see for example @xcite ) .
another question arising in this context concerns the @xmath7-_moment problem_. one wants to find a characterization of the linear functionals @xmath17\rightarrow{\mathbb r}$ ] that are integration on @xmath7 .
haviland s theorem ( @xcite , see also @xcite ) provides a necessary and sufficient condition .
namely , @xmath18 is integration on @xmath7 if and only if @xmath18 is @xmath19 on @xmath11 . as a characterization of @xmath11 is not very simple either , one wants to weaken the condition on @xmath18 and still apply haviland s theorem .
more precisely , one wants to know whether @xmath20 on @xmath2 is sufficient for @xmath18 to be integration . in view of haviland s theorem , that means to ask whether @xmath21 holds , where @xmath8 denotes the _
double dual cone _ of the preordering .
one says that the preordering has the _ strong moment property _ ( smp ) in this case .
for example , if @xmath7 is compact , then by the above mentioned schmdgen theorem , the preordering always has ( smp ) .
there are also many noncompact examples .
@xcite gives a criterion to decide this question , involving _ fibre - preorderings _ constructed from bounded polynomials .
now in @xcite , the following preordering is considered : @xmath22\mid \exists q\in{\mathbb r}[\underline{x}]\ \forall { \varepsilon}>0\ f+{\varepsilon}q\in\operatorname{po}(f_1,\ldots , f_s)\right\}.\ ] ] this definition does not use linear functionals and is therefore better accessible to algebraic methods .
@xmath23 turns out to be the _ sequential closure _ of @xmath2 with respect to the finest locally convex topology on @xmath24 $ ] , whereas @xmath8 is the _
closure_. the following relations hold : @xmath25 in @xcite , theorem 5.1 , it is shown that every element from @xmath11 belongs to the sequential closure of the preordering in a certain _ localization _ of the polynomial ring . a slightly weaker version of that is @xcite , theorem 6.2.3 .
the authors of @xcite prove @xmath26 ( and therefore ( smp ) ) for certain classes of preorderings in the polynomial ring .
the question whether in polynomial rings @xmath27 always holds remained open in these works ( see open problem 3 in @xcite and open problem 2 in @xcite ) .
we solve this problem by providing a counterexample ; @xmath23 does not equal @xmath8 in general . then it is natural to ask for conditions under which equality holds .
it is also interesting to ask whether a fibre theorem in the spirit of @xcite can be established for @xmath23 instead of @xmath8 .
this question already turned up in @xcite .
our counterexample implies that such a general theorem will require stricter assumptions than schmdgen s fibre theorem .
theorem 5.3 from @xcite is such a fibre theorem for finitely generated preorderings that describe cylinders with compact cross section .
we will generalize this result to quadratic modules in arbitrary commutative algebras .
we will then deduce criterions for @xmath27 and @xmath28 to hold .
they allow applications that go beyond the known examples from @xcite .
* acknowledgements * the author thanks murray marshall and claus scheiderer for interesting and helpful discussions on the topic of this work .
for the following results on topological vector spaces we refer to @xcite , chapter ii .
let @xmath29 be an @xmath30-vector space .
finest locally convex topology _ on @xmath29 is the vector space topology having the collection of all convex , absorbing and symmetric subsets of @xmath29 as a neighborhood base of zero .
all the linear functionals on @xmath29 are continuous , @xmath29 is hausdorff , and every finite dimensional subspace of @xmath29 inherits the canonical topology . by @xcite , chapter 2 ,
exercise 7(b ) , a sequence in @xmath29 converges if and only if it lies in a finite dimensional subspace of @xmath29 and converges there .
so for the _ sequential closure _ of a set @xmath31 in @xmath29 , defined as the set of all limits of sequences from @xmath31 , and denoted by @xmath32 , we have @xmath33 where the union runs over all finite dimensional subspaces @xmath34 of @xmath29 .
from now on , we will restrict ourself to _ convex cones _ in @xmath29 , i.e. to subsets @xmath31 that are closed under addition and multiplication with positive reals .
the following alternative characterization for @xmath32 can be found in @xcite : @xmath35 for preorderings @xmath31 in @xmath30-algebras , the object @xmath32 was first introduced in @xcite in terms of this last characterization .
it only turned out later that it is the sequential closure of @xmath31 .
the _ closure _ of a convex cone @xmath31 is denoted by @xmath36 @xcite , chapter ii.39 , corollary 5 implies @xmath37 for convex cones .
here , @xmath38 denotes the double dual cone of @xmath31 , i.e. the set @xmath39 we obviously have @xmath40 if @xmath29 has countable vector space dimension , then a set @xmath31 in @xmath29 is closed if and only if its intersection with every finite dimensional subspace of @xmath29 is closed , by @xcite , proposition 1 .
so @xmath31 is closed if and only if it is sequentially closed , i.e. if @xmath41 holds .
this implies that the ( transfinite ) sequence of _ iterated sequential closures _ of @xmath31 terminates at @xmath42 in that case .
note that in case @xmath29 is not countable dimensional , then this sequence may terminate before it reaches the closure .
it indeed always terminates at the closure with respect to the _ topology of finitely open sets _ , which can be smaller in the case of uncountable dimension .
we refer to @xcite for more information .
now let @xmath43 be a commutative @xmath30-algebra with @xmath44 .
of course @xmath43 is an @xmath30-vector space , and we equip it with the finest locally convex topology .
quadratic module _ is a set @xmath45 with @xmath46 and @xmath47 here , @xmath48 denotes the set of squares in @xmath43 .
a _ preordering _ is a quadratic module with the additional property @xmath49 quadratic modules ( and preorderings ) are convex cones , and we look at @xmath32 and @xmath50 defined as above . for a set @xmath51
, the smallest quadratic module / preordering containing @xmath52 is called the quadratic module / preordering _ generated _ by @xmath52 .
we write @xmath53 and @xmath54 for it , respectively .
@xmath53 consists of all finite sums of elements @xmath55 and @xmath56 where @xmath57 and @xmath55 is a sum of squares in @xmath43 .
@xmath54 consists of all finite sums of elements @xmath58 , where @xmath55 is a sum of squares , @xmath59 and all @xmath60 . of special interest
is the case that @xmath52 is finite .
we call @xmath53 and @xmath54 _ finitely generated _ then . an important notion , introduced in @xcite ,
is that of _ stability _ of a finitely generated quadratic module .
if @xmath31 is generated by @xmath61 , then we call @xmath31 _ stable _ , if for every finite dimensional @xmath30-subspace @xmath62 of @xmath43 there is another finite dimensional @xmath30-subspace @xmath63 of @xmath43 , such that @xmath64 a map that assigns to each @xmath62 such a @xmath63 is then called a _ stability map_. whereas the notion of stability is independent of the choice of generators of @xmath31 , the stability map is not .
we refer the reader to @xcite for proofs and details . to @xmath43
there corresponds the set of all real characters on @xmath43 , i.e. @xmath65 elements @xmath66 from @xmath43 define functions @xmath67 on @xmath68 by @xmath69 we equip @xmath68 with the coarsest topology making all these functions continuous . as the functions @xmath67 separate points , this makes @xmath68 a hausdorff space , and we have the algebra homomorphism @xmath70 if @xmath43 is finitely generated as an @xmath30-algebra , then @xmath68 embeds into some @xmath1 , by taking generators @xmath71 and sending @xmath72 to @xmath73 .
so @xmath68 is the zero set in @xmath1 of the kernel of the algebra homomorphism @xmath74\rightarrow a ; x_i\mapsto x_i$ ] .
the use of an element @xmath66 from @xmath43 as a function @xmath67 coincides with the usual use of @xmath66 as a polynomial function on that embedded variety . in particular , the topology on @xmath68 is inherited from the canonical one on @xmath1 .
note also that @xmath43 is a countable dimensional @xmath30-vector space in case it is finitely generated as an @xmath30-algebra .
now we consider arbitrary commutative @xmath30-algebras @xmath43 with @xmath44 again .
for a set @xmath75 , without loss of generality a quadratic module , we define @xmath76 if @xmath31 is finitely generated , then @xmath77 is called a _ basic closed semi - algebraic set_. for any set
@xmath78 we define @xmath79 this is a preordering . starting with a quadratic module or a preordering @xmath31 in @xmath43 , we have the following chain : @xmath80 the last inclusion comes from the fact that characters on @xmath43 are in particular linear functionals . as for example proven in @xcite , @xmath32 and @xmath38 are again quadratic modules , even preorderings if @xmath31 was a preordering .
following @xcite , we make the following definitions .
[ smpdd ] * @xmath31 has the _ strong moment property _ ( smp ) , if @xmath81 holds * @xmath31 has the _ @xmath82-property _ , if @xmath83 holds * @xmath31 is _ saturated _ , if @xmath84 holds * @xmath31 is _ closed _ , if @xmath85 holds the interest in @xmath86 and ( smp ) is motivated by a classical theorem by haviland . for certain classes of algebras ,
it states that a linear functional on @xmath43 is integration with respect to some measure on @xmath77 , if and only if it is nonnegative on @xmath87 ( @xcite in the case that @xmath43 is a polynomial algebra , see @xcite for a more general version ) .
so if @xmath31 has ( smp ) , then every functional that is nonnegative on @xmath31 is integration on @xmath77 .
nonnegativity on @xmath31 is a priori a much weaker condition than nonnegativity on @xmath88 this is one of the reasons that make ( smp ) so interesting .
a method to decide whether ( smp ) holds for a finitely generated preordering @xmath89 in @xmath24 $ ] is given in @xcite , as mentioned in the introduction .
let @xmath90 be a polynomial that is bounded on @xmath91 .
then @xmath92 holds , where @xmath93 denotes the ideal generated by @xmath94 .
this implies that @xmath89 has ( smp ) if and only if all the preorderings @xmath95 have ( smp ) .
as these so called _ fibre preorderings _ usually describe lower dimensional semi - algebraic sets , they are easier to deal with in general . the result in particular implies that every finitely generated preordering in @xmath24 $ ] describing a compact set @xmath91 has ( smp ) .
this was already part of the earlier paper @xcite .
for an alternative proof of the result from @xcite see also @xcite .
the @xmath82-property was introduced and studied in @xcite for polynomial algebras .
the authors for example show that in case @xmath91 is a cylinder with compact cross section , then the preordering @xmath89 has the @xmath82-property , under reasonable assumptions on the generators of @xmath89 .
this is also shown , using different methods , in @xcite .
it was an open problem in @xcite , whether ( smp ) and the @xmath82-property are equivalent or even @xmath96 is always true .
we start by showing that the answer to both questions is negative .
the example in this section will answer open problem 3 in @xcite and open problem 2 in @xcite to the negative .
it will also give a negative answer to the question in @xcite , whether the fibre theorem holds for the @xmath82-property instead of ( smp ) .
consider @xmath97 $ ] , the polynomial algebra in two variables , and @xmath98 these polynomials define the following basic closed semi - algebraic set @xmath99 in @xmath100 : * * [ smp ] the preordering @xmath101 in @xmath16 $ ] has ( smp ) the polynomial @xmath102 is bounded on @xmath99 , and we can apply schmdgen s fibre theorem from @xcite to @xmath89 . for any @xmath103 $ ] , the preordering @xmath104 describes a compact semi - algebraic set and therefore has ( smp ) by @xcite ( even the ( @xmath105)-property ) . for @xmath106 $ ]
, the preordering @xmath107 $ ] is saturated .
this follows from @xcite , theorem 2.2 .
so @xmath108 in @xmath16 $ ] is saturated as well . in particular
, @xmath109 has ( smp ) .
so by @xcite , the whole preordering @xmath89 has ( smp ) .
the next result is a characterization of @xmath110 .
we write @xmath111 for the set of elements having a representation in @xmath112 with sums of squares of elements of degree @xmath113 .
[ dd]a polynomial @xmath114 $ ] belongs to @xmath115 if and only if there is some @xmath116 such that for all @xmath117 $ ] , @xmath118 belongs to @xmath119 in @xmath120 $ ] .
the `` if''-part is a consequence of theorem @xmath121 below ( or can already be obtained by looking at the proof of theorem 5.3 . in @xcite ) .
for the `` only if''-part assume @xmath4 belongs to @xmath122 .
so there is some @xmath123 $ ] and sums of squares @xmath124 $ ] for all @xmath125 and @xmath126 such that @xmath127 note that the total degree of the @xmath128 may rise with @xmath129 getting smaller .
however , the degree as polynomials in @xmath130 can not rise ; it is bounded by the @xmath130-degree of @xmath131 , which does not change with @xmath129 .
this is because the set @xmath99 contains the cylinder @xmath132\times [ 1,\infty]$ ] , so whenever one adds two polynomials which are nonnegative on it , the leading terms as polynomials in @xmath130 can not cancel .
so the degree in @xmath130 of the sum is the maximum of the @xmath130-degrees of the terms . by evaluating in @xmath133 ,
this means that @xmath134 belongs to @xmath135 for some fixed @xmath136 and all @xmath117 , { \varepsilon}>0 $ ] .
but by @xcite , proposition 2.6 , this is a closed set in a finite dimensional subspace of @xmath120 $ ] .
so we get @xmath137 for all @xmath117 $ ] , the desired result .
the preordering @xmath138 does not have the @xmath105-property .
the polynomial @xmath130 is obviously nonnegative on @xmath99 .
however , it does not belong to the preordering @xmath139.\ ] ] indeed , writing down a representation and evaluating in @xmath140 , this shows that @xmath141 divides @xmath130 , a contradiction .
so in view of proposition @xmath142 , @xmath130 can not belong to @xmath110 .
note that @xmath130 is not in @xmath110 as it fails to be in the preordering corresponding to the fibre @xmath143 .
however , proposition @xmath142 even demands all the polynomials @xmath118 to have representations in the fibre - preorderings @xmath144 with _ simultaneous _ degree bounds , for @xmath4 to be in @xmath110
. indeed , there are examples of polynomials belonging to all of the fibre - preorderings , but failing the degree - bound condition ( and so also not belonging to @xmath110 ) .
we will give one here , as it gives a justification for one of the assumptions in theorems [ main ] and [ ddapp ] below .
[ exam ] take @xmath145 , which belongs to @xmath146 . for any @xmath117 $ ]
, @xmath147 belongs to @xmath148 ; for @xmath149 as @xmath118 is strictly positive on the corresponding compact semi - algebraic set ( so use @xcite ) , for @xmath150 $ ] , the fibre preordering is saturated , as mentioned in the proof of proposition @xmath151 .
however , for @xmath152 , there can be no bound on the degree of the sums of squares in the representation . indeed , for @xmath149 ,
write down a representation @xmath153 where the @xmath154 are sums of squares .
evaluating in @xmath140 , this shows @xmath155 now if the degrees of the @xmath154 could be bounded for all @xmath149 , we could write down a first order logic formula saying that we have representations as in ( @xmath156 ) for all @xmath149 .
we add the statement ( @xmath157 ) to the formula . by tarski s transfer principle
, it holds in any real closed extension field of @xmath30 .
so take such a representation in some non - archimedean real closed extension field @xmath158 for some @xmath149 which is infinitesimal with respect to @xmath30 .
the same argument as for example in @xcite , example 4.4 .
( a ) shows that we can apply the residue map @xmath159 to the coefficients of all the polynomials occurring in this representation . here
, @xmath160 denotes the convex hull of @xmath30 in @xmath158 .
this is a valuation ring with maximal ideal @xmath161 .
so we get a representation @xmath162 with sums of squares @xmath163 in @xmath120 $ ] fulfilling @xmath164 as no cancellation of highest degree terms can occur , we get @xmath165 as well as @xmath166 this last fact obviously contradicts ( @xmath167 ) .
so for @xmath168 , the degree bound condition on the fibres fails , although the polynomial belongs to all of the fibre preorderings .
in view of proposition @xmath142 , it does not belong to @xmath122 .
this shows that the `` degree bound''-assumption in theorems @xmath169 and [ ddapp ] below is really necessary .
note also that the example is an explicit illustration of theorem 8.2.6 from @xcite , where the general impossibility of a certain degree bound for schmdgen s theorem from @xcite is proved .
the above example answers the question in @xcite , whether the fibre theorem holds for the @xmath105-property instead of ( smp ) .
indeed , we have shown in the proof of proposition @xmath151 that all the fibre preorderings @xmath170 do not only have ( smp ) , but even the @xmath105-property .
as @xmath89 itself does not have the @xmath105-property , this gives a negative answer to the question .
however , we will prove a result below that allows to use a dimension reduction when examining the @xmath105-property under certain conditions .
we want to prove a fibre theorem in the spirit of @xcite , to be able to examine the sequential closure of a quadratic module in terms of ( easier ) fibre - modules .
we consider the following general setup .
let @xmath171 be commutative @xmath30-algebras with @xmath44 , let @xmath102 be a compact hausdorff space , and assume we have algebra homomorphisms @xmath172 and @xmath173 ( homomorphisms are always assumed to map @xmath44 to @xmath44 ) : @xmath174 \ar@{->}^{{\varphi}}[u ] & c(x,{\mathbb r})}\ ] ] we assume that the image of @xmath175 in @xmath176 separates points of @xmath102 , i.e. for any two distinct points @xmath177 there is some @xmath178 such that @xmath179 the stone - weierstrass theorem implies that @xmath180 is dense in @xmath176 with respect to the sup - norm then .
for @xmath181 let @xmath182 be the vanishing ideal of @xmath183 in @xmath175 , and @xmath184 the ideal in @xmath43 generated by @xmath185 .
the homomorphism @xmath186 makes @xmath43 a @xmath175-module in the usual sense of commutative algebra ( not to be confused with quadratic modules ! ) . for a @xmath175-submodule @xmath34 of @xmath43
we write @xmath187 so @xmath188 is a @xmath175-submodule of @xmath34 .
we have @xmath189 .
now let @xmath75 be a quadratic module .
for any @xmath181 , @xmath190 is again a quadratic module , called the _ fibre - module _ of @xmath31 with respect to @xmath183 .
our first goal is to prove the following abstract fibre theorem : [ main ] let @xmath171 be commutative @xmath30-algebras , @xmath102 a compact hausdorff space , @xmath191 and @xmath192 algebra homomorphisms , such that @xmath193 separates points of @xmath102 .
let @xmath75 be a quadratic module and assume @xmath194 whenever @xmath195 on @xmath102 .
then for any finitely generated @xmath175-submodule @xmath34 of @xmath43 we have @xmath196 the requirement on @xmath34 can be understood as a `` degree bound condition '' as in proposition [ dd ] .
example [ exam ] shows that @xmath197 is not true under the remaining assumptions in general ( we will discuss this in more detail below ) . to prove the theorem
, we first need the following technical proposition .
[ loc ] let @xmath171 be @xmath30-algebras and @xmath172 an algebra homomorphism .
let @xmath102 be a compact hausdorff space and @xmath198 an algebra homomorphism whose image separates points of @xmath102 .
assume @xmath199 and @xmath125 are such that for all @xmath181 there is an identity @xmath200 with @xmath201 and @xmath202 for all @xmath203 .
then there are @xmath204 with @xmath205 on @xmath102 for all @xmath203 and @xmath206 every @xmath181 has an open neighborhood @xmath207 , such that @xmath208 on @xmath207 for all @xmath209 by compactness of @xmath102 there are @xmath210 , such that @xmath211 if @xmath212 , then the result follows , so assume @xmath213 choose a partition of unity @xmath214 subordinate to that cover , i.e. all @xmath215 are continuous functions from @xmath102 to @xmath216 $ ] , @xmath217 for all @xmath218 , and @xmath219 for all @xmath220 then for @xmath221 we have @xmath222 where @xmath223 denotes the sup - norm on @xmath176 . let @xmath224 and choose a positive real number @xmath225 , big enough to bound the sup - norm of all @xmath226 .
the image of @xmath175 in @xmath176 is dense , by the stone - weierstrass theorem .
so we find @xmath227 such that @xmath228 for @xmath229 , and we define @xmath230 so we have for @xmath231 @xmath232 we define @xmath233 for @xmath209 so @xmath234 now as @xmath235 we have @xmath236 which proves the proposition
. now we can give the proof of theorem [ main ] .
it is a generalization of the proof of theorem 5.3 from @xcite .
fix a finitely generated @xmath175-submodule @xmath34 of @xmath43 .
assume @xmath237 has a representation @xmath238 with @xmath239 and @xmath240 for all @xmath220 as @xmath241 , we can assume without loss of generality @xmath242 for all @xmath183 .
let @xmath243 be generators of @xmath34 as a @xmath175-module .
due to the identity @xmath244 we can assume that all @xmath245 are squares in @xmath43 ( by possibly enlarging @xmath34 ) .
we will now show @xmath246 for all @xmath247 therefore fix one such @xmath247 we take representations @xmath248 where all @xmath249 and @xmath250 .
now each @xmath181 has an open neighborhood @xmath207 , such that @xmath251 for @xmath209 by compactness of @xmath102 we have @xmath252 for some @xmath253 let @xmath214 be a continuous partition of unity subordinate to that cover . using the stone - weierstrass theorem , we approximate the square root of each @xmath215 ( which is again a continuous function ) by elements @xmath254 from @xmath175 , such that @xmath255 holds for all @xmath209 here
, @xmath256 denotes the sup - norm on @xmath176 again .
define @xmath257 the proof is complete if we show @xmath258 fix @xmath181 .
then @xmath259 where we define @xmath260 by the above considerations we have @xmath261 for all @xmath203 .
so we can apply proposition [ loc ] to @xmath262 and find @xmath263 for some @xmath264 with @xmath265 on @xmath102 .
thus @xmath266 as all @xmath267 are strictly positive on @xmath102 and all @xmath268 are squares . we demonstrate how to apply theorem [ main ] , for a given algebra @xmath43 and a quadratic module @xmath75 .
therefore assume there are @xmath269 with @xmath270 for real numbers @xmath271 ( @xmath272 . this of course implies that each @xmath273 is bounded as a function on @xmath274 , but the converse is not always true .
let @xmath275 $ ] be the subalgebra of @xmath43 generated by the @xmath273 and let @xmath172 be the canonical inclusion .
let @xmath276 be the quadratic module in @xmath175 generated by @xmath277 it is _ archimedean _
, for example by @xcite , theorem 4.1 .
let @xmath278 , so @xmath102 is compact , and the canonical homomorphism @xmath279 fulfills the separating points condition .
now whenever some @xmath280 is strictly positive on @xmath102 , then @xmath281 by @xcite , theorem 6 ( see also @xcite theorem 5.3.6 and @xcite , theorem 5.4.4 ) .
so @xmath194 .
for any @xmath181 , we have @xmath282 $ ] , and @xmath283 holds for any @xmath175-module @xmath34 .
thus write for @xmath284 $ ] @xmath285 and @xmath286 .
if @xmath31 is finitely generated as a quadratic module , then @xmath287 is also finitely generated , by the generators of @xmath31 and @xmath288 .
the following fibre theorem is our main result .
[ ddapp ] let @xmath43 be a commutative @xmath30-algebra and @xmath75 a quadratic module .
suppose @xmath269 are such that @xmath289 for some real numbers @xmath271 ( @xmath272 . then for every finitely generated @xmath290$]-submodule @xmath34 of @xmath43 we have @xmath291 where @xmath292.$ ] in particular , if @xmath31 is finitely generated and all the ( finitely generated ) quadratic modules @xmath293 are closed and stable with the same stability map , then @xmath294 if all @xmath293 are saturated and stable with the same stability map , then @xmath31 has the @xmath82-property .
( here , the stability map with respect to the canonical generators of each @xmath293 is meant . )
the first part of the theorem is clear from the above considerations and theorem [ main ] . for the second part ,
assume @xmath31 is finitely generated , say by @xmath13 .
then @xmath293 is finitely generated as a quadratic module , by the canonical generators @xmath295 assume all @xmath293 are closed ( or saturated , respectively ) and stable with the same stability map .
suppose some @xmath4 belongs to @xmath42 ( or @xmath87 , respectively ) .
then @xmath4 belongs to all @xmath296 ( or @xmath297 , respectively ) , so to all @xmath293 by our assumption .
now by the assumed stability there is a _ fixed _ finite dimensional @xmath30-subspace @xmath34 of @xmath43 , such that @xmath4 belongs to all @xmath298 .
so the first part of the theorem yields @xmath299 [ rem](1 ) in example [ exam ] , the polynomial @xmath300 belongs to all the preorderings @xmath301 in @xmath97 $ ] .
however , there is no finitely generated @xmath302$]-submodule @xmath34 of @xmath16 $ ] such that @xmath4 belongs to all @xmath303 .
this follows from what we have shown in example [ exam ] .
we have also seen that @xmath4 does not belong to @xmath304 note that @xmath305 in @xmath306 is fulfilled , as @xmath307 , and using an easy calculation as for example in @xcite , note 2.3 ( 4 ) .
so the `` degree bound condition '' is necessary in theorem [ ddapp ] and also in theorem [ main ] .
\(2 ) example [ einsdurch ] below will show that the assumption @xmath270 for all @xmath203 can not be omitted in theorem [ ddapp ] .
so the same is true for the assumption @xmath308 in theorem [ main ] .
\(3 ) in case that all the occurring quadratic fibre - modules @xmath287 in theorem [ ddapp ] are saturated and stable with the same stability map , we get a little bit more than the @xmath82-property for @xmath31 .
we obtain that for every finite dimensional subspace @xmath63 of @xmath43 there is some @xmath309 such that whenever @xmath310 , then @xmath311 for all @xmath247 in other words , the polynomial @xmath312 from the @xmath82-property does only depend on the subspace @xmath4 is taken from , not on the explicit choice of @xmath4 .
this follows from the proof of theorem [ main ] , noting that @xmath312 does only depend on the @xmath175-module @xmath34 , which depends only on @xmath63 and the stability map here .
in this section we give some applications of the fibre theorem .
the first one is the cylinder theorem ( theorem 5.3 combined with corollary 5.5 ) from @xcite .
see @xcite for the definition of natural generators for semi - algebraic subsets of @xmath30 . [ cylinder ]
let @xmath313 be a finitely generated preordering in the polynomial ring @xmath314 $ ] .
assume @xmath315 for some @xmath316 .
now for all @xmath317 , the preordering @xmath318\ ] ] describes a basic closed semi - algebraic set @xmath319 in @xmath30 .
suppose the natural generators for @xmath319 are among the @xmath320 , whenever @xmath319 is not empty .
then @xmath89 has the @xmath82-property .
if all the fibre sets @xmath319 are of the form @xmath321,$ ] @xmath322\cup[q,\infty)$ ] or @xmath323 $ ] , then the result holds with @xmath89 replaced by @xmath324 the assumptions imply that all the preorderings @xmath325 ( or the corresponding quadratic modules , respectively ) are saturated and stable with the same stability map for all @xmath326 .
see @xcite , section 4 .
an easy calculation , as for example in @xcite , note 2.3 ( 4 ) , shows @xmath327 for all @xmath203 .
so we can apply theorem [ ddapp ] .
we can also use theorem [ ddapp ] in the case that the natural generators are not among the @xmath328 .
this can be seen as a generalization of corollary 5.4 from @xcite : let @xmath329 be a finitely generated quadratic module in @xmath314 $ ] and assume @xmath330 for some @xmath331 suppose for all @xmath332 the set @xmath319 ( defined as in corollary [ cylinder ] ) is either empty or unbounded
. then @xmath333 holds .
again @xmath334 for all @xmath203 .
furthermore , the assumptions imply that all the quadratic modules @xmath335 are closed and stable with the same stability map for all @xmath326 ( for the empty fibers use theorem 4.5 from @xcite ) . now apply theorem [ ddapp ] .
we want to get results for more complicated fibres .
@xcite gives a criterion for quadratic modules on curves to be stable and closed .
however , we need some result to obtain the _ uniform _ stability asked for in theorem [ ddapp ] .
so we consider the following setup .
let @xmath336 $ ] be a polynomial of degree @xmath337 .
we assume that the highest degree homogeneous part of @xmath90 factors as @xmath338 where all the @xmath339 are pairwise disjoint points of @xmath340 in particular , @xmath90 is square free .
let @xmath341 denote the affine curve in @xmath342 defined by @xmath90 and @xmath343 its projective closure in @xmath344 .
so @xmath343 is defined by @xmath345 , the homogenization of @xmath90 with respect to the new variable @xmath346 .
the assumption on the highest degree part of @xmath90 implies that all the points at infinity of @xmath90 , namely @xmath347 are real regular points ( of the projective curve @xmath343 ) .
so the local rings of @xmath343 at all these points are discrete valuation rings ( a well known fact , see for example @xcite , chapter 3 ) .
indeed , the projective curve @xmath343 is the so called good completion " ( see for example @xcite ) of the affine curve @xmath341 .
we denote the valuation corresponding to the local ring at @xmath348 by @xmath349 . for a polynomial @xmath350,$ ]
we write @xmath351 and mean the value with respect to the valuation @xmath352 of @xmath353 as a rational function on @xmath354 .
we start with the following result : [ curve ] let @xmath355 and @xmath343 be as above .
suppose @xmath356 for some @xmath350 , l\in{\mathbb n}$ ] and all @xmath203 .
then there is some @xmath357 $ ] with @xmath358 and @xmath359 let @xmath360 be the degree of @xmath361 and @xmath362 as well as @xmath363 be the homogenization of @xmath90 and @xmath361 , respectively
. assume without loss of generality @xmath364 for some @xmath365 for any homogeneous polynomial @xmath366 in the variables @xmath367 we have @xmath368 where @xmath369 denotes the intersection number .
this is @xcite , chapter 3.3 . as @xmath370 we have @xmath371 now whenever @xmath372 , then @xmath373 so @xmath374 must vanish at @xmath375 .
the same argument applies to all points at infinity of @xmath90 .
so if @xmath372 , then the highest degree part of @xmath90 divides the highest degree part of @xmath361 in @xmath16 $ ] .
thus @xmath361 can be reduced modulo @xmath90 to a polynomial @xmath376 of strictly smaller degree . in the following proposition ,
the pure closedness and stability result follows from @xcite , proposition 6.5 .
[ simstab ] let @xmath377 $ ] be a finitely generated quadratic module .
let @xmath336 $ ] be a polynomial whose highest degree part factors as above .
for some @xmath378 assume that all the points at infinity of the curve @xmath379 defined by @xmath380 lie in the closure of @xmath381 .
then the finitely generated quadratic module @xmath382 is closed and stable , with a stability map that depends only on @xmath90 and @xmath13 , but not on @xmath326 . without loss of generality ,
let @xmath383 be a point at infinity of @xmath379 . denote by @xmath384 the valuation with respect to the local ring of @xmath385 at @xmath375 . let @xmath350 $ ] have degree @xmath360 , and let @xmath374 as well as @xmath386 be the homogenizations , as in the previous proof . then @xmath387 where the last equality uses property ( 7 ) in @xcite , p. 75
, for intersection numbers .
so there is some @xmath225 , not depending on @xmath326 , such that @xmath388 for all the points of infinity of @xmath379 . now
the proof of proposition 6.5 from @xcite shows that whenever @xmath389 , then we can find a representation @xmath390 with sums of squares @xmath391 built of polynomials that have order greater than @xmath392 in all points at infinity of @xmath379 .
applying proposition [ curve ] we can reduce these elements modulo @xmath94 and obtain a representation as in ( [ repr ] ) with sums of squares of elements of degree less or equal to @xmath393 .
so of course also the degree of @xmath366 is bounded suitably , independent of @xmath326 .
this shows that the stability map does not depend on @xmath326 .
so the following theorem is an immediate consequence of theorem [ ddapp ] and proposition [ simstab ] .
[ infcurve ] let @xmath394 $ ] be a finitely generated quadratic module .
let @xmath336 $ ] with @xmath395 for some @xmath396 , and assume the highest degree part of @xmath90 factors as above .
suppose that for all @xmath397 $ ] all the points at infinity of the curve @xmath379 defined by @xmath380 lie in the closure of @xmath381
. then @xmath333 holds .
if all the fibre modules @xmath398 have ( smp ) in addition , then @xmath31 has the @xmath82-property .
we give some explicit examples for these last results .
[ einsdurch ] we look at the semi - algebraic set in @xmath399 defined by the inequalities @xmath400 * * a lot of interesting phenomena can be observed for this set .
there are different quadratic modules describing it , we consider the following ones : @xmath401 the quadratic module @xmath402 is stable ; one checks that no cancellation of highest degree terms can occur in a sum @xmath403 so by @xcite , theorem 5.4 , @xmath402 does not have ( smp ) . to the quadratic module
@xmath404 we can apply theorem [ infcurve ] with the polynomial @xmath405 : we have @xmath406 . for @xmath407 $ ] , the finitely generated quadratic module @xmath408 is saturated .
this is an easy calculation for @xmath149 ; for @xmath409 it is example 3.26 from @xcite .
so @xmath404 has the @xmath82-property , and in particular ( smp ) .
note that the fibre modules of @xmath402 and @xmath404 are the same : @xmath410 for all @xmath407.$ ] as @xmath402 does not have the @xmath82-property , this shows that the condition @xmath411 in theorem [ infcurve ] , as well as the corresponding conditions in theorems [ ddapp ] and [ main ] can not be omitted .
now consider @xmath412 .
the quadratic module @xmath413,$ ] obtained by factoring out the ideal @xmath414 , does not have ( smp ) ( see for example @xcite ) .
so in view of proposition 4.8 from @xcite , @xmath412 does also not have ( smp ) . on the other hand
, we can still apply theorem [ infcurve ] with @xmath405 , and obtain @xmath415 last , the preordering @xmath89 obviously contains @xmath404 and therefore also has the @xmath82-property .
this solves the question posed in @xcite , example 8.4 .
we consider the semi - algebraic set defined by the inequalities @xmath416 * * we can apply theorem [ infcurve ] to the quadratic module @xmath417 where @xmath418 .
we use @xmath90 as the bounded polynomial and obtain @xmath419 however , @xmath31 does not have ( smp ) . indeed , the quadratic module @xmath420 does not have ( smp ) .
this follows from @xcite , theorem 3.17 together with @xcite , proposition 6.5 .
so in view of proposition 4.8 from @xcite , @xmath31 does not have ( smp ) .
we want to explain how the result of theorem @xmath121 , together with the observation from remark [ rem ] ( 3 ) , can be used for constrained polynomial optimization .
we take a similar approach to the one in @xcite , see also @xcite chapter 10 and @xcite for a nice account of the topic .
however , our approach does not only apply to compact semi - algebraic sets .
assume @xmath421 $ ] , so @xmath422 .
consider a finitely generated quadratic module @xmath423\ ] ] with associated semi - algebraic set @xmath424 , which has the property @xmath425 where @xmath312 can be chosen to only depend on the degree of @xmath4 ( in other words : for any nonnegative polynomial of the same degree , the same @xmath312 works ) .
without loss of generality @xmath312 can be chosen to be from @xmath31 ( see @xcite , the note following proposition 1.3 ) .
note that for example the quadratic modules from corollary [ cylinder ] and theorem [ infcurve ] have this stronger property .
note also that if a compact semi - algebraic set is described by a preordering , or more general , by an archimedean quadratic module @xmath31 , the above condition holds , as every strictly positive polynomial belongs to @xmath31 .
so @xmath426 works for every nonnegative polynomial .
given some @xmath427 $ ] , one wants to calculate the infimum of @xmath4 on @xmath7 , denoted by @xmath428 .
this is usually a hard problem , but for compact semi - algebraic sets @xmath52 , lasserre @xcite provided a sequence of semi - definite programs ( which are much easier to solve ) , whose optimal values converge to @xmath428 .
now take the polynomial @xmath312 from above corresponding to the degree of @xmath4 .
it can be obtained explicitly in the case of theorem [ main ] from the proof .
for example , in corollary [ cylinder ] , @xmath312 can be chosen to be the sum of the elements @xmath429 where @xmath430 runs from @xmath431 to the @xmath130-degree of @xmath4 . for fixed @xmath125 and @xmath116 , we consider @xmath432 and @xmath433 here
, @xmath434 denotes the set of all elements from @xmath31 that can be represented by @xmath13 and sums of squares of degree @xmath113 .
it is clear that the sequence is monotonically increasing .
now take some @xmath326 which is feasible for @xmath438 , which means that @xmath439 belongs to @xmath31 ( if no such @xmath326 exists , then the statement is also clear , as all values are @xmath440 ) .
but then @xmath439 belongs to @xmath441 for some big enough @xmath136 .
so @xmath442 , which proves the statement .
now suppose @xmath443 on @xmath7 for some @xmath378 .
then @xmath444 is nonnegative on @xmath7 and so @xmath445 holds for all @xmath125 .
this shows @xmath446 for all @xmath125 .
we have used here that subtracting @xmath326 from @xmath4 does not change the degree , and therefore we can use the polynomial @xmath312 , no matter how big or small @xmath326 is .
this could fail if @xmath31 only has the @xmath82-property , not the stronger version we assume .
from the fact that @xmath312 is in @xmath31 it is clear that the sequence is monotonically decreasing .
now suppose @xmath428 is finite and @xmath449 for some @xmath450 and all @xmath125 .
this means that @xmath451 is feasible for all @xmath438 and so @xmath452 holds for all @xmath125 .
this clearly implies @xmath453 on @xmath7 and so @xmath454 on @xmath7 , a contradiction . | let @xmath0 be a basic closed semi - algebraic set in @xmath1 and @xmath2 the corresponding preordering in @xmath3 $ ] .
we examine for which polynomials @xmath4 there exist identities @xmath5 these are precisely the elements of the sequential closure of @xmath2 with respect to the finest locally convex topology .
we solve the open problem from @xcite , whether this equals the double dual cone @xmath6 by providing a counterexample .
we then prove a theorem that allows to obtain identities for polynomials as above , by looking at a family of _ fibre - preorderings _ , constructed from bounded polynomials .
these fibre - preorderings are easier to deal with than the original preordering in general . for a large class of examples
we are thus able to show that either _ every _ polynomial @xmath4 that is nonnegative on @xmath7 admits such representations , or at least the polynomials from @xmath8 do .
the results also hold in the more general setup of arbitrary commutative algebras and quadratic modules instead of preorderings . |
the space of all @xmath0 hermitian matrices @xmath1 with probability density function @xmath6 is known as the _ gaussian unitary ensemble _ ( or gue)@xcite . here and henceforth the variance is chosen to ensure that asymptotically for @xmath7 , the limiting mean density of the gue eigenvalues is given by the wigner semicircle law @xmath8 supported in the interval @xmath9 $ ] .
the characteristic polynomial @xmath10 of the matrix @xmath1 constitutes one of the most basic quantities of interest , encoding all eigenvalues of @xmath1 through the roots of @xmath11 .
as one varies the argument @xmath12 over an interval containing many eigenvalues for a given realization of the ensemble , the value of the polynomial @xmath13 shows huge variations by the orders of magnitude for large @xmath14 , see figure [ fig : guerealizationn50 ] for @xmath15 and figure [ fig : n3000gueres ] for @xmath16 . for @xmath15 .
the global maximum is marked with a red circle . ]
the purpose of this article is to describe the statistical properties of the _ highest peak _ displayed by the modulus of the gue polynomial @xmath17 , namely the probability density for the maximum value attained by @xmath17 over the interval @xmath18 $ ] on the real line as @xmath3 .
our main result is the following [ pred ] consider the random variable @xmath19}\bigg\{2\log|p_{n}(x)|-2\mathbb{e}(\log|p_{n}(x)|)\bigg\ } \label{maxrv}\ ] ] then in the limit @xmath3 we have @xmath20 where @xmath21 is a continuous random variable characterized by the two - sided laplace transform of its probability density : @xmath22 where @xmath23 and @xmath24 stand for the euler gamma - function and the barnes digamma - function , correspondingly .
the normalization @xmath25 can be evaluated explicitly as @xmath26 where @xmath27 is the glaisher - kinkelin constant @xmath28 .
the product form of the laplace transform offers an interesting interpretation of the above results .
noting that @xmath29 is the moment generating function of a standard gumbel random variable @xmath30 , we can write @xmath31 where @xmath32 is an independent random variable with two - sided laplace transform @xmath33 in the end of the paper we provide convincing numerical evidence that this laplace transform does indeed define a unique random variable @xmath32 .
this immediately implies that the probability density of @xmath21 is the convolution of a gumbel random variable with @xmath32 .
such a convolution structure is expected to appear universally when studying the extreme value statistics of logarithmically correlated gaussian fields , see the discussion around and after eq .
( [ lcdef ] ) . in recent years
, much interest has accumulated regarding the statistical behaviour of characteristic polynomials of various random matrices as a function of the spectral variable @xmath12 . to a large extent
this interest was stimulated by the established paradigm that many statistical properties of the riemann zeta function along the critical line , that is @xmath34 , can be understood by comparison with analogous properties of the characteristic polynomials of random matrices @xcite .
for invariant ensembles @xcite of self - adjoint matrices with real eigenvalues , statistical characteristics of @xmath11 depend very essentially on the choice of scale spanned by the real variable @xmath12 . from that end
it is conventional to say that @xmath12 spans the _ local _ ( or _ microscopic _ ) scale if one considers intervals containing in the limit @xmath7 typically only a finite number of eigenvalues ( the corresponding scale for gue in ( [ guedensity ] ) is of the order of @xmath35 ) . at such scales , standard objects of interest
are correlation functions containing products and ratios of characteristic polynomials , which show determinantal / pfaffian structures @xcite for hermitian / real symmetric matrices and tend to universal limits at the local scale .
similar structures arise for properly defined characteristic polynomials @xmath36 of circular ensembles ( like cue , coe , and cse)@xcite of unitary random matrices @xmath37 uniformly distributed with respect to the haar measure on @xmath38 ( and other classical groups ) @xcite , whose properties on the local scale are indistinguishable from their hermitian counterparts . next , when @xmath12 spans an interval containing in the limit @xmath7 typically of order of @xmath14 eigenvalues one speaks of the _ global _ ( or _ macroscopic _ ) scale behaviour .
at such a scale properties of @xmath13 display both universal and non - universal features , the latter depending on the ensemble chosen .
the study of characteristic polynomials at such a scale was initiated in @xcite where it was shown that the function @xmath39 , with @xmath37 belonging to the cue , converges ( in an appropriate sense ) to a random gaussian fourier series of the form @xmath40 where the coefficients @xmath41 are independent standard complex gaussian random variables , i.e. @xmath42 , @xmath43 and @xmath44 .
the covariance structure associated with such a process is given by @xmath45 as long as @xmath46 .
such a ( generalized ) random function @xmath47 is a representative of random processes known in the literature under the name of _
1/f noises _ , see @xcite for background discussion and further references .
recently the study of the global scale behaviour was extended to the gue polynomial @xmath13 in @xcite by using earlier insights from @xcite and @xcite . that work revealed again a structure analogous to that of ( [ 1/f ] ) , though different in detail .
namely , it was shown that the natural limit of @xmath48 is given by the random chebyshev - fourier series @xmath49 with @xmath50 being chebyshev polynomials and real @xmath51 being independent standard gaussians .
a quick computation shows that the covariance structure associated with the generalized process @xmath52 is given by an integral operator with kernel @xmath53 as long as @xmath54 .
such a limiting process @xmath52 is an example of an aperiodic @xmath55-noise .
finally , one can consider an intermediate , or _
spectral scales , with intervals typically containing in the limit @xmath7 the number of eigenvalues growing with @xmath14 , but representing still a vanishingly small fraction of the total number @xmath14 of all eigenvalues .
the properties of the characteristic polynomials at such scales were again addressed in @xcite where it was shown that for the gue , that object gives rise to a particular ( singular ) instance of the so - called fractional brownian motion ( fbm ) @xcite with the hurst index @xmath56 , again characterized by correlations logarithmic in the spectral parameter .
the discussion above serves , in particular , the purpose of pointing to an intimate connection between gaussian random processes with logarithmic correlations and the modulus of characteristic polynomials at global and mesoscopic scales .
the relation is important as logarithmically correlated gaussian ( lcg ) random processes and fields attract growing attention in mathematical physics and probability and play an important role in problems of quantum gravity , turbulence , and financial mathematics , see e.g. @xcite . in particular , the periodic @xmath55 noise ( [ 1/f ] ) emerged in constructions of conformally invariant planar random curves @xcite . among other things ,
the statistics of the global maximum of lcg fields attracted considerable attention , see @xcite and references therein . particularly relevant in the present context
are the results of ding , roy and zeitouni @xcite on the maxima of regularized lattice versions of lcg fields which we discuss informally below .
let @xmath57 be the @xmath58dimensional box of side length @xmath14 with the left bottom corner located at the origin .
a suitably normalized version of the logarithmically correlated gaussian field is a collection of gaussian variables @xmath59 with variance @xmath60 and covariance structure @xmath61 where @xmath62 and both @xmath63 and @xmath64 are continuous bounded functions far enough from the boundary of @xmath65 .
now set @xmath66 and @xmath67 .
the limiting law of @xmath68 is then expected , after an appropriate shift and rescaling , to be given by the _
gumbel distribution with random shift _ : @xmath69 where the distribution of the random shift variable @xmath70 depends on details of the behaviour of covariance ( [ lcdef ] ) for @xmath71 and @xmath72 , see the detailed discussion in @xcite .
the random variable @xmath73 is related to the so called _ derivative martingale _ associated with the lcg fields @xcite whose distribution is however not known .
recently it has been shown that the recentering term @xmath74 in also holds for a randomized model of the riemann zeta function @xcite , proved by revealing a special branching structure within the associated logarithmic correlations .
we see that our conjecture [ pred ] for the maximum of characteristic polynomial of large gue matrices fully agrees with the predicted structure of the maximum of lcg in dimension @xmath75 .
note that the expression ( [ lcmax ] ) implies that the double - sided laplace transform of the density @xmath76 for the ( shifted ) maximum @xmath21 is related to the density @xmath77 of the random variable @xmath70 as @xmath78 which is in turn equivalent to the gumbel convolution in eq.([uprime ] ) .
in fact our formula ( [ uprimelaplace ] ) provides the explicit form of the distribution for the derivative martingale of our model , thus going considerably beyond the considerations of @xcite . from a quite different perspective , processes similar to ( [ 1/f ] ) and ( [ 1/fch ] ) appeared in the context of statistical mechanics of disordered systems when studying extreme values of random multifractal landscapes supporting spinglass - like thermodynamics @xcite .
the latter link is especially important in the context of the present paper .
the idea that it is beneficial to look at @xmath79 as a disordered landscape consisting of many peaks and dips , and to think of an associated statistical mechanics problem was put forward in @xcite .
it allowed to get quite non - trivial analytical insights into statistics of the maximal value of the cue polynomial sampled over the full circle @xmath80 $ ] , or over its mesoscopic sub - intervals .
this was further used to conjecture the associated properties of the modulus of the riemann zeta - function along the critical line , see some recent advances inpired by that line of research in @xcite .
some relations between between cue characteristic polynomials and logarithmically correlated processes ( in the form of the so - called `` multiplicative chaos '' measures introduced by kahane , see @xcite for a review ) was recently rigorously verified in @xcite .
the case of gue polynomials however remained outstanding .
it is our objective in this paper to provide two separate means of supporting conjecture [ pred ] .
first , we will provide careful and explicit , albeit in part heuristic , analytical arguments .
although our technique is inspired by the approach of @xcite it contains new nontrivial features necessary to overcome challenges arising from the non - uniform eigenvalue density @xmath81 , reflecting absence of translational invariance for the gue at the global spectral scale ( note e.g. the non - trivial recentering in ) .
all this makes actual calculation for the gue much more involved in comparison to the cue and the limiting random variable @xmath82 above appears to be more complicated than its cue counterpart .
secondly , we will test our conjecture with numerical experiments for matrices of size @xmath16 and around @xmath83 realizations .
this is especially important as part of our analysis is based on very plausible but as yet not fully rigorous considerations .
finally , is natural to expect that the same distribution should be shared by the maximum modulus of characteristic polynomials for hermitian random matrices with independent entries taken from the so - called wigner ensembles , see @xcite . before giving the detail of our procedure in the next section we need to quote the following fundamental asymptotic result obtained by krasovsky @xcite which will be central for our considerations : @xmath84 \label{krasov}\end{aligned}\ ] ] where @xmath85 and @xmath24 is the barnes g - function . differentiating with respect to @xmath86
, we deduce that @xmath87 where we used that @xmath88 .
the most salient feature of the asymptotics is the product of differences on the second line , which when rewritten in the form @xmath89 , \label{kracovstruct}\ ] ] can be looked at as evidence of the limiting gaussian process ( [ 1/fch ] ) with logarithmic covariance ( [ covop ] ) in the background .
we will however stress that naively replacing the ( shifted ) @xmath90 with the corresponding @xmath55 noise ( [ 1/fch ] ) is not a valid approximation as the factors @xmath91 in ( [ krasov ] ) do play an essential role in determining the extreme value statistics of @xmath92 .
let us finally note that had we suppressed the factors @xmath93 the faithful description of @xmath90 would be that of the regularized lcg process with covariance ( [ covop ] ) , the position - dependent variance @xmath94 and the position - dependent mean @xmath95 . *
acknowledgements : * we are grateful for helpful comments from christian webb during the preparation of this manuscript .
we acknowledge support from epsrc grant ep / j002763/1 `` '' .
following the ideas of @xcite we recast the problem of computing the value of the global maximum of @xmath17 ( with an appropriate shift by the mean value ) as a statistical mechanics problem characterized by the partition function @xmath96 with the `` potential '' @xmath97 , inverse temperature @xmath98 and @xmath99-independent non - negative parameter @xmath100 . specifically
, if we define the associated `` free energy '' as @xmath101 , then @xmath102.\ ] ] note that if compared to a similar partition function for the cue case the main new feature in is the factor @xmath103 .
although naively the presence of such a factor may seem irrelevant when taking the limit @xmath104 , we will actually see that it plays a very important role in supporting our procedure of extracting the free energy for @xmath99 exceeding some critical value .
now we aim to compute the integer moments of the partition function defined in : @xmath105 in the limit @xmath7 the leading asymptotics of the above integral can be extracted by replacing the factor @xmath106 with its asymptotics from . in this way one obtains @xmath107^{k}}\prod_{j=1}^{k}(1-x_{j}^{2})^{\frac{\beta^{2}+q}{2}}\prod_{1 \leq i < j \leq k}(2|x_{j}-x_{i}|)^{-2\beta^{2}}\,dx_{1}\ldots dx_{k}\end{aligned}\ ] ] after changing variables @xmath108 the integral above assumes the form @xmath109^{k}}\prod_{j=1}^{k}y_{j}^{\frac{\beta^{2}+q}{2}}(1-y_{j})^{\frac{\beta^{2}+q}{2}}\prod_{1 \leq i < j \leq k}|y_{i}-y_{j}|^{-2\beta^{2}}\,dy_{1}\ldots dy_{k}\end{aligned}\ ] ] @xmath110 with the quantity @xmath111 being the well - known selberg integral @xcite : @xmath112^{k}}\prod_{j=1}^{m}x_{j}^{a}(1-x_{j})^{b}\prod_{1 \leq i < j \leq k}|x_{i}-x_{j}|^{-2\gamma}\,dx_{1}\ldots dx_{k}\\ & = \prod_{j=1}^{k}\frac{\gamma(a+1-(j-1)\gamma)\gamma(b+1-(j-1)\gamma)\gamma(1-j\gamma)}{\gamma(a+b+2-(k+j-2)\gamma)\gamma(1-\gamma)}\\ & = \frac{1}{\gamma^{k}(1-\gamma)}\tilde{s}_{k}(a , b,-\gamma)\end{aligned}\ ] ] it is easy to see that the found expression for the partition function moments @xmath113 in ( [ momselb],[mompart ] ) is well - defined for any @xmath114 and for an _ integer _
@xmath115 satisfying @xmath116 . to understand how to deal with the case @xmath117 , we recall that krasovsky s asymptotic formula ( [ krasov ] ) is valid only when all of the differences @xmath118 remain _ finite _ when @xmath7 , and should be replaced by a different expression when @xmath119
. one can check that the divergence of the integral for @xmath120 is due precisely to the fact that these near degeneracies become important . relying on our experience with the corresponding situation for the cue @xcite case
suggests that taking into account the correct short - scale cutoff cures the formal divergence , but changes the asymptotics of the moments @xmath113 with @xmath14 : namely , these become of the order of @xmath121 for @xmath122 whereas they are of the order of @xmath123 for @xmath124 . such a change of behaviour will lead to a log - normal ( far ) tail in the distribution .
note that for the cue case , the above behaviour conjectured in @xcite was validated by recent rigorous calculation @xcite . there is no doubt that the same mechanism is operational in present case as well and will be validated by extending the theory of @xcite from toeplitz to hankel case
actually , as argued in @xcite the moments with @xmath122 play only a secondary role when addressing the question of extreme value statistics which is controlled exclusively by moments with @xmath125 .
our next goal is to use the latter integer moments for restoring the associated part of the probability density @xmath126 for the partition function .
this will be achieved if we manage to find the distribution for a random variable @xmath127 whose positive integer moments are given by @xmath128 such a task actually requires finding a way to continue analytically those moments to _ complex _ @xmath115 .
below we will arrive at the required continuation by exploiting a relatively simple heuristic procedure suggested in @xcite .
note that in a series of insightful papers @xcite ostrovsky developed a rigorous mathematical procedure of the required continuation which provides an _ aposteriori _ justification of the results obtained via the heuristic approach .
one starts with finding a recursion satisfied by @xmath129 for integer @xmath115 which is suitable for the continuation . by writing @xmath130 one sees immediately that @xmath131 it is convenient to introduce the moments @xmath132 of the random variable @xmath133 defined for any complex @xmath134 as @xmath135 .
we then have @xmath136 and after identifying @xmath137 the recursion ( [ rec1 ] ) takes the form @xmath138 which is now assumed to be valid for any complex @xmath134 . it is convenient to further use the duplication formula for the gamma function : @xmath139 to get rid of the argument @xmath140 in the denominator . indeed , we have @xmath141 so that ( [ rec2 ] ) assumes the form @xmath142 recalling that according to ( [ zreduced ] ) in our particular case @xmath143 we now use the parameterisation @xmath144 , @xmath145 and @xmath99-independent constants @xmath146 . after this
we finally arrive at @xmath147 to determine the function @xmath132 which satisfies ( [ rec4 ] ) for any complex @xmath134 we follow @xcite and introduce a variant of the barnes function @xmath148 which for any @xmath149 is defined by : @xmath150 where @xmath151 .
this function satisfies the so - called self - duality relation @xmath152 and further posesses a shift property that is central for our studies @xmath153 one can check that @xmath148 for @xmath154 coincides with the standard barnes function @xmath155 which is a unique solution of the recursion @xmath156 satisfying @xmath157 . similarly to the standard barnes function the general barnes @xmath148 has no poles and only zeroes located at @xmath158 , @xmath159 .
a detailed discussion of properties of functions closely related to @xmath148 can be found in @xcite .
let us now define a function @xmath160 of the complex argument @xmath134 by @xmath161 where @xmath162 and @xmath163 .
then a straightforward computation which relies on the identity following from ( [ gt1 ] ) @xmath164 shows that the ratio @xmath165 reproduces the right - hand side of ( [ rec4 ] ) from which we conclude @xmath166 which finally implies that @xmath167 where @xmath168 .
together with , and the fact that @xmath169 , we obtain for @xmath170 : @xmath171 the pair ( [ momg])-([compmom ] ) solves the problem of finding the complex moments @xmath172 of the random variable @xmath127 for any complex @xmath134 , and @xmath170 .
knowledge of such moments can be used to restore the probability distribution of @xmath127 , hence of the partition function @xmath173 , and of its logarithm ( the free energy ) for large @xmath174 .
our goal is however to study the limit of the latter as @xmath175 and one therefore should have a way of extracting information on the distribution for @xmath176 . in doing this
we rely on the _ freezing transition scenario _ for logarithmically correlated random landscapes .
the background idea of such scenario goes back to @xcite and was further advanced and clarified in the series of works @xcite . in brief , this scenario predicts a phase transition at the critical value @xmath154 and amounts to the following principle : _ thermodynamic quantities which for @xmath170 are * self - dual * functions of the inverse temperature @xmath99 , i.e. functions that remain invariant under the transformation @xmath177 , retain for all @xmath176 the value they acquired at the point of self - duality @xmath178 .
_ although such a scenario is not yet proven mathematically in full generality and has the status of a conjecture supported by physical arguments and available numerics , recently a few nontrivial aspects of freezing were verified within rigorous probabilistic analysis , see e.g. @xcite for efforts in this direction . within that scenario ,
one of the main outcomes of the analysis performed in @xcite is that the self - dual object associated with the distribution of the partition function for logarithmically correlated landscapes is expected to be the appropriately defined laplace transform : @xmath179\right),\end{aligned}\ ] ] where @xmath180 is a typical scale of the partition function which is extracted from the asymptotic for the integer moments and in our case can be chosen as @xmath181^{2}}{g(2\beta+1)\gamma(1-\beta^{2})}\left(\frac{4}{\pi}\right)^{q}.\ ] ] moreover , defining the probability density @xmath182 by @xmath183 one can show that the double - sided laplace transform for such a probability density is related to the complex moments @xmath184 of the scaled partition function via the following relation ( see eq.(26 ) of @xcite ) @xmath185 actually , as shown in @xcite the freezing scenario implies that the variable @xmath21 whose probability density is given by @xmath186 is precisely the fluctuating part of the height of the global minimum of the random potential which is our main object of interest .
note however that the scale @xmath187 diverges when approaching the critical point @xmath154 , and that the associated free energy @xmath188 is self - dual only in the leading order , given by @xmath189 .
the latter term after freezing at @xmath154 yields the leading @xmath190 term in our conjecture eq.([predeq ] ) for the maximum , whereas the logarithmically divergent term @xmath191 after careful re - interpretation results in the second term @xmath192 , see @xcite for the detailed explanation of that mechanism .
the procedure leaves however a certain arbitrariness in the terms of the order of unity in the mean free energy , hence in the overall shift of the position of the maximum .
let us stress however that apart from such a shift , the shape of the distribution function recovered in the framework of the freezing paradigm is completely fixed by the procedure . our strategy therefore will be to check if self - duality holds for the right - hand side combination in ( [ denmin ] ) when we substitute our expression for the moments . before we proceed , it will be helpful to further expand our expression .
inserting and making use of the identity @xmath193 shows that ( taking into account all prefactors coming from , and ) @xmath194^{1-s}2^{2\beta^{2}+b_{2}s}\pi^{-s}\beta^{\beta^{2}s}\frac{(2\pi)^{(\beta-1)/2}\beta^{-1/2-\beta^{2}(s-1)}}{m^{(g)}_{\beta}(1)}\\ & \times\gamma(1+\beta^{2}(s-1))\frac{g_{\beta}\left(\beta\left(s+\frac{a_{2}+b_{2}+1}{2}\right)+\frac{2+a_{1}+b_{1}}{2\beta}\right ) g_{\beta}\left(\beta\left(s+\frac{a_{2}+b_{2}}{2}\right)+\frac{2+a_{1}+b_{1}}{2\beta}\right ) } { g_{\beta}\left(\beta(s+a_{2})+\frac{1+a_{1}}{\beta}\right)g_{\beta}\left(\beta(s+b_{2})+\frac{1+b_{1}}{\beta}\right)}\\ \label{momgdual } & \times \frac{g_{\beta}\left(\beta\left(s+\frac{a_{2}+b_{2}+1}{2}\right)+\frac{3+a_{1}+b_{1}}{2\beta}\right ) g_{\beta}\left(\beta\left(s+\frac{a_{2}+b_{2}}{2}\right)+\frac{3+a_{1}+b_{1}}{2\beta}\right ) } { g_{\beta}\left(\beta(s+1+a_{2}+b_{2})+\frac{2+a_{1}+b_{1}}{\beta}\right)g_{\beta}\left(\beta s+\frac{1}{\beta}\right ) } \end{split}\ ] ] a direct inspection makes it clear that the self - duality is only possible if either @xmath195 or @xmath196 . for the gue characteristic polynomials
, we have @xmath197 , @xmath198 so that duality occurs _ only if _ @xmath199 .
we therefore have to choose @xmath199 to be able to rely upon the freezing scenario allowing to interpret the function @xmath186 calculated from its laplace transform via ( [ denmin ] ) as the probability density for the ( shifted ) global minimum . using ( [ momgdual ] ) with @xmath200 we get @xmath201^{1-s}2^{(1+\beta^{2})(s-1)}\\ & \times\gamma(1+\beta^{2}(s-1))\frac{g_{\beta}\left(\beta\left(s+1\right)+\frac{3}{2\beta}\right)}{g_{\beta}\left(\beta(s+1/2)+\frac{3}{2\beta}\right)}\frac{g_{\beta}\left(\beta\left(s+1\right)+\frac{2}{\beta}\right ) g_{\beta}\left(\beta\left(s+\frac{1}{2}\right)+\frac{2}{\beta}\right ) } { g_{\beta}\left(\beta(s+2)+\frac{3}{\beta}\right)g_{\beta}\left(\beta s+\frac{1}{\beta}\right)}c_{\beta}\label{momgfinal } \end{split}\ ] ] where @xmath202 is a constant determined by the condition @xmath203 . inserting into the right - hand side of ( which is now manifestly self - dual )
leads to the following expression at @xmath154 : @xmath204 where @xmath205 and @xmath206 is a constant which determines the shift in the maximum as discussed below .
the value @xmath207 in is conjectured from the results of numerical simulations in the next section .
the latter formula constitutes our main analytical result and finally leads to our conjecture [ pred ] .
the purpose of this section is to provide a numerical test of conjecture [ pred ] . in figure
[ fig : n3000m250kgue ] we present a histogram of the recentered and rescaled maximum of the gue characteristic polynomial , defined by @xmath208 with @xmath209 defined in ( [ maxrv ] ) .
here we used the matrix size @xmath16 and @xmath83 realizations of the gue ensemble .
the dashed red line is the exact probability density of the random variable @xmath21 defined via its laplace transform in . in we
have recentered and scaled by @xmath210 and @xmath211 , presumably a consequence of finite-@xmath14 effects due to the @xmath212 terms in .
note that the influence of shift / recentering is already quite small compared with the predicted considerably larger @xmath213 shift .
the parameters @xmath214 and @xmath215 were calculated empirically from the mean and variance of @xmath21 in according to the formula @xmath216 as derived by requiring @xmath217 and @xmath218 . in table
[ table : cnsn ] we display values of the parameters @xmath214 and @xmath215 for the studied range of sizes @xmath14 , as determined empirically from the mean and variance of the random variable @xmath82 .
the observed decay with @xmath14 is certainly consistent with asymptotic validity of our conjecture [ pred ] , though the convergence to the asymptotic results is too slow to make more definite claims . to resolve further decrease of the coefficients @xmath214 and @xmath215 would require much larger matrices and is computationally demanding .
finally , we provide a numerical validation of the decomposition . in figure
[ fig : uprimefig ] we plot the inverse laplace transform of obtained by a direct numerical evaluation of the integral in the bromwich inversion formula for the laplace transform .
the positive and normalized curve clearly corresponds to a _ bona fide _ probability density of some real random variable @xmath32 .
the numerical evaluation of the maximum value may be considered quite a non - trivial problem in its own right , for at least two reasons .
firstly , the characteristic polynomial @xmath11 having zeros as the eigenvalues of @xmath219 , displays @xmath220 oscillations in the spectral interval @xmath18 $ ] with hugely varying peaks heights .
this produces considerable clusterings of ` near - maxima ' which may confuse any naive attempt to find the true maximum value .
secondly , the slow changing nature of the correction terms in conjecture [ pred ] , of order @xmath221 and @xmath222 ) respectively , require one to go to somewhat large matrices to resolve reasonable asymptotic behaviour .
the problem is further compounded by the numerical instability of calculating determinants of such matrices .
our solution to these problems heavily relies on a sparse realization of gue matrices @xmath1 originally due to trotter @xcite ( see also dumitriu and edelman @xcite ) .
he discovered that the eigenvalues of gue matrices @xmath1 have the same joint probability density as those of the following real symmetric tri - diagonal matrix : @xmath223 where @xmath224 is a normal random variable with mean @xmath225 and variance @xmath226 .
the sub - diagonal is composed of random variables @xmath227 having the same density as @xmath228 where @xmath229 is a @xmath230-square random variable with @xmath231 degrees of freedom . to compute the maximum value of @xmath232
, we begin by exploiting the known asymptotic behaviour @xmath233 so that @xmath234 further progress is now possible thanks to the fact that determinants of tri - diagonal matrices satisfy a linear recurrence relation .
furthermore , by an appropriate rescaling , the recursion computes determinants of all leading principal minors simultaneously , thus computing @xmath235 _ for all _ @xmath236 in linear time . now to find the maximum , we define a mesh @xmath237 with @xmath238 and evaluate @xmath239 at each of the points in @xmath240 . at those points where
@xmath239 is maximal the matlab function ` fminbnd ' is invoked to converge onto the global maximum .
figure [ fig : n3000gueres ] illustrates the complexity of the problem .
our algorithm is sufficiently precise to distinguish the true maximum ( located at @xmath241 in red ) from other possible candidates , _
e.g. _ @xmath242 as well as the thousands of other local maxima .
d. carpentier and p. le doussal .
glass transition of a particle in a random potential , front selection in nonlinear renormalization group , and entropic phenomena in liouville and sinh - gordon models . , 026110 , 33pp ( 2001 ) y. v. fyodorov , p. le doussal , and a. rosso .
tatistical mechanics of logarithmic rem : duality , freezing and extreme value statistics of 1/f noises generated by gaussian free fields .
10 , p10005 , 32 pp ( 2009 ) | motivated by recently discovered relations between logarithmically correlated gaussian processes and characteristic polynomials of large random @xmath0 matrices @xmath1 from the gaussian unitary ensemble ( gue ) , we consider the problem of characterising the distribution of the global maximum of @xmath2 as @xmath3 and @xmath4 .
we arrive at an explicit expression for the asymptotic probability density of the ( appropriately shifted ) maximum by combining the rigorous fisher - hartwig asymptotics due to krasovsky @xcite with the heuristic _ freezing transition _ scenario for logarithmically correlated processes .
although the general idea behind the method is the same as for the earlier considered case of the circular unitary ensemble , the present gue case poses new challenges . in particular we show how the conjectured _ self - duality _ in the freezing scenario plays the crucial role in our selection of the form of the maximum distribution .
finally , we demonstrate a good agreement of the found probability density with the results of direct numerical simulations of the maxima of @xmath5 . |
there has been a great deal of attention devoted in recent years to the properties of the hot x - ray - emitting plasma ( the intracluster medium , hereafter icm ) in the central regions of massive galaxy groups and clusters . to a large extent , the focus has been on the competition between radiative cooling losses and various mechanisms that could be heating the gas and therefore ( at least partially ) offsetting the effects of cooling . prior to the launch of the _ chandra _ and _ xmm - newton _ x - ray observatories
, it was generally accepted that large quantities of the icm should be cooling down to very low energies , where it would cease to emit x - rays and eventually condense out into molecular clouds and form stars ( fabian 1994 ) .
however , the amount of cold gas and star formation actually observed in most systems is well below what was expected based on analyses of _ rosat _ , _ asca _ , and _ einstein _
x - ray data ( e.g. , voit & donahue 1995 ; edge 2001 ) . new high spatial and spectral resolution data from _ chandra _ and _ xmm - newton _ has shown that , while cooling is clearly important in many groups and clusters ( so - called ` cool core ' systems ) , relatively little gas is actually cooling out of the x - ray band to temperatures below @xmath0 k ( peterson et al . 2003 ) .
it is now widely believed that some energetic form of non - gravitational heating is compensating for the losses due to cooling .
indeed , it seems likely that such heating goes beyond a simple compensation for the radiated energy .
as well as having a profound effect on the properties of the icm , it seems that the heat input also has important consequences for the formation and evolution of the central brightest cluster galaxy ( bcg ) and therefore for the bright end of the galaxy luminosity function ( e.g. , benson et al .
2003 ; bower et al . 2006 , 2008 ) .
the thermal state of the icm therefore provides an important probe of these processes and a concerted theoretical effort has been undertaken to explore the effects of various heating sources ( e.g. , supermassive black holes , supernovae , thermal conduction , dynamical friction heating of orbiting satellites ) using analytic and semi - analytic models , in addition to idealised and full cosmological hydrodynamic simulations ( e.g. , binney & tabor 1995 ; narayan & medvedev 2001 ; ruszkowski et al . 2004 ; mccarthy et al . 2004 ; kim et al . 2005 ;
voit & donahue 2005 ; sijacki et al .
2008 ) . in most of these approaches
, it is implicitly assumed that _ gravitationally - induced _ heating ( e.g. , via hydrodynamic shocks , turbulent mixing , energy - exchange between the gas and dark matter ) that occurs during mergers / accretion is understood and treated with a sufficient level of accuracy .
for example , most current analytic and semi - analytic models of groups and clusters attempt to take the effects of gravitational heating into account by assuming distributions for the gas that are taken from ( or inspired by ) non - radiative cosmological hydrodynamic simulations the assumption being that these simulations accurately and self - consistently track gravitational processes .
however , it has been known for some time that , even in the case of identical initial setups , the results of non - radiative cosmological simulations depend on the numerical scheme adopted for tracking the gas hydrodynamics . in particular , mesh - based eulerian ( such as adaptive mesh refinement , hereafter amr ) codes appear to systematically produce higher entropy ( lower density ) gas cores within groups and clusters than do particle - based lagrangian ( such as smoothed particle hydrodynamics , hereafter sph ) codes ( see , e.g. , frenk et al .
1999 ; ascasibar et al .
2003 ; voit , kay , & bryan 2005 ; dolag et al .
2005 ; oshea et al .
this may be regarded as somewhat surprising , given the ability of these codes to accurately reproduce a variety of different analytically solvable test problems ( e.g. , sod shocks , sedov blasts , the gravitational collapse of a uniform sphere ; see , e.g. , tasker et al .
2008 ) , although clearly hierarchical structure formation is a more complex and challenging test of the codes . at present
, the origin of the cores and the discrepancy in their amplitudes between eulerian and lagrangian codes in cosmologically - simulated groups and clusters is unclear .
there have been suggestions that it could be the result of insufficient resolution in the mesh simulations , artificial entropy generation in the sph simulations , galilean non - invariance of the mesh simulations , and differences in the amount of mixing in the sph and mesh simulations ( e.g. , dolag et al . 2005 ; wadsley et al .
we explore all of these potential causes in 4 .
clearly , though , this matter is worth further investigation , as it potentially has important implications for the competition between heating and cooling in groups and clusters ( the cooling time of the icm has a steep dependence on its core entropy ) and the bright end of the galaxy luminosity function . and
it is important to consider that the total heating is not merely the sum of the gravitational and non - gravitational heating terms .
the rankine - hugoniot jump conditions tell us that the efficiency of shock heating ( either gravitational or non - gravitational in origin ) _ depends on the density of the gas at the time of heating _ ( see , e.g. , the discussion in mccarthy et al .
this implies that if gas has been heated before being shocked , the entropy generated in the shock can actually be amplified ( voit et al .
2003 ; voit & ponman 2003 ; borgani et al . 2005
; younger & bryan 2007 ) .
the point is that gravitational and non - gravitational heating will couple together in complex ways , so it is important that we are confident that gravitational heating is being handled with sufficient accuracy in the simulations .
a major difficulty in studying the origin of the cores in cosmological simulations is that the group and cluster environment can be extremely complex , with many hundreds of substructures ( depending on the numerical resolution ) orbiting about at any given time .
furthermore , such simulations can be quite computationally - expensive if one wishes to resolve in detail the innermost regions of groups and clusters ( note that , typically , the simulated cores have sizes @xmath1 ) .
an alternative approach , which we adopt in the present study , is to use idealised simulations of binary mergers to study the relevant gravitational heating processes .
the advantages of such an approach are obviously that the environment is much cleaner , therefore offering a better chance of isolating the key processes at play , and that the systems are fully resolved from the onset of the simulation .
the relatively modest computational expense of such idealised simulations also puts us in a position to be able to vary the relevant physical and numerical parameters in a systematic way and to study their effects .
idealised merger simulations have been used extensively to study a variety of phenomena , such as the disruption of cooling flows ( gmez et al .
2002 ; ritchie & thomas 2002 ; poole et al . 2008 ) , the intrinsic scatter in cluster x - ray and sunyaev - zeldovich effect scaling relations ( ricker & sarazin 2001 ; poole et al .
2007 ) , the generation of cold fronts and related phenomena ( e.g. , ascasibar & markevitch 2006 ) , and the ram pressure stripping of orbiting galaxies ( e.g. , mori & burkert 2000 ; mccarthy et al . 2008b ) .
however , to our knowledge , idealised merger simulations have not been used to elucidate the important issue raised above , nor have they even been used to demonstrate whether or not this issue even exists in non - cosmological simulations . in the present study
, we perform a detailed comparison of idealised binary mergers run with the widely - used public simulations codes ( an amr code ) and ( a sph code ) .
the paper is organised as follows . in
2 we give a brief description of the simulation codes and the relevant adopted numerical parameters .
in addition , we describe the initial conditions ( e.g. , structure of the merging systems , mass ratio , orbit ) of our idealised simulations . in 3 , we present a detailed comparison of results from the and runs and confirm that there is a significant difference in the amount of central entropy generated with the two codes . in
4 we explore several possible causes for the differences we see .
finally , in 5 , we summarise and discuss our findings .
below , we provide brief descriptions of the and hydrodynamic codes used in this study and the parameters we have adopted .
the interested reader is referred to springel , yoshida , & white ( 2001 ) and springel ( 2005b ) for in - depth descriptions of and to fryxell et al .
( 2000 ) for the code .
both codes are representative examples of their respective amr and sph hydrodynamic formulations , as has been shown in the recent code comparison of tasker et al .
( 2008 ) .
[ [ section ] ] is a publicly available amr code developed by the alliances center for astrophysical thermonuclear flashes .
originally intended for the study of x - ray bursts and supernovae , it has since been adapted for many astrophysical conditions and now includes modules for relativistic hydrodynamics , thermal conduction , radiative cooling , magnetohydrodynamics , thermonuclear burning , self - gravity and particle dynamics via a particle - mesh approach . in this study
we use version 2.5 .
solves the reimann problem using the piecewise parabolic method ( ppm ; colella & woodward 1984 ) .
the present work uses the default parameters , which have been thoroughly tested against numerous analytical tests ( see fryxell et al .
the maximum number of newton - raphson iterations permitted within the riemann solver was increased in order to allow it to deal with particularly sharp shocks and discontinuities whilst the default tolerance was maintained .
the hydrodynamic algorithm also adopted the default settings with periodic boundary conditions being applied to the gas as well as to the gravity solver .
we have modified the gravity solver in to use an fft on top of the multigrid solver ( written by t. theuns ) .
this results in a vast reduction in the time spent calculating the self - gravity of the simulation relative to the publicly available version .
we have rigorously tested the new algorithm against the default multigrid solver , more tests are presented in tasker et al .
( 2008 ) . to identify regions of rapid flow change , s refinement and de - refinement criteria
can incorporate the adapted lhner ( 1987 ) error estimator .
this calculates the modified second derivative of the desired variable , normalised by the average of its gradient over one cell . with
this applied to the density as is common place , we imposed the additional user - defined criteria whereby the density has to exceed a threshold of @xmath2 , below which the refinement is set to the minimum @xmath3 mesh .
this restricts the refinement to the interior of the clusters and , as we demonstrate below , was found to yield nearly identical results to uniform grid runs with resolution equal to the maximum resolution in the equivalent amr run .
uses an oct - tree block - structured amr grid , in which the block to be refined is replaced by 8 blocks ( in three dimensions ) , of which the cell size is one half of that of the parent block .
each block contains the same number of cells , @xmath4 cells in each dimension in our runs .
the maximum allowed level of refinement , @xmath5 , is one of the parameters of the run . at refinement level @xmath5
, a fully refined amr grid will have @xmath6 cells on a side .
all the merger runs are simulated in 20 mpc on a side periodic boxes in a non - expanding ( newtonian ) space and are run for a duration of @xmath7 gyr . by default
, our simulations are run with a maximum of @xmath8 levels of refinement ( @xmath9 cells when fully refined ) , corresponding to a minimum cell size of @xmath10 kpc , which is small in comparison to the entropy cores produced in non - radiative cosmological simulations of clusters ( but note we explicitly test the effects of resolution in 3 ) .
the simulations include non - radiative hydrodynamics and gravity .
is a publicly available treesph code designed for the simulation of cosmological structure formation . by default , the code implements the entropy conserving sph formalism proposed by springel & hernquist ( 2002 ) .
the code is massively parallel , has been highly optimised and is very memory efficient .
this has led to it being used for some of the largest cosmological simulations to date , including the first n - body simulation with more than @xmath11 dark matter particles ( the _ millennium simulation _ ; springel et al .
2005a ) .
the sph formalism is inherently lagrangian in nature and fully adaptive , with ` refinement ' based on the density .
the particles represent discrete mass elements with the fluid variables being obtained through a kernel interpolation technique ( lucy 1977 ; gingold & monaghan 1977 ; monaghan 1992 ) .
the entropy injected through shocks is captured through the use of an artificial viscosity term ( see , e.g. , monaghan 1997 ) .
we will explore the sensitivity of our merger simulation results to the artificial viscosity in [ viscosity ] . by default
, gravity is solved through the use of a combined tree particle - mesh ( treepm ) approach .
the treepm method allows for substantial speed ups over the traditional tree algorithm by calculating long range forces with the particle - mesh approach using fast fourier techniques .
higher gravitational spatial resolution is then achieved by applying the tree over small scales only , maintaining the dynamic range of the tree technique .
this allows to vastly exceed the gravitational resolving power of mesh codes which rely on the particle - mesh technique alone and are thus limited to the minimum cell spacing .
we adopt the following numerical parameters by default for our runs ( but note that most of these are systematically varied in 3 ) . the artificial bulk viscosity , @xmath12 ,
is set to 0.8 .
the number of sph smoothing neighbours , @xmath13 , is set to 32 .
each of our @xmath14 model clusters ( see 2.1 ) are comprised of @xmath15 gas and dark matter particles within @xmath16 , and the gas to total mass ratio is @xmath17 .
thus , the particle masses are @xmath18 and @xmath19 .
the gravitational softening length is set to 10 kpc , which corresponds to @xmath20 initially .
all the merger runs are simulated in 20 mpc on a side periodic boxes in a non - expanding ( newtonian ) space and are run for a duration of @xmath21 10 gyr .
the simulations include basic hydrodynamics only ( i.e. , are non - radiative ) . in our simulations ,
the galaxy clusters are initially represented by spherically - symmetric systems composed of a realistic mixture of dark matter and gas .
the dark matter is assumed to follow a nfw distribution ( navarro et al . 1996 ; 1997 ) : @xmath22 where @xmath23 and @xmath24 here , @xmath16 is the radius within which the mean density is 200 times the critical density , @xmath25 , and @xmath26 .
the dark matter distribution is fully specified once an appropriate scale radius ( @xmath27 ) is selected .
the scale radius can be expressed in terms of the halo concentration @xmath28 .
we adopt a concentration of @xmath29 for all our systems .
this value is typical of massive clusters formed in @xmath30cdm cosmological simulations ( e.g. , neto et al .
2007 ) . in order to maintain the desired nfw configuration ,
appropriate velocities must be assigned to each dark matter particle .
for this , we follow the method outlined in mccarthy et al .
briefly , the three velocity components are selected randomly from a gaussian distribution whose width is given by the local velocity dispersion [ i.e , @xmath31 . the velocity dispersion profile itself is determined by solving the jeans equation for the mass density distribution given in eqn .
( 1 ) . as in mccarthy
( 2007 ) , the dark matter haloes are run separately in isolation in for many dynamical times to ensure that they have fully relaxed . for the gaseous component
, we assume a powerlaw configuration for the entropy is the not the actual thermodynamic specific entropy ( @xmath32 ) of the gas , but is related to it via the simple relation @xmath33 .
however , for historical reasons we will refer to @xmath34 as the entropy .
] , @xmath35 , by default .
in particular , @xmath36 where the ` virial entropy ' , @xmath37 , is given by @xmath38 this distribution matches the entropy profiles of groups and clusters formed in the non - radiative cosmological simulations of voit , kay , & bryan ( 2005 ) ( vkb05 ) for @xmath39 .
it is noteworthy that vkb05 find that this distribution approximately matches the entropy profiles of both sph ( the code ) and amr ( the code ) simulations . within @xmath40 , however , the amr and sph simulations show evidence for entropy cores , but of systematically different amplitudes .
we initialise our systems without an entropy core [ i.e. , eqn .
( 3 ) is assumed to hold over all radii initially ] to see , first ,
if such cores are established during the merging process and , if so , whether the amplitudes differ between the sph and amr runs , as they do in cosmological simulations . we leave it for future work to explore the differences that result ( if any ) between sph and amr codes when large cores are already present in the initial systems ( e.g. , the merger of two ` non - cool core ' systems ) . with the mass distribution of dark matter established ( i.e. , after having run the dark matter haloes in isolation ) and an entropy distribution for the gas given by eqn .
( 3 ) , we numerically compute the radial gas pressure profile ( and therefore also the gas density and temperature profiles ) , taking into account the self - gravity of the gas , by simultaneously solving the equations of hydrostatic equilibrium and mass continuity : @xmath41 two boundary conditions are required to solve these equations .
the first condition is that @xmath42 .
the second condition is that the total mass of hot gas within @xmath16 yields a realistic baryon fraction of @xmath43 . in order to meet the second condition
, we choose a value for @xmath44 and propagate the solution outwards to @xmath16 .
we then iteratively vary the inner pressure until the desired baryon fraction is achieved . for the simulations ,
the gas particle positions are assigned by radially morphing a glass distribution until the desired gas mass profile is obtained ( see mccarthy et al .
the entropy ( or equivalently internal energy per unit mass ) of each particle is specified by interpolating with eqn . ( 3 ) . for the simulations , the gas density and entropy of each grid cell is computed by using eqn . ( 3 ) and interpolating within the radial gas density profile resulting from the solution of eqns . ( 5 ) and ( 6 ) .
thus , for both the and simulations we start with identical dark matter haloes ( using particle positions and velocities from the isolated runs with a gravitational softening length of 10kpc ) and gas haloes , which have been established by interpolating within radial profiles that have been computed numerically under the assumption that the gas is initially in hydrostatic equilibrium within the dark matter halo . note that when varying the resolution of the simulations we simply change the maximum number of refinements @xmath5 we do not vary the number of dark matter particles in this way , the initial dark matter distribution is always the same as in the low resolution run in all the simulations we have run . in both the and simulations ,
the gaseous haloes are surrounded by a low density pressure - confining gaseous medium that prevents the systems from expanding prior to the collision ( i.e. , so that in the case of an isolated halo the object would be static ) but otherwise it has a negligible dynamical effect on the system .
isolated gas+dm haloes were run in both and for 10 gyr in order to test the stability of the initial gas and dark matter haloes .
although deviations in the central entropy develop over the course of the isolated simulations , indicating the systems are not in perfect equilibrium initially , they are small in amplitude ( the central entropy increases by @xmath45 over 10 gyr ) , especially in comparison to the factor of @xmath46 jump in the central entropy that occurs as a result of shock heating during the merger . furthermore
, we note that the amplitude of the deviations in the isolated runs are significantly decreased as the resolution of these runs is increased .
our merger simulations , however , are numerically converged ( see 3 ) , indicating that the deviations have a negligible effect on merger simulation results and the conclusions we have drawn from them .
= 8.4 cm = 8.4 cm
the existence of a discrepancy between the inner properties of the gas in groups and clusters formed in amr and sph cosmological simulations was first noticed in the santa barbara code comparison of frenk et al .
it was subsequently verified in several works , including dolag et al .
( 2005 ) , oshea et al .
( 2005 ) , kravtsov , nagai & vikhlinin ( 2005 ) , and vkb05 .
the latter study in particular clearly demonstrated , using a relatively large sample of @xmath47 simulated groups and clusters , that those systems formed in the amr simulations had systematically larger entropy cores than their sph counterparts . since this effect
was observed in cosmological simulations , it was generally thought that the discrepancy was due to insufficient resolution in the mesh codes at high redshift ( we note , however , that vkb05 argued against resolution being the cause ) .
this would result in under - resolved small scale structure formation in the early universe .
this explanation is consistent with the fact that in the santa barbara comparison the entropy core amplitude tended to be larger for the lower resolution mesh code runs .
our first aim is therefore to determine whether the effect is indeed due to resolution limitations , or if it is due to a more fundamental difference between the two types of code .
we test this using identical idealised binary mergers of spherically - symmetric clusters in and , where it is possible to explore the effects of finite resolution with relatively modest computational expense ( compared to full cosmological simulations ) . as a starting point
, we investigate the generation of entropy cores in a head on merger between two identical @xmath48 clusters , each colliding with an initial speed of @xmath49 km / s [ i.e. , the initial _ relative _ velocity is @xmath50 , which is typical of merging systems in cosmological simulations ; see , e.g. , benson 2005 ] .
the system is initialised such that the two clusters are just barely touching ( i.e. , their centres are separated by @xmath51 ) .
the simulations are run for a duration of 10 gyr , by the end of which the merged system has relaxed and there is very little entropy generation ongoing .
l l l flash sim .
cells & max .
spatial res .
+ & & ( kpc ) + @xmath52 & equiv .
@xmath53 & 78 + @xmath8 ( default ) & equiv .
@xmath9 & 39 + @xmath54 & equiv .
@xmath55 & 19.5 + @xmath56 & equiv .
@xmath57 & 9.8 + @xmath9 & @xmath9 & 39 + + gadget-2 sim . &
gas particles & max .
spatial res .
+ & & ( kpc ) + low res .
& @xmath58 & @xmath59 + default & @xmath60 & @xmath61 + hi res .
& @xmath62 & @xmath63 + our idealised test gives very similar results to non - radiative cosmological simulations there is a distinct difference in the amplitude of the entropy cores in the amr and sph simulations , with the entropy in the mesh code a factor @xmath64 higher than the sph code .
it is evident that the difference between the codes is captured in a single merger event .
an immediate question is whether this is the result of the different effective resolutions of the codes .
resolution tests can be seen in the left hand panel of figure [ entropy_radius ] , where we plot the resulting radial entropy distributions . for
, we compare runs with @xmath58 , @xmath60 ( the default ) , and @xmath62 particles .
for we compare amr runs with minimum cell sizes of @xmath65 , @xmath66 ( the default ) , @xmath67 , and @xmath68 kpc and a uniform grid run with the default 39 kpc cell size .
the simulation characteristics for these head on mergers are presented in table 1 .
to make a direct comparison with the cosmological results of vkb05 ( see their fig .
5 ) , we normalise the entropy by the initial ` virial ' entropy ( @xmath37 ; see eqn . 4 ) and the radius by the initial virial radius , @xmath16 .
the plot clearly shows that the simulations converge on two distinctly different solutions within the inner ten percent of @xmath16 , whereas the entropy at large radii shows relatively good agreement between the two codes .
the simulations performed for the resolution test span a factor of 8 in spatial resolution in and approximately a factor of 5 in .
the amr runs effectively converge after reaching a peak resolution equivalent to a @xmath69 run ( i.e. , a peak spatial resolution of @xmath70 kpc or @xmath71 ) .
we have also tried a run with a uniform ( as opposed to adaptive ) @xmath9 grid and the results essentially trace the amr run with an equivalent peak resolution .
this reassures us that our amr refinement criteria is correctly capturing all regions of significance .
the lowest resolution sph run , which only has @xmath72 gas particles within @xmath16 initially , has a slightly higher final central entropy than the default and high resolution sph runs .
this may not be surprising given the tests and modelling presented in steinmetz & white ( 1997 ) .
these authors demonstrated that with such small particle numbers , two - body heating will be important if the mass of a dark matter particle is significantly above the mass of a gas particle .
the runs converge , however , when the number of gas and dark matter particles are increased by an order magnitude ( i.e. , as in our default run ) , yielding a maximum spatial resolution of @xmath61 kpc ( here we use the minimum sph smoothing length as a measure of the maximum spatial resolution ) . a comparison of the left hand panel of fig . 1 to fig .
5 of vkb05 reveals a remarkable correspondence between the results of our idealised merger simulations and those of their cosmological simulations ( which spanned system masses of @xmath73 ) .
they find that the ratio of the amr and sph core amplitudes is @xmath64 in both the idealised and cosmological simulations .
this difference is also seen in the santa barbara comparison of frenk et al .
( 1999 ) when comparisons are made between the sph simulations and the highest resolution amr simulations carried out in that study ( ie . ,
the ` bryan ' amr results ) amr run in fig . 1 ) .
this consistency presumably indicates that whatever mechanism is responsible for the differing core amplitudes in the cosmological simulations is also responsible for the differing core amplitudes in our idealised simulations .
this is encouraging , as it implies the generation of the entropy cores can be studied with idealised simulations .
as outlined in 1 , the advantage of idealised simulations over cosmological simulations is their relative simplicity .
this gives us hope that we can use idealised simulations to track down the underlying cause of the discrepancy between particle - based and mesh - based hydrodynamic codes .
the right hand panel of figure [ entropy_radius ] shows the resulting entropy distributions plotted in a slightly different fashion .
here we plot the entropy as a function of ` enclosed ' gas mass @xmath74 .
this is constructed by simply ranking the particles / cells by entropy in ascending order and then summing the masses of the particles / cells [ the inverse , @xmath75 , would therefore be the total mass of gas with entropy lower than @xmath34 ] .
convective stability ensures that , eventually when the system is fully relaxed , the lowest - entropy gas will be located at the very centre of the potential well , while the highest entropy gas will be located at the system periphery .
@xmath74 is therefore arguably a more fundamental quantity than @xmath77 and we adopt this test throughout the rest of the paper .
it is also noteworthy that in order to compute @xmath78 one does not first need to select a reference point ( e.g. , the centre of mass or the position of the particle with the lowest potential energy ) or to bin the particles / cells in any way , both of which could introduce ambiguities in the comparison between the sph and amr simulations ( albeit likely minor ones ) . in the right hand panel of figure [ entropy_radius ] , we plot the resulting @xmath74 distributions normalised to the final entropy distribution of the default run .
here we see that the lowest - entropy gas in the runs have a higher entropy , by a factor of @xmath79 , than the lowest - entropy gas in the default run .
naively , looking at the right hand panel of figure [ entropy_radius ] one might conclude that the discrepancy is fairly minor , given that @xmath80% of the gas has been heated to a similar degree in the sph and amr simulations .
but it is important to keep in mind that it is the properties of the lowest - entropy gas in particular that are most relevant to the issue of heating vs. cooling in groups and clusters ( and indeed in haloes of all masses ) , since this is the gas that has the shortest cooling time .
the agreement between our results and those from cosmological simulations ( e.g. , frenk et al .
1999 ; vkb05 ) is striking .
the convergence of the entropy distributions in our idealised simulations negates the explanation that inadequate resolution of the high redshift universe in cosmological amr simulations is the root cause of the discrepancy between the entropy cores in sph and amr simulations ( although we note that some of the lower resolution amr simulations in the study of frenk et al .
may not have been fully converged and therefore the discrepancy may have been somewhat exaggerated in that study for those simulations ) .
we therefore conclude that the higher entropy generation in amr codes relative to sph codes within the cores of groups and clusters arises out of a more fundamental difference in the adopted algorithms . below
we examine in more detail how the entropy is generated during the merging process in the simulations and we then systematically explore several possible causes for the differences in the simulations .
we have demonstrated that the entropy generation that takes place in our idealised mergers is robust to our choice of resolution , yet a difference persists in the amount of central entropy that is generated in the sph and mesh simulations .
we now examine the entropy generation as a function of time in the simulations , which may provide clues to the origin of the difference between the codes .
figure [ 8panelimage ] shows @xmath81 in a slice through the centre of the default simulation at times 0 , 1 , 2 , 3 , 4 , 5 , 7 and 10 gyr .
this may be compared to figure [ entropygastime ] , which shows the entropy distribution of the simulations as a function of time ( this figure is described in detail below ) .
briefly , as the cores approach each other , a relatively gentle shock front forms between the touching edges of the clusters , with gas being forced out perpendicular to the collision axis .
strong heating does not actually occur until approximately the time when the cores collide , roughly 1.8 gyrs into the run .
the shock generated through the core collision propagates outwards , heating material in the outer regions of the system .
this heating causes the gas to expand and actually overshoot hydrostatic equilibrium .
eventually , the gas , which remains gravitationally bound to the system , begins to fall back onto the system , producing a series of weaker secondary shocks .
gas at the outskirts of the system , which is the least bound , takes the longest to re - accrete .
this dependence of the time for gas to be re - accreted upon the distance from the centre results in a more gradual increase in entropy than seen in the initial core collision . in a _ qualitative _ sense , the heating process that takes place in the simulations is therefore very similar to that seen in the simulations ( see 3 of mccarthy et al .
2007 for an overview of the entropy evolution in idealised mergers ) .
the top left panel in figure [ entropygastime ] shows the ratio of @xmath74 in the default run relative to @xmath74 in the default run .
the various curves represent the ratio at different times during the simulations ( see figure key note that these correspond to the same outputs displayed in figure [ 8panelimage ] ) .
it can clearly be seen that the bulk of the difference in the _ final _ entropy distributions of the simulations is established around the time of core collision .
the ratio of the central entropy in the simulation to the central entropy in the simulation converges after @xmath82 gyr .
the top right panel shows the time evolution of the lowest - entropy gas only in both the and runs .
here we see there are similar trends with time , in the sense that there are two main entropy generation episodes ( core collision and re - accretion ) , but that the entropy generated in the first event is much larger in the run than in the run . far outside the core , however , the results are very similar . for completeness ,
the bottom two panels show @xmath74 at different times for the and runs separately .
the small initial drop in the central entropy at 1 gyr in the run ( see bottom left panel ) is most likely due to interpolation errors at low resolution . this drop in entropy should not physically occur without cooling processes ( which are not included in our simulations ) , but there is nothing to prevent a dip from occurring in the simulations due to numerical inaccuracies ( the second law of thermodynamics is not explicitly hardwired into the mesh code ) . at low resolutions , small violations in entropy conservation can occur due to inaccurate interpolations made by the code .
we have verified that the small drop in entropy does not occur in the higher resolution runs .
we note that while this effect is present in default run , it is small and as demonstrated in fig .
1 the default run is numerically converged .
it is interesting that the to central entropy ratio converges relatively early on in the simulations .
this is in spite of the fact that a significant fraction of the entropy that is generated in both simulations is actually generated at later times , during the re - accretion phase .
evidently , this phase occurs in a very similar fashion in both simulations . in 4 , we will return to the point that the difference between the results of the amr and sph simulations arises around the time of core collision .
= 6.0 cm = 6.0 cm = 6.0 cm = 6.0 cm it is important to verify that the conclusions we have drawn from our default setup are not unique to that specific initial configuration . using a suite of merger simulations of varying mass ratio and orbital parameters , mccarthy et al .
( 2007 ) demonstrated that the entropy generation that takes place does so in a qualitatively similar manner to that described above in all their simulations .
however , these authors examined only sph simulations .
we have therefore run several additional merger simulations in both and to check the robustness of our conclusions .
all of these mergers are carried out using the same resolution as adopted for the default and runs . in figure
[ orbit_test ] , we plot the final to @xmath74 ratio for equal mass mergers with varying orbital parameters ( see figure caption ) . in all cases , systematically produces larger entropy cores than , and by a similar factor to that seen in the default merger setup .
interestingly , the off - axis case results in a somewhat larger central entropy discrepancy between and , even though the bulk energetics of this merger are the same as for the default case .
a fundamental difference between the off - axis case and the default run is that the former takes a longer time for the cores to collide and subsequently relax ( but note by the end of the off - axis simulation there is very little ongoing entropy generation , as in the default case ) .
this may suggest that the timescale over which entropy is generated plays some role in setting the magnitude of the discrepancy between the amr and sph simulations .
for example , one possibility is that ` pre - shocking ' due to the artificial viscosity i.e. , entropy generation during the early phases of the collision when the interaction is subsonic or mildly transonic ) in the sph simulations becomes more relevant over longer timescales .
another possibility is that mixing , which is expected to be more prevalent in eulerian mesh simulations than in sph simulations , plays a larger role if the two clusters spend more time in orbit about each other before relaxing into a single merged system ( of course , one also expects enhanced mixing in the off - axis case simply because of the geometry ) .
we explore these and other possible causes of the difference in 4 .
in addition to varying the orbital parameters , we have also experimented with colliding a cluster composed of dark matter only with another cluster composed of a realistic mixture of gas and dark matter ( in this case , we simulated the head on merger of two equal mass @xmath48 clusters with an initial relative velocity of @xmath83 km / s ) .
obviously , this is not an astrophysically reasonable setup .
however , a number of studies have suggested that there is a link between the entropy core in clusters formed in non - radiative cosmological simulations and the amount of energy exchanged between the gas and the dark matter in these systems ( e.g. , lin et al . 2006 ; mccarthy et al . 2007 ) . it is therefore interesting to see whether this experiment exposes any significant differences with respect to the results of our default merger simulation . ) .
the long - dashed green and short - dashed cyan curves represent head on collisions with initial relative velocities of @xmath84 and @xmath85 ( i.e. , at rest initially ) .
the dot dashed red curve represents an off - axis collision with an initial relative radial velocity of @xmath86 and an initial relative tangential velocity of @xmath87 ( i.e. , the total energy is equivalent to that of the default setup ) .
also shown ( dotted magenta curve ) , is the entropy ratio of a run where one of the clusters is composed of dark matter only and the other of a realistic mixture of gas and dark matter ( see text ) .
all these simulations result in a comparable difference in entropy profile between the mesh code and the sph code .
, width=317 ] the dotted magenta curve in figure [ orbit_test ] represents the final to @xmath74 ratio for the case where a dark matter only cluster merges with another cluster composed of both gas and dark matter .
the results of this test are remarkably similar to that of our default merger case .
this indicates that the mechanism responsible for the difference in heating in the mesh and sph simulations in the default merger simulation is also operating in this setup .
although this does not pin down the difference between the mesh and sph simulations , it does suggest that the difference has little to do with differences in the properties of the large hydrodynamic shock that occurs at core collision , as there is no corresponding large hydrodynamic shock in the case where one cluster is composed entirely of dark matter .
however , it is clear from figure 3 that the difference between the default mesh and sph simulations is established around the time of core collision , implying that some source of heating other than the large hydrodynamic shock is operating at this time ( at least in the simulation ) .
we return to this point in 4 .
there are fundamental differences between eulerian mesh - based and lagrangian particle - based codes in terms of how they compute the hydrodynamic and gravitational forces .
ideally , in the limit of sufficiently high resolution , the two techniques would yield identical results for a given initial setup .
indeed , both techniques have been shown to match with high accuracy a variety of test problems with known analytic solutions . however , as has been demonstrated above ( and in other recent studies ; e.g. , agertz et al . 2007 ; trac et al . 2007 ; wadsley et al .
2008 ) differences that do not appear to depend on resolution present themselves in certain complex , but astrophysically - relevant , circumstances . in what follows , we explore several different possible causes for why the central heating that takes place in mesh simulations exceeds that in the sph simulations .
the possible causes we explore include : * [ gravitysolvers ] a difference in gravity solvers - most currently popular mesh codes ( including and ) use a particle - mesh ( pm ) approach to calculate the gravitational force . to accurately capture short range
forces it is therefore necessary to have a finely - sampled mesh .
by contrast , particle - based codes ( such as and ) often make use of tree algorithms or combined tree particle - mesh ( treepm ) algorithms , where the tree is used to compute the short range forces and a mesh is used to compute long range forces .
since the gravitational potential can vary rapidly during major mergers and large quantities of mass can temporarily be compressed into small volumes , it is conceivable differences in the gravity solvers and/or the adopted gravitational force resolution could give rise to different amounts of entropy generation in the simulations .
* [ galileaninv ] galilean non - invariance of mesh codes - given explicit dependencies in the riemann solver s input states , all eulerian mesh codes are inherently not galilean invariant to some degree .
this can lead to spurious entropy generation in the cores of systems as they merely translate across the simulation volume ( e.g. , tasker et al .
* [ viscosity ] ` pre - shocking ' in the sph runs - artificial viscosity is required in sph codes to capture the effects of shock heating .
however , the artificial viscosity can in principle lead to entropy production in regions where no shocks should be present ( e.g. , dolag et al .
if such pre - shocking is significant prior to core collision in our sph simulations , it could result in a reduced efficiency of the primary shock . * [ mixing ] a difference in the amount of mixing in sph and mesh codes - mixing will be suppressed in standard sph implementations where steep density gradients are present , since rayleigh - taylor and kelvin - helmholtz instabilities are artificially damped in such circumstances ( e.g. , agertz et al .
in addition , the standard implementation of artificial viscosity will damp out even _ subsonic _ motions in sph simulations , thereby inhibiting mixing ( dolag et al .
on the other hand , one expects there to be some degree of over - mixing in mesh codes , since fluids are implicitly assumed to be fully mixed on scales smaller than the minimum cell size .
+ + we now investigate each of these possible causes in turn .
we do not claim that these are the only possible causes for the differences we see in the simulations .
they do , however , represent the most commonly invoked possible solutions ( along with hydrodynamic resolution , which we explored in 3 ) to the entropy core discrepancy between sph and mesh codes . 5 , 6 , 7 , and 8 , respectively , which correspond to peak grid cell sizes of @xmath65 , @xmath66 , @xmath67 , and @xmath68 kpc ( respectively ) . the thin solid red curve represents the uniform @xmath9 run . the gravitational softening length adopted for the run is 10 kpc . for reference , @xmath88 kpc .
the vertical dashed line indicates four softening lengths .
the dark matter distribution converges to the result when the numerical resolutions become similar : the observed differences in gas entropy are not due to differences in the underlying dark matter dynamics .
, width=317 ] in the simulations , gravity is computed using a standard particle - mesh approach . with this approach ,
the gravitational force will be computed accurately only on scales larger than the finest cell size . by contrast , the simulations make use of a combined treepm approach , where the tree algorithm computes the short range gravitational forces and the particle - mesh algorithm is used only to compute long range forces . to test whether or not differences in the gravity solvers ( and/or gravitational force resolution )
are important , we compare the final mass distributions of the dark matter in our simulations .
the distribution of the dark matter should be insensitive to the properties of the diffuse baryonic component , since its contribution to the overall mass budget is small by comparison to the dark matter .
thus , the final distribution of the dark matter tells us primarily about the gravitational interaction alone between the two clusters .
figure [ dmconverge ] shows the ratio of the final dark matter mass profiles to the final dark matter mass profile . recall that in all runs the number of dark matter particles is the same .
the differences that are seen in this figure result from solving for gravitational potential on a finer mesh .
for the lowest resolution run , we see that the final dark matter mass profile deviates significantly from that of the default run for @xmath89 .
however , this should not be surprising , as the minimum cell size in the default run is @xmath90 . by increasing the maximum refinement level , @xmath5
, we see that the discrepancy between the final and dark matter mass profiles is limited to smaller and smaller radii . with @xmath91 ,
the minimum cell size is equivalent to the gravitational softening length adopted in the default run . in this case
, the final dark matter mass distribution agrees with that of the default run to within a few percent at all radii beyond a few softening lengths ( or a few cell sizes ) , which is all that should be reasonably expected .
a comparison of the various runs with one another ( compare , e.g. , the default run with the @xmath56 run , for which there is a @xmath92% discrepancy out as far as @xmath40 ) may suggest a somewhat slower rate of convergence to the default result than one might naively have expected .
given that we have tested the new fft gravity solver against both the default multigrid solver and a range of simple analytic problems and confirmed its accuracy to a much higher level than this , we speculate that the slow rate of convergence is due to the relatively small number of dark matter particles used in the mesh simulations . in the future , it would be useful to vary the number of dark matter particles in the mesh simulations to verify this hypothesis . in summary , we find that the resulting dark matter distributions agree very well in the and simulations when the effective resolutions are comparable .
the intrinsic differences between the solvers therefore appear to be minor .
more importantly for our purposes , even though the gravitational force resolution for the default run is not as high as for the default run , this has no important consequences for the comparison of the final entropy distributions of the gas .
it is important to note that , even though the final mass distribution in the simulations shows small differences between @xmath93 and 8 , figure [ entropy_radius ] shows that the entropy distribution is converged for @xmath94 and is not at all affected by the improvement in the gravitational potential .
and 45 degrees with @xmath56 , respectively .
this comparison illustrates that the effects of galilean non - invariance on the resulting entropy distribution are minor and do not account for the difference in the entropy core amplitudes of the mesh and sph simulations .
, width=317 ] due to the nature of riemann solvers ( which are a fundamental feature of amr codes ) , it is possible for the evolution of a system to be galilean non - invariant .
this arises from the fixed position of the grid relative to the fluid . the riemann shock tube initial conditions are constructed by determining the amount of material that can influence the cell boundary from either side , within a given time step based on the sound speed
the riemann problem is then solved at the boundary based on the fluid properties either side of the boundary . by applying a bulk velocity to the medium in a given direction , the nature of the solution changes .
although an ideal solver would be able to decouple the bulk velocity from the velocity discontinuity at the shock , the discrete nature of the problem means that the code may not be galilean invariant .
since one expects large bulk motions to be relevant for cosmological structure formation , and clearly is quite relevant for our merger simulations , it is important to quantify what effects ( if any ) galilean non - invariance has on our amr simulations .
we have tested the galilean non - invariance of our simulations in two ways . in the first test
, we simulate an isolated cluster moving across the mesh with an ( initial ) bulk velocity of @xmath50 ( @xmath83 km / s ) and compare it to an isolated cluster with zero bulk velocity .
this is similar to the test carried out recently by tasker et al .
( 2008 ) . in agreement with
tasker et al .
( 2008 ) , we find that there is some spurious generation of entropy in the very central regions ( @xmath95 ) of the isolated cluster that was given an initial bulk motion .
however , after @xmath96 gyr of evolution ( i.e. , the time when the clusters collide in our default merger simulation ) , the increase in the central entropy is only @xmath97% .
this is small in comparison to the @xmath98% jump that takes place at core collision in our merger simulations .
this suggests that spurious entropy generation prior to the merger is minimal and does not account for the difference we see between the sph and amr simulations .
in the second test , we consider different implementations of the default merger simulation .
in one case , instead of giving both systems equal but opposite bulk velocities ( each with magnitude @xmath99 ) , we fix one and give the other an initial velocity that is twice the default value , so that the relative velocity is unchanged .
( we also tried reducing the size of the time steps for this simulation by an order magnitude . ) in addition , we have tried mergers that take place at oblique angles relative to the grid . if the merger is well - resolved and the dynamics are galilean invariant
, all these simulations should yield the same result .
figure [ flashgalilean ] shows the resulting entropy distributions for these different runs .
the results of this test confirm what was found above ; i.e. , that there is some dependence on the reference frame adopted , but that this effect is minor in general ( the central entropy is modified by @xmath100% ) and does not account for the discrepancy we see between entropy core amplitudes in the default and simulations . ) until @xmath101 1.6 , 1.7 , 1.8 gyr , respectively , at which point the artificial viscosity is set back to its default value .
the solid magenta squares represent a run with low artificial viscosity throughout , and the open triangles represent the default simulation . reducing the value of the artificial viscosity parameter before the cores collide delays the increase in entropy ( cyan , green and solid red ) , however as soon as the original value is restored , the entropy @xmath34 increases to a level nearly independent of when @xmath102 was restored .
therefore pre - shocking has little effect on the post - shock value of @xmath34 .
, width=317 ] artificial viscosity is required in sph codes in order to handle hydrodynamic shocks .
the artificial viscosity acts as an excess pressure in the equation of motion , converting gas kinetic energy into internal energy , and therefore raising the entropy of the gas . in standard sph implementations ,
the magnitude of the artificial viscosity is fixed in both space and time for particles that are approaching one another ( it being set to zero otherwise ) .
this implies that even in cases where the mach number is less than unity , i.e. , where formally a shock should not exist , ( spurious ) entropy generation can occur .
this raises the possibility that significant ` pre - shocking ' could occur in our sph merger simulations .
this may have the effect of reducing the efficiency of the large shock that occurs at core collision and could therefore potentially explain the discrepancy between the mesh and sph simulations .
= 8.4 cm = 8.4 cm dolag et al .
( 2005 ) raised this possibility and tested it in sph cosmological simulations of massive galaxy clusters .
they implemented a new variable artificial viscosity scheme by embedding an on - the - fly shock detection algorithm in that indicates if particles are in a supersonic flow or not . if so , the artificial viscosity is set to a typical value , if not the artificial viscosity is greatly reduced .
this new implementation should significantly reduce the amount of pre - shocking that takes place during formation of the clusters .
the resulting clusters indeed had somewhat higher central entropies relative to clusters simulated with the standard artificial viscosity implementation ( although the new scheme does not appear to fully alleviate the discrepancy between mesh and sph codes ) .
however , whether the central entropy was raised because of the reduction in pre - shocking or if it was due to an increase in the amount of mixing is unclear .
our idealised mergers offer an interesting opportunity to re - examine this test .
in particular , because of the symmetrical geometry of the merger , little or no mixing is expected until the cores collide , as prior to this time there is no interpenetration of the gas particles belonging to the two clusters ( we have verified this ) .
this means that we are in a position to isolate the effects of pre - shocking from mixing early on in the simulations .
to do so , we have devised a crude method meant to mimic the variable artificial viscosity scheme of dolag et al .
in particular , we run the default merger with a low artificial viscosity ( with @xmath103 , i.e. , approximately the minimum value adopted by dolag et al . ) until the cores collide , at which point we switch the viscosity back to its default value .
we then examine the amount of entropy generated in the large shock .
figure [ viscosityswitch ] shows the evolution of the central entropy around the time of core collision .
shown are a few different runs where we switch the artificial viscosity back to its default value at different times ( since the exact time of ` core collision ' is somewhat ill - defined ) . here
we see that prior to the large shock very little entropy has been generated , which is expected given the low artificial viscosity adopted up to this point .
a comparison of these runs to the default simulation ( see inset in figure [ viscosityswitch ] ) shows that there is evidence for a small amount of pre - shocking in the default run .
however , we find that for the cases where the artificial viscosity is set to a low value , the resulting entropy jump ( after the viscosity is switched back to the default value ) is nearly the same as in the default merger simulation . in other words , pre - shocking appears to have had a minimal effect on the strength of the heating that occurs at core collision in the default sph simulation .
this argues against pre - shocking as the cause of the difference we see between the mesh and sph codes .
lastly , we have also tried varying @xmath12 over the range @xmath104 and @xmath105 ( i.e. , values typically adopted in sph studies ; springel 2005b ) for the default run .
we find that the sph results are robust to variations in @xmath12 and can not reconcile the differences between sph and amr results .
= 8.4 cm = 8.4 cm our experiments with off - axis collisions and collisions with a cluster containing only dark matter suggest that mixing plays an important role in generating the differences between the codes .
several recent studies ( e.g. , dolag et al . 2005 ; wadsley et al .
2008 ) have argued that mixing is handled poorly in standard implementations of sph , both because ( standard ) artificial viscosity acts to damp turbulent motions and because the growth of kh and rt instabilities is inhibited in regions where steep density gradients are present ( agertz et al .
2007 ) . using cosmological sph simulations that have been modified in order to enhance mixing , dolag et al .
( 2005 ) and wadsley et al .
( 2008 ) have shown that it is possible to generate higher central entropies in their galaxy clusters ( relative to clusters simulated using standard implementations of sph ) , yielding closer agreement with the results of cosmological mesh simulations .
this is certainly suggestive that mixing may be the primary cause of the discrepancy between mesh and sph codes .
however , these authors did not run mesh simulations of galaxy clusters and therefore did not perform a direct comparison of the amount of mixing in sph vs. mesh simulations of clusters .
even if one were to directly compare cosmological sph and mesh cluster simulations , the complexity of the cosmological environment and the hierarchical growth of clusters would make it difficult to clearly demonstrate that mixing is indeed the difference .
our idealised mergers offer a potentially much cleaner way to test the mixing hypothesis .
to do so , we re - run the default merger simulation but this time we include a large number of ` tracer particles ' , which are massless and follow the hydrodynamic flow of the gas during the simulation .
the tracer particles are advanced using a second order accurate predictor - corrector time advancement scheme with the particle velocities being interpolated from the grid ( further details are given in the flash manual ( version 2.5 ) at : + http://flash.uchicago.edu/ ) .
each tracer particle has a unique i d that is preserved throughout the simulation , allowing us to track the gas in a lagrangian fashion , precisely as is done in lagrangian sph simulations . to simplify the comparison further ,
we initially distribute the tracer particles within the two clusters in our simulation in exactly the same way as the particles in our initial setup . in the left hand panel of figure [ tracers ] ,
we plot the final vs. the initial entropy of particles in the default and merger simulations .
this plot clearly demonstrates that the lowest - entropy gas is preferentially heated in both simulations , however the degree of heating of that gas in the mesh simulation is much higher than in the sph simulation .
consistent with our analysis in 3 , we find that the bulk of this difference is established around the time of core collision .
it is also interesting that the scatter in the final entropy ( for a given initial entropy ) is much larger in the mesh simulation .
the larger scatter implies that convective mixing is more prevalent in the mesh simulation . at or immediately following core collision ( @xmath106 gyr ) , there is an indication that , typically , gas initially at the very centre of the two clusters ( which initially had the lowest entropy ) has been heated more strongly than gas further out [ compare , e.g. , the median @xmath107 at @xmath108 to the median at @xmath109 .
such an entropy inversion does not occur in the sph simulations and likely signals that the extra mixing in the mesh simulation has boosted entropy production .
= 8.4 cm = 8.4 cm in the right hand panel of figure [ tracers ] we plot the final vs. the initial enclosed gas mass of particles in the default and merger simulations .
the enclosed gas mass of each particle ( or tracer particle ) is calculated by summing the masses of all other particles ( or cells ) with entropies lower than the particle under consideration , by summing the masses of all particles with potentials lower than @xmath110 for the @xmath111th particle . ] .
this plot confirms our mixing expectations based on the entropy plot in the left hand panel .
in particular , only a small amount of mass mixing is seen in the sph simulation , whereas in the mesh simulation the central @xmath112 of the gas mass has been fully mixed .
the higher degree of mixing in the simulation is shown pictorially in figure [ particles ] .
the left panel shows the final spatial distribution of the initially lowest - entropy particles in simulation , while the right panel is the analogous plot for tracer particles in the simulation ( see figure caption ) . the larger degree of mixing in the mesh simulation relative to the sph simulation is clearly evident . in the simulation , particles from the two clusters
are intermingled in the final state , while distinct red and blue regions are readily apparent in the sph calculation , a difference which arises immediately following core collision . the increased mixing boosts entropy production in the simulations , but what is the origin of the increased mixing ?
we now return to the point raised in 3 , that the bulk of the difference between mesh and sph simulation is established around the time of core collision .
this is in spite of the fact that significant entropy generation proceeds in both simulations until @xmath113 gyr .
evidently , both codes treat the entropy generation in the re - accretion phase in a very similar manner .
what is different about the initial core collision phase ? as pointed out recently by agertz et al .
( 2007 ) , sph suppresses the growth of instabilities in regions where steep density gradients are present due to spurious pressure forces acting on the particles .
could this effect be responsible for the difference we see ? to test this idea , we generate 2d projected entropy maps of the sph and mesh simulations to search for signs of clear instability development .
in the case of the sph simulation , we first smooth the particle entropies ( and densities ) onto a 3d grid using the sph smoothing kernel and the smoothing lengths of each particle computed by .
we then compute a gas mass - weighted projected entropy by projecting along the @xmath114-axis . in the case of the simulation
, the cell entropies and densities are interpolated onto a 3d grid and projected in the same manner as for the simulation .
figure [ instabilities ] shows a snap shot of the two simulations at @xmath115 gyr , just after core collision .
large vortices and eddies are easily visible in the projected entropy map of the simulation but none are evident in the simulation . in order to study the duration of these eddies ,
we have generated 100 such snap shots for each simulation , separated by fixed 0.1 gyr intervals .
analysing the projected entropy maps as a movie , we find that these large vortices and eddies persist in the simulation from @xmath116 gyr .
this corresponds very well with the timescale over which the difference between the sph and mesh codes is established ( see , e.g. , the dashed black curve in the top right panel of figure [ entropygastime ] ) .
we therefore conclude that extra mixing in the mesh simulations , brought on by the growth of instabilities around the time of core collision , is largely responsible for the difference in the final entropy core amplitudes between the mesh and sph simulations .
physically , one expects the development of such instabilities , since the kh timescale , @xmath117 , is relatively short at around the time of core collision .
we therefore conclude that there is a degree of under - mixing in the sph simulations whether or not the simulations yield the correct result , however , is harder to ascertain . as fluids are _ forced _ to numerically mix on scales smaller than the minimum cell size , it is possible that there is a non - negligible degree of over - mixing in the mesh simulations .
our resolution tests ( see 3 ) show evidence for the default mesh simulation being converged , but it may be that the resolution needs to be increased by much larger factors than we have tried ( or are presently accessible with current hardware and software ) in order to see a difference .
in this paper , we set out to investigate the origin of the discrepancy in the entropy structure of clusters formed in eulerian mesh - based simulations compared to those formed in lagrangian sph simulations .
while sph simulations form clusters with almost powerlaw entropy distributions down to small radii , eulerian simulations form much larger cores with the entropy distribution being truncated at significantly higher values .
previously it has been suspected that this discrepancy arose from the limited resolution of the mesh based methods , making it impossible for such codes to accurately trace the formation of dense gas structures at high redshift . by running simulations of the merging of idealised clusters
, we have shown that this is not the origin of the discrepancy .
we used the code ( springel 2005b ) to compute the sph solution and the code ( fryxell et al .
2000 ) to compute the eulerian mesh solution . in these idealised simulations ,
the initial gas density structure is resolved from the onset of the simulations , yet the final entropy distributions are significantly different .
the magnitude of the difference generated in idealised mergers is comparable to that seen in the final clusters formed in full cosmological simulations .
a resolution study shows that the discrepancy in the idealised simulations can not be attributed to a difference in the effective resolutions of the simulations .
thus , the origin of the discrepancy must lie in the code s different treatments of gravity and/or hydrodynamics .
we considered various causes in some detail .
we found that the difference was _ not _ due to : * the use of different gravity solvers .
the two codes differ in that uses a treepm method to determine forces , while uses the pm method alone .
the different force resolutions of the codes could plausibly lead to differences in the energy transfer between gas and dark matter .
yet we find that the dark matter distributions produced by the two codes are almost identical when the mesh code is run at comparable resolution to the sph code .
* galilean non - invariance of mesh codes .
we investigate whether the results are changed if we change the rest - frame defined by the hydrodynamic mesh .
although we find that an artificial core can be generated in this way in the mesh code , its size is much smaller than the core formed once the clusters collide , and is not enough to explain the difference between and .
we show that most of this entropy difference is generated in the space of @xmath118 gyr when the cluster cores first collide . * pre - shocking in sph .
we consider the possibility that the artificial viscosity of the sph method might generate entropy in the flow prior to the core collision , thus reducing the efficiency with which entropy is generated later . by greatly reducing the artificial viscosity ahead of the core collision ,
we show that this effect is negligible . having shown that none of these numerical issues can explain the difference of the final entropy distributions , we investigated the role of fluid mixing in the two codes .
several recent studies ( e.g. , dolag et al . 2005 ; wadsley et al .
2008 ) have argued that if one increases the amount of mixing in sph simulations the result is larger cluster entropy cores that resemble the amr results . while this is certainly suggestive
, it does not clearly demonstrate that it is the enhanced mixing in mesh simulations that is indeed the main driver of the difference ( a larger entropy core in the mesh simulations need not necessarily have been established by mixing ) . by injecting tracer particles into our simulations ,
we have been able to make an explicit comparison of the amount of mixing in the sph and mesh simulations of clusters .
we find very substantial differences . in the sph computation , there is a very close relation between the initial entropy of a particle and its final entropy .
in contrast , tracer particles in the simulation only show a close connection for high initial entropies .
the lowest @xmath119% of gas ( by initial entropy ) is completely mixed in the simulation .
we conclude that mixing and fluid instabilities are the cause of the discrepancy between the simulation methods .
the origin of this mixing is closely connected to the suppression of turbulence in sph codes compared to the eulerian methods .
this can easily be seen by comparing the flow structure when the clusters collide : while the image is dominated by large scale eddies , these are absent from the sph realisation ( see figure [ instabilities ] ) .
it is now established that sph codes tend to suppress the growth of kelvin - helmholtz instabilities in shear flows , and this seems to be the origin of the differences in our simulation results ( e.g. , agertz et al .
these structures result in entropy generation through mixing , an irreversible process whose role is underestimated by the sph method .
of course , it is not clear that the turbulent structures are correctly captured in the mesh simulations ( iapichino & niemeyer 2008 ; wadsley et al .
the mesh forces fluids to be mixed on the scale of individual cells . in nature , this is achieved through turbulent cascades that mix material on progressively smaller and smaller scales : the mesh code may well overestimate the speed and effectiveness of this process . ultimately , deep x - ray observations may be able to tell us whether the mixing that occurs in the mesh simulations is too efficient .
an attempt at studying large - scale turbulence in clusters was made recently by schuecker et al .
their analysis of _ xmm - newton _ observations of the coma cluster indicated the presence of a scale - invariant pressure fluctuation spectrum on scales of 40 - 90 kpc and found that it could be well described by a projected kolmogorov / oboukhov - type turbulence spectrum . if the observed pressure fluctuations are indeed driven by scale - invariant turbulence , this would suggest that current mesh simulations have the resolution required to accurately treat the turbulent mixing process . alternatively , several authors have suggested that icm may be highly viscous ( eg . , fabian et al .
2003 ) with the result that fluid instabilities will be strongly suppressed by physical processes .
this might favour the use of sph methods which include a physical viscosity ( sijacki & springel 2006 ) .
it is a significant advance that we now understand the origin of this long standing discrepancy .
our work also has several important implications .
firstly , as outlined in 1 , there has been much discussion in the recent literature on the competition between heating and cooling in galaxy groups and clusters .
the current consensus is that heating from agn is approximately sufficient to offset cooling losses in observed cool core clusters ( e.g. , mcnamara & nulsen 2007 ) .
however , observed present - day agn power output seems energetically incapable of explaining the large number of systems that do not possess cool cores% of all massive x - ray clusters in flux - limited samples do not have cool cores ( e.g. , chen et al .
since at fixed mass cool core clusters tend to be more luminous than non - cool core clusters , the fraction of non - cool core cluster in flux - limited samples may actually be an underestimate of the true fraction . ]
( mccarthy et al .
recent high resolution x - ray observations demonstrate that these systems have higher central entropies than typical cool core clusters ( e.g. , dunn & fabian 2008 ) . one way of getting around the energetics issue is to invoke an early episode of preheating ( e.g. , kaiser 1991 ; evrard & henry 1991 ) .
energetically , it is more efficient to raise the entropy of the ( proto-)icm prior to it having fallen into the cluster potential well , as its density would have been much lower than it is today ( mccarthy et al .
preheating remains an attractive explanation for these systems .
however , as we have seen from our idealised merger simulations , the amount of central entropy generated in our mesh simulations is significant and is even comparable to the levels observed in the central regions of non - cool core clusters .
it is therefore tempting to invoke mergers and the mixing they induce as an explanation for these systems .
however , before a definitive statement to this effect can be made , much larger regions of parameter space should be explored . in particular , a much larger range of impact parameters and mass ratios is required , in addition to switching on the effects of radiative cooling ( which we have neglected in the present study ) .
this would be the mesh code analog of the sph study carried out by poole et al .
( 2006 ; see also poole et al .
we leave this for future work . alternatively , large cosmological mesh simulations , which self - consistently track the hierarchical growth of clusters , would be useful for testing the merger hypothesis
. indeed , burns et al .
( 2008 ) have recently carried out a large mesh cosmological simulation ( with the code ) and argue that mergers at high redshift play an important role in the establishment of present - day entropy cores . however , these results appear to be at odds with the cosmological mesh simulations ( run with the code ) of nagai et al .
( 2007 ) ( see also kravtsov et al .
these authors find that most of their clusters have large cooling flows at the present - day , similar to what is seen in some sph cosmological simulations ( e.g. , kay et al .
2004 ; borgani et al .
on the other hand , the sph simulations of keres et al .
( 2008 ) appear to yield clusters with large entropy cores .
this may be ascribed to the lack of effective feedback in their simulations , as radiative cooling selectively removes the lowest entropy gas ( see , e.g. , bryan 2000 ; voit et al .
2002 ) , leaving only high entropy ( long cooling time ) gas remaining in the simulated clusters .
however , all the simulations just mentioned suffer from the overcooling problem ( balogh et al .
2001 ) , so it is not clear to what extent the large entropy cores in clusters _ in either mesh or sph _ simulations are produced by shock heating , overcooling , or both .
all of these simulations implement different prescriptions for radiative cooling ( e.g. , metal - dependent or not ) , star formation , and feedback , and this may lie at the heart of the different findings
. a new generation of cosmological code comparisons will be essential in sorting out these apparently discrepant findings
. the focus should not only be on understanding the differences in the hot gas properties , but also on the distribution and amount of stellar matter , as the evolution of the cold and hot baryons are obviously intimately linked .
reasonably tight limits on the amount of baryonic mass in the form of stars now exists ( see , e.g. , balogh et al.z 2008 ) and provides a useful target for the next generation of simulations . at present , merger - induced mixing as an explanation for intrinsic scatter in the hot gas properties of groups and clusters remains an open question . secondly , we have learnt a great deal about the nature of gas accretion and the development of hot gas haloes from sph simulations of the universe . since we now see that these simulations may underestimate the degree of mixing that occurs , which of these results are robust , which need revision ?
for example , keres et al .
( 2005 ) ( among others ) have argued that cold accretion by galaxies plays a dominant role in fuelling the star formation in galaxies .
is it plausible that turbulent eddies could disrupt and mix such cold streams as they try to penetrate through the hot halo ?
we can estimate the significance of the effect by comparing the kelvin - helmholtz timescale , @xmath117 , with the free - fall time , @xmath120 .
the kh timescale is given by ( see , e.g. , nulsen 1982 ; price 2008 ) @xmath121
where @xmath122 and @xmath123 is the density of the hot halo , @xmath124 is the density of the cold stream , @xmath125 is the wave number of the instability , and @xmath126 the velocity of the stream relative to the hot halo .
if the stream and hot halo are in approximate pressure equilibrium , this implies a large density contrast ( e.g. , a @xmath127 k stream falling into a @xmath60 k hot halo of a milky way - type system would imply a density contrast of @xmath128 ) . in the limit of @xmath129 and recognising that the mode responsible for the destruction of the stream is comparable to the size of the stream ( i.e. , @xmath130 ) , eqns .
( 7 ) and ( 8) reduce to : @xmath131 adopting @xmath132 , @xmath133 kpc , and @xmath134 km / s ( perhaps typical numbers for a cold stream falling into a milky way - type system ) , we find @xmath135 gyr . the free - fall time , @xmath136 [ where @xmath137 is the virial radius of main system and @xmath138 is the circular velocity of the main system at its virial radius ] , is @xmath118 gyr for a milky way - type system with mass @xmath139 .
on this basis , it seems that the stream would be stable because of the large density contrast in the flows .
it is clear , however , that the universality of these effects need to be treated with caution , as the free - fall and kh timescales are not vastly discrepant .
high resolution mesh simulations ( cosmological or idealised ) of milky way - like systems would provide a valuable check of the sph results . finally , the sph method has great advantages in terms of computational speed , effective resolution and galilean invariance .
is it therefore possible to keep these advantages and add additional small scale transport processes to the code in order to offset the suppression of mixing ?
wadsley et al . (
2008 ) and price ( 2008 ) have presented possible approaches based on including a thermal diffusion term in the sph equations .
although the approaches differ in their mathematical details , the overall effect is the same .
however , it is not yet clear how well this approach will work in cosmological simulations that include cooling ( and feedback ) , since the thermal diffusion must be carefully controlled to avoid unphysical suppression of cooling in hydrostatic regions ( e.g. , dolag et al .
one possibility might be to incorporate such terms as a negative surface tension in regions of large entropy contrast ( hu & adams 2006 ) . an alternative approach is to combine the best features of the sph method , such as the way that it continuously adapts to the local gas density and its flow , with the advantage of a riemann based method of solving the fluid dynamic equations ( e.g. , inutsuka 2002 ) .
clearly , there is a great need to find simple problems in which to test these codes : simple shock tube experiments are not sufficient because they do not include the disordered fluid motions that are responsible for generating the entropy core .
idealised mergers represent a step forward , but the problem is still not sufficiently simple that it is possible to use self - similar scaling techniques ( e.g. , bertschinger 1985 , 1989 ) to establish the correct solution .
one possibility is to consider the generation of turbulent eddies in a fluctuating gravitational potential .
we have begun such experiments , but ( although the fluid flow patterns are clearly different ) simply passing a gravitational potential through a uniform plasma at constant velocity does not expose the differences between sph and eulerian mesh based methods that we see in the idealised merger case .
we will tackle the minimum complexity that is needed to generate these differences in a future paper .
the authors thank the referee for a careful reading of the manuscript and suggestions that improved that paper .
they also thank volker springel , mark voit , and michael balogh for very helpful discussions .
nlm and rac acknowledge support from stfc studentships .
igmcc acknowledges support from a nserc postdoctoral fellowship at the icc in durham and a kavli institute fellowship at the kavli institute for cosmology , cambridge .
these simulations were performed and analysed on cosma , the hpc facilities at the icc in durham and we gratefully acknowledge kind support from lydia heck for computing support .
the software used in this work was in part developed by the doe - supported asc / alliance center for astrophysical thermonuclear flashes at the university of chicago . | the diffuse plasma that fills galaxy groups and clusters ( the intracluster medium , hereafter icm ) is a by - product of galaxy formation . the present thermal state of this gas results from a competition between gas cooling and heating .
the heating comes from two distinct sources : gravitational heating associated with the collapse of the dark matter halo and additional thermal input from the formation of galaxies and their black holes .
a long term goal of this research is to decode the observed temperature , density and entropy profiles of clusters and to understand the relative roles of these processes .
however , a long standing problem has been that cosmological simulations based on smoothed particle hydrodynamics ( sph ) and eulerian mesh - based codes predict different results even when cooling and galaxy / black hole heating are switched off .
clusters formed in sph simulations show near powerlaw entropy profiles , while those formed in eulerian simulations develop a core and do not allow gas to reach such low entropies . since the cooling rate is closely connected to the minimum entropy of the gas distribution , the differences are of potentially key importance . in this paper , we investigate the origin of this discrepancy . by comparing simulations
run using the sph code and the adaptive eulerian mesh code , we show that the discrepancy arises during the idealised merger of two clusters , and that the differences are not the result of the lower effective resolution of eulerian cosmological simulations .
the difference is not sensitive to the minimum mesh size ( in eulerian codes ) or the number of particles used ( in sph codes ) .
we investigate whether the difference is the result of the different gravity solvers , the galilean non - invariance of the mesh code or an effect of unsuitable artificial viscosity in the sph code .
instead , we find that the difference is inherent to the treatment of vortices in the two codes .
particles in the sph simulations retain a close connection to their initial entropy , while this connection is much weaker in the mesh simulations .
the origin of this difference lies in the treatment of eddies and fluid instabilities .
these are suppressed in the sph simulations , while the cluster mergers generate strong vortices in the eulerian simulations that very efficiently mix the fluid and erase the low entropy gas .
we discuss the potentially profound implications of these results .
[ firstpage ] hydrodynamics methods : n - body simulations galaxies : clusters : general cosmology : theory |
in the quantum chromodynamics ( qcd ) there is a very fruitful analogy between superconductivity and qcd : according to the t hooft and mandelstam assumption @xcite , @xcite a confinement in the qcd is similar to the dual meissner effect in a superconductor .
such confinement scenario assumes that the ground state of qcd is a condensate of magnetic monopoles which one confines ( pinches ) the color electric field of color charges into flux tubes .
in other words , the magnetic charges , defined as the dirac monopoles in ground state of qcd , should condense in the vacuum in just the same way as the cooper pair in a superconductor .
this analogy works in the direction : _ * superconductivity @xmath1 quantum chromodynamics . *
_ an analogy can be offered in the opposite direction : from the qcd to the high - t@xmath0 superconductivity by the following manner .
an interaction between gluons exists in the qcd and it is so strong that the force lines of color gauge field are stretched into a flux tube between quark and antiquark .
the most important in this picture is the strong interaction between gluons .
this one distinguishes the quantum electrodynamics ( qed ) from the quantum chromodynamics : qed does not have the photon - photon interaction .
thus , our basic assumption is that _ * in the high - t@xmath0 superconductor there is a strong phonon - phonon interaction*_. we assume that the jump from the ordinary superconductivity to high - t@xmath0 one is analogous to the jump from qed to qcd .
@xmath2 the consequence of such conjectural analogy is that the interaction between phonons is so strong that _ the phonons between the cooper pair are confined into a phonon tube ( pt ) _
@xcite .
the phonon hamiltonian @xmath3 in the continuum limit we can write as follows @xmath4 , \label{1 - 1}\ ] ] where @xmath5 is the mass density , @xmath6 ( @xmath7 ) is the ion deviation from the equilibrium state , @xmath8 is the potential energy , @xmath9 are some coefficients .
our assumption indicates that the potential term @xmath8 in the high - t@xmath0 superconductor can not be simplified as @xmath10 even though for the small deviations @xmath11 .
the same takes place in the qcd with lagrangian @xmath12 here @xmath13 ; @xmath14 ; @xmath15 is the su(3 ) gauge potential ; @xmath16 are the structural constants of the su(3 ) gauge group ; @xmath17 are the color indexes .
it is easy to see that in this qcd lagrangian we have @xmath18 and @xmath19 terms .
the most important is that coupling constant @xmath20 is not small , _
i.e. _ in contrast with qed we have a very strong interaction between gluons that probably leads to the appearance of a nonlocal object - flux tube stretched between quark and antiquark .
the basic idea of bcs theory is that in the presence of even a weak interaction between electrons they cooperate in the cooper pairs that leads to decreasing the ground state energy of superconductor in comparison with the normal state .
according to the above - mentioned assumption about the possible analogy with the qcd we suppose * there is a strong interaction between phonons ( nonlinear potential term in ( [ 1 - 1 ] ) expression ) which leads to appearing of a tube filled by phonons ( like the flux tube filled by chromodynamical color fields in the qcd ) .
we can name such tube as the pt .
* in the first approximation , neglecting by cross section of the pt , such object is like a string .
analogously to the qcd in the first rough approximation the pt in the high - t@xmath0 superconductor can be considered as a string . * as in the qcd we assume that the interaction between cooper electrons is described by the nonrelativistic string potential @xmath21 where @xmath8 is the potential energy of the phonon interaction between cooper electrons , @xmath22 is a coefficient and @xmath23 is a length of the string . in this model
the cooper pair is modeled as a string with two cooper electrons attached at its ends . in the qcd
such construction is a string with quark and antiquark at the ends .
the interaction quark - antiquark is so strong that we have confinement : if we increase the string length then in some moment the string is torn and again we will have the quark - antiquark pair , _
i.e. _ we can not obtain a single quark . by applying such model to the high - t@xmath0 superconductivity
we should be delicate : certainly a single electron exists but the basic idea is the same : the string interaction leads to an essential increase of the binding energy of the cooper pair .
now we would like to repeat the calculations in the microscopic bcs theory with the potential ( [ 2 - 1 ] ) ( following , for example , to ref .
@xcite ) . let us assume that @xmath24 is a quantum amplitude that the cooper pair is unoccupied and @xmath25 that it is occupied , then the wave function is @xmath26 where @xmath27 is the probability that the cooper pair @xmath28 , @xmath29 is unoccupied and accordingly @xmath30 is the probability that it is occupied ; @xmath31 is the wave vector .
we have the normalization condition @xmath32 the normal ground state is given by @xmath33 where @xmath34 is the fermi wave vector .
the energy of ground state is calculated by the variational method @xmath35 with the constraint @xmath36 , where @xmath37 is the number operator and @xmath38 is the mean number of particles in this system , @xmath39 is the fermi energy .
the calculation for kinetic energy term @xmath40 gives us @xmath41 where the single particle energies @xmath42 are measured from the fermi surface @xmath43 , the factor 2 appears as @xmath30 is the occupation probability for the cooper pair .
the expectation value of the potential energy part is @xmath44 where @xmath45 is the amplitude for the initial state in which the cooper pair @xmath31 is occupied and pair @xmath46 in unoccupied ; @xmath47 is the amplitude for the final state in which the reverse is true .
thus we have @xmath48 as usual , we minimize this functional with respect to @xmath49 that leads to the equation @xmath50 where @xmath51 as usual we simplify @xmath52 here @xmath53 is the debye energy of phonons .
now we can include our string assumption : we suppose that @xmath8 in the ( [ 2 - 14 ] ) expression is @xmath54 where @xmath22 is some constant , @xmath23 is the length of cooper pair ( length of the string ) . as usual
the further calculations give us @xmath55 where @xmath56 is the density of states in energy at the fermi surface .
we can hope that in the high - t@xmath0 superconductor the quantity @xmath8 in the ( [ 2 - 16 ] ) expression will be much more than the corresponding quantity in the ordinary superconductors @xmath57 as it takes place in the qcd : the strong interaction between quarks is much more than the electromagnetic one between electrons ( positrons ) . in the qcd
the interaction between gluons is so strong that the force lines are confined into the flux tube stretched between quark and antiquark .
the greatest difficulty here is the microscopical calculation of the quantity @xmath8 .
this difficulty is connected with the presence of the nonlinear potential term in the hamiltonian .
this means that in the presence of the strong self - interaction between phonons the expression for the exchange of one phonon between two electrons ( see fig .
[ fig1 ] ) @xmath58 now is not correct .
the same problem exists in the qcd : now we can not deduce the existence of the flux tube ( string in the first rough approximation ) from the su(3 ) lagrangian .
this problem is connected with that we do not have appropriate mathematical tools in the nonperturbative qcd , _
i.e. _ without the feynman diagram techniques
. one of the first attempts of nonperturbative calculations in the qcd was a veneziano amplitude @xcite ( see fig.[fig2 ] ) @xmath59 where @xmath60 and @xmath61 are the mandelstam variables ; @xmath62 is the gamma function ; @xmath63 ; @xmath64 and @xmath65 are some constants .
this approach has not resulted in a great success in the qcd as this amplitude is some rough approximation for the correct 4-point green s function .
it is possible to assume that such approximation can be made in the superconductivity theory : the scattering amplitude of two electrons @xmath66 is the veneziano amplitude .
certainly ( just as in qcd ) it is a rough approximation for the correct green s function . we can try to obtain an expression for @xmath8 on the basis of the dimensional reasons .
the result , for example , can be @xmath67 where @xmath68 is the cross section of the pt . if we suppose that the pt is located inside of a layer of an anisotropical superconductor then @xmath69 , where @xmath70 is the thickness of the layer . in this case
these expressions will be @xmath71 because of @xmath72 we have @xmath73 .
what is the physical meaning of the pt ? in the qcd the flux tube is filled with the color field which is parallel to the tube axis and equal to zero outside this tube . in our case
we can suppose that there can be one from the following possibilities @xmath74 the first possibility @xmath75 means that the nonzero static deviation of the ion lattice exists only inside pt .
the second one @xmath76 means that the nonzero mean square of the deviation exists only inside pt only and so on .
a possible strong phonon - phonon interaction can obstruct the application of feynman diagram techniques in our case .
some time ago w. heisenberg had conceived of the difficulties in applying an expansion in small parameters to quantum field theories having strong interactions .
he had investigated the dirac equation with nonlinear terms ( heisenberg equation ) ( see , for example , ref s @xcite - @xcite ) . in these papers
he repeatedly underscored that a nonlinear theory with a large parameter _ * requires the introduction of another quantization rule .
* _ he worked out a quantization method for strong nonlinear field unusing the expansion in a small parameter .
it is possible that in the high - t@xmath0 superconductivity the interaction between phonons is strong making it necessary take into account the interaction between phonons to correctly calculate the energy of the cooper pairs .
heisenberg s basic idea proceeds from the fact that the n - point green functions must be found from some infinity differential equations system derived from the field equation for the field operator .
for example , we present heisenberg quantization for nonlinear spinor field .
the basic equation ( heisenberg equation ) has the following form : @xmath77 = 0 , \label{a1}\ ] ] where @xmath78 are dirac matrices ; @xmath79 is the field operator ; @xmath80 is the dirac adjoint spinor ; @xmath81= \hat\psi ( \hat{\bar\psi } \hat\psi)$ ] or @xmath82 or @xmath83 or @xmath84 .
heisenberg emphasizes that the 2-point green function @xmath85 in this theory differs strongly from the propagator in linear theory .
this difference lies in its behaviour on the light cone .
@xmath85 oscillates strongly on the light cone in contrast to the propagator of the linear theory which has a @xmath70-like singularity .
then heisenberg introduces the @xmath86 functions : @xmath87 where @xmath88 is the time ordering operator .
@xmath89 is a system state characterized by the fundamental eq .
( [ a1 ] ) .
relationship ( [ a2 ] ) allows us to establish a one - to - one correspondence between the system state @xmath89 and the function set @xmath86 .
this state can be defined using the infinite function set of ( [ a2 ] ) .
applying heisenberg s equation ( [ a1 ] ) to ( [ a2 ] ) we can obtain the following infinite equations system : @xmath90 + \nonumber \\ \delta ( { \bf x}_r - { \bf y}_1 ) \tau \left ( { \bf x}_1\ldots { \bf x}_{r-1}{\bf x}_{r+1}\ldots { \bf x}_n| { \bf y}_2\ldots { \bf y}_{r-1}{\bf y}_{r+1}\ldots { \bf y}_n\right ) + \nonumber \\ \delta ( { \bf x}_r - { \bf y}_2 ) \tau \left ( { \bf x}_1 \ldots { \bf x}_{r-1}{\bf x}_{r+1}\ldots { \bf x}_n| { \bf y}_1{\bf y}_2\ldots { \bf y}_{r-1 } { \bf y}_{r+1}\ldots { \bf y}_n\right ) + \ldots . \label{a3}\end{aligned}\ ] ] heisenberg then employs the tamm - dankoff method for getting approximate solutions to the infinite equations system of ( [ a3 ] ) .
the key to this method lies in the fact that the system of equation has an approximate solution derived after cutting off the infinite equation system ( [ a3 ] ) to a finite equation system .
it is necessary to note that a method of solution to eq .
( [ a3 ] ) can be various .
for example , we can try to determine the green s functions using the numerical lattice calculations . here
the important point is the following : the technique of expansion in small parameters ( feynman diagrams ) can not be employed for strong nonlinear fields .
it is possible that as in quantum chromodynamics , where quarks are thought to interact strongly by means of flux tubes , so too in high - t@xmath0 superconductivity phonons may strongly interact among themselves . in ref .
@xcite such a mechanism is applied for the qcd . in this paper
we apply a variant of heisenberg s quantization method to solutions of the classical su(3 ) yang - mills field equations which have bad asymptotic behavior .
after quantization it has been found that the bad features ( _ i.e. _ divergent fields and energy densities ) of these solutions are moderated . from these results
is argued that in general the n - point green s functions for yang - mills theories can have nonperturbative pieces which can not be represented as the sum of feynman diagrams . in heisenberg
s theory the matter and the interacting fields are identical : fundamental spinor field @xmath91 . from a more recent perspective
this is not the case .
an interaction is carried by some kind of boson field . in superconductivity
this is the phonons , in quantum chromodynamics it is the nonabelian @xmath92 gauge field - gluons . in conclusion of this section
we emphasize again heisenberg s statement that _ * the perturbation theory is inapplicable to strong nonlinear fields . * _
thus , the basic assumption supposed here is the following : * _ the energy of cooper pair has an essential contribution coming from an interaction of phonons_*. this means that the corresponding sound wave is a nonlinear wave . in this subsection
we would like to show that heisenberg idea about quantization of strongly interacting fields in fact had been applied in the superconductivity theory for definition of green s function . here
we would like to show that the calculations which was made in ref .
@xcite is an application of the heisenberg quantization idea . in this section
we follow to ref.@xcite .
the hamiltonian of the system electrons describing the properties of a metal in the superconductivity state is @xmath93 d v , \label{green1}\ ] ] where @xmath94 is the operator of spinor field describing electrons , @xmath95 is the electron mass , @xmath96 is some constant and @xmath97 are the spinor indexes . as usually ( it is the same as heisenberg equation for nonlinear spinor field ) the operators @xmath98 and @xmath99 obey the following operator equations @xmath100 as well as in heisenberg method for nonlinear spinor field we have an equation for the 2-point green s function @xmath101 @xmath102 further we have to write an equation for term @xmath103 and so on .
after this we will have an infinite set for green s function .
the main problem here is : how can we cut off this infinite equation system ?
let us cite ref.@xcite : `` @xmath104 for interacting particles , the product of four @xmath98-operators can be expressed in terms of the vertex part , _
i.e. _ it already includes the contributions from various scattering processes . in the weak - interaction model under consideration , these scattering processes involving collisions of particles can be neglected , but at the same time , it must be borne in mind that the ground state of the system differs from the usual state with a filled fermi sphere , because of the presence of bound pairs of electrons .
@xmath104 '' .
this means that in the textbook @xcite it was made the following approximation : the operators @xmath105 and @xmath106 contain terms corresponding to the annihilation and creation of bound pairs ( cooper pairs ) .
it allows to state @xmath107 where @xmath108 and @xmath109 are ground states of system with @xmath38 and @xmath110 particles ( cooper pairs ) , respectively . we should note that this expression is approximate one and following to heisenberg idea it allows us to cut off the above - mentioned infinite equation set for green s function .
the key role here plays the last term in expression .
this expression allows us to split the 4-point green s function with the nonperturbative way not using feynman diagram techniques .
eq . means that we have neglected all effects of scattering particles by each other and the presence of the interaction has been taken into account only as leading to the formation of bound pairs . the third term in the right - hand side of eq .
has been written in complete analogy with the case of a bose gas in the correspondence with the fact that all bound pairs ( cooper pairs ) are `` condensed on the lowest level '' .
the quantity @xmath111 obviously has the same order of magnitude as the density of pairs .
we can introduce the following functions @xmath112 where @xmath39 is a chemical potential .
we now substitute into the equation for the green s function .
we can everywhere omit the first two terms in the right - hand side of since they lead to an additive correction to the chemical potential in the equations for the functions @xmath113 . as a result
we obtain the following equation connecting @xmath114 and @xmath115 : @xmath116 the quantity @xmath117 is defined as @xmath118 an equation for @xmath119 can be obtained in a similar way by using the equation @xmath120 analogously to eq .
we have @xmath121 now we have eq s . and as the finite set of equations for the 2-point green s functions @xmath114 and @xmath122 .
thus , above - mentioned reasonings should convince us that _ * this green s function method in the superconductivity theory is a realization of heisenberg idea developed by him for the quantization of non - linear spinor field .
* _ in this subsection we would like to consider a possibility for the phonons to form a static configuration .
let us write again the lagrangian ( [ 1 - 1 ] ) for phonons in continuous limit @xmath123 . \label{d1}\ ] ] thus , in this model it is assumed that operators of strong nonlinear fiels @xmath11 must satisfy the following equation ( which is implied from the lagrangian ( [ d1 ] ) ) : @xmath124 the multitime formalism of heisenberg s method ( when in @xmath125 ) allows us to investigate the scattering processes in quantum theory .
the simultaneous formalism ( when @xmath126 ) allows us to calculate the mean value of the field , the energy , or any combination of field powers .
let us consider , for example , the simplest case of the scalar field when the vector @xmath127 is replaced by the scalar field @xmath128 .
in this case the lagrangian is @xmath129 \label{d3 - 1}\ ] ] where @xmath130 and the potential term @xmath131 the classical field equation is @xmath132 the quantization of ( [ d3 - 2 ] ) equation gives us @xmath133 it is easy to see that the mean value @xmath134 satisfies the following equation : @xmath135 for the definition of @xmath136 we turn to eq.([d3 - 3 ] ) and obtain ( @xmath137 ) in heisenberg s notation ) : @xmath138 here @xmath139 .
analogously it can be used to derive the infinite equation system for calculating @xmath140 . in the first approximation
we can solve this equation system using the following assumption : @xmath141 and then we can derive the equation @xmath142 this equation is very interesting for us . derrick s
theorem @xcite states that eq . has a static solution with the finite energy only in two dimensional spacetime @xmath143 .
this solution is @xmath144 \label{d5 - 2}\ ] ] where @xmath145 and @xmath146 are some constants .
the first assumption is that such solution can be realized on the layer of high - t@xmath0 superconductor as a line ( wall ) separating regions with different vacua ( @xmath147 ) .
the existence of regions with different vacua can have very interesting physical consequences for a distribution of free electrons in the ion lattice .
following ref.@xcite we present a model example of a fermion coupled to solution .
let we have 2-dimensional dirac equation @xmath148 where @xmath149 is the two - component spinor ; @xmath20 is the coupling constant ; @xmath150 is the soliton ; @xmath151 are the dirac matrices , which we can choose to be the pauli matrices @xmath152 now we search a static solution of dirac equation in the presence of the soliton @xmath153 here we consider case with energy @xmath154 . in this case
we have the following normalizable zero mode @xmath155 \right \}^{-2m / m_\psi } \dbinom{1}{1 } \label{d5 - 5}\ ] ] where @xmath156 .
we see that it is strongly localized at the position @xmath146 of the soliton ( see , fig.[fig3 ] ) . in the context of mechanism with different vacua
this toy model can mean that the free electrons on the layer of the high - t@xmath0 superconductor will be located on the boundaries between regions with different vacua .
evidently , the appearance of localized solutions is connected with the presence of two vacuums ( @xmath147 ) .
it allows us to assume that the mechanism offered here for the formation of cooper pair in high-@xmath157 superconductivity depends on the existence of several vacuums for phonon field , _ i.e. _ on the presence of several equilibrium states ( stable and unstable ) for ions of the lattice .
such interpretation depends on the existence of different vacua .
thus , this mechanism takes place only in the presence of different equilibrium states for ions of the lattice .
it is necessary to note that can exist other mechanism of the field localization in some region .
it is possible , for example , that such mechanism ( without different vacua of gauge field ) occurs in qcd by forming a hypothesized flux tube between quark - antiquark .
another possibility can be connected with cut - in of an external gauge field . in the ref .
@xcite it was shown that the scalar field coupled with the electromagnetic field can forms the flux tube ( nielsen - olesen flux tube ) .
the detailed investigation of such possibility in the context of the gingbug - landau theory is given in ref.@xcite .
it should be pointed out that the investigation of @xmath158 gives us the information about the mean value of the field @xmath159 . for the investigation of questions on the scattering or interaction of phonons
it is necessary to explore the functions , @xmath160 .
how can this string model for the cooper pair in the high - t@xmath0 superconductor be proved ?
the best way is the direct observation of the pt between cooper electrons .
evidently it should be connected with the detection of the ion lattice between cooper electrons . if we can establish that @xmath161 inside of the tube and @xmath162 outside of the pt then probably it will indicate the existence of the pt .
certainly , such measurements of the state of ion lattice are very difficult .
another indirect way is a test for the nonlinear potential @xmath163 .
the presence of such nonlinear term can lead to : ( a ) the nonlinear effects by propagation of the sound wave in the high - t@xmath0 superconductor along the layer ; ( b ) the different equilibrium states of ions in the lattice . such effects can be : ( a ) nonlinear scattering , propagation , absorption and so on for the sound waves ; ( b ) appearance of regions with different vacua . certainly , the presence of such kind of nonlinear effects is not the direct demonstration of the pt but it can indicate the important role of the strong phonon - phonon interaction in the high - t@xmath0 superconductor .
the basic goal of this paper is an assumption that the quantum solid theory with the strong nonlinear interaction of ions in the lattice can be similar to the qcd in which there is the strong gluon - gluon interaction .
such similarity can lead to the appearance of nonlocal objects in the quantum solid theory like the flux tube in the qcd .
it is possible that such a mechanism is realized in the high-@xmath157 superconductor by such a way that the phonons are confined into a nonlocal object ( tube ) located between two cooper electrons . in this paper
we have considered only one mechanism probably leading in emergence the pt .
the problem on existence of other alternative such mechanisms was not discussed in this paper and remains while open .
it is very important to note that our analysis shows that the green s function method in the superconductivity theory is a realization of an algorithm proposed by heisenberg for quantization of nonlinear spinor field .
we have shown that green s function method is such realization of discussed heisenberg idea .
we would like to emphasize that the nonperturbative heisenberg quantization method is much more powerful than a perturbative feynman diagram techniques .
finally , we can presuppose that in the quantum field theory * _ nonlinearity can lead to nonlocality _ * : nonlinear terms @xmath18 and @xmath19 in the qcd lead to the flux tube and probably such a mechanism in the high - t@xmath0 superconductivity leads to the pt .
w. heisenberg , nachr .
gttingen , n8 , 111(1953 ) ; w. heisenberg , zs .
, * 9a * , 292(1954 ) ; w. heisenberg , f. kortel und h. mtter , zs .
naturforsch . , * 10a * , 425(1955 ) ; w. heisenberg , zs .
fr phys . ,
* 144 * , 1(1956 ) ; p. askali and w. heisenberg , zs .
naturforsch . , * 12a * , 177(1957 ) ; w. heisenberg , nucl .
phys . , * 4 * , 532(1957 ) ; w. heisenberg , rev . mod .
phys . , * 29 * , 269(1957 ) . | it is assumed that in some sense the high - t@xmath0 superconductivity is similar to the quantum chromodynamics ( qcd ) .
this means that the phonons in high - t@xmath0 superconductor have the strong interaction between themselves like to gluons in the qcd . at the experimental level
this means that in high - t@xmath0 superconductor exists the nonlinear sound waves .
it is possible that the existence of the strong phonon - phonon interaction leads to the confinement of phonons into a phonon tube ( pt ) stretched between two cooper electrons like a hypothesized flux tube between quark and antiquark in the qcd .
the flux tube in the qcd brings to a very strong interaction between quark - antiquark , the similar situation can be in the high - t@xmath0 superconductor : the presence of the pt can essentially increase the binding energy for the cooper pair . in the first rough approximation
the pt can be approximated as a nonrelativistic string with cooper electrons at the ends . the bcs theory with such potential term
is considered .
it is shown that green s function method in the superconductivity theory is a realization of discussed heisenberg idea proposed by him for the quantization of nonlinear spinor field . a possible experimental testing for the string approximation of the cooper pair is offered . |
in the standard model ( sm ) spontaneous symmetry breaking is performed making use of tree level terms in the potential , but this is only an approximate method in the full quantum theory , since it disregards the quantum corrections coming from virtual processes at the loop level , which are particularly important when considering zero and finite temperature studies . in ref .
@xcite a first effort to extend the potential considered taking into account the one - loop level corrections , in order to study the phase transition between ssb and non - ssb scenarios for different models : the authors used explicitly the landau gauge for the scalar electrodynamics theory .
later on , jackiw @xcite made the calculation in an arbitrary gauge up to two - loop level for a set of scalars with an _
o(n ) _ internal symmetry . subsequently , arnold et al .
@xcite extended the result for the whole sm with a single higgs doublet including the thermal contributions . however , as far as we know
, no one has determined from first principles the one - loop level vacuum contributions in which more than one scalar field are present and have interaction terms , such as in the two higgs doublet model ( thdm ) .
although different models have been presented in which several non interacting scalar fields are present . in the cosmological arena the scalar sector effective
potential is useful to study the electroweak phase transition @xcite , on which both vacuum and finite - temperature contributions must be taken into account . in a recent paper @xcite
a first attempt was made to have an effective potential in which two scalar fields have coupling terms present in the potential .
unfortunately the way in which multiple scalar potentials have been constructed makes use of a generic expression coming from the single scalar one - loop contribution , without an explicit and detailed derivation . in the present work ,
we obtain an effective potential for the scalar sector of a generic thdm .
we first perform a detailed analysis for the single massless and massive field theories and then work out the thdm extension , we do this only for the scalar sector , since all other sectors are equal to those of the minimal standard model , for which the effective potential has been obtained even up to the three loop level in some scenarios @xcite .
the rest of our paper is structured as follows : in sec .
[ sec : derivation ] we present the general features used in the derivation of the effective potential . and then proceed to calculate the massless self interacting real scalar field potential in sec . [ sec : massless ] .
we then work out the case of a massive self interacting real scalar field potential in sec .
[ sec : massive ] , and in sec .
[ thdm ] we obtain the expression for the thdm effective potential . our conclusions are presented in sec .
[ conclusion ] .
for a single real massless self interactive scalar , whose lagrangian is given by @xmath0 the one loop level corrections to the potential are given by an infinite series of feynman diagrams ; the first four of which are shown in fig .
[ fig : loop - a - d ] [ fig : a ] [ fig : b ] [ fig : c ] + [ fig : d ] the first diagram contains only one vertex , so we only need one factor of @xmath1 ( the factor of @xmath2 cancels with the one of the propagator ) ; and it has a single propagator inside the loop .
the second diagram , has two vertices and two propagators . in general
, the @xmath3-th diagram contains @xmath3 vertices and @xmath3 propagators , each one corresponding to a side of the polygon formed in the center of each diagram .
so , from the first diagram we obtain : @xmath4 where @xmath5 represents the momentum of the particle going around the loop .
it is a virtual particle , so we actually have to integrate over all of its possible values , that is , from minus infinity to plus infinity ( for each component ) : @xmath6 the denominator @xmath7 is conventional , and the integration limits are usually left implicit when they are @xmath8 .
now we must add combinatoric factors : the diagram stays the same if we interchange the _ two _ external legs attached to the vertex , so we add a @xmath9 ; and it is also invariant under a reflection through a vertical axis passing through the center of the loop , so we have another @xmath9 : @xmath10 finally , because of the definition of the connected generating functional , we have a factor of @xmath2 and because of the form of the taylor series expansion , we must add a @xmath11 for each vertex in the diagram , so the final expression for the first diagram is @xmath12 where the @xmath13 is defined as the functional derivative of the connected generating functional with respect to the external source .
repeating the above reasoning , we arrive at the conclusion that the @xmath3-th diagram is represented by the expression @xmath14 summing all of them , we obtain @xmath15 performing a wick rotation , that is , making the change of variable @xmath16 of the time component of the momentum ; we eliminate the pre - factor of @xmath2 and leave a minus sign .
we also identify the infinite sum with the following taylor series expansion of the logarithm : @xmath17 and dropping the @xmath18 , @xmath19 in general , we set the second derivative of the potential , evaluated at @xmath20 , equal to the ( squared ) renormalized mass @xmath21 of the particle : @xmath22 and for the the particular case of a massless scalar ( which we will treat in the following section ) , we have @xmath23 another renormalization condition that is employed concerns the fourth derivative of the potential which is equated to the coupling constant @xmath1 @xmath24 where we evaluate @xmath13 at @xmath25 to avoid ir divergences .
@xmath25 is called the _ renormalization mass _ which is an arbitrary quantity .
the lagrangian for a real massless self interacting scalar field model is of the form @xmath26 where the terms containing @xmath27 , @xmath28 , and @xmath29 are the counterterms . hence the one - loop - level effective potential is @xmath30 the integral over the sum of the diagrams can be expressed as the logarithm @xmath31 the detailed derivation of the integral is found in appendix [ app : integral1 ] .
we can drop the first term because it vanishes in the limit of large @xmath32 ; also in this limit we have @xmath33 substituting the analytical solution of the logarithmic integral that we solve we now arrive at the following expression for the potential @xmath34 the counterterm coefficient @xmath28 is obtained through eq .
[ 2dv0 ] @xmath35 and the counterterm coefficient @xmath29 , from the condition given by eq .
[ 4dv ] @xmath36}{64\pi^2},\ ] ] after substituting them into the effective potential , and a little bit of work , we get @xmath37 .\ ] ] it is important to notice that although the coefficients of the counter terms @xmath28 and @xmath29 that we obtained are different to those obtained in ref .
@xcite , our final expression for the potential is the same .
the lagrangian for a massive self interacting scalar field theory is given by @xmath38 so we identify the potential as @xmath39 once again we are left with an integral over the sum of the one loop level self interacting contributions which we write as a logarithmic integral ( the detailed derivation is found in appendix [ app : integral2 ] ) @xmath40 from which we arrive at the effective one loop level potential for a massive interacting scalar field @xmath41 .\end{aligned}\ ] ] the methodology used in obtaining the analytical solution for this case will now be extended to calculate the potential of interacting scalar fields in the thdm .
because we are interested in the one - loop vacuum corrections of the scalar sector of the thdm ( a review on the thmd can be found in ref .
@xcite ) , we will only work with the contributions arising from the scalar potential @xmath42 , \end{aligned}\ ] ] where the higgs doublets can be written as @xmath43 but the physical fields of the charged sector are defined through the transformation @xmath44 while the imaginary terms are transformed via @xmath45 and the real neutral sector is obtained with the transformation @xmath46 through the previously defined transformations we are able to obtain all of the self interaction vertices between each of the fields @xmath47 , @xmath48 , @xmath49 , @xmath50 , @xmath51 and @xmath52 found in ref .
but since we choose to work in the unitary gauge , the goldstone bosons will be `` eaten '' by the su(2)@xmath53u(1 ) gauge fields . therefore , we will only be dealing with the physical fields .
furthermore , we only consider quartic vertices .
diagrams containing triple vertices such as those shown in fig .
[ fd : propagator ] give loop corrections to the propagator , but they are not relevant to our present work , so we do not consider them . the effective potential in integral form is given in appendix [ app : integral3 ] ,
from which we obtain the effective potential as a function of the physical fields : @xmath54 the first three terms corresponding to neutral higgs of the same type @xmath55 where @xmath56 , @xmath57 , @xmath58 are given by @xmath59 next we have the coupling of three fields of the same type ( @xmath60 ) with another one of a different kind ( @xmath61 ) given by @xmath62 during the calculation for this case , we set the one of the mass counter term coefficients @xmath63 , because if one includes all the mass counter terms you end up with a divergent logarithm .
the subsequent four terms of the effective potential which have two equal pairs @xmath60 , @xmath61 are given by @xmath64 finally we have the terms in which we have a single pair of scalars of the same kind and the remaining two are of a different type @xmath65 the last term in the potential involves the coupling of fields which are all of a different kind , such as @xmath66 , this type of coupling does not permit a polygon diagram whose series is constructible .
we calculated the contributions of the thdm effective potential that come from one - loop diagrams with quartic coupling vertices which are dimensionally consistent and renormalizable .
we hope this analytical scheme will be of use to model builders interested in theories that must take in to consideration vacuum corrections to two or more scalars .
we acknowledge support from conacyt ( mxico ) , and appreciate the advice and discussions from gerardo f. torres del castillo and o. meza - aldama .
@xmath74\\ & = & \frac{\lambda^4}{64\pi^2 } \ln \left ( 1 + \frac{\lambda\phi^2}{2\lambda^2}\right ) + \frac{1}{64\pi^2 } \int_0^\lambda \rho^4\frac{2\rho^2}{2\rho^2 + \lambda\phi^2 } \frac{\lambda\phi^2}{\rho^3 } d\rho , \end{aligned}\ ] ] in this case we have that the sum of 1-loop diagrams is given by the following integral : @xmath89 we now make the change of variable @xmath90 @xmath91 @xmath92 @xmath93 .
then dropping the @xmath18 , @xmath94 let s now solve the rightmost integral ( redefining some @xmath95-independent quantities as @xmath96 and @xmath97 ) : @xmath98 setting @xmath99 and @xmath100 , to get @xmath101 we integrate : @xmath102 we now use the following known integral @xmath103 where in our case @xmath104 , @xmath105 , and @xmath106 ; thus @xmath107 dx .\end{aligned}\ ] ] for each term on the right we have @xmath108 @xmath109 putting both together , and doing some algebra , @xmath110 finally , placing a cutoff at @xmath111 , or equivalently @xmath112 , @xmath113 + x \\ & & + \left .
\frac{3}{2}(\lambda\phi^2 + 4\mu^2)\right]\big|^{2\lambda^2}_0 .\end{aligned}\ ] ] we neglect the first term since when @xmath114 it tends to zero . then
, evaluating , we re left only with @xmath115 \\ & & + \frac{\lambda\phi^2\lambda^2}{128\pi^2 } .\end{aligned}\ ] ] in the limit of large @xmath32 , we get @xmath116 .\end{aligned}\ ] ] there are essentially three types of vertices involved in the calculation : those with four equal fields , with two pair of equal fields , and with only one pair of equal fields .
we only consider the case in which only one particle goes around the loop .
thus , when we analyze a vertex with more than one particle interacting , we write only the propagators of the particle that is `` repeated '' . in the first case , the integral form of the potential is @xmath117 where @xmath1 represents genericallly the self - coupling constant and @xmath118 is the mass of the particle @xmath119 , and we ve omitted the @xmath18 . in the second case , @xmath120 where there s an obvious symmetry between @xmath60 and @xmath61 .
finally , for the last case , @xmath121 | we obtain a closed form effective potential at the one - loop level of a two higgs doublet model . through the loop expansion we reproduce the expression presented by weinberg and coleman , showing explicitly every step involved in the calculation .
the formalism is then extended to include interaction terms between two scalar doublets and a final expression of the one - loop level contributions is presented .
+ + _ keywords : _ electroweak radiative corrections , extensions of electroweak higgs sector + + pacs : 12.15.lk ; 12.60.fr |
in recent years , single - layer graphene has received much attention due to its two dimensionality , dirac - like energy spectrum and outstanding spin - coherence properties.@xcite specifically , the hyperfine interaction and spin - orbit coupling ( soc ) are considerably weak,@xcite suggesting a long spin relaxation time ( srt ) in single - layer graphene .
this long srt makes single - layer graphene a promising candidate for spintronic application .
so far , a number of experiments on spin relaxation have been carried out , with the srts being orders of magnitude shorter than theoretical prediction.@xcite this pronounced discrepancy has provoked many investigations on the extrinsic effects of spin relaxation such as adatoms,@xcite curvature,@xcite substrate effects@xcite and contacts.@xcite consisting of two layer graphene , bilayer graphene ( blg ) also possesses good spin - coherence properties.@xcite additionally , in contrast to single - layer graphene , blg has a tunable gap caused by an out - of - plane electric field,@xcite which can be used to turn on or off the electronic transport .
in addition , in the absence of space inversion symmetry , the intrinsic soc in the pseudospin space of blg provides an out - of - plane effective magnetic field near the dirac points whereas the intrinsic soc of single - layer graphene only induces a shift of the energy spectrum.@xcite moreover , the electron - phonon coupling becomes more complex in blg due to the existence of more phonon modes.@xcite so far , some efforts have been made to investigate the intravalley electron - phonon coupling.@xcite specifically , the intravalley electron - acoustic ( ac ) phonon coupling has been calculated by viljas and heikkil@xcite with the continuum theory and also by borysenko _
et al._@xcite from first principles . to explore the intravalley electron - optical ( op ) phonon coupling , the tight - binding model , density functional theory and first principles approach are employed by viljas and heikkil,@xcite cappelluti _
et al._@xcite and borysenko _
et al._,@xcite respectively .
however , the intervalley electron - phonon coupling , which has been shown to play an important role in spin relaxation in rippled single - layer graphene,@xcite has not been reported in the previous literature .
all these intriguing features make blg an attractive material for investigation .
very recently , some attention has been devoted to the spin relaxation in blg.@xcite experimentally , the srts of blg on sio@xmath0 substrate@xcite or in freely suspended blg@xcite are reported to be of the order of @xmath1-@xmath2ns . both yang _ et al._@xcite and han and kawakami@xcite measured the temperature and carrier dependences of the srt .
et al._@xcite showed that the srt is weakly dependent on temperature .
however , han and kawakami@xcite reported a marked decrease of the srt with the increase of the temperature .
as for the carrier density dependence , the decrease of the srt with increasing carrier density at low temperature is reported in both experiments . however , at room temperature , yang _
et al._@xcite observed an increase of the srt with increasing density whereas the srt measured by han and kawakami@xcite exhibits a marginal density dependence .
in addition , yang _
et al._@xcite also showed that the srt scales inversely with the mobility at both room and low temperatures .
in contrast to the above two experiments based on the micromechanically exfoliated blg samples , avsar _
et al._@xcite carried out the experiment in chemical vapor deposition grown blg . the carrier density and temperature dependences of the srt measured are comparable to the ones by yang _
et al._@xcite both quantitatively and qualitatively .
motivated by these experiments , diez and burkard@xcite calculated the electron spin relaxation due to the dyakonov - perel ( dp ) mechanism@xcite with only the long - range electron - impurity scattering included in blg .
they claimed to obtain a good agreement with the experimental data of yang _
et al._.@xcite it is noted that the effective soc ( referred to as the spin - orbit field in ref . ) used in their calculation is induced by an interlayer bias voltage together with the intrinsic soc in the pseudospin space.@xcite however , the intrinsic soc in the pseudospin space becomes incomplete in the presence of the interlayer bias since the interlayer bias breaks the space inversion symmetry and leads to additional extrinsic soc terms in the pseudospin space.@xcite also even with only the intrinsic soc in the pseudospin space in their calculation , they overlooked the out - of - plane component of the effective soc .
as will be shown later , the leading term of this component leads to a zeeman - like term in the two valleys and makes a dominant contribution to the in - plane spin relaxation .
moreover , they neglected the electron - phonon and electron - electron coulomb scatterings .
in fact , the electron - electron coulomb scattering has been demonstrated to be very important for the spin relaxation in semiconductor systems@xcite and also single - layer graphene at high temperature.@xcite the electron - phonon scattering has also been shown to play an important role in spin relaxation in both semiconductors@xcite and single - layer graphene@xcite at high temperature .
therefore , both have to be included in studying the spin relaxation . in the present work , with the electron - electron coulomb , ( both the intra- and inter - valley ) electron - phonon , possible short - range@xcite as well as long - range electron - impurity scatterings explicitly included ,
we investigate the electron spin relaxation due to the dp mechanism in blg with only the lowest conduction band being relevant by the kinetic spin bloch equation ( ksbe ) approach.@xcite the soc of the lowest conduction band near the dirac points @xmath3 is constructed from the symmetry group @xmath4 in the presence of an out - of - plane electric field.@xcite its three components
@xmath5 @xmath6 can be written as ( see appendix [ appa ] ) @xmath7,\label{eq1}\\ \omega_y^{\mu}({\bf k})&=&-\alpha_1(k)\cos\theta_{\bf k}+ \mu[\alpha_2(k)\cos 2\theta_{\bf k}\nonumber\\ & & \mbox{}-\alpha_3(k)\cos 4\theta_{\bf k}],\label{eq2}\\ \omega_z^{\mu}({\bf k})&=&\mu\beta_1(k)+\beta_2(k)\cos 3\theta_{\bf k},\label{eq3}\end{aligned}\ ] ] with @xmath8 standing for k(k@xmath9 ) valleys and @xmath10 and @xmath11 representing the magnitude and the polar angle of the momentum @xmath12 relative to k(k@xmath9 ) points , respectively .
the coefficients @xmath13 ( @xmath14-@xmath15 ) and @xmath16 are given in appendix [ appa ] by fitting to the numerical results calculated according to the latest report by konschuh _
et al._@xcite from first principles with both the intrinsic and extrinsic soc terms in the pseudospin space included .
the out - of - plane component @xmath17 [ eq .
( [ eq3 ] ) ] , which has not been reported in the previous literature , is induced by the intrinsic soc in the pseudospin space in the absence of space inversion symmetry , whereas the in - plane component is contributed by both the intrinsic and extrinsic socs in the pseudospin space .
this is very different from the case in single - layer graphene where the intrinsic soc in the pseudospin space only induces a shift of the energy spectrum in the conduction band.@xcite we find that the magnitudes of both the out - of- and in - plane components decrease with the increase of @xmath10 at large momentum , indicating a suppression of the inhomogeneous broadening@xcite by increasing momentum.@xcite this is very different from the case in both semiconductors and single - layer graphene.@xcite additionally , the leading term of the out - of - plane component , i.e. , @xmath18 , serves as a zeeman - like term with opposite effective magnetic fields perpendicular to blg plane in the two valleys , similar to the case in rippled single - layer graphene.@xcite this zeeman - like term , together with the intervalley scattering , opens a spin relaxation channel .
this spin relaxation channel suppresses the in - plane srt _ significantly _ at high temperature by the intervalley electron - phonon scattering and also at low temperature by the short - range scattering.@xcite in addition , we find that although such zeeman - like terms also exist in the in - plane components of the soc , their contributions to the out - of - plane spin relaxation are marginal since these terms are of the high order of the momentum .
this leads to a strong anisotropy between the in - plane and out - of - plane spin relaxations .
we also find that the in - plane srt shows a marked nonmonotonic dependence on temperature with a minimum down to several hundred picoseconds in the absence of short - range scatterers .
this minimum is comparable to the experimental data .
the nonmonotonic behavior results from the crossover between the weak and strong intervalley electron - phonon scattering .
moreover , we predict a peak in the density dependence at low temperature , which is very different from the one in semiconductors.@xcite as for high temperature , the srt shows a monotonic increase with increasing density .
we also find that the in - plane srt decreases rapidly with increasing initial spin polarization at low temperature , which is very different from the previous studies in both semiconductors@xcite and single - layer graphene.@xcite finally , we report detailed comparisons of the temperature and electron density dependences with the existing experiments.@xcite this paper is organized as follows . in sec .
ii , we present our model and construct the ksbes .
the main results are given in sec .
iii . in sec .
iiia , we investigate the temperature , electron density and also initial spin polarization dependences of the electron spin relaxation in the absence of short - range scatterers .
the anisotropy of the spin relaxation is also addressed in this part . in sec .
iiib , we show detailed comparisons of the temperature and electron density dependences of the spin relaxation with the existing experiments with the inclusion of short - range scatterers .
we summarize in sec . iv .
we start our investigation from the ab - stacked blg on sio@xmath0 substrate .
the spinless @xmath19-band structure can be described by the effective @xmath20 hamiltonian:@xcite @xmath21 in the on - site orbital bloch basis @xmath22 , @xmath23 , @xmath24 and @xmath25 , where a@xmath26 refers to the a sublattice in the lower layer , b@xmath0 to the b sublattice in the upper layer , etc . ,
and @xmath27 represents the two - dimensional wave vector counted from the @xmath28 point . here ,
@xmath29 and @xmath30 describe the nearest - neighbor intralayer and interlayer hoppings whereas @xmath31 and @xmath32 stand for the indirect hoppings between two layers .
@xmath33 represents the asymmetry in the energy shift of the on - site energy due to the interlayer hopping .
the potential @xmath34 with @xmath35 , @xmath36 and @xmath37 representing the electron charge @xmath38 , the out - of - plane electric field and the effective electrostatic bilayer distance , respectively . @xmath39 $ ] is the nearest - neighbor structural function of the graphene hexagonal lattice with @xmath40 denoting the lattice constant . near the dirac points , to the first order of the momentum @xmath41 originated from the k(k@xmath9 ) points , @xmath42 .
then the effective @xmath20 hamiltonian near the dirac points can be written as @xmath43 by exactly diagonalizing this hamiltonian , one obtains the eigenvalue @xmath44 and the eigenfunction @xmath45 with @xmath46 being the band index including two conduction bands and two valence ones . in our calculation , we concentrate on the heavily @xmath47-doped case with only the lowest conduction band being relevant whereas other bands are separated by large energy intervals . for convenience , the eigenvalue and the corresponding eigenfunction of the lowest conduction band at @xmath12
are denoted by @xmath48 and @xmath49 , respectively . by incorporating the spin degree of freedom ,
the effective hamiltonian of the lowest conduction band is given by @xmath50 here , @xmath51 are the pauli matrices .
@xmath52 stands for the effective magnetic field of the soc , which is constructed from the symmetry analysis with the parameters obtained approximately by fitting to numerical results based on the report by konschuh _
et al._.@xcite the analytical form of the soc with three components @xmath5 @xmath6 is shown in eqs .
( [ eq1]-[eq3 ] ) and the explicit numerical calculation is given in appendix [ appa ] .
it is noted that at large momentum , the analytical result agrees with the numerical one fairly well ( see appendix [ appa ] ) .
we then construct the microscopic ksbes@xcite to study the electron spin relaxation in blg .
the ksbes read@xcite @xmath53 where @xmath54 represent the density matrices of electrons with the diagonal terms @xmath55 describing the distribution functions and the off - diagonal ones @xmath56 denoting the spin coherence .
the coherent term is given by @xmath57 , \label{eq8}\end{aligned}\ ] ] where @xmath58\equiv ab - ba$ ] is the commutator ; @xmath59 represents the coulomb hartree - fock ( hf ) term.@xcite @xmath60 denotes the screened coulomb potential with its form given in appendix [ appb ] .
the form factor @xmath61 .
@xmath62 are the scattering terms including the electron - electron coulomb ( @xmath63 ) , long - range electron - impurity ( @xmath64 ) , intravalley electron - ac phonon ( @xmath65 ) , electron - op phonon ( @xmath66 ) , electron - remote - interfacial ( ri ) phonon ( @xmath67 ) and especially the intervalley electron - phonon scattering including the electron-@xmath68 phonon ( @xmath69 ) , electron-@xmath70 phonon ( @xmath71 ) , electron-@xmath72 phonon ( @xmath73 ) and electron-@xmath74 phonon ( @xmath75 ) scatterings .
their detailed expressions can be found in ref .
.@xcite here , @xmath68 , @xmath70 and @xmath72 are the in - plane phonon modes corresponding to the representations @xmath76 , @xmath77 and @xmath78 of group @xmath79 , respectively whereas @xmath74 stands for the out - of - plane phonon mode.@xcite it is noted that the intravalley electron - phonon scattering has been reported in the previous literature as mentioned in the introduction.@xcite however , to the best of our knowledge , there is no specific report to the intervalley electron - phonon scattering . in this work
, we derive the intervalley electron - phonon scattering matrix elements using the tight - binding model according to the arxiv version of ref . .
the scattering matrix elements are given by @xmath80,\label{eq9}\\ & & \hspace{-0.4cm}|m^{\rm ka_2^{\prime}}_{{\mu{\bf k}},{\mu^{\prime}{\bf k}^{\prime}}}|^2=\frac{3a^2{\gamma_4^{\prime}}^2}{2\rho l_4 ^ 2 } \frac{\hbar^2}{2\omega_{\rm ka_2^{\prime}}}\delta_{\mu^{\prime},-\mu } \big[|{\psi^{\mu}_{{\bf
k}}}^{\dagger}\gamma^3_{\rm d}\gamma^5_{\rm d}\psi^{\mu^{\prime}}_{{\bf k}^{\prime}}|^2\nonumber\\ & & \mbox{}\hspace{0.1cm}+|{\psi^{\mu}_{{\bf k}}}^{\dagger}\gamma^0_{\rm d}\psi^{\mu^{\prime}}_{{\bf k}^{\prime}}|^2\big],\label{eq10}\\ & & \hspace{-0.4cm}|m^{\rm ke^{\prime}}_{{\mu{\bf k}},{\mu^{\prime}{\bf k}^{\prime}}}|^2=\frac{3a^2}{\rho l_4 ^ 2 } \frac{\hbar^2}{4\omega_{\rm ke^{\prime}}}\delta_{\mu^{\prime},-\mu } \big[|{\psi^{\mu}_{{\bf k}}}^{\dagger}(\gamma_3^{\prime}\gamma^1_{\rm
d}\gamma^5_{\rm d}-i\gamma_3^{\prime}\gamma^2_{\rm d}\nonumber\\ & & \mbox{}\hspace{0.1cm}-\gamma_4^{\prime}\gamma^0_{\rm d})\psi^{\mu^{\prime}}_{{\bf k}^{\prime}}|^2+|{\psi^{\mu}_{{\bf k}}}^{\dagger } ( i\gamma_3^{\prime}\gamma^2_{\rm d}-\gamma_3^{\prime}\gamma^1_{\rm d}\gamma^5_{\rm d}-\gamma_4^{\prime}\gamma^0_{\rm d } ) \psi^{\mu^{\prime}}_{{\bf k}^{\prime}}|^2\nonumber\\ & & \mbox{}\hspace{0.1cm}+4|{\psi^{\mu}_{{\bf k}}}^{\dagger}\gamma_4^{\prime}\gamma^3_{\rm d}\gamma^5_{\rm d } \psi^{\mu^{\prime}}_{{\bf k}^{\prime}}|^2\big],\label{eq11}\\ & & \hspace{-0.4cm}|m^{\rm kz}_{{\mu{\bf k}},{\mu^{\prime}{\bf k}^{\prime}}}|^2=\frac{{\gamma_1^{\prime}}^2}{\rho } \frac{\hbar^2}{2\omega_{\rm kz}}\delta_{\mu^{\prime},-\mu } \label{eq12}\end{aligned}\ ] ] where @xmath81 is the mass density of the blg;@xcite @xmath82 , @xmath83 , @xmath84 and @xmath85 are the energy spectra of @xmath68 , @xmath70 , @xmath72 , and @xmath74 phonon modes , respectively;@xcite @xmath86 ( @xmath87 ) is the derivative of @xmath88 with respect to the corresponding hopping bond length;@xcite @xmath89 is the bond length of the interlayer hopping @xmath32 ; @xmath90 ( @xmath91-@xmath15 , @xmath92 ) , @xmath93 and @xmath94 are @xmath20 dirac matrices given in appendix [ appc].@xcite we will show that the above intervalley electron - phonon scattering plays a significant role in the in - plane spin relaxation in our investigation .
the remaining scattering matrix elements such as @xmath63 , @xmath64 , and @xmath95 ( @xmath96 ) are given in appendix [ appb ] .
it is noted that we also include the short - range scattering in our calculation with the scattering term laid out later in this paper .
.parameters used in the calculation ( from ref . unless otherwise specified ) .
[ cols= " < , < , < , < " , ]
by numerically solving the ksbes , one obtains the time evolution of spin polarization along direction @xmath97 @xmath98/n_e$ ] with @xmath99 being the electron density .
then the srt can be determined from the slope of the envelope of the spin polarization @xmath100.@xcite the initial spin polarization @xmath101 is set to be @xmath102 % and the spin - polarization direction @xmath97 is along the @xmath103-axis unless otherwise specified .
it is noted that all the parameters used in our calculation are listed in table i. it is also noted that the soc is calculated both fully numerically following the approach by konschuh _
et al._@xcite and analytically by using eqs .
( [ eq1]-[eq3 ] ) .
the srt calculated with the analytical form of the soc and with the explicit numerical one agrees fairly well with each other . in our calculation , the out - of - plane electric field is chosen to be a typical value @xmath104v / nm@xcite and the long - range impurity density is taken to be @xmath105@xmath106 .
the corresponding mobility in our investigation is of the order of @xmath107@xmath108/(vs ) , which is about @xmath109-@xmath110 orders of magnitude larger than those reported in the existing experiments in blg on sio@xmath0 substrate@xcite and comparable to the ones in the experiments in freely suspended blg.@xcite it is noted that the short - range scattering is not included in this subsection .
-axis @xmath111 ( @xmath112 ) and the calculation without the long - range electron - impurity ( @xmath113 ) , electron - ac - phonon ( @xmath114 ) , electron - ri - phonon ( @xmath115 ) , electron - electron coulomb ( @xmath116 ) or intervalley electron - phonon ( @xmath117 ) scattering as function of temperature @xmath118 .
it is noted that when @xmath119 k , only the total srt and the calculation without the intervalley electron - phonon scattering are shown since the remaining ones almost coincide with the total srt . in the calculation , the electron density @xmath120@xmath106.,width=321 ]
we first investigate the temperature dependence of the spin relaxation . in fig .
[ fig1 ] , the srt along the @xmath103-axis @xmath111 is plotted as function of temperature @xmath118 .
it is seen that @xmath111 changes little with the temperature at low temperature ( i.e. , @xmath121k ) . however , as @xmath118 becomes larger , the srt decreases dramatically with increasing temperature .
in particular , the srt reaches down to several hundred picoseconds at @xmath122k , which is comparable to the experimental values.@xcite a detailed comparison with the existing experiments will be discussed in sec .
when @xmath118 further increases , the srt begins to increase with @xmath118 . to understand the behavior of the spin relaxation , we show in fig .
[ fig1 ] the temperature dependence of the srt calculated with the electron - electron coulomb , electron - ri - phonon , electron - ac - phonon , long - range electron - impurity , or intervalley electron - phonon scattering removed , separately .
we find that at low temperature , the intervalley electron - phonon scattering is negligible .
therefore , spins relax independently in the two valleys and the srt is solely determined by the intravalley spin relaxation channel .
specifically , the long - range electron - impurity scattering is dominant , which leads to a weak temperature dependence of the srt.@xcite with the increase of the temperature , the contribution of the long - range electron - impurity scattering becomes marginal whereas the intervalley electron - phonon scattering becomes important .
this intervalley scattering opens another spin relaxation channel together with the zeeman - like term with opposite effective magnetic fields in the two valleys , which is similar to the case in rippled single - layer graphene.@xcite this zeeman - like term is given by @xmath18 [ eq .
( [ eq3 ] ) ] approximately
. then the spin relaxation due to the above intervalley spin relaxation channel can be approximated by the rate equations@xcite @xmath123 where @xmath124 , @xmath125\hat{{\bf z}}$ ] and @xmath126 represent the in - plane spin vectors in each valley , spin precession vector along the @xmath127-axis and the intervally electron - phonon scattering time , respectively . according to the report by zhang _
et al._,@xcite the in - plane srt is given by @xmath128 here , @xmath129 stands for the srt along the @xmath130-axis .
since the system we investigate is always in the degenerate regime , we take @xmath131 approximately with @xmath132 representing the fermi wave vector . in our calculation
, the srt along the @xmath103-axis is solely determined by the intervalley electron - phonon scattering at @xmath133-@xmath134k . at @xmath135k , @xmath136 , i.e.
, the intervalley electron - phonon scattering is in the weak scattering limit . as a result , the srt @xmath137 according to eq .
( [ eq14 ] ) , which decreases with the enhancement of the intervalley scattering as the temperature increases .
nevertheless , at @xmath138k , @xmath139 , i.e. , the intervalley scattering is in the strong scattering limit
. then the srt @xmath140 $ ] from eq .
( [ eq14 ] ) , which increases with increasing of @xmath118 .
therefore , the srt first decreases then increases with increasing temperature when the intervalley scattering changes from the weak to strong scattering limit .
the crossover between the weak and strong intervalley scattering limit is determined by @xmath141.@xcite at this crossover point , @xmath142 here @xmath143ps , very close to the value @xmath144ps shown in the figure .
it is noted that the electron - electron coulomb scattering is unimportant in the whole temperature regime
. then we turn to study the electron - density dependence . in fig .
[ fig2 ] , the srt along the @xmath103-axis is plotted against the electron density at @xmath145 @xmath146k .
one observes a peak in the electron - density dependence of the srt at @xmath135k whereas the srt increases with the increase of the electron density when @xmath147k .
we first focus on the case at @xmath135k and show the srt with the intervalley electron - phonon scattering excluded in fig .
it is seen that when the electron density is low , the intervalley electron - phonon scattering is marginal .
the srt is then determined by the intravalley spin relaxation channel and hence increases with the suppression of the inhomogeneous broadening@xcite as the electron density increases.@xcite the suppression of the inhomogeneous broadening can be understood from the decrease of the spin splitting with increasing momentum at high momentum [ see eqs .
( [ eq1]-[eq3 ] ) ] , as shown in fig .
[ fig9 ] in appendix [ appa ] .
this behavior is very different from the case in both semiconductors and single - layer graphene.@xcite when the electron density further increases , the intervalley electron - phonon scattering becomes important .
this intervalley scattering is always in the weak scattering limit with @xmath148 and therefore @xmath137 according to eq .
( [ eq14 ] ) . as a result ,
the srt decreases with the increase of the intervalley scattering and hence the increase of the electron density.@xcite therefore , in the whole electron - density regime , a peak is observed .
it is noted that this peak is very different from the one predicted by jiang and wu@xcite in semiconductors where the peak is attributed to the crossover from the nondegenerate - to - degenerate limit . as for the case at @xmath147k ,
the srt is always dominated by the intervalley electron - phonon scattering as shown in fig .
[ fig2 ] by comparing the srt calculated with and without the intervalley electron - phonon scattering .
since this intervalley scattering is in the strong scattering limit with @xmath149 , the srt @xmath140 $ ] from eq .
( [ eq14 ] )
. this srt increases with increasing electron density since both @xmath150 and the intervalley scattering time @xmath151 decrease with the increase of the electron density.@xcite
-axis @xmath111 as function of the electron density @xmath99 with @xmath112 ( @xmath152 ) and without @xmath115 ( @xmath114 ) the intervalley electron - phonon scattering at @xmath153k.,width=321 ] -axis @xmath111 @xmath114 ( @xmath154 ) and the calculation without the coulomb hf term @xmath152 ( @xmath112 ) or the intervalley electron - phonon scattering @xmath113 ( @xmath115 ) as function of the initial spin polarization @xmath155 at @xmath153k . in the calculation ,
the electron density @xmath156@xmath106.,width=321 ] the initial spin polarization dependences of the srt along the @xmath103-axis at @xmath145 and @xmath157k are shown in fig .
it is seen that the srt decreases rapidly with the increase of initial spin polarization @xmath155 at @xmath135k whereas the srt shows a slight increase with increasing @xmath155 at @xmath147k .
the rapid decrease of the srt at @xmath135k is very different from the previous studies in both semiconductors@xcite and single - layer graphene@xcite where the srt increases significantly with increasing initial spin polarization . in the previous studies , the increase of the srt originates from the contribution of the coulomb hf term.@xcite this term serves as an effective magnetic field along the direction of the spin polarization , which blocks the spin precession induced by the dresselhaus@xcite and/or rashba@xcite socs and slows down the spin relaxation .
however , the contribution of the coulomb hf term is marginal to the spin relaxation in our investigation , which is shown in fig .
[ fig3 ] by comparing the the srt calculated with and without the coulomb hf term . here
, the srt is determined by the intervalley electron - phonon scattering , as shown in fig .
[ fig3 ] by comparing the calculation with and without the intervalley electron - phonon scattering .
the intervalley electron - phonon scattering transfers electrons between the two valleys . the transferred electrons experience opposite effective magnetic fields induced by the zeeman - like term in the two valleys , which are not affected by the effective magnetic field induced by the coulomb hf term .
moreover , the intervalley electron - phonon scattering is in the weak scattering limit at low temperature and hence the srt @xmath137 [ see eq .
( [ eq14 ] ) ] . the intervalley scattering time @xmath151 decreases significantly with the increase of the initial spin polarization as the density of states increases with the polarization .
therefore , the srt shows a rapid decrease with the initial spin polarization at @xmath135k . as for the case of @xmath147k ,
the srt is also determined by the intervalley electron - phonon scattering in comparison with the case of @xmath135k . however ,
in contrast to the case of @xmath135k , the intervalley scattering is in the strong scattering limit .
this makes the srt present an opposite trend to the case of @xmath135k . a mild increase of the srt results from the insensitivity of @xmath151 to the initial spin polarization at @xmath147k .
@xmath158-axis @xmath111 @xmath159 as function of temperature @xmath118 . in the calculation , the electron density @xmath120@xmath106.,width=321 ]
we also address the anisotropy of the spin relaxation with respect to the spin polarization direction .
due to the existence of the out - of - plane effective magnetic field [ see eq . ( [ eq3 ] ) ] , the in - plane srts are identical.@xcite for comparison , we plot the temperature dependence of the srt along the @xmath103 ( @xmath127)-axis as the red curve with crosses ( blue curve with open squares ) in fig .
it is seen that in comparison with the srt along the @xmath127-axis , the one along the @xmath103-axis is comparable at low temperature and orders of magnitude smaller at high temperature .
as mentioned previously , the srt along the @xmath103-axis is strongly suppressed by the intervalley spin relaxation channel induced by the intervalley electron - phonon scattering and the zeeman - like term in the two valleys at high temperature . however , the contribution of this spin relaxation channel to the srt along the @xmath127-axis is marginal [ see eqs .
( [ eq1]-[eq2 ] ) ] .
therefore , a strong anisotropy between the out - of- and in - plane spin relaxations is observed at high temperature .
as function of temperature @xmath118 .
@xmath116 ( @xmath114 ) corresponds to the case of the gate voltage v@xmath160v in the experiment of han and kawakami;@xcite @xmath112 ( @xmath115 ) stands for the case of electron density @xmath161@xmath106 in the experiment of avsar _
et al._;@xcite @xmath154 ( @xmath117 ) represents the case of @xmath162@xmath106 in the experiment of yang _
et al._;@xcite @xmath163 denotes the case of @xmath164@xmath106 in the experiment of neumann _
et al._.@xcite ( a ) red solid ( orange chain ) curve stands for our calculation corresponding to the case of v@xmath160v in the absence of the short - range scatterers . in the inset , we show the temperature dependence of the long - range impurity density @xmath165 corresponding to the case of v@xmath166v .
( b ) grey dashed and skyblue double - dotted chain curves are obtained from our calculations corresponding to the case of v@xmath166v by including the short - range scatterers , with a fixed short - range impurity density @xmath167@xmath106 and a tunable @xmath168 ( shown in the inset ) , respectively .
we also show the dependence of @xmath169 on @xmath168 ( with the scale on the top of the frame ) at @xmath170k with other conditions the same as the case of v@xmath166v .
the short - range potential strength @xmath171mevm@xmath172.,title="fig:",width=332 ] as function of temperature @xmath118 .
@xmath116 ( @xmath114 ) corresponds to the case of the gate voltage v@xmath160v in the experiment of han and kawakami;@xcite @xmath112 ( @xmath115 ) stands for the case of electron density @xmath161@xmath106 in the experiment of avsar _
et al._;@xcite @xmath154 ( @xmath117 ) represents the case of @xmath162@xmath106 in the experiment of yang _
et al._;@xcite @xmath163 denotes the case of @xmath164@xmath106 in the experiment of neumann _
et al._.@xcite ( a ) red solid ( orange chain ) curve stands for our calculation corresponding to the case of v@xmath160v in the absence of the short - range scatterers . in the inset
, we show the temperature dependence of the long - range impurity density @xmath165 corresponding to the case of v@xmath166v .
( b ) grey dashed and skyblue double - dotted chain curves are obtained from our calculations corresponding to the case of v@xmath166v by including the short - range scatterers , with a fixed short - range impurity density @xmath167@xmath106 and a tunable @xmath168 ( shown in the inset ) , respectively .
we also show the dependence of @xmath169 on @xmath168 ( with the scale on the top of the frame ) at @xmath170k with other conditions the same as the case of v@xmath166v .
the short - range potential strength @xmath171mevm@xmath172.,title="fig:",width=359 ] as mentioned in the introduction , experiments on spin relaxation have been carried out in blg on sio@xmath0 substrate@xcite or in freely suspended blg@xcite very recently by different groups .
the in - plane srts in these experiments are of the order of @xmath1-@xmath2ns , which are comparable to our theoretical results at high temperature ( several hundred picoseconds ) .
this motivates us to carry out a detailed comparison with the existing experiments . in these experiments , the gate voltage , temperature and electron density@xcite ( tuned by the gate voltage )
are given explicitly .
however , some parameters such as the out - of - plane electric field and the impurity density are not available in the experiments . in our calculation
, the out - of - plane electric field is taken to be the ratio of the gate voltage to the thickness of the substrate approximately.@xcite the impurity density is obtained by fitting to the mobility or the spin diffusion coefficient given in the experiments.@xcite in addition , we assume the initial spin polarization along the @xmath103-axis without loss of generality since the in - plane spin relaxations are identical as pointed out previously .
we first compare the temperature dependence of the srts obtained from the experiments and calculation from our model . in these experiments ,
the mobilities are reported of the order of @xmath173-@xmath107@xmath108/(vs ) .
the srts from different groups are comparable as shown in fig .
however , the srt from the experiment of han and kawakami@xcite decreases with increasing temperature whereas the ones from the experiments of yang _
et al._@xcite and avsar _
et al._@xcite show a marginal temperature dependence .
in addition , han and kawakami@xcite reported the longest srt among these experiments , indicating that their experimental data may contain less extrinsic effects . motivated by this , we show our comparison with the experiment of han and kawakami@xcite in fig .
[ fig5](a ) .
the experimental data with the gate voltage v@xmath160v corresponding to the electron density @xmath174@xmath106 , is labeled as @xmath175 . by fitting to the spin diffusion coefficient given in the experiment
, we obtain the long - range impurity density @xmath165 shown in the inset as the curve with @xmath113 .
in contrast to the long - range impurity scattering , the contribution of other scatterings to the spin diffusion coefficient ( mobility ) is marginal .
then we calculate the srt corresponding to v@xmath160v shown as the red solid ( orange chain ) curve .
a crossover is observed in the temperature dependence of the srt , similar to the one shown in sec .
it is seen that the srt from our calculation is comparable to the experimental data at high temperature .
this suggests that the intervalley spin relaxation channel induced by the intervalley electron - phonon scattering and the zeeman - like term in the two valleys addressed in sec .
iiia , plays an important role at high temperature .
nevertheless , the srts are still larger than the experimental data with an observable difference at high temperature .
additionally , at low temperature , our results are orders of magnitude larger than the experimental ones since the intervalley electron - phonon scattering is marginal at low temperature and the srt is determined by the long - range electron - impurity scattering in our calculation . due to the existence of all these discrepancies , one has to take into account extrinsic effects such as adatoms,@xcite curvature,@xcite substrate effects@xcite and also contacts.@xcite in the above experiments on blg , the spin polarized electrons are all injected from ferromagnetic contacts.@xcite as reported , the ferromagnetic electrodes generate stray fields , which cause spin precession around the contact.@xcite as a result , the stray fields may have a strong influence on the observed experimental results .
additionally , electron scattering by the tunneling contacts may also be responsible for the available experimental data.@xcite in addition to the above extrinsic effects , the short - range scatterers may also be an extrinsic source.@xcite here , we discuss the contribution of short - range scatterers@xcite in detail .
the scattering term induced by the short - range scatterers is given by @xmath176 with @xmath168 and @xmath177 denoting the short - range impurity density and the constant short - range potential strength , respectively.@xcite this scattering term contributes to both the intra- and inter - valley scatterings .
the intervalley scattering opens an intervalley spin relaxation channel together with the zeeman - like term in the two valleys [ see eq .
( [ eq3 ] ) ] , similar to the intervalley electron - phonon scattering .
however , in contrast to the intervalley electron - phonon scattering , the intervalley scattering induced by the short - range scatterers is _ insensitive _ to the temperature and may play an important role in the in - plane spin relaxation especially at low temperature . here ,
we choose a fixed short - range impurity density @xmath167@xmath106 and potential strength @xmath171mevm@xmath172 .
it is noted that although with this additional short - range scattering included , the mobility is still dominated by the long - range electron - impurity scattering since the contribution of the long - range electron - impurity scattering is about four orders of magnitude larger than that of the short - range one .
we calculate the temperature dependence of the srt corresponding to the case of v@xmath166v in the experiment of han and kawakami@xcite by including the additional short - range scattering .
the result is plotted as the grey dashed curve in fig .
[ fig5](b ) .
a crossover is also observed at high temperature due to the contribution of the intervalley electron - phonon scattering as pointed out previously . by comparing our calculation with the experimental data labeled as @xmath116
, we find that our result becomes comparable at low temperature .
furthermore , by fitting to the experimental data with tunable short - range impurity density shown as the curve with @xmath152 in the inset , our result labeled as the skyblue double - dotted chain curve agrees fairly well with the experimental data in the temperature regime lower than the crossover . as for the temperature higher than the crossover , our result becomes even larger than the experimental one once the short - range scattering is included .
this is because that the intervalley electron - phonon scattering is in the strong scattering limit . with the inclusion of the short - range scattering ,
the intervalley scattering becomes stronger and the srt becomes larger .
this discrepancy may be attributed to other extrinsic effects .
as mentioned above , the srt can be tuned by the short - range impurity density .
we show the short - range impurity density dependence of the in - plane srt at @xmath170k as the purple dotted curve in fig .
[ fig5](b ) , with all other conditions remaining the same as the case of v@xmath166v .
a valley is also observed , similar to the case of the temperature dependence discussed above . at low temperature ,
the intervalley electron - phonon scattering is negligible .
the srt is then determined by the intervalley short - range scattering .
the valley originates from the crossover between the weak and strong intervalley short - range scattering limit by tuning the short - range impurity density . at the crossover point ,
the srt @xmath178ps according to eq .
( [ eq15 ] ) , close to the value @xmath179ps shown in the figure .
this crossover point gives the minimum of the srt and can be taken as a criterion for determining the possible contribution of the short - range scattering .
specifically , the crossover point @xmath180 , @xmath181 , @xmath182 , @xmath183 , and @xmath184ps corresponding to the case of the experimental data shown as @xmath114 , @xmath112 , @xmath115 , @xmath154 and @xmath117 , respectively .
we find that both the experimental data of han and kawakami@xcite and avsar _
et al._@xcite are larger than the crossover point whereas the one of yang _
et al._@xcite is smaller .
this suggests the experimental data of han and kawakami@xcite and avsar _
et al._@xcite can be explained by the short - range scatterers . as for the experiment of yang _
et al._,@xcite some other extrinsic effects have to be considered . as function of electron density .
the dotted curves with symbols correspond to the experimental data of han and kawakami ( hk)@xcite at @xmath185k and avsar _
et al._@xcite and yang _
et al._@xcite at @xmath186k .
the solid curves with symbols correspond to the experimental data at room temperature of han and kawakami,@xcite avsar _
et al._@xcite and yang _
et al._.@xcite red solid ( blue chain ) curve is from our calculation corresponding to the case of room ( low ) temperature in the experiment of han and kawakami.@xcite it is noted that the blue chain curve is calculated with the short - range scatterers with the short - range impurity density @xmath187@xmath106 and potential strength @xmath171mevm@xmath172.,width=321 ] then we turn to compare the electron density dependence of the srts from the experiments and our calculation . among the experiments , han and kawakami@xcite
measured the longest srt at low temperature ( up to several nanoseconds ) shown as the dotted curve with @xmath114 in fig .
this srt is about one order of magnitude larger than the ones from other experimental cases .
in addition , the srts from the experiments show different dependences on electron density .
specifically , at low temperature , both avsar _
et al._@xcite and yang _
et al._@xcite show a decrease of the srt with the increase of the electron density whereas a mild peak is observed by han and kawakami.@xcite at room temperature , the srts from avsar _
et al._@xcite and yang _ et al._@xcite increase with increasing electron density whereas han and kawakami@xcite reported an insensitive density dependence of the srt .
we first show our comparison with the experimental data at low temperature . as shown in fig .
[ fig5](a ) , our result is orders of magnitude larger than the experimental data of han and kawakami@xcite at low temperature without short - range impurities .
however , with the inclusion of the short - range scatterers , our result shows good agreement with the experimental data . here , with the short - range scatterers explicitly included , we also show our comparison with the experimental data of han and kawakami.@xcite with a fixed short - range impurity density @xmath187@xmath106 and potential strength @xmath171mevm@xmath172 , the electron density dependence of the srt from our calculation is shown as blue chain curve in fig .
we find that our result marginally agrees with the experimental data shown as the dotted curve with @xmath114 .
better agreement can be obtained with tunable short - range impurity density . in addition
, we also find that the experimental data of avsar _
et al._@xcite can be explained by the short - range scatterers .
however , to account for the experimental data of yang _
et al._,@xcite one has to include other extrinsic effects .
this is because the srt of yang _ et al._@xcite is smaller than the crossover point in the short - range impurity density dependence of the srt ( @xmath188-@xmath189ps ) , which is taken as a criterion for the possible contribution of the short - range scattering as mentioned previously .
we then address our comparison with the experimental data at room temperature . as shown in fig .
[ fig5](a ) , our result is larger than the experimental data of han and kawakami@xcite with an observable difference at room temperature without short - range impurities . with
the short - range scatterers included , the srt from our calculation becomes even larger than the experimental data as mentioned previously , indicating that other extrinsic effects have to be taken into account . here , we show our comparison with the experiment of han and kawakami@xcite without extrinsic effects . the experimental data and our result are shown as the solid curve with @xmath113 and the red solid curve in fig .
[ fig6 ] , respectively .
we find that the srt in our calculation is about three times as large as the experimental one in low electron density regime and one order of magnitude larger than the experimental one in high density regime .
in summary , we have investigated the electron spin relaxation due to the dp mechanism in blg with only the lowest conduction band being relevant .
the soc of the lowest conduction band is constructed from the symmetry group analysis with the parameters obtained by fitting to the numerical calculation according to the latest report by konschuh _
et al._@xcite from first principles with both the intrinsic and extrinsic soc terms in the pseudospin space included .
we find that the magnitudes of both the out - of- and in - plane components of the soc decrease with increasing momentum at large momentum , indicating a suppression of the inhomogeneous broadening with the increase of the momentum .
this is different from the case in both semiconductors and single - layer graphene .
additionally , the leading term of the out - of - plane component of the soc serves as a zeeman - like term with opposite effective magnetic fields in the two valleys , which is similar to the case in rippled single - layer graphene . this zeeman - like term , together with the intervalley electron - phonon and/or possible intervalley short - range scatterings , opens an intervalley spin relaxation channel , which has not been reported in the literature in blg .
the intervalley electron - phonon scattering is derived by using the tight - binding model .
in addition to the intervalley electron - phonon and short - range scatterings , we also include the long - range electron - impurity , electron - electron coulomb and intravalley electron - phonon scatterings to calculate the srt .
we find that the in - plane srt is strongly suppressed by the intervalley electron - phonon scattering at high temperature . in contrast to the intervalley electron - phonon scattering , the intervalley short - range scattering is insensitive to the temperature and plays an important role in the in - plane spin relaxation especially at low temperature . in the absence of short - range scatterers , a marked nonmonotonic temperature dependence of the in - plane srt
is predicted with a minimum srt down to several hundred picoseconds .
this nonmonotonic behavior originates from the crossover between the weak and strong intervalley electron - phonon scattering .
moreover , we predict a peak in the electron density dependence of the in - plane srt at low temperature , which is very different from the one in semiconductors . at high temperature ,
the in - plane srt increases monotonically with increasing density .
we also find that the in - plane srt decreases rapidly with the increase of the initial spin polarization at low temperature .
this is very different from the previous studies in both semiconductors and single - layer graphene where the srt increases significantly with increasing initial spin polarization .
the physics is understood that the spin relaxation time is determined by the intervalley electron - phonon scattering , which transfers electrons between the two valleys .
the transferred electrons experience opposite effective magnetic fields in the two valleys , which are not affected by the coulomb hf term .
in addition , a strong anisotropy between the out - of- and in - plane spin relaxations is also addressed at high temperature where the out - of - plane srt is about two orders of magnitude larger than the in - plane one .
as for low temperature , the out - of- and in - plane spin relaxation times are comparable .
we also show our comparison with the existing experiments of han and kawakami,@xcite avsar _
et al._@xcite and yang _
et al._.@xcite we find that without intervalley scattering , the srt is orders of magnitude larger than the experimental data in the whole temperature regime . with the intervalley electron - phonon scattering
explicitly included , the srt from our calculation becomes comparable to the experimental data at high temperature but still orders of magnitude larger than the experimental data at low temperature .
in addition , a crossover in the temperature dependence of the in - plane srt is shown , which also results from the crossover between the weak and strong intervalley electron - phonon scattering .
with the inclusion of the short - range scatterers , our result agrees fairly well with the experimental data in the temperature regime lower than the crossover .
as for the temperature higher than the crossover , other extrinsic effects have to be included .
moreover , a crossover point is also shown in the short - range impurity density dependence of the in - plane srt , which gives a minimum for determining the possible contribution of the short - range scattering .
the experimental srt larger than the minimum can be explained by the possible short - range scatterers whereas other extrinsic effects have to be considered for experimental data smaller than the minimum .
this work was supported by the national basic research program of china under grant no .
2012cb922002 and the strategic priority research program of the chinese academy of sciences under grant no .
one of the authors ( l.w . ) would like to thank b. y. sun for double checking the derivation of the electron - phonon scattering matrix elements .
, ( b ) @xmath190 and ( c ) @xmath191 as function of the out - of - plane electric field @xmath36 with @xmath192-@xmath92 .
it is noted that @xmath193 , @xmath194 , @xmath195 , @xmath196 and @xmath197 ( @xmath14-@xmath15 ) are dimensionless.,title="fig:",width=257 ] , ( b ) @xmath190 and ( c ) @xmath191 as function of the out - of - plane electric field @xmath36 with @xmath192-@xmath92 . it is noted that @xmath193 , @xmath194 , @xmath195 , @xmath196 and @xmath197 ( @xmath14-@xmath15 ) are dimensionless.,title="fig:",width=257 ] , ( b ) @xmath190 and ( c ) @xmath191 as function of the out - of - plane electric field @xmath36 with @xmath192-@xmath92 .
it is noted that @xmath193 , @xmath194 , @xmath195 , @xmath196 and @xmath197 ( @xmath14-@xmath15 ) are dimensionless.,title="fig:",width=257 ]
the soc term of the lowest conduction band near the dirac points can be described by a @xmath198 matrix @xmath199 in the basis @xmath200 where @xmath201 and @xmath202 ( @xmath203 ) . the coefficients @xmath204 can be obtained by exactly diagonalizing the @xmath20 effective hamiltonian given in eq .
( [ eq5 ] ) . here , @xmath205 with @xmath206 ( @xmath207 ) being the spin - orbit matrix elements near the dirac points in the pseudospin space given by konschuh
_ et al._.@xcite specifically , @xmath208+\frac{\mu-1}{2}[{c_3^{\mu}}^*c_1^{\mu}(i\lambda_4 ) \nonumber\\ & & \mbox{}\hspace{-0.4cm}+{c_1^{\mu}}^*c_2^{\mu}(-i\lambda_0)+{c_4^{\mu}}^*c_2^{\mu}(-i\lambda_4^{\prime})+{c_2^{\mu}}^*c_3^{\mu}(-i\lambda_3)\nonumber\\ & & \mbox{}\hspace{-0.4cm}+{c_3^{\mu}}^*c_4^{\mu}(i\lambda_0^{\prime})].\label{a7}\end{aligned}\ ] ] @xmath209 and @xmath210 are the strengths of single - layer - like intrinsic soc ; @xmath211 and @xmath212 represent single - layer - like extrinsic spin - orbit strengths ; @xmath213 , @xmath214 and @xmath215 stand for interlayer spin - orbit parameters.@xcite then the effective magnetic field of the soc @xmath216 in eq .
( [ eq6 ] ) is given by @xmath217 , @xmath218 and @xmath219 .
it is noted that the contribution of other energy bands to the soc of the lowest conduction band is marginal in our calculation . and ( b ) @xmath220 as function of the out - of - plane electric field @xmath36 with @xmath192-@xmath221 .
it is noted that @xmath222 , @xmath223 , @xmath224 and @xmath225 ( @xmath226 ) are dimensionless.,title="fig:",width=283 ] and ( b ) @xmath220 as function of the out - of - plane electric field @xmath36 with @xmath192-@xmath221 .
it is noted that @xmath222 , @xmath223 , @xmath224 and @xmath225 ( @xmath226 ) are dimensionless.,title="fig:",width=283 ] as reported by konschuh _
et al._,@xcite the small group of the dirac points is @xmath4 in the presence of an out - of - plane electric field .
following from the symmetry group @xmath4,@xcite we derive an analytical form of the soc near the dirac points [ see eqs .
( [ eq1]-[eq3 ] ) ] . in these equations , the coefficients @xmath13 ( @xmath14-@xmath15 ) and @xmath227 ( @xmath14-@xmath110 )
are obtained by fitting to the numerical calculation from eqs .
( [ a5 ] ) and ( [ a7 ] ) using the pad approximation.@xcite specifically , the coefficients @xmath13 ( @xmath14-@xmath15 ) in eqs .
( [ eq1]-[eq2 ] ) read @xmath228 the coefficients @xmath227 ( @xmath14-@xmath110 ) in eq .
( [ eq3 ] ) are given by @xmath229 here , the coefficient @xmath230ev , which is independent of the applied electric field .
the electric field dependences of @xmath231 ( @xmath14-@xmath15 , @xmath192-@xmath92 ) and @xmath232 ( @xmath14-@xmath110 , @xmath192-@xmath221 ) are plotted in figs .
[ fig7 ] and [ fig8 ] , respectively .
it is noted that the pad approximation@xcite can give a precise description of @xmath13 ( @xmath14-@xmath15 ) and @xmath227 ( @xmath14-@xmath110 ) at large momenta near the dirac points , which can be populated and play an important role in spin relaxation according to the experimental conditions ( heavily doped).@xcite with only a simple linear @xmath10-order approximation , @xmath13 and @xmath227 agree with the numerical results only at very small momentum near the dirac points .
we also show in fig .
[ fig9 ] the spin splitting of the lowest conduction band calculated with analytical form of the soc ( blue dashed curve ) and with the explicit numerical one ( red solid curve ) .
we find that the analytical result agrees fairly well with the numerical one at large momentum .
in addition , the spin splitting decreases with increasing momentum at large momentum , which is very different from the case in both semiconductors@xcite and single - layer graphene.@xcite near the k point as function of the momentum @xmath233 under a typical electric field @xmath104v / nm .
solid ( dashed ) curve represents the numerical ( analytical ) result .
the electron - electron coulomb scattering matrix element is given by @xmath235 .
the screened coulomb potential @xmath236 with @xmath237 being the screening under the random phase approximation.@xcite @xmath238 is the two - dimensional bare coulomb potential with @xmath239 and @xmath240 being the fermi velocity in single - layer graphene@xcite and the dimensionless wigner - seitz radius,@xcite respectively .
@xmath241 is given by@xcite @xmath242 where @xmath243 .
the long - range electron - impurity scattering matrix element @xmath244 , in which @xmath245 and @xmath246 stand for the impurity charge number and effective distance of the impurity layer to the blg sheet.@xcite it is noted that when we calculate the electron - electron coulomb and long - range electron - impurity scatterings , we take the distance between the graphene layers to be zero approximately.@xcite it is further noted that the coulomb potential in the hf term is in the static screening limit , i.e. , @xmath247 .
then we turn to the electron - phonon scattering matrix elements . for the intravalley electron - ac - phonon scattering ,
the scattering matrix element @xmath248 where @xmath249 ; @xmath250 and @xmath251 represent the deformation potential@xcite and acoustic phonon velocity,@xcite respectively . for the intravalley electron - ri - phonon scattering , @xmath252 with @xmath253 being the thomas - fermi screening length.@xcite for sio@xmath0 substrate ,
the energy spectra of the two remote phonon modes are denoted as @xmath254 and @xmath255 , respectively ; @xmath256 and @xmath257 are the corresponding dimensionless coupling parameters.@xcite the matrix elements for the intravalley electron - op - phonon scattering ( @xmath66 ) are derived using the tight - binding model according to the arxiv version of ref . .
the intravalley electron - op - phonon scattering includes both the electron in - plane ( @xmath258 ) and out - of - plane ( @xmath259 ) op phonon scatterings .
the electron in - plane op phonon scattering matrix element is given by @xmath260\psi^{\mu}_{{\bf k}^{\prime}}|^2+|{\psi^{\mu}_{{\bf k}}}^{\dagger } [ i\gamma_0^{\prime}(\sin\theta_{\bf q}\sigma^{01}_{\rm d}-\cos\theta_{\bf q}\nonumber\\ & & \mbox{}\hspace{-1.9cm}\times\sigma^{02}_{\rm d})+a\gamma_4^{\prime}(i\cos\theta_{\bf q}\gamma^3_{\rm d}- \sin\theta_{\bf q}\gamma^3_{\rm d}\gamma^5_{\rm d})/(\sqrt{3}l_4)]\psi^{\mu}_{{\bf k}^{\prime}}|^2\nonumber\\ & & \mbox{}\hspace{-1.9cm}+|{\psi^{\mu}_{{\bf k}}}^{\dagger}[\gamma_0^{\prime}(\cos\theta_{\bf q}\sigma^{23}_{\rm d } -\sin\theta_{\bf q}\sigma^{13}_{\rm d})-a\gamma_3^{\prime}(\cos\theta_{\bf q}\gamma^1_{\rm d}\gamma^5_{\rm d}\nonumber\\ & & \mbox{}\hspace{-1.9cm}-i\cos\theta_{\bf q}\gamma^2_{\rm d}-i\sin\theta_{\bf q}\gamma^1_{\rm d}-\sin\theta_{\bf q}\gamma^2_{\rm d } \gamma^5_{\rm d})/(2\sqrt{3}l_3)]\psi^{\mu}_{{\bf k}^{\prime}}|^2\nonumber\\ & & \mbox{}\hspace{-1.9cm}+|{\psi^{\mu}_{{\bf k}}}^{\dagger}[i\gamma_0^{\prime}(\cos\theta_{\bf q}\sigma^{01}_{\rm d } + \sin\theta_{\bf q}\sigma^{02}_{\rm d})-a\gamma_4^{\prime}\gamma^3_{\rm d}(i\sin\theta_{\bf q}\nonumber\\ & & \mbox{}\hspace{-1.9cm}+\cos\theta_{\bf q}\gamma^5_{\rm d})/(\sqrt{3}l_4)]\psi^{\mu}_{{\bf k}^{\prime}}|^2\big\},\end{aligned}\ ] ] where @xmath261 is the energy of the in - plane op phonon modes;@xcite @xmath262 is the bond length corresponding to the interlayer hopping @xmath31 ; and @xmath263 and @xmath264 are @xmath20 dirac matrices given in appendix [ appc].@xcite here , @xmath265 is the polar angle of the momentum @xmath266 .
the electron out - of - plane op phonon scattering matrix element reads @xmath267 being the energy of the out - of - plane op phonon mode.@xcite
the relevant dirac matrices@xcite used in the electron - phonon scattering are @xmath268 @xmath269 @xmath270 with @xmath271 being @xmath198 unit matrix .
yang , j. balakrishnan , f. volmer , a. avsar , m. jaiswal , j. samm , s. r. ali , a. pachoud , m. zeng , m. popinciuc , g. gntherodt , b. beschoten , and b. zyilmaz , phys .
lett . * 107 * , 047206 ( 2011 ) .
it is noted that due to the complicated energy spectrum of blg , we employ the single - layer graphene approximation to calculate the spin diffusion coefficient in blg according to the report by zhang and wu.@xcite | electron spin relaxation due to the dyakonov - perel mechanism is investigated in bilayer graphene with only the lowest conduction band being relevant .
the spin - orbit coupling is constructed from the symmetry group analysis with the parameters obtained by fitting to the numerical calculation according to the latest report by konschuh _
et al . _
[ phys .
rev .
b * 85 * , 115423 ( 2012 ) ] from first principles .
in contrast to single - layer graphene , the leading term of the out - of - plane component of the spin - orbit coupling in bilayer graphene shows a zeeman - like term with opposite effective magnetic fields in the two valleys . this zeeman - like term opens a spin relaxation channel in the presence of intervalley scattering .
it is shown that the intervalley electron - phonon scattering , which has not been reported in the previous literature , strongly suppresses the in - plane spin relaxation time at high temperature whereas the intervalley short - range scattering plays an important role in the in - plane spin relaxation especially at low temperature .
a marked nonmonotonic dependence of the in - plane spin relaxation time on temperature with a minimum of several hundred picoseconds is predicted in the absence of the short - range scatterers .
this minimum is comparable to the experimental data .
moreover , a peak in the electron density dependence of the in - plane spin relaxation time at low temperature , which is very different from the one in semiconductors , is predicted .
we also find a rapid decrease in the in - plane spin relaxation time with increasing initial spin polarization at low temperature , which is opposite to the situation in both semiconductors and single - layer graphene .
a strong anisotropy between the out - of- and in - plane spin relaxations at high temperature is also revealed with the out - of - plane spin relaxation time being about two orders of magnitude larger than the in - plane one .
detailed comparisons of the temperature and electron density dependences of the spin relaxation with the existing experiments [ phys .
rev .
lett . * 107 * , 047206 ( 2011 ) , phys .
rev .
lett . * 107 * , 047207 ( 2011 ) and nano lett . *
11 * , 2363 ( 2011 ) ] are reported . |
starbursts induced by major mergers are among the most extreme environments in the universe . despite their prodigious luminosities ,
local merger - driven starbursts are very compact , with most of their large gas reservoirs concentrated in dusty regions a few hundred parsecs , or less , in size .
measuring the compactness of these starbursts is critical to understanding these galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
robust size measurements allow us to translate luminosities into key physical quantities such as gas column density , optical depth , volumetric gas density , and star formation rate and luminosity surface densities .
although their luminosity renders them visible out to great distances , the present - day rarity of major mergers means that even the nearest ultra - luminous infrared galaxies ( ulirgs : defined as having @xmath22 \geq 10^{12}~l_{\odot}}$ ] ) are relatively distant ( @xmath23 70 mpc ) .
thus , measuring the true extent of their active regions requires high angular resolution .
the extraordinary extinctions present in these systems at both long ( from free - free absorption ) and short ( from dust opacity ) wavelengths complicate the interpretation of measurements at both wavelengths , compounding the difficulty of measuring sizes for such systems . given the above considerations , radio observations at centimeter wavelengths may be the best tool to study the deeply embedded , compact structures at the heart of such systems ( e.g. , * ? ? ?
* ; * ? ? ?
radio interferometers can achieve very high angular resolution and radio waves with @xmath24 5 ghz can penetrate large columns of dust and are largely unaffected by free - free absorption .
the recent upgrades to the karl g. jansky very large array ( vla ) make it particularly well - suited for such studies . in this paper , we make use of these new vla capabilities to achieve the best measurement to date of the structure of the nuclear region of the nearest ulirg , arp 220 . at a luminosity distance of 77.2 mpc , and an infrared luminosity of @xmath25 = 1.44 \times 10^{12}~l_{\odot}}$ ] , from ned ; @xmath26 using table 1 in and iras flux densities from @xcite .
however , note that the assumption of isotropic emission that leads to this luminosity has some caveats ( see section [ sec : ir ] and appendix [ app : flux ] ) . ]
arp 220 is the nearest ulirg .
co and near - ir observations indicate that arp 220 is a gas rich merger with dynamical masses of @xmath27 within @xmath5 100 pc of each nucleus .
arp 220 is obscured at optical through mid - ir wavelengths @xcite , obstructing the direct view of the nuclear energy sources at these wavelengths .
observations in the frequency range where arp 220 is optically thin have been able to resolve the system into two compact nuclear disks and find disk sizes of @xmath5 02 .
however , in each case the measured sizes remain comparable to the size of the beam .
vlbi observations at cm wavelengths by @xcite , @xcite , and @xcite provide a higher resolution view , recovering a compact distribution of point - like sources that are proposed to be a combination of radio supernovae ( rsne ) and supernova remnants ( snrs ) .
however , these observations resolve out most of the emission from the disks . from the above
, it is already clear that the disks are very compact , implying extraordinary volume densities and surface densities .
the next step is to observe the disks in the optically thin frequency range with resolution high enough to clearly resolve them , and sensitivity to recover the full extent of the emission of the system at that frequency range . in this paper
, we measure the structure of the arp 220 nuclei with sensitive , high angular resolution images obtained at 6 ghz and 33 ghz observed with the vla . based on the integrated spectral energy distribution ( sed ) model of arp 220 ( see figure 10b in * ? ? ?
* ) , the total continuum flux density at 33 ghz is a mixture of thermal and non - thermal emission with a @xmath5 1:2 ratio , while at 6 ghz this ratio is about 1:5 .
observing at these two frequencies then helps us diagnose the dominant emission mechanism at radio wavelengths in arp 220 .
we first report our observations , describe the calculations used to assess the disk structure , and then discuss the implications of our measurements . throughout this paper ,
we adopt @xmath28 , @xmath29 , @xmath30 and @xmath31 ( after correction to the cmb frame ) , such that 1 on the sky plane subtends 369 pc at the distance of arp 220 .
we observed arp 220 using the vla c ( 4 - 8 ghz ) and ka band ( 26.5 - 40 ghz ) receivers , recording emission in 1 ghz wide windows centered at @xmath5 4.7 , 7.2 , 29 and 36 ghz .
we used all four vla configurations with a total integration time ratio of 1:1:2:4 between d ( lowest resolution ) , c , b , and a. the total on - source integration time was 40 min at c band and 56 min at ka band .
we used 3c 286 as the flux density and bandpass calibrator , and j1513 + 2338 and j1539 + 2744 as the complex gain calibrators at c and ka bands , respectively .
the data were obtained in multiple observing sessions during the period 2010 august 18 to 2011 july 2 , with the c and ka band observations carried out in separate sessions .
these observations are part of a larger project ; the lower resolution ( c and d configuration ) results are presented in @xcite and @xcite , and the final results for the complete sample will be reported in barcos - muoz et al .
( in prep . ) .
we reduced the data using the common astronomy software application ( casa , * ? ? ?
* ) package following the standard procedure for vla data .
radio frequency interference ( rfi ) contaminating the c band was eliminated using the task ` flagdata ` in mode ` rflag ` .
the rfi at ka band was negligible .
after calibration , we combined the data from all configurations , weighting them in proportion to their integration time per visibility ( i.e. , 10:5:2:1 for d : c : b : a ) .
we then imaged this combined data using the task ` clean ` in mode ` mfs ` @xcite , with _
briggs _ weighting setting ` robust=0.5 ` .
we combined all the data within each receiver band and cleaned using components with a variable spectral index ( ` nterms=2 ` ) to obtain an interpolated image at an intermediate frequency ( 5.95 ghz for c band and 32.5 ghz for ka band ) . even after the initial calibration , we still observed phase and amplitude variations with time . to improve the images further , we iteratively self - calibrated in both phase and amplitude and applied extra flagging during this procedure as needed .
the solutions for the amplitude self - calibration were carefully inspected and accepted as long as the time variations in the amplitude gains for each antenna were less than @xmath5 20% .
we worked mostly with these `` combined '' images at 5.95 and 32.5 ghz , but we also separately imaged the two 1 ghz windows at c band in order to derive a robust internal c band spectral index . to check our results , we imaged the 33 ghz data separately for each vla configuration . in this test
, we primarily applied iterative phase self - calibration .
amplitude self - calibration was applied ( after @xmath32@xmath33 iterations of phase self - calibration ) , but here we only derived normalized solutions that can not change the observed flux . during this check
, we also experimented with weighting the visibilities by the measured rms noise in each data set ( using the casa task ` statwt ` ) .
these tests revealed that the combined a+b ( two most extended ) configurations recovered essentially all of the flux in the data , agreeing with the c and d data in both flux and morphology when convolved to matched resolution .
the a configuration data alone recovered less flux than the b configuration , consistent with some spatial filtering at this highest resolution .
further , the overall flux recovered agrees with an interpolation of the integrated sed @xcite .
we proceed using the full combined image with our confidence in the results reinforced by these tests ; we verified that our fitting yields consistent results using the combined image and the a+b configuration - only image .
the clean restoring beam for the combined images has a fwhm of 048 @xmath0 035 ( 177 @xmath0 129 pc ) at position angle ( p.a . ) of -40@xmath34 at 6 ghz ( c band ) , and 0081 @xmath0 0063 ( 30 @xmath0 23 pc ) at a p.a . of 65@xmath34 at 32.5 ghz ( ka band ) .
the rms noise measured from signal - free parts of the image is @xmath35jy beam@xmath36 ( c ) and @xmath37jy beam@xmath36 ( ka ) , which is within a factor of two of the expected theoretical noise .
the final dynamic ranges of the images are @xmath38 and @xmath5 280 , for c and ka band . when reporting the measured flux densities from the final images , we consider three sources of uncertainty .
first , we propagated the beam - to - beam noise ( see above ) and found its effect to be negligible at both bands .
second , we assessed the impact of the curve - of - growth technique used to measure the flux densities ( see section [ sec : calc ] ) by using the scatter of such a curve .
this is also small , but larger for the individual nuclei because they are not perfectly separable . finally , we estimated the uncertainty in the overall flux density calibration from the day - to - day variation of the flux density of the complex gain calibrator .
this scatter ( rms ) is @xmath5 12% at ka band , making the flux density calibration the dominant source of uncertainty at this band . at c band ,
the scatter in the curves of growth and the variation in the flux density calibration are comparable , i.e. , @xmath5 1% .
we sum all three uncertainty terms in quadrature and report the combined value in table [ tbl-1 ] .
note that in addition to these uncertainties , the absolute flux scale used at the vla is estimated to be uncertain by @xmath39 .
after reducing the data , we compared our 33 ghz image to vla @xcite , sma @xcite and alma archival images ( * ? ? ? * and scoville et al . _ in prep _ ) .
we found astrometric discrepancies of order 01 ( i.e. , 12 beams ) and traced the origin of these to the adopted position of our 33 ghz phase calibrator .
when revised from the nominal vla position to the position reported in the vlbi calibrator catalog , the astrometric agreement between our image and the other images improved to a fraction of a beam . for reference , in our data
the peak positions of the two nuclei at 33 ghz are @xmath40 , @xmath41 ( east ) and @xmath42 , @xmath43 ( west ) .
we derive these position via a gaussian fit ( casa s imfit ) but they also closely coincide with the positions of the highest intensity pixel for each nucleus . the uncertainties in the peak positions above may also be viewed as our overall astrometric uncertainty , which we derive from the standard deviation between the positions of the highest intensity pixels from our 6 ghz image and our shifted 33 ghz image , and the archival images from @xcite , @xcite , @xcite , @xcite and scoville et al .
_ in prep_.
in figures [ fig1 ] and [ fig2 ] , we present new vla images of arp 220 at 6 and 33 ghz and the radial profiles of each nucleus at 33 ghz . using these new data , combined with a co ( 3@xmath442 ) integrated intensity ( `` zeroth moment '' ) map from @xcite , and positional information of point sources found by @xcite , we carry out a series of calculations to determine what mechanism is producing most of the radio emission , how radio emission traces recent star formation , and the true sizes and shapes of the nuclear disks . in tables
[ tbl-1 ] , [ tbl-2 ] and [ tbl-3 ] we present the results of these calculations .
cccccc[h ] 4.7 & 222.0 @xmath45 1.9 & 92.4 @xmath45 2.1 & 61.8 @xmath45 0.5 & 114.6 @xmath45 4.3 & 89.5 @xmath45 0.7 + 7.2 & 171.4 @xmath45 2.3 & 73.2 @xmath45 1.4 & 36.0 @xmath45 0.4 & 89.5 @xmath45 1.4 & 60.4 @xmath45 0.7 + 5.95 & 197.6 @xmath45 2.8 & 81.4 @xmath45 2.8 & 49.0 @xmath45 0.7 & 94.3 @xmath45 1.7 & 73.3 @xmath45 1.0 + 32.5 & 61.8 @xmath45 7.2 & 30.1 @xmath45 3.9 & 4.1 @xmath45 0.5 & 33.4 @xmath45 4.0 & 6.5 @xmath45 0.8 + @xmath46 & -0.69 @xmath45 0.07 & -0.59 @xmath45 0.08 & & -0.61 @xmath45 0.07 + @xmath47 & -0.61 @xmath45 0.04 & -0.55 @xmath45 0.07 & & -0.58 @xmath45 0.09 in table [ tbl-1 ] , we report the flux density of the entire system , each nucleus , and the resulting spectral indices . we used a curve of growth method to derive the flux densities . for the integrated flux density , we progressively @xmath48 tapered and re - imaged the data , recording the total flux density above a signal - to - noise of 5 at each resolution .
these flux densities agree with those measured from the imaging of individual arrays ( see section [ sec : obs ] ) . for the flux densities of the individual nuclei
, we used casa s ` imstat ` task to place circular apertures around each component , varying their radii .
we plotted the flux density against aperture radius and looked for convergence in this curve - of - growth to identify the true flux density .
we also independently measured the integrated flux of the south - west component seen at c band ( see figure [ fig1]a ) , using an aperture in the casa ` viewer ` , and found that it encloses @xmath5 3% of the total 5.95 ghz flux density .
the integrated flux densities at both frequencies agree , within the reported errors , with predicted values based on the modeled integrated sed published in the literature ( e.g. , * ? ? ?
* ) . using these flux densities , we calculated the spectral indices , @xmath49 for @xmath50 , of the whole system and each nucleus via @xmath51 where @xmath52 and @xmath53 are the flux densities at frequencies @xmath54 and
@xmath55 . equation [ eq : index ] will be valid for flux densities @xmath56 over matched apertures ( or for integrated values over whole systems ) . to assess the degree to which the measured sizes are characteristic of the whole system , we compared our maps to known distributions of emission from gas and recent radio supernovae and/or supernova remnants ( rsne / snrs ) . in figure [ fig1]b , we plot c band contours over the co ( 3@xmath442 ) map of @xcite ( restoring gaussian beam with fwhm 038 @xmath0 028 at a p.a @xmath57 23@xmath34 ) . the rsne / snrs trace recent star formation , and they can accelerate cosmic ray ( cr ) electrons that emit synchrotron radiation .
we built a map of rsn / snr number density from the locations of 49 point sources identified by @xcite from 18 cm vlbi observations .
on our 33 ghz astrometric grid , we convolved delta functions with a fixed , fiducial intensity at the positions of the point sources with our 33 ghz beam , @xmath58 . ] . in figures
[ fig1]c and [ fig1]d , we compare the 33 ghz map to the distribution of recent rsne / snrs .
note that although we use rsne / snrs as signposts of recent star formation , the vlbi sources do not contribute significantly to the flux that we observe . for a typical synchrotron spectral index @xmath59 , the @xcite
rsne / snrs would contribute 1.5 mjy at 33 ghz and 4.9 mjy at 6 ghz .
this contribution would only account for @xmath5 2.5 % of the total flux density that we observe with the vla . even at 18 cm , the @xcite rsne / snrs have integrated flux only @xmath5 12 mjy at 18 cm , or @xmath5 4% of the total flux density of arp 220 at that frequency .
this contribution is small compared to the 10% expected fraction in normal spiral galaxies like m31 or the milky way @xcite .
this difference is most likely due to free - free absorption at 18 cm ( see section [ sec : od ] ) . ccccc[th ] scale length ( pc ) & 30.3 @xmath45 4.6 & 34.0 @xmath45 5.1 & 21.0 @xmath45 3.2 & 25.4 @xmath45 3.8 + peak intensity ( mjy beam@xmath36 ) & 6.0 @xmath45 0.7 & 4.8 @xmath45 0.6 & 13.4 @xmath45 1.6 & 9.0 @xmath45 1.1 + position angle ( @xmath60 ) & 54.7 @xmath45 0.6 & 55.4 @xmath45 0.6 & 79.4 @xmath45 0.8 & 77.3 @xmath45 0.8 + inclination ( @xmath60 ) & 57.9 @xmath45 0.6 & 55.4 @xmath45 0.6 & 53.5 @xmath45 0.5 & 49.1 @xmath45 0.5 + @xmath61 ( pc ) & 50.8 @xmath45 7.6 & 57.0
@xmath45 8.6 & 35.2 @xmath45 5.3 & 42.7 @xmath45 6.4 + fwhm major axis ( pc ) & 85.9 @xmath45 8.6 & 90.8 @xmath45 9.1 & 63.7 @xmath45 6.4 & 70.2 @xmath45 7.0 + fwhm minor axis ( pc ) & 46.3 @xmath45 4.6 & 51.8 @xmath45 5.2 & 38.0 @xmath45 3.8 & 44.7 @xmath45 4.5 + position angle ( @xmath34 ) & 56.0 @xmath45 1.1 & 56.5 @xmath45 1.1 & 78.7 @xmath45 1.6 & 77.1 @xmath45 1.5 + @xmath62 ( pc ) & & 73.3 @xmath45 7.5 & & 45.5@xmath45 3.7 + the smooth , ellipsoidal isophotes in figure [ fig1 ] suggest a disk - like geometry .
we modeled the 3 @xmath63 clipped image of arp 220 at 33 ghz ( outermost contour in figure [ fig1]c ) using a 2-d non - linear least - squares fitting technique . after experimenting with gaussian , exponential , srsic and hybrid profiles
, we found that the two disks are reasonably described by thin , tilted exponential disks .
we fit both nuclei simultaneously by varying , without constraints , the amplitude , position angle ( p.a . ) , inclination , center , and scale length of each nucleus .
although the parameters did not have constraints , the starting points were educated guesses of the final parameters . in each case , we construct the model image , convolve it with the synthesized beam of our observations , and compare the model and observed intensities to derive @xmath64 .
note that the results for the inclination of the disks represent lower limits because we assume thin disks .
the best fit parameters from the model fitting , along with associated uncertainties , are reported in table [ tbl-2 ] .
in addition to deriving formal uncertainties , we gauge the accuracy of our fit by varying our approach among several reasonable methods .
for example , we adopt a logarithmic , rather than linear , goodness of fit statistic and we fit the radial profile rather than the image itself . these imply an uncertainty of @xmath65 for the scale length and a few percent for p.a . and inclination .
the error in the normalization is dominated by our overall uncertainty in the amplitude calibration ( @xmath66 ) . as another point of comparison
, we also report the results of simple gaussian fitting , although , we emphasize that the residuals are substantially poorer for this approach at low and high radius . from the exponential model ,
we obtain deconvolved scale lengths of @xmath67 and @xmath68 pc for the east and west nucleus , respectively . our results for p.a . ,
inclination , and center did not vary significantly with the choice of functional form .
while the fits appear to be good descriptions , they are not perfect . from the residual images
, we found that the western nucleus showed higher residuals in the disk than the eastern nucleus , while the center of the eastern nucleus had higher residuals than the western nucleus .
both nuclei are well resolved , showing significant extent compared with the synthesized beam .
the implied deconvolved half - light radii , @xmath2 , are @xmath3 and @xmath4 pc , respectively ; that is , if viewed face - on , we would expect half the emission from arp 220 to come from nuclear disks @xmath69 ( east ) and @xmath70 ( west ) pc across . in figures
[ fig1 ] and [ fig2 ] , we have also shown that the size measurement agrees with that implied by the rsn / snr distribution @xcite .
the p.a.s of the east and west nuclei agree well with those of the kinematic major axes of the disks measured from sub - millimeter co observations @xcite , h i absorption observations from @xcite , and @xmath71 radio recombination line @xcite . the velocity gradients along these position angles on individual nuclei have also been observed in the 2 @xmath72 m @xmath73 line @xcite which traces hot molecular gas and 2.3 @xmath72 m co absorption ( the latter traces stellar velocities : @xcite ) .
we performed two checks on the size measurements .
first , as a point of comparison , we report in table [ tbl-2 ] a 2-d gaussian fit to each nucleus at 33 ghz .
we obtained deconvolved fwhm sizes of 023 @xmath0 013 ( @xmath74 ) for the eastern and 017 @xmath0 010 ( @xmath75 ) for the western nucleus .
these sizes agree fairly well with previous , marginally resolved , estimates at other frequencies .
@xcite found a deconvolved major axis size of 019 = 70 pc for the western nucleus at 1.3 mm .
@xcite found major axis sizes of 027 = 100 pc ( fwhm , east ) and 016 = 59 pc ( fwhm , west ) at 860 @xmath72 m .
however , we show through radial profiles ( figure [ fig2 ] ) that the disks are better described by an exponential morphology .
in fact , the deconvolved gaussian fit would underestimate the deconvolved half - light diameter of the disks by 9 % ( west ) and 15% ( east ) , if we account for inclination effect in the gaussian fit , i.e , the deconvolved @xmath76 is smaller by @xmath5 9% and 15% when compared to the deconvolved half - light diameter ( @xmath77 ) .
second , we calculated the area on the sky containing half of the flux associated with each nucleus .
this very basic measure still suffers from beam dilution and inclination effects , but provides a measure of size that is independent of the functional form .
we derived this image - based @xmath78 by identifying the isointensity contour that encloses 50% of the total flux density of each nucleus .
we summed the area of the pixels ( pixel size = 002 ) enclosed within that contour and estimate the observed radius for @xmath79 ( east ) and @xmath80 ( west ) , and @xmath81 .
this is @xmath5 73 pc for the eastern nucleus and @xmath5 46 pc for the western nucleus .
we report the observed @xmath82 values in table [ tbl-2 ] . as with the gaussian
, the measured area shows broad agreement with the exponential profile fitting , though differing in detail .
we created deprojected , azimuthally averaged radial profiles of 33 ghz intensity to assess the accuracy of our models and compare the structure of the nuclei to that of the rsn / snr number density map . assuming a thin tilted ring geometry , we calculated deprojected profiles for the observed emission , the emission in the convolved model , and the rsn / snr number density map . in each case , the center of the profiles correspond to the highest intensity pixel in the observed image for each nucleus .
we plot these profiles in figure [ fig2 ] .
we included only emission above a signal - to - noise ratio of 3 and then averaged the intensity in a series of inclined , @xmath50035-wide ( half the clean beam size ) rings , adopting the best - fit model inclination and p.a . from our modeling ( table [ tbl-2 ] ) .
note that these rings oversample the @xmath5 007 beam , so that adjacent bins in figure [ fig2 ] are not independent .
we normalized the rsn / snr radial profile to match the 33 ghz profile at @xmath83 .
the plotted error bars were calculated from the standard deviation of the flux within each annulus divided by the square root of the area in that annulus expressed in units of the beam size ( i.e. , the number of independent beams ) . in figure [ fig2 ] , we show that the exponential model matches the data well for both nuclei , matching slightly better for the western nucleus . meanwhile , the gaussian profile is not as good as the exponential profile when compared to the observed data , being particularly poorer in the outer parts of the disks .
the linearity of the semilog profiles also confirms ( and motivates ) our adoption of an exponential functional form .
the rsn / snr radial profiles mostly follow the integrated radio emission profiles ( and thus also the model ) . the agreement is better in the western nucleus .
the eastern nucleus lacks a bright central peak and shows a somewhat more scattered distribution , which could be caused by stochasticity and timescale effects ; i.e. , there just may not be enough rsne / snrs visible to give a smooth appearance ( compare to figure [ fig1]d to see the clumpy nature of the sne distribution ) .
the same idea applies for the outskirts of the western nucleus . in the rest of the paper
, we will consider that the rsn / snr number density distribution follow the continuum emission observed at 33 ghz closely enough that we can take our measured 33 ghz sizes as indicative of the distribution of active star formation in arp 220 .
cccccc 5.95 & 4.64 @xmath45 0.07 & 4.43 @xmath45 0.10 & 4.81 @xmath45 0.10 + 32.5 & 2.7 @xmath45 0.1 & 2.5 @xmath45 0.1 & 2.9 @xmath45 0.1 the brightness temperature , @xmath84 , can be used to constrain the emission mechanism and energy source , and may give clues regarding the optical depth .
with well resolved sizes , we can circumvent beam dilution that often confuses estimates of @xmath84 .
we calculated @xmath84 using the rayleigh - jeans approximation via @xmath85 with @xmath86 the flux density at frequency @xmath87 and @xmath88 the area subtended by the source .
we report average rayleigh - jeans brightness temperatures , @xmath89 , in table [ tbl-3 ] . from our model
, we take the area @xmath90 that we expect to enclose half the emission if the system were viewed face on ( see the `` deconvolved '' column in table [ tbl-2 ] ) . assuming this to be the true area of the disk at all frequencies , we derive average @xmath84 over the half - light region .
this means we used half of the observed flux density for @xmath91 and @xmath92 for @xmath93 in equation [ eq : tb ] .
our rationale for this assumption is that the 33 ghz image appear to be optically thin , high - resolution tracer of the distribution of recent star formation . assuming that this structure is common across wavelength regimes allows us to use a `` true '' size in place of a size observed with a much coarser beam .
we also calculated the peak @xmath84 at each band from the peak flux density and the area of the clean beam at each frequency .
this peak @xmath84 is higher at 33 ghz than at 6 ghz ; this simply reflects that the area of arp 220 at 33 ghz is smaller than the beam size at 6 ghz . for exactly this reason
the small size of arp 220 and the variable resolution at different frequencies the peak measurement has limited utility and we only report the `` average '' version . the average @xmath84 is from @xmath94 k at 6 ghz to @xmath95 ( @xmath96 ) k at 33 ghz , for the east nucleus and @xmath97 k at 6 ghz to @xmath98 ( @xmath99 ) k at 33 ghz , for the west nucleus .
figure [ fig1 ] shows that our observations clearly separated the nuclei at both 6 and 33 ghz and resolve the structure of both nuclei is resolved at 33 ghz .
we find a projected nuclear separation of 096 @xmath45 001 ( 354 @xmath45 4 pc ) , in agreement with previous works ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in the following subsections ,
we discuss the radio continuum emission processes , the correspondence with gas and dust emission , and consequences of the small sizes of the emission regions .
there is a large scale agreement between the locations of radio continuum emission , co , and rsne / snrs , suggesting that the sizes of the radio continuum sources may be viewed as characteristic of the system .
synchrotron radiation appears to produce most of the continuum emission at both 6 and 33 ghz .
the high brightness temperature of a few @xmath0 @xmath100 k , inferred at 6 ghz by using the 33 ghz nuclear sizes , argues in this direction .
this high brightness temperature can not come from hii regions , even if they are completely opaque , because in a purely thermal environment , the electron temperature of such regions should not exceed @xmath100 k. if we combined the high brightness temperature with the observed internal c band spectral index of the total system , @xmath101 , we infer that most of the emission at 6 ghz is synchrotron .
the spectral index between 6 and 33 ghz , @xmath102 , matches the internal c band @xmath49 of the total system within the errors ; the same is true for the two nuclei separately ( see table [ tbl-1 ] ) .
the similarity between @xmath103 and @xmath104 indicates no significant spectral flattening between 6 and 33 ghz , suggesting that synchrotron dominates the emission across this range of frequencies . reinforcing this point ,
our total flux density and spectral index agree with the predictions made by @xcite ( see their figure 10b ) .
they found synchrotron emission to dominate below @xmath105 ghz , and estimated the thermal fraction at 6 ghz and 33 ghz to be @xmath5 15% and 35% , respectively .
if we assume a non - thermal spectral index of @xmath106 ( see table 9 in * ? ? ?
* ) and a typical thermal spectra index of @xmath107 , we obtain @xmath108 and @xmath109 , which is consistent with what we measure and with our observation of a constant spectral index from 6 to 33 ghz .
this result is also consistent with previous results showing lower thermal fractions at 33 ghz for merging starbursts compared to normal galaxies @xcite .
the overall star formation rate of the system provides an alternate way to estimate the expected thermal radio continuum emission .
beginning with the ir ( 8 to 1000 @xmath72 m ) luminosity of arp 220 , we estimate the expected thermal luminosity of the system if the ir is all due to star formation by following star formation rate conversions from table 8 in @xcite and assuming an electron temperature of 7500 k @xcite .
this approach predicts a thermal fraction of @xmath5 55% at 33 ghz and @xmath5 20% at 6 ghz .
this is in good agreement with the thermal fraction at 33 ghz obtained in @xcite for a prototypical starburst , m82 .
however , if we derive the expected spectral index as we did at the end of the previous paragraph ( assuming 55% of thermal fraction at 33 ghz ) , we obtain @xmath110 and @xmath111 , which deviates considerably from what we observe between 6 and 33 ghz .
note that @xmath103 does not vary significantly from what we observe , this is due to the small thermal fraction expected at this frequency range .
the easiest explanation for the lower - than expected thermal flux is that a significant fraction of the ionizing photons produced by young stars are absorbed by dust before they produce ionizations .
this would lower the free - free estimate in the calculation .
we estimate that to match our observations , we would require that at least 20% of the ionizing photons be absorbed by dust .
this number seems plausible for an environment as dust embedded as arp 220 ( see section [ sec : od ] ) and is consistent with some of the arguments made when considering the apparent deficit of ir cooling line emission ( * ? ? ?
* and references therein ) .
alternatively , an imf that produces more bolometric light ( and thus ir and likely sne ) relative to ionizing photons could resolve the discrepancy .
that is , we could invoke an `` intermediate - heavy '' imf compared to that used in @xcite .
we could also reconcile our two estimates if the synchrotron spectral slope decays drastically between 6 and 33 ghz , so that the apparent 33 - 6 ghz index is a combination of very steep , curving synchrotron and emerging thermal emission .
however , this would require that thermal emission make up most of the sed at higher frequencies , which is not observed ( see * ? ? ?
* ; * ? ? ?
our best interpretation of the data is that the 33 ghz emission is mostly synchrotron , in mild contrast with a typical starburst galaxy ( see figure 1 in * ? ? ?
we suggest that the most likely cause is the suppression of thermal radio emission as dust absorbs ionizing photons . in figure [ fig1]b , we show the 6 ghz emission is largely co - spatial with co emission .
the 6 ghz emission is our more sensitive band , with a beam nearly matched to the co , and as just discussed we expect that it traces the same synchrotron emission as the 33 ghz .
the co and 6 ghz emission cover roughly the same area , have broadly coincident peaks , and both show an extended faint feature to the southwest ( * ? ? ?
* note a similar coincidence between 18 cm emission and the arp 220 starburst traced in the near - ir ) .
the distributions of co and 6 ghz emission do significantly differ in detail .
the ratio of fluxes for the two nuclei is 1:2 ( east : west ) for co and almost 1:1 for continuum . in the west nucleus , the morphology is more centrally concentrated at 6 ghz compared to the co map .
some of these differences may reflect real differences between the current gas reservoir and recent star formation , but they may also reflect temperature and optical depth effects .
the co ( 3@xmath442 ) emission in this region shows good evidence for optical thickness @xcite , and the densities are high enough that the gas temperature will be likely coupled to the dust ( @xmath112 k).therefore , making a straightforward interpretation of the co in terms of column density is challenging .
we draw the broad conclusion from figure [ fig1]b that the synchrotron originates from the same region as , and in very rough proportion to , the molecular gas supply .
a similar situation is also observed on smaller spatial scales by the comparison of the rsn / snr number density map to the 33 ghz map ( see figure [ fig1]d ) .
the distributions are co - spatial , but the continuum map appears smoother than the map made from individual rsn / snr .
this is particularly evident in the eastern nucleus , where the covering fraction of vlba point sources is small perhaps a result of stochasticity in the rate and lifetime of sn visible using vlbi measurements .
the radial profiles in figure [ fig2 ] highlight the quantitative agreement between the continuum and the rsne / snrs distribution even more .
after azimuthal averaging , the vlba point source maps are a fairly close match to the 33 ghz continuum .
such a close match between the 33 ghz continuum extent and the rsn / snr number density map may not be too surprising : if synchrotron radiation arises from cosmic ray ( cr ) electrons accelerated by sn shocks , then the 33 ghz continuum emission might be expected to resemble a `` puffed up '' version of the rsn / snr distribution due to the diffusion of cr electrons .
instead the distributions match quite well , consistent with most of the 33 ghz emission coming from very close to the original rsn / snr and little diffusion or secondary cr electron production .
this lack of significant propagation could be explained by the cooling timescales being much smaller than the diffusion time .
this is expected in compact starbursts with magnetic fields of the order of a mg ( see measurements from * ? ? ?
* ; * ? ? ?
* based on zeeman splitting of oh megamaser emission ) , like arp 220 ( see figure 1 in * ? ? ?
* ) , though not in normal galaxies @xcite .
for arp 220 , the cooling time of cr electrons at 33 ghz is @xmath5@xmath113 , which is a combination of synchrotron , bremsstrahlung , ionization and inverse compton ( ic ) losses .
we make use of equation 7 in @xcite for cr electrons with energies greater than 1 gev to estimate that the synchrotron emitting electrons at 33 ghz only have time to propagate about 5 pc , which is about 1/10th of the size that we have measured for the nuclei .
this short diffusion scale yields a synchrotron image that looks very similar to the sites of original cr production ( the rsn / snr ) and thus the sites of active star formation .
the advantage of the vla continuum in this case is that in exchange for coarser native resolution , we achieve sensitivity to most of the flux and spatial scales of interest ( and potentially still probe a longer timescale ) .
the similarity of 6 ghz , 33 ghz , co surface brightness , and the rsn / snr number density distributions lead us to view our 33 ghz measurement as indicative of the true size of the main disks of star formation and , presumably , gas and hot dust .
these morphologies are also consistent with the nuclear morphologies measured in mid - ir with the keck telescope @xcite . perhaps surprisingly , the two disks appear fairly similar in terms of profile , scale length , and observed flux .
the western nucleus appears hotter and more compact but the differences are small factors , not an order of magnitude .
the physical interpretation of such similarities is unclear .
possible explanations include a similarity in the progenitors , or some `` loss of memory '' during the process of funneling gas to the center of the galaxies during the ongoing interaction .
current best estimates of the dynamical mass per nucleus are @xmath114 m@xmath115 within @xmath69 pc of each nucleus @xcite .
these values are still uncertain , with @xmath116 m@xmath8 representing a likely lower limit in both nuclei @xcite .
the dynamical mass represents an upper limit on the gas content .
based on dynamical modeling and co imaging , estimated the gas content at @xmath117 and @xmath118 m@xmath8 for the eastern and western nucleus , but embedded in a larger gas disk with total mass @xmath119 m@xmath8 ( see also * ? ? ?
* ; * ? ? ?
these estimates mix dynamical modeling with observations of low - j co line ( up to j = 3@xmath442 ) measurements that are likely very optically thick in arp 220 .
@xcite provide an alternative , but unresolved , estimate by focusing on higher j co transitions and high critical density tracers ( e.g. , hcn ) to estimate a total molecular gas mass of @xmath120 for the entire system .
the difficulty with this estimate is apportioning this gas mass to the various components of the system .
we consider a conservative approach to be the following : we assume that half of the total molecular gas mass is equally distributed between the two nuclei and the other half in an outer disk ( e.g. , see * ? ? ?
* for evidence of an outer disk ) .
this implies @xmath121@xmath122 m@xmath8 of gas per nucleus from @xcite .
this remains in moderate tension with the dynamical masses because it would imply very high gas fractions , but given the mismatch in scales ( the dynamical masses are estimated on @xmath69 pc scales ) and uncertainties in modeling , a factor of @xmath123@xmath33 uncertainty seems plausible @xcite .
the areas that we measure for the arp 220 nuclei are stunningly small , especially when compared to the integrated properties of the system .
we adopt the literature gas mass of the nuclei as @xmath124 m@xmath8 , with the lower bound set by the @xcite values , the upper bound set by @xcite , and the best estimate consistent ( with modest tension ) with the latter .
we further expect half of the gas mass of each to be distributed within the half - light deconvolved , face - on , area ( @xmath92 ) of our radio images ( these trace star formation , so we implicitly assume that gas and star formation track one another within the system ) .
we thus compare @xmath125 m@xmath8 to our half - light areas to estimate average , nuclear , gas surface densities .
the deprojected , deconvolved gas surface densities are @xmath6 m@xmath8 pc@xmath9 ( east ) and @xmath7 ( west ) m@xmath8 pc@xmath9 .
these translate to an average , nuclear , total hydrogen column densities of @xmath126 cm@xmath9 ( east ) and @xmath127 cm@xmath9 ( west ) ( divide these numbers by 2 for h@xmath128 column densities ) .
these nuclear hydrogen columns are @xmath32@xmath129 orders of magnitude higher than those derived from x - ray observations ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , but they roughly agree with those derived from observations at @xmath130 m @xcite and @xmath131 m @xcite . the gas surface densities that we derive roughly resemble the maximum stellar surface density of @xmath132 m@xmath8 pc@xmath9 found in a compilation of literature data by @xcite .
given the large uncertainty in our mass estimate ( and the scatter in the @xcite compilation ) arp 220 appears consistent with producing such a maximal " stellar surface density system .
this is especially true when one considers that feedback and further evolution of the system may reduce the efficiency ( final fraction of gas converted to stars ) in the nuclei below unity ( a factor of @xmath133 would produce excellent agreement ) .
we do not know the thickness of the disks , but by adopting a spherical geometry we can calculate a lower limit to the h@xmath128 particle densities in the nuclei is larger than the areal correction . for simplicity , we adopt the correction appropriate for a gaussian , so that the mass within @xmath2 is @xmath134 of the total mass . ] .
this is @xmath135 @xmath16 ( east ) and @xmath136 @xmath16 ( west ) . for comparison ,
a typical milky way molecular cloud has a surface density @xmath69 m@xmath8 pc@xmath9 ( n(h)@xmath137 cm@xmath9 ) and average particle density @xmath138 .
in addition to faster free fall times , correspondingly more efficient star formation , and phenomenal opacity , potential implications of such high molecular gas densities would include the secondary production of cr electrons and confinement of cr electrons . by following the same approach
, we assume the infrared emission in arp 220 is coincident within our measured radio distribution and explore the implications .
conventionally , the infrared luminosity surface density , @xmath139 , is defined as the luminosity per unit area of the system .
we calculate @xmath139 by assuming that half of the infrared luminosity ( from 81000 @xmath72 m ) is generated within the deconvolved , face - on , half - light area ( @xmath140 ) , i.e , we calculate an average infrared luminosity surface density , within the half - light area , via @xmath141}{a_{50}}\big ) = \big(\frac{l_{50}}{a_{50}}\big)~.\ ] ] following our measurements above , we use @xmath142 to derive a _ total ( face on ) _ infrared luminosity surface density of @xmath143 .. ] if we further assume that the ratio of the fluxes between the east and west nuclei at 33 ghz ( @xmath5 1:1 ) holds at infrared wavelengths , then using the derived radio @xmath90 for the individual disks , we obtain @xmath144 and @xmath145 for the east and west nucleus , respectively . these values are more than an order of magnitude higher than those for the central 0.3 pc of the orion nebula complex and m 82 ( @xmath146 and @xmath147 , respectively @xcite ) , but are closer to those found in the brightest clusters within starburst galaxies ( @xmath148 @xcite )
. our estimated surface densities are consistent with @xcite , who estimated infrared luminosity surface densities of @xmath149 based on mid - ir keck observations and radio data from @xcite .
the deprojected sfr surface density ( defined as the sfr per unit area in the disk ) , @xmath150 , is a close corollary of the ir luminosity surface density .
we estimate this quantity within the deconvolved , face - on , @xmath92 for each nucleus at 33 ghz , using a 1:1 ratio between east and west , the radio luminosity to sfr conversion from table 8 in @xcite , and an electron temperature and non - thermal spectral index of 7500 k and 0.76 ( the spectral index in @xcite is defined with the opposite sign compared to our definition ) , respectively , from @xcite .
we obtain a @xmath150 of @xmath5 @xmath151 and @xmath152 @xmath153 within the half - light of the eastern and western nuclei , respectively ( divide these numbers by two to take into account both sides of the disks ) .
the total sfr calculated from @xmath26 , 180 @xmath154 , and from the total radio flux density at 33 ghz ( @xmath155 ) , 195 @xmath154 , differ by only @xmath5 10% , consistent with arp 220 lying on the ( 33 ghz ) radio - to - far infrared correlation ( and meaning that we would obtain essentially the same @xmath150 for either luminosity ) .
the radio sfr value differs by 20% from that derived from @xcite , which we consider to be within the uncertainties of such calculations .
@xmath139 tells us coarsely about the density of ir luminosity per unit area , but not necessarily the flux at the surface of the source , which may have important implications for feedback and depends on the detailed geometry of the system .
a spherical geometry provides a useful limit on the flux at the surface of the system . in this case
, @xmath156 will be the flux at the surface of a sphere of radius @xmath2 with the luminosity of arp 220 . using the measured radio sizes , @xmath157 for the entire system , @xmath158 for the eastern nucleus and @xmath159 for the western nucleus . a less extreme case , one that may well apply to arp 220 , is a two - sided disk . with one half of the luminosity emergent from each side , we have @xmath160 , twice the spherical case . as one would expect , these values are lower than the simple @xmath139 , but they also differ from one another , reinforcing the importance of geometry to the physics of the source .
yet another subtlety arises in the specific case where one wishes to calculate the flux through an area very close to , but just above one side of a disk .
this quantity is relevant to the often - discussed case of radiation pressure on dust ( see section [ sec : edd ] ) but because of projection effects it is not identical to any of the above quantities . in appendix [ app :
flux ] , we show that this one - sided flux perpendicular to the disk , which we call @xmath161 is equal to @xmath162 ( equation [ eq : ap7 ] ) in general . this value is further divided by an extra factor of two for the case of the two nuclei of arp 220 ( because @xmath163 combines the light from the two nuclei ) . from this calculation , we obtain @xmath164 for the entire system ( the same as for f@xmath165 , though not for the same reason ) , @xmath158 for the east nucleus and @xmath166 for the west nucleus . as discussed in appendix [ app : flux ] , these should be the most appropriate fluxes to consider when assessing the impact of pressure from radiation perpendicular to the disk .
the ratio of the flux densities between the east and west nuclei varies with frequency .
some other ratios for the east : west relation found in the literature include 1:4 at mid - ir @xcite , 1:3 at 18 cm ( if we consider only the contribution of the point sources from * ? ? ?
1:2 at sub - mm wavelengths @xcite . however , most of these observations do not offer high enough spatial resolution to truly isolate the contribution of each nucleus .
if we ignore this issue and we assume a ratio of 1:4 , and we use equation [ eq : ap7 ] , we obtain @xmath167 and @xmath168 . in every case , we obtain @xmath169 for the west nucleus , which is always the higher intensity nucleus . note that the same geometric issues discussed here raise a caveat regarding the luminosity of the source , which was derived under the assumption of isotropic emission .
an optically thick thin disk is not an isotropic emitter .
however , neither do our observations constrain the geometry of the infrared photosphere , which is the relevant surface for this calculation .
we discuss this issue in appendix [ app : flux ] . lacking information ,
we have assumed the isotropic luminosity throughout this paper , but note the uncertainty ( see also * ? ? ?
* ; * ? ? ?
note that high angular resolution infrared ( from 81000 @xmath72 m ) observations are needed in order to better constrain the true morphology of the ir photosphere of arp 220 .
@xcite and @xcite , argued that for optically thick , dense starburst galaxies , the critical feedback mechanism acting against gravitational collapse , and thus star formation , could be radiation pressure on dust .
although there is ongoing debate about whether or not radiation pressure on dust represents the dominant feedback mechanism in compact starbursts ( see * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* for further discussion ) , the maximal starburst model of @xcite and @xcite represents an interesting point of comparison for our present work . in figure [ fig3 ] , which closely follows figure 4 of @xcite , we present our new measurements for the flux near the surface of the source ( assuming a thin disk geometry , see section [ sec : ir ] and appendix [ app : flux ] ) in the context of literature observations and predictions by @xcite .
the literature observations show a sample of ulirgs with sizes based on 8.44 ghz radio maps by @xcite and luminosities from iras . following the approach presented in appendix [ app : flux ] , we calculate f@xmath170 following equation [ eq : ap7 ] assuming half of the total infrared luminosity to be enclosed within @xmath171 , where @xmath172 .
we used the deconvolved fwhm major axis , @xmath173 , from @xcite in order to account for inclination effects ( we only include resolved sources ) . following the discussion in the previous section , the fluxes for the points in figure [ fig3 ] differs by a factor of 8 compared to those in @xcite 4 that comes from the contribution of that region to the total infrared luminosity of the system @xcite . for the other systems having more than one component
, we used the relative contribution of each component to the integrated flux density observed at 8.44 ghz as a template for the relative contribution at infrared wavelengths . ] . for reasons discussed in appendix [ app : flux ] .
the error bars of our arp 220 values ( filled points in figure [ fig3 ] ) correspond to a combination of the uncertainties in @xmath2 , and uncertainty in the contribution of each nucleus to the ir luminosity , with the lower / upper limit assuming a ratio of 1:4 between east and west @xcite . to be conservative
, we also assume a 20% uncertainty in the assumption that half of the total ir luminosity is coming from @xmath140 .
the plotted values for the flux correspond to @xmath174 and @xmath175 . in this figure
, we show the `` eddington '' values for radiation pressure on dust as solid and dashed lines .
these represent an envelope for which radiation pressure on dust balances self - gravity . as a result
, no equilibrium star - forming system is expected to exist above this line , hence the `` eddington limit '' analogy .
the precise value of the limit depends on the size , gas fraction ( @xmath176 ) stellar velocity dispersion ( @xmath63 ) , rosseland mean opacity ( @xmath177 ) , and dust - to - gas ratio of the system , leading to the large spread in the model lines seen in the figure .
our measurements of arp 220 appear as solid points in figure [ fig3 ] .
there , the west nucleus of arp 220 appears among the highest brightness systems .
however , the west nucleus does not clearly stand out from the other ulirgs with respect to this value , partially because the more compact size means that the `` maximal '' value is larger for arp 220 than for larger systems .
overall , all the systems plotted in figure [ fig3 ] lie roughly around the eddington limit for a f@xmath178 = 0.1 and @xmath63 = 200 km s@xmath36 disk in @xcite , which is indicated by the blue solid line . if we adopt @xmath179 and @xmath180 200 km s@xmath36 ( e.g. , * ? ? ?
* ) , indicated by the green line in figure [ fig3 ] , then we calculate a conservative eddington limit of @xmath181 for the west nucleus and @xmath182 for the east nucleus .
refined measurements of the geometry , gas fraction , opacity , dust - to - gas ratio and kinematics are needed to specify the models more precisely . for most plausible assumed disk properties
, we can say that both arp 220 nuclei lie well below the @xcite eddington - limited starburst value , with the western nucleus being the brightest system among the local ulirgs .
if improved measurements demonstrate one or both nuclei to lie significantly above this value , one would need to consider luminosity sources other than star formation ( presumably an agn * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) or question the basic assumptions about geometry and equilibrium embedded in the model , but at present little such tension appears to exist .
another way to assess the role of radiation pressure in arp 220 is to assume that radiation pressure does represent the dominant force acting against gravity ( see appendix [ app : hydro ] ) and to calculate the required gas opacity , @xmath177 , of the system in order to be in hydrostatic equilibrium . by following equation [ eq : hydro3 ] , and using the derived values for the gas surface density and the flux for each nucleus ( see section [ sec : size ] and [ sec : ir ] ) ,
we find that @xmath183 and @xmath184 would be required for radiation pressure to balance gravity .
compared to the models from @xcite , which were used in the models from @xcite , these appear to be unrealistically high values for the gas opacity .
we interpret these high values as a reinforcement of our previous findings that the nuclei of arp 220 lie below the dusty eddington limit for @xmath176 @xmath5 1 and @xmath63 = 200 km s@xmath36 by a factor of @xmath5 10 , and then are not radiation pressure supported . the small sizes of the arp 220 nuclei also imply that optical depth effects will be important across the spectrum , even at wavelengths as long as sub - mm . by following the same approach described in section [ sec :
tb ] , we calculate the average rayleigh - jeans brightness temperatures that would be implied by combining the deconvolved half - light sizes of the nuclei with half of the 860 @xmath72 m continuum flux densities reported by @xcite from lower resolution observations . at this wavelength , we find implied average brightness temperatures of 19 k ( east ) and 76 k ( west ) , within the deconvolved half - light area calculated in this paper ( @xmath92 ) , without any accounting for optical depth ( which could change the size of the apparent emission ) . in the west nucleus , this approaches the apparent dust temperature of @xmath185 k ( see * ? ? ?
* for a more thorough idea of the dust temperature in arp 220 ) , so that optical depth must become important by this wavelength regime , with @xmath186 by 860 @xmath72 m in the western nucleus .
this neglects any more complicated geometric considerations ( e.g. , see the discussion of an inclined geometry in * ? ? ?
* ) , which are likely to make the situation even more confused .
optical depth effects appear less severe by @xmath187 mm , combining @xmath92 with half of the western nucleus dust emission from @xcite , implies an average @xmath188 45 k within the half - light area . as @xmath189 with @xmath190 the geometric mean between the deconvolved fwhm of the major and minor axis of the west nucleus .
that is , the brightness temperature is higher without accounting for inclination effects . ]
this modest mm optical depth is consistent with measurement of a spectral index steeper than @xmath191 by @xcite , implying somewhat optically thin emission .
even this simple calculation demonstrates that by sub - mm wavelengths optical depth effects can not be neglected , especially in the western nucleus . as a result
, we would expect high resolution but higher frequency observations , e.g. , at sub - mm wavelengths with alma , to observe a moderately optically thick `` photosphere '' around the galaxy and so recover a larger size than we find here ( e.g. , see the continuum observations by * ? ? ?
similarly , line observations at these wavelengths will need to consider the effects of a moderately optically thick sub - mm continuum in their interpretation .
further , we can estimate the optical depth at 33 ghz .
the brightness temperature calculated for the entire source at 33 ghz is @xmath5 500 k ( see table [ tbl-3 ] ) .
the thermal fraction at this frequency is @xmath192% , so that @xmath84 of the thermal emission is @xmath5 175 k. the thermal electron temperature ( @xmath193 ) can not exceed @xmath194 k because line cooling is high at such high temperatures .
then taking @xmath195 k and a measured @xmath196 k brightness , we estimate the average 33 ghz free - free opacity within the half - light radius to be @xmath197 . given that @xmath198
, we can calculate that @xmath199 at @xmath87 @xmath5 5 ghz increases to 0.028 and @xmath87 to 6 ghz if instead 55% of the 33 ghz emission is thermal ( see discussion in section [ sec : synch ] ) . ] . at 18 cm , @xmath200 , which might help explain the low vlba flux density from @xcite .
in addition , as @xcite note , the dust opacity is close to unity at 860 @xmath72 m .
thus , the arp 220 nuclei may be transparent _ only _ near the middle of the frequency range 5 ghz to 350 ghz . a case
can thus be made that the 33 ghz image presented here is the only existing image that is both optically thin and resolves the nuclei . in our observations ,
the western nucleus is more compact with a higher @xmath84 than the eastern nucleus .
however , consistent with previous vlbi observations , we observe no significant central excess in either arp 220 images or radial profiles ( figures [ fig1 ] and [ fig2 ] ) . @xcite discuss the possibility that one of three vlbi point sources showing a flat spectrum ( @xmath201 ) , could be an agn .
however , that is one of several possibilities that could explain the shape of their spectrum .
most of the 33 ghz emission that we observe comes from the compact , but still resolved , disks around the nuclei .
specifically , the nuclear beam of the west nucleus contains 20% of the total flux of the nucleus , while the other 80% arises from the more extended star forming regions ( table [ tbl-1 ] ) .
our measured @xmath2 contain similar information ( see table [ tbl-2 ] ) .
we can not rule out an agn in the western nucleus , but if one is present it does not make a dominant , point - like contribution to the overall 33 ghz emission on scales of @xmath202 pc .
similarly , arp 220 does not exceed the
eddington " value that might eliminate star formation as a viable power source ( see section [ sec : edd ] ) .
@xcite show that the sn rate and luminosity of arp 220 are broadly consistent with emission only generated by star formation , though uncertainty in the imf , sn rate , and sfr certainly would still allow an agn contribution .
no high brightness temperature radio core indicative of an agn is present .
however , given that most agn are radio - quiet and have weak core emission ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , the absence of a radio core does not rule out the presence of an agn . indeed ,
several studies at other wavelengths have presented evidence of a possible agn in arp 220 , but to date there is no clear evidence that the putative agn makes a significant contribution to the bolometric luminosity .
if our estimate of n(h ) @xmath5 10@xmath203 cm@xmath9 in section [ sec : size ] is correct , it would explain why evidence for agn in arp 220 has been so elusive . with such high column densities
any agn would be compton thick and undetectable by standard agn diagnostic tools .
we present new , high resolution vla observations of the nearest ulirg , arp 220 .
our 33 ghz observations measure the light distribution , which originates mostly from synchrotron emission , at a wavelength where optical depth effects are likely negligible .
we find exponential profiles with half - light radii of @xmath3 and @xmath4 pc for the eastern and western nucleus , respectively .
the distribution of 33 ghz radio emission matches the number density distribution of recent rsne / snrs very well .
this similarity may result from strong ( @xmath5 mg ) magnetic fields , which could yield cooling timescales for cosmic ray electrons that are short compared to the diffusion timescale .
adopting the measured 33 ghz sizes as characteristic of the star - forming disks , we derive implied surface densities , h column densities and volumetric gas densities that strikingly illustrate the extreme nature of the environment present in arp 220 . combining our size measurements with unresolved infrared measurements , we estimate total fluxes that , although very large , lie well below the conservative predicted values for the eddington - limited maximal starburst " , though this result is sensitive to our assumptions .
regardless , the implied luminosity surface brightness for the west nucleus of arp 220 is among the most extreme for any measured system . given the general uncertain evidence to date of a dominant agn in arp 220 , we conclude that the compact size and disk - like morphology clearly make arp 220 a prototypical example of the most extreme class of star - forming systems in the local universe .
we thank the anonymous referee for helpful comments that made this paper stronger .
we thank guillermo damke for his helpful input in the modeling process , norman murray for helpful discussions at an early stage of the project , phil arras for useful suggestions on appendix [ app : flux ] and shane davis for his input , especially in the development of appendix [ app : hydro ] , regarding radiation pressure and the eddington limit .
we thank the nrao / uva star formation group ( especially crystal brogan and kelsey johnson ) for repeated technical and scientific feedback .
we also thank nick scoville for detailed discussions of the astrometry of arp 220 .
, g.c.p . and l.b - m .
were supported by nsf grant ast 1109475 .
was also supported by fulbright and becas chile - conicyt .
this research made use of the nasa / ipac extragalactic database ( ned ) , which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration , and nasa s astrophysics data system bibliographic services .
the national radio astronomy observatory is a facility of the national science foundation operated under cooperative agreement by associated universities , inc .
consider a geometrically thin disk with radius @xmath204 viewed face on . then consider a small area parallel to the disk and at distance @xmath205 above the disk center .
the flux , @xmath206 , passing through the area will be @xmath207 where @xmath208 is the area subtended by an infinitesimal part of the disk .
the factor @xmath209 accounts for the orientation of the area relative to the patch of emitting disk under consideration with @xmath210 just as @xmath211 .
_ i _ is the specific intensity , which for the optically thick case , is just the source function of the disk and is the same for all lines of sight ( as the disk fills the beam ) . in the scenario where radiation pressure is important
, we consider that near the disk high optical depth is likely and proceed in the case of arp 220 at high viewing angles .
this factor cancels with the directional @xmath209 in equation [ eq : ap1 ] so that the optically thin case yields a different answer . ] .
the integral in equation [ eq : ap1 ] goes from 0 to @xmath212 in @xmath213 and 0 to @xmath214 in @xmath215 .
we will immediately change variables so that @xmath216 and @xmath217 .
thus , @xmath218 consider the limit where @xmath219 , i.e. , where the area the flux is passing through lies just above the disk .
then @xmath220 similar to the well known relation that the flux at the surface of a blackbody is @xmath221 will be different by a factor of two in the optically thin case . ] .
similarly , at large @xmath222 , as for an astronomical observation : @xmath223 which is nothing more than the integral of the intensity over the solid angle subtended by the disk .
the utility in this calculation is to relate @xmath224 back to the luminosity , which for an isotropic emitter is just @xmath225 ( see the last paragraph of this appendix for some caveats regarding this assumption )
. then @xmath226 so that , @xmath227 this @xmath228 is the flux through a surface near the disk and the departure from the perhaps expected @xmath229 is that we have included the @xmath209 term in the original setup to account for the projection of the incident intensity onto the unit area . when considering real observations cast in terms of the half - light area , @xmath230 ( and recall that the commonly used size at fwhm for a two dimensional gaussian is @xmath230 ) an additional factor enters from the fact that @xmath231 = @xmath232 for @xmath233 .
then in general : @xmath234 is the flux that should be used for considering radiation pressure near a large disk ( where large is defined so that @xmath219 can hold ) . in arp 220 ,
an additional factor of two comes into play if we assume the luminosity is split between the two nuclei with a ratio of 1:1 .
this decreases the relation to @xmath235 for this specific case .
note the stark difference , even for the general case , from the commonly adopted @xmath236 . in the case of an
_ optically thick _ disk , the assumption of isotropy is not valid and @xmath237 , instead @xmath238 .
further , the emission is not isotropically distributed , so that if the disk is inclined by an angle , @xmath239 , with respect to the line of sight , where @xmath240 is a face - on disk , then @xmath241 @xmath239 . in other words ,
the emission comes only from the two sides of the disks and an observer finds more flux when the disk is viewed face on ( because the constant intensity surface subtends more solid angle ) .
this is a substantial uncertainty for arp 220 ( e.g. , see * ? ? ?
* ) : it appears to host two inclined disks and shows good evidence for optical depth at ir wavelengths .
however , we are hesitant to impose any correction to the luminosity in the main analysis because we we do not know the true geometry of the infrared photosphere , which might very plausibly be more spherical and emit more isotropically than the nuclear disks picked out by our 33 ghz observations
. therefore throughout the main text we have used the conventional @xmath242 but we note this substantial uncertainty .
as a simple check on the plausibility of radiation pressure representing the main means of support , we consider vertical hydrostatic equilibrium in a simple gas disk .
we consider an infinite slab of surface density @xmath243 , so that the integrated weight of the column of gas at the midplane is @xmath244 ( here @xmath245 is the gravitational constant ) . to a coarse approximation
, radiation pressure can counteract this weight with a pressure set by the momentum flux of photons , @xmath246 ( where @xmath247 is the speed of light ) , multiplied by the total opacity of the gas column , @xmath248 , where @xmath177 is the cross section per unit gas mass . then : we can then solve for @xmath177 in terms of the other properties of the disk and the resulting @xmath177 expresses the required effective opacity in order for radiation pressure to balance the weight of the disk : if @xmath206 and @xmath243 are known , comparison of the require @xmath177 to realistic values represents a zeroth order check of whether radiation pressure represents a viable support mechanism for a system . as a close corollary , if @xmath177 is known or can be estimated , assessing the degree to which equation [ eq : hydro1 ] represents an inequality offers diagnostic of the importance of radiation pressure to the system . note that @xmath177 , as we have written it , will depend on the dust - to - gas ratio , grain properties , overall opacity , and temperature distribution . in our simplified thin disk geometry , it does not depend directly on the size of the system but all of these properties may vary substantially as a function of disk structure . as a first order approximation , we assume the @xcite model values serve as a template of typical values that would describe @xmath177 for the nuclei in arp 220 , but to our knowledge a thorough exploration of @xmath177 appropriate for the nuclear disks of merging galaxies remains lacking in the literature .
alonso - herrero , a. , rieke , g. h. , rieke , m. j. , & scoville , n. z. 2000 , , 532 , 845 anantharamaiah , k. r. , viallefond , f. , mohan , n. r. , goss , w. m. , & zhao , j. h. 2000 , , 537 , 613 armus , l. , charmandaris , v. , bernard - salas , j. , et al .
2007 , , 656 , 148 blundell , k. m. , & beasley , a. j. 1998 , , 299 , 165 clemens , m. s. , scaife , a. , vega , o. , & bressan , a. 2010 , , 405 , 887 clements , d. l. , mcdowell , j. c. , shaked , s. , et al .
2002 , , 581 , 974 condon , j. j. , huang , z .- p . ,
yin , q. f. & thuan , t. x. 1991 , , 378 , 65 condon , j. j. 1992 , , 30 , 575 davis , s. w. , jiang , y .- f . , stone , j. m. , & murray , n. 2014 , arxiv:1403.1874 daz - santos , t. , armus , l. , charmandaris , v. , et al .
2013 , , 774 , 68 downes , d. & eckart , a. 2007 , , 468 , 57 downes , d. , & solomon , p.m. 1998 , , 507 , 615 engel , h. , davies , r. i. , genzel , r. , et al .
2011 , , 729 , 58 genzel , r. , tacconi , l. j. , rigopoulou , d. , lutz , d. , & tecza , m. 2001 , , 563 , 527 gonzlez - alfonso , e. , fischer , j. , graci - carpio , j. , et al .
2012 , , 541 , a4 haas , m. , klaas , u. , mller , s. a. h. , chini , r. , & coulson , i. 2001 , , 367 , l9 hopkins , p. f. , murray , n. , quataert , e. , & thompson , t. a. 2010 , , 401 , l19 ilovaisky , s. a. , & lequeux , j. 1972 , , 20 , 347 imanishi , m. , & saito , y. 2014 , , 780 , 106 iwasawa , k. , sanders , d. b. , evans , a. s. , et al .
2005 , , 357 , 565 krumholz , m. r. , & thompson , t. a. 2013 , , 434 , 2329 leroy , a. k. , evans , a. s. , momjian , e. , et al .
2011 , , 739 , l25 lonsdale , c. j. , diamond , p. j. , thrall , h. , et al .
2006 , , 647 , 185 mazzarella , j. m. , soifer , b. t. , graham , j. r. , et al .
1992 , , 103 , 413 mcbride , j. , quataert , e. , heiles , c. , & bauermeister , a. 2014 , , 780 , 182 mcmullin , j. p. , waters , b. , schiebel , d. et al .
2007 , astronomical data analysis software and systems xvi , 376 , 127 meurer , g. r. , heckman , t. m. , lehnert , m. d. , leitherer , c. , & lowenthal , j. 1997 , , 114 , 54 mundell , c. g. , ferruit , p. , & pedlar , a. 2001 , , 560 , 168 murphy , e. j. , helou , g. , braun , r. , et al .
2006 , , 651 , l111 murphy , e. j. 2009 , , 706 , 482 murphy , e. j. , bremseth , j. , mason , b. s. , et al .
2012 , , 761 , 97 murphy , e. j. 2013 , , 777 , 58 norris , r. p. 1988
, , 230 , 345 papadopoulos , p. p. , van der werf , p. , xilouris , e. , isaak , k. g. , & gao , y. 2012 , , 751 , 10 parra , r. , conway , j. e. , diamond , p. j. , et al .
2007 , , 659 , 314 pooley , g. g. 1969 , , 144 , 101 rangwala , n. , maloney , p. r. , glenn , j. , et al .
2011 , , 743 , 94 robishaw , t. , quataert , e. , & heiles , c. 2008 , , 680 , 981 rodrguez - rico , c. a. , goss , w. m. , viallefond , f. , et al . 2005 , , 633 , 198 sakamoto , k. , scoville , n. z. , yun , m. s. , et al . 1999 , , 514 , 68 sakamoto , k. , wang , j. , wiedner , m. c. , et al .
2008 , , 684 , 957 sanders , d. b. , & mirabel , i. f. 1996 , , 34 , 749 sanders , d. b. , mazzarella , j. m. , kim , d .-
2003 , 126 , 1607 sault , r.j . , & wieringa , m.h .
1994 , , 108 , 585 scoville , n. z. , evans , a. s. , dinshaw , n. , et al .
1998 , , 492 , l107 scoville , n. 2003 , journal of korean astronomical society , 36 , 167 semenov , d. , henning , t. , helling , c. , ilgner , m. , & sedlmayr , e. 2003 , , 410 , 611 smith , h. e. , lonsdale , c. j. , lonsdale , c. j. , & diamond , p. j. 1998 , , 493 , l17 socrates , a. , davis , s. w. , & ramirez - ruiz , e. 2008 , , 687 , 202 socrates , a. , & sironi , l. 2013 , , 772 , l21 soifer , b. t. , neugebauer , g. , matthews , k. , et al . 1999 , , 513 , 207 soifer , b. t. , neugebauer , g. , matthews , k. , et al . 2000 , , 119 , 509 spoon , h.w.w . , marshall , j.a . , houck , j.r . , et al .
2007 , , 654 , l49 thompson , t.a .
, quataert , e. , & murray , n. 2005 , , 630 , 167 kellermann , k. i. , sramek , r. , schmidt , m. , shaffer , d. b. , & green , r. 1989 , , 98 , 1195 williams , p. k. g. , & bower , g. c. 2010 , , 710 , 1462 wilson , c. d. , rangwala , n. , glenn , j. , et al .
2014 , , 789 , l36 | we present new karl g. jansky very large array radio continuum images of the nuclei of arp 220 , the nearest ultra - luminous infrared galaxy .
these new images have both the angular resolution to study the detailed morphologies of the two nuclei that power the galaxy merger and sensitivity to a wide range of spatial scales . at 33 ghz
, we achieve a resolution of 0081 @xmath0 0063 ( @xmath1 ) and resolve the radio emission surrounding both nuclei .
we conclude from the decomposition of the radio spectral energy distribution that a majority of the 33 ghz emission is synchrotron radiation .
the spatial distributions of radio emission in both nuclei are well - described by exponential profiles .
these have deconvolved half - light radii ( @xmath2 ) of @xmath3 and @xmath4 pc for the eastern and western nuclei , respectively , and they match the number density profile of radio supernovae observed with very long baseline interferometry .
this similarity might be due to the fast cooling of cosmic rays electrons caused by the presence of a strong ( @xmath5 mg ) magnetic field in this system .
we estimate extremely high molecular gas surface densities of @xmath6 ( east ) and @xmath7 ( west ) m@xmath8 pc@xmath9 , corresponding to total hydrogen column densities of n@xmath10 = @xmath11 ( east ) and @xmath12 cm@xmath9 ( west ) .
the implied gas volume densities are similarly high , @xmath14 ( east ) and @xmath15 @xmath16 ( west ) .
we also estimate very high luminosity surface densities of @xmath17 ( east ) and @xmath18 , and star formation rate surface densities of @xmath19 ( east ) and @xmath20 .
these values , especially for the western nucleus are , to our knowledge , the highest luminosity surface densities and star formation rate surface densities measured for any star - forming system . despite these high values , the nuclei appear to lie below the dusty eddington limit in which radiation pressure is balanced only by self - gravity . the small measured sizes also imply that at wavelengths shorter than @xmath21 mm , dust absorption effects must play an important role in the observed light distribution while below 5 ghz free - free absorption contributes substantial opacity . according to these calculations ,
the nuclei of arp 220 are only transparent in the frequency range @xmath5 5 to 350 ghz .
our results offer no clear evidence that an active galactic nucleus dominates the emission from either nucleus at 33 ghz . |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.